Shirane's note on "mizenkei (ren'yôkei?!) of adjectives plus wa (ha)"
Klaus Pinte
Dear PMJS members,
Recently I stumbled on an intreaging note in Shirane’s ‘Classical Japanese: A Grammar’ (NY: Columbia UP, 2005). Under the heading ‘The Mizenkei of Adjectives plus WA (HA)’ in one of the ‘Advanced Study and Reference’ sections on p. 55 he mentiones the following:
When the ren’yôkei of a ku or a shiku adjective is followed by the conjunctive particle ba, indicating a hypothethical situation, the ba becomes nonvoiced (seion) wa (ha).
The sentence below is given to illustrate this:
(鯉を)切りぬべき人なくは、給べ。切らむ。Kiri-nu-beki hito NAKUWA, tabe. Kira-mu.
If there is no (naku-wa) person (hito) who can cut (kiri-nu-beki) (the carp), give it to me (tabe). I will cut it (kira-mu).
(Tsurezure, sec. 231, NKBT 30:274)
I have two questions in this respect:
1) Although a typo error can be involved, as far as I know the conjunctive particle ba is preceeded by either the mizenkei (hypothecal) or the izenkei (causal/temporal), but never before I heard of the combination with the ren’yôkei (naku then being the ren’yôkei of the ku adjective nashi). Can somebody advise me on this issue, or offer further references to sustain Shirane’s claim?
2) On p. 211 of Koide Hikari’s annotated translation of the Tsurezure (1981), the naku-ba/wa cited above is explained as the mizenkei + ba. However, am I mistaking when expecting the mizenkei of nashi not being naku, but nakara?
Looking forward to receiving your expert comments on this issue,
Drs. Klaus Pinte
Assistant lecturer Japanese language and culture
Ghent University Faculty of Arts and Philosophy
Blandijnberg 2, B-9000 Gent/Ghent (Belgium)
Tel. ++32 (09) 264 41 57, Fax. (0)9 264 41 94
guel...@waseda.jp
Kubota Jun in SNKBT and so on) read "naku wa", but there are also
editions (as Yasuraoka in his "Zenchuushaku") which read "nakuba".
The mizenkei of "nashi" is "naku", not "nakara(zu)", that works not
like "yoshi"/"yokara(zu)". So both mizenkei and renyoukei of "nashi"
have the same form; it is your interpretation to discrimate between
"nakuba" (mizenkei+ba) or "naku wa" (renyoukei+ha) because most texts
up to the 17th century don't use markers for dakuon.
Niels
>Dear PMJS members,
>
>
>
>Recently I stumbled on an intreaging note in Shirane’s ‘Classical Japanese: A Grammar’ (NY: Columbia UP, 2005). Under the heading ‘The Mizenkei of Adjectives plus WA (HA)’ in one of the ‘Advanced Study and Reference’ sections on p. 55 he mentiones the following:
>
>
>
>When the ren’yôkei of a ku or a shiku adjective is followed by the conjunctive particle ba, indicating a hypothethical situation, the ba becomes nonvoiced (seion) wa (ha).
>
>
>
>The sentence below is given to illustrate this:
>
> (鯉を)切りぬべき人なくは、給べ。切らむ。Kiri-nu-beki hito NAKUWA, tabe. Kira-mu.
>
> If there is no (naku-wa) person (hito) who can cut (kiri-nu-beki) (the carp), give it to me (tabe). I will cut it (kira-mu).
>
>(Tsurezure, sec. 231, NKBT 30:274)
>
>
>
>
>
>I have two questions in this respect:
>
>
>
>1) Although a typo error can be involved, as far as I know the conjunctive particle ba is preceeded by either the mizenkei (hypothecal) or the izenkei (causal/temporal), but never before I heard of the combination with the ren’yôkei (naku then being the ren’yôkei of the ku adjective nashi). Can somebody advise me on this issue, or offer further references to sustain Shirane’s claim?
Alexander Vovin
Please see some very brief notes below.
