Download the full version of the ebook at ebookfinal.

com

A Reference Grammar of Japanese Samuel Elmo Martin

https://ebookfinal.com/download/a-reference-grammar-of-
japanese-samuel-elmo-martin/

OR CLICK BUTTON

DOWNLOAD EBOOK

Download more ebook instantly today at https://ebookfinal.com

Instant digital products (PDF, ePub, MOBI) available
Download now and explore formats that suit you...

A Guide to Japanese Grammar A Japanese Approach to

Learning Japanese Grammar Tae Kim

https://ebookfinal.com/download/a-guide-to-japanese-grammar-a-
japanese-approach-to-learning-japanese-grammar-tae-kim/

ebookfinal.com

Tuttle Pocket Japanese Dictionary Japanese English English

Japanese Completely Revised and Updated Second Edition
Samuel E. Martin
https://ebookfinal.com/download/tuttle-pocket-japanese-dictionary-
japanese-english-english-japanese-completely-revised-and-updated-
second-edition-samuel-e-martin/
ebookfinal.com

A Reference Grammar of French R. E. Batchelor

https://ebookfinal.com/download/a-reference-grammar-of-french-r-e-
batchelor/

ebookfinal.com

A Reference Grammar of Spoken Tamil Harold F. Schiffman

https://ebookfinal.com/download/a-reference-grammar-of-spoken-tamil-
harold-f-schiffman/

ebookfinal.com
A Reference Grammar of Spanish 1st Edition R. E. Batchelor

https://ebookfinal.com/download/a-reference-grammar-of-spanish-1st-
edition-r-e-batchelor/

ebookfinal.com

A Tajik Persian Reference Grammar John R. Perry

https://ebookfinal.com/download/a-tajik-persian-reference-grammar-
john-r-perry/

ebookfinal.com

Modern Japanese Grammar A Practical Guide 1st Edition

Naomi Mcgloin

https://ebookfinal.com/download/modern-japanese-grammar-a-practical-
guide-1st-edition-naomi-mcgloin/

ebookfinal.com

Making Sense of Japanese Grammar A Clear Guide through

Common Problems Zelijko Cipris

https://ebookfinal.com/download/making-sense-of-japanese-grammar-a-
clear-guide-through-common-problems-zelijko-cipris/

ebookfinal.com

Oxford Japanese Grammar and Verbs Jonathan Bunt

https://ebookfinal.com/download/oxford-japanese-grammar-and-verbs-
jonathan-bunt/

ebookfinal.com
A REFERENCE GRAMMAR

OF JAPANESE

by Samuel E. Martin

CHARLES E. TUTTLE COMPANY

Rutland · Vermont Tokyo · Japan

REPRESENTATIVES

For the British Isles & Continental Europe

Simon & Schuster International Group. London
For Australasia
Bookwise International
1 Jeanes Street. Beverley. 5009. South Australia

t
First published in 1 975 by Yale University. New Haven and London

Published by the Charles E. Tuttle Company. Inc.

of Rutland. Vermont & Tokyo. Japan
with editorial offices at
Suido 1-chome. 2-6. Bunkyo-ku. Tokyo 11 2
Copyright© 1 987 by Charles E. Tuttle Co . Inc.

All rights reserved

Library of Congress Catalog Card No. 87-51 233

International Standard Book No. 0-8048-1 550-X

First Tuttle ed1t1on. 1 988

Third printing. 1991

Printed in Japan
 ACKNOWLEDGMENTS

I When a work takes over a decade to complete, the author enjoys the benefit of
ideas, material , counsel, and encouragement from many quarters. I have tried conscien­
tiously to acknowledge specific sources, published and unpublished, throughout this study.
Let me here mention a number of persons who have contributed in various ways to the
successful completion of the book. M uch that I have learned of the Japanese language has
been inspired by the work of my teacher the late Bernard Bloch and others of his students,
such as E lizabeth F. Gardner, Masa ko Yokoyama Lounsbury, and John J. Chew, Jr. I n
particu lar, I have benefited enormously from the studies of Eleanor Harz Jorden, whose
steady encouragement and criticisms are much appreciated. Without the devoted col labo­
ration of Hamako I to Chaplin, who has provided invaluable aid at every turn, neither this
nor other recent works on Japanese would have been possible ; some of the most intract­
able problems have found solutions through insights she has provided. Other members
of the Japa nese language faculty of Yale U niversity-Chie Chao, Kaoru Ohta, a nd Tatu­
hiko Tateyam::i-have also been of great help, and I have learned much from their kind
tutelage.
Parts of the research for this wor k, which was begun with assistance from the U .S.
Office of Education in 1961, took place while I was a visiting professor at the U niversity
of Washington in 1962-63 and at the U niversity of Hawaii in 1965-66 and 1969-70;
students and col leagues at both institutions have provided valuable information and criti­
cism. I n particular, I have incorporated specific material and ideas offered by Setsuko
Aihara , Norito F ujioka , R itva Sinikka Hayasaka, I rwin Howard, Thomas E. Huber, Shozo
Kurokawa, Leatrice Mirikitani, Zino Song, Hisami Konishi Springer, E miko Sugita,
Cecilia Y. Takaki, and Harvey Taylor. Hel pful suggestions on the manuscript were made
by Yale Graduate School students David Hughes, Patrick O 'Connor, S. R . Ram5ey, Karen
Sandness, and J. M. U nger.
Other scholars who have influenced the content of this book, by their writings or
through personal advice, include Anthony Alfonso, Mantaro Hashimoto, Shiro Hattori,
Teruo Hirayama, Yasuo l sami, Yukio l shigaki, Lewis S. Josephs, Haruhiko Kindaichi,
Susumu Kuno, S.-Y. Kuroda, James D. M cCawley, Akira Mikami, Roy A. Mil ler, Masaru
Nagano, Keiichiro O kutsu, Takeshi Shibata, I . F. Vardul, Sae Yamada, and Kanehiko
Yoshida.
Final ly, I would like to thank Marian Ash of Yale U niversity Press for hand ling
problems of publication and Elinor Clark Horne for her editorial and typographical help.
Th is book is affectionately dedicated to Norah and James, who were very patient.

June 1975 S.E.M.

This Tuttle edition has been prepared in order to make the book more widely
accessible to students of Japanese. Nearly a thousand corrections and additions have
been made to the first edition of 1975. I am grateful for the helpful reviews by N.
Akatsuka, C. Kitagawa, B. Lewin, G .B. Mathias, P . G. O'Neill, M. Shibatani, and G.E.
Wenck. I appreciate also the observations made by various other readers, and I look
forward to further comments from those who use the book, for your remarks can
improve future editions.
Jan uary 1988 S.E.M.
 CONT ENTS

0. Notational conventions : spelling, punctuation, accent, juncture. 15

0. 1 . Names and citations. 15
0.2. \Nords; sppces; hyphens. 16
0.3. Syl lables and moras. 17
0.4. Phrasing and juncture. 17
0.5. Accent i n words and phrases. 18
0 . 5 . 1 . I nherent word accents. 18
0.5.2. Variations in inherent word accents. 20
0.5.3. Variations due to vowel u nvoicing. 20
0. 5.4. \Nord accent withi n phrases. 21
0. 5.5. R egressive cancellations. 23
0.5.6. Other accent cancellations. 24
0.5.7. Accent shifts in certai n verb forms. 24
0.5.8. Accent i n dialect forms. 25
0.6. M iscel laneous conventions. 25
1 . Sentence construction : nuclear sentences ( predicates) and expanded sentences
(simplexes) ; sentence conversions. 28
Chart One. 32
Chart Two. 33
2. Predicate adjuncts. 34
2. 1 . Order of adjuncts. 35
2.2. Marking of adjuncts: noun postpositions ( "particles"). 38
2.2a. Unmarked adjuncts. 50
2.3. Focus of attention : backgrounding ( "subdui ng") and foregrounding
( "h ighl ighti ng ") of adjuncts. 52
2.3. 1 . Uses of wa. 60
2.3.2. Uses of mo . 66
2.3.3. Other particles of focus: made, sae, sura, dani ; si ka ; narade-wa. 70
2.3.4. Koso. 83
2.3.5. Some etymological speculations. 87
2.4.. R estrictives and quasi-restrictives. 90
2.4a. Yue [ n il. 1 37
2.5. The distributive (-zutu ) . 1 38
2.6. The comparator (yori) ; comparisons. 1 40
2.7. Col lectivizers; impl ied plurals. 143
2.8. Linkage of adjuncts; nom inal conjoini ng. 1 54
2.9. N ado ; nanka ; nante. 1 60
2. 1 0. Da ka, datta ka. 1 66
2. 1 1 . De mo ; d'atte. 1 67
2. 1 2. M itai (da/na/ni) . 1 72
7
8 Contents

3. Expansion constraints; noun subcategorization. 1 76

3. 1 . Classes of predicates ; subjects, objects, case valences; subjectless sentences. 1 83
3.2. D irect objects : transitivity, action, emotion, motion. 1 86
3.3. D irect objects of action verbs : kinds of affect. 1 88
3.4. I ndirect objects : beneficiaries; dative valences. 191
3.5. Existence, location, and possession ; quasi-possessives. 1 93
3.5a . Desiderative and quasi-desiderative predicates; cathexis. 1 98
3.6. Reciprocal valences. 201
3.7. Ablative, allative, and i nstrumental valences. 205
3. 7a. The set-opener ablative and the set-closer allative. 212
3.8. Locative constrai nts; stasis. 216
3.8a. Case valences with relexicalized nuclei . 22 1
3.9. Thematization. 224
3.9a. Subject delay. 234
3. 1 0. I dentification and propredication. 237
3. 1 0a. Epithematic identification. 253
3. 1 1 . M u ltiparous sentences; multiple surface-adjuncts. 254
3. 1 1 .1. M u ltiple surface-objects. 255
3. 1 1 .2. Multi ple surface-subjects; the "genitive" as a surface phenomenon,
genitivization; kinds of genitives. 256
3.1 1 .3. Thematization and focus of extruded genitives and of truncated
possessives. 267
3.1 2. I nherent verbal features : aspect and vol ition. 272
3. 1 3. Postnominal verbs and adjectives. 282
4. Voice conversions. 287
4.1 . Causatives. 292
4.2. Passives. 294
4.3. Passivized causatives. 299
4.4. Potentials. 300
4.4a. Spontaneous potentials and passives. 307
4.5. I ntransitives derived from passives. 307
4.6. Transitivity sets. 308
4. 7. Derived passive-potentials. 31 1 •
31 2
�
4.8. Derived pseudo-causatives.
4.9. Semantic skewing. 31 2

5. N uclear focus and restriction : spl it nuclei. 316

5. 1 . N uclear focus with wa and mo. 317
5.2. N uclear focus with sae and sika . 322
5.3. Restricting the nucleus with other elements. 323
5.4. Loose reference. 326
5.5. Further remarks on focus and restriction, etc. 329
6. Exaltation. 33 1
6. 1 . Subject exaltation. 336
Contents 9

6.2. Circumlocutions. 338

6.3. O bject exaltation. 342
6.4. Euphemisms. 345
6.5. Exaltation in dialects. 351
6.6. Donatory verbs. 352
7. Oesideratives. 355
7 . 1 . Desiderative adjuncts. 357
7.2. Desiderative verba lization. 358
7.3. Verbal ization of affective and cathectic adjectives and adjectival nouns. 359

8. N egation. 366
8. 1 . N egating the adjuncts. 366
8.2. Answeri ng negative questions. 368
8.3. M u ltiple negatives. 368
8.4. N egati ng the predicate. 370
8.5. The negative precopular nou n : -(a)zu (= the literary negative i nfinitive) . 376
8.6. Other negative forms borrowed from the l iterary language. 381
8.7. F u rther remarks on negation. 384
8.8. Lexical negatives. 388

9. Adverbial izations. 392

9.1 . The i nfi nitive. 392
9.1 . 1 . Verbal i nfinitive + ni + verb. 401
9.1 . 1 a. Verba l iteration with i nf initive + nL 407
9.1 . 1 b. I terated verbal i nfinitives. 408
9.1 .2. Verbal i nfinitive + --sidai (da) . 411
9.1 .3. The concurrent-concessive ( nagara) . 412
9.1 .4. Verbal i nfinitive + --tutu. 417
9.1 .5. Verbal i nfinit ive + --gati da. 418
9.1 .6. Verbal i nfinitive + --tate da. 419
9.1 . 7. Other adverbial nom i nals from verbal i nfinitives. 42 1
9.1 .7a. Verbal i nfinitive + �kko nai . 426
9. 1 . 7b. The l iterary negative infinitive -(a) zu + --zimai . 428
9. 1 .8. Facil itatives and propensives. 428
9.1 .9. The excessive (-sugiru) . 434
9.1 . 1 0. Verbal i nfi nitive + auxi liary verb. 438
9.1 . 1 1 . Conversions of adjectival and nominal i nfinitives. 455
( 1 /2) I ntransitive/transitive mutative. 455
(3) Subject-adverbial . 466
(4) N u clear-adverbia l (=der ived adverb). 467
(5) Evaluative. 468
(6) Transitive putative. 469
(7) I ntransitive putative. 469
9.1 . 1 2. Verb dropping. 470
9. 1 . 1 3. D irect adverbialization of nominals. 472
10 Contents

9.2. The gerund. 475

9.2.1. The predicated gerund . 491
9.2.1 a. Sentence-fi nal gerunds. 494
9.2.2. Gerund focus. 496
9.2.3. Verbal gerund + kara. 507
9.2.4. Verbal gerund + auxi l iary verb. 510
(1) -te [ i ] ru, -t [e] oru, -te [ i ] rassyaru, -t [e] o-ide n [ i ] naru. 514
(2) -te aru, -te gozaimasu . 523
(3) -te oku, -t' o ku . 529
(4) -te simau, -timau, -tyau. 533
(5-6) -te [ i ] ku ; -te kuru (mairu) . 536
(7) -te mfru ; -te go-ran. 541
(8) -te mieru. 544
(9) -te m iseru ; -te o-me ni kakeru . 545
(10) -te sumu ; -te su masu . 546
9.2.5. Auxi l iary constraints and sequences. 547
9.3. Cond itions: the 1-rovisional , the l iterary hypothetical, and the conditional ;
the I iterary concessive. 552
9.3.1. The provisional : -reba. 559
9.3.2. Conjoining parallel sentences with the provisional. 562
9.3.3. The conditional: -tara. 564
9.4. The representative: -tari . 566
9.5. Literary perfects : -i-tu/-turu ; -i-nu/-nuru, -i-ki/-si. 574
9.6. The l iterary perfect resu ltative : --eri/--eru ; [ -n i ]-k-eri/-eru . 575
9. 7. Phrasal postpositions. 577

10. Favors. 597

11. The perfect. 602

12. Tentatives and hortatives. 605

12.1. The tentative ; daroo. 605
12.2. The hortative ; the literary hortative . 610
12.3. The l iterary tentative. 615

13. Adnominalizations; typically adnominal and adverbial words. 616

13.1. Adnomi nal ization grammar. 619
13.1.1. Extruded adjuncts as epithemes. 624
13.1 .2. E pithematization of adju ncts extruded from converted sentences. 630
(1) Epithemes extruded from voice-converted sentences. 632
(2) Epithemes extruded from favors. 637
(3) E pithemes extruded from desiderative and quasi-desiderative
sentences. 639
(4) E pithemes extruded from facil itative-propensive sentences. 640
(5) E pithemes extruded from i ntransitivizi ng resultative sentences. 644
(6) E pithemes extruded from mutative, putative, and evaluative·sentences. 645
I
I
Contents

1 3.1 .3. Existential , locative, and possessive adnominalizations; epithemes

11

I
extruded from multiparous sentences. 646
1 3.1 .4. Extruded gen itives as epithemes. 651
1 3.1 .5. The adnominal ization of propredications a nd identifications. 652
1 3.1 .6. Subject marking and focus in adnominalized sentences. 659
1 3.2. Postadnomi nals. 664
1 3.2.1 . Mono. 725
1 3.2.2. Tokoro. 727
1 3.2.3. Toki. 730
1 3.2.4. y 00. 731
1 3.2.5. Wake. 733
1 3.2.6. Mama, manma. 735
1 3.2.7. Hazu. 736
1 3.2.8. Tumori. 738
1 3.2.9. Tame. 739
1 3.3. Tentative and hortative adnomi nals; - [ y ]oo mono/mon nara. 740
1 3.4. Possessive nominals. 741
1 3.5. R estricted nominals; ad nouns ( prenouns) .
= 742
1 3.5a. Adjectival nouns. 754
1 3.6. Quantity nominals. 766
1 3. 7. Adverbs. 782
1 3.7a . Conjunctions. 81 7
1 3.8. Precopular nouns. 821
1 3.9. Predicative and attributive forms borrowed from the l iterary language. 831
1 3.1 0. Pseudo adnomi nals; predicate-prone adjectives. 840

1 4. Nominal izations: general and specific. 841

1 4.1 . Kato. 841
1 4.1 .1 . Experiential nomi nal izations: koto ga aru/nai . 846
1 4.1 .2. Koto ni naru/suru. 849
1 4.1 .3. The circumlocutionary potentia l : koto ga dekiru. 850
1 4.1 .4. Verbal + koto ga + adject ive or adjectival noun. 851
1 4.2. No (da) . 851
1 4.2.1 . N [ol de. 856
1 4.2.2. No ni. 857
1 4.2.3. Post-appositional no. 860
1 4.2.4. Cleft sentences. 863
1 4.3. Verbal nouns. 869
1 4.4. Verbal-noun renominalizations : -tyuu da, -go(-) , -zen; -zumi. 881
1 4.5. I nfi n it ive-derived nouns. 883
14.6. D irect nominalizations. 889
1 4.6.1 . N i [wa/mo] tigai nai . 905
1 4.6.2. N i [ wa/mo] suglnai. 906
1 4.6.3. N i [wa/?mo] kimatte iru. 907
1 4.6.4. N i kagiru. 907
1 4.6.5. N i [ mo ] koto-kaite. 907
12 Contents

1 4.6.6. Verba l iteration with d irect nomi na lization + ni. 908

1 4.7. Abstract lexical nomi nal izations of adject ives and adjectival nouns (A-sa,
AN-sa ; A-mi, AN-m i ) . 909
1 4.8. Abstract lexica l and sentential nominalizations of verbs ( V-i-yoo,
V-i-kata) . 91 1

1 5. Sentence extensions.
1 5. 1 . I N e [e], na [a].
1 5.2. I Sa.
1 5.3. I Yo.
1 5.4. [ I ] Wa.
914
916
91 8
91 8
920
J
1 5.5. Zo, ze. 922
1 5.6. Ka. 923
1 5.6a. Ya. 932
1 5.7. Ka ne [e]/na [a] ; ( ka, da, -ta) f/e. 934
1 5.8. [ I ] Ka mo siren [ai). 935
1 5.9. Ka sira. 936
1 5. 1 0. K ke. 937
1 5. 1 1 . Tte; kate. 938
1 5. 1 2. Na; nakare. 942
1 5. 1 2a. -Mai, -mazi/-maziki ; --(a) masiu. 942
1 5. 1 2b. -sesi/beki, -beku ; -bekarazu/-bekarazaru ; -bei, -be [e], -be na. 943
1 5. 1 3. Koto and no ; nante. 944
1 5. 1 4. To mo. 946
1 5. 1 5. Ttara, tteba [a ] ; to, tto, to ka. 947
1 5. 1 6. [ I ] Zya nai [ ka/no ] , [ I ] zya arimasen ka ( zya nai desu ka) , [ I ] de wa
=

gozaimasen ka. 947

1 5. 1 7. oa' no. 949
1 5. 1 8. Yara. 95 1
1 5. 1 9. N ari ; nafrit [to (mo) ] . 953
1 5.20. Literary and dialect sentence-extensions. 957

1 6. Commands and requests. 959

1 6. 1 . The imperative. 959
1 6.2. Circumlocutions. 963
1 6.3. Negative commands and requests. 966
1 7. Conjunctionalizations. 968
1 7. 1 . Sentence + kara/mon [o] ; V-ru kara ni wa, V-ta kara [ n i ] wa. 968
1 7.2. Sentence + to. 974
1 7 .3. Sentence + si. 975
1 7.4. Sentence + ke [re]do, mono-no, mono [-o ] . 977
1 7.5. Sentence + ga. 979
1 7.6. Literary tentative + ga. 98 1
1 7. 7. Literary tentative + n i. 982
1 7.8. Sentence-final tentative + ni. 982
Contents 13

17.9. Sentence + nara [-ba ] . 983

18. H earsay-reporting. 984

19. The semblative : [ I ] -rasii. 986

20. The evidential: -soo/-ge da. 991

21. Quotations. 996

21.1. Special uses of quotations. 1003
21.2. Ttara, tteba. 1016
21.3. To [teL tte. 1017
21.4. Sentence + t �1 suru/naru : periphrastic putatives, subjective mutatives. 1018
21.5. Literary tentative + to. 1020
21.6. To ka; to yara. 1021
21.7. Adverbia l , i nterjectional, and mimetic quotations. 1022

22. Stylizations. 1026

22.1. Polite stylization. 1029
22.2. Honorific (hyperpo l ite, e legant) styl ization. 1035
22.3. Other stylizations: formal spoken, l iterary, formal written , and
epistolary styles. 1038

23. I nterjections; afterthoughts; m i nor sentences of various types. 1041

24. Connectors and sentence-openers; opening ellipsis. 1045

25. Apposition. 1048

26. Names, titles. 1055

27. I terative devices. 1060

28. E l l i ptical expressions. 1063

29. Demonstratives; deictics and anaphorics; pronouns. 1066

30. U sage constraints; agrammatisms; idioms; lexical compounds. 1081

B ibliography 1085

I ndex 1093
 A REFERENCE GRAMMAR OF JAPANESE

"
�-----

0 NOTAT I ONA L CONV ENT I ONS : SPELLI NG, PUN CTUATIO N ,

ACCENT, JUN CTURE

The Japanese i n this boo k is spelled i n roman letters ; the standard pronunciation
is easy to i nfer. A serious student of Japanese must be prepared to read the language in a
variety of forms, i ncluding the two kinds of romanization d istinguished here. I n the B i b­
liography and i n the spelli ng of proper names withi n E ng l ish sentences, we use the H ep­
burn romanization, which tel ls us to "pronounce the consonants as in E ngl ish and the
vowels as in I tal ian"; elsewhere the Japanese words and sentences are written l argely ac­
cordi ng to the conventions of the boo k A Manual of Japanese Writing, by Chaplin and
Martin, conventions which are almost identical with the system used in Spoken Japanese,
by B loch and Jorden, and d iffer from those used i n Beginning Japanese, by Jorden and
Chaplin, primari ly in neglecti ng to d isti ngu ish nasal from oral g and in writing "ei " for
most i nstances of what is usua. l l y pronounced "ee" in most parts of Japan . 1 In recent
loanwords the bi labial f- is represented by hw, as in hw irumu 'fi l m ' and hwan 'fan' and
byuhhwe 'buffet', except that tu is written hu, as in hurai-pan 'fryi ng pan ' and sutahhu
'staff'. In th is transcription "ti " represents the syl lable that H epburn writes as "chi "; for
the non-affricated pronunciation of such E nglish loanwords as that for 'D DT' we wil l
separate "t" and "d" from the vowel by a n apostrophe: d'ii-d 'ii-t' i i for what H epburn
writes as dldltl. ( I n A Manual of Japanese Writing this was written d e i i-d e i i-te f i , taking a
h i nt from the kana spell i ng , where a small "i" is i nserted after the syl lable for "te" or
"de", as if we were to write de i i-de i i-te i i.) The apostrophe is also used to i ndicate the be­
ginning of a syllable where doubt might arise (see §0.3). The word spel led iu 'says' is pro­
nounced /yuu/; our spel l i ng follows the native tradition, which is based on the u nderlying
form. I n explaining certai n etymologies the symbol 11 is placed i n front of a voiceless con­
sonant to show secondary voici ng ( n igori) : 11 p is pronou nced /b/, 11t /d/, 11k /g/, 11s /z/.
( But in the Middle Korean forms cited on p. 89, the symbol 11 represents the accent of
wh ich modern Seoul length is a reflex.)
0.1. NAMES A ND CI TA TIONS

