Upload
others
View
1
Download
0
Embed Size (px)
Citation preview
RENDAKU DAMPENING AND PREFIXES
Mark IrwinYamagata University
•Over the years, a number of different factors have been put forward claiming to block, dampen, constrain, or otherwise restrict rendaku
• These include (in rough order of fame):
Motoori’s / Lyman’s Law (Motoori 1822, Lyman 1894)
the right-‐branching rule (Otsu 1980)
dvandva (copulaKve) compounding (Okumura 1955)
restricKons based on lexical stratum
•What has been significantly neglected throughout these long years of research is the effect which certain compound-‐iniKal elements appear to have on blocking or dampening rendaku
• To my knowledge, the subject has been touched upon to any extent in the literature only by Nakagawa (1966), Satō (1989) and Itō (2008)
•Nakagawa (1966: 314) notes that:
数詞の基本形ヒト・フタ・ミ...の系列に接する連濁可能語彙は、<お・み>等の接頭語のつく場合と同様、原則として連清形をとる。と言っても、フタ(二)にかぎり、<フタゴコロ・フタバ・フタゴ>のごとき連濁形をも有しているのは . . .
当然追求されねばならない。
Nakagawa here claims that there are two types of compound iniKal element which block rendaku
(1)That honorifics block rendaku is accepted by all scholars in the field: I will not discuss this further in this presentaKon
(2)Numerals are more problemaKc. Nakagawa notes examples where 二 does admit rendaku. He also
stresses that his claim applies only to ‘bare’ numerals (e.g. ヒト), not full ones (e.g. ヒトツ)
I will examine Nakagawa’s claims concerning numerals in more detail later
• Satō (1989: 257) notes that:
接頭語(御、真、片、唐など)は連濁を起こしにくい
• He then goes on to list examples such as 唐傘 & 片仮名
• Satō seems to be claiming that all prefixes dampen rendaku -‐ I will have more to say on this later
• Itō (2008: 87) reiterates Nakagawa’s (1966) claim for honorifics and numerals, but also cites 唐 as a
compound-‐iniKal element where ‘連濁は起こらない’
• It is not the case, however, that 唐 totally blocks
rendaku, as pointed out by Otsu (1980), and shown also by compounds such as 唐衣 and 唐櫃
• I shall have more to say on 唐 later in this
presentaKon
• The data presented in this presentaIon are extracted from the RENDAKU DATABASE, an ongoing project being carried out by the author
• The RENDAKU DATABASE currently has two sources of data (広辞苑第六版電子版 (2008) and 新和英大辞典電子増補版
(2008)) and includes all compounds found in these dicIonaries in which at least one of the non-‐final elements may undergo rendaku
• The RENDAKU DATABASE is currently about 75% complete (the number of entries stood at 22,974 as of 1 Nov. 2011) and the data presented here should thus be regarded as provisional
The Data
0
0.5
1.0
pure
nou
ns
deve
rbal
nou
ns
verb
s
adje
ctiv
es
de-a
djec
tiva
l nou
nsALL
0.685
0.853
0.660
0.171
0.715
0.773
324 258 21,5342,5645,64512,743
rendaku rate for native japanese elements (as of 1 Nov 2011)
0
0.5
1.0
pure
nou
ns
deve
rbal
nou
ns
adje
ctiv
es
de-a
djec
tiva
l nou
nsALL
0.755
0.853
0.660
0.715
0.773
324 258 18,9705,64512,743
rendaku rate for native japanese elements (excluding verbs)
element n rendaku rate
ひと 一 51 0.000
まい 毎 4 0.000
しら 白 39 0.224
から 唐 22 0.227
ふた 二 12 0.250
ま 真 37 0.378
かた 片 41 0.429
はつ 初 32 0.469
もろ 諸 10 0.500
•Nakagawa’s claim that bare numerals dampen or block rendaku appears to hold water: 一 never
admits rendaku, while 二 does so only 25% of the
Kme
• Only a very few examples of 三 and 四 as
compound-‐iniKal elements occur in the rendaku database: the data is insignificant
element n rendaku rate
まい 毎 4 0.000
しら 白 39 0.224
から 唐 22 0.227
ま 真 37 0.378
かた 片 41 0.429
はつ 初 32 0.469
もろ 諸 10 0.500
七人の侍
•How do we explain the dampening effect of the Seven Samurai?
• TradiKonal Japanese grammar makes a disKncKon between compounds (複合語) and derivaKves (派生語), the former being composed of ‘stem+stem’, the
la_er of ‘prefix+stem’ or ‘stem+suffix’
• Perhaps we find rendaku dampening only in derivaKves?