On Wed, Oct 21, 2009 at 8:51 PM, Klaus Pinte <Klaus...@ugent.be> wrote:
>
> Dear PMJS members,
>
>
>
> Recently I stumbled on an intreaging note in Shirane’s ‘Classical Japanese: A Grammar’ (NY: Columbia UP, 2005). Under the heading ‘The Mizenkei of Adjectives plus WA (HA)’ in one of the ‘Advanced Study and Reference’ sections on p. 55 he mentiones the following:
>
>
>
> When the ren’yôkei of a ku or a shiku adjective is followed by the conjunctive particle ba, indicating a hypothethical situation, the ba becomes nonvoiced (seion) wa (ha).
grave problems here. First, voiceless consonants in intervocalic
position can become voiced, but to the best of my knowledge there is
no example of a language where voiced consonants became voiceless in
the same position. Second, there is no morphophonological trigger for
voicing pa or fa after -ku. Third, and most significantly, /ba/ was
not really voiced throughout the Heian period and likely much later --
phonetically it was prenasalized voiced [mba], and the loss of both
prenasalization and voicing at the same time is practically
impossible.
A much simpler solution is at hand (which hopefully answers your
questions below). The adjectival -ku form is essentially a non-final
verbal predicate. So is the -te form of verbs. Now, if you can have
-te fa form to function as a conditional clause, what could possibly
detain you from forming a conditional clause with -ku fa? That is, on
synchronical level, you just add topic fa to either -ku or -te to form
a conditional clause. Such 'weird' practice is actually shared by all
languages immediately in the vicinity of Japanese -- Ainu, Korean, and
Manchu.
Hope it helps,
Sasha
--
============
Alexander Vovin
Professor of East Asian Languages and Interim Chair
Department of East Asian Languages and Literatures
University of Hawai'i at Manoa, USA
========================
iustitiam magni facite, infirmos protegite
I.L. Hanami
1. the pure conditional ば is suffixed to the mizenkei. Period. Therefore the 〜く form of adjectives (なく、よく、高く) is not only a renyoukei but a (de facto) mizenkei as well--but only when attaching a conjunction 接続助詞 like ば (なくば、高くば). This thinking seems to apply to ず as well. ず is not only a shūshikei, it is also a mizenkei (行かずは) and a renyōkei (行かずて).
2. ク活用 (as opposed to the カリ活用) does not have a mizenkei. Period.Therefore, grammatically speaking, a pure conditional ば is suffixed to mizenkei of verbs (急がば) and adjectival verbs 形容動詞 (静かならば), but when attached to a ク活用 or シク活用 adjective, ば is suffixed to renyōkei (若くば、惜しくば). Of course this logic could steer students to conclude that ば is suffixed to the shūshikei of the negative ず (行かずは).
The point is moot, I think, since grammar is a construct applied after the fact, and either argument can be proved or disproved, I think.
For me, however, when teaching students bungo, consistency is important--the fewer irregulars the better. It is simpler for students to grasp 〜く as a mizenkei and renyōkei rather than ば being suffixed to one conjugating form--mizenkei--when suffixed to verbs or one type of adjective (形容動詞) and to another form--renyōkei--when suffixed to another type of adjective (ク・シク活用).
My two cents.
Ichiro Leopold Hanami
Assistant Professor of Japanese Language and Literature
Department of East Asian Languages and Literatures
ジョージ・ワシントン大学
日本語・日本文学 助教授
花見一朗リオポルド
Alexander Vovin
Pls see below
2009/10/21 <guel...@waseda.jp>:
> Most editions of the Tsurezuregusa (Nishio Minoru in Iwanami bunko,
> Kubota Jun in SNKBT and so on) read "naku wa", but there are also
> editions (as Yasuraoka in his "Zenchuushaku") which read "nakuba".
What modern editions do frequently depends on their whim (:-).
>
> The mizenkei of "nashi" is "naku", not "nakara(zu)", that works not
> like "yoshi"/"yokara(zu)". So both mizenkei and renyoukei of "nashi"
> have the same form; it is your interpretation to discrimate between
> "nakuba" (mizenkei+ba) or "naku wa" (renyoukei+ha) because most texts
> up to the 17th century don't use markers for dakuon.