I n the Bibliography and within Japanese sentences a Japanese fam i ly name d irectly pre­
cedes the personal name and title ( if these are present) , but in the E ng l ish translat ions you
wi l l find the order reversed accordi ng to the foreign custom. This means that the transla­
tion of a sentence tal king about a man referred to as Tukisi ma Syunziroo wi l l cal l him
Shunjiro Tsu kishima and if he were to turn u p as the author of a book l isted i n the B i b­
l iography you wou ld find h i m listed as Tsukish i ma Shunjiro.
Certain authors and wor ks that are quite often cited will be referred to by abbrevia­
tions, for which the ful l forms can be found i n the B i bliography. This accounts for the
frequent mention of " K K K " for Kokuritsu Kokugo Kenkyujo (National Language R esearch
I nstitute) , of "K" for Kindaichi H aruh i ko, of "H " for H irayama Teruo's Zenkoku akusento
jiten, etc. Many of the example sentences were cul led from weekly magazi nes, such as
Sande-Mainichi ( "SM ") and especially Shukan-Asahi ( "SA ") , and some are from monthl ies

1 . B ut the pronunciation ei is sti l l heard in the Ryukyus, Kyush u , southern Shikoku, parts of the Kii
peni nsu la, and the lzu I slands ( H 1 968 .95 ) .
15

�-------·-
16 § 0. Notational Conventions

such as Chuo-KOron ("C K ") .2 Others are from the works of modern novelists such as
Funabashi Sei ich i ("F n") or Kubota Mantaro ( " Kb") ; but many examples from fiction are
ta ken from secondary sources, such as l sh igaki Yu kio ( " l g ") , Yoshida Kanehi ko ( "Y"),
and I. F . Vardul ( "V ") , and are so cited . A good many examples were transcri bed from
tape record i ngs of radio and television ; these are marked "R ". U nmarked examples are
mostly the resu lt of el icitation from native speakers, but a few were ta ken from written
sources that I now find d ifficult to identify.
In maki ng the F.nglish translations, I have tried to take i nto accou nt the surrounding
context of examples from pri mary sources; this accounts for the translation of titles such
as sensei by 'you ' in more than one sentence. But I have not gone to the trouble of check­
ing the original context of examples quoted from secondary sources, since an i naccurate
reconstruction of u nexpressed elements wi l l seldom affect the grammatica l po i nt under
discussion. The purpose of the E ngl ish translations is simply to help the reader find mean­
ing i n the Japanese examples; I apo logize to those authors and translators who may be dis­
tressed at my clumsy treatment of fam i l iar sentences.
0.2. WORDS; SPA CES; H YPHENS

When a Japanese writes a sentence he leaves no spaces between the words. I f pressed to
do so, he wi l l insert a space only where it is possible to hesitate; such poi nts of hesitation
represent the surface manifestation of boundaries (ca l led "junctures" by l ingu i sts) which
separate short phrases within the sentence. In our transcribed sentences, however, you
wi l l find the spaces are placed to show a generous d iv ision i nto WO R DS as defined partly
by ru les of accent and partly by versati l ity of d istribution. R u les of accent placement are
d iscussed just below and a lso later in the book, where you will find grammatical criteria
for the various word classes that are needed to descri be the sentences.
H yphens ca l l attention to the i nternal structure of a word for any of a number of rea­
sons. When cited in isolation the verb ir-u 'needs' has a hyphen to remind you that the in­
fin itive is iri and the negative iranai ; otherwise you m ight confuse it with iru 'stays', which
has the i nfinitive i and the negative i nai. The lack of a hyphen in keizai sai ken 'to recon­
struct the economy' reflects the underlying juncture that separates the expression i nto a
two-word phrase ; the presence of a hyphen i n keizai-seikatu 'economic l ife' tel ls us that
this is a compound noun made up of the nouns keizai 'economics' and sei katu 'l ife', and
in keizai-zin 'busi ness man' and keizai-teki 'economical' we are reminded that the final ele­
ments are suffixes. Although I have tried to use the hyphen with some measure of con­
sistency, especial ly within lists, do not be dismayed to find the same word written some­
times sol id, sometimes with a hyphen, or even-under special ci rcumstances-as two words.
Withi n a word the morpheme d ivisions are not marked. When you see (or, for that mat­
ter, hear) siki there is no obvious way to tel l whether you are confronted with a one-mor­
pheme word such as that for 'ceremony' (written with a single Chinese character) or a two­
morpheme word such as that for 'the four seasons of the year' (written with two Chinese
characters) . But in certa in compounds the hyphen proves helpfu l : si ka-i means 'dentist',
being a compound of the two-morpheme noun sika 'dentistry' and the one-morpheme

2. The reference "(SA 2653.46c) " is to be read 'column c (third from top or third from right) on
page 46 of issue No. 2653 of Shukan Asahi [= the issue of 2 1 November 1 969] ' .
§ 0.4. Ph rasing and Juncture 17

abbreviation o f t h e two-morpheme noun isya 'doctor' ; sikai means 'city counci l ', a two­
morpheme nou n . The two words sound, of course, exactly the same; you can not hear a
hyphen.
0.3. S YL LABLES AND MORAS

I n reciting poetry or spe l l ing out the sound of a word , a Japanese wi l l a l low an equal
amount of time for each vowel, so that a long vowel (here written double) cou nts as two
timing un its or MO RAS : Tookyoo (Tokyo) is pronou nced as four moras to-o-kyo-o, keizai
'economics' is pronounced ke-e-za-i. When a consonant occurs without a fol lowing vowel,
it is treated as a separate mora : sinbun 'newspaper' is pronounced si-n-bu-n, and gakkoo
'school' is pronounced ga-k-ko-o. The Japanese term for mora is onsetu and th is is often
loosely translated as "syl lable". I n speech the Japanese may run two moras together to
make a single syl lable, so that the difference i n length between Tookyoo (Tokyo) and
Kyooto ( Kyoto) owes only to the extra mora, both words consisting of two syl lables. We
can think of two-mora syl lables such as those heard i n Too-kyoo, kei-zai, sin-bun, and
gak-koo as H EAVY (or LON G ) syl lables in contrast with the LIGHT (or SHO RT) syl lable
that consists of a sing le mora, such as ko or kyo or o. Foreign loanwords and mimetic ex­
pressions even contai n E XTRA-H EAVY syl lables made u p of three moras: hoon 'horn',
booi 'boy, bel lboy', sf i n 'scene', siin-to 'very qu ietly'. The accent never lands on the added
mora (or moras) of a heavy syl lable. When you see tooi desu 'it is far' you know that the
syl lable structure is too-f-de-su. The adjective ooi 'is much, are many' is pronou nced either
as three syl lables o-o-i or as a heavy syl lable fol lowed by a l ight syl lable oo-i ; ooi is the
common Tokyo version, but ooi is more widely heard e lsewhere.3 For the verb meani ng
'covers' Tokyo has both the historically expected atonic version oou (two syl lables oo-u)
and a tonic variant oou (three syl lables o-o-u) . I n general , our notation indicates syl lable
structure only by i mpl ication, but you will notice that an apostrophe marks the beginning
of a syl lable in certa in cases where doubt might arise : tan 'i 'un it' has three moras, tani 'val­
ley' has two ; ko'oo 'response' consists of a short syl lable fol lowed by a long, koo'o ' l i kes
and d islikes' consists of a long syl lable fol lowed by a short. (But in paat'ii 'party' the apos­
trophe shows that the t is not to be affricated ; the word conta ins two long syl lables paa-t'ii.)
0.4. PHRASING AND JUNCTURE

The transcriptions in th is book provide a good dea l of i nformation about accent and
phrasing. I f anything, you will probably feel overwhelmed with more i nformation than you
want, though you may eventual ly find reasons to feel grateful for certai n of the notations
wh ich seem irritati ng at first. Freely ignore as much of the notation as you see fit. The de­
vices are intended to be hel pfu l for those interested, not to browbeat those seeking other
information ; each reader wi l l have his own needs.
3. Hamako Chap l i n freely varies the pronunciation of the infinitive of ook i i 'is big' between oo kiku
(four syl lables) and o okiku ( three syl lables) , but only the latter is recognized by the dictio naries. All
sources give only o o siku as the infin itive of oos l i 'is brave ' ; there is no * oosiku. K impl ies that ook i i
and oos i i d iffer in number of syl lables, but H implies they are the same. Etymologically , the former
word is built on a morpheme reduced from two syl lables 0 [ p ]o- ( identical with the base of oo i) ' while
the latter is a redupl ication of a one-syllable morpheme o- 'male'.

�------
18 § 0 . Notational Conventions

Every spoken phrase of Japanese displays a tune that is chosen out of a l i mited stock of
arrangements of stretches of lower and h igher pitches. The phrases are separated by boun­
daries we cal l "ju nctures". A major juncture ( marked by the double bar II ) tel ls us that the
phrases on either side are pronounced rather i ndependently of each other, with ful l value
for each accent phrase. A minor boundary ( marked by the single bar I) warns us that the
pattern of the later phrase is somewhat altered by the precedi ng phrase; for example, if
there is a fal l of pitch it begins from a lower plateau. Depend ing on speed and emphasis,
you may hear the same sentence spoken with somewhat d ifferent phrasings. I n slow and
del i berate speech the sentence wi l l be broken i nto shorter phrases; in hasty speech phrases
wi l l be run together, with the major ju nctures reduced to minor and the minor junctures
often droppi ng altogether. Some of our examples, especial ly those ta ken from recordi ngs,
are transcribed with typical junctures, but most are not. An appropriate phrasi ng can usu­
ally be i nferred from the other i nformation given.
Our punctuation conventions fol low fam i l iar Eng l ish patterns. The capitalization of
proper names and of the first word in a sentence conveys no phonetic i nformation; com­
mas and semicolons are used for logical rather than phonetic purposes, though they will
often coincide with a major juncture. The original punctuation is usual ly retai ned for ex­
amples cited from written texts, even when it raises questions; you wil l f i nd side comments
on this from time to time. The sentence-final period is deferred u nt i l the end of the E ng­
l ish translation, which is set off by inverted commas (' .. .'.) , but a final question mark is in­
cluded withi n the E nglish translation ( ' . ?') , and also at the end of the Japanese sentence
. .

whenever the origi nal text contai ned the mark. The triple dot ( ... ) is used to show omis­
sions in a citation and at the end of a l ist to show that the l ist is not exhaustive . To save
space the tri ple dot is often omitted, however, when it can easily be supplied by the
reader.
1'

0.5. A CCENT IN WORDS A ND PHRASES

I n order to appreciate the devices marking accent and juncture, it is necessary to take a
brief look at certai n facts about the way Japanese is spo ken. R eaders who find themselves
puzzled by notations may wish to consu lt the fol lowing i nformation for gu idance.
0.5. 1. Inherent word accents.
I n addition to the consonants and vowels that ma ke up its moras and syllables, each
Japanese word has an inherent accent pattern, a tune that is appropriate to it i n certai n
critica l contexts. Those words which are TON I C are characterized b y an i nherent fal l of
pitch ; the poi nt of fal l is here marked with an acute accent ( ' ) over the vowel. P ROTO­
TON IC words fa l l to a lower pitch right after the first syl lable: Nara 'Nara', T l ba 'Chi ba',
A kasi 'Akashi ', Meguro 'Meguro', otukisama 'the moon', A mano-hasidate 'Amanohashi­
date'. When the first syl lable contains two vowels or a vowel followed by n, the fal l may
be heard within that syl lable: Koobe 'Kobe', Kyooto 'Kyoto', Slnbasi 'Shimbashi ',
Kyuusyuu 'Kyushu', Doitu 'Germany', Suisu 'Switzerland ', naiti 'Japan proper'.
M ESOTO N I C words have their fal l of pitch somewhere i n the midd le of the word ; the
first syl lable is pronounced rather low, but it often rises when it contai ns two vowels or a
vowel fol lowed by n : Nara-si 'Nara city', Tiba-ken 'Chi ba prefecture', Koobe-eki 'Kobe
station', Akasaka 'Akasa ka', Nagasaki 'Nagasa ki ', Harazyuku 'Harajuku', I idabasi 'I idabashi',
§ 0 .5.1 . I nherent word accents 19

Ootemati 'Otemachi ', Sinzyu-wan 'Pearl Harbor', Sendagaya 'Sendagaya', Taihei-yoo 'Pa­
cific Ocean'; I kebu kuro 'I kebu kuro', Akihabara 'Akihabara', Meguro-ku 'Meguro ward',
Sinagawa-ku 'Shi nagawa ward', l kebukuro-eki 'l kebukuro stat ion', N isi-Ogikubo-eki 'N ishi­
Ogikubo station'.
OXYTON I C words have the fal l of pitch on the last syllable, but if that contains a sin­
gle vowel you wil l hear the fal l only when the word is i mmed iately fol lowed by a particle
or copula: inu desu 'it's a dog', onna wa 'as for the woman', otoko mo 'the man also',
otooto ni 'for my younger brother', Kitizyoozi e 'to K ichijoji ', zyuuitigatu made 'til l No­
vember'. I f the fi nal syl lable contains two vowels or a vowel fol lowed by n, you may be
able to hear the fal l of pitch even without a fol lowing particle: koohii [desu ] ' [ it is] coffee',
R yuukyuu [ mo] 'the Ryu kyus [also ] ', Tyoosen [de] ' [ in] Korea', takai 'it is expensive',
yasu i 'it is cheap', kuroi 'it is black'.
ATO N IC words have no fal l of pitch even when fol lowed by a particle. All the syl lables
are pronounced rather high except for the first, which usual ly starts off low: kore [ wa]
' [as for] this', Sibuya [ e] ' [to] Shibuya', Gotanda [ n i ] ' [to] Gotanda', Sinagawa [ kara]
' [from ] Shinagawa', Otyanom izu [ mo] 'Ochanomizu [also ] '. If the f irst syl lable consists of
two vowels or a vowel fol lowed by n you may hear a rise: Oosaka [ mo ] 'Osaka [also ] ',
Taiho ku [e] ' [to] Tai pei', Kanda [de] ' [ i n ] Kanda'.
The inherent accent of a simple word cannot easily be pred icted ; it is something to
learn along with the consonants and vowels. But there are ru les by which you can pred ict
the accent of compound words, which are newly created every day, and these rules are
mentioned in the appropriate sections of the book. Simple verbs and adject ives show only
two TYPES of basic accent : aton ic or ton ic. I f you know the basic type, you can predict
the accent of a given form of the verb or adjective by rule or by analogy with simi lar forms
of the same type. Most simple verbs and almost al l adjectives are the TON IC type.
On the other hand, many nouns of three or four moras-especially and most important­
ly those written with two Chinese characters-are atonic. R ecent loanwords from
Engl ish wi l l usual l y either fol low the E ng l i sh patterns or put the accent on the third
mora from the end (or one mora earl ier if that mora is the second part of a long syl­
lable) .
I f we leave aside certain special types and particular exceptions, the accentuation of
compound nouns ( N 1 + N 2 ) can be descri bed as fol lows :
( 1) I f N 1 is tonic, the accent is removed. (Otherwise we know the structure is a syntac­
tic reduction rather than a com pound noun.) The accentuation of a compound noun de­
pends on the last element.
(2) If N 2 contai ns more than one syl lable and has an accent on any mora other than
the last, that accent is retained as the accent of the compound.
(3) Otherwise the i nherent accent is ignored and a new accent is put on the first syl lable
of N 2 provided that noun contains more than two moras; if N 2 is only one or two moras
in length the new accent retreats to the LAST syllable of N 1•

(4) There a re a n u m ber of ATON ICIZI N G SUF F IXES w h i ch except iona l l y remove a l l
i nherent accents without imparti ng a new accent. Some of these are derived from free
nouns; those of native origin are mostly oxytonic. There appear to be no more than fifty
of these suffixes, but the l ist may be growing, since several of them (such as -sya 'vehicle'
and -kin 'money') exhibit the regular pattern as wel l as the atonicizing pattern, wh ich is
probably the newer variant.

�------
20 § 0 . Notational Conventions

0.5.2. Variations in inheren t word accents.

I n d ifferent parts of Japan the same word may be heard with different accent patterns,
but the rapid spread of mass communications has led to wider and wider use of the stand­
ard accents of Tokyo speech, and these are what we mark in this book. But even withi n
the standard language certai n words are said with more than one pattern; older speakers
and younger speakers sometimes d iffer on the pattern they choose for a given word, and
one and the same speaker may find himself freely varying the accent of certai n of h is
words. I n th is book we attempt to show a l l possible varieties of accent i n standard use for
each word by placing an accent mark over the vowel at each point where a spea ker might
choose to locate the fal l of pitch . In pronounci ng the word ko koro [mo] 'the heart [also]',
some people wi l l say kokoro [ mo ] with the accent on the last syl lable of the noun, while
others-probably the majority-will say kokoro [ mo ] , with the accent i n the m iddle. I n
pronouncing zyuusan-n iti 'thirteen days' a given spea ker may find himself sometimes say­
i ng zyuusan-niti with an accent on the first syl lable and someti mes zyuusan-niti with the
accent on the second. Certai n tonic words are optiona l ly pronounced as atonic, and this is
shown by placing a raised minus sign in parentheses at the end of the word: eiga(-) 'ci nema' is
prototonic for some speakers, atonic for others. The word zidoo-sya(-) 'automobi le' may
be said by a given speaker someti mes with no fal l of pitch and sometimes with the fa l l at
the second syl lab le just as some Engl ish speakers wi l l someti mes put the heavy beat of the
English accent at the beginni ng of 'AUTomobile' and other ti mes put it at the end 'auto­
moB l L E '. The expression dono-yoo na 'what kind of' will be said as either /donoyoona/
or as /donoyoona/, and that is why we write a single hyphenated word rather than two
words. A phrase l i ke kanemotf (-) mo 'the rich man also' may be heard with any of three
patterns: kanemoti mo, kanemotf mo, or kanemoti mo. Though some patterns are more
common than others, our notation offers no prescri ptions. As a ru le of thumb, the foreign
learner would do wel l to prefer the atonic variant of a noun (whenever one is avai lable)
and the tonic variant of a verb or adjective, since this appears to be a trend toward which
the language is moving. I n Nagoya and G ifu a l l adjectives are treated as tonic, as are all
vowel verbs (Gekkan-Bumpo 2/2.1 69) ; this means that all passives, causatives, and desidera­
tives (and negatives?) are ton ic, even when they are made on an underlying verb that is
atonic.
0.5.3. Variations due to vowel unvoicing.
I n Tokyo speech when the h igh vowels i and u appear between voiceless consonants
they are usua lly unvoiced (whispered) : kusa 'grass', sita 'tongue', syuppatu 'departure',
kippu 'ticket', tukau 'uses', hitori 'one person', hutarf 'two people', kikai 'machi nery',
titf 'father', tutumu 'wraps', susumu 'advances', sisoo 'thought', hisyo 'secretary',
syusyoo 'prime minister', etc. When the unvoiced vowel is to carry an accent in certai n
verb forms, most Tokyo spea kers choose t o shift the accent over t o t h e fol lowing vowel,
so that kfta ka 'came?' is made to sound l i ke kita ka 'wore?' Other spea kers leave the ac­
cent alone, even while unvoicing the vowel ; the l istener must infer the location of the ac­
cent from the surround i ng pitch levels. We might show this variation by writing kfta 'came',
kftta 'cut', hutta 'rained', tuku 'arrives', tu kete 'attaching', and the l i ke ; but instead we wi l l
minimize the clutter o f accent marks b y regularizing o u r transcription t o accord with those
speakers who retai n the basic accent ( kita, kitta , hutta, tu ku, tu kete, etc.) and ignore the
 § 0 .5 .4. Word accent withi n phrases 21

common variant which delays the accent.4 Whenever you actual ly see two accent marks
on a verb form, you wi l l know that the basic verb has both tonic and atonic treatments ;
but two accent marks on an adjective form someti mes i ndicates variant treatments of cer­
tain classes of ton ic adjectives, as explained in Martin 1967. (Only the more prevalent ac­
centual variants are shown. )
After a voiceless consonant Tokyo speakers often unvoice i o r u at the end o f a tonic
phrase: Motfron desu 'Of course', Yosi 'O K ', Hayaku 'H urry up', N agasaki '( It's) Nagasaki'.
When the accent is expected on the immediately precedi ng syl lable, you wil l hear no fall
of pitch ; the only signal that the phrase is tonic wil l be the whisperi ng of the f i nal syllable
itself: Arimasu 'I 've got some', Kore desu 'It's this'.
0.5.4. Word accent within phrases.
Within a single phrase the pitch can fal l only once. And some phrases have no fal l of
pitch ; they rise and stay up to the very end : kore wa 'as for this', sono teeburu mo 'that
table also', akai denwa o tukau 'I wi l l use a red (= publ ic) telephone'. That is because the
component words are inherently aton ic-or because the final word is oxytonic and its
final accent must vanish when there is no further syl lable left to carry the fall . When an
oxytonic noun or adverb appears at the end of a phrase, we wi l l show that it has an i n­
herent accent by placing the accent mark i n the appropriate place, but we wi l l put brack­
ets around the mark to i nd icate the automatic cancel lation by which it sou nds as if it
were atonic:
H utal f 1 i masu 'There are two people'. Cf. H utar f ga imasu 'There are the two people'.
Takusk h tabeta 'I ate lots' . Cf. Takusan desu 'It's lots'.
l k-k�� itta 'I went one time'. Cf. l k-kai datta 'It was one time'.
l ti-d g1 sita 'I d id it once'. Cf. I ti-do datta 'It was once'.
When you have used up the sing le fal l of pitch al lowed withi n a phrase, each basic ac-
/ cent expected to turn up later is automatical ly cancel led ; once your tune goes down, it stays
down. We wil l show this by putting brackets around the cancel led accent marks. B ut i n the
extremely com mon and fami l iar situations of noun + particle and noun + copu la we will
normal ly forgo reminding you of the cancel led accent. By recalling the phrases Yokohama
made 'as far as Yokohama' and Yokohama desu ka 'is it Yokohama?' we know that the
particle made and the copula desu are basically prototon ic; accordi ngly, we wi l l not bother
to i nd icate the cancel led accents in Nagoya made (= Nagoya m kUe) 'as far as N agoya' and
Nagoya desu ka (= Nagoya d��u ka) ' I s it Nagoya?' Moreover, we will mark the final accent
of an i nherently oxytonic particle only when it is heard . From the phrase Yokohama kara
desu 'it is from Yokohama' we know that the particle kara is basical ly oxytonic so we will
not bother to i ndicate the cancel led accent in Nagoya kara desu (= Nagoya kar�1 Je1su) ' I t
is from Nagoya'. T h e phrase koko de m o ' i n t h i s place also' tells u s t h e locative particle de
has an accent-as do al l one-syl lable particles-but we will not mark the accent when it is
cancel led at the end of a phrase : koko de asobu = koko Je1 [ I ] asobu 'we wi l l play i n this
place'. 5
4. But the phrasal postpositions ni tuke, ni tuki, and ni tu kete are cited with both accentuations i n
§ 9. 7 , though elsewhere w e wr ite k i o tu kete for what Tokyo speakers usually say as k i o tukete.
5 . But in d iscussing particles or citi ng them i n isolation we wil l usually mark the accent, especially
for those such as to or no or made wh ich might otherwise be m isread as E ngl ish words. And sentence­
final particles such as ne or yo are written with the accent to ind icate that a minor j unctu re may precede
them.
22 § 0 . Notational Conventions