• But, how do we disKnguish a prefix from a compound-‐iniKal element? This is an issue fraught with problems
• One soluKon is to follow Hōjō (1973), who published a list of prefixes
• All the Seven Samurai appear in Hōjō’s list
• Unfortunately, many other prefixes appear in Hōjō’s list also, including the following:
element n rendaku rate
ま 間 15 1.000
くそ 屎 7 0.857
こ 小 109 0.842
ぎゃく 逆 6 0.833
て 手 108 0.833
お 小 22 0.773
りょう 両 13 0.769
はん 半 19 0.684
おお 大 102 0.652
浪人
•Although we can now claim the Seven Samurai are not just elements, but prefixes, why are they dampening rendaku?
•There is nothing in the phonology, phonotacIcs or accentual pa^erns of the Seven Samurai that sets them apart from other prefixes
•Neither phonology, phonotacIcs nor accent appear to be a trigger for dampening
Phonology, Phonotactics, Suprasegmentals
• Another possible explanaIon could be sought in the difference between bound and free prefixes: bound prefixes may only occur as prefixes, whilst free prefixes may also occur as independent words
• The Seven Samurai are evenly split in this respect: 4 are bound (毎-‐ 白-‐ 片-‐ 諸-‐) and 3 are free (唐 真 初)
• Among rōnin prefixes, we find both bound (大おお-‐ 両-‐ 小こ-‐ 小お-‐) and free prefixes (間 手 屎 逆)
• Boundness does not appear to be a trigger for dampening
Bound v. Free
• One of the Seven Samurai (白) is apophonic, appearing in
its covered (しら) rather than its exposed form (しろ)
• However, other apophonic nouns (雨あま .913, 木こ .846, 船ふな .722, 酒さか .700) exhibit normal rates of rendaku
• Apophony does not appear to be a trigger for dampening
Apophony
• Yet another explanaIon may lie in /Q/-‐generaIon
• Some Seven Samurai prefixes allow opIonal /Q/-‐variants (e.g. 真-‐: まふゆ~まっぴら), while others do
not
• However, such /Q/-‐variaIon can be found both among Seven Samurai (真-‐ 片-‐) and rōnin (大-‐)
• /Q/-‐generaIon does not appear to be a trigger for dampening
/Q/-generation
•More promisingly....
• Four of the Seven Samurai, while not numerals, are numerically related: 毎-‐ 片-‐ 初-‐ 諸-‐
•We have already seen that 一 blocks and 二
dampens rendaku
•However, a very few rōnin prefixes are also numerically related: 両-‐ 半-‐
Semantics
• An appeal to producIvity also holds more promise
• I asked a small number of naIve Japanese speakers, all university staff, to rank the Seven Samurai on a scale of 1-‐5 for producIvity (5 = most producIve)
• Averaged out: 真 4, 毎 3, 初 3, 片 2, 諸 2, 白 1, 唐 1
• While naIve Japanese speakers in the audience may not agree with these survey scores, most of you would agree that, with the excepIon of 真-‐, the
majority of the Seven Samurai are fairly unproducIve
• However, some rōnin prefixes are also unproducIve: 半 3, 屎3, 間2, 小お 1
Productivity
• Numbers and numerically-‐related prefixes either strongly dampen (二-‐ 片-‐ 初-‐ 諸-‐) or
block (一-‐ 毎-‐) rendaku
• UnproducIve prefixes (片-‐ 諸-‐ 白-‐ 唐-‐) tend
to dampen rendaku
Two Trends
REFERENCESLyman, Benjamin. 1894. Change from surd to sonant in Japanese compounds. Oriental Club of Philadelphia, 1-‐17.
Otsu, Yukio. 1980. Some aspects of rendaku in Japanese and related problems. Theore;cal issues in Japanese linguis;cs, ed. by Yukio Otsu and Ann Farmer, 207-‐227. MIT Working Papers in LinguisIcs.伊藤美津. 2008. 「連濁について」、 九州国際大学教養研究 15.2.
奥村三雄. 1955. 「連濁」. 国語学会(編)、「国語学事典」、東京堂(東京).
佐藤大和. 1989. 「複合語におけるアクセント規則と連濁規則」. 杉藤美代子(編)、「日本語の音声・音韻(上)」、明治書院(東京).
中川芳雄. 1966. 「連濁・連清(仮称)の 系譜」、国語国文 35.
北条正子. 1973. 「主要接辞・助数詞一覧」. 鈴木一彦・林巨樹(編)、「品詞別日本文法講座10」、 明治書院(東京).
本居宣長. 1822. 「古事記伝」、 前川六左衛門(江戸)他.
•真-‐ is neither numerically related nor
unproducIve but is worthy of further consideraIon
• Two types of 真-‐
真
•真1 conveys the noIon of an ‘ideal’
• It is most omen found with flora and fauna
•Due to semanIc drim, the compound in quesIon omen no longer represents an ideal, merely a parIcular species
•真1 exhibits standard rates of rendaku
•真鯛 真ボヤ 真鴨 真竹 真ガレイ 真鱈 etc.
•真心 真仮名
•真2 conveys the noIon of ‘directness’ or
‘centredness’
•真1 dampens rendaku
•真冬 真昼 真北 真帆 etc