This is not completely true and largely depends on script or a type of
a text. Texts in man'yoogana (which does differentiate seion and
dakuon, especially Nihonshoki type of man'yoogana) have several
examples of くは, but not くば (most are in the Man'yooshuu, but one at
least is found in Nihonshoki kayoo). While literary texts in kana do
not use dakuon, some kunten texts do. Ditto for a some dictionaries.
Shoomono texts from 15th century mostly use dakuon, although not
always consistently. Picture in shoomono is actually interesting. Here
are some examples:
1) ヨクンバ
2) セスンバ
3) ナクハ
4) ナクワ
5) ヤスクワ
Note that only ンバ, but not バ occurs. This ンバ is quite clearly a
contraction from ニハ. In examples 3-5 there is a graphic variation
between ハ and ワ reflecting of course [wa] -- but
note that バ and ワ do not alternate.
So, what was phonetically [ku pa] in 8th century, became [ku wa] in
15th, clearly showing the same path of development as the topic
particle は[pa] > [wa], but not as the conditional ば [mba] > [ba].
Consequently we can safely conclude that it is the topic particle は in
the くは construction, not the conditional ば.
Sasha
>
> Niels
>
>>Dear PMJS members,
>>
>>
>>
>>Recently I stumbled on an intreaging note in Shirane’s ‘Classical Japanese: A Grammar’ (NY: Columbia UP, 2005). Under the heading ‘The Mizenkei of Adjectives plus WA (HA)’ in one of the ‘Advanced Study and Reference’ sections on p. 55 he mentiones the following:
>>
>>
>>
>>When the ren’yôkei of a ku or a shiku adjective is followed by the conjunctive particle ba, indicating a hypothethical situation, the ba becomes nonvoiced (seion) wa (ha).
>>
>>
>>
>>The sentence below is given to illustrate this:
>>
>> (鯉を)切りぬべき人なくは、給べ。切らむ。Kiri-nu-beki hito NAKUWA, tabe. Kira-mu.
>>
>> If there is no (naku-wa) person (hito) who can cut (kiri-nu-beki) (the carp), give it to me (tabe). I will cut it (kira-mu).
>>
>>(Tsurezure, sec. 231, NKBT 30:274)
>>
>>
>>
>>
>>
>>I have two questions in this respect:
>>
>>
>>
>>1) Although a typo error can be involved, as far as I know the conjunctive particle ba is preceeded by either the mizenkei (hypothecal) or the izenkei (causal/temporal), but never before I heard of the combination with the ren’yôkei (naku then being the ren’yôkei of the ku adjective nashi). Can somebody advise me on this issue, or offer further references to sustain Shirane’s claim?
>>
>>
>>
>>2) On p. 211 of Koide Hikari’s annotated translation of the Tsurezure (1981), the naku-ba/wa cited above is explained as the mizenkei + ba. However, am I mistaking when expecting the mizenkei of nashi not being naku, but nakara?
>>
>>
>>
>>Looking forward to receiving your expert comments on this issue,
>>
>>
>>
>>Drs. Klaus Pinte
>>
>>Assistant lecturer Japanese language and culture
>>
>>Ghent University Faculty of Arts and Philosophy
>>
>>Blandijnberg 2, B-9000 Gent/Ghent (Belgium)
>>
>>Tel. ++32 (09) 264 41 57, Fax. (0)9 264 41 94
>>
>>
>>
>>>
>
> >
>
--
============
Alexander Vovin
Professor of East Asian Languages and Interim Chair
Department of East Asian Languages and Literatures
Alexander Vovin
On Thu, Oct 22, 2009 at 1:28 AM, I.L. Hanami <han...@gwu.edu> wrote:
> The point is moot, I think, since grammar is a construct applied after the
> fact, and either argument can be proved or disproved, I think.
is any description of grammar that is.
> I cannot site any specific source and am speaking only from
> memory/experience, but I believe that there are two schools of thought on
> the suffixation rule of the pure conditional ば.
