|
|
Roger Machin
Note on the transliteration of Japanese: My
transliteration follows the conventional kana
spellings as closely as possible. Thus, long vowels are
written double with the exception of /o:/, which is
rendered <ou> or <oo>
according to derivation. I write <si>,
<ti>, <tu> etc. in
preference to <shi>, <chi>,
<tsu>. The first element of the double
consonants conventionally rendered <kk>,
<ss>, <tt> etc. (ie small
っ) is written universally as <Q> (small capital), and
the syllabic nasal as <N> (small capital). Morae
in bold font are pronounced high. Words written without any
bold font are unaccented (have no fall in pitch), which is
to say the first mora is low, the second (if any) is high,
and any following word begins high.
A recent correspondent to the JAT list asked about the
etymology of the Japanese equivalent of St John's wort,
otogirisou. A respondent was able to provide the
story behind the characters 弟切草 conventionally used to
write the word. Although it may be tempting to accept
colourful explanations of this sort, I have my doubts and
would tentatively assign them to the realm of fantasy. It
is probably an example of what is called 'folk etymology',
a common enough phenomenon in any language. There is, for
instance, an English word sparrowgrass. No longer
in common use, it used to be an alternative name for
asparagus, from which it is derived by association
of sound. For all I know there may be a story or two
connected with the word, but in fact it is a folk etymology
which will have come into being as the result of people's
attempts to make some sense of a difficult and
foreign-sounding word.
I think it was Voltaire who said that in etymology the
consonants count for little, and the vowels for nothing at
all. I can see what he meant. It may be obvious that the
English word six, the Latin sex (as in
English sextuplet or sexagenarian) and
the German sechs all share the same origin,
although before you can spot even that, you have to
understand that <chs> in German is just a
way of representing the same sound as is spelt
<x> in English and Latin. It may be
somewhat more difficult to see the connection with
zes in Dutch, hex in Greek (cf. English
hexagon) or chwech in Welsh. Still on the
subject of numerals, I daresay it requires more than a
little imagination to accept that English five,
Dutch vijf, Latin quinque, Welsh
pump and Greek pente, however different
they may look or sound, are all essentially the same
word.
Etymology is an exact science, and like all exact sciences
it is governed by sets of rules which can be demonstrated
to be true. But ample care must be taken to guard against
false friends. Let me give an example. The verb 'to have'
in Latin is habere, and in German haben.
'Ah,' says the would-be etymologist, who is fully aware
that classical Latin is related to modern German as an
uncle is to a nephew (or rather as an aunt to a niece,
since languages tend to be feminine in gender), 'I can see
they're obviously from the same root.' Indeed, it looks
pretty much like a cut and dried case. But he would be
mistaken, for in this case the resemblance is purely
fortuitous. The Latin root which corresponds etymologically
to the German hab- is actually cap- (eg
the verb capere, 'to take', or English words such
as capture or, with modification of the vowel, the
-cep- in accept, both loanwords from
Latin). Meanwhile, Latin hab- corresponds to
German geb- (in geben, 'to give').
Incidentally, these examples provide us with another,
fairly simple correspondence, that between German /b/ and
English /v/ in haben, geben as against
have, give.
Coincidental resemblances outside the same family of
languages are by no means uncommon either. The word for
'eye' in modern Greek is mati, and the Malay or
Indonesian word mata means the same thing (Mata
Hari, the pseudonym of the Dutch dancer, courtesan and
First World War spy Margarete Gertrude Zelle is literally
'eye of the sun'). But few would venture to suggest on this
slender evidence that there is any connection between the
two languages, and indeed we have historical evidence to
show that the modern Greek word is derived by a process of
syncope, or shortening, from the classical Greek
ommation. Nearer to home, the Japanese
sou そう can often be translated
so in English and German, and what about the
English equivalent of the element -bone in
sebone 背骨?
The study of Japanese etymology is full of uncertainties
and fraught with pitfalls for the unwary. To begin with
there is the imprecise nature of the Japanese writing
system. Please do not misunderstand me: this is not a
criticism. In fact, I happen to believe that
kanamaziribuN 仮名混じり文 is as near perfect an
instrument for recording the Japanese language as you are
likely to get. But the fact remains that it is next to
useless when it comes to studying the phonological history
of words, and that is really what etymology is all
about.