An atonic phrase often drops its final juncture and gets pronounced as if part of the
fol lowing phrase : Akai [I] denwa o [ I ] tu katta 'I used a red telephone' consists of three u n­
derlying phrases but these are normally run together and pronounced as if one long phrase
akaidenwaotukatta. An oxytonic phrase cancels its final accent before juncture and there­
fore gets treated as an atonic phrase when the juncture, in turn, drops out : H utaiu1 [I]
tu katta 'I used two' is run together and pronounced hutatutukatta, with no fal l of pitch.
Other tonic phrases, in which the fal l of pitch is not exposed to cancel lation, wi l l retain
their accent under similar cond itions and the juncture will less readi ly drop. Should the
juncture drop, later tonic words will automatically lose their accents. I nstead of brackets
around the cancel led accent marks we wi l l someti mes use parentheses, to indicate that the
dropping of the juncture is optional ; the two phrases need not be run together , though
that may be the common practice : Kaesa'nai y�b desu 'Apparently they're not goi ng to re­
turn it' can be pronounced Kaesa'nai I yoodesu in two phrases (with a reduced fal l in the
second ) , but commonly it is run together as a single phrase Kaesanaiyoodesu . I n the sen­
tence KaganW mite kudas�� 'Look in the mirror', the brackets arou nd the first accent show
that it is automatically cancel led, being at the end of the phrase, which is a short version of
kagam f o and is run together to form a single phrase with mite ; the parentheses around the
final accent tells us that we can pronounce the sentence either as two phrases kagamimfte I
kudasa'i or as a single phrase kagamimftekudasai, the latter version bei ng more l i kely.6
The inflected forms of verb and adjective are usual ly marked for the actual accent heard
in the sentence, not for the basic accents from which this derives. The i nformation g iven
wi l l permit you to i nfer whether the underlying verb or adjective is basically ton ic or
atonic. There is actual ly an i nherent final accent on the endi ng of what appear to be u n­
accented forms of the "atonic" verb, but this wi l l be heard only when a particle fol lows
and we usual ly omit the suppressed accent, writing I tta' ka 'D id you go?' but I tta (= l tt� 1 )
'I went' and I tta kodomo wa ... (= l tt�1 [I] kodomo wa) 'The chi ld who went'. I nflected
forms of "aton ic" adjectives are simi lar, but the final accent is usual ly heard one mora
earlier : Aka'i ka ' I s it red?' and Akai denwa ... ' =Ak� � [I] denwa) 'A red telephone'-but
Toof ka ' I s it distant?' and Tooi kuni ... (=Too i1 [ I ] ku ni ... ) 'A distant land '. I n the
i nfinitive form made with the suffix -ku the final accent, when called for, may optionally
occur one syl lable early if a particie is attached : Akaku mo nai. 'It isn't red, either' but
Akaku na'tta (=Akak�1 [I] na'tta) 'It became red'.7
Some verbs and adjectives permit variant accentuations of either type, tonic or atonic.
The imperfect and perfect forms of the atonic verb ( l i ke suru 'does' and sita 'd id') and the
imperfect and i nfinitive forms of the atonic adjective (such as akai 'is red ' and akaku 'be­
ing red ') wi l l have basic accents on the end ings (for the verb -ru and -ta; for the adjective -f
and -ku but with a shift of the accent back one mora under certain circumstances) ; we wi l l
ignore this except when the form i s fol lowed b y a particle o r copu la. The verb simesu (-)
'reveals' is treated as either tonic or atonic, and that is what the notation tells us. When a
particle is added , for example ka', the two avai lable pronunciations are marked : simesu ka.
Other forms wi l l also show two pronunciations, e.g . the cond itional simesitara.
6. But in certa in very common situations, such as those i l l ustrated here, we wi l l not always put pa­
rentheses arou nd the later accents; th t reader can infer them from earlier discussions.
7 . Alo ngside Tooku natta ( = Took •] [ IJ natta) ' I t became d istant' we f i nd three versio ns of TookuI
• t t t t

mo nai ' I t isn't d istant, either', depend ing on whether the double o is taken as one heavy syl lable or
two l ight ones as wel l as on the option of anticipating the accent on -ku .
§ 0.5.5 . Regressive cancellations 23

Certain particles are attached with an underlying juncture (optional or obl igatory)
which kills the basic final accent on such forms as suru and sita, aka i and a kaku. The parti­
cle to, for example, whether used as a quotation marker or i n the meaning 'when(ever)', is
attached like ka by many Tokyo speakers, especially the younger ones, who say /suruto/
and /sitato/ for what other and more trad itional speakers say as /suruto/ and /sitato/. I n­
stead of showing this option every time it is ava i lable, we wi l l fol low the traditional speak­
ers and ignore the common variant in our notation ; rather than write "suru n to" we wi l l
write "suru to" and let the reader remember that the other version is possible. Thus when
you see "munasiW) to" you wi l l know that the adjective enjoys variant I N H E R ENT ac­
cents, independent of its occurrence with to in the given sentence. Some speakers wil l say
/munasiito/ because they treat munasf i as tonic everywhere, others because they accentu­
ate all atonic forms of verbs and adjectives before to. I n general we have tried to let our no­
tation maximize the accentual distinctions and for that reason we choose to ignore those
systematic variants which obscure the d ifferences between ton ic and atonic verbs and ad­
jectives. (See the remarks on accentuating the desiderative forms, § 7, and com pound
verbs, § 9. 1 . 1 0.)
0.5.5. Regressive cancellations.
Many of the words we ca l l restrictives and quasi-restrictives i n § 2.4 h:.:ive basic patterns
that DOM I NATE in that they take precedence over earl ier accents with i n the phrase ; such
a pattern is i ndicated by putting a raised m inus at the beginning of the domina nt word, after
the space that separates it from the precedi ng word . When you see a word l i ke -dokoro or
-gurai you know that all precedi ng words in the same phrase will lose their accents (if any) ,
so we do not bother to put brackets around the cancel led accent marks : iti-zikan -gurai ( =

iti-iikan gurai) 'about one hour' is pronounced itizikangurai, iti-zikan -han ( iti- i1ikan han)
=

'an hour and a half' is pronounced itizikanhan, and iti-zikan -han -gurai ( iti- ii�an h�h
=

gura i ) 'about an hour and a half' is pro �ounced itizi kanhangurai. N p,f all of these dominant
words have accents; some are atonic: nimotu -nami ni atukatta ( nimotu nam i ni atukatta)
=

'they treated them l i ke baggage' is pronounced nimotunamini [ I ] atukatta. And some are
oxytonic, with a final accent that wi l l be automatically cancel led u nless fol lowed by a par­
ticle or copula: Hanbun -da ke desu 'It is (exactly) half' is pronounced hanbundakedesu
and Hanbun -da ke herasoo ' Let's reduce it by half' is pronounced hanbundake [ I ]
herasoo.
The l ittle word no , regardless of which of its many meani ngs it is expressing, has a
unique effect. I t cancels the final accent of a precedi ng oxytonic noun (as if it were a junc­
ture ) , except u nder certai n circumstances. The circumstances permitting the oxytonic
noun to retai n its final accent requ ire a retai ned juncture AFT E R no e ither in the surface
form as pronounced or at that level of structure put together just before arriving at the
surface form-at the poi nt when last-mi nute phrasi ng options are to be selected . Predicting
these circumstances is tricky, so we wi l l put brackets around those final accents that are
to be cancel led : l k�1 no mawari o mawaru 'We will go around the pond ', Kin�1o no uryoo
wa 'Yesterday's rainfa l l ', N i h�h no tetudoo wa 'Japan's railroads', Otok�1 no ga kusei wa
'Male students'. When the final accent of an oxytonic noun fai ls to be cancel led before
no, you can usual ly expect a new phrase to begi n after no : H utar i no sensei wa 'The two
teachers' is usually pronounced hutarino I senseewa. But sometimes the two phrases wil l
collapse i nto one a t the last m i nute : Yuki n o y'1Jo desu ' I t looks l i ke snow' may be

�-----
24 § 0 . Notational Conventions

pronounced yukino I yoodesu as expected, but the commonly heard version is

yukinoyoodesu. ( I f the dropping of the juncture had been cal led for when packag ing the
.
constituents at an ear 1·1er stage, we wou Id expect * yu"',!d no yoo
' d esu * yu k"inoyoo ' d esu. )
=

When a dom i nant word fol lows an atonic word, appl ication of the regressive cancel la­
tion is vacuous, si nce there is no accent for it to cancel. But we wi l l usually i nd icate the
accentual dominance of the word by the ra ised minus even when it has no work to do :
kodomo -nami, nisen-en -gurai , etc.
0.5.6. Other accent cancellations.
I n certain expressions accents are cancel led, obl igatorily or optional ly, for reasons that
can be rather complicated to explain. Obl igatory cancellat ions are i ndicated b � brackets
around the accent mark: H it�1tu mo kawanakatta 'I d idn't buy even one', H it� ri mo
konakatta 'Not a soul came', D�t e mo i nai 'No one is there'. Optional cancellations are
shown by parentheses around the accent mark: N(a)n Je> mo fi 'Anything wi l l do ' can be
pronounced as nandemo I ii, as nandemo I ii, or as nandemo [ I ] li.
0.5. 1. Accent shifts in certain verb forms.
The accentuation of a g iven i nflectional form is predictable, once you know whether
the verb or the adjective is treated as basical ly tonic or atonic. The accent wi l l not neces­
sarily fal l on the same syl lable in every form of the paradigm ; the imperfect endings -ru
(for verbs) and -i (for adjectives) attract the accent of tonic bases to the syl lable just be­
fore the end ing : tabe(te) 'eati ng' but taberu 'eats', takaku (te) 'being expensive' but takai
'is expensive'. Yet for certai n verbs, when the accent is expected on the vowel before the
ending, it is sh ifted back to an immediately preced i ng vowel ; the two vowels are treated as
a single "heavy" syl lable. The relevant information is stated below.
( 1) The fol lowi ng verbs always treat the vowel dyad as a single syllable, forcing the ac­
cent in the imperfect (-ru) and the provisional (-reba) to retreat one mora : kaer-u 'returns'
(and huri-kaer-u 'looks back') , kaesu 'returns it' (and most tonic compou nds with -kaesu
such as h iki-kaesu, kiki-kaesu, etc.) ; gotta- k /g aer-u 'gets confused'; h i rugaer-u 'flutters,
reverses', h irugaesu 'reverses/waves it' ; hair-u 'enters', mair-u 'comes/goes'; tooru 'passes',
toosu 'lets pass' (and tonic compounds with -tooru and -toosu) . The i nfinitive, too, wi l l
place t h e accent a mora earlier than expected ( kaeri , toori , etc.) a s wil l t h e imperative
( kaere, toore, etc.) . The geru nd and related forms (the perfect, etc.) of the potentials will
also place the accent a mora earl i er than expected : kaerete (from kaereru) , haireta (from
haireru ) , tooretara (from tooreru) , etc.
(2) I n the fol lowi ng verbs, the vowel dyad is OPT I O NA L LY treated as a single syl lable;
most (but not all) Tokyo spea kers move the accent back one mora from its expected loca­
tion in the imperfect :8 aturaeru 'orders', humaeru 'treads', kangaeru 'thi nks', koraeru
'withstands', kotaeru 'answers', matigaeru 'mista kes', modaeru 'agonizes', osaeru 'restrai ns',
sakaeru 'flourishes', tagaeru 'violates' ( N H K also recognizes an atonic version) , tonaeru
'chants, advocates', toraeru 'captures', tu ka'eru 'clogs up'; otoroeru 'is i nferior', totonoeru
'prepares'; si-kaesu 'redoes'; ? ... . Also tonic compou nds with -kaeru (such as ki-kaeru, si­
kaeru , nori-kaeru, h i ki-kaeru) and with -tigaeru (such as iki-tigaeru) , § 9.1.1 0. The
9

8 . And also in the provi sional ( kotaereba ) and , with the exception of si-kaesu ( si-kaesanai ) , the
negative forms: kotaenai, kotaenakatta, etc.
9 . A distinction is maintained between huri-kaeru 'transfers ( money ) ' with the infin itive huri-kae
§ 0.6. Miscel laneous conventions 25

i nfinitive and the imperative of these verbs (with the exception of si-kaesu) cal l for the ac­
cent to fal l on the fi rst vowel of the dyad i n any event, but when the particle ro is added
to the imperative both options are available: humae yo and humae ro both mean 'tread ! '
and the second form perm its two d ifferent accentuations.
(3) Assuming that our l ists are comprehensive, al l other verbs with vowel dyads treat
the two vowels as separate syl lables so that the second vowel freely takes the accent:
aeru 'dresses ( vegetables)', haeru 'grows', kamaeru 'bui lds', kanaeru 'grants ( a request) ',
kitaeru 'forges', naeru 'withers'; m ieru 'seems', mazieru 'mixes'; hueru 'grows'; hoeru
'barks', kazoeru 'counts', koeru 'gets fat', kokoroeru 'realizes', oboeru 'remembers',
omoeru 'seems', soroe�u 'arranges', sueru 'sours'; ureeru 'grieves'; silru 'coerces', h i ki lru
'leads'; naoru 'recovers', naosu 'repairs', taoru 'plucks', taosu 'topples'; kooru 'packs
up'; ... . This group i ncludes the short potentials made from tonic verbs: aeru 'can meet',
haeru 'can crawl', kaeru 'can raise', kuraerun 'can eat', naeru 'can plait', naraeru 'can
learn'; kisoeru 'can vie', koeru 'can love/beg', toerun 'can i nquire', tukuroeru 'can mend ';
kueru 'can eat', nueru 'can sew'; . . . .
(4) The fol lowi ng verbs are optionally atonic. When the ton ic option is chosen the
dyad is treated as two syl lables and the second vowel freely takes the accent : amaeru (-)
'coaxes', kakaeru (-) 'embraces', kosaerun 'concocts', kuwaerun 'adds', saraeru (-) 'dredges',
sasaeru(-) 'supports', takuwaerun 'hoards' ( K also has takuw�eru and N H K has takuw�eru) ,
tataeru (-) 'brims with ; praises', tukaerun 'serves', tutaeru(-) 'communicates' ( K also has
tutaeru) , uttaeru n 'compl ai ns about' ( K and N H K both also have u ttaeru ) ; tu ieru (-)
'is wasted', katueru(-) 'hu ngers' [obsolescent] , m i-sueru (-) 'gazes'; m oyoosu (-) 'holds
(a meeting) '; ... .
0.5.8. Accent in dialect forms.
Much of the i nformation on d ialect forms is taken from secondary sources which fai led
to note the accent. I n isolated citations I have simply omitted i nformation on accent; but
in general, especial ly for sentence examples, I have marked the accent as if the sentences
were said by a Tokyo speaker, as an aid to identifying the component words. This com­
promise notation is less than satisfac_tory from a scholarly poi nt of view, but there are two
facts wh ich make it seem better than marking no accent at al l . One is that sentences with
dialect vocabu lary and grammar are sometimes heard from the lips of Tokyo speakers,
either readi ng aloud or playi ng dramatic roles, and few speakers are capable of making the
subtle adjustments necessary for an authentic and consistent version of someone else's
d ialect; none of the accentuations i nd icated here is totally artificial. Moreover, the accen­
tuation across the various d ialects is far from random ; there is a correlation by word types,
so that a speaker of a given d ialect wil l find that the Tokyo markings provide h i m with a
fairly consistent clue to many of his own pitch fal ls, even when these occur on a d ifferent
syl lable from the o ne heard in Tokyo.
0.6. MISCEL LA NEOUS CONVENTIONS

Brackets and parentheses are also used to mark various asides and shortenings i n

and h u ri-kaer-u 'looks back' with the i nfinitive huri-kaeri. ( N H K l ists only h u ri-kaeru for 't ransfers' ,
but H and K give both versio ns.) To the l i st add h i kaeru 'refrai ns' and sonaJru 'provides' .
26 § 0. Notatio nal Conventions

explanatory passages, accordi ng to fam i l iar conventions which should cause no d ifficulty
for the reader. I n d iscussions of pronunciation, brackets sometimes enclose a q uasi-phonetic
transcription, as when we say that ee is pronounced [e:], slashes sometimes enclose a quasi­
phonemic transcri ption, as when we say that ei is to be pronounced /ee/. A slash between
two forms is the fam iliar convention to show optiona l ity : a/b 'either a or b'; a(/b) 'either a
or possibly b'; a/b/c(/ ... ) 'a or b or c or possibly others unmentioned'. When more than two
sets of options are shown in a single formu la, they are usually to be taken as i ndependent
of one another : Dare ni/ga kodomo ga iru/aru ka 'Who has chi ldren?' tel ls you there are
four Japanese versions of the sentence.
The asterisk * precedes an u nattested form. I n historical discussions this refers to a form
hypothesized to have existed despite the lack of d irect evidence, but in descri ptive discus­
sions an asterisk often marks a sentence (or other formation) presented as an example of
u ngrammatical ity which is i ntended to shed l ight on the structure of those sentences
which are grammatical.1 0 Arrows are used to show synchronic relat ionships, typica l ly those
of sentence conversion: a � b 'a y ields b, a is converted i nto b, a u nderl ies b' or ( *a � b)
'b wi l l replace the unacceptable a'; b � a 'b is a conversion from a, b results from a process
applied to a'. When one form has replaced another historical ly, a d ifferent kind of arrow is
used : a > b 'the earl ier form a developed i nto the later form b'; b < a 'the h istorical source
of b is the ear lier form a'. ( I n citi ng certa i n u nderlying forms, < and > are a lso used to
mark accent shifts.)
Abbreviations for grammatical terms are genera l ly explained where they first occur;
they wil l all be found i n the I ndex. Some very frequent designations are S 'sentence', N
'noun' or 'nominal sentence (= predicated noun) ', A 'adjective' or 'adjectival sentence
(= pred icated adjective) ', V 'verb' or 'verbal sentence (= pred icated verb) ', VN 'verbal noun'
or 'verbal-noun sentence', AN 'adjectiva l noun ' or 'adjectival-nou n sentence', PcN 'precopu­
lar (=quasi-adjectival) noun', AUX (or Aux) 'auxil iary'. Although V is also used as an ab­
breviation of 'vowel ' ( i n contrast with C 'conso nant') , it shou ld be clear when this is to be
taken as 'verb'. VI stands for intransitive verb, VT for transitive verb; VN I for i ntransitive
verbal noun, VNT for transitive verbal noun. In add ition to representing 'adjective' the
letter A is also used in situat ional formulas, where A B and C stand for three d ifferent
people, X Y and Z stand for three different things, and P and 0 stand for two d ifferent
places.
In the Japanese transcriptions square brackets enclose elements that are potentially or
theoretically a part of the sentence, though not necessari ly present in the example as
given. When the brackets have a notch (or superimposed hyphen) f t the material enclosed
is optionally sayable; when the brackets have a double notch (or superimposed equal sign)
f f the materia l enclosed must be suppressed. Unnotched brackets can be taken either
way ; usually they are to be considered optional. They are used, for example, to show vari­
ous col loquial contractions, as i n M ita n [o] desu 'I 've seen it' or Kaero [o] ! 'Let's leave! ' ;
but sometimes contraction is i nd icated b y an apostrophe: M atte 'ru = Matte [ i ]ru 'I 'll be
waiting'. We must be carefu l not to use the apostrophe after the letter n u nless the nasal
forms a mora : the d ialect contraction sen [eb]a cannot be shown as "sen'a " because it is
pronounced /sena/. In mentioning certai n endings, a basic form is cited that sometimes i n­
cludes parentheses or brackets; the hortative is given as - [ y ]oo because the -y- originated

1 O. Degrees of u nacceptabil ity are suggested by mark ing a sentence with ( ? ) , ?, ( ? * ) , ( * ) , * .

§ 0.6. Miscellaneous conventions 27

by epenthesis, the negative is cited as -(a)nai because the -a- is part of the original for­
mation.
In exam ples cited from written sources, the bracketed materia l was usual ly not present
in the original text, but was added here to help explai n the overt forms. When the brack­
ets enclose a blank [ ] " what is omitted is either left u nspecified or is presumed to be
"

obvious.
Certa i n short Japanese words have romanized forms that are identica l with E ngl ish
words, and this can cause momentary confusion. Whenever it is possible to d ifferentiate
such words by writing the basic accent of the Japanese forms, I have done so: n o , t o , m ade,
site, tam e , ... . B ut for atonic nou ns such as sake 'rice wine', sore 'that', are 'that', etc., the
accent marking is not available ; I have tried to avoid letting such words fal l into positions
within E ng lish sentences where they wi l l mislead the eye of the reader, and I bel ieve there
wi l l be few occasions for discomfort.
 1 SENT ENC E CONST R U CT I ON : N UC L EAR S EN TE N C ES (PR E D I CATES)
AND E XPAN D E D SENTENCES (SIMP LEXES); SENTEN CE CONVE RSIONS

I n this book we attempt to descri be the wide variety of sentence types used by
Japanese speakers to express meanings withi n the context of situation and discourse. We wil l
concentrate o n the sentence, rather than the situation o r the discourse, though occasiona l
attention is paid to the larger setti ng. A lthough we will not try to find an overal l definition
of "sentence" as such, some l ight wi l l perhaps be shed on what might go i nto such a defini­
tion as we look at particular sentences and sentence types. The term "sentence" is used
qu ite loosely-to refer both to relatively simple clauses and to extremely involved concoc­
tions. I n general we will explai n longer sentences in terms of combinations of shorter sen­
tences; yet a number of shorter sentences wi l l be accounted for as reductions of something
larger. I n order to explai n the structure of certain sentences we rely heavily on the con-
cept of E L L I PS IS-the suppression of words or phrases presumably i ntended by the speak­
er and understood by the l istener. Typically the suppression is optiona l, and the omitted
material can be freely suppl ied to render the spoken sentence more expl icit; when the el­
lipsis is obl igatory, our i nterpretation assumes that a change has taken place (or is taking
place) in the history of the language. But i n some i nstances our assumptions, while val id
withi n their narrow frame of reference, will prove historically false : we may be accounting
for modern structures i n ways that oversi mpl ify the actual histories. E l l ipsis is shown by
putting brackets [ ] around t he omitted stretch ; for those omissions regarded as optional
we can (when we wish to be precise) use brackets with a sing le cross bar .f t and for those
regarded as obl igatory we can use brackets with a double cross bar t f . Although this
grammar does not seek to cover the history of the language, likely origins are suggested for
some of the phenomena examined. At times our historical perspective wil l be perversely
narrow; at other times we take a demand ingly wider view.
I n the mai n this work is simply a descri pt ive taxonomy of modern Japanese sentences.
The framework for the taxonomy is partly arbitrary and thus can make no clai m to u lti­
mate real ity beyond the needs of the work itself. I t is my belief, however, that the major
categories posited to describe the sentence structures are in some way functional parts of
the Japanese language that can eventually be val idated, d irectly or i nd irectly, not only as
psychologica l real ities but also as h istorical entities. I have sought to find order in the rela­
tionshi ps that seem to obtain between the categories and in the ways the categories com­
bine to yield the surface forms of the sentences I have examined. Though the reader may
occasiona lly be d ismayed by ru les and formu las and algorithmic charts, this is not a "for­
ma l " grammar in the strict sense. Nor is it an attempt to relate the structure of Japanese
to any notion of "un iversa l " grammar ; for, despite the concepts and termi nology 1 that
have been freely borrowed from descriptions of other languages (and other descriptions of
this language) as wel l as from d iscussions of grammar in general, the ideas presented here
were developed through working i ndependently with the sentences themselves. I have
stuck fairly closely to the SUR FACE forms of sentences; as far as possi ble I have chosen
1 . I f the reader misses fam i l iar labels for particular categories of the grammar, he will probably find
them listed in the I ndex, with a cross reference to the correspo nd ing terms used in this book or a defi­
nition i n those terms.
28
§ 1 . Sentence construction 29

to avoid tackl i ng the fascinating , but tricky, questions of relating these sentences to the
elements of natural logic (whether universal or language-specific) that may be expressed
withi n them.
The various sections of the book were written and rewritten at d ifferent times over a
period of many years. I have not hesitated to retai n confli cting explanations of a single
phenomenon when I felt there was somethi ng worth considering in each ; but I have tried
to pu l l the sections together as much as I could a nd to cal l attention to my own i nconsis­
tencies whenever they have come to my attention .
A t t h e outset I tried t o catalog t h e sentence types o f Japanese i n three major steps,
each encompassing a number of diverse operations. The first step was to take a relatively
sma l l stock of "nuclear sentences" (or predicates) and build them i nto a relatively large
number of "simplex sentences" (or expanded nuclear sentences) by plucking a pure noun
from the nuclear "nomina l " sentence (N da ' It's N ') and preposing it as an adju nct to some
given nuclear sentence, marking whatever grammatical relationshi ps might obtai n between
the two by a postposition of "case" relationship-either the core cases of subject or "nom­
i native" (ga) , d irect object or "accusative " (o ) , and i nd irect object or "dative" ( n i ) ; or, the
peri pheral cases of a l lative (e 'to') , ablative ( kara 'from') , instrumental (de 'with ') , locative
( nf/de 'at') , and reci procal (to 'reciprocally with') . With i n the operations of the same Step
One, the noun could be deli mited by one or more R ESTR I CT I V ES such as da ke 'on ly',
the D I STR I BUTI V E (-zutu 'd istri butively each') , the COMPARATO R (yori 'than') , a nd a
few other categories reduced from larger entities (nado, m itai , de mo, etc.) ; it cou ld also
be given FOCUS by the attachment of such particles as wa or mo and a few others of sim­
i lar function. The essential notion is that a nuclear sentence-Yobu 'Someone cal ls some­
one', Ook i i 'I t's big', Hon da 'I t's a book'-wi l l stand as a complete utterance i n Japanese,
though it may be expanded to give further i nformation : Haha ga yobu 'Mother cal ls',
Kodomo o yobu 'Someone cal ls the ch ild', etc. I n Step Two, simplexes with the same nu­
cleus can be combined to yield larger simplexes by shari ng the predicate : Haha ga kodomo
o yobu (or Kodomo o haha ga yobu ) 'The mother ca l ls the chi ld', etc. Also, sti l l withi n
Step Two, i t i s possible t o conjoi n ( by such markers a s to 'and', k a 'or', etc.) two o r more
nouns that serve joi ntly in the SAM E role : H aha to kodomo ga yobu 'Mother a nd child cal l
someone', Haha to kodomo o yobu 'Someone cal ls mother and chi ld ', etc.
Thus, the first two steps shou ld yield an enormous basic repertoire of simple sentences
of the type some linguists have cal led "kernels". Step Three takes these sentences and ex­
poses them to a number of operations whereby t hey are converted into many d ifferent
forms. Most of these sentence CONVE RSIONS enhance the u nderlying sentence with
some category of meaning-such as Negative, Perfect, etc.-though some serve primarily as
devices to relegate a sentence to a lesser role, as when Nomina lization permits a sentence
to function as a noun. To a large extent these sentence conversions correspond to what
some gram marians have cal led "general ized transformations"; those readers who find it
easier to think of "converted" sentences as "transformed" sentences are i nvited to do so.
As I worked on certai n problems it became evident that at least two additional steps
would be necessary. Step Four accounts for Thematization-roughly speaking, the source
and expression of a "topic" for the sentence-and Epithematization, the source of a target
for Adnominal ization. These rather complicated concepts are explained i n the relevant
sections of the book. F inal ly, I set up Step F ive to account for the pecu l iarities of those
sentences that express Propredication and Identification, as explained in the appropriate
30 § 1 . Sentence Construction