>
> 1. the pure conditional ば is suffixed to the mizenkei. Period. Therefore the
> 〜く form of adjectives (なく、よく、高く) is not only a renyoukei but a (de facto)
> mizenkei as well--but only when attaching a conjunction 接続助詞 like ば
> (なくば、高くば). This thinking seems to apply to ず as well. ず is not only a
> shūshikei, it is also a mizenkei (行かずは) and a renyōkei (行かずて).
of 国語学. If く and ず are not only 連用形 but also 未然形, why none of the 助動詞
following 未然形 such as む、まし、る・らる、す・さす etc. can follow く and ず? Why do
we always need a dummy ar- to be inserted?
>
> 2. ク活用 (as opposed to the カリ活用) does not have a mizenkei. Period.Therefore,
> grammatically speaking, a pure conditional ば is suffixed to mizenkei of
> verbs (急がば) and adjectival verbs 形容動詞 (静かならば), but when attached to a ク活用 or
> シク活用 adjective, ば is suffixed to renyōkei (若くば、惜しくば). Of course this logic
> could steer students to conclude that ば is suffixed to the shūshikei of the
> negative ず (行かずは).
eliminated on the basis of the fact that it is not conditional ば but
topic は that is added to ren'yookei く and ず (see my response to Niels
earlier today) to form a conditional clause.
The real problem is that mizenkei is a construct of description. Once
we leave 国語学の限界, and get rid of mizenkei which complicates enormously
not only the description, but also the acquisition of CJ, the problem
like that would not even arise.
Best wishes,
Sasha
============
Alexander Vovin
Professor of East Asian Languages and Interim Chair
Jamie Newhard
I am intrigued by this statement:
The real problem is that mizenkei is a construct of description. Once
we leave 国語学の限界, and get rid of mizenkei which complicates enormously
not only the description, but also the acquisition of CJ, the problem
like that would not even arise.
I wonder if you'd be willing to elaborate? How does the mizenkei interfere with the acquisition of CJ? What alternatives would be better?
Jamie Newhard
Richard Bowring
perhaps I may be allowed a general comment.
It is high time that everyone who teaches bungo for the purposes of
reading literature (and I include myself) also takes the time and
effort to read and digest the work of historical linguists, and by
that I mean (just to list those writing in English) Vovin, Whitman,
Frellesvig, Unger etc. It is for this reason that my heart sank when
I looked at Shirane's textbook and found little more than yet another
rehash of Japanese school grammar. It seemed to me to be a most
retrograde step and to add very little to what Morris, McCullough et
al had produced years before. This description of the language works
mechanically up to a point and we must admittedly all learn it in
order to understand the kind of annotation used by modern editors but
both the manacles of kana and the inability to conceive of different
patterns of segmentation still remain to confuse us at every turn. It
is this that gives us chimera such as meireikei plus ri, a 'final
form' apparently followed by a load of something else, and a 'not-
yetness' form (mizenkei) that is presented as the unlikely base for
negatives, suppositionals, passives and causatives. We can use it as
a mechanical tool but must do so with spades of salt.
Richard Bowring
Noel Hunt
I, too, am tantalized by your following remarks and would
dearly like to see them elaborated (I don't believe you broach
this topic in your Reference Grammar, which I have checked,
albeit cursorily):
On Sat, Oct 24, 2009 at 1:49 PM, Alexander Vovin <sasha...@gmail.com> wrote:
>
> The real problem is that mizenkei is a construct of description. Once
> we leave 国語学の限界, and get rid of mizenkei which complicates enormously
> not only the description, but also the acquisition of CJ, the problem
> like that would not even arise.
Regards,
Noel Hunt
Noel Hunt
is indeed adequately described in the Reference Grammar of Classical
Japanese Prose.
2009/10/25 Noel Hunt <noel...@gmail.com>:
Alexander Vovin
I will be happy to elaborate.