Of course, there are substantial parts of the Japanese
lexicon where the immediate etymology is plain for all to
see: the vast numbers of loanwords and loan elements from
Chinese, and the newer imports from Portuguese, Spanish,
Dutch and English. But when it comes to the native
vocabulary, the picture remains very unclear. In short, we
still do not know much about the origin of Japanese.
Numerous theories have been proposed, some of which are
more plausible than others. Perceived connections with
Tibetan dialects and the Dravidian languages of southern
India can perhaps be dismissed as being based on too little
factual evidence. What resemblances there may appear to be
are probably coincidental. There have also been the
inevitable attempts to link Japanese to other languages of
disputed pedigree, of which a fair number exist in various
parts of the world. But there are two strong contestants:
the northern link and the southern link, and the truth of
the matter would seem to lie in a combination of the two.
Japan's geographical situation argues for a link with
Korean to the north, and beyond that to Mongolian and
related languages, while the string of islands stretching
like stepping-stones from Hatizyouzima
八丈島 south through Ogasawara 小笠原 to Saipan, Guam and
beyond cannot but argue for a connection with the languages
of Micronesia. Some scholars now suggest that Japanese has
developed from a creole, a mixture of these two
strains.
Be that as it may, and the link with Korean is easily
observed in the very similar grammatical structure of the
two languages, it remains true to say that much of the
native vocabulary of Japanese continues to present
something of a mystery. It is not possible to map out a
genealogy for words of the sort that we have for English
and many other languages. As we have already seen, one of
the problems lies in the nature of the Japanese script. In
any etymological argument we may discount kaNzi 漢字 immediately.
Interesting though it may be to speculate why a given
Chinese character was assigned to a given Japanese word,
this has nothing whatsoever to do with the origin of the
word in question. All that should concern us is how the
word is written in kana 仮名. Because in a
syllabary most symbols stand for a combination of sounds
(restricted to consonant plus vowel in Japanese, but other
scripts of this type permit of more complex combinations)
rather than a single sound as in an alphabet, there is
considerably less room for change or adaptation to mirror
changing pronunciations. We know, for instance, that the
/h/ sound in はひへほ was originally a /p/, which remains
to this day after っ, and accounts for why the
corresponding voiced sound (dakuon 濁音) is /b/.
What we cannot tell, because the kana symbols
remain the same, is when and how the change took place. In
this case, as luck would have it, external evidence is
provided by the 16th-century Jesuit missionaries, who use
the letter /f/ to represent the sound, suggesting that it
was similar to the allophone still heard in ふ.
Let me conclude this first article with an illustration of
how the Chinese character conventionally assigned to a word
may hide rather than reveal its etymology. Take the word
mizuumi. The fact that it is
conventionally written with the single character 湖 will
not blind us to its obvious derivation as a compound of
mizu 水 and umi
海. The literal meaning will be '(fresh)water sea',
although the existence of a blanket term covering both
'lake' and 'sea' is not uncommon in languages. We need only
look at the Old English word mere, which appears
in Windermere and the names of several other lakes in the
English Lake District, or the German See. It is
true that the German noun is treated as masculine when it
means 'lake', and feminine when it means 'sea', but it is
still one and the same word.
However, look a little closer and you will see that even
umi 海 can be divided into two
elements, u and mi. Neither has an
independent existence any longer, but the first of them can
also be found in usio 潮, unabara (or
unabara) 海原, and perhaps
uneri うねり
(uneru うねる). The second
element appears in a whole host of words which are all
connected in one way or another with water, and we shall
look at these and some other elements in detail in the
second article in this series.
There is a postscript to this article. I had just finished
writing the last sentence when it was time to take the dog
for his afternoon walk. As often we went down to the beach,
and I was sitting on the shingle looking at the waves and
watching some large ships go by on the horizon: the Straits
of Dover is said to be the busiest stretch of water in the
world. It was then that it suddenly occurred to me that the
element u might also figure in the names of
certain seaside place names in Japan, and I thought of
Ube 宇部 in
Yamaguti-keN 山口県, Uno (or Uno)
宇野 in Okayama-keN
岡山県, and Uzina 宇品, the port for
Hirosima 広島 as possible candidates.