Ainu & Gilyak, Japanese & Korean

[Alexander Vovin]

First, I would like to voice my support in favor of Johanna Nichols
statement that Ainu and Gilyak are very different. Indeed, on the surface
they might look alike, and they even share some bizarre morphological
markers, like object prefix i-/e-, but the more you look inside the
history of these languages, the more fundamentally different they appear.
There have been contacts between Ainu and Gilyak for centuries, and, of
course, they influenced one another to a considerable extent. Late Prof.R.
Austerlitz have done a formidable research on Ainu-Gilyak contacts, but
unfortunately not everything from his scholarly heritage on the subject
has been published.
    Second, I would like to answer Johanna Nichols question of what
constitutes a nutshell evidence for the genetic relationship of Japanese
and Korean -- a quite right question to ask in the light of the present
discussion. Of course, the main evidence are the regular correspondences,
that can be established on the basis of lexical and morphological
comparisons, as this would be the only evidence acceptable for any other
language family as well. The lists of these correspondences are published
elsewhere, cf. e.g. Vovin 1994 in Diachronica XI.1:98, so I'll save myself
the labor of repeating them. I will divide the nutshell into two parts,
first dealing with morphological evidence (to please Stefan, of course,
:-), as morphological evidence is not readily available except some part
of it in S. Martin's
articles that I mentioned in one of the previous postings, which might be
difficult to digest for a person not familiar with the history of both
languages.
    As I mentioned earlier, I believe that both Japanese and Korean are
related to Tungusic, and more distantly to Mongolic and Turkic (but still
on the family, and not macrofamily level). Here, however, I will
discipline myself, and will exclude Turkic and Mongolic from further
discussion, but will add Tungusic, as I believe that Tungusic's
relationship to both Korean and Japanese is equidistant to that of
Japanese and Korean.
    Much remains to be done in this area, especially in the domain of
vowel correspondences and in reconstruction of consonant clusters in
medial position. Yet, the following, I believe can be interpreted only in
the way of genetic relationship.
    I will center on the evidence from Old Japanese language (OJ)(8th.
century) and Middle Korean (MK) (15th century), occasionally referring to
reconstructions of the corresponding Proto-Languages. As the earliest
known Tungusic language, Manchu is known only from the 17th century (still
earlier Jurchen materials are poorly deciphered and provide very little
data on morphology), I will also appeal to Prto-Tungusic and to various
other Tungusic languages, since Manchu does not always contain all
necessary data.
    Finally, the place where I access e-mail is away from my research
library, so quite a bit of what I'm going to provide is based on my memory
and I won't be able to provide chapter and verse for all data.
    NOTE 1: important addition to correspondences is provided in Vovin
1997
(forthcoming in Japanese/Korean linguistics v. 6): Proto-Japanese voiced
and voiceless opposition does not reflect voiceless and voiced of PA,
rather PA initial voiced are reflected as PJ words with low register and
PA initial voiceless (both aspirated and non-aspirated) as PJ words with
high register).
    NOTE 2: Both OJ and MK are given in Yale romanization, with the
exception of OJ otsu-rui /o/ that it typed as o2, as underlining is not
possible in my e-mail.
 
PART 1: MORPHOLOGY
 
VERBAL MORPHOLOGY
 
Both OJ and MK have very complex systems of verbal morphology, but a huge
chunk of it is of considerably later origin. Among the primary
morphological markers the following parallels can be suggested:
 
Old Japanese             Middle Korean             Tungusic
-an- < *ana              an(h)                  *ana, Manchu akU <*an-ku
negative                 negative               negative
 
accent class 2.5         -m                     ---
on some deverbal         deverbal noun
nouns, reflecting
earlier *-m
 
-ey < *(a)Ci< *(a)gi      -Gi, ki, hi etc.        -gi causative relic, cf.
transitivity flipper      transitivity flipper   e.g. Manchu al-gi- "let
                                                 know"
 
-uru                      -(V)l?                 Manchu -ra/-ro/-re
participle                irrealis participle    imperfect. participle
 
-u < *-wi < *-bi           OK -ta-wi <* -bi    -bi, Manchu, Nanai etc.-bi
final predicate marker     id                       id
 
-yi <*-i                        -e/-a                  ----
coordinative gerund         id
 
-myi < *-mi                 -mye                    -me
gerund of quality verbs     gerund                 gerund (Manchu -me,
 
 
-kyi < *-ki                MK -e/-a ?< *Ga/Ge      Manchu xa/xe/xo < *kV
retrospective              perfective              perfective
 
-te- < *-ta-Ci-            -te-/ta-                ----
perfective                 perfective
 
-(i)n-                     -(V)n                  -----
perfective                 realis participle
 
-ama-                      -----                   -Vme
tentative                                         tentative-volitive
 
-na-                       -no-                   -----
assertive                  "present tense"
 
-yi < *-i                         -i                     -----
nominalizer                 nominalizer
 
   This is the nutshell for verbal system. Of course, the above chart
represents my views, and I alone can be held responsible for it (in
particular, I do not accept Unger's views on all OJ consonantal verbs as
ending in -a). I am sure that every scholar in the J/K field may want to
add something to the chart above, or take out a couple things off it, but
overall I believe it represents the nutshell.
 
NOMINAL MORPHOLOGY AND PARTICLES ARE TO FOLLOW
 
Sincerely,
 
A.Vovin
 
On Wed, 26 Mar 1997, Johanna Nichols wrote:
 
> Miguel Carrasquer Vidal <mcv at PI.NET> wrote:
> >
> >One thing that caught my attention in the linguistic database from
> >Johanna Nichols' "Linguistic Diversity in Space and Time" was the
> >typological near-identity of Gilyak and Ainu.  Of all the typological
> >features listed, Ainu and Gilyak only differ in that Gilyak has
> >numerical classifiers (26 of them).  Is there any other evidence for
> >contacts between Gilyak and Ainu?
> >
>
> CAUTION -- That database should be used as a pointer to the descriptive
> publications, not as a full typological description in itself.  If Gilyak
> and Ainu exhibit typological near-identity, that just goes to show you how
> impoverished a typological description is compared to the real thing,
> because those two languages are very different.  (More generally, Gilyak is
> very different from any other language on earth.)
>
> Since the question of Japanese and Korean has come up on the list, I have a
> question for Alexander Vovin and/or others who have worked on these
> languages:  what, in a nutshell, is the evidence that Japanese and Korean
> are related?  I've read as much of the relevant literature as I could find,
> and the only support offered seems to be that if you assume they are
> related you can find sound correspondences and apply the comparative
> method.  But what is the evidence for assuming relatedness in the first
> place?  Thanks for any help anyone can give me.
>
> Johanna Nichols
> Slavic Languages, UC Berkeley
>