Let us first look at the 打ち消し連体形 form よまぬ. If one is operating exclusively on the basis of kana, there are only three ways to divide it:
1) よ|まぬ 2) よ|ま|ぬ 3) よま|ぬ
But there is no inherent requirement in any language that its morphemic structure would coincide with its syllabic structure. Here is where the problem with traditional approach starts: it tries to describe the grammar on the basis of assumption that all morphemic boundaries run between the syllables. And it cannot do otherwise, because how can you draw a boundary within ま and ぬ, if the only written representation you know is that of kana? Motoori and other kokugo gakusha of 18th c. who laid down the foundations of the traditional grammar could not possibly overstep this boundary.
Having said that, let me now turn to 未然形. As the name implies, it is 'imperfective base'. The first problem that we face is of the semantic nature, and was already mentioned by Richard in his excellent posting. It is conceivable that negative ず・ぬ and tentative（推量） む would be added to this 'base', but what is imperfective in the nature passives or causatives that according to traditional grammar also follow mizenkei?
Second, mizenkei is the only of the six bases that does not occur in isolation and in contrast to the other five has no morphosyntactic functions of its own. Thus, we are really dealing with entity that cannot be used by itself and has no defined function or meaning. Thus, we are entitled to arrive to suspicion that mizenkei is indeed a construct.
Now let me demonstrate on the basis of several examples that introducing mizenkei into description is contradicted by the evidence coming from the structure. Note that in the following I will write the consonant in the ha-gyoo line as /p/, in spite of the fact that I will deal with both examples from CJ (when it was f- in initial position, middle is more complicated) and Old Japanese (OJ) (when it was /p/ across the board) just in order to avoid possible confusion, and also because we are not doing phonetics here.
1) Traditional grammar teaches us that the tentative jodooshi む is added to mizenkei, thus
いはむ is analyzed as ipa-mu. Now let us look at the tentative retrospective in CJ (or the tentative past in OJ) which is けむ. We do know that OJ form was actually ke1mu, with koo-rui vowel /e1/ that goes back to the sequence *i1+a, thus ke1mu < ki1amu. We can now see two things that contradict traditional grammar: a) ki1 is the shuushikei form of past (not mizenkei!), and why we find -amu rather than simply -mu after it? Traditional grammar certainly tries to get out of this loop by announcing that there was a mizenkei ke1 for past ki1. Bear with me for a little bit longer, but note that it adds for memorization an extra mizenkei for ki1 (there is also another mizenkei -se- for -ki1; why there would be two different mizenkei in the first place?!)
2) Traditional grammar also teaches us that the nominalized form of verbs in くis added to mizenkei of yodan verbs いはく , thus ipa-ku and its variant form らく is added to shuushikei of all other verbs except kamiichidan, thus こふらく kopu-raku みらく mi-raku. This is clearly a mess. Not only we have two different forms without any explanation why in one case it is ku and in another raku, we have to memorize three different ways how to join them! And now comes this: with adjectives we have traditional nominalized form -ke1ku, which is derived from *ki1+aku by the same rule as ke1mu above. Traditional grammar claims that it is special mizenkei -ke1 for adjectives to which -ku is attached. Now we have four ways, and not only the description is completely ad hoc, but the acquisition is extremely complicated.
A much simpler and more elegant solution is to say that there is just one nominalizing suffix -aku that follows the rentaikei accross the board. Needless to say, this is exactly how nominalizing suffixes behave -- they normally follow attributive (rentaikei) forms of verbs.
Thus,
yodan
ipu-aku ==> ip-aku
kopuru-aku ==> kopur-aku
miru-aku ==> mir-aku
wosiki1-aku ==> wosike1ku
Note that in OJ two vowels normally cannot stand together, thus in the first three cases we have vowel delition and in the fourth case vowel contraction -- two simple rules much easier to remember than the mumbo-jumbo about ku/raku variation and four different ways to attach them depending on a verb class.
Note: in CJ the -aku form practically went out of use -- it mostly occurs only with ipu as ipaku. But in OJ the form was alive and kicking.