sections. Step F ive, in turn, can serve as a source for all the nominal nuclear sentences
(with the possible exception of a few non-predicable adverbs and the l i ke) that were as·
sumed in the beginning.
F rom this overview it wi l l be obvious that the present framework is inadequate in
terms of any of the currently popular theories of grammar, and I have not tried to adjust
the framework to s_µch theories, since each theory is in its way i nadequate to account for
such a wide sweep of data as I wish to encompass. It seems to me that recent attempts to
apply i nsights of l i nguistic theory to Japanese (or even to E nglish, from which most of
the theoretical specu lation stems) have been hampered by the lack of anything approach­
ing an adequate description of the language. Perhaps this book wi l l help fill some of the
gaps in our i nformation about the ingredients available to the Japanese speaker when he
coo ks up new sentences.
A word shou ld be said about three areas of inadequacy in the present treatment, even
within its own loose frame of reference. F i rstly, I have assumed that all mod ification of
nouns and of predicates is the resu lt of sentence conversions of adnominal ization and ad­
verbialization, respectively. The devices work admirably-up to a point. But there remains
a relatively small residue of non-predicable adnouns (or "prenouns") and adverbs that can
be fitted into such an explanation only at the cost of arbitrarily positing certai n nuclear
sentences that are otherwise unneeded : there is no *Goku da 'It is very'. Moreover there are
subtle problems within the processes of adnominal ization and adverbia l ization that require
special treatments. These leaky parts of the framework are patched u p by expl icit discus­
sions i n the appropriate sections of the book.
Secondly, I started off by positing only three types of nuclear sentences-verbal, adjec­
tival, and nominal. I t soon became apparent that among what looked l i ke pred icated
nouns were some words that had special verba l properties and others that had special ad­
jectival properties; these were set up as subcategories of V E R BA L NOUN and A DJEC­
T I VA L NOU N . (These terms are technical designations within the present treatment;
since "verba l nou n " is used to mean different things in other grammars, some readers
may prefer to think of these categories as "nominal verb" and "nominal adjective".)
The subcategorization of nominals, however, turned out to i nvolve a good dea l more than
distinguishing just these two obvious types ; the subject is treated extensively in §3, where
the reader wi l l find criteria to disti nguish most of the major parts of speech, including
various kinds of "pure" nouns.
Third ly, I cite the nuclear (and the simplex) sentences as fu ll-blown, usable surface ut­
terances, in the relatively unspecific "imperfect" form, sometimes cal led "nonpast" or
"present". But this form, l i ke the "perfect" form, is the resu lt of a process applied to a
more abstract entity, here called the I N F I N IT I VE. For convenience we assume that our
conversions depart directly from the imperfect forms (such as Haha ga kodomo o yobu
'Mother calls the child' or Hon ga ook i i 'The book is big', for example) rather than from
the u nderlying infin itive forms (Haha ga kodomo o yobi ... 'Mother cal l child .. .', Hon ga
ookiku ... 'Book be big .. . ') . Although the argumentation for the basic nature of the i nfini­
tives is quite convi ncing (see particularly §5) , it wi l l perhaps be easier to appreciate if ap­
proached from the outside rather than from within. When a Japanese speaker expresses a
sentence conversion, what he actually manipulates-if anythi ng-wil l be known only when
we have more revealing techniques of psychological testing ; the non-linguist foreigner (and
perhaps the Japanese himself) is l i kely to find the imperfect forms the best poi nt of de­
parture in studying the conversions.
§ 1 . Sentence construction 31

The order of presentatio n fol lows, to some extent, the order shown in the fol lowing
two charts. These are to be regarded as rough maps of u ncerta in terrain, at best, and they
may prove misleading to the unwary. The flow chart of STE P ON E gives a picture of how
a noun can be marked in various ways before joining with a nuclear sentence to form an
expanded sentence. I have not i l lustrated STE P TWO, whereby anything up to a maxi­
mally expanded simplex can be created by combi ning compati ble single expansions of the
same nuclear sentence so as to share the predicate (Haha ga yobu + Kodomo o yobu yield­
ing H aha ga kodomo o yobu or Kodomo o haha ga yobu etc.) and by conjoi ning adjuncts
that enjoy the same case marker ( H aha ga yobu + Kodomo ga yobu yielding H aha to
kodomo ga yobu or Kodomo to haha ga yobu-among other possibilities) . I n STE P
TH R E E , the major sentence conversions are ordered a n d numbered accordi ng t o a rough
scheme of applicabi l ity, spelled out in detail in the appropriate sections of the book. Chart
Two i ncludes STE P FOU R and STE P F I VE , since the epithemati zations of Step Four are
needed to account for the conversions of Adnominalization and Nominalization i n Step
Three, and the Propredication of Step F ive will perhaps explain the special nature among
the nuclear sentences of those nominal sentences containing pure nouns. The chart sug­
gests that I dentification is a special case of Propredication; that suggestion is discussed i n
§3.10.
I f the charts are confusi ng , they shou ld be disregarded; the reader is i nvited to plunge
directly into the description and look back at the charts at any poi nt where they might be
helpful in fol lowing the argument. G iven a surface sentence, you can glance through the
charts to see what operations may have gone i nto its creation, and in what order they
were applied. G iven a simple sentence that you wish to convert i nto more complicated
structures, you may need advice on how to approach the problem of orderi ng your opera­
tions; the charts wi l l not answer your questions in detail, but they provide a q uick re­
minder of the common patterns avai lable. By fol lowing the arrows you wi ll find that it is
appropriate to apply subject exa ltation ( kaku 'writes' -+ o-ka ki ni naru 'someone exalted
writes') before applying the desiderative ( ka ki-tai 'wants to write' but o-ka ki ni nari-tai
'someone exalted wants to write') , though the parenthesized arrow from desiderative to
subject exaltation a l lows for the l ess common pattern kaki-takute irassyaru. On conver­
sions applied to i nfin itive and/or auxi l iary, to gerund and/or auxiliary ( i n the structures
i nd icated by A c.I X in Chart 2) see §9.1.10 and §9 .2 4 , respectively.
.
32 §1. Sentence Construction

CHA R T ONE
S TEP ONE: From nucleus to simplex.
N UC L EAR SENTENCES
V-ru. R ecursion (to N da) is implied by each
VN suru. . .. da.
S= A-i . Zero bypasses are not all shown. 2
AN da. For special problems with possessive and
quasi-possessive predicates, see § 3.5;
N da. desiderative and quasi-desiderative, § 3.5a.
I
R ESTR I CT I VES
� I QUASI-R ESTR I CT I V ES (ate, darake, nami,
nuki, tuki, ... )
bakari
dake
gurai
�
�
N hodo da.
made

J
�
D I STR I BUTI VE

l
N zutu da. i
PE R I PH E RA L CASE MAR E RS
kara 'from'

I _D
e/ni , made 'to'
N de 'at' da. < - - - > P R E D I CATE VALENCES
:�
de 'with, using'
'reciprocally with '_

_
__

NADO d "'f::
M ITA i da. ADJU NCT FOCUS (+ S)
COM PARATO R koso [wa] I
N yori da . I ... sika
CO R E CASE MAR K E R S
+ S.
)� ... wa
... mo
... sae/sura [ mo]
... o ba
... o mo
N ga
N ni + V-ru. de mo
V N suru .
N o + transitive V-ru.
transitive VN suru.
+ quasi-intransitive V-ru.
quasi-intransiti ve VN suru.
STEP TWO: Conjoining simplexes with shared nucleus:
( 1) with identical case marker (or none)
(2) with d ifferent case markers
2. Si nce zero is a lways an option, the only mislead ing arrows are when a bypass is obligatory, as
when you are not permitted to add a core case marker after selecti ng the comparator. These and other
details are explained in the text.
 §1. Chart two 33

CHA R T TWO
STEP THREE: From nuclear or simplex sentence to converted sentence.
Major recursions and constraints are shown i n the chart; others are d iscussed separately.
Zero bypass is perm itted at every point. Each conversion (or set of conversions) is numbered .
Before the number, N V A or S means that the output is Nom i nal, Verbal , Adjectival or

/
Same as the i nput sentence. (After suru, V means 'when applied to verbal nucleus'; after
aru, N A means 'when applied to nom i na l or adjectival nucleus'.)
0. V-ru. ]-+ V T . CAUSAT I V E -+ V2. PASS I V E -+ V3. OBJECT -+ A5. D ES I D E.R AT I V E - - ->
VN suru.
A-i.
AN da.
l ( -sase) (-rare-) :
2a. pur e
2b. affective
EXAL �ATIO
(o-V-1 suru )
( V+ta-)

N da. 2c. potential

2d. short pot.
--- V5a. Desiderative
verbal i zation
( -.I
S4. SUBJ ECT E XA LTATI O N
____________
__.-/ (o-V-i n i naru)
�
- -> V6a. N U C L EAR
R ESTR I CT I O N
(ba kari/made/ ...
suru V, aru N A)
V6b. N U C L EAR F OCUS -+ Al. N EG- 8. ADVER B I A L I ZATI O N S :I nfinitive -i/-ku/ni (AUX)
(wa/mo/sae ... AT I V E Gerund -te ( AU X)

C
suru V, a r u N A) Conditio nal -tara
(sika ... aru N ) Representative -tari
Provisional -reba (+ S)
9. CO NCESS I V E-CONCUR R E NT ( nagara)
N 10. E V I D E N T I A L (-soo da) ] ./ ( 0 . )
V 1 1. EXCESS I V E (-sugiru) ] --->

[
- - > s - -> v - - - - - - - - - - - -
�2. PER F ECT (-ta) � 16. H � R ;A'v; v 19a . I M P E R ATI V E 20. QUOTAT I O N
c
� 13. ADNOM I N A L I ZA�
I 13 . Genitiviza-
(daroo) (soo da) ·� (-e) ( to)

J �
TION
a
S 15a. Literary
tion ( § 3. 1 1.2)
tentative (A 17. (rasii)
SEM B LA -
TI V E
V 19b. HORTATI V E
(- [ y ]oo)
2 1. EXTENS I O NS
(ne, sa, yo, ... )
S 18. STY L I ZAT I O N S 22 CONJUNCT I O N -
·

A L I ZATIONS
(desu, gozaimasu ) ( kara, to, si,
kedo, ga) + s

STEP FOUR: Extruded l l Extruded

\ ....... 1 ntruded
1 ntruded � Thematization : Epithematization

STEP FI VE: (S) � Propredication ( --+ N) l

( 0.
\
Identification --+ N da)
 2 PR E D I CAT E ADJ UNCTS

Although the nuclear sentences of Japanese can be used as complete utterances,

cited as they are in a finite "imperfect" form that means somethi ng l i ke 'it happens' or 'it
is', such utterances are relatively vague in content. To make the nuclear senten ce more ex­
pl icit, you can expand it by prefacing one or more A DJUNCTS, or "bui ld-up" phrases.
These adjuncts may be marked by one or more particles. The particles are often cal led
postpositions because they are placed after the word (typically a noun ) that they mark,
conveying the kind of i nformation that is sometimes carried in E ngl ish by prepositions. I n
§2.2 certai n of these postpositions are considered i n detai l , with a subcategori zation of
various uses of each. The postpositions exami ned are those which, among other uses,
someti mes mark the CAS E relationshi ps (or "valences") that obtai n between part icu lar
predicates and particular "arguments"-pure nou ns servi ng as adjuncts. Adjuncts that are
not normal ly marked with one of these postpositions are here considered to result from
one of the conversions of adverbia l i zation (with or without some overt marking ) -as if a
reduction from a nominal sentence rather than the product of propredi cation, though that
is elsewhere suggested for an alternative view. But ell i psis of case markers also occurs, lead­
i ng to another kind of "unmarked adjunct", as descri bed in §2.2a. I n §2.3 we wi l l exam­
i ne certai n particles that serve to focus emphasis, notably wa and mo. I n §2.4 - 5 we ex­
plore the distribution of those particles that serve to del i m it a noun ; those such as dake
'only' which can also be used as epitheme ( i .e. serve as the target of an adnomina l i zed
sentence) we wi l l cal l R ESTR I CTI VES, those more l i m ited in distri bution such as -okin
'skipping (every so many) ' we will cal l QUAS I -R EST R I CT I V ES. The number a nd variety
of these wi l l come as a surprise to most readers. The quasi-restrictive -zutu we wil l cal l the
D ISTR I BUTI V E ; it is given separate treatment, as is the somewhat similar COMPARATO R
yori 'than', discussed along with expressions of comparison i n §2.6. Certain q uasi-suffixes
or quasi-restrictives which serve to specify a plural or a col lective are considered in § 2. 7 ,
where you wi l l also find a d iscussion of devices avai lable t o express grammatica l number.
In § 2.8 the conjoi n i ng of nouns is descri bed, and the following three sections deal with
somewhat related phenomena i n attempting to account for nado and its synonyms, for
da ka, and for de mo. F i nal ly, § 2.1 2 d iscusses the pecul iar pseudo-restrict ive m itai.
I t is usually assumed that forms of the copula (such as da, na, no, n i, de, etc.) and the
various postnom i nal particles (such as ga, 0 , kara, made; -gurai , -do koro, etc.) are at­
tached to the noun to make a single phonological word, with no juncture between. That is
why some writers of romanized forms wi l l separate the noun from the copula a nd/or the
particles by a hyphen rather than a space. But under various circumstances a m i nor junc­
ture may be heard between a nou n and the fol lowing copu la or marker. A tonic noun
phrase is sometimes emphasized by optiona l ly inserti ng the minor juncture ( here noted by
a single bar) when a mar ker is focused with wa or m o : Nagasaki [ I 1 kara m o, Satoo san [ I ]
t o wa, Tyuugoku [ I ] -gurai wa, ... . The accent o f the marker i s suppressed when the junc­
ture is absent, except i n the i nstance of those particles that, as i nd icated by the raised minus
i n front of -gurai , cancel the accent of the attached noun-yieldi ng, i n this case,
/tyuugokuguraiwa/. Forms of the copula will sometimes be preceded by a juncture (or, in
written texts, a hyphen) when the noun phrase is a sentence that has been quoted or
34
§ 2.1 . Order of adjuncts 35
I

d irectly nomi na lized, as in this examcPJ le : Kono titi n i site, kono ko ari, NA no de aroo ka.
Sore to mo, kono ko n i site, kono titi1 ari , N A no de aroo ka ' I s it a matter of "Like father
l i ke son "? Or is it a matter of "Like son l ike father"?' (SA 2651 .72c) .
By ellipsis, forms of the copula wil l sometimes begi n a sentence, as described in § 24:
D a kara or De aru kara 'Therefore'; D a ga or D a kedo 'However'; D e mo or D '�tte 'Still'.
I n written Japanese, at least, it is possible to insert parenthetical material freely before
a marker : Sinzyuku-syo ( moto no Yodobasi-syo) NO mae o ... ' ... in front of the Sh i njuku
(formerly Yodobashi) Pol ice Station' (SA 2651 .24d) . Compare the i nsertion of parentheti­
cal material before col lectivizers mentioned in § 2.7.
Kuno ( 1 973) notices an i nteresting "stutter" effect which is sometimes heard when wa
or mo is added to a one-mora particle. I am u nfamil iar with th is phenomenon, but Hamako
Chaplin tells me that she has heard it. Apparently there is always a minor juncture i n the
surface form : Kyooto ni I ni wa = Kyooto ni wa 'in Kyoto ', N ara de I de mo = N ara de mo
'in Nara too' ( not i ntended for 'even bei ng in Nara') , otoosan to I to wa = ot�osan to wa
'with father', ... Perhaps the first particle is repeated in order to permit the juncture to
surface so as to lend extra emphasis to the fol lowing focus particle, which otherwise cou ld
not easily be set off by juncture. O r perhaps this is just a way of add i ng wa/mo as a quick
afterthought.
2. 1. ORDER OF A DJUNCTS

G ide ( 1 965. 1 07 - 8) has compared the Japanese sentence to the furoshiki, that marvel­
ous carryal l kerchief which wi l l expand or contract to just the size needed for the traveler
to carry his belongings-and which can be tucked neatly away when not in use. It is of l it­
tle concern that the contents may get rumpled in transit (or that the parts of t he sentence
may lose their overt signals of reference) , since they can always be pressed out at the end
of the journey (as the listener can i nfer the m issing marks of reference) . The E ngl ish sen­
tence, on the other hand, is l ike the unwieldy suitcase of the West-too big and too smal l
at the same time, cruel ly heavy, and cluttered with verbal coat-hangers piously designed
to keep the contents u nwri n kled to the very end of the journey.
M i kami ( 1 963a.66) describes the Japanese sentence as a dragon that is "many-headed
but sing le-tailed " (tatoo-tanbi) and that phrase is a good simile for the poi nt of v iew taken
by many Japanese grammarians who impute an equa l ity among the adjuncts of the pred i­
cate. In general , the view taken here is si milar, although we will have occasion to cal l atten­
tion to certa in facts wh ich seem to i nd icate that there is, after all, a kind of primacy en­
joyed by the u nderlying subject of the simplex.
Japanese is usually said to have a "free" word order with respect to the adju ncts. This
means that so long as you put the predicate (the nuclear sentence) at the end, where it be­
longs in a well-planned sentence, you are free to present each of the bu i ld-up phrases early
or late as you see fit. To be sure, such freedom will not always be ava i lable when we are
tal king of a sentence that is the resu lt of various conversions, rather than a simplex; thema­
tization, for example, wi l l place an adjunct at the beginni ng of the sentence ( § 3.9) .
But even within the simplex we can ask what makes the speaker see fit to choose one
order rather than another; we can a lso ask which orders are more frequent. From a study
of such questions we suggest that there is a neutral or "unmarked " order that is the nor­
mal or colorless way to present the information contributed by the adjuncts. Tentatively
36 § 2 . Predicate Adjuncts

we wil l assume that the neutral order is i l lustrated in this somewhat simpl ified d ia­
gram :
2 3 4a 5 6 7a 8
Time Place Subject R eciprocal I nstrumental Ablative Al lative Object
( locatives) N ga N to N de N kara N e N o:
(8a Affective)
4b 7b
(8b Traversal)
Nond irectional Direction- (8c Cathectic)
datives al dative
N nf N nf
As explai ned later, it is not clear that the time and place (either marked as locatives or
directly adverbialized) belong in the simplex, as such, but with respect to each other they
stand in opposite order from the neutral E ngl ish order, so that 'here and now' translates
into Japanese as f ma koko. Notice how a letter written in E nglish wi l l begi n with place
and date ; the Japanese letterwriter ends his letter with date and place, in that o rder.
By "subject" (or "nominative case") we refer to the use of ga to mark a noun as agent,
perceiver, cat hector (= emotion experiencer) , attributee, or identifier. Ga, of course, has
many other uses-e.g. , to mark the possessed in a possessive sentence, here regarded as
somethi ng more than a simplex to begin with, as explai ned in § 3. 5 and elsewhere. Simi lar­
ly, N to marks not only the reciproca l-considered an adj unct in the simplex-but also the
"comitative", here considered an abbreviation of N t�1 issyo/tomo(-) ni 'together with', i n
turn an adverbia l i zation of a separate nom i nal sentence that contai ns a reciprocal. It is
possible that the reciprocal should be given two positions in the neutra l order: d irectly af­
ter the object (thus final position) when the pred icate is a transitive verb; d irectly after the
subject (as above) with other pred icates. I have assumed that the ( i ntransitive) reciprocal
(4a) is i ncompatible with (4b) the datives of B EN E F IT (= no tame ni) , of CO N F R ONTA­
TION (= ni tai-site) , and of R E F E R ENCE (= ni totte) ; and that (7b) the dative of D I R EC­
T I O N (= -ate ni) is in complementary distribution with the al lative ( 7a) . On the several
kinds of dative, see § 3.4.
Data supporti ng the "neutral " order here posited wil l be found in a number of studies ,
notably M iyajima 1 962 and Saeki 1 960. ( I have followed Saeki rather than M iyajima i n as­
sumi ng that the dative precedes the accusative and that the ablative precedes the al lative.)
Other kinds of adjuncts occur. In add ition to time locatives ( "when it happened"),
there are temporal references of D U RATION ( "how long it happened ") and of F R E­
QU EN CY ( "how often it happened") . D uration is occasionally marked as if a traversa l
object; more often it is directly adverbial ized (san-zf kan aruku 'wal ks for three hours') ;
and sometimes it is stated as a conjoining of ablative and al lative phrases (san-zi kara roku­
zi made aru ku 'wa l ks from 3 o'clock till 6 o'clock') , as explained in § 3.7 and § 3. 7a.
You wil l probably never come across a natural sentence that contains all of the possible
adjuncts, clearly marked and i n the neutral order. There are at least two reasons. To begin
with, a Japa nese prefers sentences that are not overly expl icit; a few adjuncts are enough
to suggest what the others m ight be. And then in a typica l sentence the overwhelmingly
§ 2 . 1 . Order of adjuncts 37

common devices of focus and thematization serve to obscure (and often extrude) one or
more of the adjuncts.1
The order suggested above is the result of comparing the relative order of two or three
adjuncts in a g iven sentence. B ut natura l sentences are often more complicated than one
might think. Take, for example, this part of a longer utterance : ... konaida mo Oosaka
kara yogisya de onna-no-ko 0 turete, asa no go-zi -han -goro koko e tu ita h ito ga arimasu
'just the other day there was someone who arrived here with a l ittle girl at 5 : 30 in the
morning on the n ight train from Osaka ' (Tk 2.321 a). The first phrase konaida mo 'lately
even/too' is, I bel ieve, a h ighl ighted thematization of the time locative extruded from the
sentence ... h ito ga arimasu 'there is a person'. The subject of that sentence (hito) has been
epithematized from the subject (= agent) of the adnomi nalized sentence [ hito ga] Oosa ka
kara yogfsya de onna-no-ko 0 turete asa no go-zi -han -goro koko e tu ita [the person ] ar­ I

rived here with a l ittle girl at 5 : 30 in the morning on the n ight trai n from Osaka ', a sen­
tence with an adverbial phrase onna-no-ko 0 turete 'with a l ittle girl' that is the gerundi zed
form of onna-no-ko 0 tureru 'bri ngs along a l ittle girl ' and properly ( i.e. neutra l ly) belongs
at the beginning of the sentence; the remai nder of the sentence (Oosaka kara yog [sya de
asa no go-zi -han -goro koko e tu ita) has reversed our neutral order of instrumental and
ablative and has placed the directly adverbial ized time expression right before the a l lative.
We wou ld have expected the colorl ess version of the sentence to be: ... asa no go-zi -han
-goro [hito ga] onna-no-ko o turete yoglsya de Oosaka kara koko e tu ita. It is possible
that the ablative was placed before the instrumental u nder the i nfluence of such para­
phrases as Oosaka kara no yogfsya de 'by the night trai n (that is) from Osa ka'-as impl ied
by our original Engl ish translation-or Oosaka no yogfsya de 'by the Osa ka nig ht train'.
Since novel or critical information is saved for the end of a Japanese sentence, the reason
for delayi ng the time locative is perhaps to emphasize the i nconvenient hour of the arrival;
and the added outrage of draggi ng a chi ld along through the n ight may account for the de­
lay in expressi ng the adverbial phrase of accompaniment. B ut an a lternative explanation
wou ld treat the const ituents of the entire sentence as more equally conjoined : the sentence
[ hito ga ] Oosaka kara yogfsya de onna-no-ko 0 tureta (= turete klta) [the person ] brought
I