3) Traditional grammar teaches us also that the OJ honorific jodooshi す (merged morphophonetically with causative す・さす in CJ) follows mizenkei, thus, ex. きかす kika-su たたす tata-su あはす apa-su. All is well before we get to kamiichidan verbs, where mi1ru 'to see' and ki1ru 'wear' have honoric forms me1su and ke1su. Clearly, -su cannot follow mizenkei here, because mizenkei of mi1ru and ki1ru are mi1- and ki1- respectively. Now please recollect that vowel /e1/ goes back to *i1+a. With nidan verbs as well we get honorific forms nasu and ko2yasu from nu 'to sleep' and ko2yu 'to lie down' respectively. Clearly, -su cannot follow mizenkei here either, as mizenkei of nu is ne- and mizenkei of ko2yu is ko2yi-. Again, the simpler solution is to posit honorific suffix -as- that follows the verbal root directly:
ip- + -as- ==> ip-as-
ko2yi- + -as- ==> ko2y-as-
ne- + -as- ==> n-as-
mi1- + -as- ==> me1s-
ki1- + -as- ==> ke1s-
Note that in the first case nothing happens, because roots of yodan verbs end in a consonant. Thus, the sequence consonant (C) + vowel (V) perfectly fits the OJ phonotactics, and there is no change. In the second and third cases we have vowel deletion, and in the fourth and fifth vowel contraction -- exactly the same processes that we have already observed in (2) above.
I can go on with similar examples, but I hope I made the point clear. Once we get rid of mizenkei, we have to posit suffixes (not jodooshi) tentative -am-, negative -an-/-az-, honorific -as-, causative -sase-, passive -rare-, etc., that follow verbal roots (including roots of auxiliaries) or other suffixes directly. Let us look how the picture would look on the basis of -am-:
'say' ip-am-
'exist' ar-am-
'die' sin-am-
'long' kopi-m-
'sleep' ne-m-
'see' mi-m-
'kick' ke-m-
'come' ko-m-
'do' se-m-
One might argue that we have exchanged the complexity of verbal bases for the variation in suffix form -am- ~ -m-. But this variation is governed by a simple rule that we have already seen at work before: vowel is preserved after a consonant, but after a vowel there must be either contraction or deletion. In the case of -am- it is the vowel of the suffix which is deleted. If we could base a description of CJ verbal classes on the case of -am- ~ -m- alone, we could easily say that there are only two verbal classes in CJ: consonantal (yodan, rahen, and nahen) and vowel (all the rest) vs. nine in traditional CJ grammar. But it is not that easy, unfortunately, but the end result will be less than nine. Let us look at the traditional ren'yookei and rentaikei bases reanalized as suffixes
RY RT
'say' ip-i- ip-u
'exist' ar-i- ar-u
'die' sin-i- sin-uru
'long' kopi- < *kopi-i- kop-uru < *kopi-uru
'sleep' ne- < *ne-i- n-uru < *ne-uru
'see' mi- < *mi-i- mi-ru < *mi-uru
'kick' ke- < *ke-i- ke-ru < *ke-uru
'come' k-i- < *ko-i- k-uru < *ko-uru
'do' s-i- < *se-i- s-uru < *se-uru
Irregular forms are in bold (I am ahistorical here marking yodan and ra-hen forms as irregular, they are actually regular, but the explanation as to why will take us to much sidetracked). We can see that before attributive (rentaikei) -uru nidan, ka-hen, and sa-hen verbs lose the final vowel of their roots. Ichidan verbs, on the other hand, keep the vowel of their roots, but loose the vowel in the suffix. The ren'yookei marker -i drops after the last vowel of the root in nidan and ichidan verbs, but makes the root vowel drop in ka-hen and sa-hen. Comparing this picture with -am- above, we can actually arrive to much more simplified picture of Japanese verbal classes.
Class 1: Consonantal verbs
yodan, ra-hen, and na-hen verbs. Irregularities in all three classes are too miniscule to put them into three separate classes. Overall, all three behave as consonantal verbs.
Class 2: Weak vowel verbs
all nidans. Both kaminidan and shimonidan behave exactly in the same way, and both can lose their root final vowels in front of certain (but not all suffixes). The only difference between them is that kami roots end in -i, and shimo roots in -e -- clearly not sufficient basis to assign them to two different conjugations.
Class 3: Strong vowel verbs
all ichidans. Like class 2, the only difference between kami and shimo is the last vowel of the root. They never lose the final vowel of their roots, and always make the first vowel of a suffix drop. Kamiichidan shows quite a number of irregularities, especially in OJ (shimoichi does not exist in OJ), this is why Bjarke Frellesvig classifies these verbs as irregulars as well.
Class 4: Irregulars
ka-hen and sa-hen. These two show remarkable parallelism, so there is no sufficient basis to classify them separately.
Well, I think 4 is easier than 9 in addition to all other facts about 'mizenkei' given above.
Best,
Sasha
2009/10/23 Jamie Newhard <jamie....@gmail.com>:
Klaus Pinte
Your expert comments on my doubts about Shirane's [ren'yôkei + ba]
remark, are exceeding my initial expectations, and wish to thank all
of you, and especially prof. Vovin, for the very useful replies. I
will be more than happy to introduce this discussion to our students.
Sincerely,
Klaus.
Citeren Alexander Vovin <sasha...@gmail.com>:
> 'exist' ar-i- *ar-u*
> 'long' kopi- < *kopi-i- kop-uru < *kopi-uru
> 'sleep' ne- < *ne-i- n-uru < *ne-uru
> 'see' mi- < *mi-i- mi-ru < *mi-uru
> 'kick' ke- < *ke-i- ke-ru < *ke-uru
> 'do' *s-i-* < *se-i- s-uru < *se-uru
> ============
> Alexander Vovin
> Professor of East Asian Languages and Interim Chair
> Department of East Asian Languages and Literatures
> ========================
> iustitiam magni facite, infirmos protegite
>
> >
>
Klaus Pinte
Ghent University
Japanese Language and Culture
Blandijnberg 2
9000 Gent (Belgium)
Tel. +32 (0)9 264 4157
Fax +32 (0)9 264 4194
Jamie Newhard
What I find myself wondering now, looking back to Richard Bowring's comment, is what pedagogical applications there might be for a more linguistically correct version of CJ grammar. I imagine most of us do in fact teach bungo with full awareness that the traditional grammar is full of inadequacies and ambiguities-- that it is merely an expedient means for helping students learn to read fluently. Despite the inadequacies, though, it's hard for me to imagine a better way to _teach_ the language. As it is, a motivated student is ready to read on his or her own in little more than one semester. I imagine that having to start out with romaji representations and being unable to use dictionaries and cribs and such would complicate matters more than it simplified them. But I'd be happy to be disabused.
I in no way dispute the observation that bungo teachers ought to be much more familiar with the work of historical linguists. I'd be grateful if any list members who incorporate such work in their bungo teaching would share their experiences and strategies.
Jamie
Richard Bowring
I agree there is real problem here and that there is no alternative
but to teach the traditional framework. But there is no harm in
revealing to students as you go along that problems exist. I point
out some oddities at a very early stage (meireiki plus ri is a good
place to start because it is so obviously a contraction of the
renyokei plus ari) and it is salutary to show students the degree to
which the use of kana makes the idea that a verb stem might end in a
consonant literally unthinkable, but to head straight for Sasha's
explanation of why the mizenkei doesn't really exist would be
counterproductive. You obviously have to get the students properly
settled in one system before you upset the apple cart. I remember my
own reaction when I found out!
We talk of grammars being constructs, which is true, but I wonder if
bungo is not an extreme case, partly the result of using the kana
syllabary and partly due to the fact that the Japanese came to
grammar so late, and when they did it was European and hence largely
inapplicable. After all, sui generis traditions of analysing the
structure of one's own language along philosophical lines is a very
rare phenomenon indeed and China was no help in this regard. The
state of grammar in Japan, both modern and pre-modern, is surely
still in its infancy. Why this should still be so has much to do with
the sociology of scholarship in Japan.