a l ittle g irl by n ight train from Osaka ' attached ( by way of -te 'and') to the sentence asa
no go-zi -han -goro [ kare-ra ga] koko e tuita ' [they] arrived here at 5 : 30 in the morning'.
U nder this i nterpretation the overt adjuncts are in neutral order except for the switch be­
tween ablative and i nstrumenta l . It is also possible that the speaker i ntended to thematize
the ablative, either to go with the adnom i nalized sentence end ing i n tu ita 'arrived ' or even
as a second theme to share with konaida mo the embeddi ng sentence that ends in arimasu :
'just the other day from Osa ka there was a person who .. .'. Close semantic t ies between ad­
junct and verb wi l l sometimes favor a neutra l order that differs from what is suggested
above, e.g. the i nstrumental in Tennoo o pisutoru de ute ! 'Shoot the Emperor ( with a pis­
tol ) ! I (SA 281 7. 1 01 ab) , where pisutoru de utu '(pistol-)shoots' functions as if a simple verb.
I f, as has often been suggested , we consider the spoken sentences as representations of some
u nderlying "natural logic" that is the result of a sort of "predicate cal cu lus", we wil l want to
consider all possible adjuncts as present in the u nderlying form of each simplex , once we have
decided just which simplexes lurk beneath our surface sentence in all its spare a nd murky
1 . I ncidental l y , M ikami ( 1 963. 70) suggests that the best way to d isentangle a long written sentence
is first to mark off the quotations, and then find the topics; I wou ld have thought the first order of
busi ness wou ld be a search for the conj unctional l i nks a nd the predicates.
38 § 2 . Predicate Adjuncts

glory. The omission of particu lar adjuncts has been compared to the process of "pronominal­
ization" in languages such as Engl ish, where typica l l y an argument (= a predicate adjunct)
once stated is later referred to by an anaphoric pronoun such as "it" ( "he", "she") , "they",
"th is", "that", or "those". But there is no compulsion to omit any given adjunct in a Japa­
nese sentence; nor does Japanese suffer from the severe strictures agai nst repetition of a
nou n phrase that English i nsists upon. A suitcase is harder to pack than a furoshiki.
2.2. MA R KING OF A DJUNCTS: NOUN POSTPOSI TIONS ("PA R TICL ES")
Some of the bu i ld-up phrases can be SPEC I F I ED by particles that narrow (or sharpen)
the grammatical re-lationshi p of the phrase to the rest of the sentence. Specification by at­
tachi ng particles does not affect the freedom of order; that is control led by relat ive em­
phasis ( § 2. 1 ) . The relationa l particles have a number of uses and meani ngs, so that it
seems best to start by summing them u p i n a l ist which i ncludes d isparate uses and mean­
i ngs that wi l l appear at d ifferent points in the grammar. 2 The l ist thus i ncludes not only
all of the specification markers but also some other markers (such as the essive ni that un­
derl ies the copu la and also u nderl ies certa i n other particles-de, no, and n i-as expla ined
later) which overlap in form and/or meaning. It is not clear to what extent the fol lowing
brea kdown of uses ca n be said to be grammatical as wel l as semantic, if i ndeed it is possi­
ble to draw such a distinction to begin with .
,

ga
1 . "d irect surface-subject" of a VE R B :
( 1 ) active agent = i nitiator of a process ( including kagu 'smel ls' and mfru 'looks') .
( 2) causative agent = instigator of a process; see § 4. 1 .
(3) the one affected by a mental process = he who perceives or reacts (with omou
'thi n ks/feels', m iru 'sees/regards', etc.) .
(3a) the emotiona l ly affected (= the cathectic subject, with a desire, l i ke, d isl i ke, or
fear; see § 3.5a ) .
(4) that which exists or is located (with aru, i r u , etc.-with stasis, § 3.8) .
(5) that which is possessed (with aru etc.) or quasi-possessed (with ir-u 'needs',
wa karu 'understands', etc.) ; see § 3.5.
(6) the recipient :
of a g ift (with morau, etc.) ;
of a favor (with -te morau, etc.) -the beneficiary;
of a disfavor (with the ADVE RSAT I VE PASSI V E , §4.2) -the maleficiary, the
disobl iged, the d iscommoded , the i nconvenienced.
( 7) the affective object of a transitive verb when it has undergone PU R E PASS I V E
( § 4.2) o r I NTRANSITI V I Z I NG R ESU LTAT I VE (-te aru , § 9.2.4( .2) ) .
(8) the cathectic object of a verb of emotion u nder the PU R E PASS I V E conversion.
(9) optiona lly replacing o for the d irect object of a verb under the D ES I D E R ATI VE
conversion.
2. The l ist shou ld be glanced over for future reference; unless a particular item catches his eye, the
reader w ill probably wish to move on to the next section, return i ng to consult this section from time to
time as he conti nues through the boo k. Under certain of the head i ngs you will f ind example sentences
and explanations wh ich d id not conveniently fit into later sections.
§ 2 .2 . Marking of adjuncts : noun postpositions ( "particles") 39

( 1 0) optional l y replacing o to mark the d irect object of a verb u nder the FACI LI·
TAT I VE conversions ( § 9. 1 .8) .
( 1 1 ) optionally replacing to to mark the reciprocal under the FACI LITATl V E con·
versions ( § 9.1 .8) .
( 1 2) optionally replacing n i to mark the dative u nder the FACI LITAT I V E conver·
sions ( § 9 . 1 .8) .
( 1 3) optionally replacing de or n i marking the locative (of place or time) u nder the
FAC I L I TATI V E conversions ( § 9. 1 .8) .
( 1 4) optionally replacing the usual markers of the allative ( 'to') and perhaps the ab·
lative ( 'from ') u nder the FACI L I TATI V E conversions ( § 9. 1 .8) .
( 1 5) optionally replacing the instrumental marker de u nder the FAC I L I TAT I VE
conversions ( § 9.1 .8) .
2. "ind irect subject" of a VE R B :
( 1 6) the possessor o r quasi-possessor (= n h .
3. "genitives" of various kinds-through el l ipsis ( § 3. 1 1 .2 - 3) :
( 1 7) genitive.
4. "d irect surface-subject" of an ADJE CT I V E or ADJECT I VAL NOUN (or PR ECOPU­
LAR NOUN ) :
( 1 8) attributee of a n attribute.
( 1 9) cathectic object of an adjective of emotion ; cf (9) above. See § 7. 1 , § 1 3. 1 .2.( 3) .
5. "indirect surface-subject" of an ADJECTI VE or ADJECT I VAL NOUN (or PR ECOP­
U LAR NOU N ) :
(20) cathectic subject of an adjective of emotion.
( 2 1 ) subjectival genitive-th rough ell ipsis, see § 3. 1 1 .
6. "d irect surface-subject" of NOUN + COPU LA :
( 22) the I dentifier o f an identificational sentence ( § 3. 1 0) .
( 23) the Identified o f an identificational sentence ( § 3. 1 0) under those conversions
(such as the provisional and certain adnominalizations) which neutra l ize the marking of
I dentifier and I dentified.
7. "orphaned subject" stranded by elliptical propredication ( § 3. 1 0) :
( 24) orphaned subject.
8. "emphatic locative" =
(25) the I dentifier of a time or place i n an elliptica l identificational sentence (see end
of § 2.3. 1 ) .
( 26) ima ga ima 'right now' (= tyoodo ima ) , ima ga ima-made 'up to this very mo·
ment'-usually treated as idioms; cf. p. 267.
9. antithesis-in tokoro/no ga 'but'.
1 0. S + ga 'but/and ' ( § 1 7. 5) ; by ellipsis Ga ... 'But .. .'
.

1 1 . I iterary tentative [ y ] oo + ga/to 'whether ... or . . ' ( § 1 7 .6) .

· .

1 2. hoo ga ( § 3.9, pp. 231 - 2) .

1 3. Sore ga . . . , Hayai ha nasi ga .. ( § 24) .
.

1 4. N n i site kara ga; V-te kara ga-see kara 8.

1 5. N ni site ga (= 1 4) -see § 9.2, p. 487.
1 6. [dialect] = no 3 (pronominal 'one/fact/it') -see Note at end of this section.
40 § 2 . Predicate Adj uncts
,

1 . direct object = the affected of a verb; see § 3.3: AFF ECT I V E object, CATH ECT I C
object.
2. place traversed, with quasi-i ntransitive (= motion) verbs, both those that imply tota l
traversa l (wataru 'crosses over', tooru 'passes by/through', etc.) and those that imply par­
tial traversa l (aru ku 'wal ks', tobu 'fl ies', etc.) : TRAV E RSA L object.
3. (= kara) place departed from (with quasi-intransitive verbs of leaving such as deru
'leaves', tatu 'departs', orlru 'descends from', etc.) : ABLAT I VE object.
4. time spent: TEMPO R A L object, as in Tookyoo de issyoo o kurasu 'l ives all one's l ife
in Tokyo'; nanazyuu-nenda i o iklru ' l iving ( in) the seventies' (SA 2660.58-headi ng ) , Tuki
e mo ... ryokoo n i i keru yoo na zidai o i kiru no da kara 'For we l ive i n an era such that we
can travel a l l the way to the moon' (CK 985.372) ; N iti-yoobi o iti-ni·l?neta ' I stayed i n bed
the whole day Sunday'.
5. "orphaned object"-stranded by el l ipsis ( § 9.1 . 1 2, § 1 4.3, § 28; § 3. 1 0, § 1 3. 1 . 5) :
lti-kiro no miti o [ ] yonzyuu go-hun mo kakaru hazu ga nai 'Covering a one-kilometer
route shou ldn't ta ke a whole 45 minutes! ' (SA 2642.32d) -the ellipsis is something l i ke
[aruku no n i wa] 'to wal k'; Yu�i 1 no naka o [ ] doo-mo osore-irimasita 'Thank you for
coming [to officiate at the wa ke] in the midst of the snow' (Ariyoshi 34) .
6. antithesis: S + mono-a, § 1 7.4; S + no o, § 1 4.2; Sore o, ... 'Despite that .. .' (= Sore
na no ni) .

0. essive ( 'being') = copula i nfinitive, § 9.1 . 1 1 , a pred icator of nou ns; those uses marked
E below can be treated as th is, as probably a lso can those marked ( E ) .
1 . i nd irect object (with verbs o f giving, i nform ing, etc.) ; DAT I V E O F BEN E F IT 'for' =
[ no tame] n i 'being for the sake/case of'; DAT I VE O F CO N F RONTAT ION 'toward, with
respect to' = ni [tai-site] 'being opposed to, confronted with '-as in haha ni sinsetu da 'is
kind to mother', sa ke ni yowai 'is easi ly affected by drink', kei ken ni tobosli n 'is wanting in
experience', kei ken ni tomu 'abounds in experience'. See § 3.4; for the DAT I V E O F D I ­
R ECT I O N , see 5 below.
1 a. pseudo-reciprocal use of dative of confrontation ' [coming in contact] with' (N ni/to
+ au 'meets', wa kareru 'parts', n iru 'resembles', kuraberu 'compares', ... ) -see § 3.6.
1 b. objective stimulus; reason, cause (cf. de 6, kara 5) bi nboo n i kurusfmu 'suffers from
poverty' (cf. binboo de kurusfmu 'is d istressed at one's poverty') , byooki ni nayamu 'agon­
i zes over being i l l ' (cf. byooki de nayamu 'is afflicted with an i l lness') , kodomo ni odoroku 'is
surprised at the chi ld ' (cf. kodomo de odoroku 'is surprised by the chi ld ') , kodomo ni ga kkari
suru 'is disappoi nted in the chi ld ', te-busoku ni komaru 'is embarrassed by a shortage of per­
sonne l ', (ame de) miti ni mayou 'is confused about the road (owi ng to the ra in) ', kosame ni
nureru 'gets ( pleasantly) wet in the shower' (cf. ame de nureru 'gets drenched from the ra in') ;
Mono no neagari ni kurasi-nf ku ku natta 'With the rise in prices of thi ngs l ife became d ifficult'
(cf . ... de ... 'Due to .. . ') ; I naka-mati no tetudoo-in no mf zime na sei katu no saimoku [sic] ni,
ka nkyaku wa warau 'The onlookers laugh at the petty deta ils of the wretched l ife of a rai l road­
man in a country town ' ( lg 1 962.70) ; sore ni 'by that, for that reason'; cf. § 1 7. 1 V-ru kara ni
wa, V-ta kara [ n i ] wa.
3. Sometimes contracted to n [ i ] before t, d , or n : kore n [ i ] tomonatte, soto n [i] deru, byook i n [i ]
naru . The particle sequence n i w a i s sometimes pronounced nya [a ] , a s in Yasumi n o h i nyaa , . . . 'On
days off, . . .' ( BJ 2 .1 39 . 1 7) . Accord i ng to Jorden ( BJ 2 . 1 4 1 n 1 7 ) th is is more common i n the speech of
men. In such ph rases the kana spell ing " niya" is to be interpreted as nya [ a ] .
§ 2.2. Marking of adjuncts: noun postpositions ("particles") 41

1 c. '(dependent etc.) on' with yoru 'depends/rel ies', sitagaun 'conforms (with ) , is conse­
quent (upon)', motozuku 'is based ', ... . ( I s this from 5?)
1 d. 'from/by' an agent (= kara 4, but with subtle differences of nuance requ iring fur­
ther study) , as in sensei ni/kara piano 0 narau/osowaru 'learns piano from a teacher,
studies piano with a teacher', hito ni/kara nyuusu o kiku 'hears the news from others',
tomodati n i/kara hon o kariru 'borrows a book from a friend ', haha n i/kara tegami o
morau 'receives a letter from one's mother'. See § 1 0 (favors) , § 4.2.
1 e. 'by = at the hands of': marks the u nderlying agent i n a passive conve(sion, § 4.2
(can also be expressed by n i yor i ; cf. kara) .
1 f. 'by = at the i nstigation of': marks the i nstigator of a passivized causative.
1 g . marks the u nderlying agent of a causativized conversion.
?1 h. pseudo-agent 'by/at', with tutomeru 'works for, is employed by'-ginkoo ni
tutomeru 'is employed by the bank', cf. N [ no tame] n i tutomeru 'endeavors for N' (8) , ... .
2. ind irect subject : possessor or quasi-possessor (= ga) ; see § 3.5.
2a. [ now obsolete] a highly exalted subject : Tennoo-heika n i wa nizyuu iti-n i ti 1sitl-zi
nizip-pun kyuuzyoo o go-syutumon 'His Majesty the Emperor left the pa lace at 7: 20 on
the 21 st' ( Nagano 1 970. 1 83) .
3. specific time 'at': see § 3, § 9.1 . 1 3.
4. static location ( LOCAT I V E ) 'at' (with aru 'is', sumu 'lives', iru 'stays'; motte iru
'possesses', etc.) ; see § 3.8 for special problems.
5. DAT I V E OF D I R ECTION = [-ate] n i 'being aimed at'; target 'to', d irect ion 'toward'
(= el .
5a. change of position (M UTAT I V E- LOCATI V E ) 'onto, i nto, . . .'-probably t o be treat­
ed as an automatic reduction of locative n i + mutative use of the essive n i (6) , 'so as to be
at/on': with noru 'boards, mou nts', o ku 'puts, places'; perhaps tasu 'adds (to) , kuwaeru (-)
'adds (to) ', kasaneru 'pi les on', itaru(-) 'arrives (at) , reaches', sugiru ' exceeds', kosu 'ex­
ceeds', ... .
E 6. change of state (M UTATI VE) 'i nto', 'so as to be (someth i ng new)': with kawaru 'it
turns i nto', kaeru 'turns it into', ... , and especial ly naru 'becomes', suru 'makes it ( i nto) '­
see § 9. 1 . 1 1 . B ut the fol lowing expressions are perhaps better treated as examples of 8 ( pur­
pose) : onna 0 hisyo ni tanomu 'asks (= hires) a woman to be one's secretary'; heya 0 zimu­
syo n i kasu 'lets a room out as an office'; mono 0 miyage n i kau 'buys a thing for a souve­
nir'; Amerika o rei ni toreba 'if we take America as/for an example'; ... . (The adjuncts are
freely permutable.)
E 6a. '(treated ) as' (PUTATI V E or EVALUAT I V E , § 9. 1 . 1 1 ) .
( E ) 7. antithesis: S + no ni 'despite that S'; see § 1 4.2.2.
(E) 8. purpose 'for' = no tame [ n i ] (see § 1 3.2, § 9. 1 . 1 3, cf. § 1 7. 1 ) : iwai (o-iwai) n i 'as
a celebration, in order to celebrate'; V-i ni i ku 'goes to V', VN [si ] ni i ku 'goes to V N '
( § 9. 1 . 1 ) . Does taberu no n i y o i ' i s good for eati ng' belong here? See also 6 above; § 1 4.2.2.
E 9. appearing to be (= da to) : with omowareru, mieru , etc.-see § 9. 1 . 1 .
E 1 0. manner '-ly, -wise': see § 9. 1 . 1 1 .(3) .
( E ) 1 1 . enumerative 'and ': see § 2.8 on conjoini ng. Cf. Sore ni 'And (then/a lso) '.
1 2. See § 1 4.6 (directly nom i nal ized S + ni tigai nai, n i sugfnai, ni kimatte iru, ni kagiru,
ni koto-ka ite) ; § 1 7.8 (daroo ni) ; § 1 7. 7 ( l iterary tentative + ni with several meanings) ;
§ 9. 1 . l a ( V 1 -i n i V 1 -ru) ; ... .
( E ) 1 3. DAT I V E OF R E F E R ENCE = ni ltotte] 'taking it as being (with reference to) ' �
'for' (cf. 1 ) : Anata n i tiisa-sugiru kara ... 'Since it is too sma l l for you . . .'; Tosiyori n i wa
muzukasii desyoo ga, wakai hito n i wa heiki desu 'For old people it's probably too hard
42 § 2 . Predicate Adjuncts

but for young fol ks J t's nothing at all'.

?E 1 4. N n 1i wa n i-syu aru 'There are two kinds of N ' ; H itotu n i wa ... (moo h itotu ni
wa ... ) 'For one ... (and for another ... ) '; ... hana no sukf na h ito ni wa akunin wa nai ...
'there are no evildoers among those who love flowers' (Tk 2.64a) ; Goka i ni wa iroiro na
mono ga aru ga ... 'There are various kinds of m isunderstandings .. .' (Nagano 1 966.89) ;
Sibusawa san ni n ite 'ru hito n i , Yamada Koosaku san ga aru 'Someone who resembles Mr
Sh ibusawa [among the others who resemble him] is Mr Kosaku Yamada' (Tk 4.21 ) ; Watasi
no tizinn ni Q [ Kyuu] to i u h ito ga iru 'Among my acquaintances there's a man named
Q' (Nagano 1 966.75) ; Hurui zfnzya ya tera ni wa utukusfi tatemono ga ooi 'There are many
beautifu l bu i ld i ngs among the old shrines and temples' (this ni cou ld also be interpreted as
locative or possessor) . See pp. 25 1 - 2.
E 1 5. the CO R E LATI ON A L M UTAT I V E (465-6) : yuumei-zin o yuuzi n ni motu 'has a
celebrity for a friend', ... ..
1 6. Problematic: genki ni ahurete iru 'is overflowi ng with pep', kiboo ni m ltite iru 'is
fu l l of hope' ( ? 1 b) ; [no yoo] ni tatu 'is useful' ( ? 4 metaphorical) ; go-sankoo made n i 'for
your reference/i nformation' ( ? 8) ; is-syuukan [-goto] ni iti-do 'once a week ( ? 0, ? 3) ;
zyuu-nin ni h itori wa 'one out of ten persons' ( ? zyuu-ni n [ no utfn J n i [aru ] h itori 'one
[who is] in [the m idst of] ten') ; mizu ni oboreru 'drowns in water' ( ? 4, ? 1 b) ; zyuu-zi no
basu ni okureru 'is late for the ten o'clock bus' ( ? 1 3) ; h ito-i ki ni nomu 'drinks it at a gulp';
kam i n i tutumu 'wraps it i n paper' (cf. kamf n i kaku 'writes it on paper' 5a) ; kyoosoo ni
katu/ma keru 'wi ns/loses at the competition' (? 4 metaphorical) ; N n i husawasli 'is
su itable for/as N = ma kes a nice N' (? 6, ? 9, ? 8) ; me n i mfete 'visibly, remarkably', me ni
mienai to koro de 'in a place i nvisible to the eye' (SA 264 7.1 1 9e), me n i/de mienai hodo
tiisai 'so small it is i nvisible to the eye' (me ga mienai 'the eyes cannot see = is sightless'
wou ld be used only of a bl i nd person) ; Sono su kaato wa . .. gaisyutu nf mo hakemasu 'I can
wear that skirt ... for steppi ng out, too', Nan ni tukaimasu ka 'What wi l l you use it for?'
(? 8) ; ... tizyoo gozyuu-meetoru no taka-sa made zyoosyoo suru no n i sei koo sita 'succeed­
ed in rising to a height of 50 meters above the earth (= off the ground) ' (SA 2647.4) .
E 1 7. odoroita (etc.) koto ni 'to my su rprise (etc.)'; sinai koto n i wa 'unless' ( pp. 396,
552) ; dare no syookai ni 'at whose i ntroduction'.

de
0. [< nf-te] = gerund of copula or essive, § 9.2. The uses marked E below can be treat­
ed as this. But it has been suggested that only those uses for which the pol ite desite can be
substituted ( § 22. 1 ) are appropriately considered the COP U LA gerund as such.
1 . general locative of place : dynamic location '(happening) at' (with verbs) ; location of
a scope of reference (with ad jectives, etc.) . Cf. kara 9.
2. materia l '( made) out of' = kara 3 (but preferred to kara in speech) : kam i de tukuru
'makes it (out) of paper', kl de dekite iru 'is made (out) of wood'.
3a . means, i nstrument 'by, with, using': o-hasi de taberu 'eats with chopsticks', enpitu
de kaku 'writes in penci l '.
3b. vehi cu lar means '(rid i ng) on (a vehicle) ' = ni notte : kuruma de i ku 'goes by car'.
Cf. kara 1 0.
3c. means of communication or information 'through (the medium of) ' = o tuu-zite:
si nbu n de yomu 'reads it i n the newspaper', razion de kiku 'hears it on the radio', terebi
de m iru 'sees it on television'; eigo de hanasu 'speaks in E ngl ish', kana de kaku 'writes i n
§ 2.2. Marking of adjuncts : noun postpositions ("particles") 43

kana'; kookuu-bin de okutta 'sent it (by) air mai l '.

3d. pseudo-agent, with passive : denwa de okosareru 'is awakened by the telephone';
N i ppon wa mawari o umi de ka komarete iru 'Japan has its circumference surrounded by
sea ' ( lg 1 962.72) . Or is th is cause (6)? Cf. § 4.2; Alfonso 950 - 1 ; BJ 2.306.
?E 4. i mpers � nal ( institutional , group) su bject-always agent? : Wa-ga sya [= sinbu n­
sya ] de wa sikyoku n no koozin o iso ide iru 'Our newspaper expedites notices from
branch offices'; Keisf-tyoo de happyoo sita no to kui-tigatte iru 'There is a discrepa ncy
with what the M etropol itan Pol ice Office publ ished '; Too yakkyoku de zisin o motte
o-susume si ma'su 'This [= Our] drug store recommends it to you with confidence'. Cf.
Alfonso 991 , where de is said to be used more frequently than ga when the su bject is an
institution or moral entity ; the examples given are H itati de uri-hazimeta ... 'H itachi has
begun sell i ng .. .', Kimi no utin de ... katta zidoo-syan 'The car your fam i ly bought .. .',
Ano kaisya de ... kooi n o atumete iru 'That firm is gatheri ng workers . . .', Gakkoo de
meirei suru koto . . . 'What the school is orderi ng .. .', Ano m ise de kookoku o dasite imasu
'That shop is putti ng out ads'. Yoshida speaks of "collaborators" and (Y 399) gives the
examples: Utin -zyuu de hisyo ni iku 'The whole fam i ly flees the summer heat', San­
nin de utyuu-ryokoo suru 'Three men go on a space trip'. Cf. 9 (exclusive agent) .
?E 5a . 'within ' a time or l imit: iti-nitl de dekiru 'gets it done i n one day', yaku nlzyuu
go-hun de owatta 'fin ished i n about 25 minutes', wa'zuka mikka de modotte klta 'returned
in a mere three days', Ano nan-zlken o is-syuukan de kai ketu sita 'They got that d ifficu lt
matter settled i n a week'.
?E 5b. 'by' (a ti me) : g0 -zi de owaru 'it will be over by five o'clock' (made n i or just n i
would b e more com mon) .
E 5c. made de 'doing t i l l then': see pp. 489- 90.
E 5d. Kotosi wa n f-nen -buri de Toodai no si ken ga atte, ... 'This year for the first time
in two years Tokyo U niversity [ entrance] examinations are taking place, and . . .' (SA
2673. 1 6a) .
E 5e. N ippon-zikan de II kesa I mimei I ... 'Before dawn this morning, Japa n time, . . . '
( R ) ; Otooto wa sanzis-sai de osanai k o t o t u m a o nokosite, kootuu-zf k o de sokusi
simasita 'My younger brother at the age of thirty was ( instantly) killed i n a traffic accident,
leaving beh i nd wife and small chi ld' (SA 2684.1 23a ) -the second de is cause (6).
E 6. cause or reason (= no tame ; cf. ni l b) : soko de 'for that reason'; byooki de
yasum u 'stays home with i l lness'; N an de sinda daroo 'I wonder what he d ied of/from ';
hisyo-ti de sirarete iru mati 'a town known as/for a summer resort'; kane de komaru 'is em­
barrassed by the money' (cf. kane n i koma'ru 'is embarrassed for [= lacks] money') ; e igo de
h ito n i sugureru 'surpasses others i n E ng l ish ' ; S no de, § 1 4.2. 1 ; ti de yogorete iru 'is
stained with blood', a'se de nurete iru ( bissyori da) 'is wet (is drenched) with sweat'.
E 7. san-bon de zyuu-en 'ten yen for three (pencils, etc.) '; nama de taberu 'eats it raw';
m i nna de i koo 'let's all go together' ( ? 4, ? 9) ; anata no kangae de wa 'in your opinion';
i ma [ no zyootai/kotoba] de ieba ... 'Put i nto today's terms . . . ' (SA 2672. 1 1 Sc) ; Sore ga,
G i nza -atari o tuke-matuge de aru ite 'ru zyosei ga, kono-goro(-) i kura mo i ru 'But, lately
there are ever so many women stroll i ng around the G i nza in false eyelashes' (Tk 2. 1 03a) .
E 8. Kore de o-simai (da) 'Th is is a l l ; This is all that's left'; Siken wa kyoo de o-simai
desu 'Examinations are over with today'; Moo sukosi de wasureru tokoro datta ' I was just
about to forget'; Z i kan de harau 'We pay by [accordi ng to] the hour'.
E 9. exclusive agent: Z ibun de i u no mo hen da ga 'For me to say it is odd = Even
44 § 2. Predicate Adjuncts

if I do) say so myself'; Syuuzen wa II kariru hoo de II sinakute mo II ii desyoo ,I ne 'The

repairs wouldn't have to be made by the tenant, wou ld they?' ( BJ 2.20 1 .23) ; I i sigoto wa
minn�1 zibu n-tati de tori, wareware ni wa saitei no sigoto sika sasenakatta 'Al l the good
jobs they too k themselves, and wouldn't let us do anything but the most menial jobs'
(SA 2678.42e) ; Atasi-tati de ato de tabemasu kara ' [ Don't worry-] -we women wil l eat
it later' ( Kb 45a) ; Katyoo made wa onna de zyuubun deklru to omoimasu 'Up to the job
of section head a woman can get along quite wel l , I think' (SA 2672.62c) ; H itotu no
hakama o oya to ko de tukau, omosiroi kyoogen desu ne 'It is an amusing farce in which
father and son use a single hakama (formal skirt) ' (SA 2659.49b) -perhaps to be taken as
4 (group su bject) ; H utari de i koo 'Let's go just the two of us'. More exam ples wil l be
found in V 1 967b.44. Cf. Alfonso 992, who speaks of "exclusive amount" but gives as
examples only agents.
to
0. subjective essive4 ' [thinking it] to be': § 1 3.5a, § 2 1 .4, § 21 .7; for quotative uses, see
§ 21 , § 1 7.6, etc.
1 . 'and ': see § 2.7.
2. R EC I P ROCA L '(reciprocal ly) with' (cf. nl 1 a) ; 'from' (N to tigau 'differs from N , is
other than N ') ; 'with = agai nst' (A ga B to tata kau/arasou 'A fights/struggles with B ', A ga
B to ronsoo suru 'A argues with B ') . See § 3.6.
3. COM ITAT I VE N to [ issyo/tomo( -) ni] '(together) with N '. See § 3.6.
4. s + to 'when, if, . . .': § 1 7.2.
5. S + to '(saying/th inking) that S' (= 0.) : § 21 .
no 5
1 . 'of'-see § 1 3.4 ( possessive nomina ls) , § 3. 1 1 .3.
2. (subdued) subject i n adnominalized sentences-see § 1 3. 1 . 5- 6.
3. pronominal 'one/fact/it' etc. in nom i nal izations-see § 1 4.2, § 1 5. 1 3; also § 1 5. 1 7.
( H istorically, a truncation of mono?)
4. '(which) is (a case of) I = na (from ni [-te]C aru) in certai n adnomi nalizations of the
copula ; see § 1 3. ( H .1storical . o f ndi aru ( )
. ly, a reduction I ""

5. an obl igatory reduction(1of n�1 no (4 + 3) 'the one which is' or ( if the derivation is
.
not u ltimately the same) of no) noI ( 1 + 3) 'the one of (or belonging to) ... '. •

kara 6
1 . SO U R CE '(starting) from': a PLACE ( ko ko kara 'from here') or quasi-place ( kokoro
4. The d istinction of subj ective essive ( to ) from objectiv� essive ( n i l w i l l be d r�wn r � peated ly in this
book. Cf. Ki nosh ita 21 : "10 wa gai ken-tek i ryoo-teki de, NI wa honsitu-teki de aru 'TO is phenomeno­
log ica l and quantitative, NI is intrinsic'. I use "subj ective" to refer to what is subject to variations in in­
dividual perception rather than j udgment; I use "objective" for that which is independently j udged (or
intersubjectively verifiable) . Japanese g rammarians (e.g . Mabuchi 2 1 2 ) sometimes use the correspond ing
Japanese terms syu kan-teki 'subjective' and kyak kan-teki 'obj ective' in precisely the opposite way .
Caveat lector.
5 . Sometimes contracted to n [o ) . especially before t, d , or n : ore n [o] toko 'my place', boku-ra n [ o]
toko 'our place' ( A riyosh i 20) . And si nce /nn/ automatical ly reduces to a single /n/ al l trace of the con­
tracted n [o] is lost in Tatibana san [ ] toko wa 'Mr Tachibana's place' (Ariyosh i 71 ) . No wa sometimes
shortens to naa; N tefet naa "" N to iu no wa (an example in O kitsu 1 .96) .
6 . For some speakers prototonic kara . The particle is also pronou nced kkara (or kkara ) : mukasi kkara
§ 2 .2 . Marking of adjuncts: noun postpositions ("particles") 45

kara ai-suru 'loves from/with one's heart') ; a T I M E (kore kara 'from now on', kinoo kara
'si nce yesterday') ; an O R D E R in sequence (go-ban -me kara 'from the fifth one') ; a SUM
(sanzen-en kara suru 'it costs at least 3000 yen') or other QUANTITY (hya kuman-n in
kara no kankoo-kyaku 'over a m i l l ion tourists') ; a V I EWPO I NT : kazu kara i u to ( kazu kara
ieba) 'when it comes to numbers, from the standpoint of numbers, i n terms of number';
kanzya-suu kara ittara 'spea king from the poi nt of numbers of patients' (SA 2649.39d) ;
bunsyoo kara site (suru to, mfru to) 'judg i ng from, the (sentence) style'; kono koto , , kara
wa karuI
yoo • •..
. , 1 as1,. b etu no k a' k u d o kara k angaeru
( •) ni 'A s 1s c I ear f rom t h "1s fact ... ; S"k to, ...
'But, when considered from a d ifferent ang le, .. . ' (Tanigawa 1 39) . Notice that where
Engl ish prefers to begi n AT a time, Japanese prefers to begi n F ROM a time: yo-zi kara
hazimaru 'it starts at four'. Perhaps this accounts for yuugata kara deka keru 'wil l go out
toward evening' ( BJ 2.45) . Some uses seem a bit obscure : ... konniti kara wa yomenai
hodo ... 'to the extent that they cannot be read ( by us) today' (Ono 1 966.205) ; ... syoki
no katakana o yomu koto wa, konniti kara wa nakanaka muzu kasii no de aru 'reading
early kata kana is extremely d ifficu lt (for us) today' ( ibid.) . On kara used for ga to mark
the agent of "verbs which express the idea of TE L L I NG or I N FO R M I NG "-hanasu , tyuui
suru, tutaeru(-) , tegami o okuru, denwa o kakeru, etc.-see Alfonso 993. Vardul ( 1 967.42) ,
noti ng that the pred icate must be affirmative, i nterprets certai n of these examples of kara
as ' (doing it) oneself': Anata ga ienai to iu nara watasi kara kotowatte yarimasyoo ' I f you
ca n't say it, I will refuse h im myself'; Okaasan n i wa ore kara hanasite okoo ' I 'll let
mother know, myself'. But when the act need not rely upon the vol ition of either party, it
is possible to use both negative and affirmative predicates ( ibid.) : Teki kara no koogeki
mo nai kawari ni kotira kara mo koogeki suru tikara( - ) 0 usinatte ita kara de aru ' I t is be­
cause, to make up for the lack of any attack from the enemy, our side had also lost the
strength to attack'; Watasi kara nakigoto o ki kaseru no sae, hu-yukai ni omowareru desyoo
'You even seem unhappy that I let you hear me weep'. To Vardul ' s examples we can add :
... yoozi ga attara zibun no hoo kara iku ... 'when there was some business to attend to
THEY wou ld decide when to go do it' (SA 2642.44d) ; Watasi kara i i-tai koto wa ... 'What
I want to say is .. .'. B ut these may be extended uses of the "set-opener ablative"
(§3.7a) : Anata kara itte kudasai means both 'YO U say it [for us] = YOU do the talking'
and 'You tel l it (your way) first'; Watakusi kara hanasite mimasyoo ka means 'Shall / try
speaking [for you/us] ?' or 'Shal l I be the first to spea k? = Shal l I say my piece first?' Ac­
cordi ng to Vardul the predicate for the set-opener ablat ive must be affirmative :
"Arukimasyoo" to, M itiko kara aruki-dasita 'Say i ng " Let's walk", M ich iko was the first to
start wal king ' ( V 1 967.42) .
2. '(leaving/removi ng) from' a place-cf. o 3.
3. (= de 2) '(composed ) of' (X to Y kara natte iru 'is composed of X and Y ') , material
'( made) out of'. Notice that Danta i wa A to B (to C) kara natte iru 'The group consists of
A a nd B (and C)' amounts to the same thi ng as Dantai nf wa A to B (to C) ga aru.
4. (= nf 1 d) 'from/by' an agent. Hun '[ki kara nani ka ga umareru 'Someth ing is hatched
by the atmosphere '-or is t h i s 5?
'from way back ' (Tk 3.35a, Kb 389a ) ; hazime kkara 'from the start' ( Maeda 1 962 .209 ) ; ima kkara
'from now on' ( K b 1 4 1 b) ; kondo kkara 'from this time' ( F n 1 3 1 b, Kb 329a) ; kono-aida kkara 'from
not long ago ' ( Kb 1 40a ) ; asa k kara 'from morn ing' ( Tk 2 .307a, Zhs 2 . 1 79 ) ; o-h iru kkara hazimatte 'be­
ginn ing at noon'. Cf. -te [ k ] kara, § 9 .2.3. The vowel of a preced ing k-syllable (especially k u ) sometimes
d rops in rapid speech : . . . gaikok [ u l kara no I at� ryoku desu I nee ' it's pressu re from abroad, you see'
( R ) ; kok [ o ] kara wa 'from here on'.
46 § 2 . Predicate Adjuncts

5. (= de 6, no tame ni) cause, reaso n : R yooke no syakai-teki tii no mondai kara kono
kekkon no hanasf wa hadan ni natta 'The marriage ta lks were broken off because of the
(difference in) social position of the two famil ies'; K isy�1 no tukare kara sugu nemutta 'I
fel l right to sleep with weariness from the train'.
6. S + kara-see § 1 7 . 1 ( 'because') .
7. V-te + kara-see § 9.2.3 ( 'after doing') .
8. N ni site kara ga/mo/sae = N de sae mo 'even (being ) N ': Sensei n i site kara ga/mo/
sae, kyoositu de nondari tabetari suru n da kara, gakusei ga gyoog i ga waru ku n aru no mo
motto-mo daroo 'When even the teachers are eating and dri nki ng in the classroom, we can
surely expect the students to worsen i n deportment'.7 I n these expressions, ni site is a book­
ish equ ivalent of the copula gerund nf-te > de; the expressions with mo and sae are thus
para l lel to de mo and de sae. But the synonymous expression with ga is not easi ly explained ;
cf. l iterary V-te kara ga = V-ta to site mo 'even if, say' (Meikai kogo jiten 260c) .
9. [dialect-e.g. Tottori ] dynamic location (= de 1 ) .
1 0. [dialect] vehicular means (= d e 3b) . 8
made
1 . 'to, up to, u nti l ' a place or time or order; made n i 'by' (a time or order) , 'so as to
betill a time or a place' ( mu tative § 9. 1 . 1 1 ) ; for other uses see pp. 2 1 0, 2 1 2. On made n i
* made de * made, see p p . 489- 90.
2. 'even'; nf made 'even at/to/etc.' See § 2.3.3.
,

e
1 . target 'to', direction 'toward'. This is replaced by nf for many speakers; but note
that when the sentence is adnominal ized, nf drops : umi e no m iti 'a road to the sea' =
*umi ni no miti � umi no miti [ambiguous] ; E ki e no miti o osiete kudasaimasen ka
'Please show me the way to the station'. Cf. discussion on pp. 207 -9.
2. In downtown Tokyo e replaces n f in a number of uses, e.g . : Go-syuugi o kami e [=
ni] tutunde ... 'Wrapping the gratuity i n paper . . .' (Okitsu 1 .268) ; si nbun ni/e ni deru 'ap­
pears in the newspaper', ensoku n i/e i ku 'goes on a picnic', kokuban n i/e kaku 'writes it
on the blackboard', naka ni/e i reru 'puts it i nside', ue ni/e oku 'puts it on top', yama ni/e
noboru 'cli mbs on the mountain', hito n i/e kasu 'lends it to people', hito ni/e ageru 'gives
it to people', goruhu ni/e i ku 'goes golfing' (but not in goruhu o si ni iku 'goes to play
golf') . A sentence l i ke Anata n i/e agemasyoo ka cou ld , however, be taken as ellipsis for
Anata e/ni [ watasite] agemasyoo ka 'Shal l I hand it over to you ?' Simi lar cases: Kotira
e/nf mo kootya o kure na 'Let me have some tea too', atira-sama e/ni mo ... 'him too',
Tana ka-san no hoo e/ni mo ... 'let Tana ka have some too'. A more surprising case:
Atama mo yokatta si, ginkoo e [ = nf] de mo tutomete ireba, katyoo -kurai n i wa

1 . The only authentic example I have cul led is a bit mu rky i n meaning : Dai itl-mon ni site kara ga,
zyoonai no teizi-ban [sic ] de toppu 0 slmete i ru no wa, nizyu u-nen -go n ni z itugen, [ suru no] de aru
'Even the very first questio n, what is occupying the top of the d isplay board in the room is [something ]
t o b e realized after 2 0 years' ( S A 2678 .37c-the comma i n t h e origi nal text i s to mark t h e ell ipsis shown
in b rackets in the romanization ) .
8 . An example o f hune kara maitta 'came b y boat' is found i n D iego Collado's Ars G rammaticae
Japonicae Linguae ( 1 632) , accordi ng to l sh igaki Kenj i 1 83; this is presumably a Kyushu form.
§ 2 .2. Marking of adjuncts : noun postpositions ("particles" ) 47

natta desyoo 'He had a good head on his shoulders, too; i f he'd worked for a bank, say,
I bet he would 've become a section head at least' (SA 2665. 1 18d) .
The source of some of these morphemes is u nclear, though I have elsewhere ( Martin
1 968) suggested Korean cognates for the essive n i and the subject marker ga, relating the
latter to the morpheme ka which serves both languages as a postadnom inal noun meaning
'the question of ... ', an i ntermed iate stage perhaps bei ng -n-ka with the adnom i nal marker
-n- that is found in a number of the languages of northern Asia .
Each of the particles to and no is probably a convergence of several d ifferent etyma.
Thus some of the uses of to ( 'with '-perhaps 'and') are likely to be cognate with the Alta ic
comitative case marker, despite the apparent lack of a descendant from that etymon i n
Korean; other uses (the subjective essive, 'th inki ng/saying that . . . ') may b e the deictic to =
so 'that'-there are cognates i n the Korean deictics-or developments from the focus mark­
er -do 'even/indeed (being ) ', for which we cite R yukyu and Korean cognates i n § 2.3.4;
while sti l l other uses ( 'when/if') may be tru ncations of tokf 'time' and/or toko [ro] 'place,
circumstance', both of which enjoy Korean cognates.
Some uses of no ( 'the one wh ich .. .' etc.) may resu lt from a shortening of mono 'thing
(etc.) '; one use (as subject marker) remai ns obscure i n origin though it also may be from
mono ; the adnom i nalized forms of the copu la no/na are variant reductions of naru = ni aru
(attributive form of the literary copula, made up of essive + auxil iary 'be') ; other uses, i n­
cludi ng the genitive 'of', resu lt from el l i psis i nvolving one or more of the precedi ng forms,
as expla ined in appropriate sections of this work.
The particle e < pe is usually assumed to have developed from the noun he < pe 'vicin­
ity' ( now found mainly as a suffix -be) , sometimes taken to be an early loan from Chinese
but perhaps to be ascribed (along with herf 'edge'?) to the verb he-ru < p(e) - 'passes by'.
The particle o < wo is obscure in origin. An emphatic object-marker ba which occurs i n
the R yu kyus and a lso i n northern Honshu suggests that wo may have come from ba, by
assimi lation of the vowel to the initial labial ; it has been speculated that wo and the focus
particle wa < pa are divergent developments in the same etymon.
It is reported that e is someti mes pronounced we i n d ialects (e.g. Nagano, Zhs 2.4 73)
and o is sometimes pronou nced wo .
The etymology of made is discussed on p. 1 37 ( § 2.4) . The origin of kara is a puzzle. I
know of only one good suggestion, that described in Ono 1 966. 1 81 , accordi ng to which
the particle is related to the morpheme -kara that appears in ya-karan 'tribe' and hara-kara
'siblings' 9 and to the morpheme -gara that appears in tomogaran 'group of fellows' a nd i n
kuni-gara 'national character' a n d h ito-gara 'personal character' -with gara 'pattern, charac­
ter' itself to be considered a tru ncation?-for which there are Mongol ian and Manchu cog­
nates kar(e) , har(e) , etc., to which we can relate the Korean kyelay 'tribe'. (Note also the
obsolete kara 'stalk, stem ; handle; spatula' as wel l as kara 'husk, shell' and kara 'empty';
karada 'body' may be connected.) But I wonder if the particle may not, i nstead, turn out

9 . Also in th e o ld word � kara ' relatives' (said t o be from umi-kara with t h e n o u n derived from the
infinitive umi 'g iving b irth ' ) , best k nown in the compound u kara-y a kara 'people around o ne'. Ya-kara
contains ya i e 'house'. and hara-kara contains 'belly'. O tsu k i would take -kara as a variant of ko -ra
=

'ch ildren' in these words.

48 § 2 . Predi cate Adjuncts

to be somehow related to the Korean verb ka- 'go' with a possible cognate in the Japanese
verbs kf- < ko- 'come' and/or kayow- 'commute, go regu larly'. 1 0 See also the d i scussion on
nagara, § 9. 1 .3.
Above I have suggested that some uses of no may resu lt from a shortening of mono. But
very disturbi ng to this notion-and to the speculations about origins of the subject particle
ga-is the use i n d ialects of ga for those constructions where the standard language uses the
pronominal no 'the one/fact that... (etc.) '. Th is first came to my attention in the Kochi dia­
lect, and I confirmed with a native speaker such forms as atarasi i ga ga = atarasf i no ga 'the
one which is new (as subject) ' suggested by Doi 1 958.267 - 8. Since then I have come
across the usage i n reports of a number of d ifferent d ialects: un ga = umu no 0 'giving
birth ' ( I sh i kawa, Zhs 3. 1 1 3) ; kuru ga noo = kuru no o 'coming' (Zhs 1 .26, N iigata-noo =
nee?) ; moro ga morote = morau no o moratte 'receiving what one receives' ( I sh i kawa, Zhs
3. 1 2 1 ) ; atarasi i ga koote = atarasi i no o katte 'buyi ng a new one' (Toyama, Zhs 3.30, 35) ;
i i ga tanomu zo = fi no o tanomu yo 'Be sure to get me a good one' ( I sh i kawa, Zhs 3. 1 81 ) .
Accord ing to Kgg 82.4 1 a n. 1 3, i n Matsuzaki-machi i n Kamo-gun of Shizuoka prefec­
ture, ga is used for pronominal no in the two senses 'the one that' .. . ' and 'the act of . . . '­

as it is i n Koch i : Gakko e i ku GA o kirau 'I hate to go to school ', Watasi ga tabe-tai GA

wa ringo no aoi GA da 'What I l i ke to eat are the green ones of the apples'. But it is not
used for s no da 'it is a fact that .. .' (equ ivalent to l iterary suru nari, not to be confused
with nari 'they say .. . ') for that is said by d irect nominal ization s tnof da : Gakkoo e
SU

i ku da = Gakkoo e i ku no da. Yet i n N i igata we find S ga da : Nai ga da = Nai no da (Zhs

2.336) . And in N i igata and Toyama even no de 'since ' and no ni 'despite the fact that'
can be said with ga for no : si-tai ga de = si-tai no de 'since I want to do it' (Zhs 2.341 ) ­

N i igata) , nai ga dee = nai no de 'since I have none' (Zhs 3. 79-Toyama) ; kono attui ga nii
= kono atu i no ni 'despite its bei ng hot l i ke this' (Toyama, Zhs 3.42; on 3.45 we find atu i
n i ) . 1 1 The use of ga ni for purpose, l i ke the standard use of no n i , may be treated as ell ip­
sis ( see p. 858) V-ru ga/no {tame} ni : m i ni yuku ga n i = mf ni i ku no ni 'in order to go
look at it' ( Zhs 3.62-Toyama) ; this also occurs in Shi koku, accord i ng to Doi 1 958.27 1 .
From I shikawa these are reported : iku ga ya tee = i ku no da to ( itte) 'saying I wou ld go'
(Zhs 3. 1 73-ya is the copula) ; yattee ga de = yari-tai no de 'since I want to do it' (Zhs
3. 1 80) .
I n Ch iba n [o ] ga is apparently used to mean no [mo ] no 'the thing of . . .': hoka n ga de
= hoka no mono de 'bei ng someth ing else' (Zhs 2.220 fn) ; yoso n ga wa = yoso no
[mono] wa 'a different one' (Zhs 2.223).
In centra l N i igata prefecture for S no da you hear S ga n da and this has become known
as "Ech igo no gan-kotoba" according to Zhs 2.29.
The use of ga to mean 'the one that belongs to' is fou nd i n older Japanese ; one of the

1 0 . Aston 4 9 says "kara is contracted for k a areba 'si nce this is so', . . .' and suggests it must have been
earl ier NO kara, derivi ng ono-zukara and te-zukara from -n-tu-kara with the old ( locative-)gen itive mark­
er -tu-. Accord i ng to l sh igaki Kenj i the origi nal mean ing of the particle was abstract 'accordi ng with',
whence there developed use to mark a traversal object. The ablatives of place and t ime were always
marked by yori before the Kokin-shu ( l sh igaki K. 1 49 -50, 1 53 ) . V-te kara first appeared in col loquial
texts of 1 593 (id ., 1 7 1 -2 ) .
1 1 . Gengo-Kenkyu 47.66-7 ( 1 965) reports that nominalizing g a and g a ni/de are common in the
Suzu d ialect of the Noto Peninsula.
§ 2 .2 . Marking of adjuncts: noun postpositions ("particles") 49

examples i n Meika i kogo jiten ( 204c. 1 0) is ... Kakinomoto [ no] H itomaro ga nari = K . H .
n o mono d e aru 'is one of K . H .'s', cited from a note i n the Koki n-shu.
In general, the various d ialects of Japanese use case markers i n very simi lar ways; a few
differences have been noted i n the l ists above. For the allative e (or ni) and the dative ni
(or e) there are three d ifferent etyma which have figured prominently i n the development
I

of the language. F luctuations in the use of e and ni have been descri bed above and else­
where ; the third etymon is a group of variants that derive from the noun sama, which now
means 'appearance' but earl ier had the meanings 'way, method' and 'direction' (equ ivalent
to kata) . The forms are SA (origina lly an earmark of the Kanta plain but now of the entire
northeast, accordi ng to Zhs 2. 1 9) ; SAN E and S I N E ( East Kyushu) ; SAMY AA, SANY AA,
SAN (West Kyushu) ; SAM E , SAN , SE (South Kyushu) . Cf. Zhs 6.1 7 n 1 0. Shimmura says
SA was a d ialect form in literature of the Muromachi period ( 1 338- 1 573) . And a sixteenth­
century proverb is quoted by Maeda 1 96 1 . 1 37 and Zhs 5. 1 5 to the effect "Kyoto E,
[ Northern] Kyushu N I , and Kanto SA".
It is interesti ng that we can find examples of sa used for ni in other uses, in addition to
the dative and al lative:
( 1 ) dative : hoto ke-sama sa agete 'givi ng it to Buddha' (M iyagi , Zhs 1 . 1 70) .
(2) al lative : ya kuba sa itta 'went to the government office' ( F u kushima, Zhs 1 .278) ;
sa ihu sa ire-be [= ireyoo] to site 'when I tried to put it i n my purse' ( I wate, Ta keda
1 970.60) .
(3) mutative : N sa naru 'becomes N ' (Aomori , Zhs 1 .64 etc.) .
(4) purposive: V-i sa i ku (etc.) 'goes to V ' ( I wate, Zhs 1 .94, 96) ; asobi sa kitara 'if you
come for a visit' ( M iyagi, Zhs 1 . 1 60) ; si sa kite 'coming to do it' (Yamagata, Z hs 1 .240) ;
mukee sa igu bee ( = mukai n i i koo) 'let's go to meet h i m' ( F u kushima, Zhs 1 .264) ; m i sa
itte 'goi ng to see' ( F u kush i ma, Zhs 1 .275) . This is not to be confused with a northeastern
use of sa as an abbreviation of the d ialect particle sakai 'because (of) ' (= kara) , nor with
the sentence-fi na l sa of § 1 5.2 .
I n § 1 3.5 you wi l l find a few Osaka forms such as dona i = donna ni 'in what manner'
that appear to be the essive (or adverbialized copula) n i with the i nitial nasa l su ppressed.
Similar Osaka examples of [ n ] i i n other uses are also found, e.g . : saki i = saki n i 'ahead =
first' (temporal locative, Zhs 4.202) , tabi i det [e] = tabi ni dete 'leavi ng for a trip' ( pur­
posive, Zhs 4.227) , otumu i = atama ni 'onto the head ' ( mutative locative, Zhs 4.202) , ... .
Maeda 1 961 . 1 39 says that Osa ka also has i as a variant of e, used in situations where n i
wou ld give a different meaning as i n Soko e suwan-nahare = Soko e suwari-nasai '(Move and)
sit there', cf. Soko n i suwatte [ i- ] nahare = Soko ni suwatte i-nasai 'Stay seated there'­
Osa ka will also al low de for nf in the last example (Maeda 1 96 1 . 1 40) .
We have noted the confusion between e and nf among standard speakers, some of whom dis­
tingu ish these particles part or all of the time, wh i le others relax the d isti nction or lack it altogeth­
er . The latter category i ncludes most modern Kyushu spea kers, who lack e and use nf i nstead . I n
modern Osa ka, accordi ng t o Maeda 1 96 1 . 1 4 1 -2 , when the dative i s used with a donatory verb the
particles are distingu ished depend i ng on whether the recipient is a superior or is an i nferior such
as oneself; we can presume this reflects examples I i ke sensei E ageru ' I give it to the teacher' and
uti N I kureru 'he gives it to me'.
A fourth etymon for the dative-al lative is ge 'to a person ', found frequently in Chiba
(Zhs 2.206n) , but also in Kanagawa (Zhs 2.304 n 1 ) , and other parts of the northeast
Kanta area (Zhs 2.26) . Zhs 2.26 says th is form is probably a contraction of an old particle
or pseudo-restri ctive (-)gari 'to (where one is) ' as found in the example kimi gari i keba 'if I
50 §2. Predicate Adjuncts

come to my lord ' (MKZ 1 54a) . (The N i igata form songe = sonna ni, Zhs 2.365, must have
a d ifferent source, perhaps the evidential -ge of § 20.)
In the R yu kyus a particle saa [ n ] i is used for the i nstrumental and materia l uses of
standard de, and the northern R yukyu particle syi (Martin 1 970. 1 1 1 b) in some of its uses
may have come from a contraction of that, unless a l l uses are derived from the i nfinitive
of 'do' (cognate with sta ndard si) . Cf. H ogen-gaku gaisetsu 1 30, which states that i n the
R yu kyus [ n ] ka i is used for 'to a place' and nkai, nai , or nee for 'to a person'; a l l these
forms are probably reductions from na kai, wh ich means ' ( in)to' and perhaps derives from
a reduction of [ no ] naka [ n ] i 'to i nside of'. But reflexes of sim ple ni are fou nd in a l l
R yu kyu d ia lects, I bel ieve, as expressions o f t h e persona l dative and some o f the other
meanings found in the standard language. I n the northern R yu kyus (Martin 1 970) there
are also such forms as k [ h]atyi < kati 'to(ward) I (= e) and gadi ?< gade < kade 'up to,
as far as' (= made) .

The pronom i nal use of no 'one/fact/it' (no 3 above) is expressed by ga in some of the
d ialects, as we have observed. I n Okinawa (cf. Hogen-gaku ga isetsu 1 31 ) there is a particle
of simi lar use that has the shape si, which must derive from an earl ier su, since it has an u n­
pa latalized sibi lant. This particle appears in the northern R yu kyus i n such expressions as
was [ i ] 'mine', nas [ i ] 'yours', and qar ga si 'his/hers' (Martin 1 970. 1 23) .
2.2a. UNMA RKED A DJUNCTS

N ot a l l adjuncts are marked ; a few, adverbs par excellence, are never followed by a rela­
tional marker, as noted in § 1 3. 7. Other unmarked adjuncts are the result either of d irect
adverbia lization of a nom inal ( § 9.1 . 1 3-a process we can treat as ellipsis of the essive or
copu la i nfinitive n f) or of the application of a marker of focus, such as wa/mo, that regu­
larly suppresses the subject marker ga and the object marker o and sometimes nf ( i n sev­
era I of its uses) .
The surface versions of sentences-what we hear and see-often contai n an o ptional
omi3sion of a marker, the result of ellipsis. The object particle o is very frequently
dropped , especially in d ialects; the subject particle ga somewhat less frequently, 12 and the
so-cal led "topic" particle wa sti l l less frequently. I n western and central Japan ( most no­
ticeably in H irosh ima) the quotative to ' [saying/th inking ] that . . . ' drops read i l y before a
quotational verb. Maeda ( 1 96 1 .82 ff) suggests that the e l l i psis of o is quite old and the
other droppings are more recent but gradually gaining in frequency. For Osaka , he gives
frequencies as high as the fol lowi ng for ellipsis within discourse :
dropped o .9 1
dropped to .88 before iu 'says'
.75 before omou 'thinks'
dropped ga .60
dropped wa .36
O n the frequent droppi ng of o and quotative to by Kansa i speakers, see also Zhs 4. 1 7.

1 2. But when ga marks an u nderly ing ( semantic) obj ect, the frequency of omission is probably
the same as for o.
Exploring the Variety of Random
Documents with Different Content
his brains and his blood. Of all my company none remained but six
men and me.
To this carnage succeeded a moment of stupor. The colonel, putting
his hat on the point of his sword, was the first to scale the parapet,
crying: “Vive l’Empereur!” He was followed instantly by all the
survivors. I do not remember clearly just what followed. We entered
within the redoubt, how I do not know. We fought body to body
amid a smoke so dense that we could not see one another. I believe
that I smote, for I found my sabre was all bloody. At last I heard the
cry, “Victory!” and, the smoke diminishing, I saw blood and dead
bodies completely covering the earthworks of the redoubt. The
cannons especially were buried beneath piles of corpses. About two
hundred men, in the French uniform, were grouped without order,
some loading their muskets, others wiping their bayonets. Eleven
hundred Russian prisoners were with them.
The colonel was lying all covered with blood upon a broken caisson
near the entrance. Several soldiers bestirred themselves around him:
I approached.
“Where is the senior captain?” he inquired of a sergeant.
The sergeant shrugged his shoulders in a manner most expressive.
“And the senior lieutenant?”
“This gentleman here, who arrived yesterday,” said the sergeant, in a
perfectly calm tone.
The colonel smiled bitterly.
“Come, sir,” he said to me, “you are now in chief command;
promptly fortify the entrance of the redoubt with these wagons, for
the enemy is in force; but General C—— will see that you are
sustained.”
“Colonel,” I said to him, “you are severely wounded?”
“Pish, my dear fellow, but the redoubt is taken!”
 FOOTNOTES:
[12] Latin: not twice in the same (place).
[13] In rank.
 PIERRE LOTI, COLORIST

Pierre Loti is a cosmopolitan. Halévy was a Parisian, Maupassant was

a Norman acclimated to the habitudes of Paris, and Daudet carried
with him throughout all his experiences in the French capital the
dreamy soul of Provence; but Loti is essentially a modern. Man of
the world, not alone by temperament but by reason of wide
journeyings afield and minglings with men and women of all lands,
he typifies the spirit of to-day in French literature as few other
writers have done. He is a poetic idealist, or, perhaps more precisely,
an idealistic realist, writing at a time when realism was most potent
in France.
The externals of Loti’s life are soon recounted. Louis Marie Julien
Viaud was born in Rochefort, January 14, 1850, the same year that
gave Maupassant to the world of art. The name “Loti” is an invented
derivative of that seductive tropical flower, the lotus, and therefore
was not his by inheritance, but the affectionate gift of his South Sea
enchantress, Queen Pomaré, of Tahiti, when the young naval officer
visited the island in 1872.
The frail, prim, sensitive child described with so much self-insight in
his autobiographic Le Roman d’un enfant not only inbreathed his
love for the sea in salt-scented Rochefort, but dreamed incessantly
of the far-off lands he was destined to visit. These visions were
stimulated, if not inspired, by early reading, and by the letters of an
older brother who had long been in the navy. So at seventeen we
naturally find him a midshipman, and in due course ensign and
lieutenant, serving with distinguished bravery—as his Cross of the
Legion of Honor testifies—in the Tonquin campaign, when France
must needs re-subdue her protectorates in Asia.
M. Loti’s later life has been spent mostly ashore, serving in the
Admiralty, yet the cravings of boyhood have been indulged so often
as might be, and foreign lands, by preference oriental, visited year
by year.
In 1891 M. Loti was elected to that all-coveted distinction,
membership in the Academy, where he occupies the chair once
honored by Racine and Scribe.
Loti’s portraits show us a Gallic face, a short, pointed beard, tired,
melancholy eyes, and a general air of earnestness not quite
substantiated by his pleasure-loving life. In stature somewhat below
medium height, in form slender, he early gave himself to those
bodily exercises which once caused a professional acrobat to wonder
why our author had not begun his gymnastics early enough to turn
his steel-like muscles to spectacular account!
So much for the more patent facts of his life. But how to make just
presentment of his mental and spiritual traits I do not know, for the
task gave pause even to Loti himself.
He never learned to write; his gift was native. With reading he had
at first small commerce, preferring to turn page after page in human
hearts, and to read deep in the tome of his own nature of
bewildering variety. A composite is Loti—almost a chameleon, not
only entering into the multi-life of lands and peoples where he
chances to sojourn, but taking on their colors, and even their garbs
and customs. But of this somewhat more in due order.
Here is a character inextricable from his work, much of which is
autobiographical, since in most of his twenty-seven volumes the
author himself appears either thinly disguised under some sobriquet,
or frankly named in propria persona. So while we are at no time at a
loss for material wherewith to make up an estimate, this material is
both embarrassingly rich and—contradictory. Still, no one can
mistake the main-travelled roads in this life, they are bold and
distinct.
Loti wrote little verse, but he was a poet. He moved in the upper
layers of feeling—feeling for nature, for animals, for man, for woman
—and always he was the idealistic, though not the ideal lover. His
sympathies were positively unquenchable, and each new passion
found him fresh, tender, elemental—and as sincere as the temporary
lover can be. In elemental, primitive folk he found his personal loves
and his fictive characters; in the death of a little bird or in the lives
of two cats he centred a genuine interest; in the moods of the
uncompassed sea he felt a vast concern.
Inevitably, the religious life of such a temperament would not be
constant: Protestantism the Huguenot youth found to be cold; his
test of the ceremonial worship of Romanism satisfied him only for a
little; at last his faith was doubt.
Loti’s direct disregard of the interests of conventional life, in favor of
nature-children, constitutes one of his greatest literary charms.
Freshness, simplicity of viewpoint, naïve boyishness of spirit—these
excel all the accomplishments of the stylist’s art in an author whose
style is as subtle as gossamer, as varicolored as the evening sea he
painted with supernal beauty.
In all his work Loti greatly prefers description above dialogue. “Long
and quiet stretches of writing” abound, but their minuteness leaves
us unwearied, and though he repeats and re-repeats we are
conjured into accepting his pictures as ever new.
In style, in delicacy of nature-feeling, where in all literature will you
find aught to excel this passage from Mon frère Yves?
“Even the nights were luminous. When all was slumbering
in heavy immobility, in dead silence, the stars shone out
above, more dazzling than in any other region of earth,
and the sea also was illumined from beneath. There was a
sort of immense gleam diffused over the waters; the
 lightest motion, such as the slow gliding of the boat, or a
shark darting after it, brought out upon the warm eddies
flashes like the color of a fire-fly. Then, over the great
phosphorescent mirror of the deep, there were millions of
wild flames—they were like little lamps lighting themselves
everywhere, burning mysteriously for a second or two,
then dying out. These nights were swooning with heat,
full of phosphorescence; and in all this dim immensity
light was brooding, and all these seas held latent life, in a
rudimentary state, as did formerly the gloomy waters of
the primeval world.”
As in the foregoing, so in the following, see how this necromancer of
words accomplishes the impossible—“the planks of the ship” are the
only solid, palpable substances in this atmospheric delicacy from
Pecheur d’Islande (An Iceland Fisherman):
“Outside it was daylight, perpetual daylight. But it was a
pale, pale light, resembling nothing else; it threw dim
reflections over everything, as of a dead sun, and beyond
these, all was an immense void without color; everything
outside the planks of the ship seeming diaphanous,
impalpable, unreal.
“The eye could scarcely distinguish the sea. First it took
on the aspect of a sort of trembling mirror, with no image
reflected in it; as it spread further it seemed to become a
vaporous plain, and beyond this there was nothing—no
outline nor horizon.
“The damp freshness of the air was more intense, more
penetrating, than actual cold; and in breathing it one was
conscious of a taste of brine. All was calm, and it was no
longer raining; above, formless, colorless clouds seemed
to hold that latent, unexplained light; one could see
plainly, while conscious all the time that it was night, and
all these pallors were of no shade that can be named.”
This is not description—it is miracle; it is, in the fine phrase of M.
Doumic, “evocation;” it is music, color, subtlety, spirit, all thrown
upon the soul’s retina and sensed in some magic manner that
refuses to be classified. No one but a pantheist, sensitive to all the
moods of nature—and especially those of that abysmal enigma, the
sea—could have evoked such visions, such realities, where other
eyes see—water.
In form our author’s books are varied, following rarely any
preconceived plan, we may well suppose—only this, that the literary
wanderer with his new book every year takes us by the hand and
shows us the intimacies of his own life-experiences, discloses the
little-known beauties and sadnesses he has uncovered everywhere,
and turns into simple yet exquisitely wrought fictions the poignant
truths that have entered his own heart. Not one novel, technically
considered, did he write, but sketches strung like pearls upon a
thread: vivid impressions of home and foreign life, longer or shorter
stories of simple folk whose days dawned to labor and were twilit
with weariness, colorful pictures of men and women living under
eastern skies—and beneath and about all, the many-spirited sea.
It would require a volume to deal adequately with Loti’s many
books; but one point invites mention: each new annual volume for a
score of years discloses his life in some new land, or in the Brittany
of his affection.
His first volume, Aziyadé (1879), is the record of his love for a
beautiful Circassian slave while he sojourned in Turkey—the record,
too, of how she died of grief after his departure. Rarahu—later
issued as Le Mariage de Loti—recounts his loves in Tahiti, and much
of charm and beauty besides. Le Roman d’un Spahi transports us to
the Sahara and Senegal, Fleurs d’ennui to Montenegro, Madame
Chrysanthème to Japan, Au Maroc to Fez and Tangier, and Le désert,
and Jérusalem, and La Galilée, to Palestine.
I name these volumes not to attempt a catalogue of Loti’s works,
but to show how world-broad were the scenes he chose for his
impressionistic brush. Naturally, all of the foregoing works are more
or less oriental in tone, and the moral code revealed is not that of
“the most approved families.” But three masterpieces there are
which breathe a more wholesome air—though heavy, each one, with
the tragedy of life.
Mon frère Yves is the plotless account of “a splendid Breton sailor
and the author, his officer.” They enjoy “a sort of companionship
which finds its analogy—in a way—in the friendly relations formerly
[held] between young master and slave in our Southern States.” No
picture of the robust rollicking sailor—superstitious, drink-loving,
adventurous, warm-hearted—could be more real, none more
pathetic, and none more rich in fragments of narrative.
In Le livre de la pitié et de la mort eleven stories are brought
together to harmonize with the saddening title—“The Book of Pity
and of Death!” One of these, “The Sorrow of an Old Convict” is an
impressionistic tale of an old highwayman who is being shipped
away to exile. His only solace is a caged bird with a broken wing,
and when one day the door is opened the little bird falls into the
sea. That is all—but to read it is to feel with Yves the heart-break of
that bereft old man.
“The Wall Opposite” is a study of human tendencies. A mother, a
daughter, and an aged aunt are compelled by reverses to let out
those rooms of their apartments that faced out upon the street, but
their own little back suite had a cozy and intimate air. Its windows
overlooked a court whose walls were covered with honeysuckle and
roses. One day they were told that in the court a high wall was to be
built which would steal away the air and hide the sun. They had no
money wherewith to buy off the project, so in one short month a
grayish-white wall—almost like a twilight sky of November—shut
them in.
Long they had looked for an inheritance which would some day
come to them. Then they would buy the house and tear down that
wall—and always the old aunt used to pray that she might live to see
that day. But the bequest was long in coming.
One day a young man came, introduced by friends, and for a while
he sat at the table of these “three ladies without fortune.” He was
handsome and high-spirited, and the young girl loved him, but she
was poor, and for lack of sunlight the color had begun to fade from
her cheeks. So he went away and never returned.
Twenty years passed—the aunt had died, the mother had grown
gray, and the daughter was now past forty. Then at last the
inheritance came. They sent away their lodgers, but somehow the
two women remained in the little back salon. They had come to love
it. At last the wall which for twenty years they had endured would be
torn down. At twilight of the second day the wall was razed, but the
mother and daughter sitting at their table were bewildered at seeing
so clearly. The wall was gone—they had the light, the roses and
vines! For twenty years they had hoped for this happiness, yet now
—they were uneasy, something seemed to have gone wrong. A sort
of melancholy had come over them.
The mother, looking into her daughter’s eyes, saw tears. “It can be
built up again,” she says. “It seems to me they can try, can they not,
to make it the same again?”
“I, too, thought of that,” replied the daughter. “But no, don’t you
see? It would never be the same!”
And this was the secret: more than the power of custom in her Life
was the fact that the wall had been the background of a picture—the
face of a young man which she had watched through one short
spring-time.
This is one of Loti’s few technically perfect short-stories. His sketchy,
rambling, loosely-plotted “novels” and travel-reflections differ greatly
in manner from the compact story of plot, but his writings do
abound in easily separable fragments, or episodes—as to which a
word must now be set down, before we take up the plot and the
final scenes of Loti’s greatest work, Pecheur d’Islande.
Fortunately for the spontaneity of the novel, many authors are more
concerned for the vividness of their narration than for mere technical
form. Hence they feel free to introduce incidents which are related
more or less loosely to the plot, and serve rather as auxiliaries than
as vital parts of the action. The purpose may be to develop a tone,
suggest an atmosphere, illustrate certain traits of character, or, it
may be, to amplify an organic part of the plot. This narrative by-
path, this illuminating side-light, we technically call an episode. It
was most in vogue among the early English novelists; Defoe,
Richardson, Smollett, Fielding, and Goldsmith followed it so
habitually that all of their novels are episodic in form. But even in
the more highly organized French romances of plot—Les Miserables
and Les Mystères de Paris, for example—we find frequent episodes.
This tendency is naturally more marked in the tale and in the
prolonged sketch than in the closely plotted novel. Indeed, it is only
in the very long plot-novel that the episode can find room, since the
prosperity of the short plot-novel lies largely in the close and rapid
sequence of its incidents.
Even though “The Marriage to the Sea”—as I have ventured to
entitle this climacteric close of An Iceland Fisherman—is an essential
plot incident, and therefore an organic part of the whole, still,
considered solely for its own sake, it is easily detachable. So we may
regard it as almost a typical specimen of the episode; that is to say,
we need only have some slight prior knowledge of the setting and
the relation of the characters to invest it with the completeness and
unity of a perfect short-story. True, the crisis has occurred—unknown
to the fisherman’s wife—before this episode begins, but that could
scarcely have been arranged more artistically, with regard to
suspense, had Loti purposed to use the episode as a separate story.
Here we have the carefully laid groundwork of tone, environment,
and characters. Here, too, are the breathless expectancy, the
increasing suspense (which constitutes the complication), two false
anticipations of a happy dénouement, and then the actual
dénouement, with the artistic close.
An Iceland Fisherman is Pierre Loti’s most perfect work, and it is
gratifying to note that it is also his most popular, as witness some
three hundred and fifty French editions, and an unknown number of
translations. In form, it is less a typical novel than a brilliant
impressionistic tale. A major episode is the story of Sylvestre, which,
woven closely in its earlier part with the life of Yann and his
sweetheart Gaud, at length diverges, when the fisherboy passes into
the navy, fights a good fight in Cochin-China, and dies amid pathetic
circumstances in far-off Singapore.
The plot is very simple. It is laid in Paimpol, in Brittany, whose
dwellers rely solely upon the Iceland fisheries. Every year these
hardy Vikings of Northern France fare away to the Iceland waters
and return only after a long season there. The chief characters are
Yann Gaos, a great splendid young fisherman with handsome brown
curls, and Gaud, the daughter of “the great man” of the town. The
two are in love, and Yann ventures some hesitating advances; but
her father’s wealth deters the fisherman from making a full avowal.
However, when Gaud’s father dies she is found to be penniless; still
Yann unaccountably holds back, much to Gaud’s secret sorrow.
Homeless, she goes to live with Granny Moan, the grandmother of
the ill-fated young Sylvestre, who had been betrothed to Yann’s
sister. At length, in the little hut where Gaud lives as the bereft old
woman’s foster-grandchild, she and Yann are married.
Only a few days after their wedding, the bridegroom sails away on
the fine new Léopoldine for the Iceland fisheries. When autumn
comes the boat does not return with the others. All that is heard of
her is from the crew of the Marie-Jeanne, who report a mystic
meeting with the Léopoldine in a dense fog, when each vessel
loomed up to the other out of the mist and then passed spectre-like
away, with time for only a few quick cries of recognition from fellow-
townsmen. The final scene opens with all the town awaiting the
return of the fishers. One vessel has already come in, and then
opens this closing episode.
 THE MARRIAGE TO THE SEA
AN EPISODE FROM “AN ICELAND FISHERMAN”
(PECHEUR D’ISLANDE)
By Pierre Loti
Done into English by the Editor

The Iceland ships were returning—two the second day, four the
next, and twelve during the week following. And throughout the
country joy returned with them—there was happiness for the wives
and mothers; happiness too in the taverns where the pretty Paimpol
girls served drink to the fishermen.
The Léopoldine was in the group of belated ones; there were still ten
missing. They could not be long now, and Gaud, in the thought that
Yann would be there within a week—an extreme of delay which she
allowed for so as not to be disappointed—was in a delicious
intoxication of expectancy, keeping the home well in order—very
clean and very neat—to receive him.
Everything being in readiness, there was nothing more for her to do;
besides, in her impatience her head could hold only the one thought.
Three more of the tardy ships now arrived, and then five. Only two
were wanting from the muster.
“Come!” they said to her laughingly, “this year it is either the
Léopoldine or the Marie-Jeanne that will have to stay behind 'to
sweep up.’”
And Gaud laughed—even she—more animated and more beautiful in
her joy of anticipation.
Meanwhile the days passed by.
She continued to dress every day, to put on a gay air, to go to the
harbor a-gossiping with the others. She said that it was all quite
natural, this delay. Didn’t they see the same thing every year? Oh, as
to their coming back at all—with such good sailors, and two such
good boats!
Afterwards, when she was back home at night, the old shiver of
anxiety, of anguish, would come over her.
Could it be really possible that she began to fear—already? Was
there any cause for fear? And she trembled, for having so soon been
afraid.
The tenth of September! How the days flew by!
One morning when there was a cold mist over the earth, a true
autumn morning, the rising sun found her early seated under the
porch of the chapel of the shipwrecked mariners, at the place where
the widows go to pray—seated, she was, with eyes fixed and
temples tense as though held in a band of iron.
Two days ago these melancholy mists of dawn had begun, and on
this particular morning Gaud had awakened with a more poignant
inquietude, caused by this impression of winter. Why was it so this
day, this hour, this moment, more than the preceding? She knew
well enough that boats were often two weeks late—even a month.
But there was something different about this particular morning,
without doubt, for she had come to-day for the first time to sit under
the chapel porch and reread the names of the young men who had
died.
In Memory of
GAOS, YVON,
Lost at Sea
Near the Norden-Fjord.
Like a great shudder, a gust of wind was heard rising from the sea,
and at the same time something fell like rain upon the roof: it was
the dead leaves. A whole host of them were blown in at the porch;
the old wind-tossed trees of the graveyard were losing their foliage,
stripped by this gale from the sea. Winter was coming.
Lost at Sea,
Near the Norden-Fjord,
In the Storm of the 4th and 5th of August, 1880.
She read mechanically, and through the arch of the doorway her
eyes sought to pierce the distance over the sea: that morning it was
very vague, under the gray mist, and a suspended cloud-drapery
trailed over the horizon like a great mourning-veil.
Another gust of wind, and other dead leaves came dancing in. A
stronger squall, as if the west wind that had strewn these dead over
the sea wished to torment even the inscriptions which recalled their
names to the living.
Gaud looked with involuntary persistence at an empty space upon
the wall which seemed to wait with terrible expectancy; she was
pursued by the thought of a fresh slab that might perhaps soon be
placed there, with another name which even in spirit she did not
dare repeat in such a place.
She felt cold, but remained seated on the granite bench, her head
thrown back against the stone wall.
... Lost Near the Norden-Fjord,
In the Storm of the 4th and 5th of August,
At the Age of 23 Years,
May He Rest in Peace!
Iceland appeared to her, with its little cemetery—Iceland far, far
away, lighted from below the sea-line by the midnight sun ... and
suddenly—still in the same empty space on the wall which seemed
to be waiting—she saw with horrifying clearness the vision of that
new slab she had imagined: a fresh tablet, a death’s-head and cross-
bones, and in the centre, within a flame, a name—the adored name
of Yann Gaos! Then she drew herself up straight and stiff, with a
hoarse, wild cry in her throat like a mad creature.
Without, the gray dawn-mist still hung over the earth, and the dead
leaves continued to come dancing into the porch.
Steps on the foot-path!—Was somebody coming?—Then she arose
quickly, with a swift movement readjusting her coif, and composed
her countenance. The footsteps came nearer, as though they would
enter. At once she assumed the air of being there by chance. Not for
anything in the world would she as yet seem like the widow of a
shipwrecked mariner.
It was only Fante Floury, the wife of the mate on the Léopoldine.
She understood at once what Gaud was doing there; it was useless
to dissemble with her. And at first they stood mute, the one before
the other, these two women; all the more alarmed and angry at
being entrapped while in the same mood of fear, they almost hated
each other.
“All those from Tréguier and from Saint-Brieuc have been back for a
week,” said Fante at last, pitilessly, in a voice low and almost
irritated. She carried a taper, meaning to make a votive offering.
Ah! Yes! a votive offering—Gaud had not wished to think as yet of
that last resort of the desolate. But she entered the chapel behind
Fante, without saying anything more, and they knelt side by side,
like two sisters.
To the Virgin, Star of the Sea, they said their passionate prayers with
all their hearts. But only the sound of sobs was heard, and their
rapid tears began to fall upon the floor.
They arose together, more tender, more confident. Fante aided the
tottering Gaud, and, taking her in her arms, she kissed her.
After wiping away their tears, arranging their hair, and brushing the
saltpetre and dust of the flagstones from their skirts at the knees,
they went away without saying anything more, by different paths.

This September’s close was like another summer, only it was

somewhat melancholy. The weather was really so beautiful this year
that had it not been for the dead leaves that fell in a mournful
shower along the roadways one might have said that it was the gay
month of June. Husbands, fiancés, sweethearts, had all returned,
and everywhere was the joy of a second spring-time of love.
At last one day one of the delayed ships from Iceland was signalled
in the offing. Which one?
On the cliff, groups of mute and anxious women quickly formed.
Gaud was there, trembling and pale, by the side of the father of her
Yann.
“I firmly believe,” said the old fisher—“I firmly believe it’s them! A
red sail, a topsail that clews up—that’s jolly well like them anyhow.
What do you say, Gaud, my girl?
“And yet—it isn’t,” he went on, with sudden discouragement; “no,
we’ve made a mistake again, the boom isn’t the same, and they
have a flying jib. Well, well, it isn’t them this time, it’s the Marie-
Jeanne. Oh! but very surely, my girl, they’ll not be long now.”

And day followed day, and each night came at its appointed hour,
with inexorable tranquillity.
Gaud continued to dress every day, somewhat like a mad woman,
always in fear of seeming to be the widow of a shipwrecked sailor,
exasperated when others glanced at her compassionately and
furtively, and looking aside so that she might not meet those glances
that froze her very blood.
Now she had fallen into the habit of going of mornings right to the
end of the headland on the high cliffs of Pors-Even, passing behind
Yann’s paternal home so as not to be seen by his mother or his little
sisters. She went all alone to the extreme point of the Ploubazlanec
land, which is outlined in the shape of a reindeer’s horn against the
gray Channel, and sat there all day long at the foot of a lonely cross,
which rises above the immense expanse of waters.
There are many of these granite crosses hereabout, set up on the
uttermost cliffs of this land of mariners, as though to implore mercy,
—as though to appease that restless, mysterious thing that lures
men away and never gives them back, and by preference keeps the
bravest, the noblest.
Around this cross of Pors-Even stretched evergreen moors, carpeted
with short rushes; and at this great height the sea air was very pure,
having scarcely any of the briny smell of the seaweed, but perfumed
with the delicious ripeness of September.
Outlined in the far distance could be seen, one after another, all the
indentations of the coast, the land of Brittany terminating in ragged
edges which stretched far out into the tranquil void of the waters.
Near at hand the reefs riddled the sea, but out beyond nothing
troubled its polished mirror. There sounded over all a soft, caressing
murmur, light and infinite, arising from the deeps of its every bay.
And the distance seemed so calm, and the depths so soft! The great
blue void, the tomb of the Gaos family, guarded its inscrutable
mystery while the breezes, faint as human sighs, wafted here and
there the perfume of the gorse, which had bloomed again in the
latest autumn sun.
At certain hours regularly the sea retreated, and shallow places grew
larger everywhere, as if the Channel were slowly emptying itself;
then, with the same lazy slowness, the waters rose again, and
continued their eternal going and coming without any heed of the
dead.
And Gaud, seated at the foot of the cross, remained there, in the
midst of these tranquil scenes, gazing ever before her, until the night
fell, until she could see no more.

September had passed. Gaud could no longer take any nourishment,

she could no longer sleep.
She remained at home now, and sat crouching with her hands
between her knees, her head thrown back and leaning against the
wall behind. What was the good of getting up, what was the good of
going to bed? When she was too much exhausted she threw herself
dressed upon her bed. Otherwise she always remained seated,
benumbed; her teeth chattered with cold, in her stony quiet; always
she had that sense of a band of iron round her brows; her cheeks
felt drawn, her mouth was dry, with a feverish taste, and at times a
raucous groan rose from her breast, spasmodically repeated again
and again, while she beat her head against the granite wall.
Or else she called Yann by his name, very tenderly, in a low voice, as
if he were quite close, and whispered to him words of love.
Sometimes she would think of other things besides him—of many
little, insignificant things; she would amuse herself, for example, by
watching the shadow of the china Virgin and the holy-water basin
lengthen slowly over the high woodwork of her bed as the sun went
down. And then the thoughts of anguish returned with more horror,
and her cry broke forth again while she beat the wall with her head.
And so all the hours of the day passed, one after the other, and all
the hours of the evening, and all those of the night, and all those of
the morning. When she had reckoned how long it was since he
ought to have been back, a still greater terror laid hold upon her;
she wished to forget all about the dates and even the names of the
days.

Usually there are some indications concerning the wrecks off

Iceland: those who return have seen the tragedy from afar; or else
they have found some wreckage, or a dead body, or have some sign
from which to divine the facts. But no, of the Léopoldine nothing had
been seen, nothing was known. The men of the Marie-Jeanne, the
last to have seen her on the 2d of August, said that she was to have
gone on fishing farther towards the north, and beyond that the
mystery was unfathomable.
Waiting, always waiting, without knowing anything. When would the
moment come when she truly need wait no longer? She did not even
know that; and now she almost wished that it might be soon.
Oh! if he was dead, let them at least have pity enough to tell her!
Oh! to see him as he was at this very moment—him, or even what
remained of him! If only the Virgin, prayed to so often, or some
other such power, would grant her the blessing of showing him to
her, by some sort of second-sight—her Yann—him—living, struggling
to return to her—or else his body surrendered by the sea, so that
she might at least be sure, that she might know.
Sometimes she would suddenly have the feeling that a sail was
appearing on the rim of the horizon: the Léopoldine approaching,
hastening home! Then she would make the first involuntary
movement to rise, and rush to look out at the ocean, to see whether
it were true.
She would fall back. Alas! where was the Léopoldine now? Where
could it be? Out afar, doubtless, at that awful distance of Iceland,
abandoned, crushed, lost!
And this ended in that never-fading vision, always the same: a
wreck, gaping and empty, rocked upon the silent sea of gray and
rose—rocked slowly, slowly, without sound—with an extreme of
gentleness quite ironical—in the midst of the vast calm of the dead
waters.

Two o’clock in the morning.

It was at night especially that she held herself attentive to all the
steps that approached; at the least stir, at the slightest
unaccustomed sound, her temples vibrated; from being overstrained
that they might sense things from without, they had become terribly
sensitive.
Two o’clock in the morning. This night as on others, hands clasped
and eyes open in the dark, she listened to the wind making its well-
nigh eternal moan over the earth.
Suddenly the steps of a man—rapid steps on the path! At such an
hour, who could be passing? She drew herself up, stirred to the
deeps of her soul, her heart ceasing to beat.
Some one stopped before the door; some one mounted the small
stone steps.
He! Oh! joy of heaven, he! Some one had knocked, could it be any
other! She was up, barefooted; she, so feeble for so many days, had
sprung up nimbly as a cat, her arms outstretched to wind round her
well-beloved. Without doubt the Léopoldine had come in at night,
and anchored opposite Pors-Even Bay; and he—he had rushed
home; she arranged all this in her mind with the swiftness of
lightning. And now she tore her fingers upon the spikes of the door
—in her fury to draw the bolt it had stuck.
Ah!... And now she slowly moved back, crushed, her head fallen
upon her breast. Her sweet mad dream was over. It was no one but
Fantec, their neighbor. She could just comprehend that it was not
he, her Yann, that no part of his being had passed through the air;
she felt herself plunged again into her old abyss, to the uttermost
depths of her same awful despair.
He apologized, poor Fantec: his wife, as Gaud knew, was very ill,
and now their baby was suffocating in its cradle, seized with a
malignant sore throat; so he had come to beg for help, while he ran
to hunt up the doctor at Paimpol.
What did all this matter to her? She had gone mad in her grief, she
had nothing left to offer to others in distress. Huddled on a bench,
she sat before him with eyes glazed, as one dead, not answering
him, not hearing him, not even looking at him. What were these
things to her that the man was saying!
He understood it all; he divined why the door had been opened to
him so quickly, and he had pity for the pain he had brought about.
He stammered out an apology: Just so; he ought never to have
disturbed her—her especially.
“I!” replied Gaud quickly, “and why not I, Fantec?”
Life had returned to her suddenly, for still she did not want to
appear despairing before the eyes of others—for that she was quite
unwilling. And besides, in her turn she pitied him; she dressed to
accompany him and found strength to go see his little child.

When she returned to throw herself upon her bed, at four o’clock,
sleep laid hold upon her in a moment, for she was utterly fatigued.
But that moment of immense joy had left upon her mind an
impression which, in spite of all, was persistent; she awoke soon
with a shudder, rising a little, as remembering something.... She had
some news concerning her Yann.... In the midst of this confusion of
ideas which came back to her, rapidly she searched and searched
her mind for what it could have been.
Ah! nothing, alas, nothing but Fantec!
And a second time she fell back to the depths of the old abyss. No,
in reality, nothing was changed in her morbid, hopeless waiting.
Still, to have felt Yann there so close was as if some emanation from
him had come floating back to her; it was what they call in Breton
land a token; and she listened still more attentively for footsteps
outside, divining that some one would perhaps come who would talk
to her of him.
And indeed, when the day broke, Yann’s father entered. He took off
his cap, pushed back his beautiful white locks, which were in curls
like those of his son, and sat down beside Gaud’s bed.
His heart too was in agony, for his Yann, his splendid Yann, was his
first-born, his favorite, his glory. But he did not despair, not really, he
did not despair yet. He began to reassure Gaud in a very gentle way:
to begin with, the latest ones to return from Iceland had all spoken
of the extremely dense fogs which might easily have delayed the
vessel; and then too an idea had come to him: a stop-over at the
Faroes, which are islands situated on their route, at a great distance;
and when they sent letters from there, they took a long time to
come; the same thing had happened to himself forty years ago, and
his poor dead mother had already had a mass said for his soul....
And such a good boat, was the Léopoldine, and all those aboard
were such able mariners.
Old Granny Moan walked around them, shaking her head; the
distress of her foster grand-daughter had almost given her back her
own strength and reason; she tidied up the place, glancing from
time to time at the little faded portrait of her Sylvestre, which hung
upon the granite wall with its anchor emblems and mourning-wreath
of black beadwork; no, since following the sea had robbed her of her
grandson, she believed no longer in the safe return of sailors; she
now prayed to the Virgin only from fear, with the outside of her poor
old lips, cherishing in the bottom of her heart a grudge against her.
But Gaud listened eagerly to these consoling reasonings, her large
sunken eyes looking with deep tenderness upon this old sire who so
much resembled her well-beloved; just to have him near her was like
a hostage against death, and she felt more reassured, nearer to her
Yann. Her tears fell silently and more gently, and she repeated again
her passionate prayers to the Virgin, Star of the Sea.
A stop-over, 'way out at those islands, to repair damages, was a
likely event. She rose, brushed her hair, and made some sort of
toilet, as if he might possibly return. Doubtless all was not lost if his
own father did not yet despair. And for a few days she again took up
her waiting.
It was full autumn now, late autumn—with the nightfalls gloomy, and
all things growing dark early in the old cottage, and all the Breton
land looking sombre, too. The very days seemed but twilight;
immeasurable clouds, slowly passing, would suddenly bring darkness
at broad noon. The wind moaned constantly—it was like the sound
of a great cathedral organ at a distance, but playing profane airs, or
despairing dirges; at other times it would come close to the door,
and lift up a howl like wild beasts.
She had grown pale, pale, and became ever more dejected, as if old
age had already touched her with its featherless wing. Very often
she would finger the belongings of her Yann, his fine wedding
clothes, folding and unfolding them like some maniac—especially
one of his blue woolen jerseys, which still retained the form of his
body; when thrown gently on the table, it disclosed from long usage
the outlines of his shoulders and chest; but at last she placed it by
itself on a shelf of their wardrobe, never to remove it, so that it
might long preserve that impress.
Every evening cold mists rose from the ground; then through her
little window she would gaze over the melancholy land, where little
patches of white smoke began to rise here and there from other
chimneys: the rest of the men had returned, migratory birds driven
home by the cold. And before many of these fires the evenings
would be sweet; for the spring-time of love had begun with winter,
in all this country of “Icelanders”.
Still clinging to the thought of those islands where he might perhaps
have put in, buoyed up by a kind of hope, she had again begun to
expect him.

He never returned.
One night in August, far away in the waters of gloomy Iceland, amid
a great fury of storm, he had consummated his Marriage to the Sea
—to the Sea which had been his nurse: it was she who had cradled
him, who had made him a big and strong youth, and afterward, in
his superb manhood, had taken him back again for herself alone.
A profound mystery had surrounded the unhallowed nuptials. All the
while, dark veils trembled overhead, moving and twisting curtains
were spread so as to conceal the ceremony; and the bride gave
voice, ever seeking with louder and more awful roars to stifle his
cries.... He, thinking of Gaud, his mortal wife, had battled with giant
strength against this spouse of the tomb—until the moment when he
at last surrendered, with a great cry, deep as the roar of a dying
bull, his mouth already filled with water, his arms open, extended,
and stiffened forever.
And at his wedding were all those whom he had at one time invited.
All except Sylvestre, who himself had gone to sleep in the enchanted
gardens, far, far at the other side of the earth.
*** END OF THE PROJECT GUTENBERG EBOOK SHORT-STORY
MASTERPIECES, VOL. 1 ***

Updated editions will replace the previous one—the old editions

will be renamed.

Creating the works from print editions not protected by U.S.

copyright law means that no one owns a United States
copyright in these works, so the Foundation (and you!) can copy
and distribute it in the United States without permission and
without paying copyright royalties. Special rules, set forth in the
General Terms of Use part of this license, apply to copying and
distributing Project Gutenberg™ electronic works to protect the
PROJECT GUTENBERG™ concept and trademark. Project
Gutenberg is a registered trademark, and may not be used if
you charge for an eBook, except by following the terms of the
trademark license, including paying royalties for use of the
Project Gutenberg trademark. If you do not charge anything for
copies of this eBook, complying with the trademark license is
very easy. You may use this eBook for nearly any purpose such
as creation of derivative works, reports, performances and
research. Project Gutenberg eBooks may be modified and
printed and given away—you may do practically ANYTHING in
the United States with eBooks not protected by U.S. copyright
law. Redistribution is subject to the trademark license, especially
commercial redistribution.

START: FULL LICENSE

THE FULL PROJECT GUTENBERG LICENSE
 PLEASE READ THIS BEFORE YOU DISTRIBUTE OR USE THIS WORK

To protect the Project Gutenberg™ mission of promoting the

free distribution of electronic works, by using or distributing this
work (or any other work associated in any way with the phrase
“Project Gutenberg”), you agree to comply with all the terms of
the Full Project Gutenberg™ License available with this file or
online at www.gutenberg.org/license.

Section 1. General Terms of Use and

Redistributing Project Gutenberg™
electronic works
1.A. By reading or using any part of this Project Gutenberg™
electronic work, you indicate that you have read, understand,
agree to and accept all the terms of this license and intellectual
property (trademark/copyright) agreement. If you do not agree
to abide by all the terms of this agreement, you must cease
using and return or destroy all copies of Project Gutenberg™
electronic works in your possession. If you paid a fee for
obtaining a copy of or access to a Project Gutenberg™
electronic work and you do not agree to be bound by the terms
of this agreement, you may obtain a refund from the person or
entity to whom you paid the fee as set forth in paragraph 1.E.8.

1.B. “Project Gutenberg” is a registered trademark. It may only

be used on or associated in any way with an electronic work by
people who agree to be bound by the terms of this agreement.
There are a few things that you can do with most Project
Gutenberg™ electronic works even without complying with the
full terms of this agreement. See paragraph 1.C below. There
are a lot of things you can do with Project Gutenberg™
electronic works if you follow the terms of this agreement and
help preserve free future access to Project Gutenberg™
electronic works. See paragraph 1.E below.
1.C. The Project Gutenberg Literary Archive Foundation (“the
Foundation” or PGLAF), owns a compilation copyright in the
collection of Project Gutenberg™ electronic works. Nearly all the
individual works in the collection are in the public domain in the
United States. If an individual work is unprotected by copyright
law in the United States and you are located in the United
States, we do not claim a right to prevent you from copying,
distributing, performing, displaying or creating derivative works
based on the work as long as all references to Project
Gutenberg are removed. Of course, we hope that you will
support the Project Gutenberg™ mission of promoting free
access to electronic works by freely sharing Project Gutenberg™
works in compliance with the terms of this agreement for
keeping the Project Gutenberg™ name associated with the
work. You can easily comply with the terms of this agreement
by keeping this work in the same format with its attached full
Project Gutenberg™ License when you share it without charge
with others.

1.D. The copyright laws of the place where you are located also
govern what you can do with this work. Copyright laws in most
countries are in a constant state of change. If you are outside
the United States, check the laws of your country in addition to
the terms of this agreement before downloading, copying,
displaying, performing, distributing or creating derivative works
based on this work or any other Project Gutenberg™ work. The
Foundation makes no representations concerning the copyright
status of any work in any country other than the United States.

1.E. Unless you have removed all references to Project

Gutenberg:

1.E.1. The following sentence, with active links to, or other

immediate access to, the full Project Gutenberg™ License must
appear prominently whenever any copy of a Project
Gutenberg™ work (any work on which the phrase “Project
Gutenberg” appears, or with which the phrase “Project
Gutenberg” is associated) is accessed, displayed, performed,
viewed, copied or distributed:

This eBook is for the use of anyone anywhere in the United

States and most other parts of the world at no cost and
with almost no restrictions whatsoever. You may copy it,
give it away or re-use it under the terms of the Project
Gutenberg License included with this eBook or online at
www.gutenberg.org. If you are not located in the United
States, you will have to check the laws of the country
where you are located before using this eBook.

1.E.2. If an individual Project Gutenberg™ electronic work is

derived from texts not protected by U.S. copyright law (does not
contain a notice indicating that it is posted with permission of
the copyright holder), the work can be copied and distributed to
anyone in the United States without paying any fees or charges.
If you are redistributing or providing access to a work with the
phrase “Project Gutenberg” associated with or appearing on the
work, you must comply either with the requirements of
paragraphs 1.E.1 through 1.E.7 or obtain permission for the use
of the work and the Project Gutenberg™ trademark as set forth
in paragraphs 1.E.8 or 1.E.9.

1.E.3. If an individual Project Gutenberg™ electronic work is

posted with the permission of the copyright holder, your use and
distribution must comply with both paragraphs 1.E.1 through
1.E.7 and any additional terms imposed by the copyright holder.
Additional terms will be linked to the Project Gutenberg™
License for all works posted with the permission of the copyright
holder found at the beginning of this work.

1.E.4. Do not unlink or detach or remove the full Project

Gutenberg™ License terms from this work, or any files
 containing a part of this work or any other work associated with
Project Gutenberg™.

1.E.5. Do not copy, display, perform, distribute or redistribute

this electronic work, or any part of this electronic work, without
prominently displaying the sentence set forth in paragraph 1.E.1
with active links or immediate access to the full terms of the
Project Gutenberg™ License.

1.E.6. You may convert to and distribute this work in any binary,
compressed, marked up, nonproprietary or proprietary form,
including any word processing or hypertext form. However, if
you provide access to or distribute copies of a Project
Gutenberg™ work in a format other than “Plain Vanilla ASCII” or
other format used in the official version posted on the official
Project Gutenberg™ website (www.gutenberg.org), you must,
at no additional cost, fee or expense to the user, provide a copy,
a means of exporting a copy, or a means of obtaining a copy
upon request, of the work in its original “Plain Vanilla ASCII” or
other form. Any alternate format must include the full Project
Gutenberg™ License as specified in paragraph 1.E.1.

1.E.7. Do not charge a fee for access to, viewing, displaying,

performing, copying or distributing any Project Gutenberg™
works unless you comply with paragraph 1.E.8 or 1.E.9.

1.E.8. You may charge a reasonable fee for copies of or

providing access to or distributing Project Gutenberg™
electronic works provided that:

• You pay a royalty fee of 20% of the gross profits you derive
from the use of Project Gutenberg™ works calculated using the
method you already use to calculate your applicable taxes. The
fee is owed to the owner of the Project Gutenberg™ trademark,
but he has agreed to donate royalties under this paragraph to
the Project Gutenberg Literary Archive Foundation. Royalty