Richard Bowring
William Wetherall
The fact that kana represent morae does not, IHO, significantly prevent a grasp of the segmental changes that take place within or between the morae, so to speak. While to some extent morae may be artifacts of the graphic system imposed on the language, the development of man'yogana suggests that the language was perceived as essentially structured by morae -- with the understanding that morae will be effected by various kinds of systematic changes. Hence the development of ways to "explain" these changes.
For many years, I taught such matters -- not to native speakers of other languages -- but to native speakers of Japanese -- in a survey course on J-E translation, using texts ranging from Kojiki and Man'yoshu to Murakami and Yoshimoto (which dates my "retirement" some ten years ago). We also did journalism, manga, poetry, and pop music lyrics. All Japanese texts were recited. All English translations were recited. The focus was purely on language -- oral-aural, i.e., linguistic qualities -- whether the original was kanbun, man'yogana, mixed kana-kanji, or a form of romaji.
The first expression we tackled was always 象は鼻が長い. Students would laugh but quickly get the idea of the difference between a "topic" and a "subject". At this time I would also teach them terms like "postposition" (they would know "preposition") and "marker" (only the odd chem major would quickly relate to this metaphor).
I would then segue to the following lead of a chapter in Matsumoto Seicho's 内海の輪.
離れは、本館から独立した小さな一戸建が五つほどあった。それぞれの家の間は、高い竹垣で囲んで人眼を遮るようにしていた。中は、十畳、六畳、三畳、それに風呂場と炊事場のようなのが付いていた。奥の間の十畳は海に面し、広い縁側の硝子戸ごしに島と街の灯が真向かいにあった。
This was their homework for the next class. They were to translate this paragraph structurally -- in a way that reflected its phrasing and metaphors. And they would have to read their translations aloud -- and critique each other on both the fidelity of their English versions, and the narrative (oral, linguistic) quality of their English.
But before the class that day broke up, they had to appreciate the Japanese text as a non-graphic, i.e., oral-aural, linguistic experience. I would ask one of the students to recite the paragraph. The other students had to close their eyes and just listen. Anyone who thought they could improve on the first reading would give it another try.
Using this single paragraph, the students would master the "structural" effects of "wa" and "ga" as used in this scene.
Then I would say, Okay, what about 象は鼻が短くはない。
One student might say Iwanai. Another would say Ieru. I would say, Focus on the second は. What is it doing? And while you're at it, consider the く in relation to it. For the next item on my agenda was the introduction of 雪国, which includes the phrase 遠くへ in this line:
娘は窓いっぱいに乗り出して、遠くへ呼ぶように、
「駅長さあん、駅長さあん」
Predictably, most students would associate く with "adverbs" as presumably 短い and 遠い were "adjectives" -- again, terms they had long ago learned in their years of English studies.
I would say, Okay, but what about the は and へ? What are they doing? Once in a while a student would see where I was heading and say, They are marking nouns or nominal phrases. And I would Exactly -- and wait for the chorus of Eee?s and Uso!s.
This being the "case" -- what, then, is the structural equivalent of 短くははい？ By this time, my students would understood that は sets off what it precedes, and whatever follows は is a statement about what has been set off.
So 短くは demands that one contemplate the condition of being short, while はい declares that such a condition does not exist. How one ends of rendering this in English is another problems. Most realistic solutions would be a form of "not short" -- which is not structurally faithful, but sometimes the demands of narrative elegance trump the desire for "literal" fidelity.
Someone will then say, So 遠くis a noun -- distance. Right. Then we examine phrases like 多くはない and 多くの人々 with the same focus on く and its relationship to the following elements. And we end the class with recitations of the dynamic lines that usher us through the long tunnel into the snow country.
While obviously the above material could not be used in an introductory course tailored to non-native readers who had not yet reached a fairly high level of read-on-cite proficiency -- much less in a course that had to be presented in romanization -- I would think that the methodology of illuminating the functions of conjugation forms (like く), in conjunction with whatever might follow them (including postpositions like は), would work for any group of students.
Practically speaking.
May the linguistic gods that be forgive me my somewhat fossilized grammatical sins.
Bill Wetherall
Richard Bowring さんは書きました:
William Wetherall
William Wetherall さんは書きました: