A Descriptive and Comparative Grammar of Western Old Japanese [2 volume set, Revised, Updated and Enlarged Second Edition] 9004422811, 9789004422810

This is the revised, updated and enlarged second edition of the first detailed descriptive grammar in English (indeed, i

226 41 7MB

English Pages 1330 [1311] Year 2020

Report DMCA / Copyright

DOWNLOAD PDF FILE

Table of contents :
Contents
Acknowledgements
Preface to the First Edition, Part One
Preface to the First Edition, Part Two
Preface to the Second Edition and Acknowledgements
Abbreviations
Charts
Chapter 1 Sources and Previous Scholarship
Section 1 Sources
Section 2 Previous Scholarship
Chapter 2 Script and Phonology
Section 1 Script
1.1 Syllabic Script
1.2 Rebus Writing
Section 2 Phonetics and Phonology
2.1 Consonants
2.2 Vowels
2.3 Pitch Accent
2.4 Phonotactics
2.5 Morphophonological Processes
Chapter 3 Lexicon
Section 1 Naturalized Loanwords from Ainu
Section 2 Naturalized Loanwords from Korean
Section 3 Naturalized Loanwords from Old and Middle Chinese
Section 4 Loanwords from Chinese in Poetry
Section 5 Loanwords from Chinese in Prose
Chapter 4 Nominals
Section 1 Nouns
1.1 Prefixes
1.2 Suffixes
Section 2 Pronouns
2.1 Morphological Peculiarities of Pronouns
2.2 Personal Pronouns
2.3 Reflexive Pronoun onǝ/onǝre
2.4 Demonstrative Pronouns
2.5 Interrogative Pronouns
2.6 Collective Pronouns
Section 3 Numerals
3.1 Cardinal Numerals
3.2 Ordinal Numerals
3.3 Classifiers
3.4 Months of the Year
Chapter 5 Adjectives
Section 1 Uninflected Adjectives
1.1 Special Derived Form in -ra
1.2 -ka Adjectives
Section 2 Inflected Adjectives
2.1 Converb Form -ku
2.2 Final Form -si
2.3 Attributive Form -ki
2.4 Nominalized Form -sa
2.5 Gerund -mi
2.6 Deverbal Adjectives in -asi
Section 3 Defective Adjectives
3.1 Defective Adjective ka-
3.2 Defective Adjective sa
Chapter 6 Verbs
Section 1 Verbal Grammatical Categories
1.1 Mode
1.2 Aspect
1.3 Tense
1.4 Mood
1.5 Voice
1.6 Retrospective
1.7 Reported Action
1.8 Iterative
1.9 Predication
1.10 Honorification
1.11 Politeness
Section 2 Verbal Classes
2.1 Consonant Verbs
2.2 Vowel Verbs
2.3 Irregular Verbs
2.4 Defective Verbs
Section 3 Verbal Affixes
3.1 Verbal Prefixes
3.2 Verbal Suffixes
Section 4 Auxiliaries
4.1 Bound Auxiliaries
4.2 Lexical Auxiliaries
Chapter 7 Adverbs
Section 1 Adverb ita ~ itǝ ‘Very, Extremely’
Section 2 Adverb sǝkǝmba ‘Very’
Section 3 Adverbs paⁿda ~ panapaⁿda ‘Considerably’
Section 4 Adverbs kǝkǝmba ~ kǝkǝmbaku, kǝkǝⁿda ~ kǝkǝⁿdaku‘ So Much, Extremely’
Section 5 Adverb sapa ‘Many’
Section 6 Adverb taⁿda ‘Only’
Section 7 Adverb mata ‘Again’
Section 8 Adverb iya ‘More [and More], Plentifully, Perfectly’
Section 9 Adverb iyǝyǝ ‘More and More’
Section 10 Adverb masu-masu ‘More and More’
Section 11 Adverb yumɛ ‘At All’
Section 12 Adverb imaⁿda ‘Yet, Still’
Section 13 Adverb mǝtǝna ‘In Vain, Aimlessly, For No Reason’
Section 14 Adverb simba ~ simba-simba ‘Often’
Section 15 Adverb potǝpotǝ ‘Almost’
Section 16 Adverb napo ‘Still, More’
Section 17 Adverb sika ‘Thus, In This Way, So’
Section 18 Adverb kɛⁿdasi ~ kɛⁿdasiku ‘Probably’
Chapter 8 Conjunctions
Section 1 Conjunction tǝmǝ ‘Even If, Even Though’
Section 2 Conjunction mǝnǝ ~ mǝnǝwo ‘Although, But’
Section 3 Conjunction mǝnǝkara ‘Although’
Section 4 Conjunction mǝnǝyuwe ‘Although, Because’
Section 5 Conjunction napɛ ‘At the Same Time As, Just As’
Section 6 Conjunction ŋgane ‘So That, in Order to/That’
Section 7 Conjunction ŋgani ‘Like, As If, So That’
Section 8 Conjunction karani ‘Just Because, As Soon As’
Section 9 Conjunction toni ‘While, Before’
Chapter 9 Particles
Section 1 Focus Particles
1.1 Topic Particle pa
1.2 Focus Particle mǝ
1.3 Focus Particle sǝ ~ ⁿzǝ
1.4 Focus Particle namo
1.5 Focus Particle kǝsǝ
Section 2Interrogative Particles
2.1 Interrogative Particle ya
2.2 Interrogative Particle ka
Section 3 Desiderative Particle mǝŋga ~ mǝŋgamǝ
Section 4 Emphatic Particles
4.1 Emphatic Particle kamǝ
4.2 Emphatic Particle si
4.3 Emphatic Particle mǝ
4.4 Emphatic Particle ya
4.5 Emphatic Particle yǝ
4.6 Emphatic Particle na
4.7 Emphatic Particle wo
Section 5 Restrictive Particles
5.1 Restrictive Particle nǝmï
5.2 Restrictive Particle ⁿdani
5.3 Restrictive Particle sura ~ sora
5.4 Restrictive Particle sapɛ
5.5 Restrictive Particle mbakari
Chapter 10 Postpositions
Section 1 Postposition ⁿzi ~ ⁿzimənə ‘Like’
Section 2 Postposition sambï ‘Like’
Section 3 Postposition məkərə ‘Like’
Section 4 Postposition pa ‘Every’
Chapter 11 Interjections
Section 1 Interjection iⁿza
Section 2 Interjection ina
Section 3 Interjection wo
Section 4 Interjection ani
Section 5 Interjection ana
Section 6 Interjection wasi
Section 7 Interjection apare
Section 8 Interjection we
Bibliography
Index of Personal Names
Index of Terms
Index of Forms and Constructions
Index of the Examples from the Texts
Recommend Papers

A Descriptive and Comparative Grammar of Western Old Japanese [2 volume set, Revised, Updated and Enlarged Second Edition]
 9004422811, 9789004422810

  • 0 0 0
  • Like this paper and download? You can publish your own PDF file online for free in a few minutes! Sign Up
File loading please wait...
Citation preview

A Descriptive and Comparative Grammar of Western Old Japanese Volume 1

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

Handbook of Oriental Studies Handbuch der Orientalistik section five Japan

Edited by R. Kersten

volume 16/1

The titles published in this series are listed at brill.com/ho5

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

A Descriptive and Comparative Grammar of Western Old Japanese Revised, Updated and Enlarged 2nd Edition Volume 1

By

Alexander Vovin

LEIDEN | BOSTON

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

Cover illustration: A modern Japanese calligraphy for Man’yōshū 5.794. Library of Congress Cataloging-in-Publication Data Names: Vovin, Alexander, author. Title: A descriptive and comparative grammar of Western Old Japanese /  Alexander Vovin. Description: Revised, updated and enlarged second edition. | Leiden ;  Boston : Brill, 2020. | Series: History of Oriental studies, 0921-5239 ;  volume16 | Includes bibliographical references and index. Identifiers: LCCN 2020023510 (print) | LCCN 2020023511 (ebook) | ISBN  9789004424012 (v. 1 ; hardback) | ISBN 9789004424029 (v. 2 ; hardback) |  ISBN 9789004422117 (hardback) | ISBN 9789004422810 (ebook) Subjects: LCSH: Japanese language—Grammar, Classical. Classification: LCC PL531 .V68 2020 (print) | LCC PL531 (ebook) | DDC  495.6/7—dc23 LC record available at https://lccn.loc.gov/2020023510 LC ebook record available at https://lccn.loc.gov/2020023511

Typeface for the Latin, Greek, and Cyrillic scripts: “Brill”. See and download: brill.com/brill-typeface. ISSN 0921-5239 ISBN 978-90-04-42211-7 (hardback, set) ISBN 978-90-04-42401-2 (hardback, vol. 1) ISBN 978-90-04-42281-0 (e-book) Copyright 2020 by Alexander Vovin. Published by Koninklijke Brill NV, Leiden, The Netherlands. Koninklijke Brill NV incorporates the imprints Brill, Brill Hes & De Graaf, Brill Nijhoff, Brill Rodopi, Brill Sense, Hotei Publishing, mentis Verlag, Verlag Ferdinand Schöningh and Wilhelm Fink Verlag. Koninklijke Brill NV reserves the right to protect this publication against unauthorized use. Requests for re-use and/or translations must be addressed to Koninklijke Brill NV via brill.com or copyright.com. This book is printed on acid-free paper and produced in a sustainable manner.

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

To Sambi With all my love 賛美に 愛を込めて



Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

Contents Acknowledgements xiii Preface to the First Edition, Part One xv Preface to the First Edition, Part Two xix Preface to the Second Edition and Acknowledgements xxi Abbreviations xxiv List of Charts xxviii

Chapter 1 Sources and Previous Scholarship 1 Sources 4 2 Previous Scholarship 20

Chapter 2 Script and Phonology 1 Script 28 1.1 Syllabic Script 28 1.2 Rebus Writing 39 2 Phonetics and Phonology 41 2.1 Consonants 41 2.2 Vowels 45 2.3 Pitch Accent 48 2.4 Phonotactics 48 2.5 Morphophonological Processes 54

Chapter 3 Lexicon 1 2 3 4 5

Naturalized Loanwords from Ainu 63 Naturalized Loanwords from Korean 64 Naturalized Loanwords from Old and Middle Chinese 67 Loanwords from Chinese in Poetry 70 Loanwords from Chinese in Prose 72

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

viii

Contents

Chapter 4 Nominals 1 Nouns 84 1.1 Prefixes 85 1.2 Suffixes 109 2 Pronouns 215 2.1 Morphological Peculiarities of Pronouns 215 2.2 Personal Pronouns 215 2.3 Reflexive Pronoun onǝ/onǝre 259 2.4 Demonstrative Pronouns 261 2.5 Interrogative Pronouns 284 2.6 Collective Pronouns 319 3 Numerals 327 3.1 Cardinal Numerals 328 3.2 Ordinal Numerals 352 3.3 Classifiers 354 3.4 Months of the Year 367

Chapter 5 Adjectives 1 2

3

Uninflected Adjectives 377 1.1 Special Derived Form in -ra 383 1.2 -ka Adjectives  386 Inflected Adjectives 390 2.1 Converb Form -ku  391 2.2 Final Form -si 406 2.3 Attributive Form -ki 411 2.4 Nominalized Form -sa 425 2.5 Gerund -mi 427 2.6 Deverbal Adjectives in -asi 432 Defective Adjectives 435 3.1 Defective Adjective ka- 435 3.2 Defective Adjective sa 437

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

ix

Contents

Verbs 1

2

3 4

Chapter 6

Verbal Grammatical Categories 446 1.1 Mode 446 1.2 Aspect 446 1.3 Tense 447 1.4 Mood 447 1.5 Voice 447 1.6 Retrospective 447 1.7 Reported Action 447 1.8 Iterative 448 1.9 Predication 448 1.10 Honorification 448 1.11 Politeness 448 Verbal Classes  449 2.1 Consonant Verbs 450 2.2 Vowel Verbs 451 2.3 Irregular Verbs 453 2.4 Defective Verbs  458 Verbal Affixes 504 3.1 Verbal Prefixes 504 3.2 Verbal Suffixes  534 Auxiliaries  804 4.1 Bound Auxiliaries 804 4.2 Lexical Auxiliaries 903

Chapter 7 Adverbs 1 Adverb ita ~ itǝ ‘Very, Extremely’ 997 2 Adverb sǝkǝmba ‘Very’  1000 3 Adverbs paⁿda ~ panapaⁿda ‘Considerably’ 1001 4 Adverbs kǝkǝmba ~ kǝkǝmbaku, kǝkǝⁿda ~ kǝkǝⁿdaku ‘So Much, Extremely’  1002 5 Adverb sapa ‘Many’  1004 6 Adverb taⁿda ‘Only’  1006 7 Adverb mata ‘Again’ 1009

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

x

Contents

8 Adverb iya ‘More [and More], Plentifully, Perfectly’ 1013 9 Adverb iyǝyǝ ‘More and More’  1016 10 Adverb masu-masu ‘More and More’ 1017 11 Adverb yumɛ ‘At All’ 1018 12 Adverb imaⁿda ‘Yet, Still’  1020 13 Adverb mǝtǝna ‘In Vain, Aimlessly, For No Reason’ 1023 14 Adverb simba ~ simba-simba ‘Often’  1025 15 Adverb potǝpotǝ ‘Almost’  1026 16 Adverb napo ‘Still, More’  1027 17 Adverb sika ‘Thus, In This Way, So’ 1029 18 Adverb kɛⁿdasi ~ kɛⁿdasiku ‘Probably’ 1031

Chapter 8 Conjunctions 1 Conjunction tǝmǝ ‘Even If, Even Though’ 1037 2 Conjunction mǝnǝ ~ mǝnǝwo ‘Although, But’  1042 3 Conjunction mǝnǝkara ‘Although’ 1047 4 Conjunction mǝnǝyuwe ‘Although, Because’ 1049 5 Conjunction napɛ ‘At the Same Time As, Just As’  1050 6 Conjunction ŋgane ‘So That, in Order to/That’ 1053 7 Conjunction ŋgani ‘Like, As If, So That’ 1055 8 Conjunction karani ‘Just Because, As Soon As’  1057 9 Conjunction toni ‘While, Before’  1060

Chapter 9 Particles 1

2

Focus Particles 1067 1.1 Topic Particle pa 1067 1.2 Focus Particle mǝ 1082 1.3 Focus Particle sǝ ~ ⁿzǝ 1094 1.4 Focus Particle namo 1104 1.5 Focus Particle kǝsǝ 1109 Interrogative Particles 1118 2.1 Interrogative Particle ya 1118 2.2 Interrogative Particle ka 1127

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

Contents

3 4

5

xi

Desiderative Particle mǝŋga ~ mǝŋgamǝ 1137 Emphatic Particles 1141 4.1 Emphatic Particle kamǝ 1141 4.2 Emphatic Particle si 1154 4.3 Emphatic Particle mǝ 1162 4.4 Emphatic Particle ya 1166 4.5 Emphatic Particle yǝ 1171 4.6 Emphatic Particle na 1174 4.7 Emphatic Particle wo 1177 Restrictive Particles 1179 5.1 Restrictive Particle nǝmï 1179 5.2 Restrictive Particle ⁿdani 1182 5.3 Restrictive Particle sura ~ sora 1186 5.4 Restrictive Particle sapɛ 1188 5.5 Restrictive Particle mbakari 1192

Chapter 10 Postpositions 1 2 3 4

Postposition ⁿzi ~ ⁿzimənə ‘Like’ 1199 Postposition sambï ‘Like’ 1202 Postposition məkərə ‘Like’ 1204 Postposition pa ‘Every’ 1206

Chapter 11 Interjections 1 2 3 4 5 6 7 8

Interjection iⁿza 1213 Interjection ina 1215 Interjection wo 1217 Interjection ani 1218 Interjection ana 1219 Interjection wasi 1220 Interjection apare 1221 Interjection apare 1223

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

xii

Contents

Bibliography 1225 Index of Personal Names 1243 Index of Terms 1246 Index of Forms and Constructions 1252 Index of the Examples from the Texts 1262

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

Acknowledgements I first thought about writing this book almost a quarter of a century ago when I started to create an electronic database of Old Japanese texts, as it was clear to me that without such a database no thorough study of Old Japanese would be possible nowadays. It was a long and arduous process, requiring careful typing of both Old Japanese original texts in the man’yōgana script and its transliteration. One present and two former graduate students helped as graduate assistants during this long span of time: Patricia Welch (at the University of Michigan), John Bentley, and Kerri Russell (both at the University of Hawaii at Manoa). Many electronic databases featuring some (but not all) Old Japanese texts have appeared in Japan since then, but my students helped me create a unique database that includes practically all Old Japanese texts in phonetic writing. I am grateful for their help, as without it this book would never have been possible. I would like to express my gratitude to two institutions: the Center of Japanese Studies at the University of Hawaii which generously supported my research on Old Japanese grammar with two grants, and the Kokusai Nihon bunka kenkyū sentaa (International Center for the Japanese Studies, Kyoto), where I spent my sabbatical year in 2001–02, when the first draft of the first edition of this book was written. Various parts of this book, or some of the ideas that appear in it, were discussed with a number of colleagues. I am most grateful to Bjarke Frellesvig, with whom I had many discussions on Old Japanese grammar in United States, Japan, and Denmark, and who finally persuaded me of the reality of the category of tense in Old Japanese. Discussions of various Japanese, Ryukyuan, Altaic, Korean and Ainu issues with José Andrés Alonso de la Fuente, Bjarke Frellesvig, Ikegami Jirō,1 Juha Janhunen, Alexandra Jarosz, Itabashi Yoshizō, Karimata Shigehisa, Kazama Shinjirō, Ross King, Kirikae Hideo, Nakagawa Hiroshi, Mehmet Ölmez, Osada Toshiki, Sakiyama Osamu, Leon Serafim, Suda Jun’ichi, Tsumagari Toshiro, John Whitman, and Janick Wrona were extremely helpful. I am also grateful to Stefan Georg, who opened my eyes to the illusory nature of the ‘Altaic’ theory as it is currently enshrined. This recovery of sight resulted in a considerable rewriting of the comparative part of the book, causing about two years’ delay. However, hopefully this delay was for the best. As always, I am grateful to late Samuel E. Martin for his constant feedback on many 1  Japanese and other East Asian names are given in this book in the traditional order: family name first, given name last.

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

xiv

Acknowledgements

issues. It is the greatest pleasure for a teacher to receive feedback from former students who have now become independent scholars: John Bentley, Blaine Erickson, Hino Sukenari, Marc Miyake, and Shimabukuro Moriyo are now valuable colleagues who helped me in various ways. Needless to say, I alone am responsible for any mistakes or shortcomings in this book. Several generations of graduate students who took my seminar in Old Japanese at the University of Hawai’i at Mānoa and then at the École des hautes études en sciences sociales deserve special mention: James Baskind, John Bentley, Blaine Erickson, Timothy Harris, Hino Sukenari, Steven Ikier, Kao Hsiang-Tai, Linda Lanz, Younès M’Ghari, William Matsuda, Marc Miyake, Matthew McNicoll, Matthias Nyitrai, Kerri Russell, and Shimabukuro Moriyo. Their questions and inquisitive minds led to the improvement of many topics presented here. My foremost gratitude goes to Irene Jager of Brill, who was a very patient and wise guide for me throughout the process of the work on the manuscript of this grammar. Last, but not least, my gratitude goes to members of my family: my late mother Svetlana, my wife, Sambi, who helped me in many ways, our son Jacob Tomotatsu, and our daughter Marie Alexandra who helped simply by behaving well. This book is dedicated to Sambi.

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

Preface to the First Edition, Part One This book represents the first volume of the Descriptive and Comparative Grammar of Western Old Japanese. The first volume covers the phonology, writing system, lexicon, and nominal parts of speech. The second volume, includes chapters on adjectives, verbs, adverbs, particles, and conjunctions, as well as a comprehensive index for both volumes. I decided to split the publication into two volumes for three major reasons. First, as the reader can see, the first volume already has 400 pages, and the second volume will be twice as long. Producing a 1,200-page grammar in one volume is not only technically challenging, but it also creates a number of inconveniences for the reader who will have to handle such a heavy brick. Second, previous scholarship concentrated heavily on the verbal system; as a result Old Japanese nominals were neglected and not sufficiently described. I hope to rectify this unbalanced situation with this volume as well as present a number of innovative ideas. Third, and most importantly, in the process of my teaching Old Japanese over the last ten years at the University of Hawaii at Manoa, I found that the lack of an up-to-date Old Japanese grammar creates a serious obstacle for graduate students who want to learn the language and start reading Old Japanese texts on their own. From this point of view, publishing the first volume at an earlier date rather than delaying it until the second volume becomes available offers at least a partial solution to the problem. There is a wide misconception that Old Japanese of the Asuka–Nara periods and Classical Japanese of the Heian period are essentially one and the same language. This is not true, since the differences between Old and Classical (or, better: Middle) Japanese are very significant. These differences are found in all levels of the language—in its phonology, lexicon, morphology, and syntax. Quite a number of grammatical forms used in Old Japanese did not survive into the later language, and a great number of Middle Japanese forms are late innovations not found in Old Japanese. In addition, certain forms, common for both periods, frequently have different functions. All this demonstrates that a person who is going to read and analyze Old Japanese texts just armed with knowledge of Classical Japanese will fare no better than a person who tries to read Old English texts through the prism of Middle English. One must also keep in mind that Old Japanese and Middle Japanese are based on two geographically close and in some respects similar dialects: the dialect of the Yamato plain for Old Japanese and the dialect of Heian-kyō (Kyoto) for

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

xvi

Preface to the First Edition, Part One

Middle Japanese, both belonging to Central Japonic. Nevertheless, these are two different dialects, and there is no direct continuity between them. Finally, Old Japanese texts offer evidence for yet another different language: Eastern Old Japanese, which belongs to a completely different dialect group. This grammar deals predominantly with the description of Western Old Japanese from the Yamato plain. The data from Eastern Old Japanese are used only for comparative purposes, thus the reader will not find here a description of any Eastern Old Japanese facts that have no counterparts in Western Old Japanese. Besides Eastern Old Japanese, I also used comparative data from different temporal and geographical varieties of the Ryukyuan language group. A word of caution is in order: besides being a descriptive grammar of Western Old Japanese, this book is also a comparative grammar of Western Old Japanese, but it is not a comparative grammar of the Japonic language family. As the reader will see on several occasions, there are certain differences between early and late Western Old Japanese. Roughly speaking, the tentative boundary between these two varieties is somewhere in the late 740s or early 750s, although there are exceptions. For example, MYS 15 includes poems traditionally dated between 736–741 and is better defined as a text written in late Western Old Japanese. Besides Japonic data, I offered in some cases possible external etymologies for Western Old Japanese morphological markers. This involves data from the ‘Altaic’ languages (Turkic, Mongolic, Tungusic, and Korean), but sometimes I also used data from Austronesian and Ainu. I do not think that the Japonic language family is a member of the ‘Altaic’ language family in the genetic sense (although a remote genetic relationship cannot be denied or disproved), but I think that it is most certainly a member of the ‘Altaic’ Sprachbund, with most intimate ties to Korean. As the reader will find on more than one occasion, there are interesting morphological parallels between Western Old Japanese and Korean which are not found in Eastern Old Japanese and Ryukyuan. I believe that these parallels represent early loans from some variety of Old Korean, possibly the language of Paekche, into Western Old Japanese. There are also possible parallels between Western Old Japanese and other ‘Altaic’ languages, which, in my opinion also represent traces of ancient contacts. In addition, there are few cases when Western Old Japanese has likely parallels in Ainu or Austronesian, which I believe also presents evidence of contact. A few words about the presentation of the examples are in order. Examples from Western and Eastern Old Japanese texts are given in four lines: the first line provides the original spelling, the second the transliteration with morpheme breaks, the third glosses, and the fourth the English translation. The same is applicable to the examples from Old Ryukyuan and Old Korean texts.

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

Preface to the First Edition, Part One

xvii

Upper case letters in the second line render the semantographic spelling, and lower case letters the phonetic one. I opted for transliteration rather than transcription for the following reasons. First, although it is believed that the orthographic distinction between kō-rui 毛 /mwo/ and otsu-rui 母 /mø/ in Western Old Japanese is preserved only in the Kojiki, it has been convincingly demonstrated recently that it also at least statistically survives into some later texts (Bentley 1997, 2002). Thus, I transliterate 毛 as /mwo/ and 母 as /mø/ in all Old Japanese texts. Second, it is presumed that Eastern Old Japanese had only five phonemic vowels: /a/, /u/, /i/, /o/, and /e/. However, Eastern Old Japanese texts are written in the Western Old Japanese orthography, which differentiated seven vowels /a/, /u/, /i/, /ï/, /o/, /´/, /e/, and a diphthong /´y/. As a result, EOJ vowel /i/ is spelled as WOJ /i/ or /ï/, EOJ /o/ as WOJ /o/ or /´/, and EOJ /e/ as /e/ or /´y/. It might seem that these variations could be ignored, and transcription uniformly rendering those as /i/, /o/, and /e/ would be less confusing. However, there is a problem. While EOJ /e/ is indeed spelled as WOJ /e/ or /´y/ with almost the same frequency, there is a clear tendency for a frequent variation of other vowel spellings only in certain syllables. One has also to keep in mind that Eastern Old Japanese is a dialect continuum, which was not homogeneous and could be roughly divided into three areas. The study of spelling variation within these areas has not been done so far. Therefore, unifying the transcription of Eastern Old Japanese may be a premature enterprise: we simply cannot throw out the baby with the bathwater. Thus, transliteration, a more conservative approach, is preferable for the time being. An even more spectacular case involves Old Ryukyuan. It is generally believed that proto-Ryukyuan vowels *o and *e in Old Ryukyuan merged with *u and *i respectively, as they have done in the modern Shuri dialect. Therefore, the tendency is to transcribe Old Ryukyuan syllables spelled with kana /Co/ and /Ce/ syllables as /Cu/ and / Ci/ respectively. However, some Old Ryukyuan spellings may be very indicative and reflect the stage before the merger (Serafim, p.c.). Thus, for example, the Old Ryukyuan word for ‘dew’ is spelled predominantly as tuyo, not tuyu, and the former is supported by dialect comparative evidence, which points to PR *tuyo, not *tuyu. Therefore, before a thorough study of Old Ryukyuan spellings is done, transliteration is also preferable. Textual examples from other languages that include various modern Ryukyuan languages and dialects, and ‘Altaic’ languages are given in three lines: the first line provides transcription with morpheme breaks, the second glosses, and the third gives the English translation. Transcription rather than transliteration seemed to be a better solution here, because the data largely came either from modern languages without a writing system, or from old languages with an established practice of transcription. Whenever possible,

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

xviii

Preface to the First Edition, Part One

I tried to preserve the transcription from original sources, but in certain cases I opted for simplification and unification. For example, I used /y/ for the palatal voiced glide throughout, replacing the IPA [j] that some sources on Ryukyuan and ‘Altaic’ employed. When discussing the Eastern Old Japanese data, I frequently employ the reference to three different regions in the Eastern Old Japanese linguistic area: Region A, Region B, and Region C (Hino 2003: 197–201). The provinces are divided as follows: Region A—Simotuke, Kamitufusa; Region B—Hitati, Simotufusa, Kamituke, Musasi, Sagami, Suruga; and Region C—Sinano, Tofotafumi. The division into three regions is due to the fact that Eastern Old Japanese represented a dialect continuum rather than a single language. Region C was the one that was most influenced by Western Old Japanese; Region A was the least, with Region B representing the intermediate stage. Alexander Vovin

Hawaii, December 2004

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

Preface to the First Edition, Part Two This book represents the second part of the Descriptive and Comparative Grammar of Western Old Japanese. The first part covered the phonology, writing system, lexicon, and nominal parts of speech. The second part includes chapters on adjectives, verbs, adverbs, conjunctions, particles, and postpositions, as well as comprehensive indexes for both volumes. As already mentioned in the acknowledgements, several important technical developments happened between the publication of the first part and the completion of the second part. As a result, a number of modifications and expansions were introduced into the second part that need to be mentioned here. First, with the availability of the Man’yōshū CD-ROM (Kinoshita 2001), a new index to the Man’yōshū (Kinoshita et al. 2003) which is far superior to an older index to the Man’yōshū (Masamune 1974), and Sven Osterkamp’s Man’yōshū Searcher engine, there was no longer a necessity to treat some volumes of the Man’yōshū as major sources, and others as supplementary. In the second part all text of the Man’yōshū is treated as the major source. Still, when citing textual examples, I normally give preference to the phonetically or partially phonetically written volumes of the Man’yōshū: books 5, 15, 17, 18, 19, and 20. The reader will notice that particularly in chapter six, dedicated to the description of the Western Old Japanese verb, the examples from book 15 of the Man’yōshū are especially frequent. This is due to the fact that MYS 15, in spite of numerous phonetic misspellings found there, is probably the most important Western Old Japanese text as far as the grammatical system is concerned: many forms attested therein are not found anywhere else in the Western Old Japanese corpus. Also, I frequently opted to provide more examples from MYS 15 because it is not dominated by one single poet like, for example, books 17–20 are clearly dominated by Opotömö-nö Yakamöti. Second, before I started to write the second part, I switched over from Macintosh to PC. This gave me access to the Mojikyō map for Mojikyō fonts that was not available for Macintosh four years ago. Consequently, there was no longer the necessity to use substitute characters for the man’yōgana signs that were not easily accessible or altogether lacking in Macintosh. Thus, the second part reproduces Western Old Japanese in its original form, without any substitute characters. If a second edition of this book ever materializes, the appropriate corrections will be made to part one as well.

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

xx

Preface to the First Edition, Part Two

Third, I have introduced two minor changes into my transcriptional conventions. First, due to the same switch to PC, long vowels in modern Japanese are no longer written with an accent circumflex mark, but with a macron over the vowel (that was extremely difficult to type on the Macintosh). Thus, for example, in the second part the reader will see Man’yōshū and not Man’yôshû as in the first part. Second, the otsu-rui vowel /o₂/ that was spelled as ø in part one is transcribed as ö in part two. This is due to feedback received from some colleagues who suggested that ø looks too alien to Japanologists. Since, while I am using Yale transliteration for Old Japanese, I was and am still reluctant to use its notation o̱ for the otsu-rui vowel /o₂/, I opted for a compromise and borrowed ö from the traditional transliteration of Old Japanese. I hope that my readers will view these changes only as minor inconveniences. Fourth, over the years that this book was written my understanding and/or analysis of some Western Old Japanese expressions and passages has changed. This resulted, on some (although not numerous) occasions, in discrepancies between chapter four in the first part and the following chapters in the second part. To give an example, I used to understand the expression pyitö kuni ‘land/province of other people’ as consisting of the numeral pyitö ‘one’ and kuni ‘land, province.’ Now I prefer to analyze it as consisting of pyitö ‘person’ and kuni ‘land.’ The overall meaning of the expression does not change, but the analysis does. Rather than preserving the uniformity with the first part and repeating the erroneous analysis, I opted for the discrepancy and the correction of what I believe was a mistake. Fifth, the reader will undoubtedly notice the paucity of examples from the Norito in the second part as compared to the Senmyō. There are three main reasons for what may seem to be a neglect, partially outlined in 1.1. The Norito is a heterogeneous text, and the sixteen oldest Norito representing Western Old Japanese and not Middle Japanese are comparatively short as compared to the Senmyō. The text of Norito also uses much more semantographic writing as compared to the Senmyō, therefore it is not as valuable as the latter. Finally, there is already a grammar of Norito in English (Bentley 2001), and I was reluctant to duplicate his work. Thus, the examples from Norito were used only if a point in question could not be illustrated by any other Western Old Japanese texts. Alexander Vovin

Honolulu, October 2007

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

Preface to the Second Edition and Acknowledgements The first question that needs to be answered is: “Why the second edition?” The response to this question is multilayered. First, the first edition (Vovin 2005, 2009) was in two volumes because of the small format of the page used at the Global Oriental, the publisher for the first edition. Second, the second volume of the first edition was printed with the minimal number of copies, so it was almost immediately sold out, and almost for ten years only the electronic copy was available. Third, the second volume (2009) has special ADDITIONS chapter that contains supplemental information for the first volume (2005). This made the usage of the grammar more difficult and cumbersome. Fourth, and most importantly, there have been new developments in the scholarship on Old Japanese within last ten years which need to be reflected. The next question is: “How is the second edition different from the first except the proper incorporation of the ADDITIONS chapter?” The response is also multilayered. First, the comparative part of the grammar is greatly reduced. In the early 21th century there still was a necessity to argue against the “Altaic” hypothesis of the origins of the Japonic language family. Since this hypothesis is no longer taken seriously except by some truly faithful from the Moscow Nostratic school and Max Planck institute in Jena, there is really no need to use trees for the unnecessary discussion. In short, the comparative part has been reduced to the discussion of the internal Japonic data, Eastern Old Japanese and Ryukyuan. The only exception when any external data are discussed is made for loans from Ainu, Chinese and Korean. Also, it was becoming gradually clear in recent years that Japonic is an intrusive language family in the maritime North-East Asia, and that it probably has its ultimate Urheimat in Southern China and/or Northern South-East Asia. Japonic is not related to any of the language families found there, but it probably was for a quite long time in contact with some of the language families that are found in this region. Second, the Eastern Old Japanese comparisons were revised considerably, mostly in the sense what texts can be taken as Eastern Old Japanese and which one cannot. The basis for this decision are two volumes of my translation of the Man’yōshū books fourteen and twenty that are currently our main source of information on Eastern Old Japanese (Vovin 2012, 2013).

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

xxii

Preface to the Second Edition and Acknowledgements

Third, the romanization of Old Japanese was completely changed in such a way that it would reflect with the maximum accuracy the actual phonetics (to the extent we can know it) with the maximum accuracy. This was done mostly for the benefit of the colleagues and students who have not had a chance to study the history of the Japanese language, which is a prerequisite to understand phonologically oriented transcriptions. A single example will be sufficient: Old Japanese prenasalized voiced obstruents are most commonly transcribed as plain voiced b, d, g, z, but in this grammar the prenasalization is indicated by superscript m, ⁿ, and ŋ: mb, ⁿd, ŋg, and ⁿz. Fourth, in spite of the reductions mentioned above, I also made important additions to Chapters 7, 8, 9, and 10, that were overlooked in the first edition. I have also added a new Chapter 11 on interjections. Thus, the grammar is still almost 1,000 pages long, but it is the nature of the Old Japanese language itself that is to blame for the grammar’s size. Fifth, some technical changes were made as well. The term ‘infinitive’ that is very misleading for Old Japanese was replaced with ‘converb.’ The notation of different types of particles that were lumped together as PT (= particle) in the textual examples of the first edition was subdivided into focus, interrogative, desiderative, emphatic, and restrictive particles. The numerical notation of the Man’yōshū poems was changed from ‘MYS I: 1’ to ‘MYS 1.1’ in order to make it uniform with my translation of this anthology (Vovin 2009b–2018). In early 2014 after a quarter of the century spent at three different American universities I made an important decision to move to France for the remainder of my professional life. I can now dedicate much more time to my research while training Ph.D. students in Old Japanese, Manchu, and Middle Mongolian. It seems that in every country of my residence I am able to make longlasting friendships. Not everyone from the following list comes from or lives in France, and only few people among those are professional colleagues or students. But certainly all of them made my new life more enjoyable and fruitful: Irène Tamba, Ross Bender, Augustin de Benoist, Elena Perekhvalskaya, Bjarke Frellesvig, John Whitman, William McClure, Dieter Maue, Mehmet Ölmez, Étienne de la Vaissière, Pierre Marson, Redouane Djamouri, Christiane Babiak, Sami Saleh, Vladimir Bokarius, Alexei Egorov, Juha Janhunen, Maria Chiara Migliore, D. Zayabaatar, Wu Ying-zhe (Oyunch), Laurent Sagart, Guillaume Jacques, Guillaume Carré, Anton Antonov, Michelle Abud, Takubo Yukinori, Osada Toshiki, Suda Jun’ichi, Anna Bugaeva and my three current PhD students: Gao Hsiang-Tai (Baatar), Younès M’Ghari and Etienne Baudel. I would also like to offer my thanks to several institutions that graciously supported my research: National Institute of the Japanese Language and

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

Preface to the Second Edition and Acknowledgements

xxiii

Linguistics (NINJAL), Ecole des hautes études en sciences socials (EHESS), European Research Council (ERC), and Mongolian Government. My special thanks go to my editors at Brill: Irene Jager and Patricia Radder. Last, but not least, my gratitude goes to members of my family that is now found in three countries: France, Japan, and Russia—my parents-in-law Ishisaki Tetsuo and Ishisaki Fukiko, my wife Sambi, and my three children: Lesha (Alexei), Yasha (Jacob), and Masha (Marie). This edition is dedicated to Sambi. Alexander Vovin

Poligny, Seine-et-Marne, August 23, 2019

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

Abbreviations Languages CR EMC EOJ LHC LMC MdJ MJ MK OC OJ OK OR OT PAN PJ PK PM PR PT PTu WM WOJ



Classical Ryukyuan Early Middle Chinese Eastern Old Japanese Late Han Chinese Late Middle Chinese Modern Japanese Middle Japanese Middle Korean Old Chinese Old Japanese Old Korean Old Ryukyuan Old Turkic proto-Austronesian proto-Japonic proto-Korean proto-Mongolic proto-Ryukyuan proto-Turkic proto-Tungusic Written Mongolian Western Old Japanese

Texts and Sources

Japonic

BS Bussoku seki no uta, 753 AD GGJEG Gangōji engi, 747 AD GSWKS Goshūi wakashū, 1087 AD FK Fudoki kayō, ca. 737 AD JDB Jidai betsu kokugo dai jiten (Omodaka 1967) KGU Kagura uta, 8–9th centuries KJK Kojiki, 712 AD KK Kojiki kayō, 712 AD MYS Man’yōshū, ca. 759 AD Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

Abbreviations

xxv

NK Nihonshoki kayō, 720 AD NR Nihon ryōiki, early 9th century NS1 Nihonshoki, 720 AD NSK2 Nihonshoki, 720 AD NT Norito, 7–9th centuries OM Ochikubo monogatari, 10th century OS Omoro sōshi, 16–17th centuries RK Ryūka, 17–19th centuries RKJ Okinawa go jiten SM Senmyō, 7–8th centuries SNK Shoku nihongi kayō SSI Shōsōin documents, 7–8th centuries TN Tosa nikki, 935 CE TS Jōgū Shōtoku höō teisetsu, 7th century

Korean KKK HK Nung PT SP YP WS



Kumkang kyeng samka hay, 1482 AD Hyangka, 6–10th centuries Nungem kyeng enhay, 1461 AD Pak thongsa, 1515 AD Sekpo sangcel, 1449 AD Yongpi ethyenka, 1447 AD Welin sekpo, 1459 AD

Grammar and Literary Terms

ADJ Adjectivizer ADV Adverbilizer AFFIR Affirmative ASSER Assertive ATTR Attributive BEN Benefactive CAUS Causative CL Classifier COM Comitative 1  Cited according to Kuroita Katsumi and Matuyama Jirō (ed.) 1965–66. 2  Cited according to Sakamoto Tarō; Ienaga Saburō, Inoue Mitsusada, Ōno Susumu (eds.) 1965–67.

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

xxvi

Abbreviations

COMP Comparative CON Conjunctive gerund CONC Concessive gerund COND Conditional gerund CONJ Conjunction CONJC Conjectural CONT Continuative CONV Converb COOP Cooperative COOR Coordinative gerund COP Copula DAT Dative DEB Debitive DES Desiderative DIR Directive DLF Directive-locative focus DP Desiderative particle DV Defective verb EMPH Emphatic EP Emphatic particle EV Evidential EXCL Exclamation, Exclamative FIN Final verbal form FP Focus particle GEN Genitive HON Honorific HORT Hortative HUM Humble IF Interrogative form IP Interrogative particle INF Infinitive INT Intensive INTL Intentional INTER Interjection LOC Locative MDL Modulator MK Makura-kotoba (Pillow word) MOD Modality NEG Negative

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

Abbreviations

xxvii

NML Nominalizer NOM Nominative OSM Oblique stem marker PAST Past tense PEJ Pejorative PERF Perfective POL Polite POSS Possessive POT Potential PREF Prefix PRES Present tense PRET Preterite PREV Preverb PROG Progressive PT Particle RA Reported action REC Reciprocal-cooperative RETR Retrospective RP Restrictive particle SUB Subordinative gerund SUBJ Subjunctive SUF Suffix SUP Suppositional TENT Tentative TERM Terminative TF Transitivity flipper TOP Topic TRANSF Transferential gerund VOL Voluntative

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

Charts 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33

Major differences between various systems of transcription 27 Man’yōgana signs in varieties A and B 30 WOJ consonants 41 Distribution of LMC signs with voiced *pɦ for syllables with intervocalic /p/ in the first 100 songs of the Nihonshoki kayō 43 Western Old Japanese vocalism 45 Lexical doublets in Western Old Japanese 65 Reflexes of um-/uⁿ - in WOJ, EOJ, MJ and Shuri 68 Allomorphs of the ablative case marker in Western Old Japanese 184 Distribution of the first person pronouns wa- and a- in the early and late OJ texts 217 First person pronoun wa/ware in combination with case markers 219 The paradigm of Shuri first singular pronoun waa/waN 227 Shuri nominals in combination with a topic particle ya 228 Personal pronoun a/are in combination with case markers 239 Second person pronoun na/nare in combination with case markers Demonstrative pronouns 262 Proximal demonstrative pronoun kǝ/kǝre 263 Combination of numerals with classifier -ka ‘day’ 363 Combination of numerals with the classifier -ka ‘day’ in Ryukyuan 365 Numeral roots and irregular forms for the count of days 365 Archetypes for the count of days 366 Native names for lunar months in Old Japanese and Middle Japanese 369 Classes of inflected adjectives in Western Old Japanese 390 Main inflectional forms of consonant verbs 451 Main inflectional forms of vowel verbs 452 Main inflectional forms of strong vowel verbs 454 Main inflectional forms of kǝ- ‘to come’ 455 Main inflectional forms of se- ~ -sǝ ‘to do’ 456 Main inflectional forms of r-irregular verbs 457 Main inflectional forms of n-irregular verbs 458 Inflectional forms of defective verbs 459 Distribution of the allomorphs of the final predication suffix 538 Combinations of the final predication suffixes -u and -i with preceding suffixes and bound auxiliaries 539 Distribution of the allomorphs of the attributive suffix 550

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

Charts 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55

xxix

Combinations of the attributive suffixes -uru and -u with preceding suffixes and bound auxiliaries 551 Attributive forms in Shuri 569 Attributive forms in Old Ryukyuan 571 Distribution of the allomorphs of the evidential suffix 573 Combinations of the evidential suffixes -ure and -ɛ ~ -e with preceding suffixes and bound auxiliaries 573 Combinations of the imperative suffixes -e ~ -ǝ with preceding suffixes and bound auxiliaries 584 Combinations of the extended imperative -yǝ with preceding suffixes and bound auxiliaries 593 Combinations of the negative imperative -una with preceding suffixes and bound auxiliaries 596 Final predication and negative imperative forms in the Shuri dialect 598 Combinations of the desiderative -(a)na with other suffixes and bound auxiliaries 599 Combinations of the subjunctive suffix -(a)masi with other suffixes and bound auxiliaries 606 Combinations of the suppositional suffix -urasi ~ -asi with other suffixes and bound auxiliaries 611 Combinations of the negative tentative suffix –aⁿzi ~ -ⁿzi with other suffixes and bound auxiliaries 617 Combinations of the negative potential suffix -umasiⁿzi with other suffixes and bound auxiliaries 622 Combinations of the exclamative -umǝ with preceding suffixes and bound auxiliaries 627 Combinations of the converb -i with other suffixes and bound auxiliaries 631 Combinations of the negative converb -(a)ⁿz-u with following bound auxiliaries 646 Combinations of the conditional converb -(a)mba with preceding suffixes and bound auxiliaries 655 Combinations of the conjunctive converb -mba with preceding evidential forms of suffixes and bound auxiliaries 664 Combinations of the conjunctive converb -ndǝ[mǝ] with preceding evidential forms of suffixes and bound auxiliaries 673 Combinations of the nominalizer -(a)ku with preceding suffixes and bound auxiliaries 688 Combinations of the negative suffix -(a)n- ~ -(a)ⁿz- with following suffixes 702

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

xxx 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79

Charts Combinations of the negative suffix -(a)n- ~ -(a)ⁿz- with preceding suffixes and bound auxiliaries 702 Combinations of the tentative suffix -am- ~ -m- with following suffixes 714 Combinations of the tentative suffix -am- ~ -m- with preceding suffixes and bound auxiliaries 714 Combinations of the tentative suffix -(u)ram- with following suffixes 730 Combinations of the tentative suffix -uram- ~ -ram- with preceding suffixes and bound auxiliaries 731 Combinations of the iterative suffix -ap- ~ -ǝp- with following suffixes 739 Morphophonological variations of the passive suffix 746 Combinations of the passive suffix -aye- ~ -raye- ~ -ye- with following suffixes and auxiliaries 747 Animacy of the agent and the patient in Western Old Japanese passive constructions 750 Combinations of the passive suffix -are- with following suffixes and auxiliaries 756 Combinations of the honorific suffix -as- with following suffixes 759 Combinations of the honorific converb -as-i- with following bound auxiliaries 760 Morphophonological variations of the causative suffix -asimɛ- ~ -simɛ- 771 Combinations of the causative suffix -asimɛ- ~ -simɛ- with following suffixes and bound auxiliaries 772 Morphophonological variations of the causative suffix -as- ~ -(a)se- 777 Combinations of the causative suffix -as- ~(a)se- with following suffixes and bound auxiliaries 778 Combinations of the debitive suffix –umbɛ- ~ -mbɛ- with following suffixes 785 Combinations of the debitive suffix –umbɛ- ~ -mbɛ- with preceding suffixes 786 Combinations of the progressive suffix -er- with following suffixes and bound auxiliaries 793 Combinations of the progressive suffix -er- with preceding suffixes and bound auxiliaries 793 Combinations of the subordinative converb -te with preceding suffixes and bound auxiliaries 805 Combinations of the coordinative converb -tutu with preceding suffixes and bound auxiliaries 819 Combinations of the past auxiliaries -ki, -si, and -sika with following suffixes 827 Combinations of the past auxiliaries -ki, -si, and -sika with the verbs kǝ- ‘to come’ and se- ‘to do’ 828

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

Charts 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106

xxxi

Combinations of the past final -ki with preceding suffixes and bound auxiliaries 829 Combinations of the past attributive -si with preceding suffixes and bound auxiliaries 829 Combinations of the past evidential -sika with preceding suffixes and bound auxiliaries 830 Combinations of the perfective -n- with preceding suffixes in their converb form 845 Combinations of the perfective -n- with following suffixes and bound auxiliaries 845 Combinations of the perfective -te- with preceding suffixes in their converb form 856 Combinations of the perfective -te- with following suffixes and bound auxiliaries 857 Combinations of the perfective-progressive -tar- with preceding suffixes in their converb form 866 Combinations of the perfective-progressive -tar- with following suffixes and bound auxiliaries 867 Combinations of the retrospective -ker- with preceding suffixes and auxiliaries in their converb form 879 Combinations of the retrospective -ker- with following suffixes 880 Combinations of the potential -kate- with following suffixes 889 Combinations of the negative potential -kane- with preceding converbs 894 Combinations of the negative potential -kane- with following suffixes and auxiliaries 894 Combinations of the benefactive -kǝse- with preceding converbs 899 Combinations of the benefactive -kǝse- with following suffixes 899 Combinations of the honorific tamap- with preceding converbs 903 Combinations of the honorific wos- with following suffixes 917 Combinations of the honorific kikǝs- with following suffixes and auxiliaries 919 Combinations of the reported action auxiliary nar- with following suffixes 947 Combinations of the topic particle pa with preceding morphemes 1067 Combinations of the topic particle pa with following particles 1068 Combinations of the focus particle mǝ with preceding morphemes 1082 Combinations of the focus particle mǝ with following particles 1083 Statistics for the focus particle sǝ ~ nzǝ in the earliest texts 1094 Combinations of the focus particle sǝ ~ nzǝ with preceding morphemes 1095 Combinations of the focus particle namo with preceding morphemes 1105

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

xxxii

Charts

107 Combinations of the focus particle kǝsǝ with preceding morphemes 1110 108 Combinations of the focus particle kǝsǝ with the following particles 1110 109 Combinations of the interrogative particle ya with different verbal forms of final predication 1118 110 Combinations of the emphatic particle si with preceding morphemes 1154 111 Combinations of the emphatic particle si with following particles 1155 112 Combinations of the emphatic particle mǝ with other particles 1162 113 Combinations of the emphatic particle ya with other particles 1167

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

chapter 1 Sources and Previous Scholarship



Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

Contents of Chapter 1 Chapter 1 Sources and Previous Scholarship 4 1 Sources 4 2 Previous Scholarship 20

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

Section 1

Sources This book is predominantly based on Asuka (592–710 CE) and Nara (710–794 CE) periods literary texts written completely or almost completely phonographically in man’yōgana syllabic signs (see Chapter 2 for details), with the exception of the Senmyō and Norito, which have few elements written phonographically. Nevertheless, these two texts were used extensively in this study, as they are the only representatives of Old Japanese prose. Other texts that are written either completely logographically, or just partially phonographically were also consulted. However, I only used these sources occasionally as supplemental materials. Supplemental sources were consulted only in cases where the described grammatical forms were either not present at all in major sources or their attestation in major sources was limited and/or insufficient for descriptive purposes. All of the major sources listed below were digitized between 1993–1999 both in their romanization and man’yōgana forms, constituting the second largest database of phonographically written Old Japanese texts available electronically.1 Major sources are as follows (in chronological order): 1) The Kojiki kayō [The Songs of the Kojiki] (古事記歌謡, KK)—112 poems (113 in an alternative count) from the Kojiki [The Records of Ancient Matters] (古事記), written completely phonographically. The narrative text itself is written in hentai kanbun—Classical Chinese that is influenced considerably by Japanese, and even includes some elements, completely alien to Chinese, such as, e.g., the character 坐 used to indicate the Old Japanese honorific auxiliary -[i]mas-. Besides the above-mentioned poems, the text of the Kojiki includes a great number of personal names, place names, and even nonproper nouns recorded phonographically. However, this material is more of an interest to a lexicographer, and since the present monograph centers on Old Japanese morphosyntax rather than on Old Japanese vocabulary, for the most part it has been left out. The Kojiki text was edited by Ō-no Yasumaro on the orders of Empress Genmei (661–721 CE, ruled in 707–715 CE) in 711 CE and took its final shape in 712 CE, the third year of the Nara period. However, Ō-no Yasumaro’s work was based on the corpus of the Teiō hitsugi [The Imperial 1  Oxford-NINJAL electronic corpus of Old Japanese that appeared since then (Frellesvig 2009) is overall bigger as it includes many logographically written texts, but it presents no glossing.

© Alexander Vovin, 2020 | doi:10.1163/9789004422810_002

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

Sources and Previous Scholarship

5

Genealogies] (帝王日継) and the Sendai Kuji [Ancient Traditions] (先代旧辞), which Emperor Tenmu (d. 686 CE, ruled 673–686 CE) ordered Hieda-no Are to recite from his/her2 memory. It is quite possible that other historical sources of the seventh century, no longer extant today, such as the Tennōki [The History of the Imperial Family] (天皇記), the Kokuki [The History of the State] (国記), and the Hongi [The Official History] (本記) were used, too. All Kojiki poems are written completely phonographically, without a single use of a logogram. The language is very archaic, judging from both the spelling system and grammar. Certain grammatical forms and vocabulary items are difficult to explain, so abundant differences in commentaries and explanations exist. There are numerous modern editions with commentaries of the whole Kojiki text as well as of the texts of poems alone. The Nihon koten bungaku taikei [Series of the Japanese Classical Literature] from the Iwanami publishing house include both Tsuchihashi Yutaka’s edition and commentary of the poems texts themselves (Tsuchihashi 1957) and Kurano Kenji’s edition and commentary of the whole Kojiki text (Kurano 1958). Tsuchihasi has also published later a more extended commentary on the Kojiki poems (Tsuchihashi 1972). The Nihon koten bungaku zenshū [A Complete Collection of Japanese Classical Literature] from the Shōgakukan publishing company has Ogihara Asao’s commentary of the Kojiki text (Ogihara and Kōnosu 1973). Another comprehensive edition of the entire Kojiki text is found in the Nihon shisō taikei [Series on Japanese Thought] from the Iwanami publishing house (Aoki, Ishimoda, Kobayashi, and Saeki 1982). One more detailed commentary on the texts of the poems is by Aiso Teizō (Aiso 1962). Finally, the fundamental Kojiki taikei (Series on Kojiki) has to be mentioned: it includes an edition of the text with commentaries (Kurano 1957), as well as detailed indexes (Takagi and Toyama 1974a, 1977), and an edition of the text tailored to indexes (Takagi and Toyama 1974b). As mentioned earlier, the various commentators of the Kojiki text disagree on the song count: Tsuchihashi and Aoki et al. recognize 112 songs, while Ogihara, Kurano, and Aiso count 113. The different count starts pretty early in the text: Ogihara, Kurano and Aiso count the KK 3 in Tsuchihashi and Aoki’s count as two poems, so Tsuchihashi and Aoki’s KK 4 becomes KK 5 in Ogihara, Kurano and Aoki’s editions, etc. I have followed Tsuchihashi and Aoki et al.’s count of 112 poems. In most, but not all cases, I also followed Tsuchihashi in his interpretation of the Kojiki poems. There are numerous cases, especially when it comes to obscure 2  It is not known for sure whether Hieda-no Are was male or female, but since the female name Areko is attested in the Mino province census, the first option is much more likely.

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

6

Chapter 1

passages and words, when I opted for interpretations based on other commentaries mentioned above or even used my own solutions. This critical approach has been adopted not only for the Kojiki kayō text but also for the other Old Japanese texts as well. Although such an approach may be seen as eclectic by some, in my opinion it constitutes the only possible option, if one wants to get the most pertinent information on the language of the period, rather than to blindly follow a single commentator. 2) Four poems from the Jōgū Shōtoku hōō teisetsu [A Biography of Prince Shōtoku, King of the Law] (上宮聖徳法王帝説, TS). The Jōgū Shōtoku hōō teisetsu is a biography of the prince Shōtoku Taishi. The author is unknown. It is believed to have been compiled in the beginning or the middle of the Heian period, but it is undoubtedly based on the materials from the second part of the seventh century. The text itself is written in kanbun (Japanese version of Classical Chinese), but there are four poems and many personal names written phonographically in Old Japanese. Judging by the consistency of spellings (there is no confusion between kō-rui /mwo/ and otsu-rui /mǝ/), the poems are likely to belong to Asuka period. The Jōgū Shōtoku hōō teisetsu is published with extensive commentaries in the Shōtoku Taishi shū by Ienaga Saburō and Tsukishima Hiroshi (Ienaga, Fujieda, Hayashima, and Tsukishima 1975) in the Nihon shisō taikei series. 3) The Nihonshoki kayō [The Songs of the Nihonshoki] (日本書紀歌謡, NK)—128 poems from the Nihonshoki [The Annals of Japan] (日本書紀), written completely phonographically. The narrative portions of this text, in contrast to the Kojiki are written in standard Classical Chinese. Similar to the Kojiki, a great number of personal and place names are also written phonographically, although the number of non-proper nouns recorded is significantly less than in the Kojiki. Similar to the same kind of materials in the Kojiki, this material is more of interest to a lexicographer, and since the present monograph centers on Old Japanese morphosyntax rather than on Old Japanese vocabulary, for the most part it has been left out. The chief compiler of the Nihonshoki is believed to be prince Toneri (676?–735 CE), the third son of Emperor Tenmu, who finished the compilation of this work with his companions, including Ō-no Yasumaro, the compiler of the Kojiki, in 720 CE. A significant number of poems recorded in the Nihonshoki overlap with those from the Kojiki. However, since the Nihonshoki mostly uses a different type of spelling system than the Kojiki and other Old Japanese texts, it is actually a blessing rather than a curse for a linguist who wants to reconstruct Old Japanese phonology. Both Mori (1991) and Miyake (1999, 2003) utilized this fact to a considerable advantage in their works. The grammar of the language used in the Nihonshoki is as archaic as

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

Sources and Previous Scholarship

7

that in the Kojiki, but the spelling system in all probability reflects the phonology of the early Nara period rather than the Asuka period. There are numerous modern editions with commentaries of the entire Nihonshoki text as well as of the texts of poems alone. The Nihon koten bungaku taikei [Series of Japanese Classical Literature] from the Iwanami publishing house include both Tsuchihashi Yutaka’s edition and commentary of the poems texts themselves (Tsuchihashi 1957) and Sakamoto Tarō, Ienaga Saburō, Inoue Mitsusada, and Ōno Susumu’s edition and commentary of the entire Nihonshoki text (Sakamoto et al. 1965–67). Tsuchihashi has also published a more extended commentary on the Nihonshoki poems (Tsuchihashi 1976). The Nihon koten bungaku zenshū [A Complete Collection of Japanese Classical Literature] from the Shōgakukan publishing house has Kōnosu Hayao’s commentary of the Nihonshoki poems texts (Ogihara and Kōnosu 1973). One more detailed commentary on the texts of the Nihonshoki poems is by Aiso Teizō (Aiso 1962). An important edition without commentaries of the entire Nihonshoki text, which shows textual differences from different variants, is that by Kuroita Katsumi and Maruyama Jirō (1965–66). In most, but not all cases, I followed Tsuchihashi’s commentaries (Tsuchihashi 1957, 1976) in his interpretations of the Nihonshoki poems. There are numerous cases, especially those that contained obscure passages and words, when I opted for interpretations based on the other commentaries mentioned earlier or devised my own solutions. 4) The Fudoki kayō [The Songs of Fudoki] (風土記歌謡, FK) contains twenty poems. The Fudoki [Gazetteers] (風土記) are geographical descriptions of Old Japanese provinces compiled between 713 and 737 CE, on the orders of Empress Genmei in 713 CE. Only the Izumo fudoki (出雲風土記), the description of the Izumo province is preserved in full; somewhat less complete are the Hitachi fudoki (常陸風土記) and the Harima fudoki (播磨風土記), descriptions of Hitachi province in Kantō and Harima province in Kinki. Even less well preserved are the Hizen fudoki (肥前風土記) and the Bungo fudoki (豊後風土記), the descriptions of Hizen and Bungo provinces in Kyūshū. Descriptions of all other provinces survived only in fragments, if at all. The text of the Fudoki is written in kanbun, but it contains a significant number of personal names and place names written phonographically in addition to the aforementioned poems. Among the twenty poems, nine belong to the Hitachi fudoki. Although not universally recognized, it appears that among the nine poems from the Hitachi fudoki six (FK 2, FK 3, FK 5, FK 7, FK 8, and FK 9) are written in the Eastern Old Japanese, which should come as no surprise, since Hitachi province is located in the Azuma region. On the other hand, it is surprising that

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

8

Chapter 1

the seven other poems from Hitachi have no apparent Eastern Old Japanese features. This might suggest that the compilers of the Hitachi Fudoki did some editing. There are several modern editions of the Fudoki texts as well as of its poems. The main edition I relied on was Tsuchihashi’s edition of the poems with his commentary (Tsuchihashi 1957), but other editions were also consulted: Kōnosu’s edition of the Fudoki poems (Ogihara and Kōnosu 1973) and Akimoto Kichirō’s edition of the whole Fudoki text (Akimoto 1958). 5) The Man’yōshū [Anthology of Myriad Leaves] (万葉集) is the largest Japanese poetic anthology, consisting of 4,516 poems. The Man’yōshū is believed to be compiled chiefly by Ōtomo-no Yakamochi (717?–785 CE), a midand late-Nara period politician and poet, sometime in or soon after 759 CE.3 Although the Man’yōshū, which consists of twenty volumes (maki, 巻), is the largest extant Old Japanese text, it is also the most heterogeneous in a linguistic sense. It includes the texts written both phonographically, partially phonographically, and almost non-phonographically. Some parts of the Man’yōshū are written not in the standard Western Old Japanese dialect of the capital Nara and the surrounding areas, but in the Eastern Old Japanese dialect of the Kantō and Southern Chūbu (at that time known as Azuma) areas (considerable parts of the volumes 14 and 20). Chronologically, the Man’yōshū includes poems composed from the late sixth century, if not earlier, to 759 CE. The present study relies heavily on volumes 5, 15, 17, 18, and 20, which reflect Western Old Japanese recorded phonographically or almost completely phonographically, as well as on volumes 14 and 20, where Eastern Old Japanese has been recorded phonographically. In spite of this heavy emphasis on phonographically written texts, other volumes have not been neglected either, and a substantial number of examples from these volumes is provided as well. The heterogeneous nature of the Man’yōshū text, as well as its sheer volume, calls for a brief description of each individual volume in this anthology. MYS 1 (#1–#84) includes eighty-four miscellaneous poems (zōka, 雑歌),4 arranged in chronological order. Tradition attributes the poems of MYS 1 from the 3  The last poem in the Man’yōshū, MYS 20.4516, was composed by Ōtomo-no Yakamochi at the New Year’s banquet in 759 CE. 4  Miscellaneous poems (zōka) represent poems composed on occasions of royal trips, banquets, journeys, nature, miscellaneous objects, etc. The classification of poetic genres in the Man’yōshū is based largely on the Chinese model, but at many times it turns out to be difficult to squeeze Japanese poetry into the Procrustean bed of the Chinese one. Quite frequently, the Chinese genre classification is violated, or even outright replaced with a more appropriate Japanese one, as, e.g., mondōka ‘question and answer poems’ alien to the Chinese poetic system.

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

Sources and Previous Scholarship

9

rule of Emperor Yūryaku (雄略, 457–473 CE) to the Nara period (710–784 CE), but most likely the earliest specimens do not predate the early seventh century. There are different authors in this volume, but there is quite a significant percentage of poems written by or attributed to emperors and empresses. This volume exhibits predominantly mixed logographic and phonographic spelling, although quite a few songs are written with very few phonographic elements at all, or, just the opposite include quite a good portion of phonographic writing. My own electronic database of Old Japanese texts includes only MYS 5, 14, 15, 17, 18, and 20, which constitute the major sources for this grammar. MYS 1 was used as a supplemental source for this grammar, but I did not digitize it myself nor any of the other remaining volumes of the Man’yōshū. However, now an electronic version of it is available on a new Man’yōshū CD-ROM (Kinoshita 2001). MYS 2 (#85–#234) includes 150 poems of love and affection (sōmon, 相聞)5 and elegies (banka, 挽歌),6 also arranged in chronological order. Tradition attributes the poems of MYS 2 from the rule of Emperor Nintoku (仁徳, 313–393 CE) to the 715 CE, but most likely the earliest specimens do not predate the early seventh century, either. The poems were composed by various authors. The spelling is largely logographic, with a few phonographic elements, normally indicating particles or (more seldom) other grammatical elements. MYS 2 was used as a supplemental source for this grammar. MYS 3 (#235–#483) includes 249 poems in the genres of zōka, banka, and allegorical poems (hiyuka, 比喩歌). Not arranged in chronological order, but according to tradition the poems range from the end of the sixth century to 748 CE. As is the case with the two previous volumes, the earliest poems are unlikely to be earlier than the early seventh century. The poems were composed by various authors. The spelling is largely logographic, with few phonographic elements. The spelling system is very close to the one employed in MYS 2, although some grammatical elements appear to be spelled phonographically more frequently than in MYS 2. MYS 3 was used as a supplemental source for this grammar. MYS 4 (#484–#792) includes 309 poems in the sōmon genre, arranged chronologically according to the tradition from the rule of Emperor Nintoku (仁徳, 313–393 CE) to the 748 CE, but most likely the earliest specimens do not predate early seventh century, either. The poems were composed by various authors. The spelling system is identical with MYS 3. MYS 4 was used as a supplemental source for this grammar. 5  Besides the usual love songs, poems of love and affection (sōmon) also include poems on friendship and relationship between sovereign and retainer. 6  Elegies (banka) are poems on death, parting, sorrow, and absence from the capital.

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

10

Chapter 1

MYS 5 (#793–#906) includes 114 poems that are considered to be traditionally in the zōka genre, although some of them can be definitely classified as banka, since they deal with death and sorrow. All the poems in this volume were composed between 724 and 734 CE, which represents a much greater chronological homogeneity in comparison with MYS 1–4. Most of the poems in this volume belong to Yamanoue-no Okura (山上憶良, 660–733 CE), one of the greatest Man’yōshū poets, who was possibly a Korean from Kudara (Paekche), or at least a descendant of Kudara immigrants to Japan. Yamanoue-no Okura is also a well known outstanding scholar of the Chinese classics and Buddhism. The spelling system in this volume is predominantly phonographic, with rather few exceptions. In addition, the spelling system appears to reflect early Western Old Japanese, as demonstrated by Bentley 1997. The same can be said about its overall grammatical features. Together with the Kojiki kayō, Jōgū Shōtoku hōō teisetsu, Nihonshoki kayō, and Fudoki kayō, it constitutes one of the sources on early Western Old Japanese.7 MYS 5 was used as a major source for this grammar and was fully digitized. MYS 6 (#907–#1067) includes 161 poems in the sōmon genre. All of these poems date from 723–744 CE, representing a just slightly less tight chronological frame than MYS 5. The poems were composed by various authors. The spelling system is mostly logographic, being identical to the one employed in MYS 2 and MYS 4. MYS 6 was used as a supplemental source for this grammar. MYS 7 (#1068–#1417) includes 350 poems written in the sōmon genre. Most of the poems in this volume have no dates, but they are usually believed to be from the first part of the eighth century. The poems were composed by various authors. The spelling system is mostly logographic. Mizushima Yoshiharu believes that poem #1265 in this volume is a Sakimori poem (Mizushima 2003: 863). However, the poem in question is written almost completely in logographic script; therefore it is impossible to define whether it represents an Eastern Old Japanese text. MYS 7 was used as a supplemental source for this grammar. MYS 8 (#1418–#1663) includes 246 poems in the zōka and sōmon genres. Similar to MYS 7, most of the poems in this volume have no dates, but they are believed to be from the late seventh–early eighth centuries. The poems were composed by various authors. The spelling system is mostly logographic. MYS 8 was used as a supplemental source for this grammar.

7  Although it has become fashionable recently to speak of early and late Western Old Japanese, most likely the differences we see here are between two dialects: Asuka and Nara. Chronologically, of course, the Asuka dialect is attested earlier than the Nara dialect.

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

Sources and Previous Scholarship

11

MYS 9 (#1664–#1811) includes 148 poems in the zōka, sōmon, and banka genres. The poems in the zōka genre date up to 729 CE, and the rest are undated, but probably no poems are earlier than the late seventh century. The poems were composed by various authors. The spelling system is mostly logographic. MYS 9 was used as a supplemental source for this grammar. MYS 10 (#1822–#2350) includes 539 poems in the zōka and sōmon genres tailored to the four seasons. These poems are traditionally believed to be from the end of the seventh century, but it is also possible that some of them were composed in the eighth. The poems were composed by various authors. The spelling system is mostly logographic. MYS 10 was used as a supplemental source for this grammar. By the number of the poems included, this is the largest volume in the Man’yōshū. MYS 11 (#2351–#2840) includes 490 poems: 477 anonymous poems in the sōmon genre without dates and thirteen poems in the hiyuka genre from the late seventh–early eighth centuries. The poems were composed by various authors. The spelling system is mostly logographic. MYS 11 was used as a supplemental source for this grammar. MYS 12 (#2841–#3220) includes 380 poems in the sōmon genre. This general classification is violated though, as this volume includes also poems on travel and parting, which are not normally classified as sōmon. In addition, another genre of ‘question and answer poems’ (mondōka, 問答歌)8 is included in this volume, which does not fit into traditional Chinese classification of poetry. Dates are unknown, but most likely these poems are from the late seventh– early eighth centuries. The poems were composed by various authors. The spelling system is mostly logographic. MYS 12 was used as a supplemental source for this grammar. MYS 13 (#3221–#3347) includes 127 poems in the sōmon, zōka, mondōka, hiyuka, and banka genres. None of the poems is dated, but probably none is later than the end of the seventh century. The poems were composed by various authors. The spelling system is mostly logographic. MYS 13 was used as a supplemental source for this grammar. MYS 14 (#3348–#3577) includes 230 poems in zōka, sōmon, hiyuka, and banka genres. The majority of these poems are written in the Eastern Old Japanese (Azuma) dialect. The modern descendant of Eastern Old Japanese survives only in the Hachijō and Aogashima islands; everywhere else in the Kantō region it has been replaced by a variety of Central Japanese. Thus, Eastern Old Japanese poems from MYS 14 along with the sakimori (border guards) poems 8  Question and answer poems (mondōka) are paired poems, where person A asks a question with a poem, and person B replies to this question also with a poem.

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

12

Chapter 1

in MYS 20 are a unique source for knowing the non-Central Japanese dialect in the shape it was attested in the eighth century. The poems in this volume are all anonymous and are undated, but at the same time in slightly less than half of the poems there is an indication from which province in the East (吾妻/ 東国, Azuma) a poem comes from. Technically speaking, MYS 14 also includes five poems (#3567–#3571) composed by border guards. With a very few marginal exceptions the spelling system is entirely phonographic. MYS 14 is the major source for comparative data from Eastern Old Japanese used in this grammar, and it was fully digitized. It must be noted though, that a number of poems in MYS 14 look like they are written in flawless Western Old Japanese. MYS 15 (#3578–#3785) includes 208 poems in different genres. Among them, 145 poems belong to members of diplomatic mission to the Silla kingdom (新羅, Shiragi) in 736 CE. The remaining sixty-three poems represent the poetic exchange between Nakatomi-no Yakamori (中臣宅守) and his beloved Sano-no otogami-no wotome (狭野弟上娘子), which was probably composed before 741 CE,9 while he was in exile in Echizen (越前) province. Nineteen poems are composed by Sano-no otogami-no wotome and forty-four by Nakatomi-no Yakamori. The spelling is predominantly phonographic, but in a few occasions logographic spelling is used. MYS 15 is a major source for this grammar, and it was fully digitized. The language of MYS 15 is best defined as late Western Old Japanese. MYS 16 (#3786–#3889) includes 104 poems in a subvariety of the zōka genre: miscellaneous poems on occasion (有由縁雑歌, yoshi aru zōka). None of the poems is dated, but probably none is earlier than the eighth century. The poems were composed by various authors. The spelling system is mixed phonographic-logographic. MYS 16 was used as a supplemental source for this grammar. MYS 17 (#3890–#4031) includes 142 poems with unspecified genres. All poems in this volume were either composed or collected by Ōtomo-no Yakamochi (大伴家持) during 730–748 CE. It is believed that Ōtomo-no Yakamochi (717?–785 CE), the son of another Man’yōshū poet, Ōtomo-no Tabito (大伴 旅人, 665–731 CE) was the chief compiler of the whole Man’yōshū anthology, and he is also known as one of the best Man’yōshū poets himself. Similar to Yamanoue-no Okura and other Man’yōshū poets, he combined his poetic 9  The date of 741 CE is somewhat tentative. The name of Nakatomi-no Yakamori does not appear among those who were subject to the great amnesty in the sixth month of 741 CE, but it is suggested that he was pardoned before that, in the spring of 741 CE. More than twenty years later, Nakatomi-no Yakamori was making more or less a successful career at the court, as he was promoted from the sixth rank to the lower fifth rank in 763 CE.

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

Sources and Previous Scholarship

13

activities with political career, although in the latter he was not as successful as Ōtomo-no Tabito. Throughout his life he had to struggle with the declining fortunes of the Ōtomo clan, one of the greatest aristocratic families of the Asuka period, whose role in the Nara period greatly diminished. Ōtomo-no Yakamochi spent five years (746–751 CE) as the governor of Etchū (越中) province, and he returned to Nara in 751 CE with the title of junior counselor (少納言, shōnagon), corresponding to the junior fifth court rank—quite low for the scion of a famous and noble family. Only in 783 CE, two years before his death, was he promoted to middle counselor (中納言, chūnagon), a title corresponding to the third court rank. The spelling system in this volume is predominantly phonographic, although logographic spelling is also used. MYS 17 is a major source for this grammar, and it was fully digitized. The language of MYS 17 is best defined as late Western Old Japanese. MYS 18 (#4032–#4138) includes 107 poems with unspecified genres. All poems in this volume were either composed or collected by Ōtomo-no Yakamochi during 748–750 CE, while he was the governor of Etchū province. The spelling system is predominantly phonographic, although logographic spelling is also used. MYS 18 is a major source for this grammar, and it was fully digitized. The language of MYS 18 is best defined as late Western Old Japanese. MYS 19 (#4139–#4292) includes 154 poems with unspecified genres. Most are composed by Ōtomo-no Yakamochi (103 poems), and the rest of the poems, collected by Ōtomo-no Yakamochi belong to different authors. The poems were composed or collected between 750–753 CE, but the distinct peculiarity of this volume is that it includes a number of rather old poems, the oldest being from the end of the seventh century (#4260). Overall, however, the language of this volume is best defined as late Western Old Japanese. The spelling system of MYS 19 is somewhat unique: for the most part it is mostly logographic, but at the same time there are quite long sequences in many poems written phonographically. Due to this peculiarity, although MYS 19 was digitized, it is still used in this grammar as a supplemental and not a major source. MYS 20 (#4293–#4516) includes 224 poems with unspecified genres. Like the three previous volumes, the poems in MYS 20 are either collected or composed by Ōtomo-no Yakamochi in 753–759 CE. However, in contrast to the previous volume, MYS 19, the poems in MYS 20 are mostly collected. 101 poems, less than half of all the poems included in MYS 20 (#4321–#4359, #4363–#4394, and #4401–#4407, #4413–#4436) were composed by border guards (sakimori) in Eastern Old Japanese.10 Together with the poems from MYS 14 they 10  Since there are sometimes different points of view whether a given poem is a Sakimori poem, I followed (Mizushima 1996) in making a determination.

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

14

Chapter 1

constitute a major source on Eastern Old Japanese. However, the volume 20 of the Man’yōshū is linguistically split: poems #4293–#4320, #4360–#4362, #4395–#4400, #4408–#4412, #4433–#4435 and #4437–#4516 are written in late Western Old Japanese, while poems #4321–#4359, #4363–#4394, #4401–#4407, #4413–#4432, and #4436 are in Eastern Old Japanese. Therefore, MYS 20 is used as a major source for this grammar in two respects: poems written in Western Old Japanese constitute the basis for the main part of this grammar, concentrating on Western Old Japanese, and poems in Eastern Old Japanese are the major source for comparative data from Eastern Old Japanese. MYS 20 was fully digitized. Annotating in detail all modern editions of the Man’yōshū with commentaries will undoubtedly require a full monograph treatment, roughly comparable in size to the present book. Therefore, only major editions relevant to this study will be discussed below. The main editions I relied on are by Takagi Ichinosuke, Gomi Tomohide, and Ōno Susumu in the Nihon koten bungaku taikei (Takagi et al. 1957–62), Kojima Noriyuki, Kinoshita Masatoshi, and Satake Akihiro’s edition in the Nihon koten bungaku zenshū (Kojima et al. 1971–75); Nakanishi Susumu’s edition (Nakanishi 1984), more readily available today in paperback (Nakanishi 1978–83),11 and the twenty-two-volume edition by Omodaka Hisataka (Omodaka 1984), including a separate volume on each of the Man’yōshū’s maki, an index volume, and a volume with the original text (本文, honbun). I also occasionally consulted another multivolume edition of the Man’yōshū by Tsuchiya Fumiaki (Tsuchiya 1976–77), although in many respects it is inferior to the one done by Omodaka. Recently, a new edition by Satake Akihiro, Yamada Hideo, Ōtani Masao, and Yamazaki Yoshiyuki appeared in the Shin Nihon koten bungaku taikei series (Satake et al. 1999–2003). There are also several editions presenting only parts of the Man’yōshū written in Eastern Old Japanese: MYS 14 and border guards’ poems from MYS 20. Mizushima Yoshiharu’s edition of MYS 14 together with indexes is one of the best publications the Japanese philological tradition ever produced: it combines photocopies of all manuscript variants of the original text, Mizushima’s readings with commentaries, and all imaginable indexes (Mizushima 1984a). Almost equally superb is Mizushima 2003 edition of border guards’ poems (although no facsimiles are provided). Somewhat less detailed but still very useful is Mizushima Yoshiharu’s edition of both MYS 14 and the border guards’ poems from MYS 20 (much fewer indexes are provided) (Mizushima 1996 (1972)). 11  The paperback edition also has another important advantage: it includes the Man’yōshū jiten [The Man’yōshū encyclopedia] by the same author, which is not included in the hardback edition.

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

Sources and Previous Scholarship

15

Hoshino Yukihiko’s edition of the border guards’ poems conveniently combines many of the old commentaries on them (Hoshino 1976). To a certain extent, I also consulted Jan L. Pierson’s edition and translation of the Man’yōshū (Pierson 1929–1966). Although this monumental publication remains the only complete English translation, it is seriously dated in many respects. Any work on such an enormous text as the Man’yōshū is impossible without appropriate indexes. Besides the aforementioned 1984 index by Omodaka, I also extensively used Masamune Atsuo’s two-volume index (Masamune 1974) and the index to the poetic lines in the Man’yōshū (Hiyoshi 1992). The Man’yōshū jiten [Man’yōshū encyclopedia] by Sasaki Nobutsuna has also proven to be very useful (Sasaki 1983). Finally, when the first volume of this work was close to completion, I obtained a new CD-ROM edition of the Man’yōshū by Kinoshita Masatoshi (Kinoshita 2001), and Man’yōshū sakuin by Koten sakuin kakōkai (2003) based on it, which were also very helpful. 6) The Bussoku seki no uta or Bussoku seki ka [Poems about the stone with footprints of the Buddha] (仏足石歌) contains twenty-one poems, preserved as an inscription on a stone now in the possession of the Yakushiji temple in Nara. Poems #11 and #21 have multiple lacunae, and poem #12 is also damaged in two places. The exact date of the inscription is believed to be the same as on the adjacent stone depicting the footprints of Buddha, which was engraved on the request of prince Chinu (智奴王) in 753 CE (Tsuchihashi 1957: 7). Chinu was probably the author of these poems. The spelling system is completely phonographic. The poems were used as a major source for this study and they are completely digitized. The language of the poems is late Western Old Japanese. The edition of the Bussoku seki no uta I used is Tsuchihashi Yutaka’s edition in the Nihon koten bungaku taikei series (Tsuchihashi 1957). There is also Roy A. Miller’s critical edition (Miller 1975), but it is plagued with so many philological mistakes and inconsistencies aggravated by unjustified ad hominem attacks that most of the references made to it in the present study are of a critical nature. 7) The Shoku Nihongi kayō [Poems from the ‘Sequence to the Annals of Japan’] (続日本紀歌謡) contains eight poems from the Shoku Nihongi [Sequence to the Annals of Japan] (続日本紀). The Shoku Nihongi was compiled by Sugano-no Mamichi (菅野真道) and others in 797 CE. It comprises forty volumes (maki), and describes events from 697 to 791 CE. The poems themselves range from 738 CE to the reign of Emperor Kōnin (光仁, 770–781 CE), clearly representing late Western Old Japanese. The spelling is phonographic. The poems were used as a major source and are fully digitized.

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

16

Chapter 1

The editions of the Shoku Nihongi kayō I used are Tsuchihashi Yutaka’s edition in the Nihon koten bungaku taikei series (Tsuchihashi 1957) and Kōnosu Hayao’s edition in the Nihon koten bungaku zenshū (Ogihara and Kōnosu 1973). Also, there is a recent edition of the complete Shoku Nihongi text in five volumes in Shin Nihon koten bungaku taikei series (Aoki et al. 1989–98), by Aoki Kazuo, Inaoka Kōji, Sasayama Haruo, and Shirafuji Noriyuki, accompanied by a separate volume with an index and chronological charts by Sasayama Haruo and Yoshimura Takehiko (Sasayama and Yoshimura 2000), which was also consulted. 8) The Senmyō or Mikotonori [Imperial edicts] (宣命) contains sixty-two edicts recorded in the Shoku Nihongi and one more edict on the vow to the Asuka temple, which was discovered among the Gangōji temple documents. The earliest edict is from 697 CE and the latest one is dated by 789 CE. In spite of the fact that most of the edicts except the first eight are dated from 742 CE, they probably represent mostly early Western Old Japanese, since the language appears to be quite archaic in many respects. The archaic nature of the language is probably due to the ritual role of the imperial edicts. The spelling is predominantly logographic, with only some grammatical elements and proper names spelled phonographically. Nevertheless, the Senmyō text was treated as a major source for this study, because it is the only Old Japanese text written in prose that is quite homogeneous, and recorded quite early in contrast to another Old Japanese text, Norito [Liturgies] that is less homogeneous and was not recorded until the tenth century. The text of the Senmyō was fully digitized. I used two editions of the Senmyō: Kaneko Takeo’s edition with extensive commentaries (Kaneko 1989) and Kitagawa Kazuhide’s edition without commentaries but with several comprehensive indexes and with indication of textual variants (Kitagawa 1982), which perfectly complement each other. In addition, the full edition of the Shoku Nihongi text (Aoki et al. 1989–98) was also consulted. 9) The Norito [Liturgies] (祝詞) is a heterogeneous text, where the first fifteen liturgies found in volume eight of the Engishiki [Ceremonies of the Engi years] (延喜式) reflect pre-750 Western Old Japanese, and others from this volume were written after this date, possibly even in early Heian period (Bentley 2001: 25–26). One more Norito that appears to be ancient (Bentley 2001: 26) is found in volume sixteen of the Engishiki. The Norito uses the same type of mixed logographic-phonographic script as the Senmyō. The Norito was used as a supplemental source for this grammar, for two reasons. First, the combined text of the sixteen Norito that can be dated safely as Old Japanese is relatively short,

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

Sources and Previous Scholarship

17

and there is much longer text of the Senmyō that also represents prose. Second, there is a detailed grammar of these sixteen Norito in English (Bentley 2001). I mostly relied on Bentley’s critical text (Bentley 2001: 264–273), but in addition I also consulted Takeda Yūkichi’s critical edition (Takeda 1974). 10) The Nihon ryōiki or Nihon reiiki [Japanese tales of wonders] (日本霊 異記) is a collection of stories and legends with a didactic Buddhist flavor in three volumes (maki). It was compiled by the Yakushiji temple monk Keikai (景戒) in the beginning of the Heian period and includes stories up to the 824 CE, which is usually believed to be the year of its compilation. The Nihon ryōiki is a text written in kanbun, but it contains invaluable lexical material and also occasional phrases and sentences written phonographically in Old Japanese. The Nihon ryōiki Old Japanese fragments represent late Western Old Japanese, and the text was used as a supplemental source for this study. I used Endō Yoshimoto and Kasuga Kazuo’s critical edition of the Nihon ryōiki in the Nihon koten bungaku taikei series (Endō and Kasuga 1967). 11) The Kagura uta [Songs for the Sacred Shintō dances] (神楽歌)—102 songs, preserved in the man’yōgana writing, probably dating from the eighth to the ninth centuries.12 The language of these poems is somewhat mixed: it contains obvious innovations alongside apparent archaisms, so it is difficult to classify it as either early or late Western Old Japanese. Due to its heterogeneous nature, the Kagura uta were used as a supplemental source for the present study. I used Konishi Jin’ichi’s critical edition of the Kagura uta in the Nihon koten bungaku taikei series (Konishi 1957). 12) The Shōsōin documents, the documents dating from the seventh to the eighth centuries preserved in the Shōsōin (正倉院) imperial repository in Nara. Similar to the Nihon ryōiki, most documents are written in kanbun, and have only lexical data, occasional phrases or sentences written phonographically in the man’yōgana script. Therefore, the Shōsōin documents were used as a supplemental source for this study. To the best of my knowledge, there is no annotated edition of the Shōsōin documents. I used the multivolume photocopy edition prepared by the Imperial Household Agency (宮内庁) (Kunaichō 1990). Since this is not just a descriptive grammar, a brief outline of the texts written in languages other than Eastern Old Japanese, which are used as a source 12  Seven more are known in the later kana writing: seven are preserved in the Kokin wakashū poetic anthology (921 CE) and two more in the Utsuho monogatari (early tenth century), but these clearly represent Middle and not Old Japanese.

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

18

Chapter 1

for comparison with Western Old Japanese is in order. First, and foremost it certainly involves the sister language of Old Japanese, the Ryukyuan language. I extensively used data from both Old Ryukyuan and modern Ryukyuan dialects. The data for the latter are taken from various published sources, which are mentioned in the body of the grammar. For Old Ryukyuan, the data come from the following two sources: 13) The Omoro sōshi [Omoro Records] (おもろさうし)—1,554 poems in Old Ryukyuan, mostly reflecting the variety spoken in and around the capital city of Shuri (首里), but also possibly reflecting to a certain extent the dialects of the outer islands. The poems were recorded in the sixteenth–seventeenth centuries. Many of them have an apparent sacral nature. The poems are recorded in a pseudo-phonographic variety of the hiragana script, exhibiting very strong influence on spellings from the mainland Japanese, with very few logographic signs present. The mainland influence makes the deciphering of the proper Old Ryukyuan phonographic properties based on the Omoro sōshi a very cumbersome and unreliable business, although recently Serafim proposed some interesting insights (see preface for details). Nevertheless, the morphosyntactic structure of the language is reflected quite well even in this imperfect script, and since the morphosyntax of Old Japanese is the primary object of this study, the Omoro sōshi was used as a major comparative source. I used two modern critical editions of the Omoro sōshi: Hokama Shuzen and Saigō Nobutsuna’s edition in the Nihon shisō taikei series (Hokama and Saigō 1972), and Hokama Shuzen’s recent critical edition accompanied with translation into modern Japanese (Hokama 2000). 14) The Ryūka [Ryukyuan Songs] (琉歌)—5,100 poems from the eighteenth to the nineteenth centuries. The poems are recorded in the same cumbersome hiragana script, but with a higher percentage of logographic writing. Similar to the Omoro sōshi, it is a good source on Middle Ryukyuan morphosyntax, but not so much on its phonology. I used Shimizu Akira’s edition of the Ryūka (Shimizu 1994), which includes romanization and translation into modern Japanese. In addition, the critical edition by Shimabukuro Seibin and Onaga Toshio (Shimabukuro and Onaga 1968) was also used. Also, a number of texts written in Korean were consulted. 15) The Hyangka [Songs of the Motherland] (郷歌), twenty-six poems from the sixth to the tenth centuries. There are fourteen Hyangka poems from the sixth to the eighth centuries, recorded in the historical-mythological source

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

Sources and Previous Scholarship

19

Samkwuk yusa [Events Remaining from the Three Kingdoms] (三国遺事) compiled by monk Ilyen (一然) in 1283 CE, eleven recorded in the Kyunye cen [Biography of Kyunye] (均如伝, 1079 CE), attributed to the tenth-century Korean monk Kyunye, and one preserved in the Kolye sa [Koryo history] (高麗史, fourteenth century), representing a poem from the tenth century. I used Kim Wancin’s and Yu Changkyun’s editions of the Hyangka (Kim 1980), (Yu 1996). 16) The Yongpi echenka [Songs of the Dragons Flying in the Sky] (龍飛御 天歌), the first text written in hankul, the newly invented Korean alphabet, in 1447 CE. It is a long poem in 125 stanzas. Many archaic features of the language, not found in even slightly later texts, are preserved in this text. I used the edition with commentary by Kim Sangek (Kim 1983) and the facsimile edition from the Academy of Korean Studies (Hankwukhak yenkwuwen 1988.6). 17) The Sekpo sangcel [Buddha jataka stories] (釈譜詳節), compiled by prince Swuyang (首陽) (future king Seycwo (世祖)) in 1449 CE on occasion of queen Swohen’s (昭憲) death. I used the facsimile edition from the Academy of Korean Studies (Hankwukhak yenkwuwen 1988.3). 18) The Welin sekpo [Welin chen kang ci kwok and Sekpo sangcel combined] (月印釈譜), a combined edition of the Sekpo sangcel and Welin chen kang ci kwok [A Song for the Moon Reflected in One Thousand Rivers] (月印千江 之曲, 1449 CE) done by a group of Confucian scholars on the orders of King Seycwo in 1459 CE. I used Kang Kyusen’s facsimile edition with commentary of volumes 1–2 (Kang 1998), Kim Yengpay’s facsimile edition with commentary of volumes 7–10 (Kim 1993, 1994) and Cang Seykyeng’s facsimile edition with commentary of volumes 17–18 (Cang 1995). 19) Kumkang kyeng samka hay [An Explanation of the Diamond Sutra by Three Scholars] (金剛經三家解), an annotated edition and translation into Korean of the Diamond Sutra, published in 1482 CE. I used the facsimile edition from the Hankul Society (Hankul hakhoy 1960–61).

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

Section 2

Previous Scholarship The best studied aspect of Western Old Japanese, both inside and outside Japan, is its phonology. Numerous monographs, dissertations, and articles have been written on Old Japanese phonology, such as Arisaka (1932, 1955), Hashimoto (1938, 1950, 1966), Pierson (1929), Yoshitake (1934), Ōno (1953), Mabuchi (1957), Lange (1973), Unger (1977), Whitman (1985), Martin (1987), Mori (1991), Morishige (1993), Bentley (1997), Miyake (1999, 2003). I will not discuss in detail the previous work on Old Japanese phonology here, as the present work centers on the morphosyntax of Western Old Japanese. Interested readers can find a wonderful and most up-to-date discussion of all previous scholarship in the field of Western Old Japanese phonology in Marc Miyake’s dissertation (Miyake 1999: 90–129), or his book (Miyake 2003: 43–65). In contrast to the research on Western Old Japanese phonology, there are very few general descriptions of the Western Old Japanese language grammar published even in Japanese. Surprisingly enough, none of these descriptions is complete. The main and the most detailed of them is Yamada Yoshio’s grammar, first published in 1912, which since underwent several enlarged and corrected editions and many reprints (Yamada 1954). This grammar is accessible only to the people who have mastered Classical Japanese, as it is written in bungo. It is quite detailed, but nevertheless it is not complete: several areas of Old Japanese morphosyntax are not treated at all, for example, there is no discussion of numerals. In addition, Yamada’s grammar does not really differentiate between Western and Eastern Old Japanese, treating them as if they were data from the same language, although he dedicates a special chapter in his grammar on the special ‘grammar usage’ of Eastern Old Japanese. Besides Yamada’s grammar, which in spite of its shortcomings remains the most authoritative treatment of Western Old Japanese grammar to date, there are also several much shorter works. Saeki Umetomo’s first grammar of Old Japanese Jōko no kokugo (Saeki 1933), limited to just 117 pages, is best defined as a sketch of Old Japanese. A later version of the same book appeared under a different title Nara jidai no kokugo (Saeki 1950), which has since undergone several new enlarged and improved editions. It can be used as a very good introduction to Old Japanese, but nevertheless it still remains too sketchy in many details. The recent Old Japanese grammar by Shirafuji Noriyuki (Shirafuji 1987), also suffers from incompleteness: there is no treatment of nouns, pronouns, and numerals in his grammar. However, Shirafuji’s grammar is a welcome

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

Sources and Previous Scholarship

21

development in many aspects, such as, for example, the verbal system, and it is more detailed than Saeki’s grammar in the areas it describes. There is only one general grammatical sketch of Old Japanese written in a Western language (Syromiatnikov 1972). There are also several studies of various aspects of Old Japanese grammar that do not represent systematic grammars, such as Mabuchi Kazuo’s two books on Old Japanese, treating separate issues of phonology, lexicon, and morphosyntax (Mabuchi 1972, 1999), and two monographs by Saeki Umetomo dealing with some phonological, morphological, or even textual aspects of Old Japanese texts (Saeki 1963, 1965). In addition, there are several monographlength treatments of the language of particular Old Japanese texts (Kotani 1986; Iwai 1981; Matsuo 1978; Bentley 2001); or of particular parts of grammar (Yoshida 1973; Wada 1994). The grammatical description of Eastern Old Japanese actually fared much better than that of Western Old Japanese: there are three comprehensive monographs on the subject. Hōjō Tadao’s monograph (Hōjō 1966) is predominantly on script and phonology, although it does include a grammatical description as well. A slightly shorter description by Fukuda Yoshisuke (Fukuda 1965) has a more balanced presentation of phonology and grammar, although Hōjō’s study is much more detailed regarding phonology. A more recent study of Eastern Old Japanese with a balanced presentation of phonology and grammar is that of Mizushima Yoshiharu (Mizushima 1984b). This wonderful study is also supported with accurate statistical analysis. In addition, there are also a number of studies on particular grammar topics that do not differentiate between the Asuka–Nara period language and Heian period language, e.g. (Nakanishi 1996), lumping them together as kodaigo (古代語) ‘Old [Japanese] language,’ in opposition to the terminology that differentiates these two: jōdaigo (上代語) ‘Old Japanese’ and chūkogo (中古語) ‘late Old Japanese.’ While there are certain commonalities between these two periods, in my opinion they do represent two different historical stages of the Japanese language, corresponding to Old and Middle Japanese. Therefore, these kinds of studies were consulted minimally, since they are anachronistic, confusing two different periods of language history. Needless to say, there are hundreds of articles published on various aspects of Western Old Japanese grammar. It is neither possible nor necessary to mention them all in this introductory chapter; the reader will find references to some of them throughout the text and in the bibliography.

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

chapter 2 Script and Phonology



Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

Contents of Chapter 2 Chapter 2 Script and Phonology 26 1 Script 28 1.1 Syllabic Script 28 1.2 Rebus Writing 39 2 Phonetics and Phonology 41 2.1 Consonants 41 2.1.1 Labials 42 2.1.2 Dentals 43 2.1.3 Palatals 44 2.1.4 Velars 45 2.2 Vowels 45 2.2.1 Back Vowels 46 2.2.2 Central Vowels 46 2.2.3 Front Vowels 47 2.3 Pitch Accent 48 2.4 Phonotactics 48 2.5 Morphophonological Processes 54 2.5.1 Contraction 55 2.5.2 Monophthongization 56 2.5.3 Nasalization 57 2.5.4 *-r- Loss 58

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

This chapter presents a brief outline of Old Japanese script and phonology. I tried to concentrate here on a concise presentation of a state-of-art understanding of both Old Japanese script and phonology to the extent it is necessary for those whose primary goal is learning the Western Old Japanese language and desiring a better understanding of Western Old Japanese texts. Meanwhile, I avoided a detailed treatment of some controversial questions that are predominantly of interest to the specialists in Japanese historical phonology, since dealing with these questions merits its own monograph. Those readers who are interested in a deeper treatment of Old Japanese graphemics and phonetics are referred to an excellent recent monograph by Marc Miyake (2003). I followed Miyake’s solutions for many of the controversial questions the reader is going to see below. For the second edition of this grammar I used a new system for rendering Old Japanese data throughout this book with two exceptions. This system is even more phonetic than any of the previous ones I used, because the targeted auditorium for this book should not be limited to Japanologists, rather, should also include general linguists. It would not be fair to ask general linguists to study the history of the Japanese language in order to be able to understand the system of transcription. Thus, for example the historians of the Japanese language are well aware of the fact that OJ consonants romanized as b, d, g, and z are in fact prenasalized as mb, ⁿd, ŋg and ⁿz. But this is largely unknown by general linguists as well as by Japanologists engaged in the study of other disciplines. Ditto for the vowel /ǝ/ that is normally transcribed by /o/, sometimes with diacritics. I provide below a chart demonstrating the major differences between various systems of transcription. Note that in the current system of romanization the neutralizations of vowels i X ï = i, e X ɛ = e, and o X ǝ = o, are written as kō-vowels i, e, and o for two reasons: a) the first two i X ï = i and e X ɛ = e are found after dentals, w, r, and y, and the third one after labials and w, therefore they are all predictable; b) although ɛ and ǝ (but not ï) are much more frequent in Western Old Japanese than e and o, the result of neutralization found in the Heian period (794–1192 AD) went along with less frequent e and o for the reasons unknown to us, even though we can speculate that the vowels that are found after neutralization are less exotic and more plain than the more frequent ones. The complete Western Old Japanese syllabary will be provided below in Chart 2 in section 1.1.

© Alexander Vovin, 2020 | doi:10.1163/9789004422810_003

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

27

 Script and Phonology chart 1

Major differences between various systems of transcription

Vovin 3a

Vovin 2

Vovin 1

Traditional

Yale

Miller– Mathias

Frellesvig– Whitman

o ǝ i ï e ɛ mb ⁿd ŋg ⁿz

ô ö î ï ê ë mb ⁿd ŋg ⁿz

wo ö yi iy ye ey Np Nt Nk Ns

o ö i ï e ë b d g z

wo o̲ yi iy ye ey b d g z

ô ö î ï ê ë b d g z

wo o i iy ye e b d g z

a Vovin 3 is used in the latest volumes of the Man’yōshū translation and in this edition of the Old Japanese grammar. Vovin 1 is used in the first edition of the Old Japanese grammar and in the earliest volumes of the Man’yōshū translation. Vovin 2 is used in all other volumes of the Man’yōshū translation.

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

Section 1

Script There are several writing systems used in Western Old Japanese texts. Most of our major sources are written in either completely phonographic syllabic script usually called man’yōgana (萬葉仮名), or in a mixed logographicphonographic script, where Chinese characters can be used either logographically or phonographically. There are two varieties of the mixed script: one where both logograms and phonograms are written in the same size and the one where phonograms are written in the smaller size than the logograms. The second variety is called Senmyōgaki ‘Senmyō writing’ (宣命書), and as the name suggests, it is used in the Senmyō and also in the Norito. Other texts with mixed script use the first type with the logograms and phonograms of the same size. A number of poems in the Man’yōshū are written exclusively with logograms so that the text represents a Chinese rather than a Japanese text, where scribes attempted to sinicize the script as much as possible, rendering even Japanese grammatical elements with their approximate Chinese equivalents. For example, OJ mi-ⁿz-u ‘see-NEG-FIN’ can be written as MC 不見 [put kEnʰ] ‘NEG see,’ with the negative particle preceding the verb unlike OJ, where negative suffix follows the verbal stem. With the exception of Senmyōgaki, which seems to be restricted to the Senmyō, Norito, and one poem in the Shoku Nihongi kayō (SNK 8), there are, of course, all imaginable combinations of phonographic, mixed logographic-phonographic and ‘pure’ logographic scripts, where both ‘sinicization’ and density of logograms varies to a considerable degree. In some rare cases it may be difficult to draw exact boundary between phonographic and logographic usage. For example, the word umɛ ‘plum blossom’ is frequently written as uMƐ (宇梅) (MYS 843, 3904, 3906, etc.), where the second character is simultaneously a phonogram for syllable /mɛ/ and a logogram ‘plum tree.’ Finally, there are different types of rebus writing that are occasionally used in the Man’yōshū. 1.1 Syllabic Script The man’yōgana syllabic script is represented by two types: ongana and kungana. Ongana employs Japanicized EMC or LMC (and sometimes even LHC) readings of the characters for phonographic values of the syllables, e.g., EMC 加 [kaɨ] ‘add’ is used as a syllabic sign with the reading /ka/. There are two varieties of ongana: variety A, based on EMC readings of the Chinese characters, and variety B, based on LMC readings. Variety A is the dominant one:

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

SCRIPT AND PHONOLOGY

29

it is used in all OJ texts except the Nihonshoki, and the variety B is limited to the Nihonshoki.1 There are good reasons to believe that variety A was originally borrowed from similar systems of writing adopted on the Korean peninsula (Kim 1983; Bentley 2001b). Thus, the character readings of variety A are not directly based on EMC readings, but are conditioned by a Korean intermediary. Meanwhile, variety B appears to be modeled directly on LMC readings, although, of course, the same syllabic principle of writing underlies both varieties. There is a significant overlap between varieties A and B, since they share a number of signs. However, other signs remain specific, reflecting the differences between EMC and LMC. Different phonetic mergers within each variety represent the major difference between varieties A and B. Thus, variety A often uses the same signs for syllables containing initial voiceless obstruents /C/ and prenasalized voiced /ᶰC/,2 keeping the nasal sonorants /m/ and /n/ transcribed in a different way. On the other hand, variety B frequently employs the same signs for syllables with prenasalized voiced obstruents and nasal sonorants, keeping voiceless obstruents distinct. Kungana employs the Japanese readings of the characters for their phonographic values irrespective of their meanings. For example, character 田 ‘rice field’ (Japanese reading /ta/) and character 名 ‘name’ (Japanese reading /na/) are used to write syllables [ta] and [na] respectively. Overall, in predominantly phonographic texts kungana is used much more rarely than ongana both quantitatively and qualitatively. It is more typical for the Man’yōshū than for any other OJ texts, but even in the Man’yōshū it is comparatively infrequent with the exception of books written in the mixed logographic-phonographic script. Very few of these kungana signs are included into Omodaka et al. (1967), Osterkamp (2011), and Bentley (2016). Admittedly, some of these graphs are extremely rare, appearing only once or twice in the whole anthology, but so are some of the monosyllabic ongana included in Omodaka et al. (1967) and Bentley (2016). In addition, not every possible syllable has a corresponding kungana, while ongana exists for all syllables with a minor exception of syllable [ⁿzo] in variety B. The rare usage of kungana is especially noticeable for the most syllables that contain prenasalized obstruent initials [mb], [ⁿd], [ŋg], [ⁿz], as well as flap [r]. As with all early syllabic scripts, the polyphonic principle is largely at work in man’yōgana. The polyphonic principle manifests itself in two facts: a) a 1  There is a theory that the Nihonshoki’s ongana itself includes two varieties, alpha and beta (Mori 1991). However, I agree with Miyake that there are very significant arguments against this theory (Miyake 2003: 57–60). 2  This manifests itself primarily in the inconsistent spelling of the syllables with prenasalized consonants, which may be spelt as if they were voiceless. The reverse (that is, spelling of syllables with voiceless consonants as if they were prenasalized) is less frequent.

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

30

Chapter 2

syllable can normally be written with more than one written signs; b) a written sign can have different phonographic values. The polyphonic principle is well illustrated by the chart of the man’yōgana signs below. Thus, for example, the syllable /ka/ can be transcribed by the ongana signs 加 迦 可 賀 訶 珂 箇 架 嘉 甲 甘 敢 in variety A and by ongana signs 加 迦 可 賀 訶 河 箇 伽 舸 歌 軻 柯 介 甲 甘 in variety B. Not every given syllable has so many possibilities, but cases when only one or two signs are used, like in the case of the syllable /ⁿze/ written with only one character 是 in variety A and with only two characters 筮 噬 in variety B, are quite rare. A number of the man’yōgana signs are disyllabic, that is, they transcribe not one syllable but two. Since there were no syllable final consonants in Old Japanese, some of the Chinese characters that had readings with final consonants were provided with an extra vowel after the consonant. In some of cases, it is an echo vowel, like, for example, MC 壱 /it/ transcribing OJ sequence /iti/, but in the other cases the second vowel may be unpredictable, for example, MC 因 /in/ transcribing OJ sequence /ina/. Care must be taken, since a number of disyllabic signs can be also used as monosyllabic, thus, for example, MC 南 / nam/ can render both OJ monosyllable /na/ and a disyllabic sequence /nami/. Overall, however, disyllabic graphemes are used infrequently. They mostly appear in transcriptions of proper nouns, and sometimes in the texts written predominantly with logograms. Their usage in mostly phonographically written texts is quite close to non-existent. It is also necessary to note that some (but not all) characters are used both as ongana and kungana. For example, character 田 ‘rice field’ may function as ongana for syllable [ⁿde] and as kungana for syllable [ta], or character 世 is used as ongana for syllable [se] and kungana for syllable [yǝ]. Below, I present a chart that includes syllabic signs from both varieties. chart 2

Man’yōgana signs in varieties A and B

Transcription

Variety A

Variety B

a

ongana: 阿 安 kungana: 足 吾 鳴呼 disyllabic ongana: 英 [aŋga] disyllabic kungana: 赤 [aka] 金 [aki] 朝 [asa] 相 [api] [apu] [ara] 荒 [ara] 在 [ari, aru] trisyllabic kungana: 茜, 茜草 [akane]

ongana: 阿 安 婀 鞅 kungana: 足 吾

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

31

SCRIPT AND PHONOLOGY chart 2

Man’yōgana signs in varieties A and B (cont.)

Transcription

Variety A

Variety B

i

ongana: 伊 夷 以 怡 異 移 因 印 壱 已 kungana: 射 五十 馬声 disyllabic ongana: 因 [ina] 印 [ina] 壱 [iti] 盤 [ipa] disyllabic kungana: 勇 [isa] 乞 [iⁿde] 五百 [ipo] ongana: 于 汙 宇 有 羽 烏 雲 kungana: 氐 菟 卯 得 disyllabic ongana: 雲 [una] 鬱 [utu] disyllabic kungana: 打 [uti] 敲 [uti] 浦 [ura] ongana: 衣 愛 依 kungana: 得 榎 荏 ongana: 意 於 応 乙 憶 飫 disyllabic ongana: 乙 [otu] 邑 [opi, opu, opo] disyllabic kungana: 忍 [osi] ongana: 加 迦 可 賀 珂 箇 架 嘉 甲 甘 敢 kungana: 鹿 香 蚊 芳 歟 所 disyllabic ongana: 甘 [kamu] 敢 [kamu] 漢 [kani] 干 [kani] 葛 [katu] 甲 [kapi] 香 [kaŋgu] 高 [kaŋgu] 各 [kaku] 閑 [kana] disyllabic kungana: 借 [kasi] 方 [kata] 鴨 [kamo] 辛 [kara] 柄 [kara] 韓 [kara] trisyllabic kungana: 限 [kaŋgiru] ongana: 支 伎 吉 岐 棄 枳 企 芡 kungana: 寸 杵 來

ongana: 伊 以 怡 異

u

e o

ka

ki



ongana: 紀 幾 貴 奇 騎a 綺 寄 記 kungana: 城 木 樹

易因壱

kungana: 胆 disyllabic ongana: 因[ina] 壱 [iti] ongana: 于 汙 宇 羽 紆 禹 kungana: 鵜 菟 disyllabic ongana: 鬱 [utu] ongana: 愛 哀 埃 ongana: 意 於 淤 乙 憶飫

disyllabic ongana: 乙 [otu] 磤 [onǝ] ongana: 加 迦 可 哿 賀 訶河箇伽舸歌軻柯 介甲甘 kungana: 髪 鹿 香 蚊 disyllabic ongana: 甘 [kamu] 甲 [kapi] 香 [kaŋgu] 覚 [kaku]

ongana: 吉 岐 棄 枳 企 耆祇祁

kungana: 寸 杵 ongana: 紀 幾 奇 基 機 己既気

kungana: 城 木 樹 黄

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

32 chart 2

Chapter 2 Man’yōgana signs in varieties A and B (cont.)

Transcription

Variety A

Variety B

ku

ongana: 久 玖 口 群 苦 丘 九 鳩 君 kungana: 來 國 disyllabic ongana: 君 [kuni] disyllabic kungana: 草 [kusa] 國 [kuni]

ongana: 久 玖 句 苦 倶 区

ke



ongana: 祁 家 計 鶏 介 奚 谿 価 係 結 kungana: 異 disyllabic ongana: 兼 [kemu] 監 [kemu] 険 [kemu] ongana: 氣 既 kungana: 毛 食 飼 消

ko

ongana: 古 故 庫 祜 姑 孤 枯 kungana: 子 兒 籠小 粉



ongana: 己 許 巨 居 去 虚 忌 興 kungana: 木 disyllabic ongana: 金 [kǝmu] 今 [kǝmu] 近 [kǝnǝ] 乞 [kǝti] 興 [kǝŋgǝ] disyllabic kungana: 乞 [kǝsǝ] 言 [kǝtǝ] 來 [kǝrǝ] ongana: 何 我 賀 河 蛾 disyllabic kungana: 金 [ŋgane] 柄 [ŋgara] ongana: 藝 芸 祇 岐 伎 ongana: 疑 宜 義 ongana: 具 遇 求 隅 群 disyllabic ongana: 群 [ŋguni/ŋguri] disyllabic kungana: 晩 [ŋgura] ongana: 牙 雅 夏 ongana: 義 宜 㝵 ongana: 胡 呉 候 後 虞 吾 kungana: 籠

ŋga ŋgi ŋgï

ŋgu

ŋge  ŋgɛ  ŋgo

勾矩絇衢寠訓履 kungana: 來

disyllabic ongana: 訓 [kuni] 菊 [kuku] ongana: 祁 家 計 鶏 稽 啓

ongana: 氣 居 該 戒 階 開 慨凱穊愷 kungana: 毛 食 笥 ongana: 古 故 庫 姑 孤 固顧 kungana: 子 児 籠 小 ongana: 許 巨 居 去 虚 挙莒拠渠興 kungana: 木

disyllabic ongana: 興 [kǝŋgǝ]

ongana: 我 賀 餓 峨 俄 鵝 ongana: 藝 芸 伎 儀 蟻 𡺸 ongana: 疑 擬 ongana: 具 遇 愚 虞 群 disyllabic ongana: 群 [ŋguni/ŋguri] ongana: 霓 ongana: 㝵 礙 皚 ongana: 胡 呉 吾 誤 悟 娯

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

33

SCRIPT AND PHONOLOGY chart 2

Man’yōgana signs in varieties A and B (cont.)

Transcription

Variety A

Variety B

ŋgǝ

ongana: 其 期 碁 凝 ongana: 佐 沙 作 左 者 柴 紗 草 匝 讃

ongana: 語 御 馭 ongana: 佐 沙 作 左 瑳 磋

散尺積

舎差匝讃尺 kungana: 狭

sa

si

kungana: 狭 猨 羅 disyllabic ongana: 三 [samu] 雜 [sapa] [sapi] 障 [apa] [api] [apu] 匝 [sapi] 颯 [sapu] 讃 [sanu] 散 [sani] 薩 [sati] [satu] 相 [saŋga] [saŋgu] 尺 [saka] 作 [saka] [saku] 積 [saka] disyllabic kungana: 坂 [saka] 前 [saki] 辟 [saki] 樂 [sasa]b 刺 [sasu] 澤 [sapa] 禁 [sapɛ] 去 [sari] 核 [sane] ongana: 斯 志 之 師 紫 新 四 子 思 司 芝 詩旨寺時指此至次死偲事詞信 kungana: 爲 磯 disyllabic ongana: 信 [sina] 鍾 [siŋgu] 色 [siki/sikǝ] 餝 [sika] 式 [siki] 拭 [siki] 叔

su

se

so

[siku] disyllabic kungana: 及 [siki/siku] 科 [sina] 小竹 [sino] 塩 [sipo] 嶋 [sima] ongana: 湏 須 周 酒 洲 珠 数 kungana: 酢 簀 栖 渚 爲 disyllabic ongana: 駿 [suru] 宿 [suka] [suku] disyllabic kungana: 次 [suki] 隅 [sumi] 墨 [sumi] ongana: 勢 世 西 斉 kungana: 瀬 湍 背 脊 迫 disyllabic ongana: 瞻 [semi] ongana: 蘇 宗 祖 素 kungana: 十 麻 disyllabic kungana: 虚 [sora]

disyllabic ongana: 匝 [sapi] 戔 [sanǝ] 讃 [sanu] 薩 [sati, satu] 相 [saŋga, saŋgu] 尺 [saka]

ongana: 斯 志 之 師 紫 新 四子思司資茲芝詩 旨寺時指絁矢始尸 試伺璽辞嗣施洎信 kungana: 為 磯 disyllabic ongana: 信 [sina] 色 [siki, sikǝ]

ongana: 須 周 酒 洲 主 素秀輸殊蒭 kungana: 酢 簀 樔 disyllabic ongana: 駿 [suru] 宿 [suka, suku] ongana: 勢 世 西 斉 栖 細 制是剤

kungana: 瀬 湍 背 ongana: 蘇 素 泝 kungana: 十 麻

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

34 chart 2

Chapter 2 Man’yōgana signs in varieties A and B (cont.)

Transcription

Variety A

Variety B



ongana: 曾 所 僧 増 則 kungana: 衣 背 其 苑

ongana: 曾 所 増 則 贈

ⁿza

ongana: 射 蔵 邪 社 謝 座 disyllabic kungana: 坂 [ⁿzaka]

ⁿzi

ongana: 自 士 慈 尽 時 寺 仕

諸層賊

kungana: 衣 襲 ongana: 社 蔵 装 奘 disyllabic ongana: 蹔 [ⁿzami] ongana: 自 士 慈 尽 貳 児尓珥餌耳茸 kungana: 下

ⁿzu ⁿze ⁿzo  ⁿzǝ  ta

ti

tu

ongana: 受 授 聚 殊 洲c ongana: 是 ongana: 俗 ongana: 叙 序 賊d ongana: 多 太 他 丹 kungana: 田 手 disyllabic ongana: 丹 [tani] 塔 [tapu] 但 [tani] [taⁿdi] 當 [taŋgi] disyllabic kungana: 妙 [tapɛ] 玉 [tama] 垂 [taru] ongana: 知 智 恥 陳 珍 遲 kungana: 道 千 乳 路 血 茅 disyllabic ongana: 珍 [tinu]

ongana: 都 豆 通 追 川 kungana: 津 齋 disyllabic ongana: 筑 [tuki, tuku] 對 [tusi] disyllabic kungana: 爪 [tuma] 妻 [tuma, ⁿduma] 嬬 [tuma]e 積 [tumi] 列 [tura] 鶴 [turu]

disyllabic ongana: 甚 [ⁿzimu] ongana: 受 孺 儒 ongana: 筮 噬 ongana: — ongana: 叙 序 鐏 茹 鋤 ongana: 多 大 陁 柂 哆 駄党丹

kungana: 田 手 disyllabic ongana: 丹 [tani] 但 [tani, taⁿdi] 當 [taŋgi] ongana: 知 智 致 挃 笞 池馳珍直 kungana: 道 千 乳 路 血茅

disyllabic ongana: 直 [tiki] ongana: 都 豆 頭 菟 途 屠 突徒覩図 kungana: 津

disyllabic ongana: 筑 [tuki, tuku] 竹 [tuku]

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

35

SCRIPT AND PHONOLOGY chart 2

Man’yōgana signs in varieties A and B (cont.)

Transcription

Variety A

Variety B

te

ongana: 弖 氐 提 天 帝 底 堤 代 kungana: 手 価 直煑 disyllabic ongana: 点 [temu] ongana: 刀 斗 都 土 度 kungana: 戸 門 利 礪 速

ongana: 弖 氐 提 帝 底 堤

to



ⁿda

ⁿdi ⁿdu ⁿde ⁿdo ⁿdǝ na

ni

nu ne

ongana: 止 等 登 騰 得 kungana: 鳥 十 跡 迹 常 disyllabic ongana: 徳 [tǝkǝ] 得 [tǝkǝ] disyllabic kungana: 友 [tǝmǝ] ongana: 陁 太 大 disyllabic ongana: 弾 [ⁿdani] disyllabic kungana: 谷 [ⁿdani] ongana: 遲 治 地 ongana: 豆 頭 都 disyllabic ongana: 附 [ⁿduki] 曇 [ⁿdumi] ongana: 提 弖 代 田 低 泥 埿 ongana: 度 渡 土 ongana: 杼 騰 藤 特 ongana: 那 奈 寧 難 南 kungana: 名 魚 中 菜 七 莫 disyllabic ongana: 南 [nami, namu] 難 [nani] disyllabic kungana: 成 [nasu] 梨 [nasi] 夏 [natu] 波 [nami] 雙 [nami/namï] 楢 [nara]f trisyllabic kungana: 長柄 [naŋgara] ongana: 爾 迩 仁 日 二 而 尼 耳 人 柔 kungana: 丹 荷 似 煮 煑 disyllabic kungana: 柔 [niki] 熟 [niki]g ongana: 奴 怒 努 濃 農 kungana: 沼 宿 寝 渟 ongana: 尼 禰 泥 埿 年 kungana: 根 宿 disyllabic ongana: 念 [nemu]

諦題代

kungana: 手 ongana: 刀 斗 都 土 度 覩 妬杜図屠塗徒渡 kungana: 戸 聡 門 礪 ongana: 等 㔁 騰 苔 台 藤 kungana: 鳥 跡 迹 disyllabic: 徳 [toku] 得 [tǝkǝ] ongana: 陁 太 大 騨 娜 襄儴

ongana: 遅 治 膩 尼 泥 ongana: 豆 頭 逗 図 弩 砮 disyllabic: 曇 [ⁿdumi] ongana: 提 泥 埿 耐 弟 涅 ongana: 度 渡 奴 怒 ongana: 杼 騰 耐 廼 ongana: 那 奈 乃 儺 娜 難 kungana: 名 魚 中 disyllabic ongana: 冉 [nami] 難 [nani] 諾 [naki]

ongana: 爾 儞 迩 仁 珥 貳 kungana: 丹 瓊 ongana: 奴 怒 努 濃 農 kungana: 渟 ongana: 尼 禰 泥 埿 涅 kungana: 根

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

36 chart 2

Chapter 2 Man’yōgana signs in varieties A and B (cont.)

Transcription

Variety A

Variety B

no

ongana: 努 怒 弩 奴 kungana: 野 ongana: 乃 能 kungana: 荷 笶 箆 之h ongana: 波 播 幡 芳 婆 破 方 防 八 房 半 皤 薄 伴 泊 叵 盤i kungana: 羽 葉 歯 者 翔 disyllabic ongana: 盤 [pani] disyllabic kungana: 旗 [paⁿda] 花 [pana] ongana: 比 卑 必 臂 嬪 賔 kungana: 日 檜 氷 disyllabic kungana: 引 [pikï]j 櫃 [pitu] ongana: 非 斐 悲 飛 kungana: 火 干 乾

ongana: 努 怒 奴 弩 kungana: 野 ongana: 能 廼 kungana: 荷 ongana: 波 播 幡 芳 婆 破

nǝ pa

pi



pu

pe



po

mba mbi mbï  mbu

皤簸巴絆泮 kungana: 羽 葉 歯

disyllabic ongana: 博 [paka] ongana: 比 卑 必 臂 毘 毗譬避

kungana: 日 檜 氷 ongana: 斐 肥 悲 飛 被 彼 秘妃費

kungana: 火 熯 簸 ongana: 布 不 敷 甫 賦 府

ongana: 布 不 敷 府 賦 否 負 福 kungana: 經 歴 disyllabic ongana: 粉 [puni] 福 [puku] disyllabic kungana: 盖 [puta] ongana: 平 弊 霸 幣 敝 陛 遍 返 反 弁 kungana: 部 辺 重 隔 disyllabic ongana: 伯 [peki] ongana: 閇 倍 拝 kungana: 戸 躪 綜 經

符輔赴浮 kungana: 經 歴 乾

ongana: 平 弊 霸 幣 陛 蔽鞞𧯿

kungana: 部 辺 重 ongana: 閇 倍 沛 陪 背 杯俳珮

kungana: 戸 綜 ongana: 富 保 朋 倍 褒 裒

ongana: 保 富 宝 朋 倍 抱 方 凡 品 kungana: 帆 穂 太k disyllabic ongana: 凡 [pomu] 品 [pomu] ongana: 婆 伐 ongana: 毘 毗 鼻 妣 婢 豐l ongana: 備 肥 ongana: 夫 父 部 扶 disyllabic kungana: 吹 [mbuki] 振 [mbuki]

陪報袍譜品 kungana: 帆 穂 火 disyllabic: 品 [pomu] 法 [popu] ongana: 婆 麼 魔 磨 縻 ongana: 鼻 弥 弭 寐 ongana: 備 媚 眉 縻 ongana: 夫 父 部 矛 歩 騖

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

37

SCRIPT AND PHONOLOGY chart 2

Man’yōgana signs in varieties A and B (cont.)

Transcription

Variety A

Variety B

mbe mbɛ  mbo ma

ongana: 辨 便 別 ongana: 倍 ongana: 煩 ongana: 麻 磨 万 萬 馬 末 満 摩 kungana: 眞 間 目 信 鬼 disyllabic ongana: 望 [maŋga] [maŋgu] 莫 [] 幕 [] disyllabic kungana: 儲 [makɛ] 設 [makɛ] 松 [matu]m 圓 [mato] trisyllabic kungana: 相 [masani]n ongana: 弥 美 民 kungana: 三 御 見 水 参 視 disyllabic ongana: 敏 [minu] ongana: 微 未 味 尾 kungana: 身 実 箕 ongana: 牟 武 无 模 務 無 謀 鵡 儛 kungana: 六 disyllabic ongana: 目 [muku] disyllabic kungana: 村 [mura] ongana: 賣 咩 馬 面 kungana: 女 婦 disyllabic kungana: 食 [mesi] [mese] ongana: 米 梅 迷 昧 晩 kungana: 目 眼 ongana: 毛 ongana: 母 disyllabic kungana: 圓 [mǝrǝ] 諸 [mǝrǝ] ongana: 茂 文 聞 忘 蒙 畝 問 門 勿 木 物 kungana: 裳 藻 哭 喪 裙 disyllabic ongana: 物 [moti] disyllabic kungana: 成 [mori] ongana: 夜 移 陽 耶 益 野 楊 也 kungana: 屋 八 矢 disyllabic kungana: 山 [yama]

ongana: 謎 ongana: 倍 陪 毎 謎 ongana: 煩 ongana: 麻 磨 万 馬 麼 満

mi

mï mu

me

mɛ mo  mǝ mo

ya

魔摩莽

kungana: 真 間 目 disyllabic ongana: 望 [maŋga, maŋgu]

ongana: 弥 美 瀰 湄 弭 寐 kungana: 三 御 見 水 参 ongana: 微 未 味 kungana: 身 実 ongana: 牟 武 模 務 霧 夢茂

kungana: 六 ongana: 賣 咩 謎 迷 綿 kungana: 女 ongana: 梅 迷 昧 毎 妹 kungana: 目 眼 ongana: — ongana: — ongana: 毛 母 茂 望 暮 謀 慕梅謨悶墓 kungana: 裳

ongana: 夜 移 陽 耶 溪 益 野椰揶

kungana: 屋 八 矢 箭

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

38 chart 2

Chapter 2 Man’yōgana signs in varieties A and B (cont.)

Transcription

Variety A

Variety B

yu

ongana: 由 喩 遊 油 kungana: 弓 湯

ongana: 由 喩 愈 瑜 臾

ye

胃踰

kungana: 弓 湯 ongana: 延 曳 遙 kungana: 兄 江 枝 吉 ongana: 用 庸 kungana: 夜 ongana: 余 与 予 餘 預 誉 kungana: 世 吉

ri

ongana: 延 叡 曳 遙 要 kungana: 兄 江 枝 吉 ongana: 用 欲 容 kungana: 夜 ongana: 余 与 予 餘 誉 kungana: 世 吉 四 代 齒o disyllabic kungana: 縦 [yǝsi]p ongana: 羅 良 浪 楽 kungana: 等 disyllabic ongana: 藍 [ramu] 濫 [ramu] 覧 [ramu] 臘 [rapu] 楽 [raku] 落 [raku] ongana: 理 利 里 隣

ru

ongana: 留 流 類

ongana: 留 流 瑠 屢 盧

re

ongana: 禮 礼 例 列 烈 連 disyllabic ongana: 廉 [remu] ongana: 漏 路 ongana: 呂 侶 里 ongana: 和 丸 kungana: 輪 disyllabic ongana: 丸 [wani] disyllabic kungana: 渡 [wata] ongana: 爲 位 謂 kungana: 井 猪 居

ongana: 禮 例 黎 戻

yo yǝ

ra

ongana: 羅 囉 岱 良 楽 邏禮蘿

disyllabic ongana: 楽 [raku] ongana: 理 利 里 梨 離 挌隈 蘆楼漏婁

ro rǝ  wa

wi

we wo

ongana: 惠 廻 慧 佪 kungana: 画 座 咲 ongana: 乎 袁 烏 遠 怨 呼 越 kungana: 小 尾 少 麻 男 雄 緒 綬 𠮧 disyllabic ongana: 越 [woti] [wotǝ] disyllabic kungana: 處 [wotǝ]

ongana: 漏 盧 楼 婁 魯 露 ongana: 呂 侶 慮 盧 稜 ongana: 和 倭 涴 kungana: 輪

ongana: 韋 為 位 威 萎 委偉

kungana: 井 猪 居 ongana: 惠 廻 慧 衛 隈 穢 ongana: 乎 遠 曰 鳴 塢 弘惋越

kungana: 小 尾 少 麻 男雄

disyllabic ongana: 越 [woti] Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

SCRIPT AND PHONOLOGY

39

a Attested only twice in the Man’yōshū, both times in the place name Akï in 1.45 and 1.46 (Bent­ley 2016: 152). b Attested only in place names. c In place names only. d The phonogram 賊 as a sign for /ⁿzǝ/ is attested only in the Man’yōshū (Omodaka et al. 1967: 896). However, it appears as a phonogram in the Man’yōshū just once, in the word kiⁿzǝ (伎賊) ‘last night’ (MYS 2.150). This is the only phonographic attestation of this word in the Western Old Japanese part of the Man’yōshū. Interestingly enough, Omodaka et al. transcribe this word in the same poem as kisǝ in the entry on kisǝ ‘last night’ (1967: 241). In addition, the same word is attested in Eastern Old Japanese four times in 14.3505, 14.3522, 14.3550, and 14.3563 written as 伎曾 /kisǝ/ with a voiceless /s/. Thus, I believe that the phonogram 賊 was read /sǝ/ and I think that this word should be read as kisǝ in Western Old Japanese as well. e It is a rare phonogram in the Man’yōshū, occurring only once with certainty (1.50), and possibly in one other case. f It is used seven times in the Man’yōshū, all of them except one (12.3166), for writing place name Nara. g Strictly speaking, this is a kungana of Korean origin, without a lexical attestation in OJ, cf. OK nik- ‘to be hot,’ ‘to ripen.’ h Used only as a kungana for the attributive form n-ö of the copula n- or the comparative case marker -nö. i Bentley’s saying that “This graph appears several times in Man’yōshū” (2016: 274) is an overstatement, as it appears in this major OJ text only three times (1.22, 6.933, 11.2522). As a matter of fact, the only unambiguous example of 盤 used for pa can be found in 6.933. In 1.22 it clearly stands for ipa, and it is more likely that in 11.2522 it renders pani, and not pa. j Used to write pikï in asi pikï. k Contraction from opo ‘great.’ Attested in MYS 13.3309 and 19.4211. l Attested in MYS 19.4290. m This is a rare phonogram in the Man’yōshū, which is used as a kungana for matu ‘waits’ in 1.73, 13.3258, 13.3324, and possibly in a couple other less clear cases. n It occurs only once in the Man’yōshū (11.2507). o Listed in Bentley (2016: 533), but only as a kungana in 11.2773. Its usage as a quasi-logographic phonogram in 1.10 is not mentioned. p Attested in MYS 2.138.

1.2

Rebus Writing

There are different types of rebus writing occurring predominantly in the Man’yōshū. Rebus writing is comparatively rare, and many examples of it are unique due to its playful and/or associative nature. A student of Old Japanese should not go to the trouble of memorizing rebus writing because of its very infrequent usage. In addition, all types of rebus writing are unsystematic, and only with some difficulty can one devise a classification system of it, as done, for example in (Seeley 1991: 51–52), which I follow here also using Seeley’s examples.

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

40

Chapter 2

(1) Rebus writing based on Chinese usage: for example, characters 金風 in MYS 1700 are read not as *kane kaze ‘metal wind’ but as aki kaze ‘autumn wind,’ reflecting that MC 金風 [kim puwŋʰ] means ‘autumn wind,’ due to the association of metal as one of Five Elements with autumn. In MYS 10.1844 characters 朝烏 are read not *asaŋgarasu ‘morning raven’ but as asapi ‘morning sun,’ reflecting that in Chinese mythology the raven can symbolize the sun. (2) Rebus writing based on purely semantic links: for example, the character 寒 ‘be cold’ is used to write puyu ‘winter’ (MYS 10.1844, MYS 10.1884) and the character 暖 ‘be warm’ is used to write paru ‘spring’ (MYS 3.336, MYS 10.1884). (3) Rebus writing based on double semantic association: for example, characters 少熱 ‘a little hot’ evoke Japanese nuru- ‘be lukewarm,’ and are used to write the homonymous attributive form of the perfective marker -n-uru in MYS 11.2579. (4) Erudite rebuses: for example, the sequence tesi representing a combination of perfective -te- with past tense attributive -si could be written as 大王 (MYS 7.1321) or 羲之 (MYS 3.394, MYS 4.664, MYS 7.1324). There was a word tesi ‘calligrapher’ in Old Japanese (< te ‘hand’ + si ‘master’). Here the rebus is based on the association with the name of the famous Chinese calligrapher, Wang Xizhi (王羲之, 303(?)–361(?) CE), who was also used to be referred to as 大王 ‘the greater Wang’ in opposition to his seventh son Wang Xianzhi (344–388 CE) who was referred to as 小王 ‘the lesser Wang.’ (5) Rebus writing based on double semantic-phonographic association: for example, numeral 八十一 ‘eighty-one’ for 九九 /kuku/ ‘nine nines’ in 八十一里 /kukur-i/, a converb of kukur- ‘to tie up’ (MYS 3330); or numeral 十六 ‘sixteen’ for 四四 /sisi/ ‘four fours’ for writing homophonous sisi ‘boar’ (MYS 3.239, MYS 6.926, MYS 13.3278). (6) Rebus writing based on graphic association, for example 山上復有山 ‘on the top of a mountain there is again a mountain’ (MYS 9.1787). This character sequence represents the verb iⁿde- ‘to go out,’ written logographically as 出, the shape that resembles two signs for ‘mountain’ (山) one placed on the top of the other. There is one more variety of rebus writing that Seely does not mention, although potentially it can be viewed as a peculiar subclass of case 5). In this variety double semantic-phonographic association is used just to indicate one particular syllable. For example, characters 馬声 ‘horse voice’ are used to write syllable /i/, apparently by association with inanak- ‘to neigh’ (Omodaka et al. 1967: 891).

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

Section 2

Phonetics and Phonology On the basis of the syllabic script system described above it is possible to devise the following description of the Western Old Japanese phonetic and phonological systems. As mentioned above, the description below largely follows (Miyake 2003). 2.1 Consonants Cross-linguistically, the consonantal system of Western Old Japanese appears to be quite poor as it only includes thirteen consonants. Nevertheless, it is very interesting typologically, since voiceless stops and fricatives contrast not with plain voiced stops and fricatives, but with prenasalized voiced stops and fricatives. The prenasalized series is secondary; historically it developed from NC [-voice] (nasal + voiceless consonant) clusters, thus mb < *n+p, ⁿd < *n+t, ŋg < *n+k, and ⁿz < *n+s. The consonantal system of Western Old Japanese can be represented in the following chart: chart 3

WOJ consonants

voiceless stops voiced prenasalized stops nasals voiceless fricatives voiced prenasalized fricatives glides flap

Labials

Dentals

p mb m

t ⁿd n s ⁿz

w

r

Palatals

Velars k

ŋg

y

None of the consonants can occur in the word-final or even syllable final position, although it is possible that some representatives of the educated class were able to pronounce /p/, /t/, /k/, /m/, and /n/ in the coda position in loanwords from Chinese. The evidence for this fact during Asuka and Nara periods is rather moot and controversial, and should not entertain us here, simply

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

42

Chapter 2

because most of our sources are poetic texts and poetry was kept more or less free of foreign loans until at least the twelfth century (see chapter 3 for details). In native vocabulary only /p/, /t/, /k/, /m/, /n/, /s/, /w/, and /y/ can occur wordinitially. /mb/, /ⁿd/, /ŋg/ and /r/ appear morpheme-initially, but cannot stand in the word-initial position in native vocabulary. This limitation does not stand for loanword phonology, where /mb/, /ⁿd/, /ŋg/ and /r/ can be used in wordinitial position. A description of the consonants follows. 2.1.1 Labials /p/: bilabial voiceless stop. Miyake demonstrated beyond any reasonable doubt that this Western Old Japanese consonant was indeed a voiceless unaspirated stop [p] and not a bilabial fricative [ɸ] as it was frequently believed before (Miyake 2003: 164–166). /p/ occurs before all vowels, although the distinction between kō-rui vowel /o/ and otsu-rui vowel3 /ǝ/ after /p/ is more or less reliably preserved only in the Kojiki kayō (Bentley 1997). Examples: papa ‘mother,’ pï ‘fire,’ apu ‘meets,’ pi ‘the sun,’ po ‘ear of the grain,’ ipe ‘house.’ It has been repeatedly suggested before that Western Old Japanese voiceless obstruents in the intervocalic position might have been realized as phonetically voiced (Wenck 1959: 238ff; Frellesvig 1995: 66). Several years ago, a similar proposal that WOJ /p/ was pronounced as [b] in the intervocalic position was also put forward (Hamano 2000; Unger 2003). Both Hamano and Unger believe that this scenario better explains why in Middle Japanese /p/ > /w/ in intervocalic position; thus they posit the development p > b > w. The idea (certainly inspired by phonotactics in Korean) is imaginative, but I am afraid it is incorrect, because it is contradicted by the data. First, there is evidence to the contrary from the man’yōgana script itself. Since variety A does not differentiate clearly between /p/ and /mb/, our only bet is to analyze the data from variety B, where the distinction between /p/ and /mb/ is consistent. Miyake and I counted all occurrences of the transcription of intervocalic /p/ in the first 100 poems from the Nihonshoki kayō. There were 155 cases when syllables with intervocalic WOJ /p/ were transcribed by LMC signs with voiceless *p or *ph, and only sixty cases when they were transcribed with LMC voiced *pɦ. Among the latter we could see a very skewed distribution:

3  See the section on vowels for distinction between kō-rui and otsu-rui vowels.

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

43

SCRIPT AND PHONOLOGY chart 4

Distribution of LMC signs with voiced *pɦ for syllables with intervocalic /p/ in the first 100 songs of the Nihonshoki kayō

pa

pi



pu

pe



po

3

2

1

6

7

19

20

Thus, one can see that the distribution of LMC signs with initial *pɦ is heavily skewed towards /pɛ/ and /po/ syllables. I doubt that /pɛ/ and /po/ were selectively pronounced as /bɛ/ and /bo/. Rather, statistically it demonstrates that there is overwhelming evidence that intervocalic /p/ was not voiced, and that the skewed distribution may have a better explanation as general preferences of scribes for certain signs transcribing /pɛ/ and /po/. Second, the Hamano-Unger hypothesis is also contradicted by Old Japanese loanwords in Ainu. In Ainu intervocalic stops are realized as voiced or half-voiced, so if the OJ sipo ‘salt’ was actually pronounced as [sibo] as Hamano and Unger believe, one would expect that Ainu would borrow it as *sipo [sibo/siBo]. As a matter of fact, Ainu borrowed OJ sipo as sippo, clearly showing that the intervocalic consonant there was voiceless. /mb/: voiced prenasalized labial stop [mb]. /mb/ occurs before any vowel only in the word-medial position in native vocabulary, but the distinction between kō-rui vowel /o/ and otsu-rui vowel /ǝ/ after /mb/ is neutralized. Examples: yukɛmba ‘when/because one goes,’ tambi ‘travel,’ tǝmbu ‘flies,’ Unembï ‘Unembï (p.n.),’ sirinumbɛsi ‘must have known,’ sumbe ‘way, method,’ tumbo ‘jar.’ /m/: voiced nasal sonorant labial stop [m]. /m/ occurs before any vowel, but the distinction between kō-rui vowel /o/ and otsu-rui vowel /ǝ/ after /m/ reliably can be shown only for the Kojiki kayō and statistically for some early texts in the Man’yōshū (Bentley 1997). In Late Old Japanese texts the distinction is neutralized. Examples: mamɛ ‘bean,’ umi ‘sea,’ mune ‘breast,’ me ‘woman,’ mɛ ‘eye,’ momo ‘hundred,’ mǝnǝ ‘thing.’ /w/: voiced labial glide [w], possibly labio-velar [ɰ]. Traditionally it is believed that it does not occur before vowel /u/ (but see the section 2.4 on Phonotactics below), and in addition all kō-otsu vowel distinctions are neutralized after /w/. Examples: ware ‘I, we,’ wi ‘well,’ wemu ‘smiles,’ wotǝkǝ ‘man, male,’ awo ‘blue, green,’ awa ‘foam,’ nawi ‘earthquake,’ kǝwe ‘voice.’ 2.1.2 Dentals /t/: voiceless dental stop [t]. /t/ occurs before all vowels, but the distinction between kō-rui vowels /i/, /e/ and otsu-rui vowels /ï/, /ɛ/ are neutralized after

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

44

Chapter 2

/t/. Examples: tare ‘who,’ ti ‘blood,’ te ‘hand,’ to ‘door,’ tǝku ‘unties,’ atǝ ‘footstep,’ ita ‘board,’ otu ‘falls,’ tati ‘long sword,’ mato ‘aim.’ /ⁿd/: voiced prenasalized dental stop [ⁿd]. /ⁿd/ occurs before any vowel only in the word-medial position in native vocabulary, but the distinctions between kō-rui vowels /i/, /e/ and otsu-rui vowels /ï/, /ɛ/ are neutralized after /ⁿd/. Examples: taⁿda ‘direct,’ kaⁿdo ‘gate,’ kaⁿdi ‘rudder,’ tǝⁿdǝmu ‘makes one stay,’ taⁿdu ‘crane.’ /n/: voiced nasal sonorant dental stop [n]. /n/ occurs before any vowel, but the distinctions between kō-rui vowels /i/, /e/ and otsu-rui vowels /ï/, /ɛ/ are neutralized after /n/. Examples: nana ‘seven,’ numa ‘marsh,’ ana ‘hole,’ yǝmanu ‘[one who] does not compose,’ ne ‘sound,’ kane ‘metal,’ sirani ‘not knowing,’ nipi ‘new,’ no ‘field,’ sino ‘a type of bamboo grass,’ nǝru ‘says, declares,’ mǝnǝ ‘thing.’ /s/: voiceless dental fricative [s] with a possible palatal allophone [ɕ] before vowels /i/ and /e/ (Miyake 2003: 183). /s/ occurs before any vowel, but the distinctions between kō-rui vowels /i/, /e/ and otsu-rui vowels /ï/, /ɛ/ are neutralized after /s/. Examples: saru ‘monkey,’ sika ‘deer,’ su ‘does,’ se ‘elder brother, beloved,’ sora ‘sky,’ sǝsǝku ‘pours,’ asa ‘morning,’ asi ‘foot, leg,’ tǝse ‘year (classifier),’ osu ‘pushes,’ asǝmi ‘courtier,’ kǝsǝ, focus particle. /ⁿz/: voiced prenasalized dental fricative [ⁿz] with possible palatal allophones [ᶮj] and [ᶮʑ] before front vowels (Miyake 2003: 186). /ⁿz/ occurs before any vowel only in the word-medial position in native vocabulary, but the distinctions between kō-rui vowels /i/, /e/ and otsu-rui vowels /ï/, /ɛ/ are neutralized after /ⁿz/. In addition, /ⁿz/ before vowel /o/ is attested only in the Man’yōshū, and in one of the Shōsōin documents and only in one word: kaⁿzopu ‘counts,’ e.g., pi-wo kaⁿzopɛ-tutu ‘counting days’ (MYS 5.890). Examples: anᶻaru ‘rots (of a fish meat),’ araⁿzi ‘probably does not exist,’ kaⁿze ‘wind,’ miⁿzu ‘does not see,’ kǝⁿzǝ ‘last year,’ kaⁿzopu ‘counts.’ /r/: flap [|]. /r/ occurs before any vowel only in the word-medial position in native vocabulary, but the distinctions between kō-rui vowels /i/, /e/ and otsurui vowels /ï/, /ɛ/ are neutralized after /r/. Examples: sora ‘sky,’ tǝri ‘bird,’ paru ‘spring,’ kǝre ‘this,’ kǝkǝrǝ ‘heart,’ kuro ‘black.’ 2.1.3 Palatals /y/: palatal glide [j]. /y/ does not occur in front of high vowel /ï/, and the kō-otsu distinction between vowels /e/ and /ɛ/ is neutralized after /y/. Examples: ya ‘arrow,’ yumi ‘bow,’ ye ‘branch,’ yu ‘hot water,’ yo ‘night,’ yǝ ‘world,’ ayameŋgusa ‘iris,’ ayu ‘sweetfish,’ koyu ‘crosses,’ tuyǝ ‘strong,’ to-yo ‘from the door.’

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

45

SCRIPT AND PHONOLOGY

2.1.4 Velars /k/: voiceless velar stop /k/. /k/ occurs before all vowels. Examples: kani ‘crab,’ kiku ‘hears,’ kï ‘tree,’ kepu ‘today,’ kɛ ‘hair,’ ko ‘child,’ kǝkǝrǝ ‘heart,’ aki ‘autumn,’ kake ‘rooster,’ kukï ‘stem,’ pǝtǝkɛ ‘Buddha,’ tako ‘octopus.’ /ŋg/: voiced prenasalized velar stop [ŋg]. /ŋg/ occurs before any vowel only in the word-medial position in native vocabulary. However, there are no cases of /ŋg/ before /e/ in the poetic texts. Examples: aŋgaru ‘rises,’ kiŋgisi ‘pheasant,’ suŋgï ‘cryptomeria,’ kaŋgu ‘smells,’ suŋgɛ ‘sedge,’ aŋgora ‘high seat,’ oŋgǝru ‘behaves arrogantly.’ 2.2 Vowels The prevalent modern view on Western Old Japanese vocalism is that it includes eight vowels, including a, u and six vowels that pair in kō-rui (甲類), or A-type series: i, e, and o; and otsu-rui (乙類), or B-type series: ï, ɛ, and ǝ. Nevertheless, starting from Hattori 1977–1978 various attempts to reduce this number were presented for a long time, the most recent being Miyake theory that the mid-low vowel /ɛ/ was actually [ǝy] (2003: 227–232), thus reducing the number of vowels to seven, and Hayata’s theory arguing for only six vowels in Old Japanese (2017: 53–68). However, none of them is persuasive enough and won general recognition. chart 5

i ɛ

Western Old Japanese vocalism

e

ï [ɨ] ǝ

o

u

a

As one can see, there is really no phonetic motivation behind the division into the kō and otsu series, so these are used as convenient traditional labels only. The only reason for the existence of these terms is the fact that by Middle Japanese, kō-rui and otsu-rui series merged as /i/, /e/, and /o/, leaving only five distinctive vowels in the language: /a/, /u/, /i/, /o/, and /e/. Thus, notions of kō and otsu represent the Western Old Japanese vocalic system viewed from the point of the later language stages. This Middle Japanese vocalic system survived up till now in the Kansai dialects of Central Japanese, and the only

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

46

Chapter 2

change that occurred in Tokyo Standard Japanese was the delabialization of /u/ > /ɯ/. A description of the vowels follows. 2.2.1 Back Vowels /u/: high rounded back vowel [u]. It is believed to occur in any position in the word and after all consonants except /w/, but see section 4 on phonotactics below. Examples: u ‘cormorant,’ urupasi ‘beautiful,’ umu ‘gives birth,’ natu ‘summer,’ puyu ‘winter,’ kumo ‘cloud,’ tuŋgu ‘reports,’ tǝmbu ‘flies,’ miⁿdu ‘water.’ /o/: mid rounded back vowel [o]. /o/ occurs only in a postconsonantal position after the following consonants: /p/, /mb/, /m/, /w/, /t/, /ⁿd/, /n/, /s/, /ⁿz/, /r/, /y/, /k/, and /ŋg/. The contrast between /o/ and central mid vowel /ǝ/ after /p/ and /m/ is attested reliably only in the Kojiki (Bentley 1997; Miyake 2003: 262). There is no /o/ : /ǝ/ contrast after /w/ and /mb/, but Miyake demonstrated that phonetically it was /o/ ([o]) and not /ǝ/ ([ǝ]) after /w/ and that there was a free variation between /o/ and /ǝ/ after /mb/ (Miyake 2003: 257–258) (since there is no contrast, combinations w+o and mb+o will be spelled as wo and mbo hereafter). After /ⁿz/ the vowel /o/ is attested only in one word: kaⁿzop- ‘to count.’ Examples: po ‘ear of the grain,’ mo ‘skirt,’ wosa ‘elder,’ to ‘door,’ kaⁿdo ‘gate,’ no ‘field,’ soⁿde ‘sleeve,’ yo ‘night,’ ko ‘child.’ The vowel /o/ is usually believed to have a partially secondary origin in Western Old Japanese, being a monophthongization of proto-Japonic diphthongs *ua and *au. The origins from the diphthong *ua can be supported by an internal evidence, e.g., kaⁿzop- ‘to count’ < *kaⁿzu ap- ‘to put numbers together.’ However, the origin of /o/ from the diphthong *au is much more controversial, since it can be supported only by external evidence, which is for the most part unreliable. In addition, it seems that in a number of cases, especially in the word-final position, WOJ /o/ may represent an original PJ vowel *o, since this is supported by evidence from proto-Ryukyuan, which preserved the contrast between primary high and mid back vowels. This contrast was lost in Western Old Japanese everywhere except in the word-final position, e.g., WOJ muko ‘son-in-law,’ PR *moko ‘id.’ 2.2.2 Central Vowels /a/: low unrounded central vowel [a], although it also might be a back one [ɑ]. Miyake defines it just as a ‘low unrounded’ without specifying whether it is back or central, because the transcriptional evidence is indeed inconclusive regarding its exact phonetic nature (Miyake 2003: 203). /a/ occurs in any position in the word and after all consonants. Examples: aka ‘red,’ maneku ‘many,’ asana-asana ‘every morning, in the mornings,’ wataramasi ‘[one] would cross,’ tatu ‘stands,’ aŋgaru ‘rises,’ saⁿdamu ‘decides, determines,’ pimbari ‘lark.’ Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

SCRIPT AND PHONOLOGY

47

/ï/: high central vowel [ɨ]. /ï/ occurs only in a postconsonantal position after the following consonants: /p/, /mb/, /m/, /k/, and /ŋg/. The contrast between /ï/ and the high unrounded front vowel /i/ is not attested after coronals, /w/, and in the initial position. There is good reason to believe on phonetic grounds, as demonstrated by Miyake, that a neutralization of this contrast in initial position as well as after coronals and /w/, customarily written as i in Yale romanization, was phonetically [i] (Miyake 2003: 238–243). Examples: pï ‘fire,’ yǝrǝkǝmbï ‘joy,’ mï ‘body,’ kï ‘tree,’ suŋgï ‘cryptomeria.’ Historically, the vowel /ï/ has a secondary origin in Western Old Japanese. It represents monophthongization of PJ sequences C+y: *uy, *oy or *ǝy. Examples: WOJ tukï ‘moon’ < PJ *tukuy, WOJ kukï ‘stem’ < PJ *kukuy, WOJ kï ‘tree’ < PJ *kǝy, WOJ pï ‘fire’ < *poy. /ǝ/: mid central vowel [ǝ]. In contrast to mid back vowel /o/ it can occur in initial position, because, as demonstrated by Miyake, transcriptional evidence points to /ǝ/ rather than to /o/ in this position (Miyake 2003: 250–251). Since there is no contrast between /ǝ/ and /o/ in this position, I preserve the notation o that is used for the neutralization of this contrast. The vowel /ǝ/ also occurs in medial and/or final postconsonantal position after the following consonants: /p/, /mb/, /m/, /t/, /ⁿd/, /n/, /s/, /ⁿz/, /r/, /y/, /k/, and /ŋg/. The contrast between /o/ and central mid vowel /ǝ/ after /p/ and /m/ is attested reliably only in the Kojiki (Bentley 1997; Miyake 2003: 262). There is no /o/ : /ǝ/ contrast after /w/ and /mb/, but Miyake demonstrated that phonetically it was /o/ ([o]) and not /ǝ/ ([ǝ]) after /w/ and that there was free variation between /o/ and /ǝ/ after / mb/ (since there is no contrast, combinations w+ǝ and mb+ǝ will be spelled as wo and mbo hereafter) (Miyake 2003: 257–258). Examples: pǝsu ‘dries,’ mǝnǝ ‘thing,’ tǝkǝrǝ ‘place,’ kǝkǝ ‘here,’ tǝⁿdǝmu ‘stops, makes stay,’ nǝmboru ‘climbs,’ sǝkǝ ‘bottom,’ kǝⁿzǝ ‘last year,’ marǝ ‘round,’ yǝ ‘world,’ kǝsi ‘waist,’ oŋgǝru ‘behaves arrogantly.’ 2.2.3 Front Vowels /i/: high unrounded front vowel [i]. /i/ occurs in initial and postconsonantal positions after all consonants. There is no contrast between /i/ and /ï/ in initial position and after coronals. There is good reason to believe on phonetic grounds, as demonstrated by Miyake, that a neutralization of this contrast in initial position as well as after coronals and /w/, customarily written as i in Yale romanization, was phonetically [i] = /i/ (Miyake 2003: 238–243). Examples: ima ‘now,’ kikinikeri ‘heard,’ kiŋgisi ‘pheasant,’ wi ‘boar,’ titi ‘father,’ pi ‘the sun,’ mi ‘three,’ pimbiku ‘echoes, reverberates,’ aⁿdi ‘taste.’ /e/: mid-high unrounded front vowel [e]. /e/ occurs in postconsonantal position after all consonants except /ŋg/. There is no contrast between /e/ and /ɛ/ in initial position, after coronals, and after glides /w/ and /y/. Miyake Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

48

Chapter 2

persuasively demonstrated that a neutralization of this contrast (spelled as e in Yale romanization) was realized phonetically as [e] = /e/ after coronals and after glides, but as [ɛ] = /ɛ/ in the initial position (Miyake 2003: 244–250). Examples: kepu ‘today,’ se ‘beloved, elder brother,’ kaⁿze ‘wind,’ te ‘hand,’ soⁿde ‘sleeve,’ ne ‘sound,’ pe ‘side,’ sumbe ‘way, method,’ me ‘woman,’ ye ‘branch,’ sǝre ‘that,’ wemu ‘smiles.’ Historically, the majority of cases of the vowel /e/ have a secondary origin in Western Old Japanese. They represent a monophthongization of PJ dipththong *ia. This, of course, does not apply to all cases where /e/ merged with /ɛ/, because the majority of those have an etymological /ɛ/ and not /e/, and to small number of examples in word-final position, where /e/ reflects primary PJ *e. Examples: WOJ me ‘woman’ < PJ *mia, WOJ pe ‘side’ < PJ *pia, but WOJ kake ‘rooster’ < PJ *kake. /ɛ/: mid-low unrounded front vowel [ɛ]. It occurs in the initial position (see under /e/ above) and in postconsonantal position after velar and labial consonants. Examples: takɛ ‘bamboo,’ kaŋgɛ ‘shadow,’ pɛ- ‘to pass,’ tǝmbɛⁿdǝ ‘although [he] flies,’ amɛ ‘heaven.’ The vowel /ɛ/ has a secondary origin, going back mostly to PJ *ay, very rarely to *ǝy, examples: amɛ ‘heaven’ < *amay, mɛ ‘eye’ < *may, kɛ ‘hair’ < *kay, kaŋgɛ ‘shadow’ *kaŋgay, se ‘back’ < *sǝy (cf. sǝ-muk- ‘to turn one’s back on’). 2.3

Pitch Accent

There is a controversial theory advanced by Takayama that Chinese characters with the Late Middle Chinese even tone (pingsheng, 平聲) were used in the Nihonshoki kayō to transcribe Western Old Japanese syllables with low pitch, while Late Middle Chinese characters with other tones: rising (shangsheng, 上聲), departing (qusheng, 去聲), and entering (rusheng, 入聲) rendered Western Old Japanese syllables with high pitch (Takayama 1981, 1983). Takayama’s theory can be supported only by borderline statistics, since there are too many overlaps to the contrary. Therefore, the present study adopts the conservative attitude to the problem, maintaining that we really do not have reliable evidence on the Western Old Japanese pitch accent system. 2.4 Phonotactics It is traditionally believed that Western Old Japanese had the [V]CVCVCV … phonotactics, at least in the native vocabulary, and that there are only few exceptions to this rule, for example kai ‘oar’ or kui ‘regret’ where we have vowel clusters /ai/ and /ui/. The traditional transcription kai is based on the belief Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

49

SCRIPT AND PHONOLOGY

that there were no /yi/ or /wu/ syllables in Western Old Japanese. Nevertheless, this belief presents a significant problem: why does kai ‘oar’ not monophthongize to *kɛ, or contract to *ki as would be expected?4 Before, still sharing this belief, I suggested that kai goes back to earlier *kayi (see this section). Now I would like to revise my previous position, and posit syllables /yi/ and /wu/ for Western Old Japanese itself, and not for pre-Old Japanese. The major evidence for the existence of /yi/ and /wu/ comes from the morphophonology of the Western Old Japanese vowels verbs kǝyi- ‘to lie down,’ oyi‘to age,’ and uwe- ‘to plant.’ Notice that the first two were traditionally transcribed as kǝi- and oi-. But this traditional representation violates the general rule of the Western Old Japanese phonotactics mentioned above and it also raises the same problem as in the case of kai ‘oar’: why do kǝi- and oi- not monophthongize into *kï- and *ï-, or contract into *ki- and *i- respectively? Surely, both monophthongization and contraction are present in the Old Japanese verbal system. For example, kǝ- ‘come’ has the progressive form k-er- < *k-i-ar-, and e- ‘to get’ loses its root altogether in the final or the attributive form: u < *e-u and uru < *e-uru. The case of uwe- ‘to plant’ is even more spectacular. Its attributive form traditionally written as u-uru is attested. But it would be impossible for two identical vowels not to contract into one: examples are plentiful, and the only exceptions occur when a juncture between converb -i and the following auxiliary verb starting with vowel /i/ is present. But it is impossible to posit the juncture between a root and a suffix. Therefore, I believe that the syllable /wu/ was still present and I transcribe consequently this word form as uw-uru. Let me present the actual textual examples before we proceed to further discussion: 1) kǝyi- ‘to lie down’ written phonographically as 許伊 or 己伊: 久佐太袁利志婆刀利志伎提等許自母能宇知許伊布志提

kusa-ⁿ-da-wor-i simba tor-i sik-i-te tǝkǝ ⁿzimǝnǝ uti-kǝyi-pus-i-te grass-?-hand-break-CONV road.side.grass hold-CONV spread-CONV-SUB bed like PREF-lie.down-CONV-lie.prone-CONV-SUB [I] broke off some herbs, and taking some road side grass, [I] spread [it] and lied down on a bed-like [thing]. (MYS 5.886) 宇知奈妣伎登許尓己伊布之

uti-nambik-i tǝkǝ-ni kǝyi-pus-i PREF-stretch.out-CONV bed-LOC lie.down(CONV)-lie.prone-CONV [I] lie down, stretched out on [my] bed. (MYS 17.3969) 4  Cf. ka ~ kɛ ‘day,’ ka ~ kɛ ‘hair’ < *ka-i, and wa-ŋg-imo ‘my beloved,’ wa-ŋg-ipe ‘my house’ < wa-ŋga imo, wa-ŋga ipe. Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

50

Chapter 2

2) oyi- ‘to age’ written phonographically as 淤伊 or 意伊: 和加久閇爾韋泥弖麻斯母能淤伊爾祁流加母

waka-ku pɛ n-i wi ne-te-masi mǝnǝ oyi-n-i-ker-u kamǝ young-CONV ? DV-CONV bring(CONV) sleep(CONV)-PERF-SUBJ CONJ age(CONV)-PERF-CONV-RETR-ATTR EP [I] would have brought [her with me] and have slept [with her] if [she] were young, but [it] turned out that [she] has become old, alas! (KK 93) 意伊豆久安我未

oyi-ⁿduk-u a-ŋga mï age(CONV)-attach-ATTR I-POSS body my body that came close to aging. (MYS 19.4220) 3) The attributive uw-uru of uwe- to plant written phonographically as 宇々流: 比等能宇々流田

pitǝ-nǝ uw-uru TA person-GEN plant-ATTR paddy a paddy that people plant. (MYS 15.3746) Early Middle Chinese and Late Middle Chinese definitely had the contrast *wu : *ʔu and likely also the contrast *yi : *ʔi (Marc Miyake, p.c.). The fact that the characters 伊 and 宇 are respectively EMC and LMC *yi and *wu may further strengthen my point above. Unfortunately, the examples above are written in the variety A of man’yōgana, which was borrowed or at least filtered through a Korean intermediary, as mentioned in 2.1.1. And, as far as we know, the Korean language did not have the contrasts /i/ : /yi/ and /u/ : /wu/ throughout its history.5 Thus, if the scribes in Ancient Japan were trying to adopt a syllabic writing system that did not have these contrasts, we would expect that they may not indicate the contrasts that were present in their own language either. Nevertheless, there is one relevant example written in the variety B of man’yōgana: 俱伊 kuyi ‘regret’ (NK 124), where syllable /yi/ is written with the character 伊 /yi/. In addition, the morphophonological evidence discussed above is not going to disappear either. So far I have established that syllables / yi/ and /wu/ existed in Western Old Japanese in medial position preventing the 5  It remains to see whether the same is true for Old Korean, once we achieve a more accurate decipherment of its texts.

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

51

SCRIPT AND PHONOLOGY

violation of the phonotactic constraint [V]CVCVCV…. The legitimate question that arises next is whether the existence of syllables /yi/ and /wu/ can also be demonstrated in the initial position. Once again the verbal morphophonology comes to our help. The regular vowel verb wi- ‘to sit, to exist’ (upper bi-grade in traditional classification) has the final form that is traditionally spelled as u and is attested once in the phonographic writing in the man’yōgana gloss in the Nihonshoki: 莵岐于

tuk-i-w-u (my spelling) attach-CONV-sit-FIN [she] sat down. (NS 5.166)6 The Nihonshoki uses variety B of the man’yōgana, and the character 于 has LMC reading *wu. Thus, we find extra evidence for the existence of the syllable /wu/ in Western Old Japanese, this time in the initial position. Let me also note that analyzing the final form of wi- ‘to sit, to exist’ as w-u rather than u shows consistency with the conjugation of other vowel verbs, and also does not leave us stranded with a verbal form that has no root. In addition, there are also interesting doublets like ututu and wotutu ‘reality.’ Examples: 打乍二波更毛不得言

ututu-ni pa SARA N-I mo E-IP-AⁿZI reality-LOC TOP again DV-CONV FP POT-say-NEG/TENT [I] would not be able to say [it] again in reality. (MYS 4.784) 宇豆都仁波安布余志勿奈子

ututu-ni pa ap-u yǝsi mo na-si reality-LOC TOP meet-ATTR chance FP no-FIN There is not even a chance to meet in reality. (MYS 5.807) 久志美多麻伊麻能遠都豆尓多布刀伎呂可儛

kusi mi-tama ima-nǝ wotutu-ni taputo-ki rǝ kamu mysterious HON-stone now-GEN reality-LOC awesome-ATTR DV(ATTR) EP [these] mysterious stones are awesome in the present [day’s] reality! (MYS 5.813) 6  Cited according to Kuroita and Matsuyama 1965–66, vol. 1a.

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

52

Chapter 2

宇都追尓之多太尓安良祢婆

ututu-ni si taⁿda n-i ar-an-e-mba reality-LOC EP direct DV-CONV exist-NEG-EV-CON because [it] was not directly in the reality. (MYS 17.3978) 伊尓之敝由伊麻乃乎都豆尓

inisipe-yu ima-nǝ wotutu-ni old.times-ABL now-GEN reality-LOC from old times to the present [day’s] reality. (MYS 17.3985) 伊爾之敞欲伊麻乃乎追通爾奈我佐敞流於夜乃子等毛曾

inisipe-yo ima-nǝ wotutu-ni naŋgas-ap-er-u oya-nǝ KWO-ⁿdǝmo sǝ old.times-ABL now-GEN reality-LOC make.flow-ITER-PROG-ATTR ancestorGEN child-PLUR FP The offspring (lit.: children) of the ancestors who have passed continuously [their glorious names] from old times to the present [day’s] reality. (MYS 18.4094) The form wotutu probably represents a relic of the pre-raised form preserving primary PJ *o. Since the raising of *o > /u/ occurred around 590 AD (Miyake 2003b: 126),7 it is not very likely that initial /w-/ immediately disappeared after that. Thus, it is quite probable that ututu ‘reality’ should be actually transcribed as wututu. A similar case also involves usaŋgi ‘hare’ attested phonographically as 宇佐岐 in the Honzō wamyō (901–923 AD), which at best represents the very late Western Old Japanese attestation. But on the other hand there is EOJ wosaŋgi spelled as 乎佐藝 (MYS 14.3529) that also preserves a pre-raised form. Although here we deal with two different dialects, the case is still quite indicative. Turning now to the problem of the initial /yi-/ in Western Old Japanese, we also have evidence that is based on the following doublets in Western Old Japanese: 7  In a personal communication J. Marshall Unger suggested to me that the raising of midvowels in Pre-Old Japanese must have occurred earlier than 590 AD because if it were around 590 AD then the WOJ tera ‘Buddhist temple’ that certainly entered the language by mid-sixth century would be also raised. I do not think that this argument is valid, because WOJ tera was borrowed from an Old Korean form that was either *tiara or *taira (cf. MK tyèl ‘Buddhist temple’ and see Vovin 2007: 75–77 for a detailed discussion). Neither Pre-OJ *ai or *ia were subject to raising, as we well know, since they resulted in WOJ /ɛ/ and /e/ respectively.

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

53

SCRIPT AND PHONOLOGY

1) yuk- and ik- ‘to go’: 都祢斯良農道乃長手袁久礼々々等伊可尓可由迦牟

tune sir-an-u MITI-nǝ NA ŋGA te-wo kure-kure tǝ ika n-i ka yuk-am-u usual know-NEG-ATTR road-GEN long place-ACC dark-dark DV how DV-CONV IP go-TENT-ATTR How would [I] go along the full length of the road that [I] normally do not know, being in a dark [mood]? (MYS 5.888) 和我勢故波多麻尓母我毛奈手尓麻伎氐見都追由可牟乎於吉氐伊加婆乎思

wa-ŋga se-ko pa tama n-i mǝŋgamo na TE-ni mak-i-te MI-tutu yuk-am-u-wo ok-ite ik-amba wosi I-POSS elder.brother-DIM TOP jade DV-CONV DP EP arm-LOC wrap-CONVSUB look(CONV)-COOR go-TENT-ATTR-ACC leave-CONV-SUB go-COND be.regrettable(FIN) [I] want my elder brother to be a jade! Although [I] would go wrapping [him] around my arm and keep looking at [him], [it] is regrettable if [I] go leaving [him]. (MYS 17.3990) The variant yuk- rules supreme in Western Old Japanese: there are only two examples of ik- attested phonographically and in identical contexts. Nevertheless, ik- does exist and is further amply supported by later Middle Japanese data. 2) yu and i ‘sacred, taboo’: 波毘呂由都麻都婆岐

pa-N-pirǝ yu t-u ma-tumbaki leaf-GEN-broad sacred DV-ATTR INT-camellia a true sacred camellia with broad leaves. (KK 57) 賀美都勢爾伊久比袁宇知斯毛都勢爾麻久比袁宇知

kami-tu se-ni i kupi-wo ut-i simo-tu se-ni ma-kupi-wo ut-i top-GEN/LOC shallow-LOC sacred post-ACC hit-CONV bottom-GEN/LOC shallow-LOC INT-post-ACC hit-CONV in the upper shallows, [they] staked a sacred post, in the lower shallows [they] staked a true post. (KK 90)

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

54

Chapter 2

3) Yuki and Iki ‘Iki island’: 柯羅履儞嗚以柯儞輔居等所梅豆羅古枳駄楼武可左履楼以祇能和駄唎嗚梅 豆羅古枳駄楼

kara-kuni-wo ika n-i [i]p-u kǝtǝ sǝ Mɛⁿdurako k-i-tar-u mukasakuru Iki-nǝ watar-i-wo Mɛⁿdurako k-i-tar-u Kara-land-ACC how DV-CONV say-ATTR thing FP Mɛⁿdurako come-CONVPERF/PROG-ATTR (makura kotoba) Iki-GEN cross-NML-ACC Mɛⁿdurako come-CONV-PERF/PROG-ATTR What to call the land of Kara? Mɛⁿdurako has come; through the crossing of Iki Mɛⁿdurako has come. (NK 99) 到壹岐嶋雪連宅満忽遇鬼病死去之時作歌一首并短歌

One [long] poem and two accompanying tanka envoys composed when Yuki-nǝ muraⁿzi Yakamarǝ upon arrival to Iki island suddenly contracted an evil illness and died. (preface to poems MYS 15.3688–3690) 由吉能安末

Yuki-nǝ ama Yuki-GEN fisherman Fishermen of Yuki island. (MYS 15.3694) In addition, there is a diachronic example of the same alternation: WOJ imɛ ‘dream’ and MJ yume ‘id.’ It is quite possible that Western Old Japanese ik- ‘to go,’ i ‘sacred’ and imɛ ‘dream’ were actually pronounced as /yik-/, /yi/, and / yimɛ/. However, the evidence here is not as conclusive as in the other examples, so I spell these three words in the traditional way. The same applies to ututu ‘reality,’ but in all other cases I introduced new spellings: kayi ‘oar,’ kuyi ‘regret,’ kǝyi- ‘to lie down,’ oyi- ‘to age,’ uw-uru ‘plant-ATTR,’ and w-u ‘sit-FIN.’ Thus, I believe I was able to present some compelling evidence in favor of the fact that the syllables /yi/ and /wu/ probably were present in Western Old Japanese. 2.5

Morphophonological Processes

The morphophonological processes that affect Western Old Japanese due to the phonotactics rule outlined above are contraction, monophthongization, nasalization, and consonant cluster simplification. For the first two processes

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

SCRIPT AND PHONOLOGY

55

I follow Russell (2003), a new study of contraction and monophthongization, which supercedes an older study by Unger (1993). 2.5.1 Contraction Since Western Old Japanese did not allow vowel clusters, one of the two ways to deal with a vowel cluster when vowels were juxtaposed was to delete one of the vowels. Contraction, therefore, involves a deletion of one of the vowels V₁ or V₂ in a V₁ + V₂ sequence. According to Russell’s study, contraction is triggered when: 1) the morphemic boundary between the following categories of words is lost: – stative verb stem + noun – stative verb converb + verb – auxiliary verb + stative verb – noun + noun – particle + noun – particle + active verb – active verb + active verb 2) monophthongization cannot occur because two vowels that come together do not monophthongize. This is true when the two vowels in question are the same (a+a or i+i), when one of the vowels is already a result of monophthongization (e+a), or there is simply no known vowel that results from monophthongization of a vowel cluster, as is the case with i+u, a+ǝ, ǝ+a, and ǝ+u’ (Russell 2003: 522–523). Russell divides contraction cases into two groups (Russell 2003: 523): Group I: polysyllabic word + monosyllabic word: V₁ + V₂ > V₂ polysyllabic word + polysyllabic word: V₁ + V₂ > V₂ Group II: monosyllabic word + polysyllabic word: V₁ + V₂ > V₁ However, one can see that the division in Group I is artificial because it results in the same kind of contraction. It must be noted that the overwhelming majority of ‘monosyllabic words’ in Group II are just clitics. Therefore, I rewrite Russell’s two groups as: Group I: polysyllabic word + any word: V₁ + V₂ > V₂ Group II: monosyllabic + polysyllabic word: V₁ + V₂ > V₁ Examples for Group I: kanasi-ku+ar- → kanasikar- (MYS 5.793) sad-CONV+exist → is sad (clause) tǝ+ip- → (clause) t-ip- (MYS 5.800) (clause) DV say → (clause) DV-say (NP) n-i+ar- → n-ar- (MYS 5.805)

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

56

Chapter 2

(NP) DV-CONV exist → (NP) DV-exist V-i-te+ar- → V+i+tar- (SM 28) V-CONV-SUB exist → V-CONV-PERF/PROG V-aⁿz-u+ar- → V-aⁿz-ar- (MYS 17.3891) V-NEG-CONV exist → NEG-exist apa+umi → apumi (MYS 9.1757) light+sea → fresh water lake ara+iso → ariso (MYS 2.181) rough+rock → jagged rocks kata+uwo → katuwo (MYS 9.1740) hard+fish → katsuo (dried fish) Examples for Group II: (clause) tǝ+ip- → (clause) tǝ-p- (MYS 15.3638) (clause) DV say → (clause) DV-say ko+um- → kom- (KK 71) egg+lay → lay an egg imo-ŋga ipe → imo-ŋga-pe (MYS 5.844) beloved-POSS house → house of one’s beloved kǝkǝrǝ pa omǝp-ɛ-ⁿdǝ → kǝkǝrǝ pa mǝp-ɛ-ⁿdǝ (KK 51) heart TOP love-EV-CONC → although the heart loves mi+uma → mima (MYS 5.877) PREF(HON)-horse → honorable horse There are, however, notable exceptions to both Group I and II. First, some of the polysyllabic words sometimes retain their last vowel: are omǝp-u → are mǝpu (against the rule for the Group I) (MYS 5.852) I think-FIN → I think Second, the behavior of the possessive case marker ŋga may be unpredictable, cf. the following two examples: wa-ŋga ipe → wa-ŋga pe (as above, according to the rule for the Group II) (MYS 5.837) I-POSS house → my house wa-ŋga ipe → wa-ŋg-ipe (against the rule for the Group II, but according to the rule for the Group I) (NK 21) I-POSS house → my house 2.5.2 Monophthongization Russell proposes the following conditions for monophthongization: 1) a morpheme boundary between the following categories of words is lost: a) auxiliary verb + auxiliary verb, b) noun + verb; 2) a consonant is lost, resulting in vowel + vowel sequence (Russell 2003: 523). Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

SCRIPT AND PHONOLOGY

57

The second condition is speculative, because it involves postulating a nonspecified consonant *C (Russell 2003: 531), which cannot be reconstructed by the regular comparative method for proto-Japonic. Consequently, it is ignored in this grammar. Examples of monophthongization with a lost morphemic boundary: V-i-ki+ar- → V-i-ker- (MYS 5.836) V-CONV-PAST+exist → V-CONV-PRET V-i-ki+am- → V-i-kem- (MYS 2.104) V-CONV-PAST-TENT → V-CONV-PAST/TENT V-i-ar- → V-er- (MYS 1.16) V-CONV-exist → V-PROG kaⁿzu+ap- → kaⁿzop- (MYS 5.890) number+join → count naŋga+iki → naŋgɛki (MYS 2.199) long+breath → sigh taka+iti → takɛti (KK 101) high+market → Taketi (place name) opǝ+isi → opïsi (KK 13) big+stone → rock tǝnǝ+iri → toneri (MYS 16.3791) pavilion+entering → retainer situ+ori → sitori (NS: II) weaving pattern+weaving → native weaving pattern 2.5.3 Nasalization Nasalization is a process that occurs when a voiceless obstruent becomes a prenasalized voiced one due to vowel elision between a nasal sonorant (predominantly /n/) and a following voiceless obstruent. Since the rules of Western Old Japanese phonotactics do not allow consonantal clusters, the sequence consisting of a nasal sonorant followed by voiceless obstruent results in a prenasalized voiced obstruent ⁿC. This vowel elision occurs most frequently in genitive case marker -nǝ, but there are also cases when it involves the dative-locative case marker -ni, the comparative case marker -nǝ, and the converb form n-i or the attributive form n-ǝ of the defective verb n-. Examples: sora-ⁿ-di (MYS 15.3694) < *sora-nǝ ti ‘way to Heaven’ (cf. ti-mata ‘road fork’), no-m-biru (KK 43) < no-nǝ piru ‘field garlic,’ taku-m-busuma (KK 5) < taku-nǝ pusuma ‘taku covers,’ mi-na-ŋ-gipa (MYS 20.4462) < mi-na-nǝ kipa ‘riverbank (lit.: edge of the water),’ Yamasirǝ-ŋ-gapa (KK 58) < Yamasirǝ-nǝ kapa ‘Yamasirǝ river,’ ipe-ⁿ-zakar(MYS 5.794) < ipe-ni sakar- ‘to keep distance from home,’ mi-yama-ŋ-gakur-i-te (KK 112) < mi-yama-ni kakur-i-te ‘hiding in the deep mountains,’ miⁿdu-nǝ kaper-an-u-ŋ-gǝtǝ-ku (MYS 15.3625) < miⁿdu-nǝ kaper-an-u n-i kǝtǝ-ku ‘like the Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

58

Chapter 2

water that does not return,’ osu-m-pur-ap-i (KK 2) < *osu n-i pur-ap-i ‘repeatedly pushes and shakes,’ pikǝ-ⁿ-dur-ap-i (KK 2) < *pikǝ n-i tur-ap-i ‘repeatedly pulls and shoves,’ ma-tama-ⁿ-de (KK 3) < ma-tama-nǝ te ‘hands like jewels.’ There is one case of the elision of the last vowel in the root after a nasal: siⁿ-tu ye (KK 100) < simo-tu ye ‘bottom branches.’ This also appears to be the only case when the nasal /m/ and not /n/ is the source of the nasalization. However, in some rare cases the prenasalized forms in compounds reflect not the vowel elision between a nasal and a voiceless obstruent, but indicate that originally the first element of compound ended with a nasal, e.g., wombune ‘small boat’ (KK 52), woⁿ-dani ‘little valley’ (KK 111), woⁿdi ‘old man’ (MYS 17.4014) < woⁿ- ‘small’ + ti ‘father,’ wa-ⁿ-dǝri ‘my bird’ (KK 3), na-ⁿ-dǝri ‘thy bird’ (KK 3), a-ŋ-ko ‘my children’ (NK 8). 2.5.4 *-r- Loss The process of *-r- loss (Whitman’s law) is a residue from the previous stage of the language, no longer active in Western Old Japanese.8 However, it is important to keep in mind, as it affected verbal morphophonology. Whitman argued that *-r- was lost in pre-OJ if the preceding vowel was short, but retained if the preceding vowel was long (Whitman 1985: 190–201). At first glance, this explanation may seem to be circular, because the evidence for pre-OJ vowel length, especially in non-first syllables, is moot, to say the least. However, it turns out that it works in the case of verbal morphophonology if we treat vowels resulting from contraction or monophthongization as long. Thus, pre-OJ *-uro > WOJ -uru,9 attributive, and pre-OJ *-ure > WOJ -ure, evidential, undergo the following developments depending whether they follow consonantal or vowel verbal stems, cf. the following examples involving consonantal verb omǝp- ‘to think, love’ and vowel verb kopï- ‘to love.’ *omǝp-uru → *omǝp-u[r]u → omǝp-u *omǝp-urɛ → *omǝp-u[r]ɛ → omǝp-ɛ *kopï-uro → *kop-ūru → kop-uru *kopï-urɛ → *kop-ūrɛ → kop-ure

8  There seems to be two possible exceptions to this rule, reflected by WOJ kusurisi ~ MJ kususi ‘medicine man,’ and WOJ urutapu ~ MJ utapu ‘to complain’ (no direct phonographic Western Old Japanese attestations of urutapu, the word is usually cited on the basis of later glosses (Omodaka et al. 1967: 137). 9  In Western Old Japanese pre-OJ *o underwent raising to *u, thus pre-OJ *-uro > WOJ -uru.

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

chapter 3 Lexicon



Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

Contents of Chapter 3 Chapter 3 Lexicon 62 1 Naturalized Loanwords from Ainu 63 2 Naturalized Loanwords from Korean 64 3 Naturalized Loanwords from Old and Middle Chinese 67 4 Loanwords from Chinese in Poetry 70 5 Loanwords from Chinese in Prose 72

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

As a lexicon of any human language, the lexicon of Western Old Japanese surely consisted of both native vocabulary and loanwords. However, the usage of loanwords was restricted in poetry (although there were some exceptions), and since most of our sources on Western Old Japanese represent poetic texts, we will probably never know the extent to which loanwords were employed in the everyday language of the time. This is especially true since the only prose source that uses loanwords more or less extensively, the Senmyō, has two significant limitations. First, many of the loanwords from Chinese are written in this text logographically by Chinese characters. Second, the vocabulary of the Senmyō is very specific, since it represents imperial edicts. The only other Western Old Japanese prose text, the Norito, is hardly helpful at all. Since it is a collection of the Shintō prayers it cannot be expected to contain a substantial amount of foreign vocabulary. However, there are two major exceptions to this otherwise almost complete absence of loanwords. The first one concerns those loanwords that are frequently labeled ‘prehistoric,’ although not all of them are prehistoric in the true sense of the word. Rather, it involves those borrowings that have already become the flesh and blood of the Old Japanese language, and were no longer perceived as foreign elements. These words, therefore, were treated as an organic part of the native vocabulary, and they were freely used in poetry. There are three recognizable sources of naturalized loanwords: Ainu, Korean, and Chinese. However, there were undoubtedly many more, including languages that were once spoken in the Japanese archipelago. Since Japonic absorbed these languages comparatively early, almost no traces of them remain today. Identification of these potentially foreign elements in Japanese is an exciting, albeit controversial topic, which falls outside of the scope of the present work; therefore, my focus here will be on the immediately identifiable loanwords. The second major exception is due to the usage of Chinese loanwords in the so-called ‘playful poems’ (tawamure-no uta) that are heavily concentrated in volume sixteen of the Man’yōshū (Saeki 1959: 35).

© Alexander Vovin, 2020 | doi:10.1163/9789004422810_004

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

Section 1

Naturalized Loanwords from Ainu Apart from place names, e.g., WOJ Nǝtǝ, a name for the Noto peninsula (< Ainu not ‘cape,’ or noto ‘big cape’), there are very few Ainu loanwords in Western Old Japanese.1 One of them is WOJ kanipa (or possibly kanimba) ‘sakura bark,’ likely a loanword from Ainu karinpa ‘bark of mountain sakura’ (Saeki 1959: 32). Another example is the diminutive prefix wo- < woⁿ- < *bon-, a likely loanword from Ainu pon ‘small’ (see chapter 4, section 1.1.3 for details).

1  But there are plenty of Ainu loans in Eastern Old Japanese, see Vovin 2012: 11–12 and 2013: 13–15.

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

Section 2

Naturalized Loanwords from Korean Loanwords from Korean are much more numerous than loanwords from Ainu and they are found in virtually all layers of Western Old Japanese vocabulary, including basic vocabulary. This is contrary to the general belief that most of the loanwords from Korean are limited to Buddhist religious vocabulary. The difficulty of their identification is due to the fact that Western Old Japanese borrowed heavily from now extinct language of the Paekche kingdom, from which only few glosses remain. Nevertheless, the tentative identification of certain ‘native’ Western Old Japanese vocabulary items as Korean loanwords is still possible due to two factors. First, the Old Korean language, the language of the Silla kingdom, is attested much better, with several texts surviving. The Koreanic language of Paekche is related reasonably close to Old Korean. In addition, Middle Korean, which is not quite a direct descendant of Old Korean, but which is based on an Old Korean dialect very close to the Silla language, is attested quite perfectly. Second, a number of Western Old Japanese ‘native’ vocabulary items have parallels in Old and/or Middle Korean, but do not have any cognates either in Eastern Old Japanese or in Ryukyuan. Rather than view them as Koreo-Japonic cognate sets, supporting the genetic relationship of Japonic and Korean, I prefer to identify them as loanwords from some variety of Old Korean (possibly Paekche). If they were genuine Koreo-Japonic cognates, one would expect them to show up in Eastern Old Japanese and/or in Ryukyuan. The very fact that they are absent from the two branches of Japonic that were not in direct contact with ancient languages of the Korean peninsula, speaks in favor of the loanword scenario. Besides, these words usually have semantic doublets in Western Old Japanese. The doublets normally have cognates in Eastern Old Japanese and/or Ryukyuan, but no parallels in Old and/or Middle Korean. The following chart illustrates these two vocabulary layers in Western Old Japanese lexicon. The first word in each pair (indexed with subscript ‘₁’) is apparently native, because it has obvious cognates in Eastern Old Japanese and/or Ryukyuan. The second word in each pair (indexed with subscript ‘₂’) is likely a loanword from Korean, since there are no Eastern Old Japanese and/or Ryukyuan cognates, but there are Korean parallels.

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

65

LEXICON chart 6

Lexical doublets in Western Old Japanese

Father₁ Father₂ mother₁ mother₂ elder brother₁ elder brother₂ head₁ head₂ eye₁ eye₂ snow₁ snow₂ sea₁ sea₂

WOJ

EOJ

Ryuk.

OK

MK

titi kaso papa omo ani se MJ tub/muri -matab mɛ na-mita ‘tear’c yuki MJ na-dared umi wata

titi — papa amo ani se — — mɛ — yoki — umi —

cici — fafa — ?ani — çiburu — *me nada *yoki — *omi —

— kaso (Paekche) — *EME? — ? — EMK mati — *NWUN? — ? — ?

— — — emi — hyenga — mari — nwun — nwun — parar, patah

a Not a Chinese loan, as frequently assumed. Chin. 兄 ‘elder brother’ has MK Sino-Korean reading /hywuyeng/, modern Sino-Korean /hywung/. b Preserved in ya-mata nǝ woroti ‘eight-headed serpent’ (Vovin 2000: 142–143). c Lit.: ‘water-eye,’ and not ‘eye-water’ (Vovin 2010b). d Nadare ‘snowslide, avalanche.’.

Certainly, naturalized Korean loanwords are not limited to basic vocabulary alone. On the contrary, many more are found in cultural vocabulary. Some examples: WOJ tera ‘Buddhist temple,’ borrowed from some Old Korean dialect, probably Paekche: cf. MK tyél ‘id.’ MK form apparently underwent apocope, and WOJ preserves the old form of the word. WOJ pǝtǝkɛ ‘Buddha,’ borrowed from some Old Korean dialect, probably Paekche: cf. MK pwuthye ‘id.’ MK form apparently underwent syncope, resulting from OK *pwutVhye (cf. Manchu fucihi ‘Buddha,’ also borrowed from some OK dialect, probably Parhae). WOJ kusirǝ ‘bracelet made of precious stones’ borrowed from some Old Korean dialect, probably Paekche: cf. MK kwùsúl ‘bead, precious stone.’ WOJ katana ‘single blade sword’ is an apparent compound consisting of kata‘one’ and na ‘blade.’ This compound as such does not present itself in Modern

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

66

Chapter 3

Korean or Middle Korean, but this can be easily explained by the fact that Koreans in recent historical times used only double-blade swords. Although kata- ‘one’ is found in a number of compounds even in modern Japanese, such as kata-miti ‘one way,’ kata-asi ‘one foot,’ kata-ude ‘one arm,’ kata-omoi ‘unreciprocated love,’ kata-oya ‘one parent,’ kata-gawa ‘one side,’ kata-toki ‘single moment,’ etc., it is not a native word. The native vocabulary item is WOJ pitǝ‘one.’ Kata- ‘one’ is an apparent loanword from Korean: although MK has honah (Modern Standard Korean hana) ‘one,’ pre-fifteenth-century Korean materials clearly indicate earlier *xata- with *-t- rather than *-n-: EMK xatun (Kyeylim yusa #19), OK HAtʌn (一等) ‘one’ (Kwangumka 6, 8). WOJ na ‘blade’ < MK nolh [nʌrh] ‘id.,’ cf. native WOJ pa ‘blade.’ WOJ pati ‘bowl for alms.’ The ultimate source for this word is Sanskrit pātra or pātrī ‘bowl,’ but the immediate source is certainly Korean, the word being borrowed from Korean before the lenition -t- > -r- took place, cf. MK pal [par] ‘bowl for alms’ < OK *pati. Unless the immediate Korean source is assumed for this word (since no other direct loans from Sanskrit to Western Old Japanese are known), it would be difficult to explain the following: a) the same restricted semantics (‘bowl for alms’ vs. ‘bowl’), b) final vowel -i (that can be easily explained as Korean diminutive -i) rather than -u, which would be more likely to be expected if the word were a direct loan from the Chinese intermediary (EMC pat ‘bowl for alms’).

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

Section 3

Naturalized Loanwords from Old and Middle Chinese There are few naturalized loanwords from Old and Middle Chinese, and even fewer are completely unproblematic. To illustrate the possibilities and the problems that besiege them, I will use the following four examples. WOJ we ‘picture’ is a loanword from EMC ɣwɛ:jʰ (畫) ‘id.’ It freely appears in different kinds of texts, including poetry, which probably points to the fact that even then this word was not perceived as an alien element. WOJ tǝnǝ ‘pavilion’ is likely to be a borrowing of OC *tˁǝ[n]-s (殿) ‘pavilion.’ Phonetically it is quite plausible, with an echo vowel added after the Chinese final consonant. The echo vowel principle also appears in some ongana signs that were used as disyllables, e.g., MC 作 (EMC tsak) or 尺 (MC tɕʰiajk) rendering the WOJ sequence /saka/. It is also very plausible in a cultural context, because the building of pavilions was certainly an enterprise borrowed from China, whether directly or indirectly through a Korean intermediary. The surviving samples of native Japanese architecture certainly do not include any pavilions, that all appear to be based on existing Chinese models. At any rate, this etymology is much more certain that the other two that will be discussed below. WOJ uma ‘horse’ is almost universally believed to be an early loanword from OC *mra (馬) ‘horse’ (Saeki 1959: 33). It is not quite clear why OC cluster *mrwould be reflected as *um- in Western Old Japanese, one would expect more realistic reflexes *m- or *n- < *r-, thus, WOJ *ma or *na would be expected. It is highly unlikely that the word was borrowed from MC *ma, because in this case initial *u- in Western Old Japanese would be still crying for an explanation. An even greater problem confronting this comparison involves the reconstruction of the proper proto-Japonic archetype of the word. There is a diachronic alternation ∅- ~ m- in this word: WOJ uma ~ MJ muma ‘horse.’ The same alternation between WOJ form without initial m- and MJ with initial m- is observed in some other words, e.g., WOJ umɛ ~ MJ mume ‘plum’ (discussed below), WOJ umaŋgwo ~ MJ mumago ‘grandchild,’2 WOJ uma- ~ MJ muma- ‘to be good, excellent, splendid, delicious,’ WOJ umbap- ~ MJ mubaf- ‘to take away,’ WOJ umbara 2  Forms muma ‘horse’ and mumaŋgo ‘grandchild’ with initial m- are also attested in Western Old Japanese, but they occur more seldom than forms without m-.

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

68 chart 7

Chapter 3 Reflexes of um-/uⁿ - in WOJ, EOJ, MJ and Shuri

Gloss

WOJ

EOJ

MJ

Shuri

horse plum grandchild good/tasty take away thorn necessity sea pus give birth bury

uma umɛ umaŋgwo umaumbapumbara umbɛ umi umi umum-

mumaa – – – – – – umi – – –

muma mume mumago mumamubafmubara mube umi umi umume-

’Nma ’Nmi ’Nmaga maasa-N ’Nbay-uN –b – ’umi ’Nmi ’Nmari’Nbee-y-uN

a The form uma is attested also, but only in those Eastern Old Japanese texts that have no or very few real Eastern Old Japanese features (MYS 14.3439, 3537a, 3538). b Shuri ’ibayaa ‘thorn’ is a late loan from Japanese ibara ‘id.’

~ MJ mubara ‘thorn, brier,’ WOJ umbɛ ~ MJ mube ‘really, necessity.’ It is usually believed that the Western Old Japanese forms without m- are more archaic than Middle Japanese forms with m-. However, it is necessary to explain why there are also a number of words where WOJ um-/uⁿ- corresponds to MJ um/-u˜-3 and not to MJ mum-/mu˜-, e.g.: WOJ umi ~ MJ umi ‘sea,’ WOJ umi ~ MJ umi ‘pus,’ WOJ um- ~ MJ um- ‘to give birth,’ WOJ um- ~ MJ ume- ‘to bury.’4 Let us compare these two groups of words with their reflexes in Eastern Old Japanese and Ryukyuan (Modern Shuri will be used as an example). With the exception of Shuri maasa-N ‘is delicious’ that represents an irregular development, we can see three distinct types of correspondences: 1) WOJ um- ~ EOJ mum- ~ MJ mum- ~ Shuri ’Nm2) WOJ um- ~ EOJ um- ~ MJ um- ~ Shuri um3) WOJ um- ~ EOJ ? ~ MJ um- ~ Shuri ’Nm-

3  Middle Japanese consonants transcribed as b, d, g, z were either prenasalized voiced obstruents like WOJ mb, ⁿd, ŋg, ⁿz, or they triggered nasalization over a preceding vowel. 4  There are intransitives mumare- ‘to be born’ and mumore- ‘to be buried’ attested with initial m- in Middle Japanese, but these are probably analogical forms, since transitive verbs um- ‘to give birth’ and ume- ‘to bury’ never appear as *mum- and *mume-.

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

69

LEXICON

Serafim suggested that the second correspondence should be interpreted as a reflex of PJ sequence *om-, with primary PJ *o that underwent raising to /u/ in WOJ (unconditional), EOJ (in this case), and in initial position in Shuri (Serafim, personal communication). The third correspondence is then likely to reflect PJ *um-, which was reflected as such in Central (WOJ, MJ) branch of Japanese, but shifted to ’Nm- in Shuri. That leaves only one option for interpretation of the first correspondence: as PJ *mum-. Thus, the reconstruction of the PJ archetype for WOJ uma ‘horse’ should be *muma, and this short discussion demonstrates again that Western Old Japanese is not the Latin of the Japonic language family. It turns out to be the most innovative among all the Japonic languages as far as its treatment of PJ sequences *mum-, *um- and *om- is concerned: it is the only Japonic language among those used above that merged them all. Therefore, it is quite doubtful that PJ *muma ‘horse’ (> WOJ uma) can be a loanword from OC *mra. The reflex of OC initial cluster *mr- as PJ *mum- is not very likely. This kind of a phonetically strange development is placed even more in doubt by an apparent irregularity that arises if we take into consideration the hypothesis that WOJ umɛ ‘plum’ is also a loanword from Old Chinese, which is discussed next. WOJ umɛ ‘plum’ is usually considered to be an early loanword from OC *C.mˁǝ5 (梅) ‘plum’ (Saeki 1959: 33). However, as demonstrated above the archetype of WOJ umɛ ‘plum’ was PJ *mumay with an initial m-. This creates a problem in regularity of the correspondences if we compare this word with WOJ uma ‘horse’ and its alleged OC counterpart: gloss WOJ PJ OC horse uma *muma *mˁraɁ plum umɛ *mumay * C.mˁǝ We have two almost identical shapes in proto-Japonic, but the alleged Old Chinese sources are very different. It is really inconceivable how a PJ *mumay or WOJ umɛ could develop an extra initial syllable if the word for ‘plum’ was really borrowed from Old Chinese. Thus, I must conclude that neither WOJ uma ‘horse’ nor WOJ umɛ ‘plum’ are loanwords from Chinese. Their etymology remains obscure.

5  I follow here the traditional Old Chinese reconstruction by Baxter and Sagart (2014: 351).

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

Section 4

Loanwords from Chinese in Poetry Most Chinese loanwords found in Western Old Japanese poetry represent religious vocabulary, in particular vocabulary that is related to Buddhism. The lion’s share of all Chinese loanwords in poetry is concentrated in volume sixteen of the Man’yōshū. The usual reason cited is that MYS 16 contains many tawamure-no uta, or ‘playful poems’ (Saeki 1959: 35) although not all the poems in volume sixteen containing Chinese loanwords are playful in their content. The following loanwords from Chinese are attested in Western Old Japanese poetic texts:6 suŋguroku ‘sugoroku (name of a game)’ < EMC ʂɶ:wŋ luwk (雙六) (MYS 16.3827, MYS 16.3838), kau ‘fragrance’ < EMC xɨaŋ (香) (MYS 16.3828), kuu ‘achievement’ < EMC kǝwŋ (功) (MYS 16.3858), ŋgowi ‘fifth rank’ < EMC ŋɔ’ wiʰ (五位) (MYS 16.3858), popusi ‘Buddhist priest’ < EMC puap ʂi (法師) (MYS 16.3846), ŋgakï ‘hungry demon’ < EMC ŋaʰ kuj’ (餓鬼) (MYS 4.608, MYS 16.3840), puse ‘offering (at a Buddhist temple)’ < EMC pɔʰ ɕiǝ̌ (布施) (MYS 5.906), kuwasǝ [kwasǝ] ‘travel pass’ < EMC kwaʰ ʂiǝ̌’ (過所) (MYS 15.3754), rikiⁿzi ‘Buddhist guardian’ in rikiⁿzi-mapi ‘Buddhist guardian dance’ < EMC lik dʑi’ (力士) (MYS 16.3831), muŋgau ‘non-existing’ in muŋgau n-ǝ sato ‘non-existing village’7 < EMC muǝ̌ ɣa wuw’ (無何有) (MYS 16.3851), pakwoya ‘Pakoya, p.n.’ in pakwoya n-ǝ yama ‘mythical mountain Pakoya where the immortals live’8 < EMC mjiaw’ kɔ jiaʰ (藐孤射) (MYS 16.3851), teusamu ‘going to the court’ < EMC driaw tsʰam (朝参) (MYS 18.4121). Some of the Chinese loanwords can form compounds with native vocabulary items. Thus, the loanword ŋgakï ‘hungry demon’ can occur independently (MYS 4.608), and also can form compounds with native vocabulary items: me-ŋgakï ‘female hungry demon,’ wo-ŋgakï ‘male hungry demon’ (MYS 16.3840). 6  The following list is based mostly on the list found in Saeki (1959: 34), with one addition from Shirafuji (1987: 203). Shirafuji provides several other cases, not included in the list in Saeki (1959: 34), but they are more controversial, as native readings may in fact hide behind the logographic script in these cases. For example, the characters 生死 in MYS 16.3849 are read as a loanword syaunzi by Shirafuji, but native iki-sini fits better in the meter of the poem (Takagi et al. 1962: 151). 7  An image taken from Zhuangzi that symbolizes the ideal unspoiled abode of immortals. 8  Another image taken from Zhuangzi.

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

LEXICON

71

The loanword rikiⁿzi ‘Buddhist guardian,’ on the other hand, occurs in poetry only in the compound rikiⁿzi-mapi ‘Buddhist guardian dance’ (MYS 16.3831). There are also three loanwords borrowed from Sanskrit via Chinese that also appear in poetry. All of them occur only in the volume sixteen of the Man’yōshū: tapu ‘pagoda’ < EMC tʰap (塔) < Sanskrit stūpa (MYS 16.3828), ⁿdaniwoti ‘almsgiver’ < EMC dan wuat (檀越) < Sanskrit dānapati (MYS 16.3847), mbaramoni ‘Brahman’ < EMC ba la mǝn (婆羅門) < Sanskrit brāhmaṇa (MYS 16.3856).

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

Section 5

Loanwords from Chinese in Prose Chinese loanwords in prose are much more numerous in spite of the fact that they all come from the same text: the Senmyō, which is the only prose text that contains borrowings from Chinese. The following Chinese loanwords (including loanwords from Sanskrit via Chinese) are attested in the Senmyō:9 uwi ‘[people] with a rank’ < EMC wuw’ wiʰ (有位) (SM 42), yeni ‘karma’ < EMC jwian (縁) (SM 41), yemimi ‘witchcraft’ < EMC Ɂjiam miʰ (厭魅) (SM 54), omuyau ‘Ying and Yang’ < EMC Ɂim jɨaŋ (陰陽) (SM 42), kaundoku ‘reading’ < EMC kaɨwŋ dǝwk (講讀) (SM 42), ŋgaku ‘music’ < EMC ŋaiwk (樂) (SM 9), kitiⁿzyauteni ‘name of Bishamon’s spouse’ < EMC kjit ziaŋ thɛn (吉祥天) (SM 42), ŋgipu ‘righteous husband’ < EMC ŋiʰ puǝ̌ (義夫) (SM 42), kiyau ‘sutra’ < EMC kɛjŋ (經) (SM 28, SM 38), ŋgiyau ‘deed’ < EMC γaɨjŋʰ (行) (SM 41), kwanuseomu mbosati ‘bodhisattva Kannon’ < EMC kwanʰ ɕiajʰ Ɂim bɔ sat (観世音菩薩) (SM 19, SM 43), keiuni ‘auspicious (lit.: scenic) cloud’ < EMC kijaŋ’ wun (景雲) (SM 42), keuŋgi ‘filial piety’ < EMC xɛ:wʰ ŋiʰ (孝義) (SM 13, SM 48, SM 50, SM 55, SM 61), keuⁿzi ‘respectful son’ < EMC xɛ:wʰ tsɨ’ (孝子) (SM 42), keumbu ‘virtuous wife’ < EMC xɛ:wʰ buw’ (孝婦) (SM 42), kekwa ‘penance’ < EMC xwǝj’ kwaʰ (悔過) (SM 42), kokuwau ‘country’s king’ < EMC kwǝk wuaŋʰ (國王) (SM 28), ŋgopopu ‘protect the Buddhist Law’ < EMC γɔ’ puap (護法) (SM 19), samuseni ‘triple selection’ < EMC sam swian’ (三選) (SM 42), samumbau ‘Three Treasures’ < EMC sam paw’ (三宝) (SM 12, SM 13, SM 38, SM 42, SM 45), ⁿzamu ‘beheading’ < EMC tʂɛ:m’ (斬) (SM 53, SM 62), si ‘mentor, teacher’ < EMC ʂi (師) (SM 24, SM 28, SM 35, SM 36, SM 48), sikiⁿzi ‘office manager’ < EMC tɕik dʑɨʰ (職事) (SM 24), siNdaiteniwau ‘Four Great Heavenly Kings’ < EMC siʰ daj’ thɛn wuaŋʰ (四大天王) (SM 19, SM 43), ⁿziyaukai ‘Buddhist precepts’ < EMC dziajŋʰ kɛ:jʰ (浄戒) (SM 28), siyari ‘Buddha’s relics’ < EMC ɕiaʰ liʰ (舎利) 9  The following list is based on the list compiled by Shirafuji (1987: 204–205) on the basis of the index to Senmyō by Kitagawa Kazuhide (1982) with a few additions based on Kitagawa’s index that for some reason were not included into Shirafuji’s list. I also added Early Middle Chinese sources. We know virtually nothing about the phonology of Chinese loanwords in Western Old Japanese prose, because their kana spellings are sadly lacking, therefore the transcription below is just a speculation. I have added the vowels -i or -u after all final consonants in the tentative transcription of the examples above: judging by some relic loanwords surviving into later Japanese, it is quite possible that this was the case, i.e., e.g., the pronunciation of character 人 ‘person’ was [ⁿzini] with a final echo vowel. It is possible, however, that members of the educated class were able to pronounce the Chinese syllable final -p, -t, -k, -m, -n, and -N. However, none of these hypotheses can be really proven.

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

LEXICON

73

(SM 41), ⁿziyunisonu ‘obedient grandchild’ < EMC ʑwinʰ swǝn (順孫) (SM 42), siyouraku ‘bliss’ < EMC ɕiŋ lak (勝楽) (SM 45), siyosiyau ‘all sages’ < EMC tɕɨǎ ɕiajŋʰ (諸聖) (SM 44, SM 45), siyoteni ‘all Heavens’ < EMC tɕɨǎ thɛn (諸天) (SM 42), sinitai ‘advance and retreat’ < EMC tsinʰ thwǝjʰ (進退) (SM 62), ⁿzuwisiyo ‘auspicious writing’ < EMC dʑwiʰ ɕɨǝ̌ (瑞書) (SM 42), sekeni ‘world’ < EMC ɕiajʰ kɛ:nʰ (世間) (SM 45), setimbu ‘faithful wife’ < EMC tsɛt buw’ (節婦) (SM 42), ⁿzenisi ‘meditation teacher’ < EMC dʑian ʂi (禅師) (SM 28, SM 41), ⁿdaiŋgiyaku ‘great crime’ < EMC dajʰ ŋiajk (大逆) (SM 54), ⁿdaisi ‘great teacher’ < EMC dajʰ ʂi (大師) (SM 26, SM 41), taisiyaku ‘Taisaku (= Indra)’ < EMC tɛjʰ ɕiajk (帝釈) (SM 19, SM 43), ⁿdaiⁿzini ‘minister’ < EMC dajʰ dʑin (大臣) (SM 28, SM 35, SM 36), ⁿdaiNzuwi ‘great wonder’ < EMC dajʰ dʑwiʰ (大瑞) (SM 42, SM 48), ⁿdaipau ‘Minister of the Right (lit.: Great Protector)’ < EMC dajʰ paw’ (大保) (SM 25, SM 26), ⁿdaipopusi ‘great teacher of the Buddhist Law’ < EMC dajʰ puap ʂi (大法師) (SM 41, SM 42), ⁿdairitusi ‘teacher of Great Principles’ < EMC dajʰ lwit ʂi (大律師) (SM 41), tiŋgiyau ‘educated monk’ < EMC triʰ ɣɛ:jŋʰ (智行) (SM 61), tiniⁿziyu ‘garrison commander’ < EMC trinʰ ɕuwʰ (鎮守) (SM 62), ⁿdesi ‘disciple’ < EMC dɛj’ tsɨ’ (弟子) (SM 27), ⁿdokuⁿziyu ‘reading [sutras]’ < EMC dǝwk zuawŋ (讀誦) (SM 45), naisau ‘internal minister’ < EMC nwǝjʰ sɨaŋʰ (内相) (SM 19), niyorai ‘Tathagata’ < EMC ɲɨǝ̌ lǝj (如来) (SM 41), ninikeu ‘filial piety’ < EMC ɲin xɛ:wʰ (仁孝) (SM 59), ⁿziniteni ‘human and Heavenly worlds’ < EMC ɲin thɛn (人天) (SM 45), pakase ‘scholar’ < EMC pak dʑɨ’ (博士) (SM 11), pukasiŋgi ‘wonderful’ < EMC put kha’ siʰ ŋiʰ (不可思議) (SM 19, SM 43), peniⁿziyu ‘border guard’ < EMC pɛn swit (辺戍) (SM 62), mbosati ‘bodhisattva’ < EMC bɔ sat (菩薩) (SM 28, SM 38, SM 41), mboⁿdai ‘enlightenment’ < EMC bɔ dɛj (菩提) (SM 27), popusamuŋgi ‘councilor of the Buddhist Law’ < EMC puap sam ŋiʰ (法参議) (SM 41), popuⁿzini ‘minister of the Buddhist Law’ < EMC puap dʑin (法臣) (SM 41), popuwau ‘king of the Buddhist Law’ < EMC puap wuaŋʰ (法王) (SM 41), mbomuwau ‘Bonten (= Brahma)’ < EMC buamʰ wuaŋʰ (梵王) (SM 19, SM 43), muponu ‘rebellion’ < EMC muw puan’ (謀反) (SM 18, SM 34, SM 53, SM 54), muwi ‘[people] without a rank’ < EMC muǝ̌ wiʰ (无位) (SM 42), yaupuku ‘prosperity’ < EMC wiajŋ puwk (栄福) (SM 45), rai ‘ritual’ < EMC lɛj’ (礼) (SM 9), rikiⁿdeni ‘rice field lot’ < EMC lik dɛn (力田) (SM 13, SM 42), rusiyana ‘Vairocana Buddha’ < EMC lɔ ɕiaʰ naʰ (廬舎那) (SM 12, SM 19, SM 43), wiⁿzini ‘divine’ < EMC Ɂuj ʑin (威神) (SM 19, SM 43). It can be immediately seen that the majority of the Chinese loanwords in Western Old Japanese prose are also from the domain of Buddhist vocabulary, as we have already seen in the case of the poetic texts. Interesting enough, a considerable part of the Chinese loanwords attested in prose is heavily concentrated in edicts 41–43 of the Senmyō, and especially in edict 42.

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

CHAPTER 4 Nominals



Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

Contents of Chapter 4 Chapter 4 Nominals 83 1 Nouns 84 1.1 Prefixes 85 1.1.1 Honorific Prefix mi- 85 1.1.2 Intensive Prefix ma- 90 1.1.3 Diminutive Prefix wo- 95 1.1.4 Diminutive Prefix ko- 98 1.1.5 Locative Prefix sa- 101 1.1.5.1 Exception (1) 105 1.1.5.2 Exception (2) 105 1.1.5.3 Exception (3) 106 1.1.5.4 Exception (4) 108 1.2 Suffixes 109 1.2.1 Number and Plural Markers 110 1.2.1.1 Plural Marker -ra 111 1.2.1.2 Plural Marker -ⁿdǝmǝ 115 1.2.1.3 Plural Marker -tati 117 1.2.1.4 Plural Marker -na 119 1.2.1.5 Reduplication 123 1.2.2 Case Markers 125 1.2.2.1 Case Marker -i 126 1.2.2.2 Case Marker -ŋga 131 1.2.2.2.1 Possessive Marker 132 1.2.2.2.2 Subject Marker in a Dependent Clause 133 1.2.2.2.3 Subject Marker in a Main Clause 135 1.2.2.3 Genitive Case Marker -nǝ 139 1.2.2.3.1 Genetive Case Marker 140 1.2.2.3.2 Special Compressed Genitive Form 141 1.2.2.3.3 Subject Marker in a Dependent Clause 142 1.2.2.3.4 Subject Marker in a Main Clause 143 1.2.2.4 Dative-Locative Case Marker -ni 145 1.2.2.4.1 Dative Case Marker 145 1.2.2.4.2 Locative Case Marker 147 1.2.2.4.3 Special Compressed Locative Form 149 1.2.2.4.4 Special Reduced Locative Form 149 1.2.2.4.5 Agent Marker in Passive Constructions 150

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

78

Contents of Chapter 4

1.2.2.4.6 Directive Case Marker 151 1.2.2.4.7 Special Usage 152 1.2.2.5 Dative-Locative Case Marker -ra 158 1.2.2.6 Locative Case Marker -na 162 1.2.2.7 Genitive-Locative Case Marker -tu 163 1.2.2.8 Accusative-Absolutive Case Marker -wo 168 1.2.2.8.1 Accusative Case Marker 168 1.2.2.8.2 Emphatic Accusative -womba 170 1.2.2.8.3 Absolutive Case Marker 172 1.2.2.8.4 Peculiarities of Usage 175 1.2.2.8.5 Special Usage 176 1.2.2.9 Comitative Case Marker -tǝ 180 1.2.2.10 Ablative Case Marker -yu/-yuri/-yo/-yori 183 1.2.2.10.1 Allomorph -yu 184 1.2.2.10.2 Allomorph -yuri 185 1.2.2.10.3 Allomorph -yo 186 1.2.2.10.4 Allomorph -yori 187 1.2.2.11 Directive Case Marker -ŋgari 192 1.2.2.12 Directive Case Marker -pe 194 1.2.2.13 Terminative Case Marker -maⁿde 196 1.2.2.14 Comparative Case Marker -nǝ/-ni/-nasu/-nǝsu 200 1.2.2.14.1 Allomorph -nǝ 201 1.2.2.14.2 Allomorph -ni 202 1.2.2.14.3 Allomorph -nasu 203 1.2.2.14.4 Allomorph -nǝsu 204 1.2.2.14.5 Special Form -ⁿ 205 1.2.2.15 Origins of the New Ablative Case Marker -kara 207 1.2.3 Diminutive Suffixes 208 1.2.3.1 Diminutive Suffix -ra 208 1.2.3.2 Diminutive Suffix -rǝ 209 1.2.3.3 Diminutive Suffix -na 211 1.2.3.4 Diminutive Suffix -ko 212 1.2.3.4.1 Diminutive Meaning 212 1.2.3.4.2 Endearment Meaning 213 2 Pronouns 215 2.1 Morphological Peculiarities of Pronouns 215 2.2 Personal Pronouns 215

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

Contents of Chapter 4

2.2.1

2.2.2

2.2.3 2.2.4 2.2.5

2.2.6 2.2.7 2.2.8

79

First Person Pronoun wa/ware 218 2.2.1.1 Stem wa- 219 2.2.1.1.1 Oblique Stem wa-ⁿ- in a Compound 220 2.2.1.1.2 Possessive wa-ŋga 220 2.2.1.1.3 Accusative wa-wo 222 2.2.1.1.4 Directive wa-ŋgari 222 2.2.1.2 Stem ware 222 2.2.1.2.1 Isolated Form ware 222 2.2.1.2.2 Dative ware-ni 223 2.2.1.2.3 Accusative ware-wo 223 2.2.1.2.4 Ablative ware-yori 224 First Person Pronoun a/are 229 2.2.2.1 Stem a 230 2.2.2.1.1 Isolated Form a 230 2.2.2.1.2 Oblique Stem a-ⁿ- in a Compound 231 2.2.2.1.3 Possessive a-ŋga 231 2.2.2.1.4 Accusative a-wo 232 2.2.2.2 Stem are 232 2.2.2.2.1 Isolated Form are 232 2.2.2.2.2 Dative are-ni 233 2.2.2.2.3 Accusative are-wo[mba] 233 First Person Pronoun marǝ 236 First Person Pronoun na 237 Second Person Pronoun na/nare 238 2.2.5.1 Stem na 240 2.2.5.1.1 Isolated Form na 240 2.2.5.1.2 Oblique Stem na-ⁿ- in a Compound 240 2.2.5.1.3 Possessive na-ŋga 240 2.2.5.1.4 Accusative na-wo 242 2.2.5.1.5 Special Form na-ne 242 2.2.5.2 Stem nare 242 2.2.5.2.1 Isolated Form nare 243 2.2.5.2.2 Accusative nare-wo 243 Second Person Pronoun masi/mimasi/imasi 246 Second Person Pronoun ore 248 Third Person Pronoun si 250 2.2.8.1 Special Forms 258

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

80

Contents of Chapter 4

2.3 Reflexive Pronoun onǝ/onǝre 259 2.3.1 Stem onǝ 260 2.3.1.1 Isolated Form onǝ 260 2.3.1.2 Possessive onǝ-ŋga 260 2.3.2 Stem onǝre 261 2.3.2.1 Isolated Form onǝre 261 2.4 Demonstrative Pronouns 261 2.4.1 Proximal Demonstrative Pronouns 262 2.4.1.1 Proximal Demonstrative Pronoun kǝ/kǝre 262 2.4.1.1.1 Stem kǝ 263 2.4.1.1.1.1 Isolated Form kǝ 263 2.4.1.1.1.2 Modifier Form kǝnǝ 264 2.4.1.1.1.3 Emphatic Accusative kǝ-womba 265 2.4.1.1.1.4 Ablative kǝ-yo and kǝ-yu 265 2.4.1.1.2 Stem kǝre 265 2.4.1.1.2.1 Isolated Form kǝre 265 2.4.1.1.2.2 Modifier Form kore n-ǝ 266 2.4.1.1.2.3 Accusative kǝre-wo 266 2.4.1.2 Proximal Demonstrative Pronoun kǝkǝ 268 2.4.1.3 Proximal Demonstrative Pronoun kǝti 271 2.4.1.4 Proximal Demonstrative Pronoun kǝnata 272 2.4.2 Mesial Demonstrative Pronouns 273 2.4.2.1 Mesial Demonstrative Pronoun sǝ/sǝre 274 2.4.2.1.1 Stem sǝ 274 2.4.2.1.1.1 Accusative sǝ-wo 274 2.4.2.1.1.2 Modifier Form sǝnǝ 275 2.4.2.1.2 Stem sǝre 276 2.4.2.2 Mesial Demonstrative Pronoun sǝkǝ 277 2.4.3 Distal Demonstrative Pronouns 278 2.4.3.1 Distal Demonstrative Pronoun ka/kare 279 2.4.3.1.1 Stem ka 279 2.4.3.1.1.1 Isolated Form ka 279 2.4.3.1.2 Stem kare 280 2.4.3.2 Distal Demonstrative Pronoun kanata 281 2.4.3.3 Distal Demonstrative Pronoun woti/wotǝ/wote 282

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

81

Contents of Chapter 4

2.5 Interrogative Pronouns 284 2.5.1 Interrogative Pronoun ta-/tare 285 2.5.1.1 Stem ta 285 2.5.1.1.1 Possessive ta-ŋga 285 2.5.1.1.2 Dative ta-ni 286 2.5.1.2 Stem tare 286 2.5.1.2.1 Isolated Form tare 287 2.5.1.2.2 Dative tare-ni 288 2.5.1.2.3 Accusative tare-wo 288 2.5.2 Interrogative Pronoun nani 291 2.5.2.1 Special Form 295 2.5.3 Interrogative Pronoun ika 299 2.5.3.1 Uncontracted Form ika n-i ar- 301 2.5.4 Interrogative Pronoun iⁿdu/iⁿduku 304 2.5.5 Interrogative Pronoun iⁿduti 307 2.5.6 Interrogative Pronoun iⁿdure 308 2.5.7 Interrogative Pronoun itu 310 2.5.8 Interrogative Pronoun iku 313 2.5.8.1 Special Forms 314 2.5.9 Interrogative Pronoun naⁿdǝ, naⁿzǝ 316 2.6 Collective Pronouns 319 2.6.1 Collective Pronoun mïna 319 2.6.2 Collective Pronoun morǝ 321 2.6.2.1 Special Form 324 3 Numerals 327 3.1 Cardinal Numerals 328 3.1.1 Cardinal Numeral pitǝ ‘One’ 328 3.1.2 Cardinal Numeral puta ‘Two’ 330 3.1.3 Cardinal Numeral mi ‘Three’ 332 3.1.4 Cardinal Numeral yǝ ‘Four’ 334 3.1.5 Cardinal Numeral itu ‘Five’ 334 3.1.6 Cardinal Numeral mu ‘Six’ 336 3.1.7 Cardinal Numeral nana ‘Seven’ 336 3.1.8 Cardinal Numeral ya ‘Eight’ 337 3.1.9 Cardinal Numeral kǝkǝnǝ ‘Nine’ 339

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

82

Contents of Chapter 4

3.1.10 Cardinal Numeral tǝwo ‘Ten’ 340 3.1.11 Count above ‘Ten’ 341 3.1.12 Tens 342 3.1.13 Hundreds 346 3.1.14 Higher Numerals 350 3.2 Ordinal Numerals 352 3.3 Classifiers 354 3.3.1 Classifier -tu 355 3.3.2 Classifier -ti 357 3.3.3 Classifier -ri 358 3.3.4 Classifier -mǝtǝ 359 3.3.5 Classifier -pe 360 3.3.6 Classifier -ka 362 3.4 Months of the Year 367

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

As in Modern Japanese, the class of nominals in Old Japanese comprises three major subclasses: nouns, pronouns, and numerals. These three subclasses may be contrasted to verbs, a class of words, which denote action or condition and have formal markers distinguishing them from the non-verbal class. Also, verbal stems can never be used independently without suffixes with the exception of a small group of verbs that can have a suffix-less imperative form in Western Old Japanese (but not in any later stage of the language). On the other hand, nominal stems can be used independently, and nominals do not have formal markers, which further distinguishes them from verbs. Two other important characteristics of Old Japanese nominals include their relatively simple morphology and almost complete lack of fusion. On the first point, typically no more than two suffixal positions are possible in a nominal paradigm; usually a plural marker and case marker, if any, and only two prefixal positions are possible. Thus, the maximal nominal paradigm form will look like this: PREF-PREF-NOMINAL-SUF-SUF.1 Compared to this, the verbal paradigm with its two possible prefixal positions and theoretically unlimited number of suffixal positions is much more complex. In addition, the nominal paradigm exhibits pure agglutination, with no traces of fusion (with the possible exception of rendaku (sequential voicing), but even in the case of rendaku it is still possible to draw a morphological boundary). In contrast, verbal paradigms exhibit some cases of fusion where it becomes difficult or even impossible to draw a morphological boundary between the constituents of a paradigmatic form.

1  Theoretically, three suffixal positions should be possible, since there are plural markers, diminutive suffixes, and case markers. However, there are no examples where plural markers occur together with diminutive suffixes, so the actually attested paradigms are limited to PLUR-CASE and DIM-CASE. There is one example in Eastern Old Japanese, when two different diminutive suffixes are used in front of a case marker: imo-na-rǝ-ŋga ‘beloved-DIM-DIMPOSS’ (MYS 14.3446, see section 1.2.3.3 for details).

© Alexander Vovin, 2020 | doi:10.1163/9789004422810_005

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

Section 1

Nouns From the standpoint of word formation, all nouns can be divided into nouns and deverbal nouns. Nouns in Western Old Japanese as well as in Classical Japanese or Modern Japanese do not have a formal marker distinguishing them from other classes. Therefore, nouns are classified on the basis of two criteria: semantics and distribution. The distributional criterion means that nouns combine with certain morphemes, which are connected with grammatical categories typical only of nouns. Nouns in Old Japanese are characterized by the categories of case and number. There are two types of deverbal nouns: nouns derived from action verbs and nouns derived from quality verbs. The first type on the segmental level coincides with the converb (but on the suprasegmental level the converb and deverbal nouns have different accents) and has the formal marker -i, -ï or -ɛ. The second type always has the formal markers -sa or -mi, which are added directly to the stem of a quality verb. I will discuss all deverbal nouns in chapter 6, which is dedicated to the verb. Nouns as a class in Western Old Japanese have a formal feature that largely disappears from the later stages of the language. Some nouns ending in vowels /ï/ and /ɛ/ (or in /i/ and /e/ which had underlying *ï and *ɛ) in Western Old Japanese have two stems: a free stem, and a bound stem. Bound stems occur in compounds and for some of these nouns also tend to occur before genitive or possessive case marker, although free stems can also be used before these two case markers. Thus, e.g., the word kï ‘tree’ (free stem) has also a bound stem kǝ-; the possible genitive forms are both kǝ-nǝ and kï-nǝ. In Classical Japanese and later stages of the language bound stems have been preserved only in nonproductive compounds, many of which became synchronically indivisible from the morphological point of view. Not all OJ nouns ending in /ï/ or /ɛ/ have two stems, e.g., such words as patakɛ ‘dry field,’ potǝkɛ ‘Buddha,’ or mï ‘winnow’ always occur in one form only. Below is the list of nouns that have free and bound stems.

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

85

NOMINALS

free stem amɛ amɛ2 kɛ kɛ kamï kï kukï kuti ( ∅/__i, Ryukyuan -nakai/-Nkai perfectly corresponds to Old Japanese -ŋgari. As implied by the Ryukyuan uncontracted form -nakai, unless it is an innovation, the nasalization of Old Japanese -ŋgari and the -N- part of Ryukyuan -Nkai probably represents the locative case marker -na, attested in Eastern Old Japanese and various Ryukyuan languages (see section 1.2.2.6). That allows us to reconstruct the directive marker as a bimorphemic *-na-kari and leaves *-kari as a potential proto-Japonic directive marker.46 1.2.2.12 Directive Case Marker -pe The directive case marker -pe is a result of the grammaticalization of the noun pe ‘side.’ This noun is used frequently in both early and later Old Japanese texts as a quasi-postposition in combination with the following dative-locativedirective case marker -ni in the form pe-ni ‘to the side’ or ‘at the side.’ 淤岐弊迩波袁夫泥都羅羅玖

oki-pe-ni pa wom-bune turar-aku offing-side-LOC TOP DIM-boat be stretched in a line-NML small boats are stretched in a line in the offing (KK 52) 夜麻登弊迩爾斯布岐阿宜弖

Yamatǝ-pe-ni nisi puk-i-aŋgɛ-te Yamatǝ-side-LOC western.wind blow-CONV-raise(CONV)-SUB Western wind blows towards Yamatǝ, and … (KK 55) 46  This form *-na-kai, actually attested as -nakai in Ryukyuan, rules out Martin‘s etymology deriving this marker from muka[w]i < *muka-pa-Ci ‘facing’ (Martin 1990: 498). Another Martin’s suggestion that it may be from naka ni ‘at/to within’ (Martin 1990: 498) is also questionable, since in those Ryukyuan dialects, where dative-locative case marker -ni is present, and which also have locatives -nakai or -na such as Kumejima and Tokunoshima, -ni is reflected as -ni, and not as *-i.

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

195

NOMINALS 夜麻登弊迩由玖波多賀都麻

Yamatǝ-pe-ni yuk-u pa ta-ŋga tuma Yamatǝ-side-LOC go-ATTR TOP who-POSS spouse Whose spouse goes towards Yamatǝ? (KK 56) 山辺爾草乎思香奈久毛

YAMA-PE-ni sa-wǝ-sika nak-umo mountain-side-LOC PREF-male-deer cry-EXCL a male deer cries at the mountain side! (MYS 15.3674) There are, however, a number of examples where -pe appears to be completely grammaticalized as a directive case marker: 摩佐豆古和藝毛玖迩弊玖陀良須

Masaⁿduko wa-ŋg-imo kuni-pe kuⁿdar-as-u Masaduko I-POSS-beloved province-DIR descend-HON-FIN Masaduko, my beloved, goes towards [her] province (KK 52) 阿米弊由迦婆

amɛ-pe yuk-amba heaven-DIR go-COND if [you] go to heaven (MYS 5.800) 宮弊能保留等

MIYA-pe nǝmbor-u tǝ palace-DIR ascend-FIN DV thinking of going to the palace (MYS 5.886) 新羅奇敞可伊敞爾可加反流

Sirakï-pe ka ipe-ni ka kaper-u Silla-DIR IP home-LOC IP return-ATTR [Will I go] to Silla, or will [I] return home? (MYS 15.3696) 和我勢古我久爾敞麻之奈婆

wa-ŋga se-ko-ŋga kuni-pe [i]mas-i-n-amba I-POSS beloved-DIM-POSS land-DIR come(HON)-CONV-PERF-COND If my beloved would have returned to [his] land (MYS 17.3996) It is interesting that -pe as a directive case marker in Western Old Japanese is used only with an animate agent, who moves by himself/herself to some point.

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

196

Chapter 4

This limitation on -pe usage was suggested by Syromiatnikov for Classical Japanese (Syromiatnikov 1983: 65), where it is not true (Vovin 2003: 75–76), but apparently it seems to be true for Old Japanese. Comparative Data Level A: Other Japonic A1: Eastern Old Japanese There is one example in Eastern Old Japanese with -pe as a grammaticalized directive case marker: 和我世古乎夜麻登敞夜利弖

wa-ŋga se-ko-wo Yamatǝ-pe yar-i-te I-POSS beloved-DIM-ACC Yamatǝ-DIR send-CONV-SUB Having sent my beloved to Yamatǝ (MYS 14.3363) 1.2.2.13 Terminative Case Marker -maⁿde The terminative case marker -maⁿde indicates the limit of a movement or another action in time or space, or in other words, the point beyond which the action does not reach. It must be noted that -maⁿde can be followed by the locative case marker -ni, with no apparent change in meaning. This usage predominantly occurs with -maⁿde-ni used after a nominalized clause (see below), although there are two examples of the same usage after a noun. That differs from the modern Japanese double case marker -made-ni, which expresses a terminative-directive meaning ‘by,’ as in getsuyōbi-made-ni ‘by Monday.’ Examples: 美夜故摩提意久利摩遠志弖

miyako-maⁿde okur-i-mawos-i-te capital-TERM see.off-CONV-HUM-CONV-SUB [I would] see [you] off to the capital, and … (MYS 5.876) 寝屋度麻弖來立呼比奴

NE-YA-ⁿ-do-maⁿde K-i-TAT-I YOmB-Ap-i-n-u sleep(CONV)-house-GEN-door-TERM come-CONV-stand-CONV-call-ITER -CONV-PERF-FIN [the village headman] came out up to our bedroom door and called [us] repeatedly (MYS 5.892)

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

197

NOMINALS 伊豆礼能日麻弖安礼古非乎良牟

iⁿdure n-ǝ PI-maⁿde are kopï wor-am-u which DV-ATTR day-TERM I long.for(CONV) exist-TENT-FIN until what day should I be longing for [you]? (MYS 15.3742) 伊都麻弖可安我故非乎良牟

itu-maⁿde ka a-ŋga kopï-wor-am-u when-TERM IP I-POSS long.for(CONV)-exist-TENT-ATTR until when should I long for [you]? (MYS 15.3749) 安里我欲比都加倍麻都良武万代麻弖爾

ari-ŋgayop-i tukapɛ-matur-am-u YƏRƏⁿDU YƏ-maⁿde-ni ITER-go back and forth-CONV serve(CONV)-HUM-TENT-FIN ten thousand generation-TERM-LOC constantly going back and forth, [we] will serve the [emperor] until ten thousand generations (MYS 17.3907) 之路髪麻泥爾大皇爾都可倍麻都礼婆

siro KAMI-maⁿde-ni OPO-KIMI-ni tukapɛ-matur-e-mba white hair-TERM-LOC great-lord-DAT serve(CONV)-HUM-EV-CON when [one] serves the emperor until [one gets] white hair (MYS 17.3922) Similar to the case markers -wo, -ni and -yuri/-yo/-yori, -maⁿde can also mark a nominalized clause, meaning ‘until.’ It is interesting that in the oldest Old Japanese texts the terminative case marker -maⁿde occurs only after a nominalized sentence, and not after a noun, like in the following two examples from the Nihonshoki kayō and the Bussoku seki no uta. Besides, even in the later texts from the Man’yōshū, its usage after a nominalized clause occurs much more frequently than after a noun (with a ratio almost 3:1). Yamada Yoshio suggested that -maⁿde might be originally an adverb (Yamada 1954: 474); however the more frequent and ancient usage after nominalized clauses rather points to the fact that -maⁿde is originally some kind of a bound noun. Its etymology is obscure. Examples after a nominalized clause:

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

198

Chapter 4

志我都矩屢麻泥爾飫裒枳瀰爾柯柂倶都柯陪麻都羅武

si-ŋga tukur-u47-maⁿde-ni opo kimi-ni kata-ku tukapɛ-matur-am-u they-POSS come.to.an.end-ATTR-TERM-LOC great lord-DAT strong-CONV serve(CONV)-HUM-TENT-FIN Until they come to an end, [I] intend to serve faithfully to the emperor (NK 78) 阿止乎美都都志乃波牟多太爾阿布麻弖爾麻佐爾阿布麻弖爾

atǝ-wo mi-tutu sinǝp-am-u taⁿda n-i ap-u-maⁿde-ni masa n-i ap-u-maⁿde-ni footstep-ACC see(CONV)-COOR yearn-TENT-FIN direct DV-CONV meet-ATTRTERM-LOC real DV-CONV meet-ATTR-TERM-LOC looking at [Buddha’s] footprints, [I] will yearn [for him], until [I] meet [him] directly, until [I] really meet [him] (BS 6) 布流由岐得比得能美流麻提烏梅能波奈知流

pur-u yuki tǝ pitǝ-nǝ mi-ru-maⁿde uMƐ-nǝ pana tir-u fall-ATTR snow DV person-GEN see-ATTR-TERM plum-GEN flower fall-FIN plum blossoms fall to such extent that people will perceive them as falling snow (MYS 5.839) 伊波敞和我勢古多太爾安布末低爾

ipap-e wa-ŋga se-ko taⁿda n-i ap-u-maⁿde-ni pray-IMP I-POSS beloved-DIM direct DV-CONV meet-ATTR-TERM-LOC pray, my beloved, until [we] meet directly (MYS 15.3778) 朕高御座爾坐始由理今年尓至麻低六年尓成奴

WARE TAKA MI-KURA-ni IMAS-I-SƏMƐS-U-yuri KƏ TƏSI-ni ITAR-U-maⁿde MU TƏSI n-i NAR-I-n-u I high HON-seat-LOC be(HON)-CONV-begin-ATTR-ABL this year-LOC reachATTR-TERM six year DV-CONV become-CONV-PERF-FIN It has been six years (reaching up to) this year since I have been on the high throne (SM 7)

47  The verb tukur- is an intransitive counterpart of tukus- ‘to exhaust, to exert (oneself).’ It occurs only in the Old Japanese, and it is interesting that it did not make its way into any dictionaries, including (Omodaka et al. 1967). The much more widespread intransitive equivalent of tukus- is OJ tukï- w/___a is unknown both in Old Japanese in particular and in the history of Japanese in general. Moreover, a comparison with cognates from Ryukyuan demonstrates that the proto-Japonic form of OJ wa- was *wa/*wa-n, which would make the fortition explanation even more difficult.57 Certainly, one can claim a prothesis of [w-] in front of /a/, but again no other cases of such a prothesis are known. This, together with the aforementioned tendency of replacing wa- with a- in Old Japanese, should point to the fact that a- forms are secondary as compared to wa- forms, and are due to the general process of /w/ lenition in Japanese. This lenition explanation, although more plausible than the fortition one, faces two problems. First, there are no other examples of w > ∅/___a attested in Old Japanese, although there is one more in Middle Japanese: wamek-/amek- ‘to cry, to shout,’ so the change is irregular. 57  The South Ryukyuan *b- corresponding to OJ w-, which is sometimes cited as evidence for PJ *b- and not *w-, is a result of a fortition. The evidence for this point of view is largely based on several Chinese loanwords in South Ryukyuan, such as, e.g., Miyako boo ‘king’ (used as child’s name) < MJ wau < Chinese wang (王) ‘id.’ as well as on distribution patterns (see Karimata 1999: 55–56 for details).

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

218

Chapter 4

Second, on the basis of the above statistics in Chart 9 and the fact that a- forms did not survive into Middle Japanese, including the earliest Classical Japanese texts, and have much more limited Ryukyuan attestations than wa- forms, it looks like that the a- forms were short-lived, probably originating in protoJaponic, but quickly fading away after Old Japanese in mainland Japanese and surviving after Old Ryukyuan only in few Ryukyuan dialects. All these facts bring reasonable doubt in traditional belief that both wa- and a- forms represent genetically related phonetic variants of the same pronoun, in spite of their significant morphological similarity: both have an extended stem in -re, and it appears that both have an oblique stem marker -ⁿ-, although in the case of a- forms there is no enough evidence due to the lack of support from modern Ryukyuan languages, which is crucial. Yoshizō Itabashi was the first linguist who proposed to view wa- and a- forms as etymologically different (Itabashi 1999). Although I do not personally agree with his Austronesian etymology for a- forms (see section 2.2.2 below), I believe that his general proposal to separate the two forms genetically is probably a step in a right direction. The crucial evidence that leads to the understanding of the difference between a- and wa- as well as to the reconstruction of the PJ situation comes from some South Ryukyuan languages, where a- is always singular ‘I’ and wa- is always plural ‘we.’ Thus, in this respect South Ryukyuan is more archaic than Old Japanese or any other variety of Japonic. 2.2.1 First Person Pronoun wa/ware The first person pronoun wa/ware is a neutral pronoun from a sociolinguistic point of view, which can be used without any regards to the gender or social status. By the beginning of the Heian period, the extended stem wa-re with the obsolete suffix -re was used in the nominative case and all oblique cases except the genitive, where the unextended stem wa- was used. In Old Japanese the situation was different: there are examples when the unextended stem is used independently or with the following dative, directive and accusative case markers, thus we have: wa ‘I,’ wa-ni ‘I-DAT,’ wa-ŋgari ‘I-DIR’ and wa-wo ‘I-ACC,’ alongside with more common ware ‘I,’ ware-ni ‘I-DAT,’ and ware-wo ‘I-ACC.’ There is no directive form *ware-ŋgari, however. All forms in isolation and dative forms are attested only in Eastern Old Japanese, as well as the majority of examples of accusative forms. In one example (KK 3) unextended stem wa- appears in the compound wa-ⁿ-dǝri ‘my bird’ as if it were followed by the compressed genitive -ⁿ-. Since this is one of two extant examples (the other one is found after second person pronoun na-) and since we know that personal pronouns combine with possessive case marker -ŋga and not genitive case marker -nǝ, it is safer to assume that this -ⁿ- is a marker of the oblique

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

219

NOMINALS

stem of a pronoun. This proposal will be further supported below by Eastern Old Japanese and Ryukyuan data, where the oblique stem of the first person pronoun appears as wanu, waN, banu (see the example with wanu ‘I’ from MYS 14.3476 below).58 While modern Japanese strictly differentiates the number of personal pronouns: thus, watakusi means only ‘I,’ but watakusi-tati means only ‘we,’ it is usually considered that OJ wa-/ware can mean both. However, the unextended stem wa- occurs in the plural meaning in Western Old Japanese only in the possessive form wa-ŋga: otherwise it is always singular. In all other cases plural meaning is associated with the extended stem ware only, which can be used in both singular and plural meanings. This leads to a hypothesis that extending suffix -re goes back to a plural suffix -ra plus active case marker -i: wa-re < *wa-ra-i. By the time when the first Old Japanese texts were recorded the distinction between singular wa- and plural wa-re was lost in most cases, but the former singular stem has managed to retain its original meaning in isolation and before all case markers except the possessive -ŋga. The following chart summarizes the distribution of both unextended stem wa and extended stem ware in isolation and with the following case markers in Western and Eastern Old Japanese. chart 10

Isolation Possessive Dative Accusative Directive Ablative

First person pronoun wa/ware in combination with case markers

Unextended stem wa-

Extended stem ware

wa* S wa-ŋga SP wa-ni* S, wanu-ni* S wa-wo S wa-ŋgari S —

ware SP — ware-ni S ware-wo** SP — ware-yori** S

Note: S – singular, P – plural, * – Eastern Old Japanese only, ** – Western Old Japanese only

2.2.1.1 Stem waThe unextended stem wa- in Western Old Japanese never occurs without following case markers, except in one case in a compound (see below). In most cases it is followed by the possessive case marker -ŋga, although accusative wa-wo and directive wa-ŋgari are also attested. 58  Martin reconstructs PJ *ba[nu] ‘I’ (Martin 1987: 567).

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

220 2.2.1.1.1

Chapter 4

Oblique Stem wa-ⁿ- in a Compound

伊麻許曾婆和杼理迩阿良米能知波那杼理爾阿良牟遠

ima kǝsǝ pa wa-ⁿ-dǝri n-i ar-am-ɛ nǝti pa na-ⁿ-dǝri n-i ar-am-u-wo now FP TOP I-OSM-bird DV-CONV exist-TENT-EV after TOP you-OSM-bird DV-CONV exist-TENT-ATTR-ACC Now [I] am my bird, later [I] will be your bird, so … (KK 3) 2.2.1.1.2 Possessive wa-ŋga 淤曾夫良比和何多多勢禮婆比許豆良比和何多多勢禮婆

osǝ-m-bur-ap-i wa-ŋga tat-as-er-e-mba pikǝ-ⁿ-dur-ap-i wa-ŋga tat-as-er-e-mba push-DV(CONV)-shake-ITER-CONV I-POSS stand-HON-PROG-EV-CON pullDV(CONV)-shove-ITER-CONV I-POSS stand-HON-PROG-EV-CON [I] was pushing and shaking [the door], when I was standing [there], [I] was pulling and shoving [it], when I was standing [there] (KK 2) 和何許許呂

wa-ŋga kǝkǝrǝ I-POSS heart my heart (KK 3) 志藝和那波留和賀麻都夜

siŋgi-wana par-u wa-ŋga mat-u ya snipe-trap stretch-FIN we-POSS wait-FIN EP [we] set a snipe trap. We waited! (KK 9) 許能美岐波和賀美岐那良受

kǝnǝ mi-ki pa wa-ŋga mi-ki nar-aⁿz-u this HON-rice.wine TOP I-POSS HON-rice.wine be-NEG-FIN This rice wine is not my rice wine (KK 39) 和賀由久美知

wa-ŋga yuk-u miti we-POSS go-ATTR way the way we go (KK 43) 和我惠比爾祁牟

wa-ŋga wep-i-n-i-k-em-u I-POSS get intoxicated-CONV-PERF-CONV-PAST/FIN-TENT-FIN I got intoxicated (FK 6) Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

221

NOMINALS 和何世古我多那礼乃美巨騰都地爾意加米移母

wa-ŋga se-ko-ŋga ta-nare n-ǝ mi-kǝtǝ tuti-ni ok-am-ɛ ya mǝ I-POSS beloved-DIM-POSS hand-accustom(NML) DV-ATTR HON-koto groundLOC put-TENT-EV IP EP would anyone dare to put the favorite koto of my beloved on the ground?! (MYS 5.812) 伊波敞和我勢古多太爾安布末低爾

ipap-e wa-ŋga se-ko taⁿda n-i ap-u-maⁿde-ni pray-IMP I-POSS beloved-DIM direct DV-CONV meet-ATTR-TERM-LOC pray, my beloved, until [we] meet directly (MYS 15.3778) The possessive form wa-ŋga can occasionally lose the vowel /a/ in the possessive marker -ŋga in front of the words imo ‘beloved,’ and ipe ‘home,’ producing close-to-compounds forms wa-ŋg-imo ‘my beloved’ and wa-ŋg-ipe ‘my home.’59 Examples: 摩佐豆古和藝毛玖迩弊玖陀良須

Masaⁿduko wa-ŋg-imo kuni-pe kuⁿdar-as-u Masaⁿduko I-POSS-beloved province-DIR descend-HON-FIN Masaⁿduko, my beloved, goes toward [her] province (KK 52) 和藝幣能伽多由区毛位多知区暮

wa-ŋg-ipe-nǝ kata-yu kumowi tat-i-k-umo I-POSS-home-GEN side-ABL cloud rise-CONV-come-EXCL Clouds rise from the side of my home (NK 21) The possessive form wa-ŋga can assimilate the vowel /a/ to /ǝ/ in the possessive marker -ŋga in front of the compound opo kimi ‘emperor’ (lit.: ‘great lord’), becoming wa-ŋgǝ:60

59  The uncontracted forms wa-ŋga imo ‘my beloved’ and wa-ŋga ipe ‘my home’ are also plentifully attested in Old Japanese texts. Interesting enough, when the word imo ‘beloved’ is followed by the diminutive suffix -ko, there is just one example of uncontracted wa-ŋga imo-ko, attested in Eastern Old Japanese (MYS 20.4405); in all other known attestations there is always contracted wa-ŋg-imo-ko. 60  Unassimilated wa-ŋga opo kimi ‘my emperor’ is also attested in Old Japanese texts, although wa-ŋgǝ opo kimi seems to be more frequent.

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

222

Chapter 4

和期於保伎美余思努乃美夜乎安里我欲比賣須

wa-ŋgǝ opo kimi Yǝsino-nǝ miya-wo ari-ŋgayop-i mes-u I-POSS great lord Yǝsino-GEN palace-ACC ITER-go back and forth-CONV look(HON)-FIN My emperor constantly visits the palace in Yǝsino, and looks around (MYS 18.4099) 2.2.1.1.3 Accusative wa-wo 和乎待難尓

wa-wo MAT-I-kate-n-i I-ACC wait-CONV-POT-NEG-CONV [she] is unable to wait for me, and … (MYS 11.2483) 何爲牟尓吾乎召良米夜

NANI SE-m-u-ni WA-wo MES-Uram-ɛ ya what do-TENT-ATTR-LOC I-ACC summon-TENT2-EV IP should you have summoned me in order to do something? (MYS 16.3886) 2.2.1.1.4 Directive wa-ŋgari 今夜可君之我許來益武

KƏ YƏPI ka KIMI-ŋGA WA-ŋgari K-i-[i]mas-am-u this night IP lord-POSS I-DIR come-CONV-HON-TENT-ATTR Is it tonight that [my] lord will come to me? (MYS 8.1519) 2.2.1.2 Stem ware The extended stem ware in Western Old Japanese occurs in isolation, and also followed by the dative case marker -ni, accusative case marker -wo, and ablative case marker -yori. 2.2.1.2.1

Isolated Form ware

和禮波和須禮士

ware pa wasure-ⁿzi I TOP forget-NEG/TENT I will not forget (KK 12) 和禮波夜惠奴

ware pa ya we-n-u we TOP EP starve-PERF-FIN we are starving (KK 14) Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

223

NOMINALS 迦美斯美岐迩和禮惠比迩祁理

kam-i-si mi-ki-ni ware wep-i-n-i-ker-i brew-CONV-PAST/ATTR HON-rice.wine-LOC I be.drunk-CONV-PERF-CONVRETR-FIN I got drunk with brewed holy rice wine (KK 49) 和礼由久

ware yuk-u I go-FIN I will go (MYS 15.3706) 2.2.1.2.2 Dative ware-ni 伊慕我堤鳴倭例爾魔柯斯毎

imo-ŋga te-wo ware-ni mak-asimɛ beloved-POSS hand-ACC I-DAT use as a pillow-CAUS(CONV) [my] beloved will let me use [her] hands as a pillow (NK 96) 和礼爾於止礼留比止乎於保美

ware-ni otǝr-er-u pitǝ-wo opo-mi I-DAT be worse-PROG-ATTR person-ABS many-GER because there are many people who are worse than me (BS 13) 比故保思母和礼爾麻佐里弖於毛布良米也母

Pikoposi mǝ ware-ni masar-i-te omop-uram-ɛ ya mǝ Altair FP I-DAT surpass-CONV-SUB long.for-TENT2-EV IP EP Will Altair long for [his beloved] more than I [do]?! (lit.: surpassing me) (MYS 15.3657) 2.2.1.2.3 Accusative ware-wo 和例烏斗波輸儺

ware-wo top-as-u na I-ACC ask-HON-FIN EP [You] asked me! (NK 63) 烏麼野始爾倭例烏比岐例底

wom-bayasi-ni ware-wo pik-i-[i]re-te DIM-wood-LOC I-ACC pull-CONV-insert(CONV)-SUB Taking me into the small wood … (NK 111)

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

224

Chapter 4

伊敞妣等乃伊豆良等和礼乎等波婆伊可爾伊波牟

ipe-m-bitǝ-nǝ iⁿdu-ra tǝ ware-wo tǝp-amba ika n-i ip-am-u home-GEN-person-GEN where-LOC DV I-ACC ask-COND how DV-CONV say-TENT-FIN If people from [your] home ask me where [are you], how should [I] answer? (MYS 15.3689) 和礼乎於吉弖比等波安良自

ware-wo ok-i-te pitǝ pa ar-aⁿzi we-ACC leave-CONV-SUB person TOP exist-NEG/TENT except us, there would be no [other] men (MYS 18.4094) 2.2.1.2.4 Ablative ware-yori 和礼欲利母貧人乃父母波

ware-yori mǝ MAⁿDUSI-KI PITƏ-nǝ TITI PAPA pa I-ABL FP poor-ATTR person-GEN father mother TOP fathers and mothers of people poorer than me (MYS 5.892) Comparative Data Level A: Other Japonic A1: Eastern Old Japanese As was mentioned above, Eastern Old Japanese has more combinations of the unextended stem wa with case markers than Western Old Japanese, and it also can be used in isolation. The plural meaning of the first person pronoun wa/ware is not attested in Eastern Old Japanese. Stem wa

The unextended stem wa is used very frequently in Eastern Old Japanese: it occurs eleven times in isolation, two times with the dative case marker -ni (one more time the dative form wanu-ni based on the oblique stem wanu-), four times with accusative marker -wo, and one time with directive marker -ŋgari. Isolated Form wa 和可加敞流弖能毛美都麻弖宿毛等和波毛布

waka kaperute-nǝ momit-u-maⁿde NE-m-o tǝ wa pa [o]mop-u young maple-GEN leaves.turn.red/yellow-ATTR-TERM sleep-TENT-ATTR DV I TOP think-FIN I think that [we] should sleep [together] until the young maple becomes red (MYS 14.3494) Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

225

NOMINALS 己許呂能未伊母我理夜里弖和波己許爾思天

kǝkǝrǝ nǝmï imo-ŋgari yar-i-te wa pa kǝkǝ-ni s-i-te heart RP beloved-DIR send-CONV-SUB I TOP here-LOC do-CONV-SUB I [have] to stay here, sending just [my] heart to [my] beloved (MYS 14.3538) Possessive wa- ŋga 和我世故乎安杼可母伊波武

wa-ŋga se-ko-wo aⁿ-dǝ kamǝ ip-am-u I-POSS beloved-DIM-ACC what-DV EP say-TENT-ATTR What shall [I] say about my beloved, I wonder? (MYS 14.3379) Dative wa-ni, wanu-ni

和爾奈多要曾祢

wa-ni na-taye-sǝ-n-e I-DAT NEG-break-do-DES-IMP Do not become estranged from me (MYS 14.3378) 和爾余曾利

wa-ni yǝsǝr-i I-DAT become intimate-CONV becoming intimate with me (MYS 14.3408) 和奴爾故布奈毛

wanu-ni kop-unam-o I-DAT long.for-TENT2-ATTR [you] will probably long for me (MYS 14.3476) In this example wanu, in all probability, represents in the Eastern Old Japanese a trace of the first person pronoun oblique stem *wa-n[u]-, which is well attested in Ryukyuan. Accusative wa-wo 和乎布利弥由母阿是古志麻波母

wa-wo pur-i-mi-y-umǝ Aⁿze ko si map-am-ǝ I-ACC swing-CONV-look-PASS-EXCL Aⁿze girl EP dance-TENT-ATTR the girl from Aⁿze is going to dance, suddenly looking back at me! (FK 7) This song from Hitachi province has other typical Eastern Old Japanese features, such as attributive in -o, so I consider it to be an Azuma text. Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

226

Chapter 4

和乎可麻都那毛伎曾毛己余必母

wa-wo ka mat-unam-wo kisǝ mo kǝ yǝpi mǝ I-ACC IP wait-TENT2-ATTR last night FP this night FP will [she] have waited for me, both last night and tonight? (MYS 14.3563) Directive wa- ŋgari 伊豆由可母加奈之伎世呂我和賀利可欲波牟

iⁿdu-yu kamǝ kanasi-ki se-rǝ-ŋga wa-ŋgari kayop-am-u where-ABL EP beloved-ATTR husband-DIM-POSS I-DIR visit-TENT-ATTR where will my beloved husband visit me from, I wonder? (MYS 14.3549) Stem ware

The extended stem ware is rather rare and has limited usage in Eastern Old Japanese compared to the unextended stem wa: it occurs only two times in isolation and once with the following dative-locative case marker -ni in MYS 14 (all examples are below). The accusative form ware-wo is not attested. There is an isolated example of warǝ ‘I,’ probably a phonetic variant of ware in MYS 20.4343. In addition, there are two attestations in MYS 20.4344 and MYS 20.4348. Isolated Form ware 久君美良和礼都賣杼

kuku-mira ware tum-e-ⁿdǝ stalk-leek I pick up-EV-CONC Although I pick stalk-leeks … (MYS 14.3444) 和礼左倍爾伎美爾都吉奈那

ware sapɛ n-i kimi-ni tuk-i-n-ana I even DV-CONV lord-DAT cling-CONV-PERF-DES Even I want to cling to [my] lord (MYS 14.3514) Variant warƏ 和呂多比波多比等於米保等

warǝ tambi pa tambi tǝ omɛp-o-ⁿdǝ I journey TOP journey DV think-EV-CONC Although I think that [my] journey is [just] a journey … (MYS 20.4343)

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

227

NOMINALS Dative ware-ni 麻許登可聞和礼爾余須等布

ma-kǝtǝ kamo ware-ni yǝs-u tǝ [i]p-u INT-thing EP I-DAT bring close-FIN DV say-ATTR I wonder [whether it is] true that [people] say that [she] has an intimate relationship with me (MYS 14.3384) A2: Ryukyuan The Ryukyuan cognate waa/baa/waN/banu to the Old Japanese wa/ware is well supported by all Ryukyuan languages.61 Except in some Southern Ryukyuan languages, as already mentioned above, it is always used as a first person singular pronoun ‘I.’ There are two morphological peculiarities of this pronoun in Ryukyuan, which are significant for the reconstruction of the proto-Japonic archetype. First, there are practically no traces in Ryukyuan of the extended stem ware,62 which together with the rarity of ware in Eastern Old Japanese further supports the claim that this stem is an innovation. Second, in some Ryukyuan dialects this pronoun has two stems: the unextended stem waa- and the extended stem ending in -N: waN. Examples from the Northern Ryukyuan dialects: Shuri waa/waN, Sesoko waa/waN, Kunigami waa/waN (Uchima 1984: 45–46). There are other dialects, however, particularly in the Southern Ryukyus, which do not have this distinction, employing either an unextended or an extended stem only: Kawahira banu, but Hateruma baa (Uchima 1984: 48–50). As an example, let us look at the paradigm of the first person waa/waN in Shuri:63 chart 11

The paradigm of Shuri first singular person pronoun waa/waN

Topic form Nominative Genitive (in compounds) Dative Accusative

waN-nee waa-ga waawaN-ni waN

61  Interesting enough, this cognate has survived in Ryukyuan up to this day, long after the first person pronoun wa/ware had disappeared from Central Japanese itself (except in a limited number of idiomatic expressions). 62  ware ‘I’ appears once in Ryukyuan plays (Hokama 1995: 740), but it is a singular example in a group of texts highly influenced by mainland Japanese. 63  The sample of paradigm is taken from Kaneshiro and Hattori (1955: 531), with some simplification of their transcription, which mainly involves converting their phonetic transcription into a phonological and omitting accent notations.

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

228 chart 11

Chapter 4 The paradigm of Shuri (cont.)

Comitative Directive

waN-tu waN-niNkai

On the surface, this chart looks like a perfect counterpart to the Classical Japanese paradigm, where the unextended stem wa- appears only before the possessive case marker -ga (corresponding to the nominative case marker -ga in Shuri), and the extended stem ware is used in all other cases. However, there are at least three problems facing this superfluous approach. First, as shown above, the distribution of extended and unextended stems in Old Japanese was different from Classical Japanese, especially if Eastern Old Japanese data is taken into consideration. Second, Classical and Old Japanese /re/ corresponds regularly to Shuri /ri/, not /N/. Third, Shuri extended stem waN never has the plural meaning ‘we,’ which is well attested in both Old and Classical Japanese ware. Comparing the above chart with the data from Old Japanese, we find the following major discrepancies: a) the form before a topic marker, as well as accusative and directive forms use the extended form waN, not the unextended form waa-; b) genitive form in compounds is unextended waa-, but not the special oblique form wa-ⁿ- as in Old Japanese. The major problem is the difference between the unextended form wa in Eastern Old Japanese and extended form waN in Shuri before the topic particle, since this is the form that in Old Japanese can be considered as equivalente to a form in isolation. There are two arguments that, as I believe, point to the fact that the Shuri usage of the extended form before the topic particle is an innovation. First, most combinations of Shuri nominals with the following topic particle ya result in fusion. When fusion occurs, it is very difficult, if not impossible, to draw a proper morphological boundary. Examples: chart 12

Shuri nominals in combination with a topic particle ya

Nominal

Nominal + ya

Meaning

sima ’ami kumu saNsiN

simaa ’amee kumoo saNsinoo

island rain cloud shamisen

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

NOMINALS

229

Second, in Old Ryukyuan the unextended stem wa, and not the extended stem waN, appears in isolation and before the topic marker (Hokama 1995: 730), although this usage is not found in the Omoro sōshi, where unextended wa- appears only before nominative -ga (Nakahara and Hokama 1978: 364). Nevertheless, usage in isolation of the unextended stem wa, not in the possessive function, occurs in the Ryūka, although very infrequently, compared to the usage of the same stem in the possessive function: 誰もわていやは

TARU mo wa te iy-aba who FP I DV say-COND if everyone says ‘I’ (RK 617) あれやわ自由しゆすが

are ya wa ziyu shiyu-su ga she TOP I freedom do-NML IP Will I be [able to] do with her as I want? (RK 1531) Therefore, it seems most likely that the proto-Ryukyuan paradigm of the first person pronoun could be reconstructed in the following way: non-oblique stem: wa < *wa oblique stem: waN < *wa-n[u] The fact that Eastern Old Japanese has wanu- as an oblique stem before the dative case marker -ni suggests that a similar system should be reconstructed for proto-Japonic. The only obstacle to this reconstruction is the fact that the Old Japanese possessive case marker -ŋga, which should appear after an oblique stem, appears after a non-oblique stem wa. However, it is necessary to keep in mind that the possessive case marker -ŋga itself begins with an -ŋ- < *-n-, and it is quite possible that the possessive case marker -ŋga should be reanalyzed as -ŋ-ga, with -ŋ- belonging to the stem of the pronoun instead to the following case marker. Needless to say, such an analysis is going to upset the possible alternative etymological analysis of the possessive case marker -ŋga and genitive case marker -nǝ offered in sections 1.2.2.2 and 1.2.2.3; and it will be necessary to explain why possessive -ŋga follows also other nominals except pronouns in Old Japanese and Ryukyuan. 2.2.2 First Person Pronoun a/are The first person pronoun a/are exhibits in Old Japanese more limited distribution compared to wa/ware, as it combines with less number of case markers Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

230

Chapter 4

than the latter. In contrast to wa/ware, there are no examples when the unextended stem a, even in combination with the following possessive marker -ŋga, it has a plural meaning, and the extended stem are is attested in plural function only once. Both the unextended stems a and the extended stem are can occur in isolation and with the accusative case marker -wo; the only difference between them concerns the usage with the possessive case marker -ŋga, which occurs only after the unextended stem a, and the dative case marker -ni, which is used exclusively after the extended stem are, as the following chart demonstrates. chart 13

Personal pronoun a/are in combination with case markers

Isolation Possessive Dative Accusative

Unextended stem a

Extended stem are

a* a-ŋga — a-wo

are — are-ni* are-wo

Note: * – only in Western Old Japanese

2.2.2.1 Stem a 2.2.2.1.1 Isolated Form a 阿波母與賣迩斯阿禮婆那遠岐弖遠波那志

a pa mǝ yǝ me n-i si ar-e-mba na-wo [o]k-i-te wo pa na-si I TOP EP EP woman DV-CONV EP exist-EV-CON you-ACC leave-CONV-SUB man TOP no-FIN Because I am a woman, I have no [other] man, besides you (KK 5) 比登都麻都阿勢袁

pitǝ-tu matu a se wo one-CL pine I beloved EP Oh, lone (lit.: one) pine, my beloved! (KK 29) 吾耳也之可流

A64 NƏMÏ ya sik[a]-ar-u I RP IP thus-exist-ATTR is [it] so only [for] me? (MYS 5.892) 64  Takagi et al. read this pronoun as ARE (Takagi et al. 1959: 100–101). Since it is written logographically, it is not possible to be absolutely sure whether the reading is ARE or A, but Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

231

NOMINALS

2.2.2.1.2 Oblique Stem a-ⁿ- in a Compound Similar to the oblique stem waⁿ-, there is just one reliable example of the oblique stem aⁿ- in a compound: 阿誤予阿誤予

aⁿ-go yǝ aⁿ-go yǝ I-child EP I-child EP Oh, my children! Oh, my children! (NK 8) There are no other uncontroversial examples for the oblique stem a-ⁿ- in Old Japanese texts. Yamada Yoshio cites aŋgi ‘my lord’ and Aⁿduma ‘East’ which allegedly contains a-ⁿ- (Yamada 1954: 32–33), but the first word in all likelihood just means ‘child’ and does not include a-ⁿ- (cf. MK ákí ‘child’ < *anki), since there is no coherent phonological explanation for the contraction of kimi ‘lord’ > ki. The etymology of Aⁿduma ‘East’ as a-ⁿ-duma ‘my wife’ is just a folk etymology. 2.2.2.1.3 Possessive a-ŋga 阿賀美斯古迩

a-ŋga mi-si ko-ni I-POSS see(CONV)-PAST/ATTR girl-DAT to the girl I saw (KK 42) 伊弊爾由伎弖伊可爾可阿我世武

ipe-ni yuk-i-te ika n-i ka a-ŋga se-m-u home-LOC go-CONV-SUB how DV-CONV IP I-POSS do-TENT-ATTR What will I do, when [I] go back home? (MYS 5.795) 久須利波牟用波美也古弥婆伊夜之吉阿何微麻多越知奴倍之

kusuri pam-u-yo pa miyako mi-mba iyasi-ki a-ŋga mï mata woti-n-umbɛ-si medicine eat-ATTR-ABL TOP capital see-COND ignoble-ATTR I-POSS body again rejuvenate(CONV)-PERF-DEB-FIN rather than taking the medicine, my ignoble body would have rejuvenated again if [I] saw the capital (MYS 5.848)

reading it as ARE breaks the meter of the poem, making this line eight syllables long instead of seven. Since this poem does not show other irregularities in meter, I prefer to follow the reading A, suggested by Nakanishi Susumu (Nakanishi 1978: 403). Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

232

Chapter 4

伊都麻弖可安我故非乎良牟

itu-maⁿde ka a-ŋga kopï-wor-am-u when-TERM IP I-POSS long.for(CONV)-exist-TENT-ATTR until when should I long for [you]? (MYS 15.3749) 2.2.2.1.4 Accusative a-wo 阿袁麻多周良武知知波波良波母

a-wo mat-as-uram-u titi papa-ra pa mǝ I-ACC wait-HON-TENT2-ATTR father mother-DIM TOP EP dear mother and father, who will probably wait for me (MYS 5.890) 安乎忘為莫

a-wo WASUR-As-uNA I-ACC forget-HON-NEG/IMP Do not forget me (MYS 12.3013) 吾妹子之阿乎偲良志

WA-ŋG-IMO-KO si a-wo SINOP-Urasi I-POSS-beloved-DIM EP I-ACC long-SUP It seems that my beloved longs for me (MYS 12.3145) 2.2.2.2 Stem are 2.2.2.2.1 Isolated Form are 佐泥牟登波阿禮波意母閇杼

sa-ne-m-u tǝ pa are pa omǝp-ɛ-ⁿdǝ PREF-sleep-TENT-FIN DV TOP I TOP long-EV-CONC Although I long to sleep together [with you] (KK 27) 阿禮許曾波余能那賀比登

are kǝsǝ pa yǝ-nǝ naŋga pitǝ I FP TOP world-GEN long person I, [the most] long[-living] man in the world (KK 72) 比等奈美爾安礼母作乎

pitǝ nami-ni are mǝ NAR-ER-U-wo person ordinary-COMP I FP born-PROG-ATTR-ACC although I also was born like an ordinary person (MYS 5.892)

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

NOMINALS

233

安礼麻多無

are mat-am-u I wait-TENT-FIN I will wait [for you] (MYS 15.3747) 2.2.2.2.2 Dative are-ni 安礼爾都氣都流

are-ni tuŋgɛ-t-uru I-DAT report(CONV)-PERF-ATTR [thus he] reported to me (MYS 17.3957) 佐刀毘等能安礼迩都具良久

sato-m-bitǝ-nǝ are-ni tuŋg-ur-aku village-GEN-person-GEN I-DAT report-ATTR-NML what the village people report to me (MYS 17.3973) 2.2.2.2.3 Accusative are-wo[mba] 阿礼乎婆母伊可爾世与等可

are-womba mǝ ika n-i se-yǝ tǝ ka I-ACC(EMPH) FP how DV-CONV do-IMP DV IP What do [you] think I should do? (MYS 5.794) 伊多豆良爾阿例乎知良須奈

Itaⁿdura n-i are-wo tir-as-una in.vain DV-CONV we-ACC fall-CAUS-NEG/IMP Do not let us fall in vain (MYS 5.852a) This is the only example when are is attested as the first person plural pronoun. 安礼乎於伎弖人者安良自等富己呂倍騰

are-wo ok-i-te PITƏ pa ar-aⁿzi tǝ pǝkǝr-ǝp-ɛ-ⁿdǝ I-ACC leave-CONV-SUB person TOP exist-NEG/TENT DV boast-ITER-EV-CONC although [I] repeatedly boast that there are probably no other persons besides me (MYS 5.892)

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

234

Chapter 4

Comparative Data Level A: Other Japonic A1: Eastern Old Japanese The first person pronoun a-/are is attested also in Eastern Old Japanese, but with a slightly more limited distribution: the unextended stem a- occurs only with the possessive marker -ŋga and accusative marker -wo, and the extended stem are appears only in isolation and with the accusative marker -wo. There is one example of a special usage in Eastern Old Japanese, when possessive form a-ŋga means not the possessive form ‘my’ of the first person pronoun, but the possessive form of the second person reflexive pronoun ‘your own’ (see MYS 20.4420 below). Stem a Possessive a- ŋga 可伎武太伎奴礼杼安加奴乎安杼加安我世牟

kaki-muⁿdak-i n-ure-ⁿdǝ ak-an-u-wo aⁿ-dǝ ka a-ŋga se-m-u PREF-embrace-CONV sleep-EV-CONC satisfy-NEG-ATTR-ACC what-DV IP I-POSS do-TENT-ATTR although [I] slept [with her] keeping [her] in my arms, since it was not enough [for me], what should I do? (MYS 14.3404) 奈氣伎曾安我須流

naŋgɛk-i sǝ a-ŋga s-uru lament-NML FP I-POSS do-ATTR I lament (MYS 14.3524) Special Usage

There is one example in Eastern Old Japanese, when possessive a-ŋga means not the first person, but the second person reflexive pronoun ‘your own’: 比毛多要婆安我弖等都氣呂許礼乃波流母志

pimo taye-mba a-ŋga te-tǝ tukɛ-rǝ kǝre n-ǝ paru mǝs-i cord tear-COND your.own-POSS hand-COM attach-IMP this DV-ATTR needle hold-CONV if the cords [of your garment] tear, attach them with your own hand, holding this needle (MYS 20.4420)

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

235

NOMINALS Accusative a-wo 安乎麻知可祢弖

a-wo mat-i-kane-te I-ACC wait-CONV-NEG/POT(CONV)-SUB [she] cannot wait for me (MYS 14.3563) Stem are Isolated Form are 許呂安礼比毛等久

kǝ-rǝ are pimo tǝk-u beloved-DIM I cord untie-FIN [my] beloved [and] I will untie our [garment] cords (MYS 14.3361) Accusative are-wo 和我世奈阿礼乎之毛波婆

wa-ŋga se-na are-wo si [o]mop-amba I-POSS beloved-DIM I-ACC EP love-COND my beloved, if [you] love me (MYS 20.4426) A2: Ryukyuan The first person pronoun a is attested in Old Ryukyuan, (namely in the Omoro sōshi and the Ryūka), in the Sesoko dialect on Okinawa island, in Ogami, Irabu, Minna, and Nakasuji subdialects of the Miyako dialect (Thorpe 1983: 218), and in the Yonaguni dialect. The Sesoko forms: ’a, ’agan (< ’agami), and ’agami are used only as plural ‘we’ (Uchima 1984: 85). The Yonaguni form ’anu ‘I’ is used as first person singular pronoun (Hirayama 1967: 239), but ba- is used in plurals (Thorpe 1983: 218). Such limited distribution, confined to few dialects, mostly in South Ryukyus and to Old Ryukyuan texts, the latter being known to be highly influenced by mainland Japanese, could have provoked reasonable doubts in the authenticity of a in Old Ryukyuan, but since a/are did not survive even into early Classical Japanese, the possibility that authors of Old Ryukyuan texts were using Old Japanese texts as their desk reference tools is very slim. Even more unrealistic is direct borrowing from Old Japanese to the Sesoko, Miyako and Yonaguni dialects. Similar to the Ryukyuan counterpart of OJ wa, there are no traces of the extended stem are in Ryukyuan. The a form in Old Ryukyuan appears either in isolation (followed by a topic marker), or before the case marker -ga (< *-ŋga). Examples:

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

236

Chapter 4

Old Ryukyuan あはいのて

a fa ino-te I TOP pray-SUB I pray, and … (OS 13.747; OS 22.1535) あがおとぢや

a-ga oto-diya I-POSS younger brother-person my younger brother (OS 16.1155) Sesoko Dialect

’a-ga ya we-POSS house our house (Uchima 1984: 85) ’agan ik-a we go-HORT Let us go (Uchima 1984: 85) ’agami ya wur-a we TOP stay-HORT Let us stay (Uchima 1984: 85) 2.2.3 First Person Pronoun marǝ This pronoun is used only as a first-person personal pronoun. It occurs once in the Kojiki kayō and once in the Nihonshoki kayō in the identical context. Ōno Susumu believes that this pronoun is etymologically related to -marǝ, a suffix of male names, e.g., Yasu-marǝ, Kotu-marǝ, Uta-marǝ, etc. (Ōno 1990: 1241). In Old Japanese it is attested only as male first person pronoun, but there are too few examples to draw any decisive conclusions.65 In Classical Japanese it can 65  R  oy A. Miller, in an apparent haste to secure an additional Altaic etymology for Japonic, goes as far as to claim that the marǝ form has external Altaic connections, and that it is an alternative variant of EOJ warǝ ‘I’ (Miller 1971: 157, 177–178), which has a singular attestation in MYS 20.4343, cited above in section 2.2.1. Unfortunately for Miller’s Altaic theory, both WOJ marǝ ‘I’ and EOJ warǝ ‘id.’ have simpler internal explanations. EOJ warǝ apparently corresponds to WOJ ware, as can be seen from the verbal form omɛp[-]o-ⁿdo, corresponding to WOJ omǝp-ɛ-ⁿdǝ ‘although [I] think,’ which occurs in the next line in the same poem (MYS 20.4343). As for Miller’s statement that WOJ marǝ ‘I ‘later becomes Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

237

NOMINALS

be used as a first person pronoun by both men and women. It is not attested in Eastern Old Japanese or in Ryukyuan. Examples: 岐許志母知袁勢麻呂賀知 (KK script) 枳居之茂知塢勢摩呂餓智 (NK script)

kikǝs-i-mǝt-i-wos-e marǝ-ŋga ti drink(HON)-CONV-hold-CONV-HON-IMP I-POSS father Deign to drink [it], my father (KK 48, NK 39) 2.2.4 First Person Pronoun na The first person pronoun na ‘I’ occurs slightly more frequently than the first person pronoun marǝ. Apart from several controversial examples, the following attestations seem quite certain, in spite of Yamada Yoshio’s treating them as second person pronouns (Yamada 1954: 60–61). First person pronoun na is always used as a first person singular ‘I.’ 名兄乃君

na SE n-ǝ KIMI I beloved DV-ATTR lord my beloved lord (MYS 16.3885) 奈弟乃美許等

na OTƏ n-ǝ mi-kǝtǝ I younger brother DV-ATTR HON-thing my dear younger brother (MYS 17.3957) These examples lacking the possessive case marking (we would expect na-ŋga se ‘my beloved’ and na-ŋga otǝ ‘my younger brother’), could have been probably taken as na ‘you’ as an address form plus the rest: ‘you, my beloved and lord,’ ‘you, my dear younger brother,’ but this seems unlikely in the light of a se ‘my beloved’ attested in KK 29 (see section 2.2.2).

a common element used in second position in aristocratic given names’ (Miller 1971: 177; italics are mine—A.V.), it is fallacious in two respects. First, -marǝ as a suffix of male names is attested before the examples of marǝ as a first person pronoun in the Kojiki kayō (712 CE) and the Nihonshoki kayō (720 CE): Ō-no Yasumaro, the author of the preface to the Kojiki, and Ōtomo-no Yasumaro (d. 714), the grandfather of the great Man’yōshū poet Ōtomo-Yakamochi, both have -marǝ in their personal names. Second, -marǝ as the last element of male names, frequently occurs in the peasant names of the Mino province census of 702 (Shōsōin 2 1990), so I doubt that there is anything ‘aristocratic’ about it at all.

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

238

Chapter 4

Comparative Data Level A: Other Japonic A1: Eastern Old Japanese The first person pronoun na ‘I’ occurs in Eastern Old Japanese also in two examples: 奈西乃古何夜蘇志麻加久理

na se n-ǝ ko-ŋga yaso sima kakur-i I beloved DV-ATTR child-POSS eighty island hide-CONV my beloved hides [behind] eighty islands … (FK 8) 奈勢能古

na se n-ǝ ko I beloved DV-ATTR child my beloved (MYS 14.3458) Level B: External Comparisons Limited attestation of the first person na and its limited usage as a singular first person pronoun probably point to the fact that it is a loan. The likeliest source is some form of Old Korean (Paekche?), cf. MK nà ‘I.’66 Examples: na non icey silum-i kiph-e I TOP now sadness-NOM deep-CONV I am very sad now, and … (WS II.5) na-y nimkum kuli-sy-a I-GEN king miss-HON-CONV My king is lonely, and … (YP 50) 2.2.5 Second Person Pronoun na/nare The second person pronoun na/nare ‘you’67 is believed to be a non-honorific familiar pronoun, used towards one’s friends and family members or towards 66  There are no phonetic attestations of this pronoun in Old Korean. 67  There is a later Middle Japanese form namuti/namudi ‘you,’ which is obviously derived from OJ na. Namuti is attested in Old Japanese as well, but the attestations are limited to either glosses in the texts, or the transcriptional usage. Because there are no textual examples, I exclude namuti from this grammar.

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

239

NOMINALS

people ranking lower on the social scale (Omodaka et al. 1967: 512). This definition is, I believe, incorrect. Although the majority of the examples below reflect such distribution, e.g., it is frequently used in both Western and Eastern Old Japanese by a man to a woman, by a husband to his wife, etc., but there are also examples when it is used by a woman towards a man. The most important counter-example to the traditional definition can be found in the KK 73 below, when it is used by a courtier towards the emperor. Thus, I believe that na/nare has to be redefined as a neutral second person pronoun, pretty much as its first person counterpart wa/ware, which also can be used without any regards to the gender or social status. There is one example demonstrating that the second person pronoun na, like its first person counterpart wa, might have an oblique stem marker -N (see KK 3 below, where the oblique stem na-ⁿ is found in a compound). The extended stem nare occurs exceedingly rare in comparison to the unextended stem na (two examples in Western Old Japanese texts and three examples in Eastern Old Japanese texts). The unextended stem na never has a plural meaning, the extended stem nare is attested in plural meaning only once. That can probably be taken as the further evidence to the proposal discussed in section 2.2.1 in connection with the first person pronoun wa/ware, that the extended form in -re was originally a plural form. Similar to the situation with the first person unextended stem wa, the second person unextended stem na also shows more versatility in combination with different case markers in Eastern Old Japanese than in Western Old Japanese. That demonstrates that Eastern Old Japanese is likely to be more archaic in this respect. The following chart summarizes possible combinations of the unextended stem na and the extended stem nare with following case markers as well as their respective occurrence in Western and Eastern Old Japanese texts. chart 14

Isolation Possessive Dative Accusative Comitative

Second person pronoun na/nare in combination with case markers

Unextended stem na

Extended stem nare

na S na-ŋga, na-ne-ŋga** na-ni* S na-wo S na-tǝ* S

nare SP — — nare-wo** S —

Note: S – Singular, P – Plural, * – only in Eastern Old Japanese, ** – only in Western Old Japanese

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

240

Chapter 4

2.2.5.1 Stem na 2.2.5.1.1 Isolated Form na 那迦士登波那波伊布登母

nak-aⁿzi tǝ pa na pa ip-u tǝmǝ weep-NEG/TENT DV TOP you TOP say-FIN CONJ Even though you say that [you] would not weep (KK 4) (the husband to his wife) 那許曾波遠迩伊麻世婆

na kǝsǝ pa wo n-i imas-e-mba you FP TOP man DV-CONV exist(HON)-EV-CON since you are a man (KK 5) (the wife to her husband) 那許曾波余能那賀比登

na kǝsǝ pa yǝ-nǝ naŋga pitǝ you FP TOP world-GEN long person you, [the most] long[-living] man in the world (KK 71) (the emperor to his courtier) 儺波企箇輸揶

na pa kik-as-u ya you TOP ask-HON-FIN IP Shall [I] ask you? (NK 62) (the emperor to his subject) 2.2.5.1.2

Oblique Stem na-ⁿ- in a Compound

能知波爾阿良牟遠伊能知波那志勢婆多麻比曾

nǝti pa na-ⁿ-dǝri n-i ar-am-u-wo inǝti pa na-si-se-tamap-i-sǝ later TOP you-OSM-bird DV-CONV exist-TENT-ATTR-ACC life TOP NEG-die-CAUS(CONV)-HON-CONV-do because [I] will be your bird later, do not kill [your] life [with desire] (KK 3) (the female deity to the male deity) 2.2.5.1.3 Possessive na-ŋga 那賀那加佐麻久

na-ŋga nak-as-am-aku you-POSS weep-HON-TENT-NML Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

241

NOMINALS

the fact that you will weep (KK 4) (the husband to his wife) 佐和佐和爾那賀伊弊勢許曾

sawa-sawa n-i na-ŋga ip-es-e kǝsǝ noisily-noisily DV-CONV you-POSS speak(CONV)-HON-EV FP you spoke noisily (KK 63) (the emperor to the empress) 那賀美古夜那毘迩斯良牟登加理波古牟良斯

na-ŋga miko ya tumbi n-i sir-am-u tǝ kari pa ko (u)m-urasi you-POSS68 prince IP ? DV-CONV rule-TENT-FIN DV wild goose TOP egg bear-SUP Wild goose probably laid an egg to show that your prince will rule ?-ly (KK 73) (a courtier to the emperor) 奈何名能良佐祢

na-ŋga NA nǝr-as-an-e you-POSS name name-HON-DES-IMP [I] wish [you] tell [me] your name! (MYS 5.800) (didactic poem to a friend or a junior official) 奈我奈氣婆

na-ŋga nak-ɛ-mba you-POSS cry-EV-CON when you cry (MYS 15.3785) (a man to a cuckoo) 68   This poem presents a certain difficulty in interpretation of na-ŋga as ‘you-POSS.’ Traditional interpretation of na- here as a second person pronoun (Tsuchihashi 1957: 81; Tsuchihashi 1972: 275) is in contradiction with the traditional definition of na/nare as a familiar second person pronoun used towards relatives, friends, and lower-ranking people, since in this case it used by a courtier (although old) towards an emperor. I thought about a possibility to interpret it as a first person na ‘I,’ especially since this poem is from a section of the Kojiki dedicated to Emperor Nintoku, who is believed to have strong Korean connections (Ledyard 1975: 251). This interpretation, although plausible at the first glance, also meets with difficulty: if na-ŋga miko is to be understood as ‘my prince,’ it remains unclear why the old man addresses Nintoku, who is already an emperor, as a prince. To be on the conservative side, I decided to follow the tradition here, and interpret na-ŋga miko as ‘your prince,’ believing that it refers to Nintoku’s successor, who will rule after his father, not being replaced by an outsider. Thus, the traditional definition of na/nare as a familiar second person pronoun probably has to go in the light of this example. I redefine it as a neutral second person pronoun. Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

242

Chapter 4

2.2.5.1.4 Accusative na-wo 阿波母與賣迩斯阿禮婆那遠岐弖遠波那志

a pa mǝ yǝ me n-i si ar-e-mba na-wo [o]k-i-te wo pa na-si I TOP EP EP woman DV-CONV EP exist-EV-CON you-ACC leave-CONV-SUB man TOP no-FIN Because I am a woman, I have no [other] man, besides you (KK 5) (a wife to her husband) 2.2.5.1.5 Special Form na-ne There are two examples, where the unusual form na-ne appears. It is usually explained as a contraction of na-ane ‘my beloved,’ lit. ‘my elder sister’ (Takagi 1957: 305), but the problem is that the female beloved is usually called imo, lit. ‘younger sister.’ In addition, in the second example (MYS 9.1800) there is imo na-[a]ne-ŋga ‘younger sister you-elder sister-POSS,’ where both imo ‘younger sister’ and ane ‘elder sister’ are used together, which makes little or no sense at all. I think that it is more likely that in both cases we have the extended stem of na ‘you’ ending in -ne (compare also sǝ-ne 2nd/3rd(?) person pronoun in section 2.2.6): 名姉之恋曾

na-ne-ŋGA KOP-URE sǝ you-?-POSS love-EV FP you love [me] (MYS 4.724) (a man to a woman) 妹名根之作服異六白細乃紐

IMO na-ne-ŋGA TUKUR-I KI-SE-k-em-u SIRO-TAPƐ-nǝ PIMO beloved you-?-POSS make-CONV wear-CAUS-PAST/FIN-TENT-ATTR whitecloth-GEN cord the cord of the white cloth that you, beloved, made and made [me] wear (MYS 9.1800) (a man to a woman) 2.2.5.2 Stem nare The extended stem nare very rarely occurs in Western Old Japanese texts: only two examples are attested, one as the isolated form, and another in the accusative form nare-wo. The latter example occurs in the Nihon ryōiki, which is a very late text.

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

243

NOMINALS

2.2.5.2.1

Isolated Form nare

於夜那斯爾奈礼奈理鶏迷夜

oya na-si n-i nare nar-i-k-em-ɛ ya parent no-FIN DV-CONV you be born-CONV-PAST/FIN-TENT-EV IP Were you possibly born without parents? (NK 104) (the prince to the beggar) 2.2.5.2.2 Accusative nare-wo 奈禮乎曾與咩爾保師登多禮

nare-wo sǝ yǝme n-i posi tǝ tare you-ACC FP bride DV-CONV be desirable DV who who wants you as [his] bride? (NR 2.33) Comparative Data Level A: Other Japonic A1: Eastern Old Japanese The second person pronoun na/nare is also well attested in Eastern Old Japanese. Similar to Western Old Japanese, the unextended stem na occurs much more frequently. Stem na

The unextended stem na occurs in Eastern Old Japanese not only in isolation and with the possessive case marker -ŋga and the accusative case marker -wo, as in Western Old Japanese, but also with the dative case marker -ni and the comitative case marker -tǝ. Isolated Form na

Unfortunately, the only two examples of the isolated form na in Eastern Old Japanese are not attested phonographically, but an educated guess on the basis of the meter of poems and widespread usage of the unextended stem na in Eastern Old Japanese suggests that this logographic usage represents na pretty accurately. 汝者故布婆曾毛

NA pa kop-u[re]- mba sǝ mo you TOP love-EV-CON FP EP because you love [me] (MYS 14.3382) (a man to a woman)

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

244

Chapter 4

汝波安杼可毛布

NA pa aⁿ-dǝ ka [o]mop-u you TOP what-DV IP think-ATTR What do you think? (MYS 14.3494) (a man to a woman) Possessive na- ŋga 奈我目保里勢牟

na-ŋga MƐ por-i se-m-u you-POSS eye desire-NML do-TENT-FIN [I] wish [to see] your eyes (MYS 14.3383) (a man to a woman) 奈我己許呂能礼

na-ŋga kǝkǝrǝ nǝr-e you-POSS heart tell-IMP Reveal [to me] your heart (MYS 14.3425) (a man to a woman) Dative na-ni 奈爾与曾利鶏米

na-ni yǝsǝr-i-k-em-ɛ you-DAT approach-CONV-PAST/FIN-TENT-EV [I] approached you (MYS 14.3468) (a man to a woman) 奈爾己曾与佐礼

na-ni kǝsǝ yǝs-ar-e you-DAT FP approach-PROG-EV [I] approached you (MYS 14.3478) (a man to a woman) Accusative na-wo 奈乎許曾麻多賣

na-wo kǝsǝ mat-am-e you-ACC FP wait-TENT-EV [I] will wait for you (MYS 14.3493) (a woman to a man) Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

NOMINALS

245

可波刀爾奈乎麻都

kapa-to-ni na-wo mat-u river-door-LOC you-ACC wait-FIN [I] wait for you at the ford (MYS 14.3546) (probably a woman to a man, but it is unclear) Comitative na-tƏ 勢奈那登布多理左宿而久也思母

se-na na-tǝ puta-ri sa-NE-TE kuyasi-mǝ beloved-DIM you-COM two-CL PREF-sleep(CONV)-SUB be.regrettable-EXCL it is regrettable that [I], my beloved, slept with you, two [of us]! (MYS 14.3544) (a woman to a man) Stem nare

Isolated stem nare is attested in Eastern Old Japanese only in one example. There are no other examples of EOJ nare. Isolated Form nare 和我可度乃可多夜麻都婆伎麻己等奈礼

wa-ŋga kaⁿdo-nǝ kata yama tumbaki ma-kǝtǝ nare I-POSS gate-GEN side mountain camellia INT-thing you [Oh,] camellia [flowers] at the mountain near my gate, [it is] really you (MYS 20.4418) (a man to flowers, this is the only apparent usage of nare as a plural form) A2: Ryukyuan The second person pronoun na is attested comparatively well in Ryukyuan: there are no attestations in the Omoro sōshi or the Ryūka: the Old Ryukyuan naa ‘you’ seems to be confined to Ryukyuan plays (Hokama 1995: 472). However, there is no lack of attestations in the modern dialects, mostly in the Northern Ryukyus, where such forms as Koniya nam¡; Namizato naa; Sesoko naa, naN; Shuri naa, etc. are attested (Uchima and Arakaki 2000: 358). Hateruma has an aberrant form daa, which is probably not related. It seems that in all these dialects this second person pronoun has a similar or close function to the Shuri pronoun naa, which is a familiar pronoun used towards older people of lower status (RKJ 1983: 399).69 An example from the Shuri dialect: 69  Uchima and Arakaki gloss their Ryukyuan dialect attestation as modern Japanese anata ‘you’ (Uchima and Arakaki 2000: 358), which is a semi-formal pronoun used towards Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

246

Chapter 4

naa ya /içi mooca ga you TOP when come/PERF IP When did you come? (RKJ 1983: 399) Level B: External Comparisons The well-known comparison for Old Japanese second person pronoun na is Middle Korean second person pronoun nè, also well attested in modern standard Korean and the dialects. Examples from Middle Korean: ne-y na-y mal-ol ta tul-ul-tta hoy-a-nol you-NOM I-GEN word-ACC all hear-IRR/PART-QF say-CONV-CON When [I] say: ‘Are you going to listen to all my words?’ (SP 6.7) ne non kisk-e two you rejoice-CONV FP Although you rejoiced … (KKK 2.5) Given the fact that in Ryukyuan the second person pronoun na ‘you’ is a kind of honorific, it is likely that the Korean second person pronoun was borrowed by Western Old Japanese from Korean, and then was borrowed from the former into Eastern Old Japanese, and into Ryukyuan. 2.2.6 Second Person Pronoun masi/mimasi/imasi The second person pronoun masi/mimasi/imasi occurs in Old Japanese texts in the phonographic spelling quite rarely. In Western Old Japanese it occurs only in the forms imasi and mimasi. The first of these forms is attested in the Man’yōshū and the Senmyō once each, the second appears exclusively in the Senmyō, although much more frequently—nine times altogether. Yamada Yoshio believes that the original form of this pronoun is just masi, which appears only in Eastern Old Japanese, with mi- and i- being prefixes (Yamada 1954: 95). He is probably right: mi- in mi-masi is likely to be an honorific prefix mi-, which could have irregularly changed to i- before the following /m/, although this is difficult to prove with certainty. Yamada gives an ambivalent definition to this pronoun as either honorific or pejorative (Yamada 1954: 94). Omodaka et al. define masi as an originally honorific second person pronoun, which was also used as a familiar pronoun toward people with a lower social status (Omodaka et al. 1967: 673). At the same time Omodaka et al. claim that relatives (e.g., wife to her husband), or towards lower-ranking people, when some minimal degree of respect is maintained.

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

247

NOMINALS

mimasi is more honorific than imasi or masi, although mimasi can also be used towards ministers and children (Omodaka et al. 1967: 715). This claim seems to be contradicted by the example from SM 29 below, where it is used by Empress Kōken towards the dethroned Emperor Jūnin. Ōno et al., on the other hand, define masi/imasi/mimasi as second person pronoun used towards equal or socially lower people, although at the same time they suggest that masi/imasi/ mimasi is derived from the nominalization of the honorific verb imas- ‘to exist’ (Ōno 1990: 1220). It is unclear, however, why a nominalization of an honorific verb is used for a non-honorific pronoun. In addition, the form mi-masi remains unexplained under this scenario. In Eastern Old Japanese the second person pronoun masi occurs as a textual variant of na ‘you,’ which probably demonstrates that at least in the Eastern Old Japanese na and masi did not have any functional differences. The overall paucity of examples precludes any exact definition of socio-linguistic function of the second person pronoun masi/imasi/mimasi in Old Japanese, although it is extremely likely that it was not an honorific pronoun as alleged, since in most examples it is found in the speech directed by a person higher in the social hierarchy to a person lower in the same hierarchy. 伊麻思毛吾毛事応成

imasi mo WARE mo KƏTƏ NAR-UmBƐ-SI YA you FP I FP thing be-DEB-FIN IP Should [it] be something for both you and me? (MYS 11.2517) (probably a man to a woman) 天下方朕子伊麻之仁授給

AMƐ-NƏ SITA pa WA-ŋGA KO imasi-ni SAⁿDUKƐ-TAMAP-U heaven-GEN bottom TOP I-POSS child you-DAT give(CONV)-HON-FIN [I] give to you, my child, the land under the Heaven (SM 29) (Empress Kōken to Emperor Jūnin on the occasion of the latter’s dethronement) 美麻斯乃父止坐天皇乃美麻斯尓賜志天下

mimasi-nǝ TITI tǝ IMAS-U SUMERA-MIKƏTƏ-nǝ mimasi-ni TAMAP-i-si AMƐ-NƏ SITA you-GEN father DV exist(HON)-ATTR emperor-deity-GEN you-DAT give(HON)-CONV-PAST/ATTR heaven-GEN bottom the land under the Heaven that the emperor-deity, who is your father, gave to you (SM 5) (Empress Genshō to Emperor Shōmu)

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

248

Chapter 4

美麻斯王乃齢乃弱爾

mimasi MIKO-nǝ YƏPAPI-nǝ YOWA-KI-ni you prince-GEN age-GEN weak-ATTR-LOC because you, prince, were in infancy … (SM 5) (Empress Genshō to Emperor Shōmu) 美麻志大臣乃仕奉状

mimasi OPOMAPETUKIMI-nǝ TUKAPƐ-MATUR-I-KƏ-SI SAMA you minister-GEN serve(CONV)-HUM-CONV-come-PAST/ATTR shape the way how you, minister, served [us] until now (SM 52) (the emperor to the deceased minister) Comparative Data Level A: Other Japonic A1: Eastern Old Japanese In Eastern Old Japanese only the form imasi is attested once. In all probability, it is neutral second person pronoun. Again, the paucity of data precludes any exact definition of its socio-linguistic function. 伊麻思乎多能美波播爾多我比奴

imasi-wo tanǝm-i papa-ni taŋgap-i-n-u you-ACC trust-CONV mother-DAT go against-CONV-PERF-FIN [I] went against [my] mother, trusting you (MYS 14.3359) (a woman to a man) 2.2.7 Second Person Pronoun ore The second person pronoun ore can be both singular and plural. It probably has the same structure as other pronouns ending in -re, namely that the form ore represents an extended stem. Unfortunately, the unextended stem *o is not attested in Western Old Japanese texts, but it seems that one can make an educated guess and introduce a morphemic boundary into o-re by analogy with wa-re ‘I, we’ and na-re ‘you.’ Once again we see that the extended stem in -re probably has a connection with plural marker -ra (see sections 2.2.1 and 2.2.5 on the tendency to use the extended stems ware and nare for plural usage). The second person pronoun ore clearly functions as a pejorative pronoun. This is probably the reason why it is not attested in poetic texts. Examples:

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

NOMINALS

249

意禮二字以音爲大國主神

ore [ni ⁿzi i on] OPO KUNI NUSI KAMÏ NAR-I you [two characters by sound] great land master deity become-FIN you ([these] two characters [to be read] phonographically) will become the deity Opo kuni nusi (lit.: the master of the great land) (KJK 1.30a) Susanowo speaking to Opo kuni nusi who eloped with the daughter of the former. 所作仕奉於大殿内者意禮此二字以音先入

OPO TƏNƏ-NƏ UTI-NI pa TUKUR-I-TUKAPƐ-MATUR-U PA ore [ni ⁿzi i on] MAⁿDU IR-I great palace-GEN inside-LOC TOP make-CONV-serve(CONV)-HUM-ATTR TOP you [these two characters by sound] first enter-CONV you will enter first into the great palace that [you] have built (KJK 2.5b) Retainers of the emperor Jinmu speaking to the person who plotted to kill the emperor. 意禮熊曾建二人不伏無禮聞看而取殺意禮詔而遣

ore kumasǝ TAKERU PUTA-RI MATUR-AP-AⁿZ-U REI NA-SI TƏ KIKƏS-I-MES-I-TE ore TƏR-I-KƏRƏS-E TƏ NƏTAMAP-i-TE TUKAP-AS-ER-I you Kumasǝ brave two-CL serve-ITER-NEG-CONV ritual no-FIN DV hear(HON)CONV-HON-CONV-SUB you take-CONV-kill-IMP DV say(HON)-CONV-SUB send-HON-PROG-FIN [the emperor] heard that you two Kumasǝ brave [brothers] do not submit and have no manners, so [he] sent [me] ordering to kill you (KJK 2.39a) Prince Yamatǝ Takeru talking to a Kumasǝ leader about the latter and the latter’s elder brother whom he killed earlier. Comparative Data Level A: Other Japonic A2: Ryukyuan Both Old Ryukyuan and most modern Ryukyuan languages have a second person singular pronoun that is apparently cognate to WOJ ore ‘you,’ for example: OR o, u; Shuri ʔyaa; Nase ʔya; Koniya, Chabana ʔura; Hentona ya; Kuroshima ʔuva; Psara vva, etc. (Hirayama 1966: 303; Hirayama 1967: 241). This pronoun functions as an informal or pejorative pronoun. Examples:

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

250

Chapter 4

Old Ryukyuan おがやへよりおわよりな

o-ga ya-fe-yori owa-yor-i na you-POSS home-side-ABL come(HON)-approach-CONV EP Did [you] come from your home? (OS 14.998) Shuri

ʔari-ga yum-aa ʔyaa-N yum-ee he-NOM read-COND you-FP read-IMP If he reads [it], you read it, too (RKJ 69) Miyako

ʔuva-ga du basïka-N thou-NOM FP bad-FIN You are bad (Nohara 1998: 372) vva ndza-ŋkai ga ik-ï you where-LOC IP go-FIN Where are you going? (Nohara 1998: 381) The familiar and pejorative pronouns very often reflect the original pronominal system. Given also the functional and geographical limitations of the Western Old Japanese second person pronouns na ~ nare and masi ~ mimasi ~ imasi discussed in sections 2.2.5 and 2.2.6, it seems reasonable to assume that WOJ ore and PR *o-ra represent the original proto-Japonic second person pronoun: singular *o and plural *o-ra. 2.2.8 Third Person Pronoun si This pronoun in Old Japanese occurs exclusively with the following possessive case marker -ŋga, with both animate and inanimate referents. Whether the personal pronoun si represents second or third person is a debated question. Tradition generally holds it to be the third person pronoun (Yamada 1954: 70), cf. also the commentaries on the Old Japanese texts discussed below (Tsuchihashi 1957; Tsuchihashi 1976; Takagi et al. 1959; Nakanishi 1978; Kojima et al. 1972; Satake et al. 1999); although there is also another point of view stipulating that si represents a mesial (chūshō) demonstrative pronoun (Omodaka et al. 1967: 346). In the light of absence of earlier commentaries on Old Japanese texts where this pronoun appears, the value of the traditional Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

251

NOMINALS

point of view, starting at best from the eighteenth century, can be debated. The first known challenge to the traditional point of view known to me was presented by Roy A. Miller (Miller 1971: 159), although unfortunately he did not present any philological discussion of texts, where this pronoun appears, limiting himself to a bare statement that OJ si is a second person pronoun supported by a reference to Yamada Yoshio’s grammar (Yamada 1954: 75–76), which is an unfortunate mistake, because Yamada clearly states his point of view earlier that si is a third person pronoun (Yamada 1954: 70). Samuel E. Martin adopted the traditional point of view defining si as ‘it, that, he’ (Martin 1987: 521). This controversy can be resolved only by a linguistic and philological analysis of texts where the pronoun si appears. One general consideration in favor of Miller’s position is that it is typologically odd that a third person pronoun would disappear completely from a language where it once appeared, and that would be the case with Old Japanese: if si is a genuine third person pronoun it is unclear why it disappears in the later history of the language. Generally, the typology of third person pronouns works in the opposite direction: there are many languages that originally did not have third person pronouns, but acquire them for one reason or another; however, I am not aware of any cases where a language unconditionally loses original third person pronouns. Nevertheless, language typology is only the secondary evidence as compared to the internal language data. Miller gives two examples in support of his point of view that OJ si represents a second person pronoun, NK 80 and MYS 5.904, without providing any evidence in support of his point of view (Miller 1971: 159). Let us start the discussion from these two as well. I will translate si as second person ‘you’/‘he’ until the data positively prove whether it is second person or third person pronoun in the following examples: 偉儺謎能陁倶彌柯該志須彌儺皤旨我那稽麼柂例柯柯該武預

Winambɛ-nǝ takumi kakɛ-si sumi-napa si-ŋga na-k-emba tare ka kakɛ-m-u yǝ Winambɛ-GEN carpenter apply(CONV)-PAST/ATTR ink-cord you/he-POSS noATTR-COND who IP apply-TENT-ATTR EP the ink-cord that the carpenter of Winambɛ applied: if you/he are/is no more, who would apply it? (NK 80) This poem is preceded by the following text in the Nihonshoki: The carpenter, Mane, of the Wina-be, planed timber with an axe, using a stone as a ruler. All day long he planed and never spoiled the edge by mistake. The emperor visited the place, and, wondering, asked of him, saying: Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

252

Chapter 4

‘Dost thou never make a mistake and strike the stone?’ Mane answered and said: ‘I never make a mistake!’ Then the emperor called together the Uneme, and made them strip off their clothing and wrestle in open view with only their waistcloths on. Hereupon Mane ceased for a while and looked upon them, and then went on with his planing. But unawares he made a slip of the hand, and spoilt the edge of his tool. The emperor accordingly rebuked him, saying:—‘Where does this fellow come from that, without respect to us, he gives such heedless answers with unchastened heart?’ So he handed him over to the Mononobe to be executed on the moor. Now against his comrades there was a carpenter who lamented for Mane, and made a song, saying …70 On the basis of this introduction to the poem, it is really difficult to choose between the second or third person personal pronoun: Mane’s comrade could have made an address in either second or the third person. It seems, however, that the third person ‘he’ would be more natural, given the fact that Mane is no longer present at the scene, but this is difficult to prove, unless some additional evidence is found. Miller’s next example is from MYS 5.904, a Yamanoue-no Okura’s poem dedicated to the untimely death of his son, Purupi. For the sake of the following philological analysis, the poem has to be cited in full: 世人之貴慕七種之宝毛我波何為和我中能産礼出有白玉之吾子古日者明星 之開朝者敷多倍乃登許能辺佐良受立礼杼毛居礼杼毛登母爾戯礼夕星乃 由布弊爾奈礼婆伊射祢余登手乎多豆佐波里父母毛表者奈佐我利三枝之 中爾乎祢牟登愛久志我可多良倍婆何時可毛比等等奈理伊弖天安志家口 毛与家久母見武登大船乃於毛比多能無爾於毛波奴爾横風乃爾布敷可爾 覆來礼婆世武須便乃多杼伎乎之良爾志路多倍乃多須吉乎可氣麻蘇鏡弖 爾登利毛知弖天神阿布藝許比乃美地祇布之弖額拝可加良受毛可賀利毛 神乃末爾麻爾等立阿射里我例乞能米登須臾毛余家久波奈之爾漸漸可多 知都久保利朝朝伊布許等夜美霊剋伊乃知多延奴礼立乎杼利足須里佐家 婢伏仰武祢宇知奈氣吉手爾持流安我古登婆之都世間之道

YƏ-NƏ PITƏ-NƏ TAPUTOP-i-NE ŋG-AP-U NANA-KUSA-NƏ TAKARA mo WARE pa NANI SE-M-U wa-ŋga NAKA-nǝ UMARe-IⁿDE-TAR-U SIRA TAMA-NƏ WA-ŋGA KO PURUPI pa AKA-POSI-NƏ AK-URU ASITA pa SIK-i-tapɛ-nǝ tǝkǝ-nǝ PE sar-aⁿz-u TAT-Er-e-ⁿdǝmo WOr-e-ⁿdǝmo tǝmǝ n-i TAPAmB-Ure YUPU-TUⁿZU-nǝ yupu-m-be n-i nar-e-mba iⁿza ne-yǝ tǝ TE-wo taⁿdusapar-i 70  I cite William G. Aston’s translation of the Nihonshoki (Aston 1896: 361–362).

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

NOMINALS

253

TITI-PAPA mǝ UPƐ pa na-saŋgar-i SAKI-KUSA-NƏ NAKA-ni wo ne-m-u tǝ UTUKUSI-ku si-ŋga katar-ap-ɛ-mba ITU SI kamo pitǝ tǝ nar-i-iⁿde-te asi-k-eku mo yǝ-k-eku mǝ MI-m-u-tǝ OPO PUNE-nǝ omop-i-tanǝm-u-ni omǝp-an-u-ni YƏKƏ-SIMA-KAⁿZE-nǝ nipumbuka n-i OPOP-i-K-i-TAr-e-mba se-m-u sumbe-nǝ taⁿdǝki-wo sir-an-i siro-tapɛ-nǝ tasuki-wo kakɛ ma-so KA ŋGAMI te-ni tǝr-i-mot-i-te AMA-TU KAMÏ apuŋg-i-kǝp-i-nǝm-i KUNI-TU KAMÏ pus-i-te NUKA TUK-i kakar-aⁿz-u mo kakar-i mo KAMÏ-nǝ manima n-i tǝ TAT-I-aⁿzar-i ware KƏP-i-nǝm-ɛ-ⁿdǝ SIMASI-KU mǝ yǝ-k-eku pa na-si-ni YAKU-YAKU N-I katati tukupor-i ASA-NA ASA-NA ip-u kǝtǝ yam-i TAMA-KIPARU inǝti taye-nure TAT-i-woⁿdǝr-i ASI sur-i sakemb-i FUS-I APU ŋG-i mune ut-i naŋgɛk-i TE-ni MƏT-Er-u a-ŋga ko tǝmb-as-i-t-u YƏ-NƏ NAKA-NƏ MITI world-GEN person treasure-CONV-desire-ITER-ATTR seven-type-GEN treasure I TOP what do-TENT-FIN we-POSS inside-GEN be born(CONV)-exitPERF-ATTR white jewel-COMP we-POSS child Purupi TOP bright-star-GEN be bright-ATTR morning TOP spread-CONV-cloth-GEN bed-GEN side go away-NEG-FIN stand-PROG-EV-CONC sit-PROG-EV-CONC together DVCONV play-IMP evening-star-GEN evening-GEN-side DV-CONV becomeEV-CON INTER sleep-IMP DV hand-ACC take-CONV father-mother FP top TOP NEG-leave-CONV three-grass-COMP middle-LOC EP sleep-TENT-FIN beautiful-CONV you/he-POSS talk-ITER-EV-CON when EP EP person DV become-CONV-exit(CONV)-SUB bad-ATTR-NML FP good-ATTR-NML FP see-TENT-FIN DV big boat-COMP think-CONV-request-ATTR-LOC think-NEG-ATTR-LOC side-island-wind sudden DV-CONV cover-CONVcome-CONV-PERF/PROG-EV-CON do-TENT-ATTR way-GEN measure-ACC know-NEG-CONV white-cloth-GEN cord-ACC tuck up(CONV) INT-clear mirror hand-LOC take-CONV-hold-CONV-SUB heaven-GEN/LOC deity look up-CONV-request-CONV-pray-CONV land-GEN/LOC deity prostrateCONV-SUB forehead touch-CONV work-NEG-CONV FP work-CONV FP deity-GEN according DV-CONV DV stand-CONV-?-CONV I request-CONVpray-EV-CONC for a while-CONV good-ATTR-NML TOP no-FIN-LOC gradually-gradually DV-CONV facial form get emaciated-CONV morningPLUR morning-PLUR say-ATTR thing stop-CONV jewel-? life cease-PERF-EV stand-CONV-jump-CONV foot rub-CONV scream-CONV prostrate-CONV look up-CONV chest hit-CONV lament-CONV hand-LOC hold-PROG-ATTR I-POSS fly-CAUS-CONV-PERF-FIN world-GEN inside-GEN way What shall I do with the Seven Treasures that people of the world value and desire? Our child Purupi, born between us, who was like a white jewel, [you/he] did not leave the side of our white-clothed bed in the morning when the bright stars become pale. Whether standing or sitting, [you/he]

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

254

Chapter 4

played [with us]. When it became evening with the evening stars, and [we] told [you/him]: ‘Go to sleep!’; taking [our] hands, as you/he used to say beautifully: ‘Father and mother, do not leave [me]! [I] would like to sleep between you, [three of us], like ‘three [stems] grass,” [I] would like to see what bad things and what good things [were in store for you/him] when [you/he] becomes a [grown-up] person—thus [I] thought and prayed, [confident] as a big boat. But suddenly a[n evil] crosswind from the island came and covered [us]. Not knowing what to do, [I] tucked up the [sleeves of my kimono] with white cloth cords, and taking in [my] hands a clear mirror, [I] prayed looking up with a request to the deities of Heaven; and [I] prostrated [myself] before the deities of the Earth, touching [the floor] with [my] forehead. Although I stood and prayed so it would be according to the wishes of deities, whether [my prayers] work or not, for a while there was no improvement and [your/his] face get gradually thinner and thinner, and [you/he] stopped to say things in the mornings, and [your/his] jewel-like life ended. [I] jumped, and [I] rubbed my feet, and [I] screamed, [I] prostrated [myself], [I] looked up, and hit [myself] in the chest, and [I] lamented. [I] let my child, whom [I] was holding in [my] hands, fly away. [Alas, this is] the way of this world (MYS 5.904) Without exception, modern commentators of the Man’yōshū interpret si-ŋga in this poem as ‘he-POSS’ (Takagi et al. 1959: 119; Kojima et al. 1972: 117; Nakanishi 1978: 423; Satake et al. 1999: 524). And, it seems to me, that there is serious philological evidence supporting this interpretation. There is a poem (MYS 5.794) composed by the same poet, Yamonoue-no Okura, as an elegy for the death of Ōtomo-no Tabito’s wife: 大王能等保乃朝庭等斯良農比筑紫國爾泣子那須斯多比枳摩斯提伊企陀爾 母伊摩陀夜周米受年月母伊摩他阿良祢婆許許呂由母於母波奴阿比陀爾 宇知那毘枳許夜斯努礼伊波牟須弊世武須弊斯良爾石木乎母刀比佐氣斯 良受伊弊那良婆迦多知波阿良牟乎宇良売斯企伊毛乃美許等能阿礼乎婆 母伊可爾世与等可爾保鳥能布多利那良毘為可多良比斯許許呂曾牟企弖 伊弊社可利摩須

OPO KIMI-nǝ tǝpo n-ǝ MIKAⁿDO tǝ siranupi TUKUSI-NƏ KUNI-ni NAK-U KO-nasu sitap-i k-i-mas-i-te iki ⁿdani mǝ imaⁿda yasumɛ-ⁿz-u TƏSI TUKÏ mǝ imaⁿda ar-an-e-mba kǝkǝrǝ-yu mǝ omǝp-an-u apiⁿda-ni uti-nambik-i kǝy-as-i-n-ure ip-am-u sumbe se-m-u sumbe sir-an-i ipa-kï-wo mǝ top-i sakɛ sir-aⁿz-u ipe-n-ar-amba katati pa ar-am-u-wo uramesi-ki imo n-ǝ mikǝtǝ-nǝ are-womba mǝ ika n-i se-yǝ tǝ ka nipo-ⁿ-DƏRI-nǝ puta-ri naramb-i-wi katarap-i-si kǝkǝrǝ sǝmuk-i-te ipe-ⁿ-zakar-i-imas-u

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

255

NOMINALS

Great Lord-GEN far DV-ATTR Imperial.Court DV Tukusi-GEN land-LOC cryATTR child-COMP long-CONV come-CONV-HON-CONV-SUB breath RP FP yet rest-NEG-FIN year month FP yet be-NEG-EV-CON heart-ABL FP thinkNEG-ATTR interval-LOC PREF-stretched.out-CONV lie.down-HON-CONVPERF-EV say-TENT-ATTR way do-TENT-ATTR way know-NEG-CONV rock tree-ACC FP ask-CONV-split(NML) know-NEG-FIN home-LOC-exist-COND form TOP exist-TENT-ATTR-ACC dear/cruel-ATTR beloved DV-ATTR spiritGEN I-ACC(EMPH) FP how DV-CONV do-IMP DV IP grebe-DV(ATTR)-birdlike two-CL be.side.by.side-sit(CONV) talk-ITER-CONV-PAST/ATTR heart turn.back-CONV-SUB home-LOC-be.separated-CONV-HON-FIN [She] came longing like a crying baby to the land of Tukusi, which is called a distant Palace of the great Lord, not taking a rest even for one breath. But while years and months have not yet elapsed, while even in [my] heart [I] did not think, stretched out [she] was lying. [I] do not know what to say and what to do, nor do I know how to escape [my grief even if I] ask even the trees and rocks. If [you] were [still] at home, [your] form would be there [at least]! [My] dear and cruel beloved, what do [you] think [I] should do, like two grebes [we] swore to live together, but [you] left home, turning [your] back on that promise (MYS 5.794) The parallelism between MYS 5.904 and the first part of MYS 5.794 is too striking to ignore. Apparently, we have in both cases a lament about the deceased member of the family, and it is made in the third person. There is, however, a considerable difference as well: MYS 5.794 also has the second part, representing an incantation to the soul of the deceased, made in the second person, but MYS 5.904 has no analog of it: it ends with a description of the author’s grief. Still, additional evidence is needed to establish si as a third person pronoun in Old Japanese. Next comes combined linguistic and textual evidence from KK 57 and KK 101, which have very close but not identical contexts: 迦波能倍迩淤斐陀弖流佐斯夫袁佐斯夫能紀斯賀斯多迩淤斐陀弖流波毘呂 由都麻都婆岐斯賀波那能弖理伊麻斯芝賀波能比呂理伊麻須波淤富岐美 呂迦母

kapa-nǝ [u]pɛ-ni opï-ⁿdat-er-u sasimbu-wo sasimbu n-ǝ kï si-ŋga sita-ni opïⁿdat-er-u pa-m-birǝ yu t-u ma-tumbaki si-ŋga pana-nǝ ter-i-imas-i si-ŋga pa-nǝ pirǝr-i-imas-u pa opǝ-kimi rǝ kamǝ river-GEN top-LOC grow(CONV)-stand-PROG-ATTR sasimbu-ABS sasimbu DV-ATTR tree it-POSS bottom-LOC grow(CONV)-stand-PROG-ATTR

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

256

Chapter 4

leaf-GEN-broad sacred DV-ATTR INT-camellia it-POSS leaf-GEN shineCONV-HON-CONV it-POSS leaf-GEN be.broad-CONV-HON-ATTR TOP greatlord DV EP under the sasimbu tree, growing above the river, below it [there is] a sacred true camellia growing, its flowers are shining, its leaves are broad, as the great lord! (KK 57) It seems impossible to interpret all si-ŋga in this poem as the second person ‘you-POSS’: we will get a nonsense text with second person pronouns referring to two different trees, like ‘sasimbu tree, below you grows camellia, your flowers and leaves’—whom does second ‘your’ refer to? In addition, the next passage, also from the Kojiki kayō, proves almost without any doubt that we are dealing with the third person pronoun: 淤斐陀弖流波毘呂由都麻都婆岐曾賀波能比呂理伊麻志曾能波那能弖理伊 麻須多加比迦流比能美古爾登余美岐多弖麻都良勢

opï-ⁿdat-er-u pa-m-birǝ yu t-u ma-tumbaki sǝ-ŋga pa-nǝ pirǝr-i-imas-i sǝ-nǝ pana-nǝ ter-i-imas-u taka-pikar-u pi-nǝ mi-ko-ni tǝyǝ mi-ki tatematur-as-e grow(CONV)-stand-PROG-ATTR leaf-GEN-broad sacred DV-ATTR INT-camellia it-POSS leaf-GEN be.broad-CONV-HON-CONV it-GEN flower-GEN shineCONV-HON-FIN high-shine-ATTR sun-GEN HON-child-DAT abundant HON-rice.wine present(HUM)-HON-IMP growing broad-leafed sacred true camellia, its leaves are broad, its flowers are shining. Present the abundant rice wine to the honorable child of the highshining Sun (KK 101) Looking at this text, one can clearly see that si-ŋga appears to be a textual variant of sǝ-ŋga, a third person pronoun possessive form also attested in Classical Japanese (Vovin 2003: 114), which is derived from the demonstrative sǝ- ‘that.’ Moreover, there are the following two examples in the Kojiki kayō, which have si-ŋga that can be interpreted only as a third person pronoun: 加良怒袁志本爾夜岐斯賀阿麻里許登爾都久理

Karano-wo sipo-ni yak-i si-ŋga amar-i kǝtǝ-ni tukur-i Karano-ACC salt-LOC burn-CONV it-POSS remain-NML koto-LOC make-CONV [They] burned [the ship] Karano for salt, and made its remainders into a koto (KK 74) It would be a complete folly to imagine that the ship is addressed in the second person in this example.

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

257

NOMINALS 意布袁余志斯毘都久阿麻余斯賀阿禮婆宇良胡本斯祁牟斯毘都久志毘

op[ǝ]-uwo yǝsi simbi tuk-u ama yǝ si-ŋga ar-e-mba ura-koposi-k-em-u simbi tuk-u Simbi big-fish ? tuna harpoon-ATTR fisherman EP she-POSS go away-EV-CON inside-sorrowful-ATTR-TENT-FIN tuna harpoon-ATTR Simbi Oh, fisherman, who harpoons tuna, the big fish! [You] will be sorrowful when she gets away, Simbi who harpoons tuna (KK 110) If we are interpret si-ŋga ar-e-mba as ‘when you get away,’ the poem would not make any sense at all. The final and the most conclusive evidence for the fact that si represents the third, and not the second person pronoun comes from the Senmyō texts, where it is written not only phonographically, but also logographically using the Old Chinese oblique third person pronoun 其 ‘he, she, it’: 先仁之我奏之事

SAKI-ni si-ŋga MAWOS-I-si KƏTƏ before-LOC he-POSS say(HUM)-CONV-PAST/ATTR thing things he said before (SM 28) 清麻呂其我姉法均止甚大尓悪久奸流妄語乎作弖

KIYOMARƏ SI-ŋga ANE POPUKUN-tǝ ITƏ OPO-KI n-i ASI-ku KAⁿDAM-Er-u ITUPAR-I- ŋ-GƏTƏ-wo TUKUR-I-te Kiyomarǝ he-POSS elder.sister Popukun-COM very big-ATTR DV-CONV badCONV be.insincere-PROG-ATTR lie-NML-GEN- word-ACC make-CONV-SUB Kiyomarǝ with his elder sister Popukun created an extremely big, bad and insincere lie … (SM 44) There is also one example in late Old Japanese, where si also can be only interpreted as the third person pronoun: 老人毛女童兒毛之我願心太良比爾

OYI-PITƏ mo WOMINA WARAPA mo si-ŋga NE ŋG-AP-U KƏKƏRƏ-ⁿ-dar-ap-i n-i old(CONV)-person FP woman child FP they-POSS desire-ITER-ATTR heartLOC-be.enough-ITER-NML DV-CONV old people, women, and children, all [of them get] enough what they desire to [their] hearts’ [content] (MYS 18.4094)

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

258

Chapter 4

2.2.8.1 Special Forms Besides the special form sǝ-ŋga, mentioned above in KK 101, there is also a special form sǝ-ne-ŋga and sǝ-ne, attested just once each in the same text. These forms appear to be parallel to the second person pronoun special form na-ne (see section 2.2.5): 曾泥賀母登曾泥米

sǝ-ne-ŋga mǝtǝ sǝ-ne mɛ it-?-POSS root it-? bud its roots, its buds (KK 11)71 It seems that none of the examples attested in Old Japanese offers uncontroversial support for si as the second person pronoun; and quite to the contrary, most of the attested examples can be safely interpreted as the cases of the third person pronoun. There is one example from an early Middle Japanese text, known to me, that might be an evidence for Miller’s point of view on si as a second person pronoun, but it falls out of the scope of the grammar dedicated to Old Japanese: しが身の程知らぬこそいと心憂けれ

si-ga mi-no fodo sir-an-u koso ito kokoro u-kere you-POSS body-GEN time know-NEG-ATTR FP very heart sad-EV [I] feel very sad to the extent you never felt (OM 1) The construction si-ga mi ‘your body’ (= ‘you’) seems to be perfectly parallel to the construction wa-ga mi ‘my body’ (= ‘I’) that frequently appears in Middle Japanese texts in the function of the first person pronoun. However, a singular attestation of si as a second person pronoun in this example does not allow us to reconstruct the second person pronoun si for the earlier stages of the Japanese language. Comparative Data Level A: Other Japonic A1: Eastern Old Japanese There is no second or third person pronoun si in Eastern Old Japanese, but the genitive form sǝ-nǝ ‘it-GEN’ of third person sǝ- ‘it’ is attested once, and there 71  The same song is attested in NK 13 with the phonetic variant sǝ-nǝ-ŋga mǝtǝ sǝ-ne mɛ.

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

259

NOMINALS

is also one example of the accusative form sǝ-wo, which does not occur in Western Old Japanese. 曾能可抱与吉爾

sǝ-nǝ kapo yǝ-ki-ni it-GEN face good-ATTR-LOC its face is good, but [it does not move forward] (MYS 14.3411) 比登豆麻等安是可曾乎伊波牟

pitǝ-ⁿ-duma tǝ aⁿze ka sǝ-wo ip-am-u person-GEN-wife DV why IP she-ACC say-TENT-ATTR Why should [I] call her the wife of another? (MYS 14.3472) 2.3

Reflexive Pronoun onǝ/onǝre

There is only one reflexive pronoun in Old Japanese: onǝ/onǝre, which can be used as a reflexive involving any person: first, second, or third. The forms of the reflexive pronoun onǝ/onǝre seem to be parallel morphologically (although not exactly) to personal pronouns wa/ware, a/are, and na/nare: the unextended stem onǝ occurs almost exclusively in combination with the following possessive case marker -ŋga, although there are two examples of it occurring in isolation found in the same text. The unextended stem onǝ occurs in Western Old Japanese only six times, and the extended stem onǝre is even rarer (only two examples of phonographic spelling are attested) and it occurs only in isolation.72 None of the case markers are attested after onǝ/onǝre except possessive -ŋga after the unextended stem onǝ.

72  In the early Heian texts both stems of the personal-reflexive pronoun ono and onore are used, while in the later Heian texts there is a distinct preference for the extended stem onore, the stem ono- being used very rarely. In the Taketori monogatari possessive forms ono-ga and onore-ga co-occur (Vovin 2003: 102). The Heian period personal-reflexive pronoun onodukara, derived from onǝ, is not attested in phonetic spelling in Old Japanese texts, therefore, it is excluded from this grammar. Another personal reflexive pronoun, midukara, becomes prominent in Classical Japanese, especially in the later Heian texts, but it occurs in Old Japanese texts only in later glosses, and never appears in the texts themselves, therefore it is excluded from the present description.

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

260

Chapter 4

2.3.1 Stem onǝ 2.3.1.1 Isolated Form onǝ 於乃毛於乃毛貞仁能久清伎心乎以天奉仕

onǝ mo onǝ mo SAⁿDAKA n-i YƏ-ku KIYO-ki KƏKƏRƏ-wo MOT-I-teTUKAPƐMATUR-E yourself FP yourself FP truthful DV-CONV good-CONV clear-ATTR heart-ACC hold-CONV-SUB serve(CONV)-HUM-IMP All of you, serve truthfully and well, with a clear heart (SM 33) 2.3.1.2 Possessive onǝ-ŋga 意能賀袁袁奴須美斯勢牟登

onǝ-ŋga wo-wo nusum-i si-se-m-u tǝ yourself-POSS cord-ACC steal-CONV die-CAUS-TENT-FIN DV [they] are going to steal your [own life-]cord and murder [you] (KK 22)73 意乃何身志伊多波斯計礼婆

onǝ-ŋga MÏ74 si itapasi-kere-mba myself-POSS body EP ill-EV-CON as my [own] body became ill (MYS 5.886) 於能我於弊流於能我名負弖

onǝ-ŋga op-er-u onǝ-ŋga NA OP-i-te himself-POSS carry-PROG-ATTR himself-POSS name carry-CONV-SUB [everyone of] themselves carrying their own names (MYS 18.4098) 人祖乃意能賀弱兒乎養治事乃

PITƏ N-Ə OYA-nǝ onǝ-ŋga WAKU-ŋ-GO-wo PITAS-U KƏTƏ-nǝ person DV-ATTR parent-GEN themselves-POSS young-DV(ATTR)-child-ACC rear-ATTR thing-COMP like human parents rear their own young children (SM 3)

73  A variant of this poem also appears in NK 18 as onǝ-ŋga wo-wo si-se-m-u tǝ nusum-aku ‘the fact that [they] will steal with intention to kill your [own life-] cord.’ 74  Ono-ga mi occurs as a personal reflexive pronoun in early Heian texts (Vovin 2003: 104); however, there is just a singular example in Old Japanese in this context that does not allow to treat the OJ construction onǝ-ŋga mï as a grammaticalized pronoun.

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

261

NOMINALS

2.3.2 Stem onǝre 2.3.2.1 Isolated Form onǝre 於能礼故所詈而居者

onǝre YUWE NOR-AYE-TE WOR-E-mba myself reason scold-PASS(CONV)-SUB exist-EV-CON As [you] were scolded because of me (MYS 12.3098) 伊夜彦於能礼神佐備

IyaPIKO onǝre KAMU-sambï Iyapiko itself deity-like [The mountain] Iyapiko itself [is] deity-like (MYS 16.3883) Comparative Data Level A: Other Japonic A2: Ryukyuan It seems that a possible cognate to OJ onǝ/onǝre is attested only in the Shuri literary language as ’uga (RKJ 1983: 541), and in Old Ryukyuan as oga (おが、おか) (Hokama 1995: 123; Nakahara and Hokama 1978: 73–74). However, this limited distribution, along with its predominant function as a second person and not a reflexive pronoun (it has the reflexive function only in the Shuri literary language), as well as its peculiar phonetic shape ’uga < *onǝ-ŋga, with no other morphological variants attested, strongly suggest that in Ryukyuan it is a loan from Classical Japanese, and not a genuine cognate dating back to Japonic. 2.4

Demonstrative Pronouns

Demonstrative pronouns in Old Japanese show considerable morphological similarity to the major personal pronouns and the reflexive pronoun, since all of them also have unextended stem and extended stem in -re. However, there are also two significant differences: first, demonstrative pronouns do not allow the possessive case marker -ŋga to be added to an unextended stem. They have a modifier form in -nǝ, which looks like an unextended stem, followed by a genitive case marker -nǝ, but in reality it is not the genitive case marker, but the attributive form n-ǝ of the defective verb n-.75 Thus, e.g., historically proximal modifier form kǝnǝ < *kǝ n-ǝ ‘this DV-ATTR.’ Second, they do not have the oblique stem 75  There is also a modifier form kǝre n-ǝ, although it is used much less frequently than kǝnǝ. Neither *sǝre n-ǝ nor *kare n-ǝ are attested in Old Japanese texts.

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

262 chart 15

Chapter 4 Demonstrative pronouns

isolated extended modifier place place/direction

Proximal

Mesial

Distal

kǝ kǝre kǝnǝ kǝkǝ kǝti*, kǝnata*

sǝ*⁺ sǝre sǝnǝ sǝkǝ —

ka* kare* kanǝ** — kanata*, woti*, wotǝ*, wote

Note: * – only in Western Old Japanese, ** – only in Eastern Old Japanese, ⁺ – only in combination with the following accusative marker.

marker -ⁿ. Similar to modern and Classical Japanese, demonstrative pronouns in Old Japanese can be subdivided into three groups: proximal, mesial, and distal. However, in contrast to modern and Classical Japanese, these three groups do not seem to be connected to the speaker/addressee axis, denoting just general levels of proximity: proximal, mesial, and distal. 2.4.1 Proximal Demonstrative Pronouns It is considered that proximal demonstrative pronouns denote something that is closer to the speaker than to his/her addressee, however this speakeroriented function becomes apparent only in Classical Japanese. The usage in Old Japanese seems to provide evidence only for the general proximity, without any regard to a speaker or an addressee’s position. All pronouns of this group incorporate as their first element the stem kǝ, the unextended stem of the proximal demonstrative pronoun kǝ/kǝre. 2.4.1.1 Proximal Demonstrative Pronoun kǝ/kǝre The proximal demonstrative pronoun kǝ/kǝre indicates that something is located in close proximity. In contrast to Classical Japanese, the unextended stem kǝ occurs in Old Japanese more frequently than the extended stem kǝre, and not only in the modifying form kǝnǝ, but also in isolation and with the following case markers. Furthermore, the unextended stem kǝ never occurs in plural usage, but there is one example when the extended stem kǝre has a plural meaning (see MYS 15.3638 below). This is quite reminiscent of the similar usage of the extended stems ware and are of the personal pronouns wa/ware and a/are described above (see sections 2.2.1 and 2.2.2). Therefore, this plural usage of kǝre can be used as further evidence in the favor of the hypothesis

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

263

NOMINALS chart 16

Proximal demonstrative pronoun kǝ/kǝre

isolated modifier accusative ablative

Unextended stem kǝ

Extended stem kǝre

Kǝ⁺ S kǝnǝ S kǝ-womba kǝ-yu*, kǝ-yo* S

kǝre* SP kǝre n-ǝ S kǝre-wo* S —

Note: S – singular, P – plural, * – only in Western Old Japanese, ⁺ – can also be used as a modifier

outlined in section 2.2.1 that the extending suffix -re originally incorporated plural marker -ra. The chart above summarizes the distribution of both the unextended stem kǝ and the extended stem kǝre in isolation, modifier form, and with the following case markers in Western and Eastern Old Japanese. 2.4.1.1.1 Stem kǝ 2.4.1.1.1.1

Isolated Form kǝ

Isolated form kǝ is attested in Western Old Japanese in three examples in the Kojiki kayō with the following particles, and once in the Nihonshoki kayō, modifying the following noun yǝpi ‘night.’ Both usages survived into Classical Japanese, but there they become even more infrequent. In modern Japanese this isolated form survived only in compounds like kotosi ‘this year’ (OJ kǝ tǝsi). 許母布佐波受

kǝ mǝ pusap-aⁿz-u this FP be suitable-NEG-FIN This is not suitable, either (KK 4) 許斯與呂志

kǝ si yǝrǝsi this EP be good this is good (KK 4) 許斯母阿夜爾加志古志

kǝ si mǝ aya n-i kasiko-si this EP FP very DV-CONV be awesome-FIN This is very awesome, too (KK 100)

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

264

Chapter 4

区茂能於虚奈比虚予比辞流辞毛

kumo-nǝ okǝnap-i kǝ yǝpi siru-si mo spider-GEN perform-NML this night be distinctive-FIN EP the spider’s performance is distinctive tonight (NK 65) In this example kǝ has a modifier function without the following n-ǝ. 2.4.1.1.1.2

Modifier Form kǝnǝ

許能登理母宇知夜米許世泥

kǝnǝ tǝri mǝ uti-yamɛ-kǝse-n-e this bird FP PREF-stop(CONV)-BEN-DES-IMP [I] wish [you] would stop [the singing] of these birds (KK 2) 許能美岐波和賀美岐那良受

kǝnǝ mi-ki pa wa-ŋga mi-ki nar-aⁿz-u this HON-rice.wine TOP I-POSS HON-rice.wine be-NEG-FIN This rice wine is not my rice wine (KK 39) 許能迦迩夜伊豆久能迦迩

kǝnǝ kani ya iⁿduku-nǝ kani this crab IP where-GEN crab This crab, where it is from? (KK 42) 許能久斯美多麻

kǝnǝ kusi mi-tama this precious HON-jewel these precious jewels (MYS 5.814) 許能安我家流伊毛我許呂母能阿可都久見礼婆

kǝnǝ a-ŋga ker-u imo-ŋga kǝrǝmǝ-nǝ aka tuk-u MI-re-mba this I-POSS wear(PROG)-ATTR beloved-POSS garment-GEN dirt attach-ATTR see-EV-CON when [I] see that this garment of [my] beloved which I wear became dirty (MYS 15.3667)

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

265

NOMINALS 2.4.1.1.1.3

Emphatic Accusative kǝ-womba

加多理其登母許遠婆

katar-i-ŋ-gǝtǝ mǝ kǝ-womba talk-NML-GEN-thing FP this-ACC(EMPH) [this] story, too, this [one] (KK 2; also in KK 3 (twice), KK 4, KK 100, KK 101, KK 102) Since -womba occurs after kǝ in the same context, it probably suggests a rudimental usage. 2.4.1.1.1.4 Ablative kǝ-yo and kǝ-yu

There are only three examples of the unextended stem kǝ, used with the following ablative case marker -yu/-yo. All three occur in very similar contexts, that probably suggests a rudimental usage. In all three cases kǝ refers to a place and not an object, thus having a usage similar to kǝkǝ ‘here.’ 保登等伎須和我須武佐刀爾許欲那伎和多流

potǝtǝŋgisu wa-ŋga sum-u sato-ni kǝ-yo nak-i watar-u cuckoo I-POSS live-ATTR village-LOC this-ABL cry-CONV cross-FIN the cuckoo crosses over, crying, from here to the village where I live (MYS 15.3783) 許由奈伎和多礼

kǝ-yu nak-i watar-e this-ABL cry-CONV cross-IMP [Oh, cuckoo,] cross from here, crying (MYS 18.4035) 保等登藝須許欲奈枳和多礼

potǝtǝŋgisu kǝ-yo nak-i watar-e cuckoo this-ABL cry-CONV cross-IMP [Oh,] cuckoo, cross from here, crying (MYS 18.4054) 2.4.1.1.2 Stem kǝre

2.4.1.1.2.1

Isolated Form kǝre

許禮婆布佐波受

kǝre pa pusap-aⁿz-u this TOP be suitable-NEG-FIN This is not suitable (KK 4)

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

266

Chapter 4

巨礼也己能名爾於布奈流門能宇頭之保爾多麻毛可流登布安麻乎等女杼毛

kǝre ya kǝnǝ NA-ni op-u NaruTO-nǝ uⁿdu-sipo-ni tama-mo kar-u tǝ [i]p-u ama wotǝme-ⁿdǝmǝ This IP this name-LOC carry-ATTR Naruto-GEN whirl-pool-LOC jewelseaweed cut-ATTR DV fisher maiden-PLUR Are these the fisher-maidens who are said to cut precious seaweeds in the whirlpool bearing this name of Naruto? (MYS 15.3638) 波里夫久路己礼波多婆利奴

pari-m-bukuro kǝre pa tambar-i-n-u needle-GEN-bag this TOP receive(HUM)-CONV-PERF-FIN Needle bag—this [I] have received (MYS 18.4133) 2.4.1.1.2.2

Modifier Form kore n-ǝ

Only three examples of the modifier form kǝre n-ǝ are attested in Western Old Japanese texts: 許礼能水嶋

kǝre n-ǝ MIⁿDUSIMA this DV-ATTR Miⁿdusima This Miⁿdusima (MYS 3.245) 己礼乃与波宇都利佐留止毛

kǝre n-ǝ yǝ pa utur-i-sar-u tǝmo this DV-ATTR world TOP change-CONV-go away-FIN CONJ Even though this world changes and goes away … (BS 10) 己礼乃微

kǝre n-ǝ mï this DV-ATTR body this body (BS 20) 2.4.1.1.2.3 Accusative kǝre-wo 許礼乎於伎低麻多波安里我多之

kǝre-wo ok-i-te mata pa ar-i-ŋ-gata-si this-ACC leave-CONV-SUB again TOP exist-NML-GEN-be.hard-FIN [It] is difficult to find again [a falcon of the same quality], except this [one] (MYS 17.4011)

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

267

NOMINALS

Comparative Data Level A: Other Japonic A1: Eastern Old Japanese In Eastern Old Japanese the distribution of the unextended stem kǝ is more limited than in Western Old Japanese: it occurs only in isolation and in modifier form kǝnǝ, but that can probably be explained by the limited amount of the Eastern Old Japanese corpus. The extended stem kǝre, on the other hand, occurs just once, and this seems quite consistent with limited usage of the extended stems of other pronouns in Eastern Old Japanese. Ultimately, this is additional oblique evidence for the secondary nature of the extended stem, which likely was originally a plural form. Stem kƏ Isolated Form kƏ

Isolated form kǝ occurs in Eastern Old Japanese only once as a modifier of the following word yǝpi ‘night’: 和乎可麻都那毛伎曾毛己余必母

wa-wo ka mat-unam-wo kisǝ mo kǝ yǝpi mǝ I-ACC IP wait-TENT2-ATTR last night FP this night FP will [she] have waited for me, both last night and tonight? (MYS 14.3563) Modifier Form kƏnƏ 伎美我見延奴己能許呂

kimi-ŋga MI-ye-nu kǝnǝ kǝrǝ lord-POSS see-PASS-NEG-ATTR this time lately (lit.: this time), when [I] cannot see [my] lord (MYS 14.3506) Stem kƏre Modifier Form kƏre n-Ə

There is one example in Eastern Old Japanese of the modifier form kǝre n-ǝ: 比毛多要婆安我弖等都氣呂許礼乃波流母志

pimo taye-mba a-ŋga te-tǝ tukɛ-rǝ kǝre n-ǝ paru mǝs-i cord tear-COND your.own-POSS hand-COM attach-IMP this DV-ATTR needle hold-CONV if the cords [of your garment] tear, attach them with your own hand, holding this needle (MYS 20.4420)

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

268

Chapter 4

A2: Ryukyuan In Ryukyuan, only the extended stem kuri and modifier form kunu are attested, including Old Ryukyuan. The unextended stem *ku does not appear in isolation: it is found only in the modifier form kunu and in the proximal demonstrative pronoun kuma ‘here,’ indicating the place. This situation is typical for both Old Ryukyuan and modern languages. Examples: Old Ryukyuan これはつにしや

kore facu nisi ya this first northern wind COP this is the first northern wind (OS 13.899) ちゑねんもりくすくこの世まさり

tiwenenmori kusuku kono YO masar-i Tiwenenmori castle this world surpass-CONV Tiwenenmori castle surpasses [all in] this world … (OS 19.1302) Shuri

kuree nuu yaibii ga this(TOP) what is(POL) IP What is this? (Nishioka and Nakahara 2000: 44) kunu maNgaa umu-sa-N doo this cartoon be interesting-NML-FIN FP this cartoon is interesting (Nishioka and Nakahara 2000: 18) Thus, there is sufficient internal evidence for reconstruction of proto-Japonic proximal demonstrative pronoun *kǝ ‘this.’ 2.4.1.2 Proximal Demonstrative Pronoun kǝkǝ The proximal demonstrative pronoun kǝkǝ ‘here’ is a combination of the unextended stem kǝ- of the proximal demonstrative pronoun kǝ/kǝre and the obsolete word *kǝ ‘place.’76 It is basically a stative demonstrative pronoun, showing location, not direction, but it can also be used as a directional demonstrative 76  It is possible that *kǝ ‘place’ in kǝkǝ ‘here’ and sǝkǝ ‘there’ is a result of progressive vowel assimilation, cf. -ko ‘place’ with kō-rui vowel /o/ in miya-ko ‘capital (lit.: palace-place),’ where the conditions for such progressive assimilation are not present; as well as -ka Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

NOMINALS

269

when followed by the locative/directive case marker -ni, although this usage is rare in comparison to Classical Japanese. Examples: 母登弊波岐美袁淤母比傅須惠幣波伊毛袁淤母比傅伊良那祁久曾許爾淤母 比傅加那志祁久許許爾淤母比傅

mǝtǝ-pe pa kimi-wo omǝp-i-[i]ⁿde suwe-pe pa imo-wo omǝp-i-[i]ⁿde irana-k-eku sǝkǝ-ni omǝp-i-[i]ⁿde kanasi-k-eku kǝkǝ-ni omǝp-i-[i]ⁿde root-side TOP lord-ACC think-CONV-exit(CONV) top-side TOP beloved-ACC think-CONV-exit(CONV) be.regrettable-ATTR-NML there-LOC think-CONVexit(CONV) be.sorrowful-ATTR-NML here-LOC think-CONV-exit(CONV) at the root [of the tree, I] remember [my] lord, at the top [of the tree, I] remember [my] beloved, [I] remember [my lord] there with a regret, [I] remember [my beloved] here with sorrow (KK 51)77 許己乎志毛間細美香母

kǝkǝ-wo si mo ma-KUPASI-mi kamǝ here-ABS EP FP INT-beautiful-GER EP It is beautiful here! (MYS 13.3234) 安我未許曾世伎夜麻故要弖許己爾安良米許己呂波伊毛爾与里爾之母能乎

a-ŋga mï kǝsǝ seki yama koye-te kǝkǝ-ni ar-am-ɛ kǝkǝrǝ pa imo-ni yǝr-i-n-isi mǝnǝwo I-POSS body FP barrier mountain cross(CONV)-SUB here-LOC exist-TENT-EV heart TOP beloved-DAT approach-CONV-PERF-PAST/ATTR CONJ My body has crossed barriers and mountains, and is probably here. But [my] heart is near [my] beloved! (MYS 15.3757) 可奈之家口許己爾思出伊良奈家久曾許爾念出

kanasi-k-eku kǝkǝ-ni OMƏP-i-[I]ⁿDE irana-k-eku sǝkǝ-ni OMƏP-I-[I]ⁿDE be sorrowful-ATTR-NML here-LOC think-CONV-exit(CONV) be regrettableATTR-NML there-LOC think-CONV-exit(CONV) [I] remember [the parting with you] here with sorrow, [I] remember [the parting with you] there with a regret (MYS 17.3969)

‘place’ in ar-i-ka ‘place to be’ and sum-i-ka ‘place to live,’ where -ka can be a truncation of *-kau (> -ko). 77  N K 43 represents a very close textual variant of this poem.

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

270

Chapter 4

保等登藝須許許爾知可久乎伎奈伎弖余

potǝtǝŋgisu kǝkǝ-ni tika-ku-wo k-i nak-i-te-yǝ cuckoo here-LOC be close-CONV-ACC come-CONV cry-CONV-PERF-IMP Cuckoo! Come close to here, and cry! (MYS 20.4438) This is the only example attested in Western Old Japanese where kǝkǝ with the following locative/directive case marker -ni has a function of a directional demonstrative pronoun. In the following four examples kǝkǝ retains much of its original meaning ‘this place’ or ‘this location’: 許己念者胸許曾痛

kǝkǝ [O]MƏP-Ɛ-mba MUNE kǝsǝ ITAM-Ɛ here think-EV-CON chest FP is.painful-EV when [I] think of this situation, my heart hurts (MYS 8.1629) 許己乎之母安夜爾多敷刀美宇礼之家久伊余与於母比弖

kǝkǝ-wo si mǝ aya n-i taputo-mi uresi-k-eku iyǝyǝ omǝp-i-te here-ABS EP FP extreme DV-CONV be venerable-GER be glad-ATTR-NML more.and.more think-CONV-SUB thinking with joy more and more how extremely venerable our position is (MYS 18.4094) 許己乎之母安夜爾久須之弥

kǝkǝ-wo si mo aya n-i kususi-mi here-ABS EP FP extreme DV-CONV be.wonderful-GER this situation is extremely wonderful, and … (MYS 18.4125) 許己見礼婆宇倍之神代由波自米家良思母

kǝkǝ MI-re-mba umbɛ-si KAMÏ-YƏ-yu paⁿzimɛ-ker-asi-mǝ here look-EV-CON be.indeed-FIN deity-age-ABL begin(CONV)-RETR-SUP-EXCL when [you] look at this place, it indeed looks like [they] began [building of the palaces] from the Age of Gods! (MYS 20.4360)78 78  Opinions differ, whether this poem belongs to a border-guard, or to Ōtomo-no Yakamochi himself. The latter seems more probable, since the poem is preceded by the line in Chinese: 陳私拙懐一首 ‘a poem stating my own humble thoughts,’ and also because it is written in perfect Western Old Japanese, without any elements typical for Eastern Old Japanese. Therefore, I treat it as a Western Old Japanese text.

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

NOMINALS

271

Comparative Data Level A: Other Japonic A1: Eastern Old Japanese There is just one example of the proximal demonstrative pronoun kǝkǝ in Eastern Old Japanese: 己許呂能未伊母我理夜里弖和波己許爾思天

kǝkǝrǝ nǝmï imǝ-ŋgari yar-i-te wa pa kǝkǝ-ni s-i-te heart RP beloved-DIR send-CONV-SUB I TOP here-LOC do-CONV-SUB I [have] to stay here, sending just [my] heart to [my] beloved (MYS 14.3538) A2: Ryukyuan There are no direct cognates in Ryukyuan of the Old Japanese proximal demonstrative pronoun of place kǝkǝ, since the Ryukyuan counterpart of OJ kǝkǝ is kuma, consisting of the proximal demonstrative ku < *kǝ and ma ‘interval.’ Ryukyuan kuma is amply attested in Old Ryukyuan, and in all modern languages, except Hateruma (Uchima and Arakaki 2000: 397), so apparently it goes back to proto-Ryukyuan, as well as OJ kǝkǝ goes back to proto-Japanese. Thus, both Japanese kǝkǝ and Ryukyuan kuma seem to be late independent formations in two branches of Japonic, and it is not possible to reconstruct a common proto-Japonic archetype. 2.4.1.3 Proximal Demonstrative Pronoun kǝti The proximal demonstrative pronoun kǝti in Western Old Japanese indicates the place, and not the direction, contrary to what might have been expected on the basis of its etymology. It tends to occur in the set phrase woti kǝti ‘there [and] here,’ or in the reduplicated form kǝti-ŋgǝti ‘here [and] there (lit.: here and here).’ Etymologically kǝti probably goes back to the proximal demonstrative pronoun kǝ ‘this’ plus ti ‘way,’ which is attested exclusively in compounds, like mi-ti ‘HON-road’ or ti-mata ‘road-fork’ (see section 1.1.1).79

79  As far as internal evidence is concerned, this *ti ‘road’ is probably from otsu-rui *tï: the distal demonstrative pronoun woti ‘the place over there’ occurs in compounds as wotǝ-, which suggests that OJ ti < *tǝ-[C]i.

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

272

Chapter 4

許知能夜麻登多多美許母弊具理能夜麻能許知碁知能夜麻能賀比爾

kǝti-nǝ yama-tǝ tatami-kǝmǝ peŋguri-nǝ yama-nǝ kǝti-ŋgǝti-nǝ yama-nǝ kapi-ni here-GEN mountain-COM mat-rush Peguri-GEN mountain-GEN here-hereGEN mountain-GEN interval-LOC between the mountains here [and] there, the mountain here and the Peguri mountain [which is like] a rush mat (KK 91) 己知其智乃國之三中從

kǝti-ŋgǝti-nǝ KUNI-NƏ mi-NAKA-YU here-here-GEN province-GEN HON-middle-ABL from the provinces here [and] there (MYS 3.319) 己智其智乃花

kǝti-ŋgǝti-nǝ pana here-LOC-here-GEN flower flowers here [and] there (MYS 9.1749) 和可伎兒等毛波乎知許知爾佐和吉奈久良牟

waka-ki KO-ⁿdǝmo pa woti kǝti-ni sawak-i-nak-uram-u young-ATTR child-PLUR TOP there here-LOC make. noise-CONV-cry-TENT2-FIN young children will probably cry loudly here [and] there (MYS 17.3962) 都麻母古騰母毛乎知己知爾左波爾可久美為

tuma mǝ ko-ⁿdǝmǝ mo woti kǝti-ni sapa n-i kakum-i wi spouse FP child-PLUR FP there here-LOC many DV-CONV surround-CONV exist(CONV) both my spouse and children are around [me] in great numbers here and there (MYS 20.4408) Comparative Data Level A: Other Japonic There are no examples of the proximal demonstrative pronoun kǝti in Eastern Old Japanese or Ryukyuan. Thus, kǝti appears to be a pure Western Old Japanese formation, and a reconstruction of the Japonic archetype is not feasible. 2.4.1.4 Proximal Demonstrative Pronoun kǝnata The proximal demonstrative pronoun kǝnata ‘here,’ ‘this side’ originated as a result of a contraction of the modifier form kǝnǝ of the proximal demonstrative

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

NOMINALS

273

pronoun kǝ/kǝre- and the word kata ‘side,’ ‘direction.’ It is not attested in the phonographic script in Western Old Japanese, but there is one example in the logographic script that allows us to read the characters 此方 as kǝnata and not kǝti due to the meter of the poem: 壯士墓此方彼方二造置有故

WOTƏKƏ-ⁿ-DUKA KƏNATA KANATA-ni TUKUR-I-OK-ER-U YUWE [young] man-GEN-grave here there-LOC make-CONV-put-PROG-ATTR reason because the graves of [young] men were built here [and] there (MYS 9.1809) Comparative Data Level A: Other Japonic A2: Ryukyuan It seems that the proximal demonstrative pronoun kogata < *kun[u] kata < *kǝnǝ kata is attested only in Old Ryukyuan: although RKJ lists an attestation in Shuri, the example provided there (RKJ 1983: 327) is from the Ryūka given below: 恩納嶽あがた里が生まれ島森もおしのけてこがたなさな

UNNA-DAKI agata SATU-ga Umare-JIMA MUI mo os-i-noke-te kogata nas-ana Unna-peak over there village-POSS born(CONV)-island mountain FP pushCONV-put away(CONV)-SUB here make-DES On the other side of the peak Unna [there is] the village on the island where [my beloved] was born. Pushing the mountain aside, [I] would like to place [it] here (RK 1243) Isolated attestations in Old Ryukyuan not supported by the modern dialect data are always suspicious, but in this particular case OR kogata cannot be a loan from Classical Japanese konata (< OJ kǝnata), since the patterns of contraction are different in Japanese (kǝnata < *kǝn[ǝ k]ata) and in Ryukyuan (kugata < *kun[u] kata < *kǝnǝ kata). Thus, it is possible to reconstruct protoJaponic phrase *kǝnǝ kata ‘this side’ with a deictic usage ‘here.’ 2.4.2 Mesial Demonstrative Pronouns Similar to proximal demonstrative pronouns, mesial demonstrative pronouns are considered to denote that something is closer to the addressee than to the speaker, however this addressee-oriented function becomes apparent only in Classical Japanese. The usage in Old Japanese seems to provide the evidence only for the generally based remoteness, without any regard to a speaker or

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

274

Chapter 4

an addressee’s position. All pronouns of this group incorporate as their first element the stem sǝ, the unextended stem of the proximal demonstrative pronoun sǝ/sǝre. In contrast to the proximal demonstratives with kǝ, which have the full set of pronouns, there are no mesial demonstratives *sǝti and *sǝnata attested in Old Japanese texts. 2.4.2.1 Mesial Demonstrative Pronoun sǝ/sǝre The mesial demonstrative pronoun sǝ/sǝre indicates that something is relatively remote in a general sense, without reference to a speaker/addressee axis. This pronoun can also have an anaphoric usage, referring to something previously mentioned in the text. The forms of the mesial demonstrative pronoun sǝ/sǝre ‘that’ occur in Old Japanese with a very skewed distribution. The most frequent form is the modifier form sǝnǝ,80 based on the unextended stem sǝ. Other usages of the unextended stem sǝ are extremely rare: it never occurs in isolation,81 and the only example with the following case marker is with the accusative case marker -wo attested twice in Western Old Japanese. The extended stem sǝre is also very rare, and what is more important, it occurs only once in the phonographic spelling. There are several more examples where it is written logographically with the character 其,82 but we can surmise on the basis of the meter in the poems that this character indeed represents *sǝre. 2.4.2.1.1 Stem sǝ

2.4.2.1.1.1 Accusative sǝ-wo 其乎奈何不來者來者其乎

SƏ-wo naⁿDƏ?/naⁿZƏ? KƏ-ⁿZ-U pa KƏ-mba SƏ-wo that-ACC why come-NEG-CONV TOP come-COND that-ACC 80  The modifier form sǝnǝ, similar to the proximal demonstrative modifier form kǝnǝ, historically consists of the unextended base of the demonstrative pronoun sǝ or kǝ plus the attributive form n-ǝ of the defective verb n- ‘to be.’ 81  Yamada Yoshio gives one Eastern Old Japanese example in isolation from MYS 14.3451 (Yamada 1954: 73), but he apparently misreads so, an exclamation chasing away a horse, as sǝ ‘that,’ cf. Takagi et al. 1960: 431, and Kojima et al. 1973: 474, who agree that so in this poem is an exclamation. 82  Yamada Yoshio believes that the character 彼 also was read as sǝre and not kare, because, according to his opinion kare was not yet used extensively in Old Japanese (Yamada 1954: 77–78). He even points out a place name 彼杵 which has reading /sǝnǝkï/. However, kare ‘that over there’ is actually attested in Old Japanese (e.g., in MYS 18.4045) (this example is cited by Yamada himself (Yamada 1954: 77), and it is well known how slippery the evidence from place names could be. Thus, I think that only the cases where the character 其 is used can be treated as sǝre with a certain degree of confidence. Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

275

NOMINALS

if [you] do not come or if [you] come, why [should I be impatient] about that? (MYS 11.2640) 曾乎見礼婆許己呂乎伊多美

sǝ-wo MI-re-mba kǝkǝrǝ-wo ita-mi that-ACC see-EV-CON heart-ABS be painful-GER when [one] sees that, [one’s] heart aches (MYS 18.4122) This is an example of the anaphoric usage, referring to a situation previously described in the text. 2.4.2.1.1.2

Modifier Form sǝnǝ

夜弊賀岐都久流曾能夜弊賀岐袁

ya-pe-ŋ-gaki tukur-u sǝnǝ ya-pe-ŋ-gaki-wo eight-fold-GEN-fence make-FIN that eight-fold-GEN-fence-ACC [I] am making eight-folded fence, that eight-folded fence (KK 1) This is an example of the anaphoric usage, referring to the fence previously mentioned in the text. 和賀淤岐斯都流岐能多知曾能多知波夜

wa-ŋga ok-i-si turuki n-ǝ tati sǝnǝ tati pa ya I-POSS put-CONV-PAST/ATTR sword DV-ATTR long.sword that long.sword TOP EP I put [down my] long sword, that long sword! (KK 33) This is an example of the anaphoric usage, referring to the sword previously mentioned in the text. 美母呂能曾能多迦紀

mi-mǝrǝ-nǝ sǝnǝ taka kï HON-mountain-GEN that high fortress that high fortress of the sacred mountain (KK 60) 曾能阿牟袁阿岐豆波夜具比

sǝnǝ amu-wo akiⁿdu paya-ŋ-gup-i that horsefly-ACC dragonfly quick-DV(CONV)-eat-CONV a dragonfly quickly ate that horsefly (KK 97) 秋風左無美曾乃可波能倍爾

AKI KAⁿZE samu-mi sǝnǝ kapa-nǝ [u]pɛ-ni autumn wind be cold-GER that river-GEN top-LOC above that river, the autumn wind is cold (MYS 17.3953) Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

276

Chapter 4

曾能許己呂多礼爾見世牟

sǝnǝ kǝkǝrǝ tare-ni MI-se-m-u that feeling who-DAT see-CAUS-TENT-FIN to whom will [I] show that feeling? (MYS 18.4070) In this example sǝnǝ refers to the author’s (speaker’s) feeling, demonstrating well that there is no apparent connection of sǝnǝ with addressee, as in later Japanese. 2.4.2.1.2 Stem sǝre The stem sǝre occurs only in isolation if we exclude all dubious cases. The only phonographic example of sǝre is attested in a Shōsōin document: 之可流可由惠尓序礼宇氣牟比止良

sik[a]-ar-u-ŋga yuwe n-i sǝre ukɛ-m-u pitǝ-ra thus-exist-ATTR-POSS reason DV-CONV that receive-TENT-ATTR person-PLUR by this reason, the people who will receive that (SSI)83 All other examples of sǝre are in logographic spelling. Some of them are questionable, or have competing readings, however. The following example is the only reliable case: 吾勢子尓令見常念之梅花其十方不所見

WA-ŋGA se-KO-ni MI-SE-M-U tǝ OMƏP-i-si UMƐ-NƏ PANA SƏRE tǝ mo MI-YE-ⁿZ-U I-POSS beloved-DIM-DAT see-CAUS-TENT-FIN think-CONV-PAST/ATTR plumGEN flower that DV FP see-PASS-NEG-CONV The plum blossoms that [I] was going to show to my beloved, do not look like those (i.e., like plum blossoms) (MYS 8.1426) Comparative Data Level A: Other Japonic A1: Eastern Old Japanese There is only one example of the mesial demonstrative pronoun sǝ/sǝre in Eastern Old Japanese, occurring as the modifier form sǝnǝ.

83  I cite this example according to Omodaka et al. (1967: 407), who, unfortunately does not indicate what exactly the document is. Since there is no wider context, it is not possible to tell whether sǝre in this example refers to a single or multiple objects.

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

277

NOMINALS Stem sƏ Modifier Form sƏnƏ 曾能可奈之伎乎刀爾多弖米也母

sǝnǝ kanasi-ki-wo to-ni tate-m-ɛ ya mǝ that beloved-ATTR-ACC door-LOC place-TENT-EV IP EP would [I] let that beloved [of mine] stand at the door?! (MYS 14.3386) A2: Ryukyuan Ryukyuan cognates of the Western Old Japanese mesial demonstrative pronoun sǝ/sǝre are questionable, since in all Ryukyuan languages as well as in Old Ryukyuan we have reflexes of *’u-/’uri ‘that,’ e.g., Shuri ’unu/’uri ‘id.,’ corresponding to Old Japanese sǝnǝ/sǝre. The dubious point is the correspondence of Ryukyuan initial *’- to Japanese *s-, which is irregular. There is only one other example where the same correspondence can be observed: OJ sak- ‘to split’ (intr.) vs. Shuri ac-uN ‘id.’ It is worth mentioning that the Ryukyuan second person pronoun *’ura is probably based on the same pronominal root, although with irregular development of the initial consonant. 2.4.2.2 Mesial Demonstrative Pronoun sǝkǝ The proximal demonstrative pronoun sǝkǝ ‘here’ is a combination of the unextended stem sǝ- of the proximal demonstrative pronoun sǝ/sǝre and the obsolete word *kǝ ‘place.’ It is a stative demonstrative pronoun showing location. Examples: 母登弊波岐美袁淤母比傅須惠幣波伊毛袁淤母比傅伊良那祁久曾許爾淤母 比傅加那志祁久許許爾淤母比傅

mǝtǝ-pe pa kimi-wo omǝp-i-[i]ⁿde suwe-pe pa imo-wo omǝp-i-[i]ⁿde irana-k-eku sǝkǝ-ni omǝp-i-[i]ⁿde kanasi-k-eku kǝkǝ-ni omǝp-i-[i]ⁿde root-side TOP lord-ACC think-CONV-exit(CONV) top-side TOP beloved-ACC think-CONV-exit(CONV) be regrettable-ATTR-NML there-LOC think-CONVexit(CONV) be sorrowful-ATTR-NML here-LOC think-CONV-exit(CONV) at the root [of the tree, I] remember [my] lord, at the top [of the tree, I] remember [my] beloved, [I] remember [my lord] there with a regret, [I] remember [my beloved] here with sorrow (KK 51)

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

278

Chapter 4

可奈之家口許己爾思出伊良奈家久曾許爾念出

kanasi-k-eku kǝkǝ-ni OMƏP-i-[I]ⁿDE irana-k-eku sǝkǝ-ni OMƏP-i-[I]ⁿDE be sorrowful-ATTR-NML here-LOC think-CONV-exit(CONV) be regrettableATTR-NML there-LOC think-CONV-exit(CONV) [I] remember [the parting with you] here with sorrow, [I] remember [the parting with you] there with a regret (MYS 17.3969) 曾己乎之毛安夜爾登母志美

sǝkǝ-wo si mo aya n-i tǝmǝsi-mi there-ABS EP FP strange DV-CONV attractive-GER those places are strangely attractive, and … (MYS 17.4006) 則許母倍婆許己呂志伊多思

sǝkǝ [o]mǝp-ɛ-mba kǝkǝrǝ si ita-si there think-EV-CON heart EP be.painful-FIN when [I] think of those places, [my] heart aches (MYS 17.4006) 曽己由惠尓情奈具也

sǝkǝ yuwe n-i KƏKƏRƏ naŋg-u ya there reason DV-CONV heart calm down-FIN IP will [my] heart calm down due to those circumstances? (MYS 19.4154) Comparative Data Level A: Other Japonic A2: Ryukyuan Similar to the case with kǝkǝ, described above (section 2.4.1.2), there are no direct cognates in Ryukyuan of the Western Old Japanese mesial demonstrative pronoun of place sǝkǝ, since the Ryukyuan counterpart of OJ sǝkǝ is ’uma, consisting of the proximal demonstrative ’u and ma ‘interval.’ Ryukyuan ’uma is amply attested in Old Ryukyuan, and in all modern languages, so apparently it goes back to proto-Ryukyuan. Thus, both Japanese sǝkǝ and Ryukyuan ’uma seem to be late independent formations in two branches of Japonic, and it is not possible to reconstruct a common proto-Japonic archetype. It is a different problem, whether OJ mesial sǝ and Ryukyuan mesial ’u are related (section 2.4.2.1): if they are not, then these formations are completely unrelated in Japanese and Ryukyuan. 2.4.3 Distal Demonstrative Pronouns Distal demonstrative pronouns denote that something is located at a considerable distance in a general sense. This general orientation is typical in Old

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

279

NOMINALS

Japanese for all demonstrative pronouns, as was noted above, and is quite different from the speaker/addressee orientation that started to be prevalent from Classical Japanese. There are two sets of distal demonstrative pronouns: the pronouns of the first set group are based on the stem ka-, and the pronouns of the second set on the stem wo-.84 It is worth mentioning that among distal demonstratives indicating place there is a kind of complimentary distribution of derivatives: thus, the -ti form is based on the wo- set (woti), but the -nata form is based on the ka- set (kanata).85 2.4.3.1 Distal Demonstrative Pronoun ka/kare The distal demonstrative pronoun ka/kare indicates that something is located at a considerable distance. Both its unextended stem ka and extended stem kare are very rare in Old Japanese compared to later periods in Japanese language history. For example, in contrast to proximal and mesial modifier forms kǝnǝ and sǝnǝ, which are the most frequent forms of the respective proximal and mesial sets, the distal modifier form kanǝ occurs only once in Eastern Old Japanese, but is unattested in Western Old Japanese. This low frequency could potentially indicate that distal demonstrative pronouns are a comparatively recent innovation in Old Japanese,86 and that the proto-Japanese deictic system probably included just the bipartite (kǝ and sǝ) system, but not the tripartite (kǝ, sǝ, ka), which flourishes only during the Heian period. However, the existence of ka- series in South Ryukyuan dialects prevents such a conclusion, and probably indicates that there was distal demonstrative *ka- in proto-Japonic. 2.4.3.1.1 Stem ka The unextended stem ka occurs in Western Old Japanese only in isolation. 2.4.3.1.1.1

Isolated Form ka

The isolated form ka occurs in Western Old Japanese only in the construction ka … ka-ku ‘that [way/much]… this [way/much].’ This construction may also include the converb form n-i of the defective verb n- ‘to be’ after each of its members: ka n-i … ka-ku n-i.

84  There are two sets of distal demonstrative pronouns in Classical Japanese as well, but only the first set based on the stem ka- is identical with Old Japanese. Another Classical Japanese set of distal demonstrative pronouns is based on the stem a- (Vovin 2003: 120). 85  Cf. the proximal demonstratives where both -ti form and -nata form are based on kǝ-: kǝti, kǝnata. 86  This possibility was suggested by Bjarke Frellesvig (personal communication).

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

280

Chapter 4

阿波志斯袁美那迦母賀登和賀美斯古良迦久母賀登

ap-as-i-si womina ka mǝŋga tǝ wa-ŋga mi-si ko-ra ka-ku mǝŋga tǝ meet-HON-CONV-PAST/ATTR woman that DP DV I-POSS see(CONV)-PAST/ ATTR girl-DIM thus-CONV DP DV [I] wish that [much] the woman [I] met; [I] wish this [much] the girl I saw (KK 42) 可爾迦久爾保志伎麻爾麻爾斯可爾波阿羅慈迦

ka n-i ka-ku n-i posi-ki manima n-i sika n-i pa ar-aⁿzi ka that DV-CONV thus-CONV DV-CONV be desirable-ATTR according DV-CONV thus DV-CONV TOP exist-NEG/TENT IP [acting] that [way] and this way, according to [your] desires, [it] would not be that way, [would it]? (MYS 5.800) 可由既婆比等爾伊等波延可久由既婆比等爾迩久麻延

ka yuk-ɛ-mba pitǝ-ni itǝp-aye ka-ku yuk-ɛ-mba pitǝ-ni nikum-aye that go-EV-CON person-DAT detest-PASS(CONV) thus-CONV go-EV-CON person-DAT hate-PASS(CONV) when [they] go that [way], [they] are detested by people, and when [they] go this way, [they] are hated by people … (MYS 5.804) 2.4.3.1.2 Stem kare There is one example of the extended stem kare in Western Old Japanese attested in phonographic spelling: 支見我弥不根可母加礼

kimi-ŋga mi-pune kamǝ kare lord-POSS HON-boat EP that [Is] that [my] lord’s boat, I wonder? (MYS 18.4045) Besides this example, there are several others spelled logographically with the character 彼. In the examples from the poetry, we can speculate that this character renders kare, but chances are that it also may be sǝre: all we can say on the basis of the meter is that the word this character renders is disyllabic. Examples: 誰彼我莫問

TA SƏ KARE WARE-WO NA-TOP-I-SƏ who FP that I-ACC NEG-ask-CONV-do Do not ask me: ‘Who is that?’ (MYS 10.2240)

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

281

NOMINALS 誰彼登問者

TA SƏ KARE tǝ TOP-Amba who FP that DV ask-COND if [someone] asks: ‘Who is that?’ (MYS 11.2545) Comparative Data Level A: Other Japonic A1: Eastern Old Japanese There is only one example of the distal demonstrative pronoun ka/kare in Eastern Old Japanese. It occurs in the modifier form kanǝ: Modifier Form kanƏ 可能古呂等宿受夜奈里奈牟

kanǝ ko-rǝ-tǝ NE-ⁿz-u ya nar-i-n-am-u that girl-DIM-COM sleep-NEG-CONV IP become-CONV-PERF-TENT-ATTR Will it become [so that I] will not sleep with that girl? (MYS 14.3565) A2: Ryukyuan Ryukyuan data represent an interesting picture from the comparative perspective. Most of the Ryukyuan languages, as well as Old Ryukyuan, have the distal demonstrative pronoun ’ari/’anu, based on the stem ’a-, which does not have a counterpart in Old Japanese, but has parallels in Classical Japanese and later historical stages of Japanese. However, in dialects of South Ryukyus we find the distal demonstrative pronoun kari/kanu, based on the stem ka-, such as the Agarinakasone and the Yonaha dialects on Miyako island, the Tonoshiro dialect on Ishigaki island (Uchima and Arakaki 2000: 359), the Yonaguni dialect, and other Yaeyama dialects (Hirayama 1967: 195). Since direct borrowing from Old Japanese into South Ryukyuan is out of the question, this distribution suggests that ka- forms in Ryukyus are original, and that the ’a- forms were probably spread by the language interference with Classical Japanese via Old Ryukyuan. However, the unnaturally low frequency of ka- forms in Old Japanese as compared to Classical Japanese, still presents a puzzle, although not an obstacle for reconstructing distal demonstrative *ka- on the proto-Japonic level. 2.4.3.2 Distal Demonstrative Pronoun kanata The distal demonstrative pronoun kanata ‘there,’ ‘that side’ originated as a result of a contraction of the modifier form kanǝ of the distal demonstrative pronoun ka/kare- and the word kata ‘side,’ ‘direction.’ It is not attested in the phonographic script in Western Old Japanese, but there is one example in the

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

282

Chapter 4

logographic script that allows us to read the characters 彼方 as kanata due to the meter of the poem: 壯士墓此方彼方二造置有故

WOTƏKƏ-ⁿ-DUKA KƏNATA KANATA-ni TUKUR-I-OK-ER-U YUWE [young] man-GEN-grave here there-LOC make-CONV-put-PROG-ATTR reason because the graves of [young] men were built here [and] there (MYS 9.1809) Comparative Data Level A: Other Japonic A2: Ryukyuan In contrast to proximal demonstrative kogata, attested in Old Ryukyuan (section 2.4.1.3), its distal counterpart *kagata does not appear in Old Ryukyuan texts. 2.4.3.3 Distal Demonstrative Pronoun woti/wotǝ/wote The distal demonstrative pronoun woti/wotǝ/wote indicates the remote place or time. Only the form woti can be used independently (but always within the construction woti kǝti ‘there [and] here’), but all three forms can be used as a modifier without the following attributive forms n-ǝ and t-u of the defective verbs n- and t- ‘to be’ for the nouns kata ‘side,’ ‘direction’ (woti kata), mo ‘side,’ ‘direction’ (wote mo), tǝsi ‘year’ (wotǝ tǝsi),87 and with the following attributive form t-u for the noun pi ‘day’ (wotǝ t-u pi) and one noun pataⁿde ‘edge’ (?) (see KK 105 below). Although it almost went out of active usage by the Heian period,88 its reflexes still survive in the modern language in the compounds oto-to-i89 ‘the day before yesterday’ < OJ wotǝ t-u pi and oto-tosi ‘the year before last’ < OJ wotǝ tǝsi. It is quite apparent that even in Old Japanese woti/wotǝ/ wote has become a kind of bound form that probably represents a remainder of an old form from the earlier stages of the language history.

87  The form wotǝ tǝsi does not occur in the phonetic writing, therefore, other readings may be possible; however, this particular reading is supported by the data from the later periods of the Japanese language history. 88  Ōno Susumu gives one attestation of woti in the Kokin wakashū (380) for the Heian period, and one in the Uji shūi monogatari for the Kamakura period (Ōno et al. 1990: 1457). 89  The modern standard Japanese from ototoi ‘the day before yesterday’ is, of course, irregular. However, the regular form ototui is preserved in dialects, e.g., in the Fukumitsu dialect in Western Toyama.

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

283

NOMINALS 意富美夜能袁登都波多傅須美加多夫祁理

opǝ miya-nǝ wotǝ t-u pataⁿde sumi katambuk-er-i great palace-GEN that DV-ATTR edge(?) corner incline-PROG-FIN The edge corners of that side of the great palace are falling apart (KK 105) 前年之先年從至今年恋跡

WOTƏ TƏSI-NƏ SAKI-TU TƏSI-YORI KƏ TƏSI-MAⁿDE KOP-URE-ⁿdǝ that year-GEN before-GEN/LOC year-ABL this year-TERM long.for-EV-CONC although [I] long for [you] from the year before two years ago (lit.: from the year before the year before last) up to this year (MYS 4.783) 波都世乃加波乃乎知可多尓伊母良波多多志己乃加多尓和礼波多知弖

Patuse-nǝ kapa-nǝ woti kata-ni imǝ-ra pa tat-as-i kǝnǝ kata-ni ware pa tat-i-te Patuse-GEN river-GEN that side-LOC beloved-DIM TOP stand-HON-CONV this side-LOC I TOP stand-CONV-SUB [My] beloved is standing on that side of the Patuse river, and I stand on this side (MYS 13.3299a) This is a variant of the poem MYS 13.3299. 乎登都日毛昨日毛今日毛由吉能布礼礼婆

wotǝ t-u PI mo KINƏPU mo KEPU mo yuki-nǝ pur-er-e-mba that DV-ATTR day FP yesterday FP today FP snow-GEN fall-PROG-EV-CON when the snow has been falling the day before yesterday, yesterday, and today (MYS 17.3924) 和可伎兒等毛波乎知許知爾佐和吉奈久良牟

waka-ki KO-ⁿdǝmo pa woti kǝti-ni sawak-i-nak-uram-u young-ATTR child-PLUR TOP there here-LOC make.noise-CONV-cryTENT2-FIN young children will probably cry loudly here [and] there (MYS 17.3962) 二上能乎弖母許能母爾安美佐之弖

PUTA ŋGAMI-nǝ wote mǝ kǝnǝ mǝ-ni ami sas-i-te Putagami-GEN that side this side-LOC net spread-CONV-SUB Spreading nets on that side [and] this side of the [Mt.] Putagami (MYS 17.4013)

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

284

Chapter 4

都麻母古騰母毛乎知己知爾左波爾可久美為

tuma mǝ ko-ⁿdǝmǝ mo woti kǝti-ni sapa n-i kakum-i wi spouse FP child-PLUR FP there here-LOC many DV-CONV surround-CONV exist(CONV) both my spouse and children are around [me] in great numbers here and there (MYS 20.4408) Comparative Data Level A: Other Japonic A1: Eastern Old Japanese Only the form wote occurs in Eastern Old Japanese, and only once in the construction wote mo kǝnǝ mo-ni ‘on that side [and] this side’: 安思我良能乎弖毛許乃母爾佐須和奈

Asiŋgara-nǝ wote mo kǝnǝ mǝ-ni sas-u wana Asiŋgara-GEN that side this side-LOC set-ATTR trap the traps [they] set on that side [and] this side of the [Mt.] Asiŋgara (MYS 14.3361) A2: Ryukyuan There are no cognates of the Old Japanese distal demonstrative pronoun woti/ wotǝ/wote in Ryukyuan. As mentioned above, woti, by analogy with kǝti should consist of the distal deictic element wo, and the remainder -ti, which probably represents the word ti ‘way’ < *tï.90 Here, however, we come to a halt, because to the best of my knowledge there are no reliable external cognates of this distal deictic *wo-. 2.5

Interrogative Pronouns

Western Old Japanese has the following interrogative pronouns: ta-/tare ‘who’; nani ‘what’; ika ‘how’; iⁿdu/iⁿduku, iⁿduti ‘where’; iⁿdure ‘which’; itu ‘when’; naⁿzǝ, naⁿdǝ ‘why,’ iku/ikura/ikuⁿda ‘how many,’ ‘how much.’ Indefinite pronouns are derived from interrogative pronouns with adding to the latter the interrogative particle ka. In a similar way, collective pronouns are combinations of interrogative pronouns with the following particle mǝ.

90  The otsu-rui nature of the vowel /i/ in ti is supported by the forms wotǝ and wote, which allow us to reconstruct *wo-tǝy.

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

NOMINALS

285

2.5.1 Interrogative Pronoun ta-/tare Similar to the most personal and demonstrative pronouns, interrogative pronoun ta/tare ‘who’ also has the unextended stem ta- and the extended stem tare. The unextended stem ta- never occurs in isolation, but is always followed by a case marker, in most cases the possessive case marker -ŋga. The extended stem tare occurs both in isolation and in combination with the following case markers, except the possessive -ŋga. Although the extended stem tare is used more extensively than extended stems of other pronouns, the unextended stem ta- in contrast to the unextended stem of other pronouns has much more limited usage. However, it is still likely that in the period preceding first written texts, the unextended stem ta- had a wider usage since, for example, both dative ta-ni and tare-ni are attested once each in Western Old Japanese, but ta-ni occurs in earlier Old Japanese, and tare-ni in later Old Japanese. Therefore, it is possible to believe that the extended stem tare was gradually replacing the unextended stem ta-. The combination of tare ‘who’ with the following interrogative particle ka in Old Japanese always has the function of an emphatic question, in contrast to Classical Japanese, where the combination of tare with particle ka: tare ka has the function of an indefinite pronoun and usually means ‘somebody’ or ‘anybody’ (in questions), although there are also cases when the interrogative particle ka simply emphasizes the question asked about tare. There is, however, one possible exception, when the construction ta-ŋga … ka, attested in the Fudoki kayō, probably has the function of an indefinite pronoun. Combination of tare with particle mǝ: tare mǝ which in Classical Japanese has the function of a collective pronoun meaning ‘everybody,’ or ‘nobody’/‘anybody’ (in negative sentences), does not occur in Old Japanese texts. Examples: 2.5.1.1 Stem ta The unextended stem ta- occurs only in combination with possessive case marker -ŋga and dative case marker -ni. 2.5.1.1.1 Possessive ta-ŋga The possessive form ta-ŋga in Western Old Japanese occurs only in the function of a modifier of the following noun, and never in the function of a subject of the dependent clause. 夜麻登弊迩由玖波多賀都麻

Yamatǝ-pe-ni yuk-u pa ta-ŋga tuma Yamato-side-LOC go-ATTR TOP who-POSS spouse Whose spouse goes towards Yamato? (KK 56)

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

286

Chapter 4

賣杼理能和賀意富岐美能淤呂須波多他賀多泥呂迦母

Meⁿdǝri n-ǝ wa-ŋga opǝ kimi-nǝ or-ǝs-u pata ta-ŋga tane rǝ kamǝ Meⁿdǝri DV-ATTR I-POSS great lady-GEN weave-HON-ATTR fabric who-POSS material DV EP The fabric my lady Meⁿdǝri weaves, for whom (lit. whose) the material is, I wonder? (KK 66) 柂我佐基泥佐基泥曾母野倭我底騰羅須謀野

ta-ŋga sakï-ⁿ-de sakï-ⁿ-de sǝ mǝ ya wa-ŋga te tǝr-as-umo ya who-POSS chap(NML)-DV(ATTR)-hand chap(NML)-DV(ATTR)-hand FP EP EP I-POSS hand take-HON-EXCL IP whose chapped hand, chapped hand will take my hand?! (NK 108) 多我多米爾奈礼

ta-ŋga tamɛ n-i nare who-POSS for DV-CONV you for whom? [For] you! (MYS 17.4031) 波奈爾保比弖里低多弖流波波之伎多我都麻

pana nipop-i ter-i-te tat-er-u pa pasi-ki ta-ŋga tuma flower be fragrant-CONV shine-CONV-SUB stand-PROG-ATTR TOP be lovelyATTR who-POSS spouse whose lovely spouse is the one who is standing, fragrant and shining [as] a flower? (MYS 20.4397) 2.5.1.1.2 Dative ta-ni There is only one example of the dative form ta-ni in Western Old Japanese texts. 多爾加母余良牟

ta-ni kamǝ yǝr-am-u who-DAT EP rely-TENT-ATTR On whom will [you] rely, I wonder? (KK 94) 2.5.1.2 Stem tare The extended stem tare occurs in Western Old Japanese in isolation and with dative case marker -ni and accusative case marker -wo.

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

287

NOMINALS

2.5.1.2.1

Isolated Form tare

多禮曾意富麻弊爾麻袁須

tare sǝ opǝ mape-ni mawos-u who FP great front-DAT say(HUM)-ATTR Who will report to the emperor? (KK 97) 柂例柯擧能居登飫褒磨陛爾麻鳴須

tare ka kǝnǝ kǝtǝ opo mape-ni mawos-u who IP this matter great front-DAT say(HUM)-ATTR Who will report this matter to the emperor? (NK 75) In this example, which represents a textual variant of the example from KK 97 above, the combination tare ka is used in the function of an emphatic question, not as an indefinite pronoun. 偉儺謎能陁倶弥柯該志須弥儺皤旨我那稽麼柂例柯柯該武預

Winambɛ-nǝ takumi kakɛ-si sumi-napa si-ŋga na-k-emba tare ka kakɛ-m-u yǝ Winambɛ-GEN carpenter apply(CONV)-PAST/ATTR ink-cord he-POSS noATTR-COND who IP apply-TENT-ATTR EP the ink-cord that carpenter of Winambɛ applied: if he is no more, who would apply it? (NK 80) Like in the example above (NK 75), tare ka in this example is not the indefinite pronoun ‘somebody’: the interrogative particle ka simply emphasizes the question asked with tare. 美多多志世利斯伊志遠多礼美吉

mi-tat-as-i s-er-i-si isi tare mi-ki HON-stand-HON-NML do-PROG-CONV-PAST/ATTR stone who see(CONV)-PAST/FIN who has seen the stone on which [she] took [her] stand? (MYS 5.869) 念意緒多礼賀思良牟母

OMƏP-U KƏKƏRƏ-wo tare ka sir-am-umǝ love-ATTR heart-ACC who IP know-TENT-EXCL who will know [my] loving heart?! (MYS 17.3950) Like in the examples above (KK 97) and (NK 75), tare ka in this example is not the indefinite pronoun ‘somebody’: the interrogative particle ka simply emphasizes the question asked with tare.

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

288

Chapter 4

山人夜多礼

YAMA-m-BITƏ ya tare mountain-GEN-person IP who Who is the hermit? (MYS 20.4294) 奈禮乎曾與咩爾保師登多禮

nare-wo sǝ yǝme n-i posi tǝ tare you-ACC FP bride DV-CONV be desirable DV who who wants you as [his] bride? (NR 2.33) 2.5.1.2.2 Dative tare-ni Only one example is attested in Western Old Japanese texts: 曾能許己呂多礼爾見世牟等於母比曾米家牟

sǝnǝ kǝkǝrǝ tare-ni MI-se-m-u tǝ omǝp-i-sǝmɛ-k-em-u that feeling who-DAT see-CAUS-TENT-FIN DV think-CONV-begin(CONV)PAST/FIN-TENT-FIN [I] started to think: ‘Whom will [I] show that feeling?’ (MYS 18.4070) 2.5.1.2.3 Accusative tare-wo Only one example is attested in Western Old Japanese texts: 袁登賣杼母多禮袁志摩加牟

wotǝme-ⁿdǝmǝ tare-wo si mak-am-u maiden-PLUR who-ACC EP pillow-TENT-FIN whom will you take to bed [among] the maidens? (KK 15) Comparative Data Level A: Other Japonic A1: Eastern Old Japanese In Eastern Old Japanese the unextended stem ta- occurs only with the following possessive case marker -ŋga, and extended stem tare occurs in isolation and with the following accusative case marker -wo. Stem ta Possessive ta- ŋga

Only three examples are attested in Eastern Old Japanese: 都久波尼爾阿波牟等伊比志古波多賀己等岐氣波加弥尼阿波巣氣牟

Tukumba-ne-ni ap-am-u tǝ ip-i-si ko pa ta-ŋga kǝtǝ kik-ɛ-mba ka mi-ne ap-aⁿz-u-k-em-u Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

289

NOMINALS

Tukumba-peak-LOC meet-TENT-FIN DV say-CONV-PAST/ATTR girl TOP whoPOSS word listen-EV-CON IP HON-peak/sleep(NML)91 meet-NEG-CONVPAST/FIN-TENT-FIN The girl, who promised [to me] that [we] will meet at the Tukumba peak, did not meet [me] at the peak/for sleeping, because [she] listened to somebody’s words (FK 2) This is the only example in Eastern Old Japanese where ta-ŋga with the following interrogative particle ka probably is used in the indefinite pronoun function. 兒呂波多賀家可母多牟

KO-rǝ pa ta-ŋga ke ka mǝt-am-u girl-DIM TOP who-POSS container IP hold-TENT-ATTR whose container will the girl hold? (MYS 14.3424) 佐伎毛利爾由久波多我世

sakimori-ni yuk-u pa ta-ŋga se border guard-LOC go-ATTR TOP who-POSS beloved Whose beloved is going to (be a) border guard? (MYS 20.4425) Stem tare Isolated Form tare

Only one example is attested in Eastern Old Japanese: 多礼曾許能屋能戸於曾夫流

tare sǝ kǝnǝ YA-nǝ TO osǝ-m-bur-u who FP this house-GEN door push-DV(CONV)-shake-ATTR Who pushes and shakes the door of this house? (MYS 14.3460) Accusative tare-wo

Only one example is attested in Eastern Old Japanese: 多礼乎可伎美等弥都都志努波牟

tare-wo ka kimi tǝ mi-tutu sinop-am-u who-ACC IP lord DV see(CONV)-COOR long for-TENT-ATTR whom shall [I] long for, viewing [him] as [my] lord? (MYS 20.4440)

91  The word ne here represents a play on words: ne ‘peak’ and ne ‘sleep(CONV).’ Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

290

Chapter 4

A2: Ryukyuan The interrogative pronoun cognate to OJ ta-/tare is attested in all Ryukyuan languages and in Old Ryukyuan. On the basis of comparative evidence from the modern languages we can reconstruct the proto-Ryukyuan form *taa, which is attested in most languages as taa or ta (Uchima and Arakaki 2000: 416). In addition, there is good evidence for reconstructing an extended stem *taro (Thorpe 1983: 223), which is attested in Old Ryukyuan as taru and has reflexes mostly in the Northern Ryukyus, but the form taru is also attested in the Tonoshiro dialect on Ishigaki island. Several dialects also have the form taN, and on the basis of the data from the Sesoko dialect it appears that this form represents a contraction of the extended form taru, although the final conclusion has to wait until the descriptive data from other dialects become available. Old Ryukyuan data present evidence for the unextended stem ta/taa and extended stem taru. The crucial difference from the Old Japanese data is that both of these stems mostly occur with the following possessive case marker -ga, although there is one example when ta is attested with the following dative case marker -ni, and one example when taru is attested in isolation (before the particle mo). Stem ta たかとりよらたかうちよら

ta-ga tor-i-yor-a ta-ga ut-i-yor-a who-POSS take-CONV-exist-TENT who-POSS hit-CONV-exist-TENT Who will take [it]? Who will hit [it]? (OS 16.1157) おの玉やたあに呉ゆが

ono tama ya taa-ni KWIy-u ga that jewel TOP who-DAT give-FIN IP to whom will [I] give that jewel?92 Stem taru たるかさちへにせる

taru-ga satife ni-se-ru who-POSS wear(COND) fit-CAUS-FIN for whom will [it] be appropriate to wear? (OS 11.594) 92  This example from the Ryūka is cited according to Hokama (1995: 380).

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

NOMINALS

291

誰もわていやは

TARU mo wa te iy-aba who FP I DV say-COND If everyone says ‘I’ (RK 617) In the Sesoko dialect the unextended stem taa can be used as: 1) a modifier of the following noun; 2) before the certain case markers and particles, such as the comitative case marker, the topic marker or the restrictive particle bakee; 3) before the case marker -ga used in the function of a subject marker. The extended stem taru appears only before the case marker -ga used in the possessive function (Uchima 1984: 107–108). The form taN is used exactly as the extended stem taru. Sesoko Dialect Stem taa

taa muN ga who thing IP Whose thing [is this]? (Uchima 1984: 107) taa-tu taa-tu ’ik-uu ga who-COM who-COM go-FIN IP Who and who will go? (Uchima 1984: 108) taa-ga hatʃaa ga who-NOM write(PAST) IP Who wrote? (Uchima 1984: 107) Stem taru, taN < taru

taru-ga muN ga who-POSS thing IP Whose thing [is this]? (Uchima 1984: 107) taN-ga ʃida yee ga who-POSS elder brother be(FIN) IP Whose elder brother is [it]? (Uchima 1984: 141) 2.5.2 Interrogative Pronoun nani The interrogative pronoun nani ‘what’ can modify a noun with or without the following attributive form n-ǝ of the defective verb n- ‘to be.’ Similar to ta-/tare, in most cases combinations of nani with the following interrogative particle Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

292

Chapter 4

ka functions as an emphatic question, although there is one example below (MYS 15.3733) where it can be treated as an indefinite pronoun. A combination of nani with particle mǝ is not attested in Old Japanese texts. Out of all of the case markers, only the combination of nani with the following accusative case marker -wo is attested. As noted by Yamada Yoshio, in a number of Old Japanese examples nani means ‘why’ rather than ‘what’ (1954: 85), however this alone cannot be a sufficient basis to claim that in OJ nani was not a pronoun, but an adverb, as he suggests (Yamada 1954: 84). Yamada Yoshio believes that nani ‘what’ is segmentable into na+ni, where -ni is the same ni as in ika ni ‘how’ (that, is, in our terminology the converb form n-i of the defective verb n- ‘to be’). To support his point of view, he provides two examples from MYS 14.3430 and MYS 15.3684 (Yamada 1954: 87). However, in MYS 15.3684 we apparently deal with the OJ naⁿzǝ ‘why’ (Omodaka et al. 1967: 523, (Kojima et al. 1975: 73), spelled as 奈曾 /nasǝ/. Bear in mind that as was noted in chapter 2, the prenasalized obstruents in the script variety A are not always spelt consistently. In any case, more uncontroversial examples other than combinations with the following particle sǝ would be necessary to prove Yamada’s hypothesis, but they apparently do not exist. The example from MYS 14.3430, an Eastern Old Japanese text, is even less reliable, because the alleged *na there spelled 奈 is likely to be a scribal mistake of 余 /yǝ/,93 occurring in other manuscripts, a possibility acknowledged by Yamada himself (Yamada 1954: 87). Even if it is not, *na there again occurs before syllable si, with the resulting form nasi, which is treated by some commentators as an Azuma dialect form of naⁿzǝ ‘why’ (Pierson 1961: 80). Finally, a combination of nani ‘what’ with the following converb form n-i of the defective verb n- ‘to be’ is attested in Western Old Japanese (see MYS 13.3265 below). A double sequence of defective verb converbs n-i n-i in *na n-i n-i is ungrammatical. Below I will also demonstrate on the basis of internal Japonic evidence that segmentation of nani into na-ni is not justified. 那爾騰柯母于都倶之伊母我磨陀左枳涅渠農

nani tǝ kamǝ utukusi imǝ-ŋga mata sak-i-[i]ⁿde-kǝ-n-u what DV EP beautiful beloved-POSS again bloom-CONV-exit(CONV)-come-NEG-ATTR for what [reason], I wonder, does [my] beautiful beloved not bloom again? (NK 114) 93  Both NKBT and NKBZ adopt the variant /yǝ/ (Takagi et al. 1960: 427; Kojima et al. 1973: 468–469).

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

293

NOMINALS 那爾柯那皚柯武

nani ka naŋgɛk-am-u what IP lament-TENT-ATTR why would [I] lament? (NK 116) The combination of nani with the interrogative particle ka here expresses an emphatic question. 奈爾能都底擧騰多柂尼之曳鶏武

nani n-ǝ tute-kǝtǝ taⁿda n-i si ye-k-em-u what DV-ATTR report(CONV)-word direct DV-CONV EP good-ATTR-TENT-FIN What message [do you have]? It would be better [to say it] directly (NK 128) In this example nani modifies the following noun tute-kǝtǝ ‘message’ with the attributive form n-ǝ of the defective verb n- following after nani. Cf. MYS 15.3733 and MYS 17.3912 below. 銀母金母玉母奈爾世武

SIROKANE mǝ KU ŋGANE mǝ TAMA mǝ nani se-m-u silver FP gold FP jewel FP what do-TENT-ATTR What shall [I] do with silver, gold, and jewels? (MYS 5.803) 奈爾可佐夜礼留

nani ka sayar-er-u what IP prevent-PROG-ATTR what will be preventing [me]? (MYS 5.870) The combination of nani with the interrogative particle ka here expresses an emphatic question. 吾哉難二加還而将成

WARE YA nani n-i ka KAPER-I-TE NAR-AM-U I EP what DV-CONV IP return-CONV-SUB become-TENT-ATTR Why should I return? (MYS 13.3265) The combination of nani with the interrogative particle ka here expresses an emphatic question. Also, note that nani is followed by the converb n-i of the defective verb n- ‘to be,’ which prevents analyzing ni in nani as this verb.

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

294

Chapter 4

奈爾之可母奇里爾多都倍久奈氣伎之麻佐牟

nani si kamǝ kïri-ni tat-umbɛ-ku naŋgɛk-i s-i-mas-am-u what EP EP fog-COMP rise-DEB-CONV lament-NML do-CONV-HON-TENT-ATTR I wonder why should [you] lament so that [your tears] have to rise as a fog? (MYS 15.3581) 和伎毛故我可多美能許呂母奈可里世婆奈爾毛能母弖加伊能知都我麻之

wa-ŋg-imo-ko-ŋga katami n-ǝ kǝrǝmǝ na-k-ar-i-s-emba nani monǝ mǝt-e ka inǝti tuŋg-amasi I-POSS-beloved-DIM-POSS keepsake DV-ATTR garment no-CONV-exist-CONVPAST/ATTR-COND what thing hold-EV IP life continue-SUBJ If [I] did not have a garment of my beloved as a keepsake, would having anything keep me alive? (MYS 15.3733) In this example nani with the following ka probably has a function of an indefinite pronoun, although other analyses are not impossible either. Also, in this example nani modifies the following noun mǝnǝ ‘thing’ without having the attributive form n-ǝ of the defective verb n- ‘to be’ after it. Cf. NK 128 above and MYS 17.3912 below. 何爲牟尓吾乎召良米夜

NANI SE-m-u-ni WA-wo MES-Uram-ɛ ya what do-TENT-ATTR-LOC I-ACC summon-TENT2-EV IP should you have summoned me in order to do something? (MYS 16.3886) In this example nani has a function of an indefinite pronoun without the following interrogative particle ka. 保登等藝須奈爾乃情曾

potǝtǝŋgisu nani n-ǝ KƏKƏRƏ sǝ cuckoo what DV-ATTR feeling FP [Oh,] cuckoo! What feelings [do you have?] (MYS 17.3912) In this example nani modifies the following noun kǝkǝrǝ ‘feeling’ with the attributive form n-ǝ of the defective verb n- ‘to be’ after it. Cf. NK 128 and MYS 15.3733 above.

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

NOMINALS

295

奈爾乎可於母波牟

nani-wo ka omǝp-am-u what-ACC IP think-TENT-ATTR what should [I] think? (MYS 17.3967) The combination of nani with the interrogative particle ka here expresses an emphatic question. 2.5.2.1 Special Form Besides nani ‘what,’ there is a special form aⁿ- ‘what’ of the same pronoun occurring once in front of the defective verb tǝ ‘to say’ in Western Old Japanese.94 安杼毛倍香許己呂我奈之久伊米爾美要都流

aⁿ-dǝ [o]mop-ɛ ka kǝkǝrǝ-ŋ-ganasi-ku imɛ-ni mi-ye-t-uru what-DV think-EV IP heart-GEN-sad-CONV dream-LOC see-PASS(CONV)-PERF-ATTR what was [I] thinking [about]? As [I] was sad in [my] heart, [she] suddenly has appeared in [my] dream. (MYS 15.3639) The loss of the initial n- does not occur in Old Japanese, so this word an- might have been viewed as another word for ‘what,’ etymologically unconnected with nani ‘what,’ but as I will demonstrate below the comparative evidence offers a cogent explanation for this seeming irregularity. In addition, this form is found in book fifteen of the Man’yōshū, that was likely compiled by a speaker of Eastern Old Japanese (Vovin 2012: 14–16). Comparative Data Level A: Other Japonic A1: Eastern Old Japanese In Eastern Old Japanese the interrogative pronoun nani ‘what’ is attested only once in a Sakimori poem:

94  Influenced by the kana syllabary, Japanese linguists usually list this pronoun as aⁿdǝ (Omodaka et al. 1967: 32). However, it is quite clear from its syntactic pattern that this -tǝ is nothing but a defective verb ‘to say’ with a prenasalization from the first element aⁿ‘what.’ Some Japanese scholars also believe that 阿跡念 used in MYS 10.2140 to spell the verb aⁿdǝmop-u ‘to lead’ also supports aⁿ-dǝ (Omodaka et al. 1967: 32; Kojima et al. 1973: 119), but this is more controversial, since first two characters 阿跡 may be just a case of phonetic spelling, while the last character 念 alone represents rebus writing.

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

296

Chapter 4

奈爾須礼曾波波登布波奈乃佐吉泥己受祁牟

nani s-ure sǝ papa tǝ [i]p-u pana-nǝ sak-i-[i]ⁿde-kǝ-ⁿz-u-k-em-u what do-EV FP mother DV say-ATTR flower-GEN bloom-CONV-exit(CONV)-come-NEG-CONV-PAST/FIN-TENT-ATTR why (lit.: having done what) has the flower called ‘Mother’ not bloomed? (MYS 20.4323) This limited distribution is suspicious, especially since in Eastern Old Japanese there is the aforementioned interrogative pronoun aⁿ- ‘what,’ ‘why’95 occurring seven times before the defective verb tǝ ‘to say’ or the defective verb tǝ ‘to be.’ 和我世故乎安杼可母伊波武

wa-ŋga se-ko-wo aⁿ-dǝ kamǝ ip-am-u I-POSS beloved-DIM-ACC what-DV EP say-TENT-ATTR What shall [I] say about my beloved, I wonder? (MYS 14.3379) 阿杼可多延世武

aⁿ-dǝ ka taye se-m-u why-DV IP break(NML) do-TENT-ATTR why should [we] break up? (MYS 14.3397) 安加奴乎安杼加安我世牟

ak-an-u-wo aⁿ-dǝ ka a-ŋga se-m-u satisfy-NEG-ATTR-ACC what-DV IP I-POSS do-TENT-ATTR since it was not enough [for me], what should I do? (MYS 14.3404) 奴流我倍爾安杼世呂登可母

n-uru-ŋga [u]pɛ-ni aⁿ-dǝ se-rǝ tǝ kamǝ sleep-ATTR-POSS top-LOC what-DV do-IMP DV EP besides that [I] slept with [her], what [else do I] do, I wonder? (MYS 14.3465) 汝波安杼可毛布

NA pa aⁿ-dǝ ka [o]mop-u you TOP what-DV IP think-ATTR What do you think? (MYS 14.3494)

95  J DB glosses aⁿ-dǝ as ‘how,’ ‘why’ instead of ‘what’ (Omodaka et al. 1967: 32). However, the meaning ‘how’ does not occur in the texts at all, and in only one example out of seven aⁿ- means ‘why,’ while it certainly means ‘what’ in the remaining six. Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

297

NOMINALS 安騰須酒香可奈之家兒呂乎於毛比須吾左牟

aⁿ-dǝ s-u s-u ka kanasi-ke KO-rǝ-wo omop-i-suŋgos-am-u what-DV do-FIN do-FIN IP lovely-ATTR girl-DIM-ACC think-CONV-passTENT-ATTR how (lit.: doing what) would I forget the lovely girl? (MYS 14.3564) 安杼毛敞可

aⁿ-dǝ [o]mop-e ka what-DV think-EV IP Thinking what? (MYS 14.3572) Since in Old Japanese the sequence /ni/ can become /ⁿ/ before the following obstruent, cf. sir-an-i ‘not knowing’ (know-NEG-CONV) and sir-aⁿz-u ‘not knowing’ < *sir-an-i s-u (know-NEG-NML do-FIN), it is logical to presume that aⁿ- < *ani ‘what.’ I believe that the interrogative pronoun aⁿ- which occurs in Eastern Old Japanese seven times more frequently than nani is the genuine Eastern Old Japanese form for the interrogative ‘what,’ while the existence of EOJ nani is probably due to the Western Old Japanese influence. This is further confirmed by the fact that an Eastern Old Japanese poem containing nani ‘what’ appear to be in almost perfect Western Old Japanese. On the other hand, the single instance of aⁿ- in Western Old Japanese is in MYS 15.3639, is probably due to the native language of the compiler of book fifteen. In any case, a single example of aⁿ- in the entire Western Japanese corpus is suspicious. Thus, if my supposition is correct, we probably have a Western Old Japanese nani ‘what’ vs. Eastern Old Japanese an[i] ‘id.’ A2: Ryukyuan Most Ryukyuan languages have nuu ‘what’ or noo ‘id.’ (the latter is attested in the South Ryukyus on the islands of Miyako and Ishigaki) (Uchima and Arakaki 2000: 428). At first glance, it seems that it is impossible to establish a cognateship on the basis of regular correspondences between these forms and Western Old Japanese nani, because they share only initial consonant n-, while the rest of both forms seems to be unrelated. However, in Old Ryukyuan we have the form nau ‘what’ (spelled なお /nao/),96 which demonstrates that the modern 96  Hokama believes that the word was read /nu/ in Old Ryukyuan like the modern dialect forms (Hokama 1995: 473). However his belief seems to be based on modern dialect data, which is circular, and on the spelling 루욱/lwuwuk/ found in the Korean source Haytong ceykwukki, where we have Old Ryukyuan transcribed with Hankul. However, to suppose that this form /lwuwuk/ transcribes Ryukyuan /nu:k/ is unrealistic because of the final -k which is not attested anywhere else. In addition, the spelling /au/ for /u/ in the Omoro sōshi is not used in other cases. Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

298

Chapter 4

Ryukyuan dialectal forms nuu and noo go back to noo < nau. There are also two modern Ryukyuan dialect forms that support this reconstruction: Sawada nau and Ōgami naɷ ‘what’ (Ōshiro 1972: 492–493). Now we have to remember that OJ /ni/ may be an otsu-rui *nï which goes back to *nuy, *noy, or *nǝy. The Old Ryukyuan and Sawada form nau as well as Ōgami form naɷ perfectly support the otsu-rui nature of /ni/ in WOJ nani, which can be reconstructed as *nanuy ~ *nanoy or *nanǝy. The first variant is more plausible, because the vowel sequences a-u or a-o occur much more frequently than the vowel sequence a-ǝ; therefore I tentatively reconstruct it as *nanu. At this stage, the Eastern Old Japanese form aⁿ- has to be brought into play, demonstrating that the initial *n- was in all likelihood a segmentable element. Thus, the reconstruction of the proto-Japonic archetype should be *n-anuy ~ *nanoy. The development in the daughter languages then must be as follows. In proto-Central Japanese, the ancestor of Western Old Japanese, the morphological boundary between *n- and *-anu was obliterated, and the *oy ~ *uy blended into otsu-rui *ï, which then merged with a kō-rui *i, resulting in WOJ nani, which is attested in the texts. In the ancestor of Eastern Old Japanese, the initial morpheme *n- was lost. Since EOJ aⁿ- is not attested with its original final vowel, it is not completely clear what happened to the end of this interrogative pronoun in Eastern Old Japanese. In proto-Ryukyuan, the morphological boundary between *n- and *-anu was obliterated like in Western Old Japanese, and the medial *-n- in *nanuy ~ *nanoy was lost (an irregular change), resulting in the form nau, attested in Old Ryukyuan, which further developed (quite regularly) into nuu or noo in modern languages. Examples from Ryukyuan: Old Ryukyuan やまとたひなおかいかのほてやましろたひなおかいかのほて

yamato tabi nao ka-i ga nobo-te yamasiro tabi nao ka-i ga nobo-te Yamato journey what buy-CONV IP go up(CONV)-SUB Yamasiro journey what buy-CONV IP go up(CONV)-SUB [she] goes to Yamato to buy what, [she] goes to Yamasiro to buy what (OS 11.637) Shuri

nu-nu ’a ga what-GEN exist IP what is there? (RKJ 1983: 427)

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

299

NOMINALS

2.5.3 Interrogative Pronoun ika In most cases the interrogative pronoun ika ‘how,’ ‘what’ in Western Old Japanese occurs with the following converb form n-i of the defective verb n- ‘to be.’ Outside of this usage it is used also with the following defective verb tǝ ‘to be’ (twice in the same text), in the construction ika sama n-i ‘in what manner’ (seven examples are attested),97 and once in the construction ika mbakari ‘how,’ ‘to what extent.’98 The first usage with defective verb tǝ did not survive into the Heian period. 高照日之皇子何方尓所念食可

TAKA TER-AS-U PI-NƏ MIKO IKA SAMA n-i OMƏP-OS-I-MES-E ka high shine-ATTR sun-GEN prince what manner DV-CONV think-HON-CONVHON-EV IP in what manner does the prince of the high-shining Sun think? (MYS 2.162) 何方爾念座可

IKA SAMA n-i OMƏP-i-[I]MASE ka what manner DV-CONV think-CONV-HON(CONV) IP in what manner did [he] think? (MYS 3.443) 伊加登伊可等有吾屋前尓

ika tǝ ika tǝ AR-U WA-ŋGA YAⁿDO-ni how DV how DV exist-ATTR I-POSS dwelling-LOC how, how, in my (existing) dwelling (MYS 8.1507) 伊加婆加利故保斯苦阿利家武麻都良佐欲比売

ika mbakari koposi-ku ar-i-k-em-u Sayo-pime how RP be longing-CONV exist-CONV-PAST/FIN-TENT-FIN Sayo-pime how Sayo-pime must have been longing (MYS 5.875)

97  Unfortunately, none of these examples is written phonetically. We can guess that the spellings 何方尓 or 何方 indeed represent ika sama n-i on the basis of the meter of the poems and also on the basis of the fact that ika sama n-i occurs in the later stages of the Japanese language, e.g., in the Genji monogatari (Ōno et al. 1990: 84). However, since nani-sama also occurs, we cannot be completely sure that these examples represent ika sama n-i and not nani-sama n-i. 98  Since the particle mbakari ‘only,’ ‘about’ goes back to N-pakar-i, where pakar-i is an converb of the verb pakar- ‘to measure,’ it is quite possible that N- historically represents a contracted form of the converb form n-i of the defective verb n- ‘to be,’ thus further reducing the number of cases when ika occurs without the following n-i.

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

300

Chapter 4

The most frequently occurring form ika n-i ‘how DV-CONV’ also appears with the following interrogative particle ka in the form ika n-i ka.99 There is no apparent difference in meaning between ika n-i and ika n-i ka. It is possible that the interrogative particle ka is used here to make question more emphatic, as with other interrogative pronouns. 柯羅履爾鳴以柯爾輔居等所

Kara-kuni-wo ika n-i [i]p-u kǝtǝ sǝ Korea-country-ACC how DV-CONV say-ATTR matter FP What (lit.: how) will [you] say about Korea? (NK 99) 阿礼乎婆母伊可爾世与等可

are-womba mǝ ika n-i se-yǝ tǝ ka I-ACC(EMPH) FP how DV-CONV do-IMP DV IP What (lit.: how) do [you] think I should do? (MYS 5.794) 伊弊爾由伎弖伊可爾可阿我世武

ipe-ni yuk-i-te ika n-i ka a-ŋga se-m-u home-LOC go-CONV-SUB how DV-CONV IP I-POSS do-TENT-ATTR What (lit.: how) will I do, when [I] go back home? (MYS 5.795) 伊可爾可由迦牟可利弖波奈斯爾

ika n-i ka yuk-am-u karite pa na-si-ni how DV-CONV IP go-TENT-ATTR food supply TOP no-FIN-LOC because [I] do not have a food supply, how would [I] go? (MYS 5.888) 伊敞妣等乃伊豆良等和礼乎等波婆伊可爾伊波牟

ipe-m-bitǝ-nǝ iⁿdu-ra tǝ ware-wo tǝp-amba ika n-i ip-am-u home-GEN-person-GEN where-LOC DV I-ACC ask-COND how DV-CONV say-TENT-FIN If people from [your] home ask me where [are you], what (lit.: how) should [I] answer? (MYS 15.3689) In contrast to Classical Japanese, where the construction ika nar-u ‘what kind of’ occurs as a set modifier, in Old Japanese the verb nar- ‘to be,’ used after ika

99  O J ika n-i ka is, of course, the predecessor of Classical and modern Japanese ikaga ‘how’ (< ikaŋga < ika n[-i] ka).

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

301

NOMINALS

‘how’ still has a number of paradigmatic forms, and may be separated from the word it modifies by an emphatic particle ya. Historically, ika nar- ‘to be of what kind,’ ‘to be how’ certainly goes back to a combination ika n-i ar- ‘how DV-CONV exist,’ since the verb nar- ‘to be’ originated from the contraction of the converb form n-i of the defective verb n- ‘to be’ with the verb ar- ‘to exist.’ In Western Old Japanese there are two examples where ika nar- still occurs in its uncontracted form ika n-i ar- (see MYS 5.810 and MYS 18.4036 below). 伊可奈留夜比止爾伊麻世可

ika nar-u ya pitǝ n-i imas-e ka how be-ATTR EP person DV-CONV be(HON)-EV IP Oh, what kind of person is [he]? (BS 5) 空事何在云

MUNA-ŋ-GƏTƏ-WO IKA NAR-I TƏ IP-I-TE empty-DV(ATTR)-word-ACC how be-FIN DV say-CONV-SUB how [can you] call [it] an empty word? (MYS 11.2466) 面忘何有人之為者焉

OMO WASURE IKA NAR-U PITƏ-NƏ S-URU MƏNƏ SƏ face forget(NML) how be-ATTR person-GEN do-ATTR thing FP What kind of person will forget [her] face? (MYS 11.2533) 如何有哉人之子故曾

IKA NAR-U YA PITƏ-NƏ KO YUWE sǝ how be-ATTR EP person-GEN girl reason FP due to what kind of girl (MYS 13.3295) 2.5.3.1

Uncontracted Form ika n-i ar-

伊可爾安良武日能等伎爾可母

ika n-i ar-am-u PI-nǝ tǝki-ni kamǝ how DV-CONV exist-TENT-ATTR day-GEN time-LOC EP I wonder, in what time of the day … (MYS 5.810) 伊可爾安流布勢能宇良曾毛

ika n-i ar-u Puse-nǝ ura sǝ mo how DV-CONV exist-ATTR Puse-GEN bay it EP Bay of Puse, how [beautiful] it is! (MYS 18.4036)

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

302

Chapter 4

Comparative Data Level A: Other Japonic A1: Eastern Old Japanese The interrogative pronoun ika ‘how,’ ‘what’ occurs in Eastern Old Japanese texts three times. There are two examples of it in the form ika n-i and one example in the form ika nar-u. 伊可爾思弖古非波可

ika n-i s-i-te kopï-mba ka how DV-CONV do-CONV-SUB long.for-COND IP if [I] long for [you], in what way … (MYS 14.3376) 伊麻波伊可爾世母

ima pa ika n-i se-m-ǝ now TOP how DV-CONV do-TENT-ATTR What (lit.: how) shall [I] do now? (MYS 14.3418) 伊可奈流勢奈可和我理許武等伊布

ika nar-u se-na ka wa-ŋgari kǝ-m-u tǝ ip-u how be-ATTR beloved-DIM IP I-DIR come-TENT-FIN DV say-ATTR what kind of beloved is [he], who says that [he] will come to me? (MYS 14.3536) A2: Ryukyuan There are a number of different forms, corresponding to Old Japanese ika in the modern Ryukyuan dialects in Northern Ryukyus: Nase kyasi (Uemura and Suyama 1993: 403), Yuwan, Koniya ’ikyaʃi; Izena ’itʃantu; Oku ’isaaʃi, tʃaaʃi; Benoki kaaʃi; Kijōka, Sesoko, Namizato, Maejima, Tokashiki tʃaa; Arumi tʃaNgutu; Kowan tʃantuu; Tomigusuku tʃanʃi, and Kumejima ’itʃa (Uchima and Arakaki 2000: 423). There are several things that have to be noted here: first, the Northern Amami dialects point to initial proto-Ryukyuan *e-, not *i-. Second, -ʃi, -ntu[u], and -Ngutu are historically suffixes, because there are forms without them, and because the last two suffixes, -ntu[u] and -Ngutu each have attestations in one or two particular dialects. Third, there are Ryukyuan forms with and without initial i-. Fourth, the Benoki form kaaʃi without initial i- is the only form that does not show any traces of the progressive palatalization /ika/ > /ikya/ > /itʃa/ > /isa/. Since Benoki does not regularly lose initial *i- in front of a velar (cf. Benoki ’iki ‘pond,’ Jp. ike), the logical conclusion is that the initial i- < *e- in the Ryukyuan archetype *e-ka is segmentable, and that it is responsible

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

303

NOMINALS

for this palatalization: *e-ka > *i-ka > ikya > itʃa > isa.100 It is also quite apparent that some of the forms above with palatalization without the initial i- just lost it at a later stage after the palatalization was completed. The chronology of this change is well demonstrated by the Shuri dialect, where caa ‘how’ occurs in the colloquial, and ’icaa ‘how’ in the literary variety (RKJ 1983: 244). Old Ryukyuan represents a considerable array of different forms, almost all of which are already affected by the progressive palatalization: ikya (いきや), ita (いた), itya (いちや, いちゃ), tya (ちや, ちゃ), tyaa (ちやあ) in the Omoro sōshi and the Ryūka, and ika (いか), ikya (いきや), kya (きや), kyaa (きやあ) in the Ryūka (Hokama 1995: 53). In contrast to both Western Old Japanese, where the interrogative pronoun ika ‘how’ is followed by the converb form n-i of the defective verb n- ‘to be’ in the majority of cases, and Eastern Old Japanese, where n-i is always used after ika, Old Ryukyuan ika/ikya and Shuri ’icaa/caa are always used in isolation. Examples: Old Ryukyuan いきやおさうずめしやいが

ikiya osauzu-mesiyai ga how think-HON IP what (lit.: how) do [you] think? (RK 20) いろいろに言ちもいかなしも行かぬ

iro-iro n-i I-ti-mo ika nas-i-mo ik-an-u different DV-CONV say(CONV)-SUB-FP how do-CONV-FP go-NEG-FIN whatever [you] say, and whatever [you] do, [I] will not go (RK 725) Shuri

caa debiru ga how be(POL) IP How is it? (RKJ 1983: 140) ’icaa usoozi-mishee ga how think-HON IP what (lit.: how) do [you] think? (RKJ 1983: 244) 100  The progressive palatalization of a velar after a preceding /i/ is well known in Ryukyuan, cf. case marker -ga > -gya in Old Ryukyuan after nominals ending in /i/ (Hokama 1995: 177).

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

304

Chapter 4

Apparently, the addition of n-i to ika represents an innovation in both Western and Eastern Old Japanese. Taking into consideration what was said earlier about the segmentability of *eka into *e-ka on the basis of Ryukyuan dialect data, it is proper to reconstruct the proto-Japonic archetype as *e-ka. Level B: External Comparisons Given the above reconstruction of the proto-Japonic archetype as *e-ka, the likeliest external parallels for both *e- and *-ka are found in Korean. For the first element *e- a possible external comparison seems to be MK e:styé, e:sté, e:styéy ‘how,’ which probably goes back to a complex formation *e-is-ti ‘how-exist-ADV.’ Comparative example for *e- : Middle Korean icey estye Lahwula-lol aski-no-n-ta now how Rahula-ACC take away-PRES-ATTR-QS how [can you say that I] am taking away Rahula now? (WS 6.9) 2.5.4 Interrogative Pronoun iⁿdu/iⁿduku The interrogative pronoun iⁿdu ‘where’ occurs in Western Old Japanese only three times, once followed by the locative case marker -ra, and twice by the noun pe ‘side’: 何時邊乃方二我戀將息

iⁿdu101-PE-nǝ KATA-ni A-ŋGA KOPÏ YAM-AM-U where-side-GEN side-LOC I-POSS love(NML) end-TENT-FIN where will my love end? (MYS 2.88) 伊敞妣等乃伊豆良等和礼乎等波婆伊可爾伊波牟

ipe-m-bitǝ-nǝ iⁿdu-ra tǝ ware-wo tǝp-amba ika n-i ip-am-u home-GEN-person-GEN where-LOC DV I-ACC ask-COND how DV-CONV say-TENT-FIN If people from [your] home ask me where [are you], how should [I] answer? (MYS 15.3689)

101  In this poem the Japanese reading (kun-yomi) itu of the characters 何時 used ideographically for the interrogative pronoun itu ‘when’ is used as a rebus writing to write the interrogative pronoun iⁿdu ‘where.’

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

305

NOMINALS 伊頭敝能山乎鳴可将超

iⁿdu-pe-nǝ YAMA-wo NAK-I ka KOY-URAM-U where-side-GEN mountain-ACC cry-CONV IP cross-TENT2-ATTR on what side will [the cuckoo] cross the mountains, crying? (MYS 19.4195) In this example the interrogative particle ka is used for emphasizing the question. In all other examples only the form induku appears, with the obsolete suffix -ku of unclear origin. 許能迦迩夜伊豆久能迦迩

kǝnǝ kani ya iⁿduku-nǝ kani this crab IP where-GEN crab This crab, where it [is] from? (KK 42) 都奴賀能迦迩余許佐良布伊豆久迩伊多流

Tunuŋga-nǝ kani yǝkǝ sar-ap-u iⁿduku-ni itar-u Tunuŋga-GEN crab side go.away-ITER-FIN where-LOC reach-FIN Crab from Tunuŋga goes all the time along the side[way]. Where [to] will [it] arrive? (KK 42) 伊豆久欲利枳多利斯物能曾

iⁿduku-yori k-i-tar-i-si mǝnǝ sǝ where-ABL come-CONV-PERF/PROG-CONV-PAST/ATTR thing FP Where did [you] come from? (MYS 5.802) 伊豆久由可斯和何伎多利斯

iⁿduku-yu ka siwa-ŋga k-i-tar-i-si where-ABL IP wrinkle-POSS come-CONV-PERF/PROG-CONV-PAST/ATTR Where did the wrinkles come from? (MYS 5.804) 烏梅能波奈知良久波伊豆久

uMƐ-nǝ pana tir-aku pa iⁿduku plum-GEN flower fall-NML TOP where Where will the plum blossoms fall? (MYS 5.823)

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

306

Chapter 4

Comparative Data Level A: Other Japonic A1: Eastern Old Japanese Only the interrogative pronoun iⁿdu ‘where’ is attested in Eastern Old Japanese, but not the extended form iⁿduku. 伊豆由可母加奈之伎世呂我和賀利可欲波牟

iⁿdu-yu kamǝ kanasi-ki se-rǝ-ŋga wa-ŋgari kayop-am-u where-ABL EP beloved-ATTR husband-DIM-POSS I-DIR visit-TENT-ATTR where will my beloved husband visit me from, I wonder? (MYS 14.3549) A2: Ryukyuan Old Ryukyuan has idu-ma/zuma (< *eⁿdu-ma) ‘where,’ consisting of idu- ‘where’ and -ma ‘place,’ which is attested in the Ryūka (Hokama 1995: 74). The actual pronunciation of OR iduma is supposed to be [dzïma], which is corroborated by the Middle Korean transcription coma, found in the Haytong ceykwukki of 1501 CE (Hokama 1995: 74). Among the modern Ryukyuan languages, the apparent cognate to the Old Ryukyuan iduma [dzïma] is found in the Tonoshiro dialect on Ishigaki island in the Southern Ryukyus. Other likely cognates in the Southern Ryukyus are Agarinakasone ndza, Yonaha ndzan (on the Miyako island), and Hateruma dza (Uchima and Arakaki 2000: 424). Apparent cognates in the Northern Ryukyus are attested in the China and Wadomari dialects on Okinoerabu island: ’uda ‘where’ (Hirayama 1986: 508). Also cf. Kijōka dʒaa in the Northern part of Okinawa island and forms da, daa, and raa attested in different localities in Northern Ryukyus (Uchima and Arakaki 2000: 424).102 South and central Okinawa, including the Shuri dialect, as well as some neighboring islands have apparently innovative maa ‘where’ (Uchima and Arakaki 2000: 424; RKJ 1983: 353), which still may be partially related to OJ iⁿdu/iⁿduku ‘where,’ if, as Hokama believes, this maa developed from *dzïma through the loss of the first syllable (Hokama 1995: 74). Thorpe reconstructed *ezu(ro) as the proto-Ryukyuan archetype for ‘where’ and ‘which’ (Thorpe 1983: 347), but preservation of -d- in a number of languages probably allows a reconstruction of PR *edu.

102  All these various forms ending in -a probably underwent peculiar irregular developments. They are all likely to be derived from the archetype *eⁿduma through the following processes: *eⁿduma > *endzïma > *ndzïa > ndza; *eⁿduma > *endzïma > *dzïa > dzaa; *enduma > *idua > daa; *enduma > *idua > *ida > *ira > raa.

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

307

NOMINALS Old Ryukyuan 北京お主てだやずまにそなれゆが

FICIN o-SHU-teda ya zuma-ni so nar-i-yu ga Ficin-GEN HON-lord-sun TOP where-LOC next be-CONV-FIN IP Where does the emperor of Beijing live? (RK 3888) Level B: External Comparisons The proto-Japonic archetype should be reconstructed as *entu ‘where’ on the basis of Old Japanese and Ryukyuan data. The most striking external parallel is MK ètúy ‘where.’ Since MK ètúy does not show lenition -t- > -l- in the typical lenition environment V˙Cu/ₒ (Martin 1996: 2–3), it should go back to PK *èntúy, with a nasal sonorant blocking the lenition (Vovin 2005: 89–103). This reconstruction allows us to see that PK *èntúy ‘where’ is in fact bimorphemic, going back to ènú ‘which’ and túy ‘place’: *ènú + túy > *èntúy > ètúy. We do not have similar internal evidence to segment PJ *entu, and this probably indicates that PJ *entu is a loan from Korean. 2.5.5 Interrogative Pronoun iⁿduti The interrogative pronoun iⁿduti ‘where to’ is apparently related to iⁿdu/iⁿduku ‘where’ in the same way as the demonstrative pronouns kǝkǝ and kǝti ‘here’ are related: they share the first element. The second element -ti, as was discussed above, probably represents the word -ti ‘way’ < *tǝy, as evidenced by distal demonstrative pronouns woti and wotǝ. Only one example of iⁿduti in phonographic writing is attested in Western Old Japanese,103 where it is used as an interrogative pronoun referring to direction, and not to stative location.104 伊豆知武伎提可阿我和可留良武

iⁿduti muk-i-te ka a-ŋga wakar-uram-u where face-CONV-SUB IP I-POSS part-TENT2-ATTR facing what direction will I part [with this world]? (MYS 5.887) 103  There are two other examples written logographically as 何処 (MYS 3.287, MYS 7.1412), but it is not clear whether these characters should be read as iⁿduti or iⁿduku. The argument by Omodaka et al. that they must represent *iⁿduti and not *iⁿduku because they denote direction and not location (Omodaka et al. 1967: 83) could have been valid only if we had several reliable examples of iⁿduti in Western Old Japanese all pointing exclusively to direction. Besides, these characters in MYS 3.287 obviously denote location, and not direction: ipe ya mo iⁿduku ‘where is [my] home?’ Takagi et al. read these characters as iⁿduku in both cases (Takagi et al. 1957: 160–161; 1959: 266–267). 104  Note the contrast to the proximal and distal demonstrative pronouns kǝti and woti that in Old Japanese refer only to stative location. Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

308

Chapter 4

Comparative Data Level A: Other Japonic A1: Eastern Old Japanese The interrogative pronoun iⁿduti ‘where to’ occurs only in Western Old Japanese part of the Azuma corpus. The real Eastern Old Japanese form iⁿdusi ‘where to’ is attested only once in Eastern Old Japanese. 伊豆思牟伎弖可伊毛我奈氣可牟

iⁿdusi muk-i-te ka imo-ŋga naŋgɛk-am-u where face-CONV-SUB IP beloved-POSS lament-TENT-ATTR facing what direction, will [my] beloved lament? (MYS 14.3474) A2: Ryukyuan In Ryukyuan I could find formations of interrogative pronouns in -ti/-tʃi attested only in three dialects: Kunigami dzaaNgatʃi, Sesoko raaNgati, and Jashiki dzaatʃi ‘where to’ (Uchima 1984: 45–47). The initial elements dzaa- and possibly raa- as well are related to OJ iⁿdu ‘where’ (see section 2.5.4 above). On reconstruction of the proto-Japonic archetype also see section 2.5.4. Level B: External Comparisons For external parallels see section 2.5.4. 2.5.6 Interrogative Pronoun iⁿdure The interrogative pronoun iⁿdure ‘which’ is attested four times in late Western Old Japanese in phonographic writing. It always appears in the extended form iⁿdure including the obsolete suffix -re 105 found in other pronouns, and in most cases iⁿdure is found followed by the attributive form n-ǝ of the defective verb n- ‘to be.’ Iⁿdure can be used in combination with the interrogative particle ka which emphasizes a question. Its first element iⁿdu- is apparently related to iⁿdu/iⁿduku and iⁿduti ‘where,’ discussed above in sections 2.5.4 and 2.5.5. Yamada Yoshio suggested that iⁿdure ‘which’ and itu ‘when’ should have the same origin (Yamada 1954: 91–92), but this seems unlikely, because itu always appears as /itu/ without the prenasalization -ⁿ-. It is true that iⁿdure is spelled as /iture/ once in Western Old Japanese (see MYS 15.3593 below) and in the only available example in Eastern Old Japanese (see MYS 20.4392 below), but this can be attributed to the occasional inconsistency in spelling prenasalized voiced obstruents in the Old Japanese writing (see chapter 2). 105  There is no unextended stem *iⁿdu meaning ‘which,’ since as was noted in section 2.5.4, iⁿdu means ‘where.’

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

309

NOMINALS 伊都礼乃思麻爾伊保里世武和礼

iⁿdure n-ǝ sima-ni ipor-i se-m-u ware which DV-ATTR island-LOC stay.temporarily-NML do-TENT-FIN I on which island will I stay for a night? (MYS 15.3593) 伊豆礼能日麻弖安礼古非乎良牟

iⁿdure n-ǝ PI-maⁿde are kopï wor-am-u which DV-ATTR day-TERM I long.for(CONV) exist-TENT-FIN until what day should I be longing for [you]? (MYS 15.3742) 伊頭礼乃時加吾孤悲射良牟

iⁿdure n-ǝ TƏKI ka WA-ŋGA kopï-ⁿz-ar-am-u which DV-ATTR time IP I-POSS love-NEG(CONV)-exist-TENT-ATTR which time will I not love [her]? (MYS 17.3891) The interrogative particle ka emphasizes the question. 伊豆礼乎可和枳弖之努波无

iⁿdure-wo ka wakite sinop-am-u which-ACC IP specially long-TENT-ATTR for which [bird] would [we] specially long? (MYS 18.4089) The interrogative particle ka emphasizes the question. Comparative Data Level A: Other Japonic A1: Eastern Old Japanese The interrogative pronoun iⁿdure ‘which’ occurs in Eastern Old Japanese just once in a Sakimori poem. 阿米都之乃以都例乃可美乎以乃良波加

amɛ-tusi-nǝ i[ⁿ]dure n-ǝ kami-wo inǝr-amba ka heaven-earth-GEN which DV-ATTR deity-ACC pray-COND IP if [I] pray to which deities of Heaven and Earth … (MYS 20.4392) The interrogative particle ka emphasizes the question. A2: Ryukyuan The cognates of OJ iⁿdure ‘which’ are well attested in most modern Ryukyuan languages: Wadomari, China ’uduru (Hirayama 1986: 519); Yuwan dërï; Koniya diṛ; Izena diru; Oku, Kijōka, Arumi, Namizato, Toguchi, Kowan, Maejima, Tomigusuku, Tokashiki dʒiru; Uezu duri; Henza dʒuri; Agarinakasone, Yonaha ndʒi; Tonoshiro dʒiri ‘which’ (Uchima and Arakaki 2000: 426). The Modern

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

310

Chapter 4

Standard Shuri dialect has ziru ‘which’ (RKJ 1983: 603). OR idure/[i]zure (probably pronounced as [izïri]/[izuru] or [zïri]/[ziru]) is plentifully attested in various Old Ryukyuan texts (Hokama 1995: 75). Examples: Old Ryukyuan ずれがててとうな

zure-ga te [i]-te to-una which-POSS DV say-SUB ask-NEG/IMP do not ask which [boat] (OS 13.882) Shuri

ziru-Nkai syu ga yaa-N di ’umu-too-N which-LOC do(HORT) IP be-FIN DV think-PERF/PROG-FIN [I] thinking on which [one] to decide (RKJ 1983: 603) The reconstruction of the proto-Japonic archetype should be *entu-re ‘which’ on the basis of Old Japanese and Ryukyuan data. Since this interrogative pronoun is based on the same stem *entu- as iⁿdu/iⁿduku and iⁿduti ‘where,’ for external cognates to *intu- see section 2.5.4 for an external comparison. 2.5.7 Interrogative Pronoun itu The interrogative pronoun itu ‘when’ occurs in Western Old Japanese quite frequently, but examples of it in phonographic writing in earlier Western Old Japanese are rare. One is found in the Harima Fudoki, where 伊都 /itu/ is used as a phonographic gloss to the logographically written character sequence 何 時 ‘when’ (Omodaka et al. 1967: 80). Two others are attested in MYS 5.804 and MYS 5.886, both composed by Yamanoue-no Okura. All other examples are found in later texts. 伊都斯可母京師乎美武

itu si kamǝ MIYAKO-wo mi-m-u when EP EP capital-ACC see-TENT-ATTR when, I wonder, will [I] see the capital? (MYS 5.886) 和伎毛故波伊都登加和礼乎伊波比麻都良牟

wa-ŋg-imo-ko pa itu tǝ ka ware-wo ipap-i mat-uram-u I-POSS-beloved-DIM TOP when DV IP I-ACC pray-CONV wait-TENT2-ATTR

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

311

NOMINALS

My beloved will probably wait for me, praying [to the gods], and thinking: ‘When [will he return]?’ (MYS 15.3659) In this example the interrogative particle ka emphasizes the question. 多都多能山乎伊都可故延伊加武

Tatuta-nǝ YAMA-wo itu ka koye-ik-am-u Tatuta-GEN mountain-ACC when IP cross(CONV)-go-TENT-ATTR When will [I] go, crossing Tatuta mountain? (MYS 15.3722) In this example the interrogative particle ka emphasizes the question. 伊都麻弖可安我故非乎良牟

itu-maⁿde ka a-ŋga kopï-wor-am-u when-TERM IP I-POSS long.for(CONV)-exist-TENT-ATTR until when should I long for [you]? (MYS 15.3749) 宇梅乃花伊都波乎良自

uMƐ-no PANA itu pa wor-aⁿzi plum-GEN flower when TOP break-NEG/TENT when [one] should not pick plum blossoms (MYS 17.3904) 伊都可聞許武

itu kamo kǝ-m-u when EP come-TENT-ATTR I wonder, when will [he] come? (MYS 17.3962) 伊都之加登奈氣可須良牟曾

itu si ka tǝ naŋgɛk-as-uram-u sǝ when EP IP DV lament-HON-TENT2-ATTR FP [she] probably laments: ‘When [will he return]?’ (MYS 17.3962) In this example the interrogative particle ka emphasizes the question. 多麻久之氣伊都之可安氣牟

tama-kusi-ŋ-gɛ itu si ka akɛ-m-u jewel-comb-GEN-container when EP IP dawn/open-TENT-ATTR When will it dawn?/When will one open a jewel comb-box? (MYS 18.4038)106 In this example the interrogative particle ka emphasizes the question. 106  T  ama-kusi-ŋ-gɛ ‘jewel comb-box’ is used as a makura-kotoba to the verb akɛ- ‘to open’ (tr.), ‘to dawn’ (intr.).

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

312

Chapter 4

Comparative Data Level A: Other Japonic A1: Eastern Old Japanese The interrogative pronoun itu ‘when’ is attested in Eastern Old Japanese in two examples in Sakimori poems. With the following focus particle mǝ, itu has the function of a collective pronoun meaning ‘always.’ 都久之閇爾敞牟加流布祢乃伊都之加毛都加敞麻都里弖久爾爾閇牟可毛

Tukusi-pɛ-ni pe muk-ar-u pune-nǝ itu si kamo tukape-matur-i-te kuni-ni pɛ muk-am-o Tukusi-side-LOC bow turn-PROG-ATTR boat-GEN when EP EP serve(CONV)HUM-CONV-SUB province-LOC bow turn-TENT-ATTR I wonder when the boat which is [now] turning [its] bow towards Tukusi will turn its bow towards [my home] province, after [I finish my] service (MYS 20.4359) 以都母以都母於母加古比須須

itu mǝ itu mǝ omǝ-ŋga kopi-susu when FP when FP mother-POSS love(CONV)-COOR [my] mother always, always loves [me] (MYS 20.4386) Here the combination of itu ‘when’ with the focus particle mǝ functions as a collective pronoun ‘always.’ A2: Ryukyuan The cognates to OJ itu ‘when’ are well attested in modern Ryukyuan languages: Wadomari, China ’itʃi; Tokunoshima ’itsï (Hirayama 1986: 74); Yuwan ’itsï; Koniya ’itii; Izena, Oku, Benoki, Kijōka, Arumi, Namizato, Toguchi, Sesoko, Uezu, Henza, Kowan, Maejima, Tomigusuku, Tokashiki ’itʃi; Kumejima ’itʃii; Agarinakasone, Yonaha, Tonoshiro, Hateruma itsï ‘when’ (Uchima and Arakaki 2000: 364), Shuri ’içi ‘when’ (RKJ 1983: 246). In Old Ryukyuan we have itu (いつ) ‘when,’ probably pronounced as [itsï], cf. Korean transcription itco from the Haytong ceykwukki (1501 CE) (Hokama 1995: 72). Old Ryukyuan いつよりもまなしや

itsu-yori mo manasi ya when-ABL FP dear be(FIN) [it] is dear from any time (OS 3.99)

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

NOMINALS

313

Shuri

’içi ca ga when come(PERF) IP When did [he] come? (RKJ 1983: 246) Thorpe reconstructs the proto-Ryukyuan archetype as *etu (Thorpe 1983: 223), and for proto-Japonic we also have to reconstruct *etu ‘when,’ since Japanese raised proto-Japonic *e to *i. I am less convinced by a recent proposal by Whitman to analyze *etu ‘when’ as bimorphemic *e-tu in the same way as *eka ‘how’ < *e-ka (Whitman 2001: 9). While we have internal Japonic evidence for the bimorphemic nature of *e-ka, discussed above (section 2.5.3), it seems to me that such evidence is lacking for *etu, although such possibility definitely exists since the interrogative root *e- can be reconstructed for proto-Japonic. Level B: External Comparisons No apparent external cognates, unless *etu is really segmentable (see section 2.5.3 for external cognates of the interrogative root *e-). 2.5.8 Interrogative Pronoun iku The interrogative pronoun iku ‘how many,’ ‘how much’ in phonographic spelling occurs in Western Old Japanese eight times in seven examples (it appears twice in NK 15). Five of these examples are provided below. In addition, there are two special forms ikura and ikuⁿda ‘how many’ which include the obsolete suffixes -ra and -ⁿda with unclear meaning.107 In phonographic spelling each of these special forms occurs just once. 伊久用加泥都流

iku yo ka ne-t-uru how.many night IP sleep(CONV)-PERF-ATTR how many nights have [we] slept? (KK 25) 於朋望能農之能介瀰之瀰枳伊句臂佐伊句臂佐

opo-monǝ-nusi-nǝ kam-i-si mi-ki iku pisa iku pisa great-thing-master-GEN brew-CONV-PAST/ATTR HON-rice.wine how much long how much long The holy rice wine brewed by [the deity] Great Master of Things, [flourish] eternally (lit.: how much long, how much long) (NK 15) 107  It is possible to speculate that -ra represents plural marker -ra, but the paucity of examples prevents any positive identification. Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

314

Chapter 4

伊久与乎倍弖加

iku yǝ-wo pɛ-te ka how.many generation-ACC pass(CONV)-SUB IP passing through how many generations (MYS 15.3621) 伊久与布流末弖伊波比伎爾家牟

iku yǝ p-uru-maⁿde ipap-i k-i-n-i-k-em-u how.many generations pass-ATTR-TERM pray-CONV come-CONV-PERFCONV-PAST/FIN-TENT-FIN for how many generations did [Ipapîsima] pray? (MYS 15.3637) 伊久欲布等余美都追伊毛波和礼麻都良牟曾

iku yo p-u tǝ yǝm-i-tutu imo pa ware mat-uram-u sǝ how.many night pass-FIN DV count-CONV-COOR beloved TOP I wait-TENT2ATTR FP [My] beloved will probably wait for me, counting: ‘How many nights have passed?’ (MYS 18.4072) 2.5.8.1 Special Forms The peculiarity of the usage of the special forms ikuⁿda and ikura is that they occur with the focus particle mǝ and with the following verb in the negative form, which makes them function as negative collective pronouns with the meaning ‘[not] so many,’ ‘not at all.’ 佐祢斯欲能伊久陀母阿羅祢婆

sa-ne-si yo-nǝ ikuⁿda mǝ ar-an-e-mba PREF-sleep-PAST/ATTR night-GEN how.many FP exist-NEG-EV-CON when not so many nights when [they] slept [together] have passed (MYS 5.804) 左尼始而何太毛不在者

sa-ne-SƏMƐ-TE IKUⁿda mo AR-AN-E-mba PREF-sleep(CONV)-start(CONV)-SUB how.many FP exist-NEG-EV-CON because no [time] at all has passed since [we] started to sleep [together] (MYS 10.2023) 年月毛伊久良母阿良奴爾

TƏSI TUKÏ mo ikura mǝ ar-an-u-ni year month FP how.many FP exist-NEG-ATTR-LOC although not so many years [and] months have passed (MYS 17.3962)

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

315

NOMINALS

Comparative Data Level A: Other Japonic A2: Ryukyuan The Ryukyuan cognates of OJ ikura and MJ ikutu (not attested in OJ) ‘how much,’ ‘how many’ are attested in practically all modern Ryukyuan languages. Old Ryukyuan still has iku ‘how much.’ Some dialects, like Namizato and Kowan do not differentiate between PR *ekura and *ekutu: Namizato tʃassaa, Kowan tʃassa ‘how much,’ ‘how many,’ although most others do, e.g., Kowan ’ikutsï < ikutu ‘how many,’ but ’ikyassaa < ikura ‘how much’ (Uchima and Arakaki 2000: 363). In case of *ekura, all Ryukyuan languages show the same progressive assimilation ky > tʃ > s after /i/ as in the case of *e-ka, described above in section 2.5.3, with the same exception of Benoki kassaa ‘how much’ (cf. Benoki kaaʃi ‘how’).108 Similar to its reflex of kaaʃi ‘how,’ the Benoki dialect does not show any trace of the initial i- in kassaa ‘how much’ either, therefore it is reasonable to suggest that *eku in proto-Japonic was also segmentable like *e-ka ‘how.’ Thus, I reconstruct proto-Japonic archetype as *e-ku ‘how many,’ ‘how much.’ Examples: Old Ryukyuan 幾度ながめゆが

IKU DU nagame-yu ga how.many times look(CONV)-FIN IP how many times will [I] look on [autumn nights]? (RK 4666) Shuri

caQsa-N kooy-uN how.much-FP buy(CONV)-FIN [I] will buy no matter how (much) expensive (RKJ 1983: 142) Level B: External Comparisons The proto-Japonic archetype *e-ku ‘how many,’ ‘how much’ suggests that this interrogative pronoun is related to the interrogative pronoun *e-ka ‘how.’ The element *-ku has an unclear meaning. See section 2.5.3 for external cognates of the interrogative root *e-. 108  Three dialects in the Southern Ryukyus also do not show palatalization in their reflexes of *ikura: Agarinakasone isïka, Yonaha isaka, and Hateruma ikooba. However, Agarinakasone and Yonaha forms obviously exhibit metathesis which blocked the palatalization. In any case, all three of these dialects do not have cognates of OJ ika, therefore we do not have a basis for comparison, in contrast to other dialects. Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

316

Chapter 4

2.5.9 Interrogative Pronoun naⁿdǝ, naⁿzǝ The interrogative pronouns naⁿdǝ, naⁿzǝ ‘why’ are historically derived from nani ‘what’: naⁿdǝ is a contraction of nani ‘what’ and defective verb tǝ ‘to say,’ while naⁿzǝ is a contraction of nani ‘what’ and focus particle sǝ.109 The uncontroversial examples in phonographic spelling include only two examples of naⁿdǝ and one example of naⁿzǝ in Western Old Japanese. Omodaka et al. claim that naⁿdǝ is used with a following verb, while naⁿzǝ is used predominantly with a following adjective (Omodaka et al. 1967: 523). The number of examples in Western Old Japanese is too limited to verify or disprove this claim, but overall I am skeptical: notice the example from MYS 15.3684 below, where naⁿzǝ is used with a following verb. Examples: 那杼佐祁流斗米

naⁿdǝ sak-er-u to mɛ why tattoo-PROG-ATTR sharp eye Why the tattooed sharp eyes? (KK 17) 奈騰可聞妹尓不告來二計謀

naⁿdǝ kamo IMO-ni NƏR-AⁿZ-U K-I-n-i-k-em-u why EP beloved-DAT tell-NEG-CONV come-CONV-PERF-CONV-PAST/ FIN-TENT-ATTR why, I wonder, did [I] come [here] without telling my beloved [the words of good-bye]? (MYS 4.509) 奈曾許許波伊能祢良要奴毛

naⁿzǝ kǝkǝmba i-nǝ ne-raye-n-umo why extremely sleep-GEN sleep-PASS-NEG-EXCL why cannot [I] sleep at all? (MYS 15.3684) In addition, there are two examples in Western Old Japanese where this interrogative pronoun is spelled half-phonographically and half logographically as 奈何. Since the crucial second syllable is spelled logographically with the character 何 ‘what,’ there is no way to tell whether this spelling renders naⁿDƏ or naⁿZƏ.

109  Another etymological possibility is that this sǝ is a stem of the verb se-/sǝ- ‘to do.’ In the light of MJ naze < *naⁿze ‘why’ this etymology is likely to be more plausible.

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

317

NOMINALS 奈何其玉之手二巻難寸

naⁿDƏ?/naⁿZƏ? SƏNƏ TAMA-NƏ TE-ni MAK-I-ŋGATA-ki why that jewel-COMP hand-LOC wind-CONV-difficult-ATTR why is that [jewel bracelet] difficult to wind around the arm that is like a jewel? (MYS 3.409) 其乎奈何不來者來者其乎

SƏ-wo naⁿDƏ?/naⁿZƏ? KƏ-ⁿZ-U pa KƏ-mba SƏ-wo that-ACC why come-NEG-CONV TOP come-COND that-ACC if [you] do not come or if [you] come, why [should I be impatient] about that? (MYS 11.2640) Comparative Data Level A: Other Japonic A1: Eastern Old Japanese The Eastern Old Japanese cognate to the Western Old Japanese naⁿ zǝ ‘why’ is aⁿze, which occurs eight times (seven times in Azuma texts and once in a Sakimori poem). In six Azuma poems and in the Sakimori poem aⁿze means ‘why,’ but in one Azuma poem it means ‘how.’ The interrogative pronoun aⁿze in all examples but one occurs with the following interrogative particle ka, which apparently just emphasizes the question. Eastern Old Japanese aⁿze ‘why’ is derived from the Eastern Old Japanese interrogative pronoun aⁿ- ‘what’ in the same way as Western Old Japanese naⁿzǝ is derived from the Western Old Japanese interrogative pronoun nani ‘what.’ The difference in vocalism of the second syllables in Western and Eastern Old Japanese words can probably be explained by their slightly different morphological composition: WOJ naⁿzǝ consists of nani ‘what’ and the stem sǝ- of the verb se-/sǝ- ‘to do,’ while EOJ aⁿze consists of aⁿ- ‘what’ and the stem se- of the verb se-/sǝ- ‘to do.’ There also might be an alternative phonological explanation: some /e/ in Old Japanese go back to *ǝy, e.g., OJ se ‘back’ < *sǝy, but sǝ-muk- ‘to turn one’s back on somebody.’ Ultimately, these two explanations should probably be combined, as the stem se- of the verb se-/sǝ- ‘to do’ is likely to include historically sequence *ǝy, thus OJ se < *sǝy. The proto-Japanese archetype should be reconstructed as *n-anu sǝ or *n-anï sǝ (see section 2.5.2 for reconstruction of the archetype of nani ‘what’ as *n-anu). 須我麻久良安是加麻可左武

suŋga-makura aⁿze ka mak-as-am-u sedge-pillow why IP pillow-HON-TENT-ATTR why do [you] put your head [on] a sedge pillow? (MYS 14.3369)

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

318

Chapter 4

波比爾思物能乎安是加多延世武

pap-i-n-i-si monǝ-wo aⁿze ka taye se-m-u crawl-CONV-PERF-CONV-PAST/ATTR thing-ACC why IP break(NML) do-TENT-ATTR why do [you] want to break off our relationship (lit.: the thing) which spread [like a vine] (MYS 14.3434) 安是登伊敞可

aⁿze tǝ ip-e ka why DV say-EV IP for what reason (lit.: saying: ‘Why?’) (MYS 14.3461) 阿是曾母許与比与斯呂伎麻左奴

aⁿze sǝ mǝ kǝ yǝpi yǝs-i-rǝ k-i-[i]mas-an-u why FP EP this night approach-CONV-? come-CONV-HON-NEG-ATTR why did not [you] come tonight? (MYS 14.3469) In this example aⁿze ‘why’ is used without the following interrogative particle ka. 比登豆麻等安是可曾乎伊波牟

pitǝ-ⁿ-duma tǝ aⁿze ka sǝ-wo ip-am-u person-GEN-wife DV why IP she-ACC say-TENT-ATTR Why should [I] call her the wife of another? (MYS 14.3472) 安是可多要牟等伊比之兒呂

aⁿze ka taye-m-u tǝ ip-i-si KO-rǝ why IP break.off-TENT-ATTR DV say-CONV-PAST/ATTR girl-DIM the girl who said: ‘Why should [our relationship] be broken off?’ (MYS 14.3513) 安是可加奈思家

aⁿze ka kanasi-ke why IP be dear-ATTR Why is [she so] dear [to me]? (MYS 14.3576) This is a Sakimori poem. In the following example aⁿze means ‘how’ rather than ‘why’:

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

NOMINALS

319

思良久毛能多要爾之伊毛乎阿是西呂等

sira kumo-nǝ taye-n-i-si imo-wo aⁿze se-rǝ tǝ white cloud-COMP break off(CONV)-PERF-CONV-PAST/ATTR beloved-ACC how do-IMP DV what (lit.: how) should [I] do about [my] beloved who separated from [me] like a white cloud (MYS 14.3517) A2: Ryukyuan There are no cognates in Ryukyuan of WOJ naⁿzǝ and EOJ aⁿze ‘why,’ which appears to be an exclusive proto-Japanese formation. Level B: External Comparisons Since proto-Japanese *n-anu sǝ or *n-anï sǝ ‘why’ is derived from proto-Japonic *n-anu ‘what,’ see section 2.5.2 for external cognates. 2.6

Collective Pronouns

There are two collective pronouns in Western Old Japanese: mïna and morǝ ‘all.’ 2.6.1 Collective Pronoun mïna The collective pronoun mïna ‘all’ is attested in phonographic spelling in Western Old Japanese four times. It may refer to both animate and inanimate nouns. When modifying a noun, mïna ‘all’ is always found in postposition. It is interesting to note that the case marker may be found both after the modified noun and after the modifying mïna with no apparent change in meaning as in pitǝ-nǝ mïna and pitǝ mïna-nǝ ‘all people’ found in the examples below. 伽遇破志波那多智麼那辞豆曳羅波比等未那等利

ka-ŋgupasi pana tatimbana siⁿ-du ye-ra pa pitǝ mïna tǝr-i PREF-beautiful flower mandarin.orange bottom-GEN/LOC branch-PLUR TOP person all take-CONV As for the lower branches of beautiful flowering mandarin oranges, all people break [them], and … (NK 35)

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

320

Chapter 4

比等未奈能美良武麻都良能多麻志末乎

pitǝ mïna-nǝ mi-ram-u Matura-nǝ Tamasima-wo person all-GEN see-TENT2-FIN Matura-GEN Tamasima-ACC All people will probably see Tamasima in Matura (MYS 5.862) 可治能於等波於保美也比等能未奈伎久麻泥爾

kaⁿdi-nǝ otǝ pa opo miya-m-bitǝ-nǝ mïna kik-u-maⁿde-ni rudder.oar-GEN sound TOP great palace-GEN-person-GEN all hear-ATTRTERM-LOC until all people in the Great Palace will hear the sound of a rudder oar (MYS 20.4459) 古非波未奈和我宇弊邇於知奴

kopï 110 pa mïna wa-ŋga upɛ-ni oti-n-u love(NML) TOP all I-POSS top-LOC fall(CONV)-PERF-FIN As for love, [it] all has descended on me (NR 1.2) Comparative Data Level A: Other Japonic A1: Eastern Old Japanese Only one example of the collective pronoun mïna ‘all’ is attested in Eastern Old Japanese: 比等未奈乃許等波多由登毛

pitǝ mïna-nǝ kǝtǝ pa tay-u tǝmo person all-GEN word TOP break off-FIN CONJ Even if the words of all people are broken off … (MYS 14.3398) A2: Ryukyuan Cognates of the Old Japanese collective pronoun mïna ‘all’ are attested in several modern Ryukyuan dialects: Uezu, Maejima nna; Kowan n:na; Agarinakasone ṃ:na (Uchima and Arakaki 2000: 410), Shuri ’Nna (RKJ 1983: 436). There is also the Oku form minna, but it looks like a loanword from modern Standard Japanese minna ‘all,’ especially since there is another collective pronoun muttuu in Oku. Muttuu is apparently related to another Old Japanese collective pronoun morǝ ‘all,’ which will be discussed in the section 2.6.2. In Old Ryukyuan the collective pronoun in question is attested only in the Haytong ceykwukki and the Ryūka. It probably was pronounced as mina in the late fifteenth–early 110  Different manuscripts of the Nihon ryōiki show the variation between etymological spelling 古非 /kopï/ and unetymological 古比 /kopi/ for this word. Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

321

NOMINALS

sixteenth century according to the Korean transcription mina (미나) in the Haytong ceykwukki (1501 CE), becoming later ’Nna or ’Nnya, as different existing spellings in the Ryūka indicate: mina (みな), nna (んな), nnya (んにや) (Hokama 1995: 639). Examples: Old Ryukyuan つくづくと浮世みな思て

Tsukuzuku to UCI-YU mina UMU-te Intently DV floating-world all think-SUB Everybody is thinking intently about the floating world (RK 2766) Shuri

shidini naa ’Nna-yaka ‘ippee diki-too-N already already all-ABL very be able-PERF/PROG-FIN [he] is already more skillful than everybody (else) (Nishioka and Nakahara 2000: 26) The reconstruction of the proto-Japonic archetype should be *mǝyna or *muyna. It is possible to speculate that *-na represents a plural marker -na, as was discussed in section 1.2.1.4, while the remainder *mǝy-/*muy- is the root of the collective pronoun. If this is true, and if the original vowel was *ǝ and not *u, it is quite possible that *mǝy can be derived from *mǝrǝy according to Whitman‘s law,111 and is therefore a variant of the another Old Japanese collective pronoun morǝ ‘all’ < *mǝrǝ, which will be discussed in the next section 2.6.2.112 2.6.2 Collective Pronoun morǝ The collective pronoun morǝ ‘all’ is attested four times in phonographic spelling in Western Old Japanese. Besides, there are a number of examples when it is written logographically with the Chinese character 諸 as in the example 111  The possibility of an etymological relationship between OJ mïna and morǝ is further supported by the congruency in their accent register (both belong to high register accent classes): mïna is 2.2b (HL) and morǝ is 2.1 (HH). Martin further suggests that mïna 2.2b could go back to 2.1 (HH) (Martin 1987: 479). 112  Martin also treats mï-as a possible derivation from *morǝ-Ci ‘many,’ but interprets -na as ‘person’ (Martin 1987: 479), which occurs as a second element in the compounds okina ‘old man’ and oto-na ‘adult.’ He further adds wo-mina ‘woman’ (Martin 1987: 490), but -mina probably means ‘woman’ by itself, as suggested by Whitman (cf. OJ me ‘woman’ < *mina) (Whitman 1985: 238). Since mïna ‘all’ can refer also to inanimate nouns, it is more likely that -na is a plural marker rather than a bound noun with the meaning ‘person.’ Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

322

Chapter 4

from SM 17 below. In the phonographic spelling it is predominantly written as 毛呂 /morǝ/ with the kō-rui syllabic sign 毛 /mo/. The spelling 母呂 /mǝrǝ/ occurs only once in the MYS 5.832, once in the Bussoku seki no uta (in the reduplicated form spelled morǝ-mǝrǝ, see BS 18 below) and in Eastern Old Japanese. However, the contrast between /mo/ and /mǝ/ is generally believed to be lost after the Kojiki kayō, which is the only text that preserves it consistently. Therefore, the spelling in later texts may be unetymological, confusing kō-rui 毛 /mo/ and otsu-rui 母 /mǝ/.113 The crucial point here is, however, the well-known fact that the kō-rui vowel /o/ and the otsu-rui vowel /ǝ/ cannot combine within the same morpheme, the only possible sequences being o-o and ǝ-ǝ, e.g., momo ‘hundred,’ but kǝkǝrǝ ‘heart.’ Since the contrast between the kō-rui syllable /ro/ and the otsu-rui syllable /rǝ/ was much more long-lived than the contrast between /mo/ and /mǝ/, we must arrive at the inescapable conclusion that the original vocalism is preserved in the second syllable /rǝ/ of the word in question. Given the above-mentioned constraint on distribution of /o/ and /ǝ/, the pre-Old Japanese form of the word must be reconstructed as *mǝrǝ, which is further supported by Ryukyuan data (see below). There is a tendency to gloss OJ morǝ as ‘various,’ ‘several’ (Martin 1987: 485), or as ‘several’ in addition to ‘all’ (Omodaka et al. 1967: 752), or even puzzling ‘two’ in addition to ‘many’114 (Ōno et al. 1990: 1331). However the data do not support this interpretation: all examples below indicate that OJ morǝ means just ‘all,’ with a possible exception of MYS 5.843, where the meaning ‘many’ can be argued for. The example from BS 1 below is especially important, since the Buddhist idea of salvation extends to all sentient beings. The fact that OJ morǝ is a collective pronoun ‘all’ is further supported by the Ryukyuan data, where its cognate muru also functions as a collective pronoun. It is quite possible to believe that the collective pronoun developed from the original word meaning ‘many’ on the basis of external comparisons, but this does not necessarily exactly define the function of this word in Old Japanese. Contrary to the collective pronoun mïna ‘all,’ discussed above in section 2.6.1, the collective pronoun morǝ ‘all’ almost always is found preceding the noun it modifies. The only apparent exception is SM 17 below, but since the pronoun in question is written there logographically with the Chinese

113  John Bentley demonstrated that in MYS 5 the difference between kō-rui /mo/ and otsu-rui /mǝ/ is preserved to a certain extent, but only statistically (Bentley 1997). 114  This ‘two’ is obviously inspired by turuŋgi tati morǝ-pa ‘two blades of a sword’ in MYS 11.2636. Unfortunately, sword cannot have more than two blades, so ‘two blades’ are certainly ‘all blades.’

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

323

NOMINALS

character 諸, one cannot be completely sure that this character must be read as morǝ, and not as mïna. It is also possible that there might be a slight functional difference between mïna ‘all’ and morǝ ‘all’ in Western Old Japanese, if it is not the effect of attestation. While the former is used with both animate and inanimate nouns, the latter refers only to animate beings, at least in the examples written phonographically, where we can be completely sure that we deal with morǝ, and not with mïna. There is no such a difference in Eastern Old Japanese, where morǝ can refer to both animate and inanimate nouns. 母呂比得波家布能阿比太波多努斯久阿流倍斯

mǝrǝ pitǝ pa kepu-nǝ apiⁿda pa tanosi-ku ar-umbɛ-si all person TOP today-GEN interval TOP be joyful-CONV exist-DEB-FIN all people today must be joyful (MYS 5.832) 宇梅能波奈乎理加射之都都毛呂比登能阿蘇夫遠美礼婆弥夜古之叙毛布

uMƐ-nǝ pana wor-i kaⁿzas-i-tutu morǝ pitǝ-nǝ asomb-u-wo mi-re-mba miyako si ⁿzǝ [o]mop-u plum-GEN flower break.off-CONV put.in.the.hair-CONV-COOR all person-GEN play-ATTR-ACC see-EV-CON capital EP FP think-ATTR When [I] see that all people enjoy themselves breaking off plum blossoms and putting [them] in [their] hair, [I] think of the capital (MYS 5.843) 美阿止都久留伊志乃比鼻伎波阿米爾伊多利都知佐閇由須礼知知波波賀多 米爾毛呂比止乃多米爾

mi-atǝ tukur-u isi-nǝ pimbik-i pa amɛ-ni itar-i tuti sapɛ yusur-e titi papa-ŋga tamɛ n-i morǝ pitǝ-nǝ tamɛ n-i HON-footprint make-ATTR stone-GEN echo-NML TOP heaven-LOC reachCONV earth RP shake-EV father mother-POSS for DV-CONV all person for DV-CONV The echo of the stone, where [I] carved the footprints [of the Buddha], reaches the Heaven, and shakes the earth as well, for father and mother, for all people (BS 1)115 115  R  oy A. Miller translates this poem as ‘The ringing of the rock on which the holy footprints are cut reaches to heaven and even the earth resounds: for father and mother, for all men’ (Miller 1975: 75). To give Miller partial credit, he mentions that the verb yusur- ‘is clearly transitive’ in MYS 1239, but then he obscures that matter with a long and irrelevant discussion of examples from Middle Japanese, which he clearly misunderstands (Miller 1975: 77), and ends up translating yusur- as an intransitive verb, in spite of the fact that all extant Old Japanese examples of yusur- are transitive. This is further aggravated by

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

324

Chapter 4

吾大王乃毛呂比登乎伊射奈比多麻比

WA-ŋGƏ OPƏ KIMI-nǝ morǝ pitǝ-wo iⁿzanap-i-tamap-i I-POSS great lord-GEN all person-ACC induce-CONV-HON-CONV Our great lord induced all people … (MYS 18.4094) 汝多知諸者吾近姪奈利

IMASI-tati MORƏ pa WA-ŋGA TIKA-KI WOPI nar-i you-PLUR all TOP I-POSS close-ATTR nephew be-FIN All [of] you are my close nephews (SM 17) 2.6.2.1 Special Form The collective pronoun morǝ ‘all’ also occurs in a special reduplicated form morǝ-morǝ ‘all.’ There is no apparent difference in meaning. 毛呂毛呂須久比和多志多麻波奈

morǝ-morǝ sukup-i-watas-i-tamap-ana all-all save-CONV-carry.across-CONV-HON-DES [I] wish [the Buddha’s footprints] will save and lead everybody across (BS 4) 与伎比止乃伊麻須久爾爾波和礼毛麻胃弖牟毛呂毛呂乎為弖

yǝ-ki pitǝ-nǝ imas-u kuni-ni pa ware mo mawi-te-m-u morǝ-morǝ-wo wi-te good-ATTR person-GEN exist(HON)-ATTR land-LOC TOP I FP go(HUM) (CONV)-PERF-TENT-FIN all-all-ACC lead(CONV)-SUB I would also have gone to the land where the Buddha (lit.: good person) resides, leading everybody (BS 8) Miller’s mistaken belief that the particle sapɛ ‘even’ is involved in kakari-musubi, triggering the final verbal form to acquire the evidential form (Miller 1975: 78); if Miller bothered to check all existing Old Japanese examples of sapɛ usage, he would undoubtedly discover that this is not the case. In addition, among multiple philological mistakes one is especially revealing: the translation of yusur- as Chinese dòng (動) ‘to move’ (tr.) does not go back to the Tokugawa period scholar Kariya Ekisai as Miller alleges (Miller 1975: 77), but to the Heian period dictionary Ruijū myōgishō (1081 CE). Listing the much more superior and better translation of the Bussoku seki no uta by Douglas E. Mills (Mills 1960) as one of the three major disasters that has befallen the stone with the poems (Miller 1975: 3–5), Miller apparently forgot to mention the greatest of them all: his own ad hoc interpretations and translations of the poems. In fact, the only evidence of understanding yusur-e as an evidential form rather than an imperative, is the use of the topic particle pa after pimbik-i ‘echo-NML,’ which would seem to be a strange device, if the form yusur-e would be indeed an imperative.

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

325

NOMINALS 都止米毛呂毛呂須須賣毛呂母呂

tutǝmɛ morǝ-morǝ susum-e morǝ-mǝrǝ strive(IMP) all-all go forward-IMP all-all strive, everybody, go forward, everybody (BS 18) Comparative Data Level A: Other Japonic A1: Eastern Old Japanese The collective pronoun morǝ ‘all’ is attested in Eastern Old Japanese only twice: once in a basic form mǝrǝ and once in a reduplicated form mǝrǝ-mǝrǝ. In Eastern Old Japanese morǝ can refer to both animate and inanimate nouns, as can be seen from the following examples. 武蔵野乃久佐波母呂武吉

MUSASI-NO-nǝ kusa pa mǝrǝ muk-i Musasi-field-GEN grass all face-CONV The grass on the Musasi plain faces all [sides]… (MYS 14.3377) 母呂母呂波佐祁久等麻乎須

mǝrǝ-mǝrǝ pa sake-ku tǝ mawos-u all-all TOP safe-CONV DV say(HUM)-FIN [I] will ask [the deities] that everbody [would return] safely (MYS 20.4372) A2: Ryukyuan The cognates of the Old Japanese pronoun morǝ ‘all’ are well attested in the modern Ryukyuan languages: Kikaigashima muru/munu ‘all’ (Iwakura 1941: 302), Izena, Benoki, Kijōka, Arumi, Namizato, Toguchi, Sesoko, Uezu, Henza, Kowan, Maejima, Tomigusuku, Tokashiki, Kumejima muru; Tonoshiro, Hateruma muuru ‘all’ (Uchima and Arakaki 2000: 410), Shuri muru ‘all’ (RKJ 1983: 395). The collective pronoun muru ‘all’ is also attested in Old Ryukyuan (spellings: moro (もろ), muru (むる), moru (もる)) (Hokama 1995: 677). Examples: Old Ryukyuan 野辺のもろ人も雪の噂

NU-BI-nu moro-BITU mo YUCI-no UWASA field-(GEN)side-GEN all-(DV-ATTR)person FP snow-GEN rumor all people at the field, too, are talking about snow (RK 2673)

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

326

Chapter 4

Shuri

muru mii-too-N all grow-PERF/PROG-FIN everything has grown (RKJ 1983: 395) The crucial Northern Amami attestations such as Kikaigashima muru/munu ‘all’ show that the word clearly goes back to PJ *mǝrǝ, since in Northern Amami PJ *o (> OJ /o/) is reflected as /o/, and PJ *ǝ (> OJ /ǝ/) is reflected as /u/.

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

Section 3

Numerals Regarding the numerical system, Western Old Japanese differs considerably from Middle Japanese and Modern Japanese in two respects. First, in sharp contrast to the later stages of the Japanese language history, the numerals of Chinese origin are not attested in the texts. It is quite possible that they already penetrated the colloquial language of the time, but this cannot be proven either way. The overwhelming majority of Old Japanese texts are poetry, and the poetic language was not generally integrating contemporary loanwords from Chinese at least until the twelfth century. Second, as a partial sequence of the lack of the Chinese numerals in the system,116 the ordinal numerals as a separate class are practically non-existent, with the major exception of the ordinal numeral patu ‘first.’ There is no distinction between cardinal and ordinal numerals among the numerals of Japanese origin, which are all essentially cardinal, but also can be theoretically used in an ordinal function. The cardinal and ordinal functions of native Japanese numerals should have been differentiated exclusively by context, but in Old Japanese texts there are in fact no examples of cardinal numerals used in the ordinal function (in the phonographic spelling). It seems that the classifier system so typical for the modern Japanese language was still in the process of formation in Old Japanese. Although Old Japanese possesses several ‘real’ classifiers that are bound nouns used only in combination with the preceding numeral, there are also regular nouns that are used in free form elsewhere that can assume the classifier function. Needless to say, there are no classifiers of Chinese origin attested in the texts. Besides, numerals can be used with nouns without the intervening classifiers, which is atypical for the modern language, where such formations mostly represent idiomatic expressions. In the Japanese linguistic tradition, it is considered that a number of numerals, such as ya ‘eight,’ yaso ‘eighty,’ ipo ‘five hundred,’ yapo ‘eight hundred,’ ti ‘thousand,’ ya-ti ‘eight thousand,’ and yǝrǝndu ‘ten thousand’ in most cases meant just ‘many.’ It is a possibility, but it leaves the question why there were so many numeral forms with the same meaning ‘many’ unanswered. In the case of ya ‘eight’ and its derivatives, it is possible to surmise that the number ‘eight’ carries the magic meaning of unity because it is a sum of the number ‘three,’ which symbolises female, and the number ‘five,’ which symbolises male. It is 116  The suffix -me, marking native Japanese ordinal numerals in modern Japanese appeared only in the Early Modern Japanese.

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

328

Chapter 4

less apparent in the case of other numerals, and it might be safer to treat them at face value before we have an exhaustive socio-cultural-historical explanation for this usage. 3.1

Cardinal Numerals

3.1.1 Cardinal Numeral pitǝ ‘One’ The numeral pitǝ ‘one’ can occur with and without the following classifiers. Examples: 夜麻登能比登母登須須岐

Yamatǝ-nǝ117 pitǝ-mǝtǝ susuki Yamato-GEN one-CL silver grass one silver grass in Yamato (KK 4) 袁都能佐岐那流比登都麻都比登都麻都阿勢袁

Wotu-nǝ saki-n-ar-u pitǝ-tu matu pitǝ-tu matu a se wo Wotu-GEN cape-LOC-exist-ATTR one-CL pine one-CL pine I beloved EP a lone (lit.: one) pine that stands at the cape of Wotu. Oh, lone (lit.: one) pine, my beloved! (KK 29) 夜多能比登母登須宜波古母多受

Yata-nǝ pitǝ-mǝtǝ suŋgɛ pa ko mǝt-aⁿz-u Yata-GEN one-CL sedge TOP child hold-NEG-FIN One sedge from Yata does not have children (KK 64) 等之爾安里弖比等欲伊母爾安布比故保思

tǝsi-ni ar-i-te pitǝ yo imǝ-ni ap-u Pikoposi year-LOC exist-CONV-SUB one night beloved-DAT meet-ATTR Altair Altair, who meets [his] beloved one night in a year (MYS 15.3657)

117  Tsuchihashi treats Yamatǝ as yama-tǝ ‘mountain place’ (Tsuchihashi 1957: 39–40). However, ‘place’ is to with kō-rui vowel /o/, not otsu-rui /ǝ/. Ogihara adopts the same explanation, noting that the syllabic sign with otsu-rui vowel is a ‘mistake’ here (Ogihara 1973: 104–105). One is just left to wonder why there are no other mistakes substituting / to/ with /tǝ/ in the Kojiki kayō text. Thus, I adopt a text-based interpretation of Yamatǝ as ‘Yamato,’ the only one which seems to be feasible in this context.

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

329

NOMINALS

Comparative Data Level A: Other Japonic A1: Eastern Old Japanese The numeral pitǝ ‘one’ is attested once in Eastern Old Japanese without the following classifier (see the examples with classifiers under the respective classifiers in section 3.3). 比登祢呂爾伊波流毛能可良

pitǝ ne-rǝ n-i ip-ar-u monǝkara one peak-DIM DV-CONV say-PROG-ATTR CONJ Although [they] say that [we] are one peak … (MYS 14.3512) A2: Ryukyuan As a rule, in modern Ryukyuan languages all digit numerals that take the classifier -tu in Old Japanese are attested only in a bound form with this classifier, although the stems of these numerals could be found in compounds, like in Modern Standard Japanese. All Ryukyuan dialects in Northern and Central Ryukyus have reduced form of the numeral *pitǝ-tu ‘one’ which has truncated the first syllable *pi-: Shuri tiiçi (RKJ 1983: 517),118 Yuwan t‘iitsï, Izena t‘iitʃi, Uezu tiitʃi etc. ‘one’ (Uchima and Arakaki 2000: 442). However, in the Sakishima islands the untruncated forms are still found: Agarinakasone, Yonaha psï�t̥ iitsï, Tonoshiro pï�t̥ iidzï, Hateruma pï�t̥ sï�t̥ sï ‘one’ (Uchima and Arakaki 2000: 442). It is believed that the Old Ryukyuan form was *hwitu-çi (Hokama 1995: 565). Although Hokama does not provide examples of phonographic spellings in his dictionary of Old Ryukyuan (Hokama 1995), some can be found in the Ryūka.119 Combined Southern Ryukyuan and Old Ryukyuan data constitute a good basis for reconstructing PR *pitu- ‘one.’ There are also the Old Ryukyuan form fitu-ri and Shuri written form hwicui (colloquial cui) ‘one person,’ representing a combination of the numeral *pitu- ‘one’ with the classifier -ri ‘person,’ that offer further support to this reconstruction. Examples: Old Ryukyuan ふてつたむななのおとぢや

fute-tsu ta mu nana n-o otodiya one-CL two six seven DV-ATTR brother one, two, six, seven brothers (OS 13.898) 118  The archaic form hwituçi ‘one’ is also found in Shuri, but only in the written language (RKJ 1983: 241b–242). 119  There is also an intriguing form futetsu ‘one’ in the OS 13.898. Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

330

Chapter 4

月やひとつ

TSUCHI ya fito-tsu moon TOP one-CL [there is only] one moon (RK 485) ただひとり

tada fito-ri just one-CL just one person (OS 14.997) Shuri

sjuutee tii-çi nay-uN household one-CL become-FIN [Members of] the household represent one unit (RKJ 1983: 517) 3.1.2 Cardinal Numeral puta ‘Two’ The cardinal numeral puta ‘two’ can occur with or without the following classifiers. Examples: 美多迩布多和多良須阿治志貴

mi-tani puta watar-as-u Aⁿdisikï HON-valley two cross-HON-ATTR Aⁿdisikï Aⁿdisikï [deity], who crosses two valleys (KK 6) 袁夜迩須賀多多美伊夜佐夜斯岐弖和賀布多理泥斯

wo-ya-ni suŋga-tatami iya-saya sik-i-te wa-ŋga puta-ri ne-si DIM-house-LOC sedge-mat more(?)-rustling spread-CONV-SUB we-POSS two-CL sleep(CONV)-PAST/ATTR in the little hut, spreading very rustling sedge mats, two of us slept (KK 19) 阿波泥辞摩異椰敷多那羅弭阿豆枳辞摩異椰敷多那羅弭予呂辞枳辞摩之魔

Apaⁿdi-sima iya puta naramb-i Aⁿduki-sima iya puta naramb-i yǝrǝsi-ki sima-sima Apaⁿdi-island perfectly two line.up-CONV Aⁿduki-island perfectly two line. up-CONV good-ATTR island-island Apaⁿdi island perfectly lines up [with another island as] two, Aⁿduki island perfectly lines up [with another island as] two, [they are] good islands (NK 40)

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

331

NOMINALS 麻多麻奈須布多都能伊斯乎

ma-tama-nasu puta-tu n-ǝ isi-wo INT-jewel-COMP two-CL DV-ATTR stone-ACC two stones like real jewels (MYS 5.813) 弥蘇知阿麻利布多都乃加多知夜蘇久佐等曾太礼留比止

miso-ti amar-i puta-tu n-ǝ katati yaso kusa tǝ sǝⁿdar-er-u pitǝ thirty-CL exceed-CONV two-CL DV-ATTR mark eighty type DV be completePROG-ATTR person a person, who is complete with thirty-two marks and eighty [lesser sign] types (BS 2) Comparative Data Level A: Other Japonic A1: Eastern Old Japanese In Eastern Old Japanese the numeral puta ‘two’ can be used with and without a following classifier, as the example below demonstrates: 奴麻布多都可欲波等里賀栖安我己許呂布多由久奈母

numa puta-tu kayop-a tǝri-ŋga su a-ŋga kǝkǝrǝ puta yuk-unam-ǝ marsh two-CL go.over-ATTR bird-POSS nest I-POSS heart two go-TENT2-ATTR Should my heart go [to] two [different places like] nests of birds that go over two marshes? (MYS 14.3526) A2: Ryukyuan For the most part, the same development as that of the *pitǝ ‘one,’ outlined above in 4.3.1.1, befell the *puta ‘two’ in Ryukyuan: in the modern North Ryukyuan dialects the first syllable was lost, resulting in such forms as Yuwan t‘aatsï, Izena t‘aatʃi, Uezu taatʃi (Uchima and Arakaki 2000: 442), Shuri taaçi (RKJ 1983: 501) etc. ‘two.’ Meanwhile, the original disyllabic form is well preserved in the Southern Ryukyuan dialects: Agarinakasone, Yonaha fu̥ taatsï; Tonoshiro Fu̥ taadzï; Hateruma ɸu:taatsï ‘two’ (Uchima and Arakaki 2000: 442). Old Ryukyuan here appears of less help, since almost all the examples of the numeral ‘two’ in the Ryūka are not in the phonographic spelling, and only the older text of the Omoro sōshi has a clear example of ta [taa] in phonographic spelling. Examples:

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

332

Chapter 4

Old Ryukyuan ふてつたむななのおとぢや

fute-tsu ta mu nana n-o otodiya one-CL two six seven DV-ATTR brother one, two, six, seven brothers (OS 13.898) Shuri

taa-çi sy-uN two-CL do-FIN to double (lit.: make [it] two) (RKJ 1983: 501) 3.1.3 Cardinal Numeral mi ‘Three’ The cardinal numeral mi ‘three’ occurs in Western Old Japanese in phonographic spelling only with the following general classifier -tu ‘thing’ in two almost identical contexts in the Kojiki kayō. However, we can conclude that it was used independently without the following classifiers before the noun like the cardinal numerals pitǝ ‘one’ and puta ‘two’ on the basis of several pieces of oblique evidence: 1) it appears in logographic writing as a Chinese character in the poetic texts where it cannot be read as bisyllabic Sino-Japanese /samu/ < MC sam (三) ‘three’ on the basis of poetic meter; 2) it is attested in combinations ‘three + noun’ that are verified by Heian period texts as including mi ‘three’ without the following classifier -tu, as in the mi tǝse ‘three years’ below; 3) it is attested phonographically in the proper name Miwa (lit.: mi-wa ‘three rings’), occurring in NK 16 and NK 17 in the phrase Miwa-nǝ tǝnǝ (瀰和能等能) ‘the shrine of Miwa’; 4) mi is attested in the numeral mi-so-ti ‘thirty’ (three+ten+classifier) written phonographically in BS 2 as 弥蘇知 (see section 3.1.2 above); 5) the kun-yomi /mi/ of the Chinese character 三 ‘three’ is used to render OJ syllable /mi/, as, for example, in NA-mi (無三) ‘there is no’ (no-GER) in MYS 4.714. 美都具理能曾能那迦都迩

mi-tu-ŋ-guri-nǝ sǝnǝ naka-tu ni three-CL-DV(ATTR)-chestnut-COMP that middle-GEN/LOC clay that clay from the middle that is like three chestnuts (KK 42) 美都具理能那迦都延

mi-tu- ŋ-guri-nǝ naka-tu ye three-CL-DV(ATTR)-chestnut-COMP middle-GEN/LOC branch the middle branches that are like three chestnuts (KK 43) Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

333

NOMINALS 三歳之間爾

MI TƏSE-NƏ POⁿDƏ-ni three year-GEN time-LOC in the interval of three years (MYS 9.1740) Comparative Data Level A: Other Japonic A2: Ryukyuan The reflexes of the cardinal numeral ‘three’ are attested in all Ryukyuan languages without any irregular developments: Yuwan miitsï; Izena miitʃi; Agarinakasone, Hateruma miitsï (Uchima and Arakaki 2000: 442); Shuri miiçi (RKJ 1983: 372) ‘three,’ etc. The cardinal numeral mitu [miçi] ‘three’ is also attested in Old Ryukyuan. Examples: Old Ryukyuan あおりみつたてて

aori mi-tu tate-te sun umbrella three-CL put(CONV)-SUB putting up three sun umbrellas (OS 19.1321) Shuri

’ikuçi-N mii-çi-N how.many-FP three-CL-FP Many (lit.: even how many, even three) (RKJ 1983: 372) Level B: External Comparisons The proto-Japonic archetype can be reconstructed as *mi- ‘three.’ The only reliable cognate seems to be in the pseudo-Koguryǒ language, which has 密 (EMC mit, LHC *mrit ‘three’). If the use of the Old Chinese pronounciation for this Koguryǒ numeral is correct,120 the pre-proto-Japonic form should be reconstructed as *miri-. No other external parallels to this numeral exist to the best of my knowledge.

120  There seem to be many instances in the ‘Koguryo’ language when the Late Han Chinese rather than the Late Middle Chinese reading is applicable (see, e.g., Vovin 1999).

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

334

Chapter 4

3.1.4 Cardinal Numeral yǝ ‘Four’ It appears that the numeral yǝ ‘four’ in phonographic spelling is found in original Western Old Japanese texts (not in later glosses121) only in one example in combination with the following classifier -tu ‘thing’: 与都乃閇美伊都都乃毛乃乃阿都麻礼流伎多奈伎微乎婆

yǝ-tu n-ǝ pɛmi itu-tu n-ǝ monǝ-nǝ atumar-er-u kitana-ki mï-womba four-CL DV-ATTR snake five-CL DV-ATTR demon-GEN gather-PROG-ATTR dirty-ATTR body-ACC(EMPH) the dirty body which accumulates four snakes and five demons (BS 19) However, there is oblique evidence that the numeral yǝ could be used independently in Western Old Japanese, since the kun-yomi /yǝ/ of the Chinese character 四 ‘four’ is used as kungana to write the imperative suffix -yǝ: 愛寸事盡手四

URUPASI-ki KƏTƏ TUKUS-I-te-yǝ splendid-ATTR word exhaust-CONV-PERF-IMP exhaust [your] splendid words (MYS 4.661) Comparative Data Level A: Other Japonic A2: Ryukyuan The reflexes of the cardinal numeral ‘four’ are attested in all Ryukyuan languages without any irregular developments: Yuwan yuuttsï; Izena yuutʃi; Agarinakasone, Hateruma yuutsï (Uchima and Arakaki 2000: 442); Shuri ’yuuçi (RKJ 1983: 294) ‘four,’ Yonaguni duutʃi (Hirayama 1967: 227) etc. The cardinal numeral yo-tu ‘four’ is also attested in Old Ryukyuan, although there seems to be no examples with phonographic spelling. 3.1.5 Cardinal Numeral itu ‘Five’ The numeral itu ‘five’ in phonographic spelling is found in original Western Old Japanese texts (not in later glosses) only in one example in combination with the following classifier -tu ‘thing’:

121  For attestation in later glosses see Omodaka et al. 1967: 791, 799.

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

335

NOMINALS 与都乃閇美伊都都乃毛乃乃阿都麻礼流伎多奈伎微乎婆

yǝ-tu n-ǝ pɛmi itu-tu n-ǝ monǝ-nǝ atumar-er-u kitana-ki mï-womba four-CL DV-ATTR snake five-CL DV-ATTR demon-GEN gather-PROG-ATTR dirty-ATTR body-ACC(EMPH) the dirty body which accumulates four snakes and five demons (BS 19) There is no phonographic evidence that itu ‘five’ could be used independently without classifiers, but we still can speculate it could on the basis of analogy with other numerals. Comparative Data Level A: Other Japonic A1: Eastern Old Japanese The cardinal numeral itu ‘five’ is attested once in Eastern Old Japanese: 和加加都乃以都母等夜奈枳

wa-ŋga katu-nǝ itu-mǝtǝ yanaŋgi I-POSS gate-GEN five-CL willow five willow trees at my gate (MYS 20.4386) A2: Ryukyuan The reflexes of the cardinal numeral ‘five’ are attested in all Ryukyuan languages without any irregular developments: Yuwan Ɂitsïsï; Izena Ɂitʃi̥tʃi; Agarinakasone itsï�t̥ sï; Hateruma issïï (Uchima and Arakaki 2000: 442); Shuri Ɂiçiçi (RKJ 1983: 246) ‘five,’ etc. The cardinal numeral itu-tu [iti-ti] ‘five’ is also attested in Old Ryukyuan. Old Ryukyuan さしふいつつころ

sasifu itu-tu koro priestess five-CL person five priestesses (OS 6.312) Level B: External Comparisons The proto-Japonic archetype can be reconstructed as *itu- ‘five’ on the basis of both Old Japanese and Ryukyuan data. The only more or less reliable external comparison is pseudo-Koguryǒ 于次 *utsï ‘five.’

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

336

Chapter 4

3.1.6 Cardinal Numeral mu ‘Six’ The cardinal numeral mu ‘six’ is not attested in phonographic writing in Western Old Japanese texts, but there is oblique evidence for its existence, since the kun-yomi reading /mu/ of the Chinese character 六 ‘six’ is used as kungana to write syllable /mu/ on a number of occasions:122 妹名根之作服異六白細乃紐

IMO NA-ne-ŋGA TUKUR-I KI-SE-k-em-u SIRO TAPƐ-nǝ PIMO beloved you-?-POSS make-CONV wear-CAUS-PAST/FIN-TENT-ATTR white cloth-GEN cord the cord of the white cloth that you, [my] beloved, made and made [me] wear (MYS 9.1800) Comparative Data Level A: Other Japonic A2: Ryukyuan The reflexes of the cardinal numeral ‘six’ are attested in all Ryukyuan languages without any irregular developments: Yuwan muutsï; Izena muutʃi; Agarinakasone mmtsï; Hateruma nntsï (Uchima and Arakaki 2000: 442); Shuri muuçi (RKJ 1983: 397) ‘six,’ etc. The cardinal numeral mu ‘six’ is also attested in Old Ryukyuan. Old Ryukyuan ふてつたむななのおとぢや

fute-tsu ta mu nana n-o otodiya one-CL two six seven DV-ATTR brother one, two, six, seven brothers (OS 13.898) 3.1.7 Cardinal Numeral nana ‘Seven’ The cardinal numeral nana ‘seven’ is attested in Western Old Japanese in phonographic spelling in three examples with and without the following classifiers. The only classifier attested phonographically with nana ‘seven’ is -pe ‘layer.’ 多加佐士怒袁那那由久袁登賣杼母

Takasaⁿzi-no-wo nana yuk-u wotǝme-ⁿdǝmǝ Takasaⁿzi-field-ACC seven go-ATTR girl-PLUR seven girls going across the Takasaⁿzi field (KK 15) 122  See also MYS 3.302 and MYS 12.2940 for examples of the same usage.

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

337

NOMINALS 飫瀰能古簸多倍能波伽摩鳴那那陛鳴糸

omi-nǝ ko pa tapɛ-nǝ pakama-wo nana-pe wos-i noble-GEN child TOP mulberry.tree.bark.cloth-GEN pants-ACC seven-CL wear(HON)-CONV The son of a noble wears seven layers of mulberry tree bark cloth pants, and … (NK 74) 麻都良我波奈奈勢能與騰

Matura kapa nana se-nǝ yǝⁿdǝ Matura river seven rapid-GEN pool Pools at the seven rapids of the Matura river (MYS 5.860) Comparative Data Level A: Other Japonic A1: Eastern Old Japanese The cardinal numeral nana ‘seven’ is attested in Eastern Old Japanese only once with the following classifier -pe ‘layer.’ 志毛用爾奈奈弁加流去呂毛

simo yo-ni nana-pe k-ar-u kǝrǝmo frost night-LOC seven-CL wear-PROG-ATTR garment in the frosty night [I] wear seven layers of garments (MYS 20.4431) A2: Ryukyuan The reflexes of the cardinal numeral ‘seven’ are attested in all Ryukyuan languages without any irregular developments: Yuwan nanatsï; Izena nanatʃi; Agarinakasone, Hateruma nanatsï (Uchima and Arakaki 2000: 443); Shuri nanaçi (RKJ 1983: 409) ‘seven,’ etc. The cardinal numeral nana ‘seven’ is also attested in Old Ryukyuan. Old Ryukyuan ふてつたむななのおとぢや

fute-tsu ta mu nana n-o otodiya one-CL two six seven DV-ATTR brother one, two, six, seven brothers (OS 13.898) 3.1.8 Cardinal Numeral ya ‘Eight’ The cardinal numeral ya ‘eight’ is attested in Western Old Japanese in phonographic spelling in three examples with and without the following classifiers.

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

338

Chapter 4

Traditionally it is considered that ya ‘eight’ may mean just ‘many,’ however some linguists who support this point of view, (e.g., Omodaka et al. 1967: 754), overlook the important fact that the number ‘eight’ is the sum of ‘three’ (female number) and ‘five’ (male number). Therefore, its meaning in KK 1 (in essence, a bridal song) below as well as in other examples should be just ‘eight’ (as a symbol of marital union). Still, there might be examples where ya means ‘many’ rather than ‘eight,’ e.g., NK 91 below. It is more likely that a woven brushwood fence had many meshes rather than just ‘eight,’ however, we do not know for sure since no samples of these fences from the Asuka or Nara period survive. 夜久毛多都伊豆毛夜弊賀岐都麻碁微爾夜弊賀岐都久流

ya kumo tat-u Iⁿdumo ya-pe-ŋ-gaki tuma-ŋ-gǝmï-ni ya-pe-ŋ-gaki tukur-u eight cloud rise-ATTR Iⁿdumo eight-CL-DV(ATTR)-fence spouse-GEN-be secluded(NML)-LOC eight-CL-DV(ATTR)-fence make-FIN [I] am making eight-folded fence for [my] spouse to seclude herself, eightfolded fence in Iⁿdumo, where eight clouds rise (KK 1) 夜斯麻久爾都麻麻岐迦泥弖

ya sima kuni tuma mak-i-kane-te eight island country spouse pillow-CONV-NEG/POT(CONV)-SUB not being able to obtain a spouse in the country of Eight Islands (KK 2) 耶賦能之魔柯枳

ya pu-nǝ simba kaki eight (many?) mesh-GEN brushwood fence a woven brushwood fence with eight (many?) meshes (NK 91) 於弥能古能野陛能比母騰倶

omi-nǝ ko-nǝ ya-pe n-ǝ pimǝ tǝk-u noble-GEN son-GEN eight-CL DV-ATTR cord untie-FIN The sons of nobles were going to untie eight cords (NK 127) 夜都代爾母安礼波和須礼自許乃多知婆奈乎

ya-tu YƏ-ni mǝ are pa wasure-ⁿzi kǝnǝ tatimbana-wo eight-CL generation-LOC FP I TOP forget-NEG/TENT this mandarin. orange-ACC I will not forget these mandarin orange [flowers] even in eight (many?) generations (MYS 18.4058)

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

339

NOMINALS

Comparative Data Level A: Other Japonic A1: Eastern Old Japanese The cardinal numeral ya ‘eight’ is attested in Eastern Old Japanese only once in the combination with the following classifier -pe ‘layer’ (misspelled as pɛ). 多妣己呂母夜倍伎可佐祢弖

tambi kǝrǝmǝ ya-pɛ ki-kasane-te travel garment eight-CL put on(CONV)-pile(CONV)-SUB [I] put on eight layers of travel clothes one onto another … (MYS 20.4351) A2: Ryukyuan The reflexes of the cardinal numeral ‘eight’ are attested in all Ryukyuan languages without any irregular developments: Yuwan yaatsï, Izena yaatʃi, Agarinakasone and Hateruma yaatsï (Uchima and Arakaki 2000: 443) Shuri yaaçi (RKJ 1983: 270), Yonaguni daatʃi (Hirayama 1967: 229) ‘eight,’ etc. The cardinal numeral ya ‘eight’ is also attested in Old Ryukyuan. Old Ryukyuan やここのことをいぬつれて

ya kokono-ko towo inu ture-te eight nine-CL ten dog lead(CONV)-SUB leading eight, nine, ten dogs (RK 11)123 3.1.9 Cardinal Numeral kǝkǝnǝ ‘Nine’ The cardinal numeral kǝkǝnǝ ‘nine’ is attested in Western Old Japanese twice in phonographic spelling in the Kojiki kayō and the Nihonshoki kayō, but in an identical context: 迦賀那倍弖用迩波許許能用比迩波登袁加袁 (KK 26) 伽餓奈倍低用珥波虚虚能用比珥波苔塢伽塢 (NK 26)

ka-ŋga nambɛ-te yo n-i pa kǝkǝnǝ yo pi n-i pa tǝwo-ka-wo day-day put side by side(CONV)-SUB night DV-CONV TOP nine night day DV-CONV TOP ten-CL-ACC counting all the days, [it] is nine nights and ten days (KK 26, NK 26)

123  Cited according to Hokama 1995: 678.

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

340

Chapter 4

Comparative Data Level A: Other Japonic A2: Ryukyuan The reflexes of the cardinal numeral ‘nine’ are attested in all Ryukyuan languages without any significant irregular developments (the following Yuwan form appears to be slightly irregular): Yuwan kuunutsï, Izena kukunutʃi, Agarinakasone and Tonoshiro kukunutsï (Uchima and Arakaki 2000: 443); Shuri kukunuçi (RKJ 1983: 330) ‘nine,’ etc. The cardinal numeral kokono ‘nine’ is also attested in Old Ryukyuan. Old Ryukyuan やここのことをいぬつれて

ya kokono-ko towo inu ture-te eight nine-CL ten dog lead(CONV)-SUB leading eight, nine, ten dogs (RK 11)124 Level B: External Comparisons The proto-Japonic archetype can be reconstructed as *kǝkǝnǝ ‘nine’ on the basis of both Old Japanese and Ryukyuan data. There are no apparent external etymologies. 3.1.10 Cardinal Numeral tǝwo ‘Ten’ The cardinal numeral tǝwo ‘ten’ in phonographic spelling is attested in Western Old Japanese twice but in an identical context found in the Kojiki kayō and the Nihonshoki kayō. 迦賀那倍弖用迩波許許能用比迩波登袁加袁 (KK 26) 伽餓奈倍低用珥波虚虚能用比珥波苔塢伽塢 (NK 26)

ka-ŋga nambɛ-te yo n-i pa kǝkǝnǝ yo pi n-i pa tǝwo-ka-wo day-day put side by side(CONV)-SUB night DV-CONV TOP nine night day DV-CONV TOP ten-CL-ACC counting all the days, [it] is nine nights and ten days (KK 26, NK 26)

Comparative Data Level A: Other Japonic A2: Ryukyuan The reflexes of the numeral ‘ten’ are practically uniform in all modern Ryukyuan languages: tuu (Uchima and Arakaki 2000: 443; RKJ 1983: 533). The numeral 124  Cited according to Hokama 1995: 678. Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

NOMINALS

341

too ‘ten’ is also attested in Old Ryukyuan (with spelling variants とう /tou/ and とを /towo/). Old Ryukyuan ふてつたむとうのおとぢや

fute-tsu ta mu tou n-o otodiya one-CL two six ten DV-ATTR brother one, two, six, ten brothers (OS 13.898) やここのことをいぬつれて

ya kokono-ko towo inu ture-te eight nine-CL ten dog lead(CONV)-SUB leading eight, nine, ten dogs (RK 11)125 Level B: External Comparisons The proto-Japonic archetype can be reconstructed as *tǝwo ‘ten’ on the basis of both Old Japanese and Ryukyuan data. The well-known comparison is with pseudo-Koguryǒ *tǝk (徳) ‘ten.’ However, this comparison, likely as it is, is not as unproblematic as the comparisons of the other Japonic and Koguryǒ numerals, due to the unusual correspondence of PJ *-w- to Koguryǒ *-k-. 3.1.11 Count above ‘Ten’ The text of the Kojiki offers a number of later glosses to the numbers of deities above the number ‘ten,’ all of them in combination with the classifier pasira ‘column,’ used to count deities. The pattern is identical: tǝwo ‘ten’ + [a]mar-i ‘exceed-CONV’ + digital numeral + classifier pasira (Takagi and Toyama 1974a: 269): – tǝwo [a]mar-i pitǝ-pasira 11 deities – tǝwo [a]mar-i puta-pasira 12 deities – tǝwo [a]mar-i mi-pasira 13 deities – tǝwo [a]mar-i yǝ-pasira 14 deities – tǝwo [a]mar-i itu-pasira 15 deities – tǝwo [a]mar-i mu-pasira 16 deities – tǝwo [a]mar-i nana-pasira 17 deities – tǝwo [a]mari kǝkǝnǝ-pasira 19 deities The only number absent from this list below ‘twenty’ is *tǝwo [a]mar-i ya‘eighteen.’ The problem is, of course, that the Kojiki kana glosses are of a late origin, and cannot be trusted independently. However, there is one example 125  Cited according to Hokama 1995: 678. Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

342

Chapter 4

in the Bussoku seki no uta that includes the numeral 32 in phonographic spelling: miso-ti amar-i puta-tu ‘thirty-CL exceed-CONV two-CL,’ that demonstrates that the above glosses in all probability reflect the Western Old Japanese system of count above ‘ten’: 弥蘇知阿麻利布多都乃加多知夜蘇久佐等曾太礼留比止

miso-ti amar-i puta-tu n-ǝ katati yaso kusa tǝ sǝⁿdar-er-u pitǝ thirty-CL exceed-CONV two-CL DV-ATTR mark eighty type DV be.completePROG-ATTR person a person, who is complete with thirty-two marks and eighty [lesser sign] types (BS 2) Therefore, numerals above ‘ten’ and other tens were built on the pattern: ten number [+CL] + exceed-CONV + unit number [+CL]. Comparative Data Level A: Other Japonic A2: Ryukyuan A similar, but not identical system, of counting above ‘ten’ is found in South Ryukyuan dialects, e.g.: Agarinakasone and Yonaha tuu-psitiitsï, Tonoshiro tuupïtiidzï, Hateruma tuu-pï�t̥ utsï ‘eleven (lit.: ten + one)’ (Uchima and Arakaki 2000: 443). The South Ryukyuan system apparently does not include amar-i ‘exceed-CONV’ placed after tens and before digits as in Western Old Japanese. In other Ryukyuan languages, as well as in Old Ryukyuan, the system of counting above ‘ten’ seems to have been already supplanted by Chinese loans, as in Middle Japanese. 3.1.12 Tens All tens in Western Old Japanese with the exception of pata ‘twenty’ are built on the pattern: corresponding digit number + -so ‘ten.’ Not all tens are actually attested in phonographic spelling in Western Old Japanese texts; those that are not attested in phonographic writing or in borrowed kun-yomi writing (see below) are marked in the following list by an asterisk. Nevertheless, we can tentatively reconstruct the missing tens on the basis of the fact that attested tens follow the above structural pattern, as well as on the basis of poetic meter and evidence from Heian period Middle Japanese. – pata 20 – miso 30 – yǝso 40

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

NOMINALS

343

– *iso 50 – *muso 60 – *nanaso 70 – yaso 80 – *kǝkǝnǝso 90 The form iso ‘fifty’ is irregular (instead of the expected *itu-so). Possibly, a truncation occurred here. The tens can be followed by the general classifier -ti, similar to the general classifier -tu that occurs after digit numbers, e.g., miso-ti ‘thirty-CL.’ The second element -so ‘ten’ in the tens has an unclear origin. Connecting it etymologically with WOJ tǝwo ‘ten’ is dubious for phonetic reasons: first, it would require setting a unique internal correspondence of /t/:/s/ in Western Old Japanese: second, there seems to be no other known cases of contraction of the sequence /ǝwo/ to /o/. Third, there are no traces of -so in Ryukyuan except in apparent loans from Japanese. That suggests that the origin of -so ‘ten’ in Western Old Japanese may be later than the split between Japanese and Ryukyuan. I suspect that this -so ‘ten’ may be really a loan from the Paekche or another Korean-type language on the Korean peninsula. Cf. the tens in Middle Korean: – súmúlh, súmú126 20 – syèlhún 30 – màzòn 40 – swuy:n 50 – yèswyuy:n 60 – nìlhún 70 – yètún 80 – àhón 90 These tens except súmúlh ‘twenty’ are likely to include the suffix *-son/-sun in its different phonetic variants: -hun, -zon, -un, -on, -Vn. I think that this suffix may have a close genetic link to the source (probably Paekche) from which Western Old Japanese -so ‘ten’ was borrowed. PATA ‘twenty,’ strictly speaking, is not attested phonographically in Western Old Japanese texts. There is, however, one example in the Man’yōshū when the kun-yomi pata of the character 廿 ‘twenty’ is used as a borrowed kun-yomi writing to spell the homonymous word pata ‘loom.’ Since the word pata ‘loom, cloth’ is attested phonographically (e.g., in KK 66), there is little doubt that the word for ‘twenty’ was also pata in Western Old Japanese. 126  M K súmú ‘20’ is an attributive form.

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

344

Chapter 4

我廿物白麻衣

WA-ŋGA pata-MƏNƏ SIRO ASA KƏRƏMƏ I-POSS loom-thing white hemp garment white hemp garment [from] my loom (MYS 7.1298) MISO- ‘thirty’ is attested phonographically in Western Old Japanese in one example: 弥蘇知阿麻利布多都乃加多知夜蘇久佐等曾太礼留比止

miso-ti amar-i puta-tu n-ǝ katati yaso kusa tǝ sǝⁿdar-er-u pitǝ thirty-CL exceed-CONV two-CL DV-ATTR mark eighty type DV be completePROG-ATTR person a person, who is complete with thirty-two marks and eighty [lesser sign] types (BS 2) WOJ miso- ‘thirty’ is the only decade numeral that survived into the modern language, where it is still found in the compounds miso-ka ‘the last day of the month,’ misoji ‘thirty years old,’ and oo-miso-ka ‘the last day of the year.’ YOSO ‘forty,’ like pata ‘twenty’ is attested in Western Old Japanese only indirectly in the borrowed kun-yomi writing when the kun-yomi yǝso of the character 廾廾 ‘forty’ is used to write the word yǝsǝ ‘other place.’ There is certainly a problem that these two words are only nearly-homophonous, since yǝso ‘forty’ has a kō-rui vowel /o/ in the second syllable, while yǝsǝ ‘other place’ has an otsu-rui vowel /ǝ/, but since kungana sometimes allows for inexactitude, I will adopt this traditional explanation. 築羽根矣廾廾耳見乍

Tukumba-NE-WO yǝsǝ NƏMÏ MI-TUTU Tukumba-peak-ACC other.place RP look(CONV)-COOR while looking at the Tukumba peak only [from] the other place (MYS 3.383) ISO ‘fifty’ is not attested in Western Old Japanese texts (neither phonographically, nor logographically). The first phonographic attestation is only Middle Japanese one, found in the preface to the Heian period poetic anthology Goshūi wakashū, compiled in 1087 CE:

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

345

NOMINALS たへなるうたももちあまりいそちをかきいだして

tafe nar-u uta momo-ti amar-i iso-ti-wo127 kak-i-idas-ite excellent be-ATTR song hundred-CL exceed-CONV fifty-CL-ACC writeCONV-exit-SUB [he] recorded one hundred and fifty excellent poems, and … (GSWKS PREF 8.10) MUSO ‘sixty’ is not attested in Western Old Japanese texts (neither phonetically, nor in character writing). There seems to be no phonetic attestations in Heian period texts, either. NANASO ‘seventy’ is not attested in Western Old Japanese texts (neither phonographically, nor logographically). The first phonographic attestation is only Middle Japanese one, found in the Tosa nikki (935 CE): ななそちやそちはうみにあるものなりけり

nanaso-ti yaso-ti fa umi-ni ar-u mono nar-iker-i seventy-CL eighty-CL TOP sea-LOC exist-ATTR thing be-RETR-FIN Being at sea makes [you] look like seventy [or] eighty [years old] (TN 43.15) YASO ‘eighty’ is attested three times in Western Old Japanese and twice in Eastern Old Japanese. This better rate of attestations compared to the other tens is probably connected to the sacral importance attached to the number ‘eighty.’ 弥蘇知阿麻利布多都乃加多知夜蘇久佐等曾太礼留比止

miso-ti amar-i puta-tu n-ǝ katati yaso kusa tǝ sǝⁿdar-er-u pitǝ thirty-CL exceed-CONV two-CL DV-ATTR mark eighty type DV be completePROG-ATTR person a person, who is complete with thirty-two marks and eighty [lesser sign] types (BS 2) 海原乎夜蘇之麻我久里伎奴礼杼母

UNA-PARA-wo yaso sima-ŋ-gakur-i k-i-n-ure-ⁿdǝmǝ sea-plain-ACC eighty island-LOC-hide-CONV come-CONV-PERF-EV-CONC Although [I] came across the sea plain hiding [between] eighty (many?) islands … (MYS 15.3613) 127  The modern commentators transcribe this numeral as iso-di (Kubota and Hirata 1994: 8), but their choice for -di rather than -ti is not quite clear.

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

346

Chapter 4

夜蘇登毛乃乎波宇加波多知家里

yaso tǝmo n-ǝ wo pa u kapa tat-i-ker-i eighty companion DV-ATTR man TOP cormorant river stand-CONV-RETR-FIN eighty male companions [of mine] were cormorant fishing [at] the river (MYS 17.4023) Comparative Data Level A: Other Japonic A1: Eastern Old Japanese The cardinal numeral yaso ‘eighty’ is also attested in Eastern Old Japanese. 夜蘇久爾波那爾波爾都度比

yaso kuni pa Nanipa-ni tuⁿdop-i eighty province TOP Nanipa-LOC gather-CONV Eighty provinces gather at Naniwa, and … (MYS 20.4329) 夜蘇志麻須義弖

yaso sima suŋgï-te eighty island pass(CONV)-SUB [I] passed eighty islands, and … (MYS 20.4349) 3.1.13 Hundreds All hundreds in Western Old Japanese with the exception of momo ‘hundred’ are built on the pattern: corresponding digit number + -po128 ‘hundred.’ Only momo ‘hundred,’ ipo ‘five hundred,’ and yapo ‘eight hundred’ are actually attested in phonographic spelling in Western Old Japanese texts; those that are not attested are marked in the following list by an asterisk. Nevertheless, we can tentatively reconstruct the missing hundreds on the basis of the fact that the attested hundreds follow the structural pattern mentioned above, as well as on the basis of poetic meter. – momo 100 – *putapo 200 – *mipo 300 – *yǝpo 400 – ipo 500 – *mupo 600 128  The form -po with the kō-rui vowel [o] can be deduced on the basis of the KK spelling 本.

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

347

NOMINALS

– *nanapo 700 – yapo 800 – *kǝkǝnǝpo 900 The form ipo ‘five hundred’ is irregular (instead of the expected *itu-po). Possibly, a truncation occurred here. The hundreds can be followed by the general classifiers -tu, the one that follows units, or -ti, the one that follows tens, e.g., ipo-tu, ipo-ti ‘five hundred-CL.’ The second element -po ‘hundred’ in the hundreds has an unclear origin. Connecting it etymologically with WOJ momo ‘hundred’ seems to be impossible: there is no internal correspondence p : m in Western Old Japanese. In addition, there are no traces of -po in Ryukyuan. That suggests that the origin of -po ‘hundred’ in Western Old Japanese may be later than the split between Japanese and Ryukyuan. MOMO ‘hundred’ is well attested in Western Old Japanese: 毛毛知陀流夜迩波母美由

momo-ti-ⁿ-dar-u ya nipa mǝ mi-y-u hundred-thousand-GEN-be.enough-ATTR house garden FP see-PASS-FIN [I] can see flourishing (lit.: plentiful with hundreds and thousands) houses and gardens (KK 41) 意比久留母能波毛毛久佐爾勢米余利伎多流

op-i-k-uru mǝnǝ pa momo kusa n-i semɛ-yǝr-i-k-i-tar-u pursue-CONV-come-ATTR thing TOP hundred kind DV-CONV assault(CONV)-approach-CONV-come-CONV-PERF/PROG-ATTR the things that pursue [us], come assaulting [us] in hundred varieties (MYS 5.804) 毛毛等利能己惠能古保志枳波流岐多流良斯

momo tǝri-nǝ kǝwe-nǝ koposi-ki paru k-i-tar-urasi hundred bird-GEN voice-GEN be.missing-ATTR spring come-CONV-PERF/ PROG-SUP It looks like that the spring [with] voices of hundred birds, that [I] missed, [finally] has come (MYS 5.834) 毛毛可斯母由加奴麻都良遅

momo-ka si mǝ yuk-an-u Matura-ⁿ-di hundred-CL EP FP go-NEG-ATTR Matura-GEN-way [on] the way to Matura, [one] does not go one hundred days (MYS 5.870)

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

348

Chapter 4

PUTAPO ‘two hundred’ is not attested in Old Japanese texts (neither phonographically, nor logographically). There seems to be no phonographic attestations in Heian period texts, either. MIPO ‘three hundred’ is not attested in Old Japanese texts (neither phonographically, nor logographically). There seems to be no phonographic attestations in Heian period texts, either. YƏPO ‘four hundred’ is not attested in Old Japanese texts (neither phonographically, nor logographically). There seems to be no phonographic attestations in Heian period texts, either. IPO ‘five hundred’ is attested in Western Old Japanese in five examples. 伊制能奴能娑柯曳鳴伊裒

ise-nǝ no-nǝ saka-ye-wo ipo Ise-GEN field-GEN flourishing-branch-ACC five.hundred five hundred branches of flourishing [trees] on the Ise plain (NK 78) 安波妣多麻伊保知毛我母

apambi tama ipo-ti moŋgamǝ abalone pearl five.hundred-CL DP [I] wish [to have] five hundred abalone pearls (MYS 18.4101) 朝獦尓伊保都登里多氐

ASA-ŋ-GAR-I-ni ipo-tu tǝri tate morning-GEN-hunt-NML-LOC five.hundred-CL bird raise(CONV) At the morning hunt, [it] raised five hundred birds (MYS 17.4011) 思良多麻能伊保都追度比乎手爾牟須妣

sira tama-nǝ ipo-tu tuⁿdop-i-wo TE-ni musumb-i white pearl-GEN five.hundred-CL collect-NML-ACC hand-LOC tie-CONV to tie a set of five hundred white pearls to [your] arms (MYS 18.4105) 五百都綱波布

IPO-tu TUNA pap-u five.hundred ropes creep-FIN five hundred ropes are creeping (MYS 19.4274)

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

349

NOMINALS

MUPO ‘six hundred’ is not attested in Old Japanese texts (neither phonographically, nor logographically). There seems to be no phonographic attestations in Heian period texts, either. NANAPO ‘seven hundred’ is not attested in Old Japanese texts (neither phonographically, nor logographically). There seems to be no phonographic attestations in Heian period texts, either. YAPO ‘eight hundred’ is attested once in Western Old Japanese. 夜本爾余志伊岐豆岐能美夜

yapo ni yǝ-si i-kiⁿduk-i n-ǝ miya eight.hundred ground good-FIN PREF-build-NML DV-ATTR palace a palace build on an eight hundred [times] good soil (KK 100) There is also a logographic attestation: 八百日往濱之沙毛吾戀二豈不益歟

YAPO-KA IK-U PAMA-NƏ MASA ŋGO mo A-ŋGA KOPÏ-ni ANI MASAR-AⁿZI KA Eight.hundred-CL go-ATTR beach-GEN sand FP I-POSS longing-LOC INTER be.superior-NEG/TENT How would my longing be less than sand [grains] on the beach, on which one walks for eight hundred days? [—Certainly not!] (MYS 4.596) KƏKƏNƏPO ‘nine hundred’ is not attested in Old Japanese texts (neither phonographically, nor logographically). There seems to be no phonographic attestations in Heian period texts, either. Comparative Data Level A: Other Japonic A1: Eastern Old Japanese The cardinal numeral momo ‘hundred’ is also attested in Eastern Old Japanese in one example: 毛母久麻能美知波紀爾志乎

momǝ kuma-nǝ miti pa k-ï-n-i-si-wo hundred corner-GEN road TOP come-CONV-PERF-CONV-PAST/ATTR-ACC although [I] came along the road with a hundred corners (MYS 20.4349)

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

350

Chapter 4

A2: Ryukyuan The Ryukyuan cardinal number mumu ‘hundred,’ seems to be attested only in Old Ryukyuan as momo and in the Shuri written language as mumu (RKJ 1983: 390). Thus, it apparently represents a loan from Middle Japanese. Old Ryukyuan かみのふねももおうね下のふね

kami-no fune momo o-une SITA-no fune top-GEN boat hundred HON-boat bottom-GEN boat top boats, hundred boats, lower boats (OS 13.813) 3.1.14 Higher Numerals Only two higher cardinal numerals are attested in Western Old Japanese texts, ti ‘thousand’ and yǝrǝⁿdu ‘ten thousand.’ TI ‘thousand’ is well attested in Western Old Japanese: 夜知富許能迦微能美許登

ya-ti pǝkǝ n-ǝ kamï-nǝ mi-kǝtǝ eight-thousand spear DV-ATTR deity-GEN HON-deity the deity Yatipoko (lit.: Eight thousand spears) (KK 2) 毛毛知陀流夜迩波母美由

momo-ti-ⁿ-dar-u ya nipa mǝ mi-y-u hundred-thousand-GEN-be.enough-ATTR house garden FP see-PASS-FIN [I] can see flourishing (lit.: plentiful with hundreds and thousands) houses and gardens (KK 41) 知余珥茂訶勾志茂餓茂

ti yǝ-ni mo ka-ku si moŋgamo thousand year-LOC FP thus-CONV EP DP [I wish that my sovereign] will be [like] that in a thousand years, too (NK 102) 志良久毛能知弊仁辺多天留都久紫能君仁波

sira kumo-nǝ ti-pe-ni peⁿdat-er-u Tukusi-nǝ kuni pa white cloud-GEN thousand-CL-LOC be separated-PROG-ATTR Tukusi-GEN land TOP land of Tukushi, which is separated [from me] by a thousand white clouds (MYS 5.866)

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

351

NOMINALS 伎美乎見麻久波知登世爾母我母

kimi-wo MI-m-aku pa ti tǝse n-i mǝŋgamǝ lord-ACC see-TENT-NML TOP thousand year DV-CONV DP [I] would like to see [my] lord for a thousand years (MYS 20.4304) 伊爾志加多知与乃都美佐閇保呂夫止曾伊布

in-i-si kata ti yǝ-nǝ tumi sapɛ porǝmb-u tǝ sǝ ip-u go-CONV-PAST/ATTR side thousand life-GEN sin RP disappear-FIN DV FP say-ATTR [they] say that even the sins of one thousand former lives will disappear (BS 17) YƏRƏⁿDU ‘ten thousand’ is well attested in Western Old Japanese: 予呂豆余珥訶勾志茂餓茂

yǝrǝⁿdu yǝ-ni ka-ku si moŋgamo ten thousand year-LOC thus-CONV EP DP I wish that [my sovereign] will be [like] that in ten thousand years (NK 102) 余呂豆余爾伊麻志多麻比提

yǝrǝⁿdu yǝ-ni imas-i-tamap-i-te ten thousand year-LOC exist(HON)-CONV-HON-CONV-SUB [may you] live for ten thousand years (MYS 5.879) 餘呂豆代爾可多理都具倍久

yǝrǝⁿdu YƏ-ni katar-i-tuŋg-umbɛ-ku ten.thousand year-LOC talk-CONV-continue-DEB-CONV [it] would be talking for ten thousand years (MYS 17.3914) 己乃美阿止夜与呂豆比賀利乎波奈知伊太志

kǝnǝ mi-atǝ ya-yǝrǝⁿdu pikari-wo panat-i-iⁿdas-i this HON-footstep eight-ten thousand light-ACC emit-CONV-exit-CONV This footstep [of the Buddha] emits eighty thousand lights (BS 4) 爾詩乃美夜古波与呂豆与乃美夜

nisi-nǝ miyako pa yǝrǝⁿdu yǝ-nǝ miya west-GEN capital TOP ten thousand generations-GEN palace The Western Capital [is] the palace [for] ten thousand generations (SNK 6)

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

352

Chapter 4

Comparative Data Level A: Other Japonic A2: Ryukyuan The cardinal numeral ti ‘thousand’ is attested only in Old Ryukyuan and exclusively in the combination ya-ti-yo/ya-di-yo/ya-ti-ya ‘eternally,’ ‘for ever’ (lit.: ‘eight thousand generations/years’). Such a limited distribution makes it a probable loan from mainland Japanese. やちよかけて

ya-ti-yo kake-te eight-thousand-year hang(CONV)-SUB being mindful for ever (lit.: eight thousand years) (OS 3.147) The cardinal numeral yorodu ‘ten thousand’ is attested only in Old Ryukyuan. Due to this limited attestation it is likely to be a borrowing from mainland Japanese. よろづ物事

yorodu MUNU-GUTU ten thousand thing-thing ten thousand things (RK 1173) Level B: External Comparisons There are no apparent external etymologies for OJ ti ‘thousand.’ OJ yǝrǝⁿdu ‘ten thousand’ has been frequently compared with MK yèléh ‘many.’ The different suffixation in Old Japanese and Middle Korean suggests that it may be a loanword from some variety of Old Korean (Paekche?). 3.2

Ordinal Numerals

The only ordinal numeral attested in Old Japanese texts is patu ‘first.’ The other ordinal numerals were probably formally indistinguishable from cardinals, like in Classical Japanese, where one can tell the difference on the basis of the context alone. However, the cases like that are not attested in the Old Japanese texts: one can only speculate that they indeed existed in the colloquial.

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

353

NOMINALS 波都迩波波陀阿可良氣美

patu ni pa paⁿda aka-ra-kɛ-mi first clay TOP surface red-NML-COMP-GER As for the first clay, [its] surface was red-like (KK 42) 相見婆登許波都波奈爾

API-MI-RE-mba tǝkǝ patu pana n-i REC-look-EV-CON eternal first flower DV-CONV when [we] looked at each other, it was [always] like eternal first flowers (MYS 17.3978) 波都波奈乎延太爾多乎理弖

patu pana-wo yeⁿda-ni ta-wor-i-te first flower-ACC branch-LOC hand-break-CONV-SUB breaking off the first flowers from the branch (MYS 18.4111) 波都由伎波知敞爾布里之家

patu yuki pa ti-pe n-i pur-i-sik-e first snow TOP thousand-CL DV-CONV fall-CONV-cover-IMP First snow, fall in a thousand layers! (MYS 20.4475) 波都波流能家布敷流由伎

patu paru-nǝ kepu pur-u yuki first spring-GEN today fall-ATTR snow snow that is falling today, on the first spring [day] (MYS 20.4516) Comparative Data Level A: Other Japonic A1: Eastern Old Japanese The ordinal numeral patu ‘first’ is also attested in one example in Eastern Old Japanese: 夜麻杼里乃乎呂能波都乎

yama-ⁿ-dǝri-nǝ wo-rǝ-nǝ patu wo mountain-GEN-bird-GEN tail-DIM-COMP first hemp the first hemp [of the year] that is like the tail of a mountain bird (MYS 14.3468)

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

354

Chapter 4

A2: Ryukyuan The ordinal number haçi/faci/pacii ‘first’ is attested in the two best-described modern Ryukyuan dialects: Nakijin and Shuri, as well as in Old Ryukyuan. This makes one suppose that it probably exists in other dialects as well, at least in the Northern Ryukyus, and is simply not reflected in existing grammars and dictionaries of those dialects. Nakijin examples: pacii tabi ‘first trip,’ pacii maaga ‘first grandchild’ (Nakasone 1983: 393). Shuri examples: haçi ’aQcii ‘first walk after the birth,’ haçi mumu ‘first willows/plum trees [in the year]’ (RKJ 1983: 198). Old Ryukyuan はつにしがおしいぢへば

facu nisi-ga os-i-id-ife-ba first northern wind-NOM push-CONV-exit-EV-CON when the first northern wind starts to blow (OS 7.349) Level B: External Comparisons There are no apparent internal or external etymologies for patu ‘first.’ Some kokugogaku scholars want it to be derived from pitǝ-tu ‘one’ by ubiquitous ‘vowel alternation,’ but the accents are incongruent: while patu ‘first’ is highregister 2.1 class, pitǝ-tu is low-register 3.7b class (Martin 1987: 402, 411). 3.3 Classifiers The system of classifiers in Old Japanese is much less developed than in modern Japanese. Only six following classifiers are attested in the texts in the phonographic writing: -tu (classifier for objects used with digits and hundreds), -ti (classifier for objects used with tens and hundreds, starting with pata ‘twenty’129), -ri (classifier for persons), -mǝtǝ (classifier for grassy plants), -pe 129  There is no form pata-ti ‘twenty-CL’ attested phonetically in Old Japanese texts, but the Middle Japanese form fata-ti appears in the phonetic writing in one of the earliest Classical Japanese texts Ise monogatari:   その山はここにたとへばひえの山をはたちばかりかさねあげたらんほどして   ono yama fa koko-ni tatof-e-ba Fiye n-o yama-wo fata-ti bakari kasane-age-tar-an fodo s-ite   that mountain TOP here-LOC compare-EV-CON Hiei DV-ATTR mountain-ACC twenty-CL RP pile up(CONV)-raise-PERF/PROG-TENT/ATTR degree do-SUB   That mountain, when [one] compares [it] with [mountains] here, is as high as Mt. Hiei, piled up about twenty [times]… (IM 9.117.14). Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

355

NOMINALS

(classifier for layers and folds), and -ka (classifier for days). Two more are possible: -pasira (classifier for deities) and -tambi (classifier for times), but none of these appears in phonographic script, and they are therefore excluded from the present description. The word tǝse ‘year’ comes close to being a classifier, because it replaces a regular word tǝsi ‘year’ when preceded by a numeral, but since it never loses its lexical meaning, I do not consider it as a full classifier. 3.3.1 Classifier -tu The classifier -tu ‘thing’ is a general classifier for inanimate objects used with digits (even if these digits appear after tens, see BS 2 below), and after ipo- ‘five hundred.’ Probably a more exact definition of this classifier would be as a classifier for non-humans and non-deities, since it appears in one example from the Bussoku seki no uta as a classifier for snakes and demons (see BS 19 below). 袁都能佐岐那流比登都麻都比登都麻都阿勢袁

Wotu-nǝ saki-n-ar-u pitǝ-tu matu pitǝ-tu matu a se wo Wotu-GEN cape-LOC-exist-ATTR one-CL pine one-CL pine I beloved EP a lone (lit.: one) pine that stands at the cape of Wotu. Oh, lone (lit.: one) pine, my beloved! (KK 29) 美都具理能曾能那迦都迩

mi-tu-ŋ-guri-nǝ sǝnǝ naka-tu ni three-CL-DV(ATTR)-chestnut-COMP that middle-GEN/LOC clay that clay from the middle that is like three chestnuts (KK 42) 弥蘇知阿麻利布多都乃加多知夜蘇久佐等曾太礼留比止

miso-ti amar-i puta-tu n-ǝ katati yaso kusa tǝ sǝⁿdar-er-u pitǝ thirty-CL exceed-CONV two-CL DV-ATTR mark eighty type DV be completePROG-ATTR person a person, who is complete with thirty-two marks and eighty [lesser sign] types (BS 2) 与都乃閇美伊都都乃毛乃乃阿都麻礼流伎多奈伎微乎婆

yǝ-tu n-ǝ pɛmi itu-tu n-ǝ monǝ-nǝ atumar-er-u kitana-ki mï-womba four-CL DV-ATTR snake five-CL DV-ATTR demon-GEN gather-PROG-ATTR dirty-ATTR body-ACC(EMPH) the dirty body which accumulates four snakes and five demons (BS 19)

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

356

Chapter 4

藤原朝臣麻呂等伊負図亀一頭献止奏賜不爾 PUⁿDIPARA-NƏ ASƏMI MARƏ-RA-i PUMI-WO OP-ER-U KAMƐ-WO PITƏ-TU TATEMATUR-AKU tǝ MAWOS-I-TAMAp-u-ni … Puⁿdipara-GEN retainer Marǝ-PLUR-ACT writing-ACC bearPROG-ATTR tortoise-ACC one-CL offer(HUM)-NML DV say(HUM)-CONV-HON-ATTR-LOC [They] said that the retainer Puⁿdipara Marǝ and others had offered a (lit.: one) tortoise bearing writing [on its back] … (SM 6) 麻多麻奈須布多都能伊斯乎

ma-tama-nasu puta-tu n-ǝ isi-wo INT-jewel-COMP two-CL DV-ATTR stone-ACC two stones like real jewels (MYS 5.813) 朝獦尓伊保都登里多氐

ASA-ŋ-GAR-I-ni ipo-tu tǝri tate morning-GEN-hunt-NML-LOC five.hundred-CL bird raise(CONV) At the morning hunt, [it] raised five hundred birds (MYS 17.4011) 夜都代爾母安礼波和須礼自許乃多知婆奈乎

ya-tu YƏ-ni mǝ are pa wasure-ⁿzi kǝnǝ tatimbana-wo eight-CL generation-LOC FP I TOP forget-NEG/TENT this mandarin. orange-ACC I will not forget these mandarin orange [flowers] even in eight (many?) generations (MYS 18.4058) 思良多麻能伊保都追度比乎手爾牟須妣

sira tama-nǝ ipo-tu tuⁿdop-i-wo TE-ni musumb-i white pearl-GEN five.hundred-CL collect-NML-ACC hand-LOC tie-CONV to tie a set of five hundred white pearls to [your] arms (MYS 18.4105) 五百都綱波布

IPO-tu TUNA pap-u five.hundred ropes creep-FIN five hundred ropes are creeping (MYS 19.4274)

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

357

NOMINALS

Comparative Data Level A: Other Japonic A1: Eastern Old Japanese The classifier -tu ‘thing’ is attested only once in Eastern Old Japanese: 奴麻布多都可欲波等里賀栖安我己許呂布多由久奈母

numa puta-tu kayop-a tǝri-ŋga su a-ŋga kǝkǝrǝ puta yuk-unam-ǝ marsh two-CL go.over-ATTR bird-POSS nest I-POSS heart two go-TENT2-ATTR Should my heart go [to] two [different places like] nests of birds that go over two marshes? (MYS 14.3526) A2: Ryukyuan The classifier -çi/-ci ‘thing,’ a cognate of OJ -tu, is well attested in various modern Ryukyuan languages and in Old Ryukyuan. The numeral digit stem with this classifier appears to be more bound than the corresponding stem in the Old Japanese. See section 3.1 for more details and more examples. Old Ryukyuan 月やひとつ

TSUCHI ya fito-tu moon TOP one-CL [there is only] one moon (RK 485) Shuri

sjuutee tii-çi nay-uN household one-CL become-FIN [Members of] the household represent one unit (RKJ 1983: 517) 3.3.2 Classifier -ti The classifier -ti ‘thing’ is a general classifier for inanimate objects used with tens and hundreds. Only two examples are attested in Old Japanese texts in phonographic writing. Probably a more exact definition of this classifier would be as a classifier for non-humans and non-deities, as in one of the examples below it is used with the word apambi ‘abalone.’

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

358

Chapter 4

弥蘇知阿麻利布多都乃加多知夜蘇久佐等曾太礼留比止

miso-ti amar-i puta-tu n-ǝ katati yaso kusa tǝ sǝⁿdar-er-u pitǝ thirty-CL exceed-CONV two-CL DV-ATTR mark eighty type DV be completePROG-ATTR person a person, who is complete with thirty-two marks and eighty [lesser sign] types (BS 2) 安波妣多麻伊保知毛我母

apambi tama ipo-ti moŋgamǝ abalone pearl five.hundred-CL DP [I] wish [to have] five hundred abalone pearls (MYS 18.4101) 3.3.3 Classifier -ri The classifier -ri ‘person’ occurs in phonographic script in Old Japanese texts only in combination with the preceding numerals pitǝ ‘one’ and puta ‘two,’ like in the modern language. Later glosses on logographically written combinations that indicate also mi-tari ‘three people,’ yon-dari ‘four people,’ and iku-tari ‘how many people’ (Ōno 1990: 830), would suggest that -ri is a truncation of -tari, but these glosses’ authenticity is very doubtful. It is quite possible that these glosses are the later creations inspired by the incorrectly analyzed forms pitǝri ‘one person’ and putari ‘two persons.’ In the absence of the contrary evidence, it is safer to rely on actual phonographic attestations in Old Japanese texts. It is also significant that just the evidence for -ri, and not for -tari is corroborated by both Eastern Old Japanese and Ryukyuan. 袁夜迩須賀多多美伊夜佐夜斯岐弖和賀布多理泥斯

wo-ya-ni suŋga-tatami iya-saya sik-i-te wa-ŋga puta-ri ne-si DIM-house-LOC sedge-mat more(?)-rustling spread-CONV-SUB we-POSS two-CL sleep(CONV)-PAST/ATTR in the little hut, spreading very rustling sedge mats, two of us slept (KK 19) 夜多能比登母登須宜波比登理袁理登母意富岐弥斯與斯登岐許佐婆比登理 袁理登母

Yata-nǝ pitǝ-mǝtǝ suŋgɛ pa pitǝ-ri wor-i tǝmǝ opǝ kimi si yǝ-si tǝ kikǝs-amba pitǝ-ri wor-i tǝmǝ Yata-GEN one-CL sedge TOP one-CL exist-FIN CONJ great lord EP be.goodFIN DV say(HON)-COND one-CL exist-FIN CONJ Even if one sedge from Yata is alone, if the great lord says [it] is fine, even if [she] is alone (KK 65)

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

359

NOMINALS 烏梅能波奈比等利美都都

uMƐ-nǝ pana pitǝ-ri mi-tutu plum-GEN flower one-CL see(CONV)-COOR looking alone at the plum blossoms (MYS 5.818) Comparative Data Level A: Other Japonic A1: Eastern Old Japanese The classifier -ri ‘person’ is attested in one example in Eastern Old Japanese: 兒良波安波奈毛比等理能未思弖

KO-ra pa ap-ana-m-o pitǝ-ri nǝmï s-i-te girl-DAT TOP meet-DES-TENT-ATTR one-CL RP do-CONV-SUB being absolutely alone, [I] wish to meet [this] girl (MYS 14.3405) A2: Ryukyuan The classifier -ri is also attested in Old Ryukyuan and its regular reflex -i in modern Shuri dialect, e.g., cu-i ‘one person’ < pitǝ-ri (Shuri written language still has hwicu-i ‘one person,’ with preservation of the first syllable). Old Ryukyuan ただひとり

tada fito-ri just one-CL just one person (OS 14.997) Level B: External Comparisons There are no apparent external parallels for the classifier -ri ‘person.’ Since it begins with the /r/, which is unusual even for morpheme-initial position in Old Japanese, I think that it probably has a perfect internal etymology, being a truncation of ar-i ‘being’ (lit.: exist-NML). 3.3.4 Classifier -mǝtǝ The classifier -mǝtǝ ‘root’ is used for grassy plants. It is well attested in Western Old Japanese. 夜麻登能比登母登須須岐

Yamatǝ-nǝ pitǝ-mǝtǝ susuki Yamato-GEN one-CL silver grass one silver grass in Yamato (KK 4) Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

360

Chapter 4

阿波布爾波賀美良比登母登

apa-pu-ni pa ka-mira pitǝ-mǝtǝ millet-tussock-LOC TOP smell-leek one-CL one smelly leek in the millet tussock (KK 11) 夜多能比登母登須宜波古母多受

Yata-nǝ pitǝ-mǝtǝ suŋgɛ pa ko mǝt-aⁿz-u Yata-GEN one-CL sedge TOP child hold-NEG-FIN One sedge from Yata does not have children (KK 64) 比登母等能奈泥之故宇惠之

pitǝ-mǝtǝ n-ǝ naⁿdesiko uwe-si one-CL DV-ATTR carnation plant(CONV)-PAST/ATTR [I] planted one carnation (MYS 18.4070) Comparative Data Level A: Other Japonic A1: Eastern Old Japanese -mǝtǝ is attested once in EOJ as a classifier for yanaŋgi ‘willow.’ 和加加都乃以都母等夜奈枳

wa-ŋga katu-nǝ itu-mǝtǝ yanaŋgi I-POSS gate-GEN five-CL willow five willow trees at my gate (MYS 20.4386) A2: Ryukyuan The classifier -mǝtǝ ‘root’ is not attested in Ryukyuan, although there is certainly the word mu[u]tu ‘root,’ ‘base’ in Okinawan. However, it is likely to be a Japanese loan, as there are no attestations in the Southern Ryukyuan dialects. Level B: External Comparisons There are no apparent external etymologies of -mǝtǝ as a classifier for grassy plants, although the word mǝtǝ ‘root’ itself may be a possible Korean loan: < MK mìth ‘base,’ ‘bottom.’ Nevertheless, the correspondence of OJ /ǝ/ to MK /i/ in the first syllable represents a considerable problem. 3.3.5 Classifier -pe The classifier -pe ‘layer,’ ‘fold’ is well attested in Western Old Japanese. It is used to count fences, garments, cords, and clouds.

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

361

NOMINALS 伊豆毛夜弊賀岐都麻碁微爾夜弊賀岐都久流

Iⁿdumo ya-pe-ŋ-gaki tuma-ŋ-gǝmï-ni ya-pe-ŋ-gaki tukur-u Iⁿdumo eight-CL-DV(ATTR)-fence spouse-GEN-be secluded(NML)-LOC eightCL-DV(ATTR)-fence make-FIN [I] am making eight-folded fence for [my] spouse to seclude herself, eightfolded fence in Iⁿdumo (KK 1) 虚呂望虚曾赴多幣茂予耆

kǝrǝmǝ kǝsǝ puta-pe mo yǝ-ki garment FP two-CL FP good-ATTR [It] is good [to wear] two layers of garments (NK 47) 虚呂望赴多幣枳低

kǝrǝmo puta-pe ki-te garment two-CL wear(CONV)-SUB wearing two layers of garments (NK 49) 飫瀰能古簸多倍能波伽摩鳴那那陛鳴糸

omi-nǝ ko pa tapɛ-nǝ pakama-wo nana-pe wos-i noble-GEN child TOP mulberry.tree.bark.cloth-GEN pants-ACC seven-CL wear(HON)-CONV The son of a noble wears seven layers of mulberry tree bark cloth pants, and … (NK 74) 於弥能古能野陛能比母騰倶

omi-nǝ ko-nǝ ya-pe n-ǝ pimǝ tǝk-u noble-GEN son-GEN eight-CL DV-ATTR cord untie-FIN The sons of nobles were going to untie eight cords (NK 127) 志良久毛能知弊仁辺多天留都久紫能君仁波

sira kumo-nǝ ti-pe-ni peⁿdat-er-u tukusi-nǝ kuni pa white cloud-GEN thousand-CL-LOC be separated-PROG-ATTR Tukusi-GEN land TOP land of Tukushi, which is separated [from me] by thousand white clouds (MYS 5.866)

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

362

Chapter 4

Comparative Data Level A: Other Japonic A1: Eastern Old Japanese The classifier -pe ‘layer’ is attested twice in Eastern Old Japanese: 志毛用爾奈奈弁加流去呂毛

simo yo-ni nana-pe k-ar-u kǝrǝmo frost night-LOC seven-CL wear-PROG-ATTR garment in the frosty night [I] wear seven layers of garments (MYS 20.4431) 多妣己呂母夜倍伎可佐祢弖

tambi kǝrǝmǝ ya-pɛ ki-kasane-te travel garment eight-CL put on(CONV)-pile(CONV)-SUB [I] put on eight layers of travel clothes one onto another … (MYS 20.4351) Level B: External Comparisons There are no apparent external etymologies of -pe as a classifier, although the word pe ‘layer’ itself has a reliable Korean parallel: Early Modern Korean pel ‘set [of clothes].’ Its absence in Ryukyuan suggests that OJ pe is a loan from Korean. 3.3.6 Classifier -ka The classifier -ka ‘day’ occurs in Western Old Japanese in phonographic writing in a few examples. It is used starting from the numeral ‘two.’ ‘One day’ is pitǝ pi ‘one day,’ without any classifiers used. Unfortunately, most of the numbers occurring with the classifier -ka are not attested in phonographic writing in Old Japanese texts (the phonographically attested ones are only putu-ka ‘two days,’ nanu-ka ‘seven days,’ tǝwo-ka ‘ten days,’ and momo-ka ‘hundred days’). The majority of them probably were irregular as they are in the language of later periods, including modern Japanese, but we simply cannot tell due to the absence of their attestation in the phonographic writing. The best thing we can do is to reconstruct their phonographic values relying on the data from Middle and Early Modern Japanese, and in the case of *yau-ka ‘eight days’ even from modern Japanese. The reconstructed entries are marked with an asterisk (*) in the chart below.

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

363

NOMINALS chart 17

1 day 2 days 3 days 4 days 5 days 6 days 7 days 8 days 9 days 10 days 20 days 100 days

Combinations of numerals with classifier -ka ‘day’

OJ

MJ and EMJ

Modern Japanese

pitǝ pi putu-ka *mi-ka *yǝ-ka *itu-ka *mui-ka nanu-ka *yau-ka *kǝkǝnu-ka tǝwo-ka *patu-ka momo-ka

MJ fito fi MJ futu-ka MJ mi-ka EMJ yo-ka EMJ itu-ka EMJ mui-ka MJ nanu-ka ? MJ kokonu-ka MJ towo-ka MJ fatu-ka MJ momo-ka

(iti-n[i]ti) futu-ka mik-ka yok-ka itu-ka mui-ka nano-ka/nanu-ka yoo-ka kokono-ka too-ka hatuka (hyaku-niti)

Note: OJ forms in italics are irregular. Loans from Chinese are included in parentheses.

Examples: 迦賀那倍弖用迩波許許能用比迩波登袁加袁

ka-ŋga nambɛ-te yo n-i pa kǝkǝnǝ yo pi n-i pa tǝwo-ka-wo day-day put side by side(CONV)-SUB night DV-CONV TOP nine night day DV-CONV TOP ten-CL-ACC counting all the days, [it] is nine nights and ten days (KK 26) 毛毛可斯母由加奴麻都良遅

momo-ka si mǝ yuk-an-u Matura-ⁿ-di hundred-CL EP FP go-NEG-ATTR Matura-GEN-way [on] the way to Matuura, [one] does not go one hundred days (MYS 5.870) 比登比母伊毛乎和須礼弖於毛倍也

pitǝ-pi mǝ imo-wo wasure-te omop-ɛ ya one-day FP beloved-ACC forget(CONV)-SUB think-EV IP would [I] imagine forgetting [my] beloved even for one day? (MYS 15.3604)

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

364

Chapter 4

知加久安良婆伊麻布都可太未等保久安良婆奈奴可

tika-ku ar-amba ima putu-ka ⁿdamï tǝpo-ku ar-amba nanu-ka close-CONV exist-COND now two-CL ? far-CONV exist-COND seven-CL if [it] is soon, [from] now only (?) two days; if [it] is long, seven days (MYS 17.4011) Comparative Data Level A: Other Japonic A2: Ryukyuan The classifier -ka ‘day’ is attested in Old Ryukyuan as well as in the modern Shuri and Nakijin dialects. This distribution makes it possible that this classifier is present in other Ryukyuan dialects as well, at least in the Northern Ryukyus, but is not reflected in descriptions or dictionaries. Nevertheless, the possibility of borrowing is also quite strong, and will remain a likelier explanation unless some attestations of -ka are found in the Southern Ryukyus. Similar to Western Old Japanese, the classifier -ka is used only starting from number ‘two.’ ‘One day’ is hwiQcii in Shuri and phiQcui in Nakijin, both being regular correspondences to OJ pitǝ pi. In Shuri the native numbering of days is preserved only for numerals 1–4, 10, and 20. The combinations for days 5–9 are replaced by Chinese loanwords. The situation is more complicated in Nakijin, where the native count is attested for numerals 1–4, 7, 10, and 20. However, these are the only combinations recorded in Nakosone Seizen’s 1983 Nakijin dictionary. It remains unclear whether the combinations for 5–6 and 8–9 were simply not recorded or are replaced by Chinese loans like in Shuri. In Old Ryukyuan only the combinations for ‘two days,’ ‘three days,’ ‘four days,’ ‘seven days,’ ‘ten days,’ and ‘twenty days’ are attested. The following chart represents a combined evidence from Old Ryukyuan, Shuri, and Nakijin. Examples: Old Ryukyuan 月よかたてば

CIKI yo-ka tat-e-ba moon four-CL rise-EV-CON when the moon rises on the fourth day (OS 14.1032) Shuri

hwiiQcii ya kuutuguutu nukubaay-uN

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

365

NOMINALS chart 18

1 day 2 days 3 days 4 days 5 days 6 days 7 days 8 days 9 days 10 days 20 days

Combinations of numerals with the classifier -ka ‘day’ in Ryukyuan

OR

Shuri

Nakijin

— fucika mika yoka — — nanuka — — tooka facika

fiQcii huçika mica/miQca/mika ’yuQka (gunici) (dukunici/rukunici) (sicinici) (hacinici) (kunici) tuka haçika

phiQcui puciika mikha yuQkha ? ? nankhaa ? ? tuhwaa paciikhaa

Note: Chinese loanwords are in parentheses.

one-day TOP every get warm-FIN Every (single) day it gets warmer (RKJ 1983: 237) Level B: External Comparisons Before attempting any external comparisons, it is important to reconstruct the proto-Japonic archetype of the classifier -ka ‘day.’ Although it is routinely believed to be just -ka in the Japanese grammatical tradition (Omodaka et al. 1967: 170; Ōno 1990: 266), it remains unclear why a simple /ka/ form would cause considerable irregularities in combinations with preceding numeral roots. Let us review the irregular forms vs. corresponding numeral roots. chart 19

Numeral roots and irregular forms for the count of days

Numeral

Number of days

puta- ‘2’ mu- ‘6’ nana- ‘7’ ya- ‘8’ kǝkǝnǝ- ‘9’ pata- ‘20’

putu-ka mui-ka nanu-ka yau-ka kǝkǝnu-ka patu-ka

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

366

Chapter 4

chart 20 Archetypes for the count of days

Number of days

Archetype

put-uka ‘3 days’ mu-ika ‘6 days’ nan-uka ‘7 days’ ya-uka ‘8 days’ kǝkǝn-uka ‘9 days’ pat-uka ‘20 days’

*puta-uka *mu-ika *nana-uka *ya-uka *kǝkǝnǝ-uka *pata-uka

In all these cases except mui-ka ‘six days’ we see a ubiquitous /u/ added to the stem of a numeral, which replaces either a final original low central vowel /a/ or a final original mid-central vowel /ǝ/ in the numerical stem. The exception mui-ka which exhibits not /u/, but /i/ is not attested in the phonographic writing until the sixteenth century. No other combination of these numeral roots with other classifiers or nouns provides evidence for an /u/ inherent to any of these numeral roots. Under these circumstances it is logical to surmise that this /u/ really belongs to a classifier rather than to a numeral root. The above chart presents this reanalysis together with a reconstruction of the archetypes for the number of days. I suggest that the archetype of this classifier is really *-uka, not *-ka. The aberrant form mu-ika, attested only in Early Modern Japanese, probably represents a late and unique case of a progressive vowel dissimilation. The remaining ‘regular’ cases can be easily explained as cases where /u/ in *-uka is lost following not a final central but a final front or back vowel in a numeral root. The cases of mi-ka ‘three days’ < *mi-uka, itu-ka ‘five days’ < *itu-uka, and momo-ka ‘hundred days’ < *momo-uka are quite transparent: they all involve final front or back vowel in the numeral stem. Less clear are the cases of yǝ-ka ‘four days’ and tǝwo-ka ‘ten days’: both numeral stems yǝ- ‘four’ and tǝwo- ‘ten’ end in a mid-central vowel /ǝ/.130 However, the process of *yǝ-uka > *y-uka ‘four days’ may be blocked due to the fact that OJ yǝ- ‘four’ is a monosyllabic morpheme: all other numeral morphemes undergoing the loss of a final vowel in the numerical root before the classifier *-uka are either di- or tri-syllabic (note that neither ya- ‘eight’ or mu- ‘six’ lose their final vowels, either). Under this additional condition the outcome of *tǝwǝ-uka should be *tǝw-uka, and 130  Since OJ vowels /ǝ/ and /wo/ cannot combine within a single morpheme, the archetype of OJ tǝwo ‘ten’ must be *tǝwǝ, and not *tǝwo.

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

NOMINALS

367

since a syllable /wu/ existed in Old Japanese (see chapter 2), this certainly represents a problem. Therefore, it is likely that many WOJ and PJ *-wu- were lowered to *wo, which gives us then the following scenario of the development: *tǝwǝ-uka > *tǝw-uka > *tǝw-oka, with a subsequent reanalysis as *tǝwo-ka. There is one possible counter-argument to the analysis presented above. It has long been suggested under the traditional analysis that the classifier -ka ‘day’ is related to the independent word kɛ ‘day’ < *kay well attested in Old Japanese. The problem is, however, that -ka is only used as a classifier, and never appears as the first element of compounds or in front of a genitive marker -nǝ, which is typical for other words that show [-]Ca-/[-]Cɛ alternation as described in section 1. This distribution with -ka only as a classifier and with kɛ only as an independent word seems to be quite unique. Therefore, there is a formidable internal obstacle in relating both of them directly. Both the classifier -ka ‘day’ and the independent word kɛ ‘day’ have been compared to the MK word hóy ‘sun.’ In modern Korean the word hay < hóy also means ‘daytime’ besides ‘sun,’ and the meaning ‘day’ for proto-Korean could be also reconstructed from numerical complexes denoting the number of days in Middle Korean and modern Korean: MK it-hul ‘two days,’ sa-Gol < *sa-hol ‘three days’ (cf. modern Korean sa-hul), na-Gol < *na-hol ‘four days’ (cf. modern Korean na-hul), yel-hul ‘ten days,’ etc., with its phonetic archetype as *hol-i. As I tried to argue, MK /h/ represents a lenited form of *k, which is found in the complimentary distribution with MK /G/ (Vovin 2003). It is quite possible that initial MK h- represents these lenition cases after the initial vowel was lost; thus PK *Vkol-i > MK hóy ‘sun,’ -hol ‘day’ (classifier). Therefore, it looks like PJ *-uka ‘day’ is an earlier loan from PK *Vkol ‘sun,’ ‘day,’ while the OJ independent noun kɛ ‘day’ looks like a later loan from a Korean-type language which had a word *hoy ‘sun,’ ‘day,’ similar to MK hóy ‘sun.’131 There are no other reliable external parallels to PJ *-uka ‘day.’ 3.4

Months of the Year

The Old Japanese historical texts, such as the Nihonshoki and the Shoku Nihongi, widely use the Chinese system of notation for lunar months, representing the combination of the corresponding numeral with the character 月 ‘moon,’ ‘month’: 一月 ‘first month,’ 二月 ‘second month,’ 三月 ‘third month,’ 131  Both must be loans, because they are doublets with the native Old Japanese word pi ‘day.’ I believe that the latter is native, because it has a broader distribution, while OJ -uka and ka/kɛ have much more limited usage.

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

368

Chapter 4

四月 ‘fourth month,’ 五月 ‘fifth month,’ 六月 ‘sixth month,’ 七月 ‘seventh month,’ 八月 ‘eighth month,’ 九月 ‘ninth month,’ 十月 ‘tenth month,’ 十一月 ‘eleventh month,’ and 十二月 ‘twelfth month.’ These are, of course, logographic representations, and we have no means of knowing whether the actual readings of these combinations were already borrowed from Chinese, as in the later stages of the language, or they still were the native Japanese words for the months, represented with Chinese characters. Since the native names for the months of the year are still in use in the Classical Japanese language, the second possibility is quite strong, although it does not rule out completely the possibility of the parallel usage of both Chinese loanwords and native words denoting the months of the year. The native Japanese names for the months are not based on numerals like Chinese and modern Japanese names for months, which were borrowed from Chinese. However, this is not surprising since neither Latin names for the months that are adopted in most European languages, nor the Slavic names for the months, still used in virtually all Slavic languages except Russian, are based on a numerical system. Only three native names of the months are attested in phonographic writing in Old Japanese: mutukï ‘first lunar month,’ satukï ‘fifth lunar month,’ and sipasu ‘twelfth lunar month.’ One more, simoTUKÏ ‘eleventh lunar month’ is attested in only partial phonographic spelling, with its -tukï ‘month’ part spelled logographically as 月. However, the absent names of lunar months can be tentatively reconstructed on the basis of Middle Japanese evidence. This combined evidence is provided in the chart below along with the alternative character spellings that came into being during the Heian period, with reconstructed readings marked by an asterisk (*). Since these spellings are late, in most cases they probably reflect folk etymologies, but in some cases they might be following some ancient tradition. Examples: 武都紀多知波流能吉多良婆

mu-tukï tat-i paru-nǝ k-i-tar-amba first lunar month rise-CONV spring-GEN come-CONV-PERF/PROG-COND When the first lunar month begins, and the spring has come … (MYS 5.815) 保等登藝須奈可牟佐都奇波佐夫之家牟可母

potǝtǝŋgisu nak-am-u sa-tukï pa sambusi-k-em-u kamǝ cuckoo cry-TENT-ATTR fifth-lunar.month TOP be.lonely-ATTR-TENT-ATTR EP [in] the fifth lunar month when the cuckoo will cry, [I] would be lonely! (MYS 17.3996)

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

369

NOMINALS chart 21

Native names for lunar months in Old Japanese and Middle Japanese

1st month 2nd month 3rd month 4th month 5th month 6th month 7th month 8th month 9th month 10th month 11th month 12th month

OJ

MJ

Character spelling

mutukï *kisaraŋgi *yayopi *uⁿdukï satukï *minaⁿdukï *pumitukï *pa[N]tukï *naŋgatukï *kamunaⁿdukï simoTUKÏ sipasu

mutuki kisaragi yayofi uduki satuki minaduki fuduki faduki nagatuki kamunaduki simotuki sifasu

睦月/正月 如月 弥生 卯月 五月 水無月 文月 葉月 長月 神無月 霜月 師走

乎美名古乃左衣八志毛月志波須乃加伊古本千

womina-ko-nǝ sae pa simo-TUKÏ sipasu-nǝ kai kopot-i woman-DIM-GEN talent TOP eleventh lunar month twelfth lunar monthGEN fence break-CONV Breaking fences in the eleventh [and] twelfth lunar months [is the] women’s talent (KGU 68) Comparative Data Level A: Other Japonic A2: Ryukyuan The native names for the lunar months are attested only in Old Ryukyuan, and not all of them. Only the following are extant, according to Hokama 1995: mutuki /muçici/ ‘first lunar month,’ uduki ‘fourth lunar month,’ fuduki ‘seventh lunar month,’ faduki ‘eighth lunar month,’ nagatuki /nagaçici/ ‘ninth lunar month,’ simutuki ‘eleventh lunar month,’ and sifasu ‘twelfth lunar month.’ Among those only uduki, fuduki, faduki, and simotuki occur in the oldest of Old Ryukyuan texts—the Omoro sōshi.132 Examples:

132  It is believed that uduki ‘fourth lunar month’ in OS 14.1000 is a mistake for mutuki, because it occurs in the phrase: uduki yuki fur-i-ni ‘when it snows in the fourth lunar month,’ which is, of course, unlikely.

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

370

Chapter 4

ふつきしめあらしにはつきしめあらしに

fuduki sime-arasi-ni faduki sime-arasi-ni seventh lunar month storm-LOC eighth lunar month storm-LOC during the storm in the seventh lunar month, during the storm in the eight lunar month (OS 13.961) しも月たてよれば

simoTUKI-ga tat-e-yor-e-ba eleventh.lunar.month-NOM rise-CONV-exist-EV-CON As the eleventh lunar month begins … (OS 13.817) Level B: External Comparisons A quick glance over the native names for the lunar months reveals that they have heterogeneous origins. Most of these names end in -tukï ‘month,’ but three do not: kisaraŋgi ‘second lunar month,’ yayopi ‘third lunar month,’ and sipasu ‘twelfth lunar month.’ Some of the names ending in -tukï seem to have transparent etymologies, although not necessarily in accordance with their alternative character spellings that appeared in the Heian period. Thus, minandukï is apparently the ‘month of water’ (mina-n[ǝ]-tukï), since in this month water was brought to paddies, and not the ‘month without water,’ as the alternative character spelling 水無月 suggests. Kamunaⁿdukï is probably the ‘month of deities’ (kamu-na-n[ǝ] tukï ‘deity-PLUR-GEN month’), since in this month many important ancient rites involving offerings to deities were performed, and not the ‘month without deities,’ as the alternative character spelling 神無月 suggests. The names of paⁿdukï ‘leaf month,’ simotukï ‘frost month,’ and possibly naŋgatukï ‘long month’ are likely to have the original meanings suggested by the character spelling. However, the etymologies of other months’ names are not that obvious. Let me deal only with three of the remaining names here, because I suspect that they might have at least partial external connections. Satukï ‘fifth month’ is the only one among the native names for the lunar months that does not have an alternative character spelling: the only existing character spelling for it is the one based on the Chinese system: 五月 ‘fifth month.’ It is, therefore, quite likely, that sa- in satukï means just ‘five,’ which has been pointed out before (Gluskina 1979). But there is no word *sa ‘five’ in Japonic, so it probably represents a foreign loanword. The likely source is some Central Asian language of the ‘Altaic’ type, which is extinct, but which had a initial s- instead of t- in the numeral ‘five’: *sa-.133 Thus, contrary to what has 133  There is some intimate variation between /t/ and /s/ in various Altaic languages, especially when it gets to ‘fifty’ vs. ‘five.’ Thus, MK has ta-sos ‘five’ but swuyn ‘fifty’; Written

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

NOMINALS

371

been said above regarding the fact that Old Japanese native names for months do not represent a numerical system like in Chinese, it may not be completely true—some native names for lunar months actually can represent vestiges of a system typologically close to Chinese. Two other likely candidates for loanwords are kisaraŋgi ‘second lunar month’ and mutukï ‘first lunar month.’ Mu- in mutukï ‘first lunar month’ can be tentatively identified with Tungusic *emu ‘one.’ Since kisaraŋgi ‘second lunar month’ is not attested in phonographic writing in Old Japanese, it is impossible to know whether it has kō-rui /ki/ or otsu-rui /kï/. The hypothesis which I published many years ago treats this word as *kï saran-ki ‘second month-?,’ being a possible loan from Mongolic, cf. Mongolian qoyar ‘two’ and saran ‘moon,’ ‘month’ (Vovin 1993). It is, of course, based on the assumption that at least the first /ki/ syllable in kisaraŋgi has an otsu-rui vowel /ï/ < *ǝy, *uy, *oy. The last element -ki is left unexplained, but I believe that this might be at least a working etymology.

Mongolian has ta-bun ‘five’ (Kidan tau), but Kidan word for ‘fifty’ probably had s- initial, as witnessed by Manchu susai, which likely to be a Kidan loan.

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

CHAPTER 5 Adjectives



Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

Contents of Chapter 5 Chapter 5 Adjectives  376 1 Uninflected Adjectives 377 1.1 Special Derived Form in -ra 383 1.2 -ka Adjectives  386 1.2.1 Adverbial Usage  388 1.2.2 Special Derived Form in -kɛ-  389 2 Inflected Adjectives 390 2.1 Converb Form -ku  391 2.1.1 Special Form -ku-te  394 2.1.2 Special Usage: -ku + nar- ‘To Become’ 395 2.1.3 Converb Form -ku + ar- ‘Exist’ 396 2.1.3.1 Special Constructions with Emphatic Particles mǝ and si, and Topic Particle pa 398 2.1.4 Contracted Form -k-ar- 400 2.2 Final Form -si 406 2.3 Attributive Form -ki  411 2.3.1 Nominalized Form -k-eku 418 2.3.1.1 Special Negative Nominalized Form -k-en-aku 419 2.3.2 Evidential Forms -ke- and -kere 421 2.3.3 Tentative Form -k-em- 422 2.3.4 Conditional Form -k-emba 423 2.4 Nominalized Form -sa  425 2.5 Gerund -mi  427 2.6 Deverbal Adjectives in -asi 432 3 Defective Adjectives 435 3.1 Defective Adjective ka-  435 3.2 Defective Adjective sa 437

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

In Western Old Japanese, adjectives form a transitional class between nominals and verbs in contrast to Modern and Middle Japanese, where adjectives are best defined as a special subclass of verbs, namely quality verbs (Vovin 2003: 187). The crucial distinction is that in both Modern and Middle Japanese, adjectives are always inflected similar to verbs. First, they always have attributive forms such as MdJ -i and MJ -ki ~ -i when modifying the following head noun, with very few exceptions that are limited to non-productive usage in obsolete compounds like sira-nami ‘white waves,’ where a bound adjectival stem can modify a following head noun. It is usually believed that Western Old Japanese had a similar usage, namely that an adjectival stem preceding a following head noun was also bound, and that all the existing cases represent obsolete compounds as well (Yamada 1954: 118; Shirafuji 1987: 149). As I will demonstrate below, this is certainly not the case in Western Old Japanese. Second, while in Middle Japanese the attributive and predicative functions are strictly differentiated between attributive form -ki and final predication form -si (when there are no focus particles in the sentence); this is not the case for Western Old Japanese, where especially -si can have an attributive function (Martin 1987: 807–809). All this probably points out the fact that the system of verb-like inflection in Western Old Japanese was still in the process of establishing itself, and that originally pre-Old Japanese adjectives behaved quite similar to adjectives in Turkic, Mongolic, and Tungusic, where they are nominals rather than verbs.

© Alexander Vovin, 2020 | doi:10.1163/9789004422810_006

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

Section 1

Uninflected Adjectives As mentioned above, uninflected adjectives could modify following head nouns in Western Old Japanese, and this usage is not limited to a few obsolete compounds as in Middle Japanese. Below I provide a list of the most typical adjectives that could be used as uninflected modifiers in Western Old Japanese: aka ‘red,’ ara ‘rough,’ arata ‘new,’ awo ‘green, blue,’ i ~ yu ‘sacred,’ kata ‘hard,’ kupasi ~ ŋgupasi ‘beautiful,’ kura ‘dark,’ kuro ‘black,’ kusi ‘precious,’ muna ‘empty,’ masura ‘excellent,’ miⁿdu ‘fresh,’ naŋga ‘long,’ nipi ‘new,’ opo ‘big,’ osǝ ‘slow,’ paya ‘quick, fast,’ puru ‘old,’ puto ‘majestic, great,’ pirǝ ‘broad,’ pisa ‘long,’ sakasi ‘wise,’ sikǝ ‘stupid,’ siⁿdu ‘low-class,’ sira ~ siro ‘white,’ taka ‘high,’ take ‘brave,’ tawaya ‘slender, delicate,’ tǝkǝ ‘eternal,’ tǝpo ‘distant,’ tǝyǝ ‘abundant,’ to ‘sharp,’ uma ‘sweet,’ utu ‘real,’ utu ‘empty,’ utukusi ‘beautiful,’ yasu ‘easy,’ waka ‘young,’ wawara ‘frayed.’ Although at first glance the list may not seem to be long, it is necessary to keep in mind that it is longer than the list of bound adjectives occurring in compounds in Middle Japanese, which most frequently appear only with opo- ‘big’ and sira- ‘white,’ if we exclude proper nouns and titles. It is also worth noting that one of the arguments for the bound nature of these adjectives in Western Old Japanese is based on the observation that sira ‘white’ never occurs with a following suffix, such as converb -ku, final -si, or attributive -ki; in other words its usage is limited to compounds. Meanwhile, siro- is always inflected, and never occurs in compounds. While this observation is certainly true for Middle Japanese, it does not work for Western Old Japanese, where both siro and sira could modify following nouns. Interestingly enough, even in Middle and Modern Japanese we can see traces of the original nominal nature of adjectival stems. Thus, both MJ waka-gimi ‘child of a nobleman’ and MdJ ao-zora ‘blue sky’ involve sequential voicing (連濁 rendaku), which can only be explained as a result of the reduction of the following structures: waka n-o kimi ‘young DV-ATTR lord’=> *waka ŋ-gimi > waka-gimi and awo n-o sora ‘blue DV-ATTR sky’ > *ao-ⁿ-sora > ao-zora. The historical presence of the copula n- that can only follow nominals or nominalized forms of verbs in these constructions clearly demonstrates that historically adjectival stems were just nouns.

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

378

Chapter 5

The best evidence for the independent and free nature of adjectival stems comes from the fact that another word or morpheme, including another adjective, can separate an adjective stem from a following head noun. Examples: 波毘呂由都麻都婆岐

pa-m-birǝ yu t-u ma-tumbaki leaf-GEN-broad sacred DV-ATTR INT-camellia a true sacred camellia with broad leaves (KK 57) 因名其剣謂川上部亦名曰裸伴阿箇潘娜我等母

Therefore [they] called these swords ‘Kapakami set,’ their other name is aka paⁿdaka tǝmǝ, ‘completely naked companions, (lit: red naked ­companions)’ (NS 6.189) 布刀御幣

puto mi-TE ŋGURA great HON-offering great offerings (KJK 1.20a) 阿邏瀰多摩

ara mi-tama rough HON-spirit rough spirit (NS 9.246) 登余美岐多弖麻都良勢

tǝyǝ mi-ki tatematur-as-e abundant HON-rice.wine present(HUM)-HON-IMP present the abundant rice wine (KK 101) 高照日之皇子何方尓所念食可

TAKA TER-AS-U PI-NƏ MIKO IKA SAMA n-i OMƏPOS-I-MES-E ka high shine-ATTR sun-GEN prince what manner DV-CONV think(HON)-CONVHON-EV IP in what manner does the prince of the high-shining Sun think? (MYS 2.162) 可敝里許牟麻須良多家乎尓美伎多弖麻都流

kaper-i-kǝ-m-u masura take wo-ni mi-ki tatematur-u return-CONV-come-TENT-ATTR excellent brave man-DAT HON-rice.wine present-FIN [I] will present the rice wine to the excellent, brave men who will come back (MYS 19.4262) Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

379

Adjectives 和己於保支美波多比良気久那何久伊末之弖等与美岐麻都流

wa-ŋgǝ opo kimi pa tapirakɛ-ku naŋga-ku imas-i-te tǝyǝ mi-ki matur-u I-POSS great lord TOP safe-CONV long-CONV exist(HON)-CONV-SUB abundant HON-rice.wine present(HUM)-FIN [I] present the abundant rice wine so that my sovereign (lit.: great lord) will live safely and long (SNK 4) In the following two examples tǝpo naŋga-ku and tǝpo naŋga n-i ‘for a long, long time’ are used adverbially, but they still present the first adjectival stem separated from the modified verb by another adjective. Cf. also the EOJ example from MYS 14.3356 below, where tǝpo naŋga-ki ‘distant and long’ modifies a noun phrase, and not a verb. 遠長久思将徃

TƏPO NA ŋGA-ku SINOP-i YUK-AM-U long long-CONV yearn-CONV go-TENT-FIN [I] will go yearning for a long, long [time] (MYS 2.196) 都可倍麻都良米伊夜等保奈我尓

tukapɛ-matur-am-ɛ iya tǝpo naŋga n-i serve(CONV)-HUM-TENT-EV plentifully long long DV-CONV I will serve [you] plentifully and for a long, long time (MYS 18.4098) In the following five examples the nominal nature of the adjectival stems tǝpo ‘distant’ and taka ‘high’ is also confirmed by the fact that they can be followed directly by the copulas n- or tǝ: 等保乃朝庭

tǝpo n-ǝ MIKAⁿDO distant DV-ATTR court distant court (MYS 5.794) 等保能久尓

tǝpo n-ǝ kuni distant DV-ATTR country distant country (MYS 15.3688) 登保都比等

tǝpo t-u pitǝ distant DV-ATTR person people from far away (lit.: distant people) (MYS 17.3947) Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

380

Chapter 5

等保追可牟於夜

tǝpo t-u kamu-oya distant DV-ATTR deity-ancestor distant divine ancestors (MYS 18.4096) 高津神乃災高津鳥乃災

TAKA t-u KAMÏ-nǝ WAⁿZAPAPI TAKA t-u TƏRI-nǝ WAⁿZAPAPI high DV-ATTR deity-GEN calamity high DV-ATTR bird-GEN calamity calamities from deities high [in Heaven], calamities from birds high [in the sky] (NT 10) Cf. the following example with the adjective sikǝ ‘stupid’ that has no inflected forms: 之許都於吉奈

sikǝ t-u okina stupid DV-ATTR old.man stupid old man (MYS 17.4011) The same construction with the first adjective stem not immediately preceding the modified noun can be observed in the transparent pillow-word (枕詞 makura-kotoba) pisa kata n-ǝ ‘eternally strong,’ which mostly applies to the word amɛ ~ ama- ‘heaven,’ but also can modify some other words, such as tukï ‘moon,’ and amɛ ‘rain’ (homophonous with amɛ ‘heaven’). Examples: 比佐迦多能阿米能迦具夜麻

pisa kata n-ǝ amɛ-nǝ Kaŋgu-yama eternal hard DV-ATTR heaven-GEN Kaŋgu-mountain eternal and strong Amɛ-no Kaŋguyama (lit.: Heavenly Kaŋgu mountain) (KK 27) 比佐箇多能阿梅箇儺麼多

pisa kata n-ǝ amɛ kana-m-bata eternal hard DV-ATTR heaven metal-GEN-loom eternal and strong heaven[’s] metal loom (NK 59) 比佐可多能阿米欲里由吉能那何列久流加母

pisa kata n-ǝ amɛ-yori yuki-nǝ naŋgare-k-uru kamǝ eternal hard DV-ATTR heaven-ABL snow-GEN flow(CONV)-come-ATTR EP I wonder [whether it is] snow that flows down from the eternal and strong heaven (MYS 5.822) Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

381

Adjectives

Now that the nominal nature of uninflected adjectives is established, I provide below more examples of their usage as modifiers of following head nouns or noun phrases. This usage occurs more frequently in Early Old Japanese than in Late Old Japanese, and that probably indicates that this was an archaic usage on its way out: 故志能久邇邇佐加志賣遠阿理登岐加志弖久波志賣遠阿理登岐許志弖

Kosi-nǝ kuni-ni sakasi me-wo ar-i tǝ kik-as-i-te kupasi me-wo ar-i tǝ kik-ǝs-i-te1 Kosi-GEN province-LOC wise woman-ABS exist-FIN DV hear-HON-CONV-SUB beautiful woman-ABS exist-FIN DV hear-HON-CONV-SUB [Opo kuni nusi] heard that there is a wise woman in the Kosi province, heard that there is a beautiful woman (KK 2) 斗迦麻迩佐和多流久毘

to kama-ni sa-watar-u kumbi sharp scythe-COMP PREF-cross-ATTR swan a swan flying across [with wings] like sharp scythes (KK 27) 多迦紀那流意富韋古賀波良

taka kï-n-ar-u opǝ wiko-ŋga para high place-LOC-exist-ATTR big boar-POSS plain the plain of the Great Boar that is at the high place (KK 60) 美母呂能曾能多迦紀

mi-mǝrǝ-nǝ sǝnǝ taka kï HON-mountain-GEN that high fortress that high fortress of the sacred mountain (KK 60) 阿禮許曾波余能那賀比登

are kǝsǝ pa yǝ-nǝ naŋga pitǝ I FP TOP world-GEN long person I, [the most] long[-living] man in the world (KK 72)

1  Martin treats sakasi ‘wise’ and kupasi ‘beautiful’ in this text as haplological final forms in the attributive function (Martin 1987: 807). While I agree with his argumentation that these do not form compounds with the following word me ‘woman’ because the Nihonshoki variant of the same text shows that they are accentuated as independent words (Martin 1987: 807), I believe that in the light of the evidence presented above they are better treated as free adjectival stems rather than haplological final predicative forms. Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

382

Chapter 5

那爾騰柯母于都倶之伊母我磨陁左枳涅渠農

nani tǝ kamǝ utukusi imǝ-ŋga mata sak-i-[i]ⁿde-kǝ-n-u what DV EP beautiful beloved-POSS again bloom-CONV-exit(CONV)come-NEG-ATTR I wonder why [my] beautiful beloved does not bloom again. (NK 114) 等伎波奈周迦久斯母何母等意母閇騰母

tǝk[ǝ]-ipa-nasu ka-ku si mǝŋgamǝ tǝ omǝp-ɛ-ⁿdǝmǝ eternal-rock-COMP thus-CONV EP DP DV think-EV-CONC Although [I] think that [I] would like to be (thus) like an eternal rock … (MYS 5.805) 許能久斯美多麻

kǝnǝ kusi mi-tama this precious HON-jewel these precious jewels (MYS 5.814) 志良久毛

sira kumo white cloud white clouds (MYS 5.866) 安可等吉能安左宜理其問理可里我祢曾奈久

aka tǝki-nǝ asa-ŋ-gïri-ŋ-gǝmor-i kari-ŋga ne sǝ nak-u bright time-GEN morning-GEN-fog-LOC-hide-CONV wild goose-POSS sound FP cry-ATTR wild geese cry loudly being hidden in the morning (lit.: bright time) fog at the dawn (MYS 15.3665) 之路髪

siro KAMI white hair grey hair (MYS 17.3922) 敷刀能里等其等

puto nǝritǝ-ŋ-gǝtǝ majestic Nǝritǝ-GEN-word words of the majestic Nǝritǝ (MYS 17.4031)

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

383

Adjectives 牟奈許等母於夜乃名多都奈

muna2 kǝtǝ mǝ oya-nǝ NA tat-una empty word FP ancestor-GEN name break-NEG/IMP do not destroy the name of [your] ancestors [with] empty words (MYS 20.4465) There is also one example where an adjective stem is used as a predicate without any following copula: 於朋望能農之能介瀰之瀰枳伊句臂佐伊句臂佐

opo monǝ nusi-nǝ kam-i-si mi-ki iku pisa iku pisa great thing master-GEN brew-CONV-PAST/ATTR HON-rice.wine how much long how much long The holy rice wine brewed by [the deity] Great Master of Things, [flourish] eternally (lit.: how much long, how much long) (NK 15) 1.1

Special Derived Form in -ra

There is a special derived adjectival form in -ra with an unclear meaning3 that occurs only after four adjectival stems: aka ‘red, bright,’ sakasi- ‘wise,’ usu- ‘thin’ (Yamada 1954: 123), and yǝ- ‘good’ forming derivatives aka-ra, sakasi-ra, usu-ra and yǝ-ra- respectively. Among these three sakasi-ra is used only as a nominal ‘wisdom,’ or adverbially as sakasi-ra n-i ‘wisely’4 but two other words appear as modifiers, although usu-ra ‘thin’ is attested only in one example in Western Old Japanese texts: 阿迦良袁登賣袁伊邪佐佐婆余良斯那

aka-ra wotǝme-wo iⁿza sas-amba yǝ-ra-si na ruddy-? maiden-ACC INTER stick-COND good-?-FIN EP hey, if [you take and] stick the ruddy maiden [as the ornament in your hair], [it] will be good! (KK 43)

2  The word muna ‘empty’ in this example is interesting, because it involves a usage of the root muna, rather than the stem muna-si-. 3  It could potentially be diminutive -ra, discussed in Chapter 4, section 1.2.3.1. Omodaka et al. argue that sakasi-ra, for example, actually means ‘pretended wisdom’ (Omodaka et al. 1967: 318). In another case aka-ra clearly means ‘ruddy’ (KK 43). But in other cases we do not have a strong basis to argue for a diminutive nature of this suffix. 4  In MYS 16.3860, 16.3864.

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

384

Chapter 5

痛醜賢良乎為跡

ana MINIKU SAKASI-ra-wo S-U tǝ INTER ignoble wisdom-?-ACC do-FIN DV thinking [that they] have such an ignoble wisdom, ouch (MYS 3.344) 赤羅小船

AKA-ra WOⁿ-BUNE red-? DIM-boat red little boat (MYS 16.3868) 安加良多知婆奈

aka-ra tatimbana red-? mandarin.orange red mandarin oranges (MYS 18.4060) 安可良我之波

aka-ra-ŋ-gasipa red-?-DV(ATTR)-oak red oak (MYS 20.4301) 宇須良婢

usu-ra-N-pi thin-?-DV(ATTR)-ice thin ice (MYS 20.4478) The last two examples imply that the reduced attributive form N- of the defective verb n- ‘to be’ could also be used after the special adjectival form in -ra. Comparative Data Level A: Other Japonic A1: Eastern Old Japanese Similar to Western Old Japanese, adjective stems can also be used in Eastern Old Japanese to modify following nouns or noun phrases. 麻具波思児呂

ma-ŋgupasi KO-rǝ INT-beautiful girl-DIM really beautiful girl (MYS 14.3424)

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

385

Adjectives 和可加敞流弖能毛美都麻弖宿毛等和波毛布

waka kaperute-nǝ momit-u-maⁿde NE-m-o tǝ wa pa [o]mop-u young maple-GEN leaves.turn.red/yellow-ATTR-TERM sleep-TENT-ATTR DV I TOP think-FIN I think that [we] should sleep [together] until the young maple becomes red (MYS 14.3494) 意保枳美能美己等可之古美阿乎久牟乃等能妣久夜麻乎古与弖伎怒加牟

opo kimi-nǝ mi-kǝtǝ kasiko-mi awo kumu-nǝ tǝnǝmbik-u yama-wo koyǝ-te k-i-n-o kamu great lord-GEN HON-word awesome-GER blue cloud-GEN trail-ATTR mountain-ACC cross(CONV)-SUB come-CONV-PERF-ATTR EP Since the emperor’s (lit.: great lord’s) order is awesome, [I] came [here] crossing mountains where dark clouds trail (MYS 20.4403) A2: Ryukyuan In modern Ryukyuan dialects, as far I can tell, all adjectives are inflected, and there are no cases when an adjectival stem can modify a following noun or a noun phrase. In Old Ryukyuan, however, we find the same situation as in Western and Eastern Japanese: しらにしやがおしいぢへば

sira nisiya-ga os-i-idife-ba white northern wind-NOM push-CONV-exist-CON when the first (lit.: white) northern wind blows (OS 7.349) 大と

OPO to big sea big sea (OS 13.956) たかかはのみづ

taka kafa-no midu high river-GEN water water from an upper [part] of a river (OS 17.1222) Certainly, one can argue that since this construction is limited only to Old Ryukyuan it was likely borrowed from Middle Japanese. Usually this is how I treat isolated Old Ryukyuan attestations in this grammar. However, this time

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

386

Chapter 5

the situation is different. First of all, we should not forget that by the time of Middle Japanese, all cases of adjectival stem + noun were already compounds. Therefore, we would expect that these compounds would be borrowed as such into Old Ryukyuan. Undoubtedly, some of them, like sira-tama ‘white jewel’ were borrowed into Old Ryukyuan, because (a) sira-tama is attested only in the Ryūka and the Kumi-odori, but not in the Omoro sōshi, and (b) because siratama is frequently used in Middle Japanese as a compound. However, none of the cases appearing in the above examples: sira nisiya ‘white northern wind,’ opo to ‘big sea,’ and taka kafa ‘upstream of a river’ could be loans from Middle Japanese, because none of these exist as compounds in Middle Japanese. In addition, in the Japanese branch of Japonic to ‘sea’ does not exist, nisi ‘means’ ‘west,’ and not ‘northern wind,’ and ‘upstream of a river’ is kapa kami, not taka kapa. Thus in all three cases we deal with pure Ryukyuan phrases, and it would be highly improbable if Ryukyuans borrowed certain unproductive Middle Japanese compounds, analyzed them, and then proceeded to the creation of hybrid Japanese-Ryukyuan compounds that follow a Middle Japanese pattern that was not productive itself. Thus, it seems that a construction that allowed an uninflected adjectival stem to be used as a modifier of following nouns and noun phrases represents an archaism in all branches of Japonic. Consequently, the inflection of adjectives may be a secondary development. Level B: External Comparisons Since we do not have any morphological markers here, strictly speaking there is no need to argue for or against any external connections. Let me just note that the archaic Japonic structure with adjective stems modifying following nouns typologically agrees well with Turkic, Mongolic, and Tungusic, but radically goes against the typology found in Korean. I will argue below that the adjectival inflection in Japonic could have originated under strong structural pressure from Korean. 1.2

-ka Adjectives

In Modern Japanese there is a special class of words like sizuka-na ‘quiet’ and taisetu-na ‘important’ traditionally defined by most Japanese linguists as ‘adjectival verbs’ (形容動詞 keiyōdōshi) (Tōjō 1937: 313–315; Tsukishima 1968: 123), although other treat them as nouns (Tokieda 1950: 131). Similarly, there is no agreement among Western scholars on the nature of these words. Some

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

387

Adjectives

linguists define them as adjectival nouns (Martin 1988: 179–181), or as adjectives (Henderson 1948: 179; Fel’dman 1960: 5; Alpatov 1979b: 44–45). I believe that it is appropriate to classify these words in Modern Japanese as adjectives. However, in Middle Japanese they are better defined as adjectival nouns.5 Similar to Middle Japanese, in Western Old Japanese there is a construction consisting of -ka adjective + attributive form nar-u ‘be-ATTR’ of the copula nar- + noun phrase.6 Nevertheless, in contrast to both Modern and Middle Japanese, in Western Old Japanese -ka adjectives can directly modify following nouns or noun phrases without any following forms of a copula or any suffixes. In other words, they syntactically behave in the same way as the uninflected adjectival stems described above. However, they also have a morphological peculiarity: most of them consist of a bound stem + suffix -ka, although in a few cases the bound stem can be etymologically traced to an independent noun or another adjectival root: cf. for example paⁿdaka ‘naked’ in the first example below, which is probably derived from paⁿda ‘skin.’ 因名其剣謂川上部亦名曰裸伴阿箇潘娜我等母

Therefore [they] called these swords ‘Kapakami set,’ their other name is aka paⁿdaka tǝmǝ, ‘completely naked companions, (lit: red naked companions)’ (NS 6.189) 世間之愚人乃吾妹兒尓告而語久

YƏ-NƏ NAKA-NƏ ORƏKA PITƏ-nǝ WA-ŋG-IMO-KO-ni NƏR-I-TE KATAR-Aku world-GEN middle-GEN stupid person-GEN I-POSS-beloved-DIM-DAT sayCONV-SUB speak-NML the stupid person of this world told his beloved: … (MYS 9.1740) 7使乎無跡 TASIKA NAR-U TUKAPI-wo NA-MI tǝ certain be-ATTR messenger-ABS no-GER DV thinking that there is no reliable messenger (MYS 12.2874)

5  For detailed discussion see Vovin 2003: 93–94. 6  If a noun is separated from a -ka adjective by another modifier, the -ka adjective is marked by the converb form n-i of the copula n-, see the example from SM 7 below. 7  The first character is not available in bold font, so the color red is used here as alternative.

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

388

Chapter 5

今米豆良可爾新伎政者不有本由利行來迹事曽

ΙΜΑ mɛⁿduraka8 n-i ARATASI-ki MATURI ŋGƏTƏ N-I pa AR-AⁿZ-U MƏTƏ-yuri OKƏNAP-I-KƏ-SI ATƏ KƏTƏ sǝ now strange DV-CONV new-ATTR governance DV-CONV TOP exist-NEG-CONV root-ABL conduct-CONV-come(CONV)-PAST/ATTR trace matter FP The present [one] is not a strange and new act of governance, [it is] an act that was conducted from the beginning (SM 7) 汝多知方貞仁明伎心乎以天

IMASI-tati pa SAⁿDAKA n-i AKA-ki KƏKƏRƏ-wo MƏT-I-te you-PLUR TOP loyal DV-CONV bright-ATTR heart-ACC hold-CONV-SUB you, with loyal and clear hearts … (SM 37) Overall, -ka adjectives in Western Old Japanese occur very infrequently compared to Middle Japanese. The most frequent usage seems to be an adverbial one, which is discussed below. 1.2.1 Adverbial Usage -ka adjectives are followed by the converb form n-i of the defective verb nwhen used adverbially: 於毛波奴爾横風乃爾布敷可爾覆來礼婆

omop-an-u-ni YƏKƏSIMA-KAⁿZE-nǝ nipumbuka n-i OPOP-I-K-I-TAr-e-mba think-NEG-ATTR-LOC cross-wind-GEN sudden DV-CONV cover-CONV-comeCONV-PERF/ROG-EV-CON when [I] did not think [about it], a cross wind suddenly came (MYS 5.904) 将死命尓波可尓成奴

SIN-AM-U INƏTI nipaka n-i NAR-I-n-u die-TENT-ATTR life sudden DV-CONV become-CONV-PERF-FIN [I] will die suddenly (MYS 16.3811) 於呂可尓曽和礼波於母比之乎不乃宇良能

orǝka n-i sǝ ware pa omǝp-i-si Wopu-nǝ ura insufficient DV-CONV FP I TOP think-CONV-PAST/ATTR Wopu-GEN bay the bay of Wopu about which I did not think much (lit. thought being insufficient) (MYS 18.4049) 8  Meyⁿduraka ‘strange’ is derived from the same independently unattested root *meyNtura- as in the adjective meyⁿdura-si- ‘rare, strange.’

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

389

Adjectives 於乃毛於乃毛貞仁能久清伎心乎以天奉仕

onǝ mo onǝ mo SAⁿDAKA n-i YƏ-ku KIYO-ki KƏKƏRƏ-wo MOT-I-te TUKAPƐMATUR-E yourself FP yourself FP truthful DV-CONV be.good-CONV clear-ATTR heartACC hold-CONV-SUB serve(CONV)-HUM-IMP All of you, serve truthfully and well, with a clear heart (SM 33) 1.2.2 Special Derived Form in -kɛThere are inflected adjectives with the stem-forming suffix -kɛ- that are clearly derived from -ka adjectives, although the -ka adjectives themselves are not attested in Western Old Japanese texts, e.g., akirakɛ- ‘bright’ (< *akira-ka), siⁿdukɛ- ‘quiet’ (< *siⁿdu-ka), tapirakɛ- ‘safe’ (< *tapira-ka), etc. 安伎良氣伎名

akirakɛ-ki NA bright-ATTR name bright names (MYS 20.4466) 尓波母之頭氣師

nipa mǝ siⁿdukɛ-si sea surface FP quiet-FIN the surface of the sea is also quiet (MYS 3.388) 和己於保支美波多比良気久那何久伊末之弖等与美岐麻都流

wa-ŋgǝ opo kimi pa tapirakɛ-ku naŋga-ku imas-i-te tǝyǝ mi-ki matur-u I-POSS great lord TOP safe-CONV long-CONV exist(HON)-CONV-SUB abundant HON-rice.wine present(HUM)-FIN [I] present the abundant rice wine so that my sovereign (lit.: great lord) will live safely and long (SNK 4)

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

Section 2

Inflected Adjectives In this section I will discuss the morphology that is unique to inflected adjectives. Therefore, the morphological markers that are shared by both inflected adjectives and verbs, such as prefixes, will be treated in Chapter 6. There are two classes of inflected adjectives in Western Old Japanese: class 1 (traditional く活用 ku-katsuyō type), which shows no irregularities and has suffixes attached to the adjectival stem that is identical to the adjectival root; and class 2 (traditional しく活用 shiku-katsuyō type), which shows one slight irregularity (haplological contraction of the final -si-si form into -si-∅) and has suffixes attached to the adjectival stem in -si- that represents an extension after an adjectival root. The two adjectival classes can be summarized in the following chart that presents two typical adjectives, aka- ‘red’ and utukusi- ‘beautiful,’ from each class followed by all suffixes that can directly follow an adjectival stem in Western Old Japanese. chart 22 Classes of inflected adjectives in Western Old Japanese

Forms

Class 1

Class 2

converb -ku final -si attributive -ki evidential 1 -ke evidential 2 -kere nominalizer 1 -keku nominalizer 2 -sa gerund -mi

aka- ‘red’ aka-ku aka-si aka-ki aka-ke aka-kere aka-keku aka-sa aka-mi

utuku-si- ‘beautiful’ utuku-si-ku utuku-si-∅ utuku-si-ki utuku-si-ke utuku-si-kere utuku-si-keku utuku-si-sa utuku-si-mi

Note: both evidential forms and the nominalizer in -keku are secondary forms based on the attributive form -ki.

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

391

Adjectives

The adjectival forms from Chart 22 above will be discussed in detail below. 2.1 Converb Form -ku The converb form of inflected adjectives is formed by attaching the suffix -ku to the stem: aka-ku ‘red-CONV,’ siro-ku ‘white-CONV,’ utukusi-ku ‘beautifulCONV,’ saŋgasi-ku ‘steep-CONV,’ etc. The converb form has two main functions in Western Old Japanese: (1) non-final predicate and (2) adverbial. Examples of -ku as a non-final predicate: 多陀爾阿波須阿良久毛於保久志岐多閇乃麻久良佐良受提伊米爾之美延牟

taⁿda n-i ap-aⁿz-u ar-aku mo opo-ku sik-i-tapɛ-nǝ makura sar-aⁿz-u-te imɛ-ni si mi-ye-m-u direct DV-CONV meet-NEG-CONV exist-NML FP many-CONV spread-CONVmulberry.tree.bark.cloth-GEN pillow go.away-NEG-CONV-SUB dream-LOC EP see-PASS-TENT-FIN There are many occasions when [we] do not meet directly, and [I] want to see [you] in [my] dream without going away from [your] mulberry tree bark cloth pillow (MYS 5.809) 萬世爾得之波岐布得母烏梅能波奈多由流己等奈久佐吉和多流倍子

YƏRƏⁿDU YƏ n-i tǝsi pa k-i p-u tǝmǝ uMƐ-nǝ pana tay-uru kǝtǝ na-ku sak-i-watar-umbɛ-si ten.thousand generation DV-CONV year TOP come-CONV pass-FIN CONJ plum-GEN blossom be interrupted-ATTR matter no-CONV bloom-CONV-cross-DEB-FIN Although years will come and pass for ten thousand generations, plum blossoms would continue to bloom without interruption (MYS 5.830) 雪布流欲波為部母奈久寒之安礼婆

YUKI pur-u yo pa SUmBE mǝ na-ku SAMU-KU si ar-e-mba snow fall-ATTR night TOP way FP no-CONV cold-CONV EP exist-EV-CON because in the night when snow is falling nothing can be done and [it] is indeed cold (MYS 5.892)

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

392

Chapter 5

都追牟許等奈久波也可敞里麻勢

tutum-u kǝtǝ na-ku paya kaper-i-[i]mas-e have.diffuculty-ATTR matter no-CONV fast return-(CONV)-HON-IMP Return quickly, without having difficulties (MYS 15.3582) 清麻呂其我姉法均止甚大尓悪久奸流妄語乎作弖

KIYOMARƏ SI-ŋga ANE POPUKUN-tǝ ITƏ OPO-KI n-i ASI-ku KAⁿDAM-Er-u ITUPAR-I-ŋ-GƏTƏ-wo TUKUR-I-te Kiyomarǝ he-POSS elder.sister Popukun-COM very big-ATTR DV-CONV badCONV be insincere-PROG-ATTR lie-NML-GEN-word-ACC make-CONV-SUB Kiyomarǝ with his elder sister Popukun created an extremely bad and insincere lie … (SM 44) 狭國者広久峻國者平久

SA-KI KUNI pa PIRƏ-ku SA ŋGASI-KI KUNI pa TAPIRAKƐ-ku narrow-ATTR land TOP broad-CONV steep-ATTR land TOP level-CONV The narrow land will be broad, and the steep land will be level, and … (NT 1) Examples of an adverbial usage of -ku: 志我都矩屢麻泥爾飫裒枳瀰爾柯柂倶都柯陪麻都羅武

si-ŋga tukur-u9-maⁿde-ni opo kimi-ni kata-ku tukapɛ-matur-am-u they-POSS come.to.an.end-ATTR-TERM-LOC great lord-DAT strong-CONV serve(CONV)-HUM-TENT-FIN Until they come to an end, [I] intend to serve faithfully to the emperor (NK 78) 乎武例我禹杯爾倶謨娜尼母旨屢倶之多多婆

wo-mure-ŋga upɛ-ni kumo ⁿdani mǝ siru-ku si tat-amba DIM-mountain-POSS top-LOC cloud RP FP distinct-CONV EP rise-COND if even the clouds distinctly rise above the small mountain (NK 116)

9  The verb tukur- is an intransitive counterpart of tukus- ‘to exhaust, to exert (oneself).’ It occurs only in Old Japanese, and it is interesting that it did not make its way into any dictionaries, including Omodaka et al. (1967). The much more widespread intransitive equivalent of tukus- is OJ tukiy- u, typical for Central Japanese, but this is not the case, as PR form is definitely *-ku, not *-ko.12 2.2 Final Form -si As it was mentioned at the beginning of this chapter, in contrast to Middle Japanese, Western Old Japanese -si does not necessarily have an exclusive function of a final predication marker, as it can be sometimes used as an attributive marker as well. This specific usage certainly points to the fact that the final marker -si and the attributive marker -ki were not as strictly differentiated functionally in Western Old Japanese as they were in Middle Japanese. This phenomenon in its own turn strongly speaks in favor of a point of view that both -si and -ki are relatively late innovations in Japanese, especially given the fact that both markers are completely absent from the Ryukyuan branch of Japonic. Nevertheless, on the synchronic level, the function of -si in most 12  Martin also adds to this comparison OJ nominalizer -ku used after verbs, e.g., ip-aku ‘the fact that [someone] says,’ but I believe it does not belong here, because it is rather -aku than just -ku, as can be seen from such forms as mi-r-aku ‘the fact that [someone] sees’ and s-ur-aku ‘the fact that [someone] does,’ which make it clear that it is -aku and not -ku that historically follows the attributive form of verbs, thus, we have the following developments: ip-aku < *ip-u-aku, mi-r-aku < *mi-ru-aku, and s-ur-aku < *s-uru-aku. The same is true for adjectives that have nominalized forms in -keku < attributive -ki + nominalizer -aku (see section 2.3.2).

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

407

Adjectives

Western Old Japanese textual examples is firmly connected with final predication. This phenomenon probably means that -si was on its way to be established as a marker of final predication of inflected adjectives. Examples: 阿波母與賣迩斯阿禮婆那遠岐弖遠波那志

a pa mǝ yǝ me n-i si ar-e-mba na-wo [o]k-i-te wo pa na-si I TOP EP EP woman DV-CONV EP exist-EV-CON you-ACC leave-CONV-SUB man TOP no-FIN Because I am a woman, [I] have no [other] man, besides you (KK 5) 夜多能比登母登須宜波比登理袁理登母意富岐弥斯與斯登岐許佐婆比登理 袁理登母

Yata-nǝ pitǝ-mǝtǝ suŋgɛ pa pitǝ-ri wor-i tǝmǝ opǝ kimi si yǝ-si tǝ kikǝs-amba pitǝ-ri wor-i tǝmǝ Yata-GEN one-CL sedge TOP one-CL exist-FIN CONJ great lord EP be.goodFIN DV say(HON)-COND one-CL exist-FIN CONJ Even if one sedge from Yata is alone, if the great lord says [it] is fine, even if [she] is alone (KK 65) 許斯母阿夜爾加志古志

kǝ si mǝ aya n-i kasiko-si this EP FP very DV-CONV awesome-FIN This is very awesome, too (KK 100) 区茂能於虚奈比虚予比辞流辞毛

kumo-nǝ okǝnap-i kǝ yǝpi siru-si mo spider-GEN perform-NML this night distinctive-FIN EP the spider’s performance is distinctive tonight (NK 65) 情佐麻祢之

KƏKƏRƏ sa mane-si thought so many-FIN [sad] thoughts are so many (MYS 1.82) 尓波母之頭氣師

nipa mǝ siⁿdukɛ-si sea.surface FP quiet-FIN the surface of the sea is also quiet (MYS 3.388)

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

408

Chapter 5

則許母倍婆許己呂志伊多思

sǝkǝ [o]mǝp-ɛ-mba kǝkǝrǝ si ita-si there think-EV-CON heart EP be.painful-FIN when [I] think of those places, [my] heart aches (MYS 17.4006) 許礼乎於伎低麻多波安里我多之

kǝre-wo ok-i-te mata pa ar-i-ŋ-gata-si this-ACC leave-CONV-SUB again TOP exist-NML-GEN-hard-FIN [It] is difficult to find again [a falcon of the same quality], except this [one] (MYS 17.4011) The attributive usage of -si is attested as well, albeit in fewer examples, most of which belong to Early Old Japanese, which again speaks in favor of the archaic nature of this type of usage. The clear-cut cases can only be found in the case of the adjectives belonging to class 1, since the adjectives of class 2, as I mentioned above, do not have a final predication form different from their stem. Examples: 登富登富斯故志能久迩迩

tǝpǝ tǝpǝ-si Kosi-nǝ kuni-ni distant distant-FIN Kosi-GEN land-LOC in the distant, distant land of Kosi (KK 2) 夜本爾余志伊岐豆岐能美夜

yapo ni yǝ-si i-kiⁿduk-i n-ǝ miya eight.hundred ground be.good-FIN DLF-build-CONV DV-ATTR palace a palace built on an eight hundred [times] good soil (KK 100) 意布袁余志斯毘

op[o] uwo yǝ-si simbi big fish be.good-FIN tuna tuna that is good big fish (KK 110) 波辞枳豫辞和芸幣

pasi-ki yǝ-si wa-ŋg-ipe lovely-ATTR be.good-FIN I-POSS-house my lovely and good house (NK 21)13 13  Traditional Japanese scholarship treats yǝsi in this text as an exclamation on the basis of the fact that the Kojiki textual variant of the same poem has pasike yasi (KK 32) (Tsuchihasi

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

409

Adjectives 婀鳴爾與志乃楽

awo ni yǝ-si Nara green earth good-FIN Nara Nara [mountain, where] the green earth is good (NK 95) 蚊黒為髪尾信櫛持於是蚊寸垂

kaŋ-GURO-si KAMI-wo MA-KUSI MƏT-I kaki-TARE INT-black-FIN hair-ACC INT-comb hold-CONV PREF-make.hang. down(CONV) making pitch black hair hang down with a comb (MYS 16.3791) 安良志乎須良尓奈氣枳布勢良武

ara-si wo sura n-i naŋgɛk-i pus-er-am-u be.rough-FIN man RP DV-CONV lament-CONV lie.prone-PROG-TENT-FIN even a rough man would be lying down and lamenting (MYS 17.3962) A clear relic of the usage of -si in attributive function is found in the special construction with locative marker -ni, which normally appears after the attributive form (see chapter 4, section 1.2.2.4 for details). However, in the case of the negative existential verb na- ‘there is no, not to exist,’ -ni follows not the attributive -ki, but the final -si: 伊豆毛多祁流賀波祁流多知都豆良佐波麻岐佐味那志爾阿波禮

Iⁿdumo Takeru-ŋga pak-er-u tati tuⁿdura sapa mak-i sa-mï na-si-ni apare Iⁿdumo Takeru-POSS wear-PROG-ATTR long.sword vine many wrap-CONV PREF-body no-FIN-LOC INTER the long sword that Iⁿdumo Takeru is wearing is wrapped with many vines, [but] because there is no blade, alas! (KK 23)

1957: 138). But the Kojiki text, which has a worse history of textual transmission than the Nihonshoki, appears to be corrupted: it is highly unlikely that Western Old Japanese would have preserved the pre-raised form of attributive form *-ke (> WOJ -ki) as in pasike; therefore there is a great chance that yasi is also a textual corruption of yǝ-si ‘good-FIN.’ Here, as well as in other cases when the Kojiki and the Nihonshoki have discrepancies, I mostly rely on the Nihonshoki text.

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

410

Chapter 5

余家久波奈之爾漸漸可多知都久保利

yǝ-k-eku pa na-si-ni YAKUYAKU N-I katati tukupor-i be.good-ATTR-NML TOP no-FIN-LOC gradually DV-CONV facial features get emaciated-CONV as there was no improvement, [his] face became gradually emaciated, and … (MYS 5.904) Comparative Data Level A: Other Japonic A1: Eastern Old Japanese In Eastern Old Japanese -si is attested only in the function of a final predication marker. 阿爾久夜斯豆之曾能可抱与吉爾

ani k-u ya siⁿdu-si sǝ-nǝ kapo yǝ-ki-ni INTER come-FIN IP calm-FIN it-GEN face good-ATTR-LOC [contrary to my expectations,] does [it] come? [It] is calm. Its face is good, but [it does not move forward] (MYS 14.3411) 多妣己呂母夜倍伎可佐祢弖伊努礼等母奈保波太佐牟志伊母尓志阿良祢婆

tambi kǝrǝmǝ ya-pɛ ki-kasane-te i n-ure-ⁿdǝmǝ napo paⁿda samu-si imǝ si ar-an-e-mba travel garment eight-CL wear(CONV)-pile(CONV)-SUB sleep sleep-EV-CONC still skin be.cold-FIN beloved EP exist-NEG-EV-CON Because [my] beloved is not [here], although [I] sleep piling up eight layers of [my] travel garments, [my] skin is still cold (MYS 20.4351) 多妣波久流之

tambi pa kuru-si travel TOP be.hard-FIN [My] travel is hard (MYS 20.4406) A2: Ryukyuan The final predicative form -si is not attested in Ryukyuan. This invites two possible solutions for proto-Japonic: either Japanese -si is an innovation, developed in Japanese independently of Ryukyuan, or it is an archaism that was lost in Ryukyuan. Since, as I have demonstrated above, Western Old Japanese still preserves the vestiges of an earlier system when uninflected adjectival stems could modify following nominals or noun phrases, the former solution

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

411

Adjectives

maintaining that -si represents a secondary development in Japanese that did not affect Ryukyuan seems to be more viable and realistic. Level B: External Comparisons Given the fact that OJ -si can serve as both a final predication form and an attributive, as well as the fact that it is not present in Ryukyuan, it is likely to be a loan from the Old Korean irrealis attributive marker -l (phonetically probably voiceless [lh] as indicated by the character 尸 with which it was written):14 慕理尸心未行乎尸道尸

KUli-l MOSOm-i NYE-wo-l KIl long for-ATTR/IRR mind-GEN go-MOD-ATTR/IRR way the way that [my] mind, longing for [you], is going (Hyangka 1.7) 臣隱愛賜尸母史也

SIN-un TOSO-si-l Esi I-LA retainer-TOP love-HON-ATTR/IRR mother be-FIN Retainers are loving mothers (Hyangka 3.2) The case for borrowing is further supported by the fact that two other adjectival markers, uniquely attested in Japanese, but not in Ryukyuan, attributive -ki and gerund -mi, also likely have a Korean origin (see 2.3 and 2.5). 2.3 Attributive Form -ki The main function of the marker -ki is attributive. Adjectives with this marker normally modify following nominals or nominal phrases. 久路岐美祁斯遠麻都夫佐爾登理與曾比 … 阿遠岐美祁斯遠麻都夫佐迩登理 與曾比

kuro-ki mi-kes-i-wo ma-tumbusa n-i tǝr-i-yǝsǝp-i … awo-ki mi-kes-i-wo matumbusa n-i tǝr-i-yǝsǝp-i be.black-ATTR HON-wear(HON)-NML-ACC INT-without fail DV-CONV take-CONV-dress-CONV … be.blue-ATTR HON-wear(HON)-NML-ACC INT-without fail DV-CONV take-CONV-dress-CONV wearing neatly a black garment … wearing neatly a blue garment (KK 4) 14  There are other apparent loans from OK into Japanese that reflect the correspondence OK -l ~ OJ -si, e.g., MK kàlàp ‘oak’ ~ OJ kasi ‘id.,’ MK kál ‘cangue’ ~ MJ kasi ‘id.’

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

412

Chapter 5

志祁志岐袁夜迩

sikesi-ki wo-ya-ni be.quiet-ATTR DIM-house-LOC in the quiet little house (KK 19) 於朋耆妬庸利于介伽卑氐

opǝ-ki to-yori ukakap-i-te be.big-ATTR door-ABL peek-CONV-SUB peeking from the big door (NK 18) 人毛奈吉空家者草枕旅尓益而辛苦有家里

PITƏ mo na-ki MUNASI-KI IPE pa KUSA MAKURA TA ͫBI-ni MASAR-I-TE KURUSI-K-ar-i-ker-i person FP no-ATTR be.empty-ATTR house TOP grass pillow journey-LOC exceed-CONV-SUB be.painful-CONV-exist-CONV-RETR-FIN It turned out that an empty house with nobody in [it] is more painful than a journey [when one uses] grass for a pillow (MYS 3.451) 余能奈可波牟奈之伎母乃等志流等伎子伊与余麻須万須加奈之可利家理

yǝ-nǝ naka pa munasi-ki mǝnǝ tǝ sir-u tǝki si iyǝyǝ masu-masu kanasi-k-ar-i-ker-i world-GEN middle TOP be.empty-ATTR thing DV know-ATTR time EP more. and.more more.and.more be.sad-CONV-exist-CONV-RETR-FIN When [I] realized that the world is (an) empty (matter), [it] turned out to be more and more sad (MYS 5.793) 和可伎児等毛波乎知許知爾佐和吉奈久良牟

waka-ki KO-ⁿdǝmo pa woti kǝti-ni sawak-i-nak-uram-u be.young-ATTR child-PLUR TOP there here-LOC make. noise-CONV-cry-TENT2-FIN young children will probably cry loudly here [and] there (MYS 17.3962) 安伎良氣伎名

akirakɛ-ki NA be.bright-ATTR name bright names (MYS 20.4466)

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

413

Adjectives 伎多奈伎微乎婆伊止比須都閇志波奈礼須都倍志

kitana-ki mï-womba itǝp-i-sut-umbɛ-si panare-sut-umbɛ-si be.dirty-ATTR body-ACC(EMPH) hate-CONV-discard-DEB-FIN leave(CONV)-discard-DEB-FIN [I] should hate and discard [my] unclean body. [I] should leave and discard [it] (BS 19) In the prose texts several adjectival modifiers marked with -ki may follow one after another as in the following example from SM 1: 貴支高支広支厚支大命

TAPUTO-ki TAKA-ki PIRƏ-ki ATU-ki OPO MI-KƏTƏ be.awesome-ATTR be.high-ATTR be.broad-ATTR be.thick-ATTR great HON-word awesome, high, broad, and strong imperial edict (SM 1) 今米豆良可爾新伎政者不有本由利行來迹事曽

ΙΜΑ mɛⁿduraka n-i ARATASI-ki MATURI ŋGƏTƏ N-I pa AR-AⁿZ-U MƏTƏ-yuri OKƏNAP-I-KƏ-SI ATƏ KƏTƏ sǝ now strange DV-CONV be.new-ATTR governance DV-CONV TOP exist-NEGCONV root-ABL conduct-CONV-come(CONV)-PAST/ATTR trace matter FP The present [one] is not a strange and new act of governance, [it is] an act that was conducted from the beginning (SM 7) 汝多知方貞仁明伎心乎以天

IMASI-tati pa SAⁿDAKA n-i AKA-ki KƏKƏRƏ-wo MƏT-I-te you-PLUR TOP loyal DV-CONV be.bright-ATTR heart-ACC hold-CONV-SUB you, with loyal and clear hearts … (SM 37) The attributive form -ki also replaces the final predication form -si if the particles sǝ ~ ⁿzǝ, ya, ka, kamǝ, and kǝsǝ are found previously in the sentence.15 This rule is known in traditional grammar as 係り結び kakari musubi ‘the rule 15  Strictly speaking, the particles ka and kamǝ trigger the change of the final predication form -si to the attributive -ki irrespective of their position in the sentence when action verbs are involved. However, to the best of my knowledge, the particle ka is not attested after adjectival predicates in attributive form in Western Old Japanese. It must also be mentioned that examples with adjectival attributives triggered by the particle namo (a cognate of MJ namu) are considered to be unattested (Yamada 1954: 115), but there is

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

414

Chapter 5

of linking.’ Note that in contrast to Middle (Classical) Japanese the particle kǝsǝ triggers in Western Old Japanese a change of the adjectival final predication -si into the attributive -ki, and not into the evidential form.16 瀰致喩區茂能茂多遇譬低序豫枳

miti yuk-u monǝ mo taŋgupi-te ⁿzǝ yǝ-ki road go-ATTR person FP companion-hand FP be.good-ATTR [It] is also good for a person who travels to have a companion (NK 50) 比等余里波伊毛曽母安之伎

pitǝ-yǝri pa imo sǝ mǝ asi-ki person-ABL TOP beloved FP EP be.bad-ATTR [my] beloved is worse than [other] people (MYS 15.3737) 枳彌波夜那祇

kimi pa ya na-ki lord TOP IP no-ATTR Do not [you] have a lord? (NK 104) 可牟加良夜曽許婆多敷刀伎

kamu-kara ya sǝkǝmba taputo-ki deity-origin IP very be.awesome-ATTR Is [not] the origin of deities very awesome? (MYS 17.3985) 椋橋乃山乎高可夜隠尓出來月乃光乏寸

KURAPASI-nǝ YAMA-wo TAKA-MI ka YO-ŋ-GƏMƏR-I-ni IⁿDE-K-URU TUKÏ-nǝ PIKARI TƏMƏSI-ki Kurapasi-GEN mountain-ABS high-GER IP night-GEN-hide-NML-LOC exist(CONV)-come-ATTR moon-GEN light be.scanty-ATTR Is [it] because Mount Kurapasi is high that the light of the moon that goes out in the dead of the night is scanty? (MYS 3.290)

at least one controversial example in logographic spelling (see the example from MYS 12.2877 in chapter 6, section 3.2.1.2.3 and chapter 9, section 1.4). 16  This does not affect verbs that in both Western and Eastern Old Japanese change their final predication forms into evidential similar to Middle Japanese.

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

415

Adjectives 片念為歟比者之吾情利乃生戸裳名寸

KATA-OMƏP-I S-URE KA KƏNƏ KƏRƏ-NƏ WA-ŋGA KƏKƏRƏ-to-nǝ IK-ER-U to mo na-ki one-love-NML do-EV IP this time-GEN I-POSS heart-place-GEN live-PROGATTR place FP no-ATTR Is [not it because] of unrequited love that I have no intention to live at this time (MYS 11.2525) 美母呂能伊都加斯賀母登加斯賀母登由由斯伎加母加志波良袁登賣

mi-mǝrǝ-nǝ i t-u kasi-ŋga mǝtǝ kasi-ŋga mǝtǝ yuyusi-ki kamǝ kasi-para wotǝme HON-mountain-GEN sacred DV-ATTR oak-POSS below oak-POSS below be.awesome-ATTR EP oak-field maiden Under the sacred oaks of the sacred mountain, is [she not] awesome, a maiden from the oak field, I wonder? (KK 92) In this example kamǝ follows the adjectival form while in the next two examples it precedes it. 奈何鴨目言乎谷裳幾許乏寸

naNI SI kamo MƐ KƏTƏ-wo ⁿdani mo KƏKƏⁿDA TƏMƏSI-ki what EP EP eye word-ABS RP FP so.much be.scarce-ATTR why even [our] meetings are so scarce, I wonder? (MYS 4.689) 天漢敝太而礼婆可母安麻多須辨奈吉

AMA-NƏ ŋGAPA peⁿdat-ure-mba kamǝ amata sumbe na-ki heaven-GEN river separate-EV-CON EP many way no-ATTR I wonder, is [it not] because the Heavenly River separates [them], there is not much that can be done (MYS 8.1522) 虚呂望虚曾赴多幣茂予耆

kǝrǝmǝ kǝsǝ puta-pe mo yǝ-ki garment FP two-CL FP be.good-ATTR [It] is good [to wear] two layers of garments (NK 47) 野乎比呂美久佐許曽之既吉

NO-wo pirǝ-mi kusa kǝsǝ siŋgɛ-ki field-ABS wide-GER grass FP be.thick-ATTR because the fields are wide, grass grows thickly (MYS 17.4011)

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

416

Chapter 5

The attributive form in -ki can also be used as a noun: 武都紀多知波流能吉多良婆可久斯許曽烏梅乎々利都々多努之岐乎倍米

mu-tukï tat-i paru-nǝ k-i-tar-amba ka-ku si kǝsǝ uMƐ-wo wor-i-tutu tanosi-ki wopɛ-m-ɛ first lunar month rise-CONV spring-GEN come-CONV-PERF/PROG-COND be.thus-CONV EP FP plum.blossom-ACC break-CONV-COOR be.pleasantATTR finish-TENT-EV When the first lunar month begins, and the spring has come, let [us] enjoy the pleasure to the end while picking plum blossoms (MYS 5.815) 安布倍伎与之能奈伎我佐夫之佐

ap-umbɛ-ki yǝsi-nǝ na-ki-ŋga sambusi-sa meet-DEB-ATTR chance-GEN no-ATTR-POSS be.sad-NML sadness of the non-existence of a chance to be able to meet (MYS 15.3734) Comparative Data Level A: Other Japonic A1: Eastern Old Japanese Both forms -ki and -ke (spelled variously as -kɛ or -ke) are attested as attributives in Eastern Old Japanese. Although the first one is statistically more frequent, the second is almost invariably found in the poems that have other Eastern Old Japanese features. In addition, only -ke is found in region A, both -ke and -ki co-occur in region B, and only -ki is found in region C (Hino 2003: 200). Thus, in all likelihood, the real Eastern Old Japanese form was -ke, that shows the pre-raised stage of the vowel *e that underwent raising to /i/ in Western Old Japanese. 於毛思路伎野乎婆奈夜吉曾

omosiro-ki NO-womba na-yak-i-sǝ be.beautiful-ATTR field-ACC(EMPH) no-burn-CONV-do Do not burn the beautiful field (MYS 14.3452) 可奈之家伊母

kanasi-ke imǝ be.pretty-ATTR beloved pretty beloved (MYS 20.4369)

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

417

Adjectives 阿志氣比等奈里

asi-kɛ pitǝ nar-i be.bad-ATTR person be-FIN [he] is a bad person (MYS 20.4382) 奈賀氣己乃用

naŋga-kɛ kǝnǝ yo be.long-ATTR this night this long night (MYS 20.4394) 麻可奈之伎西呂

ma-kanasi-ki se-rǝ INT-beloved-ATTR husband-DIM [my] (really) beloved husband (MYS 20.4413) Kakari-musubi is also present in Eastern Old Japanese: 阿母志々尓己等麻乎佐受弖伊麻叙久夜之氣

amǝ sisi-ni kǝtǝ mawos-aⁿz-u-te ima ⁿzǝ kuyasi-kɛ mother father-DAT word say(HUM)-NEG-CONV-SUB now FP be.regretful-ATTR now [I] regret that [I] did not tell [my] mother and father (MYS 20.4376) 安是可加奈思家

aⁿze ka kanasi-ke why IP be.dear-ATTR Why is [she so] dear [to me]? (MYS 14.3576) A2: Ryukyuan The attributive marker -ki is not attested in Ryukyuan. For the same reasons as already outlined above regarding final predication marker -si, it is better treated as an internal innovation in Japanese that did not affect Ryukyuan. Level B: External Comparisons As I mentioned above, since WOJ -ki is paralleled by EOJ -ke, we have to reconstruct PJN *-ke. Because the distribution of this marker is limited to Japanese, and because the origin of adjectival inflection appears to be a recent phenomenon, it is possible that PJN *-ke is a loan from OK -kuy, an attributive marker of quality verbs:

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

418

Chapter 5

東京明期月良

TWONG-KYENG POLK-kuy TOLAL-la Eastern Capital bright-ATTR moon-LOC At the bright moon in the Eastern Capital … (Hyangka 5.1) Two obvious problems are present here. First, although OK 明期 POLK-kuy ‘bright’ is clearly used in the attributive function, this is the only example where OK 期 -kuy occurs as an attributive marker of a quality verb in Old Korean. Second, the -k- portion of -kuy may really belong to the stem POLK- ‘bright.’ There is really no way to tell, because both cases like 心未 MOSOm-i ‘mindGEN’ in Hyangka I.7, where the phonographically used character 未 renders both final consonant of the stem and a suffix, and 月良 TOLAL-la ‘moon-LOC’ in the just cited Hyangka 5.1 where only the suffix is written phonographically, are present in Old Korean texts. 2.3.1 Nominalized Form -k-eku Western Old Japanese has a special nominalized form -k-eku, an apparent contraction of attributive -ki and nominalizer -aku. The uncontracted form *-kiaku is not attested in the Western Old Japanese texts. This form did not survive into the later stages of the language. Examples: 伊良那祁久曾許爾淤母比傅加那志祁久許許爾淤母比傅

irana-k-eku sǝkǝ-ni omǝp-i-[i]ⁿde kanasi-k-eku kǝkǝ-ni omǝp-i-[i]ⁿde be.sorrowful-ATTR-NML there-LOC think-CONV-exit(CONV) be.sad-ATTRNML here-LOC think-CONV-exit(CONV) [I] recollect that [with] sorrow, [I] recollect that [with] sadness (KK 51) 宇麼能耶都擬播嗚思稽矩謀那斯

uma-nǝ ya-tu-ŋgï pa wosi-k-eku mo na-si horse-GEN eight-CL-? TOP grudge-ATTR-NML FP no-FIN eight horses are not even to be grudged (NK 79) 平氣久安久母阿良牟遠事母無裳無母阿良牟遠世間能宇計久都良計久

TAPIRAkɛ-ku YASU-ku mǝ ar-am-u-wo KƏTƏ mǝ NA-KU MƏ NA-KU mǝ ar-am-uwo YƏ-NƏ NAKA-nǝ u-k-eku tura-k-eku peaceful-CONV quiet-CONV FP exist-TENT-ATTR-ACC matter FP no-CONV misfortune no-CONV FP exist-TENT-ATTR-ACC world-GEN inside-GEN be.sad-ATTR-NML be.painful-ATTR-NML although [I] want to live peacefully and quietly, although [I] want to live uneventfully and without misfortunes, the fact that life is sad and painful … (MYS 5.897) Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

419

Adjectives 余家久波奈之爾漸漸可多知都久保利

yǝ-k-eku pa na-si-ni YAKUYAKU N-I katati tukupor-i be.good-ATTR-NML TOP no-FIN-LOC gradually DV-CONV facial features get emaciated-CONV as there was no improvement, [his] face became gradually emaciated, and … (MYS 5.904) 可之故伎美知乎也須家口母奈久奈夜美伎弖

kasiko-ki miti-wo yasu-k-eku mǝ na-ku nayam-i k-i-te awesome-ATTR road-ACC be.peaceful-ATTR-NML FP no-CONV suffer-CONV come-CONV-SUB [he] came along the awesome road, suffering and without peace [of mind] (MYS 15.3694) 伊多家苦乃日異麻世婆

ita-k-eku-nǝ PI-NI KE N-I mas-e-mba be.painful-ATTR-NML-GEN day-LOC unusual DV-CONV increase-EV-CON as [my] pain unusually increases [every] day (MYS 17.3969) 2.3.1.1 Special Negative Nominalized Form -k-en-aku There is also a special adjectival negative nominalized form -k-en-aku,17 that consists of the adjectival attributive -ki, negative -an-, and nominalizer -aku. The vowel /i/ of the attributive and the vowel /a/ of the negative monophthongize into /e/, so the indication of the morphemic boundary between the attributive and negative is artificial. This form occurs only with two inflected adjectives: na- ‘to be non-existent’ and yasu- ‘to be easy.’ There are only three examples where this form is attested phonographically, all of them in the Man’yōshū. Two of these examples represent the identical context, although in two different texts. 多婢等伊倍婆許等尓曽夜須伎須久奈久毛伊母尓戀都々須敝奈家奈久尓

tambi tǝ ip-ɛ-mba kǝtǝ n-i sǝ yasu-ki sukuna-ku mo imǝ-ni KOPÏ-tutu sumbe na-ken-aku n-i journey DV say-EV-CON word DV-CONV FP be.easy-ATTR be.few-CONV FP beloved-DAT long.for(CONV)-COOR way no-ATTR-NEG-NML DV-CONV It is easy to talk about a journey, but at least continuing to long for [my] beloved is a big deal (MYS 15.3743)

17  I am grateful to Janick Wrona who pointed out to me the existence of this form. Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

420

Chapter 5

奈氣久蘇良夜須家奈久尓於母布蘇良久流之伎母能乎

naŋgɛk-u sora yasu-k-en-aku n-i omǝp-u sora kurusi-ki monǝwo lament-ATTR RP be.easy-ATTR-NEG-NML DV-CONV long.for-ATTR RP be.hard-ATTR CONJ although even to lament is not easy, and even to long is difficult (MYS 17.3969) 嘆蘇良夜須家奈久尓念蘇良苦伎毛能乎

NA ŋGƐK-U sora yasu-k-en-aku n-i OMƏP-U sora KURUSI-ki monǝwo lament-ATTR RP be.easy-ATTR-NEG-NML DV-CONV long.for-ATTR RP be.hard-ATTR CONJ although even to lament is not easy, and even to long is difficult (MYS 19.4169) Comparative Data Level A: Other Japonic A1: Eastern Old Japanese There is only one example of -k-eku found in Eastern Old Japanese texts, but it occurs in a poem with strong dialectal features, therefore it should be accepted as a genuine one: 安奈由牟古麻能乎之家口母奈思

a nayum-u koma-nǝ wosi-k-eku mǝ na-si foot suffer-ATTR stallion-GEN regrettable-ATTR-NML FP no-FIN [I] have no sorry feelings for [my] stallion that will hurt [his] feet (MYS 14.3533) In addition, there is another cognate Eastern Old Japanese form -k-aku < *-ke-aku that represents contraction and not a monophthongization. It is also attested only in one example: 夜麻邊能之牙可久尓伊毛呂乎多弖天左祢度波良布母

yama-m-BE-nǝ siŋge-k-aku n-i imo-rǝ-wo tate-te sa-ne-ⁿ-do parap-umǝ mountain-GEN-side-GEN thick-ATTR-NML DV-CONV beloved-DIM-ACC make stand(CONV)-SUB PREF-sleep(NML)-DV(ATTR)-place clean-EXCL as the mountain (side) is overgrown [with bush], [I] let my beloved stand, and I am clearing a place to sleep [for us]! (MYS 14.3489)

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

421

Adjectives

2.3.2 Evidential Forms -ke- and -kere There are two evidential forms in Western Old Japanese: -ke- and -kere, with no apparent difference in function. The origin of the latter form is quite transparent: it is a contraction of the attributive -ki and the evidential form ar-e of the verb ar- ‘to exist.’ The derivation of the former is not absolutely transparent, but in all likelihood it represents the contraction of -kere to -ke- with -rloss. Only the form -kere survived into Middle Japanese. The form -ke- is never found in the word-final position: it is always followed either by the concessive converb -ⁿdǝ[mǝ] or the conjunctive converb -mba. On the contrary, the longer form -kere occurs in one example as word-final (see the example from MYS 18.4118 below). Examples: 久良波斯夜麻波佐賀斯祁杼伊毛登能煩禮波佐賀斯玖母阿良受

Kurapasi yama pa saŋgasi-ke-ⁿdǝ imo-tǝ nǝmbǝr-e-mba saŋgasi-ku mǝ ar-aⁿz-u Kurapasi mountain TOP be.steep-EV-CONC beloved-COM climb-EV-CON be.steep-CONV FP exist-NEG-FIN although Mount Kurapasi is steep, when I climb [it] with [my] beloved, [it] is not steep at all (KK 70) 于泥備椰摩虚多智于須家苔多能彌介茂

Unembï yama kǝ-tat-i usu-ke-ⁿdǝ tanǝm-i kamo Unembï mountain tree-stand-NML be.thin-EV-CONC rely-NML EP Although the lines of trees on Mount Unembï are thin, are [they not] reliable, I wonder? (NK 105) 挂文由遊志計礼杼母言久母綾尓畏伎

KAKƐ-M-AKU mo yuyusi-kere-ⁿdǝmǝ IP-AM-Aku mǝ AYA n-i KASIKO-ki think-TENT-NML FP be.reserved-EV-CONC say-TENT-NML FP extreme DV-CONV be.awesome-ATTR Although [it] is unthinkable even to think [about it], and to say [it], too, is extremely awesome … (MYS 2.199a) 意乃何身志伊多波斯計礼婆

onǝ-ŋga MÏ si itapasi-kere-mba myself-POSS body EP be.ill-EV-CON as my [own] body became ill (MYS 5.886)

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

422

Chapter 5

吾戀不止本之繁家波

WA-ŋGA KOPÏ YAM-AⁿZ-U MƏTƏ-NƏ SI ŋGƐ-ke-mba I-POSS love stop-NEG-FIN root-GEN be.thick-EV-CON My love will not stop, because [its] roots are growing densely (MYS 10.1910) 美知能等保家婆間使毛遣縁毛奈美

miti-nǝ tǝpo-ke-mba MA-ⁿ-DUKAPI mo YAR-U YƏSI mo na-mi way-GEN be.far-EV-CON interval-GEN-messenger send-ATTR chance FP no-GER since there is not even a chance to send messengers between [us] because the way is far (MYS 17.3969) 安比見流毛乃乎須久奈久母年月經礼波古非之家礼夜母

api-mi-ru monǝwo sukuna-ku mǝ TƏSI TUKÏ P-Ure-mba kopïsi-kere ya mǝ REC-see-ATTR CONJ be.few-CONV FP year month pass-EV-CON miss-EV IP EP although [we] see each other, as the time goes by, do [we still] miss [each other] just a bit?! (MYS 18.4118) 2.3.3 Tentative Form -k-emThe tentative form -k-em- represents a rather transparent contraction of the attributive -ki and tentative -am-. Since the tentative -am- follows stems of verbs, and not their attributive forms, it naturally leads to the hypothesis that unless this usage of the tentative -am- after the attributive -ki is an innovation, the tentative -am- itself must go back to some kind of auxiliary. However, since there are no Eastern Old Japanese examples supporting this form, the solution that takes it as an exclusive Western Old Japanese innovation is more viable. 奈爾能都底擧騰多柂尼之曳鶏武

nani n-ǝ tute-kǝtǝ taⁿda n-i si ye-k-em-u what DV-ATTR report(CONV)-word direct DV-CONV EP be.good-ATTR-TENT-FIN What message [do you have]? [It] would be better [to say it] directly (NK 128) 和可礼奈波宇良我奈之家武

wakare-n-amba ura-ŋ-ganasi-k-em-u part(CONV)-PERF-COND heart-LOC-sad-ATTR-TENT-FIN If [we] part, [I] would be sad in [my] heart … (MYS 15.3584)

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

423

Adjectives 将若異子等丹所詈金目八

WAKA-k-eM-U KO-RA-ni NOR-AYE-kane-m-ɛ ya young-ATTR-TENT-ATTR girl-PLUR-LOC abuse-PASS(CONV)-NEG/ POT-TENT-EV IP would [you] be able not to be abused by girls who would be younger? [Certainly you would be abused!] (MYS 16.3793) 保等登芸須奈可牟佐都奇波佐夫之家牟可母

potǝtǝŋgisu nak-am-u sa-tukï pa sambusi-k-em-u kamǝ cuckoo cry-TENT-ATTR fifth lunar month TOP be.lonely-ATTR-TENT-ATTR EP [in] the fifth lunar month when the cuckoo will cry, [I] would be lonely! (MYS 17.3996) 故布流比於保家牟

kop-uru pi opo-k-em-u long.for-ATTR day be.many-ATTR-TENT-FIN there would be many days when [I] long for [you] (MYS 17.3999) 孤悲之家久氣乃奈我家牟

kopïsi-k-eku kɛ naŋga-k-em-u long.for-ATTR-NML day be.long-ATTR-TENT-FIN the days of longing for [you] would be long (MYS 17.4006) 2.3.4 Conditional Form -k-emba Care must be taken to differentiate between the evidential form -ke followed by the conjunctive converb -mba and the conditional form -k-emba that is obviously derived from a contraction of the attributive -ki and the conditional converb -amba. Since the conditional converb -amba follows stems of verbs, and not their attributive forms, it naturally leads to the hypothesis that unless this usage of the conditional converb -amba after the attributive -ki is an innovation, the conditional converb -amba itself must go back to some kind of auxiliary. There is just one example of -k-emba in Eastern Old Japanese, which supports the auxiliary hypothesis, especially that there are only two examples of the conditional form -k-emba in Western Old Japanese texts and also the special EOJ -k-amba. WOJ examples:

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

424

Chapter 5

許智多鶏波乎婆頭勢夜麻能伊波帰爾母為弖許母郎奈牟

kǝt[ǝ]-ita-k-emba wom-Batuse-yama-nǝ ipa kï-ni mǝ wi-te kǝmǝr-ana-m-u rumor-painful-ATTR-COND DIM-Patuse-mountain rock fortress-LOC FP lead(CONV)-SUB hide-DES-TENT-FIN if rumors are painful, [I] want to take [you] along to a rocky fortress on the Small Patuse mountain and hide away (FK 1) 戀之家婆形見尓将為

KOPÏsi-k-emba KATAMI n-i SE-M-U long.for-ATTR-COND farewell present DV-CONV do-TENT-FIN If [you] long for [me], [I] want to make a farewell present [for you] (MYS 8.1471) Comparative Data Level A: Other Japonic A1: Eastern Old Japanese There is only one example of the conditional form -k-emba in Eastern Old Japanese. 古非思家波素弖毛布良武乎

kopïsi-k-emba soⁿde mo pur-am-u-wo miss-ATTR-COND sleeve FP wave-TENT-ATTR-ACC If [you] miss [me], [I] will wave my sleeves, but … (MYS 14.3376) There is, however, the Eastern Old Japanese conditional form -k-amba that represents a contraction -si has taken place after adjectival and verbal roots that end in a vowel, but the original shape -asi was preserved after the consonantal roots. This can be further confirmed by such an adjectival form as kopï-si ‘to be longed for,’ clearly derived from the vowel verb kopï- ‘to love, to long for.’ The process of deriving adjectives from verbs with -asi is not productive in Western Old Japanese, as -asi is only found after a handful of verbs. It can be added either to a consonantal verb root, or it can follow the iterative suffix -ap-. In certain cases the lexicalization has occurred, for example see below ikiⁿduk-asi ‘lamentable,’ which is derived from ikiⁿduk‘to breathe [hard].’ After the verb omǝp- ‘to think, to love,’ the suffix -asi is assimilated to -osi. Also, in addition to the deverbal adjective kopï-si mentioned above, there is a different form kop-osi, derived with a deletion of a final vowel of the vowel verb kopï- ‘to love, to long for’ with a subsequent assimilation of -asi to -osi.19 母智騰利乃可可良波志母与

mǝti-ⁿ-dǝri-nǝ kakar-ap-asi-mǝ yǝ mochi-GEN-bird-COMP be.stuck-ITER-ADJ-EXCL EP [You] are stuck like a bird on a mochi [trap-stick]! (MYS 5.800) 毛々等利能己惠能古保志枳波流岐多流良斯

momo tǝri-nǝ kǝwe kop-osi-ki paru k-i-tar-urasi hundred bird-GEN voice-GEN long.for-ADJ-ATTR spring come-CONV-PERF/ PROG-SUP [It] seems that the spring with voices of a hundred birds that [I] missed has arrived (MYS 5.834)

19  Historically we, of course, have two different assimilations in omǝp-osi and kop-osi, since the first has the root vowel /ǝ/, and the second the root vowel /o/. But since in all texts except the Kojiki kayō the kō-rui vowel /o/ and the otsu-rui vowel /ǝ/ merged after /p/, the forms attested in the texts both have neutralized /o/. Thus, the pre-Western Old Japanese forms were *ǝmǝp-ǝsi and *kop-osi respectively.

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

433

Adjectives 意乃何身志伊多波斯計礼婆

onǝ-ŋga MÏ si itap-asi-kere-mba self-POSS body EP be.painful-ADJ-EV-CON because [my] own body was in pain … (MYS 5.886) 世間乎宇之等夜佐之等於母倍杼母

YƏ-NƏ NAKA-wo u-si tǝ yas-asi tǝ omǝp-ɛ-ⁿdǝmǝ world-GEN middle-ABS be.sad-FIN DV get.emaciated-ADJ DV think-EV-CONC Although [I] think that [this] life is sad and poor … (MYS 5.893) 空氣衝之相別去者

ANA IKIⁿDUK-Asi API-WAKARE-N-Amba INTER lament-ADJ(FIN) REC-part(CONV)-PERF-COND Oh, how lamentable! If [we] part with each other … (MYS 8.1454) 従君毛吾曽益而伊布可思美為也

KIMI-YORI mo ARE sǝ MASAR-I-TE imbuk-asi-mi S-URU lord-COMP FP I FP increase-CONV-SUB feel.uneasy-ADJ-GER do-ATTR I feel more uneasy than [my] lord (MYS 12.3106) 安奈伊伎豆加思美受比佐尓指天

ana ikiⁿduk-asi mi-ⁿz-u pisa n-i s-i-te INTER lament-ADJ(FIN) see-NEG-CONV long DV-CONV do-CONV-SUB Oh, how lamentable! It has been a long [time] without seeing [you] (MYS 14.3547) Although this poem is found in book 14, there are no Eastern Old Japanese features in it, so I treat it as a Western Old Japanese text. 於母保之伎許登都氐夜良受

omǝp-osi-ki kǝtǝ tute-yar-aⁿz-u think-ADJ-ATTR word send.a.message(CONV)-send-NEG-FIN [I] do not send a message that [I] think of (MYS 17.3962) 佐由利能波奈能惠麻波之伎香母

sa-yuri-nǝ pana-nǝ wem-ap-asi-ki kamǝ PREF-lily-GEN flower-GEN smile-ITER-ADJ-ATTR EP the lily flowers are smile-provoking! (MYS 18.4086)

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

434

Chapter 5

After an otsu-rui vowel /ǝ/ in the root adjectivizer -asi assimilates to -ǝsi (with /ǝ/ neutralized to /o/ in the environments where the contrast between /ǝ/ and /o/ is lost): 伊豫国与利白祥鹿乎献奉天在礼方有礼志与呂許保志止奈毛見流

Iyǝ-NƏ KUNI-yǝri SIRO-KI SIRUSI N-Ə SIKA-wo TATEMATUR-I-te Ar-e-mba uresi yǝrǝkǝmb-osi tǝ namo MI-ru Iyǝ-GEN province-ABL white-ATTR mark DV-ATTR deer-ACC present(HUM)CONV-SUB exist-EV-CON be.glad rejoice-ADJ DV FP see-ATTR when [they] presented [us] with a deer with white marks from the province of Iyǝ, [we] regarded this as a joyful and auspicious [event] (SM 46) Comparative Data Level A: Other Japonic A1: Eastern Old Japanese There is only one example of -asi in Eastern Old Japanese: 比古布祢乃斯利比可志母與

pik-o pune-nǝ siri pik-asi-mǝ yǝ pull-ATTR boat-COMP buttocks pull-ADJ-EXCL EP [I] am in the mood of pulling the buttocks [of my husband] like a tow boat! (MYS 14.3431)

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

Section 3

Defective Adjectives There are two defective adjectives in Western Old Japanese: ka- ‘to be so, thus’ and sa ‘id.’ 3.1

Defective Adjective ka-

The defective adjective *ka- ‘to be so, thus’ is attested in two forms: the converb form ka-ku and the subordinative gerund form ka-ku-te. The latter form appears in phonographic writing only once in the Senmyō. In most cases the converb ka-ku has an adverbial usage, modifying a following verb, but on some occasions it can also be used as a quasi-predicate with the following desiderative particle mǝŋga~ mǝŋgamǝ. The converb ka-ku can also be nominalized when it is followed by the defective verb n- ‘to be.’ 和賀美斯古良迦久母賀

wa-ŋga mi-si ko-ra ka-ku mǝŋga I-POSS see(CONV)-PAST/ATTR girl-DIM thus-CONV DP [I] desire this much the girl I saw (KK 42) 加久能碁登那爾淤波牟登

ka-ku-nǝ ŋgǝtǝ na-ni op-am-u tǝ thus-CONV-GEN like name-LOC carry-TENT-FIN DV in order to perpetuate that [it] was like that (KK 97) 豫呂豆余珥訶勾志茂餓茂知余珥茂訶勾志茂餓茂

Yǝrǝⁿdu yǝ-ni ka-ku si moŋgamo ti yǝ-ni mo ka-ku si moŋgamo ten.thousand year-LOC thus-CONV EP DP thousand year-LOC FP thus-CONV EP DP [I] wish [that my sovereign] will be [like] that in ten thousand years, will be [like] that in a thousand years, too (NK 102)

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

436

Chapter 5

企許斯遠周久爾能麻保良叙可爾迦久爾保志伎麻爾麻爾斯可爾波阿羅慈迦

kikǝs-i-wos-u kuni-nǝ ma-po-ra ⁿzǝ ka n-i ka-ku n-i posi-ki manima n-i sika n-i pa ar-aⁿzi ka rule(HON)-CONV-HON-ATTR country-GEN INT-top-LOC FP thus DV-CONV thus-CONV DV-CONV desire-ATTR according DV-CONV thus DV-CONV TOP exist-NEG/TENT IP in the highest place of the country, where [the emperor] rules, [it] would not be thus according to what [you] wish in this way and that way, [would it]? (MYS 5.800) 可由既婆比等爾伊等波延可久由既婆比等爾迩久麻延

ka yuk-ɛ-mba pitǝ-ni itǝp-aye ka-ku yuk-ɛ-mba pitǝ-ni nikum-aye that go-EV-CON person-DAT detest-PASS(CONV) thus-CONV go-EV-CON person-DAT hate-PASS(CONV) when [they] go that [way], [they] are detested by people, and when [they] go this way, [they] are hated by people … (MYS 5.804) 等伎波奈周迦久斯母何母等意母閇騰母

tǝk[ǝ]-ipa-nasu ka-ku si mǝŋgamǝ tǝ omǝp-ɛ-ⁿdǝmǝ eternal-rock-COMP thus-CONV EP DP DV think-EV-CONC Although [I] think that [I] would like to be (thus) like an eternal rock … (MYS 5.805) 故之能吉美良等可久之許曾楊奈疑可豆良枳多努之久安蘇婆米

Kosi-nǝ kimi-ra-tǝ ka-ku si kǝsǝ YAnaŋgï kaⁿdurak-i tanosi-ku asomb-am-ɛ Kosi-GEN lord-PLUR-COM thus-CONV EP FP willow wear.as.a.wig-CONV pleasant-CONV amuse-TENT-EV [I] will thus amuse [myself] with lords from Kosi by putting willow [branches] in our hair in this way (MYS 18.4071) 日継波加久弖絶奈牟止為

PI TU ŋG-I pa ka-ku-te TAYE-n-am-u tǝ S-U sun follow-NML TOP thus-CONV-SUB be.interrupted(CONV)-PERF-TENT-FIN DV do-FIN the inheritance of the Sun is going to be interrupted in this way (SM 27)

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

437

Adjectives

Comparative Data Level A: Other Japonic A1: Eastern Old Japanese The converb ka-ku is attested several times in Eastern Old Japanese, but only in book 14. All examples represent an adverbial usage. 可久太尓毛久尓乃登保可婆奈我目保里勢牟

ka-ku ⁿdani mo kuni-nǝ tǝpo-k-amba na-ŋga MƐ por-i se-m-u thus-CONV RP FP province-GEN far-ATTR-COND you-POSS eye want-NML do-TENT-FIN if the [home] province is just so far, [I] want [to see] your eyes (MYS 14.3383) 古奈宜可久古非牟等夜

ko-naŋgï ka-ku kopï-m-u tǝ ya DIM-water.hollyhock thus-CONV long-TENT-FIN DV IP Do [you] think [I] will long so [strongly] for a small water-hollyhock? (MYS 14.3415) 安豆左由美須惠尓多麻末吉可久須酒曽

aⁿdusa yumi suwe-ni tama mak-i ka-ku s-u s-u sǝ catalpa bow end-LOC jewel wrap-NML thus-CONV do-FIN do-FIN FP [I] wrapped the ends of the catalpa bow with jewels in this way (MYS 14.3487) 3.2

Defective Adjective sa

The defective adjective sa ‘to be so, such’ possibly appears in two forms: the root form sa and the subordinative gerund sa-te, although the evidence for the latter is mostly logographic, and, therefore, quite slim. Examples: 情佐麻祢之

KƏKƏRƏ sa mane-si thought so many-FIN so many [sad] thoughts (MYS 1.82) 比等里佐奴礼婆

pitǝ-ri sa n-ure-mba one-CL thus sleep-EV-CON when I thus sleep alone (MYS 15.3626)

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

438

Chapter 5

佐夜麻太乃乎治我其日爾母等米安波受家牟

sa Yamaⁿda n-ǝ woⁿdi-ŋga SƏNƏ PI-ni mǝtǝmɛ ap-aⁿz-u-k-em-u so Yamaⁿda DV-ATTR old man-POSS that day-LOC search(CONV) meet-NEG-CONV-PAST/FIN-TENT-FIN So, old man Yamaⁿda searched for [him] on that day, but did not find [him] (MYS 17.4014) 住吉之岸乎田尓墾蒔稲乃而及苅不相公鴨

SUMINƏYE-NƏ KISI-wo ta-ni PAR-I MAK-I-SI INE SA-TE KAR-U-MAⁿDE-NI AP-AN-U KIMI kamo Suminǝye-GEN shore-ACC paddy-LOC make-CONV sow-CONV-PAST/ATTR rice be.thus(CONV)-SUB cut-ATTR-TERM-LOC meet-NEG lord EP Alas, [I] will not meet with [my] lord until [they] turn the shores of Suminǝye into paddies and thus cut the rice that [they] sowed! (MYS 10.2244) 雪寒三咲者不開梅花縦比來者然而毛有金

YUKI SAMU-mi SAK-I N-I pa SAK-AⁿZ-U UMƐ-NƏ PANA YƏ-SI KƏNƏ KƏRƏ pa SA-TE mo AR-U ŋgane snow cold-GER bloom-NML DV-CONV TOP bloom-NEG-CONV plum-GEN blossom good-FIN this time TOP be.thus(CONV)-GER FP exist-ATTR CONJ [It] is good that plum blossoms do not even bloom because the snow is cold. So that [it] will be so at the appropriate time (MYS 10.2329) 然弖己丑年稲目大臣薧已後

SA-te KƏ-NƏ USI TƏSI Inamɛ OPOMAPETUKIMI MAKAR-I-TAMAP-I-SI NƏTI be.thus(CONV)-SUB Kǝ-GEN ox year Inamɛ minister pass.away-CONV-HONCONV-PAST/ATTR after Thus, after the minister Inamɛ passed away in the year Kǝ-nǝ usi of the sexagenary cycle … (GGJEG)20 20  Cited according to Omodaka et al. 1967: 333.

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

chapter 6 Verbs



Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

Contents of Chapter 6 Chapter 6 Verbs 441 1 Verbal Grammatical Categories 446 1.1 Mode 446 1.2 Aspect 446 1.3 Tense 447 1.4 Mood 447 1.5 Voice 447 1.6 Retrospective 447 1.7 Reported Action 447 1.8 Iterative 448 1.9 Predication 448 1.10 Honorification 448 1.11 Politeness 448 2 Verbal Classes  449 2.1 Consonant Verbs 450 2.2 Vowel Verbs 451 2.3 Irregular Verbs 453 2.3.1 Strong Vowel Verbs 453 2.3.2 Verb kǝ- ‘To Come’ 454 2.3.3 Verb se- ~ -sǝ ‘To Do’ 455 2.3.4 R-irregular Verbs 456 2.3.5 N-irregular Verbs 457 2.4 Defective Verbs  458 2.4.1 Defective Verb n- ‘To Be’ 459 2.4.1.1 Converb Form n-i 459 2.4.1.1.1 Special Compressed Form -n- 466 2.4.1.2 Attributive Form n-ǝ 468 2.4.1.2.1 Special Compressed Form -n- 472 2.4.1.3 Subordinative Converb Form n-i-te 474 2.4.1.4 Special Form nar- < n-i ar- 475 2.4.1.4.1 Non-contracted Form n-i ar- 477 2.4.2 Defective Verb tǝ ‘To Be’ 483 2.4.2.1 Converb Form tǝ 483 2.4.2.2 Attributive Form t-u 490 2.4.3 Defective Verb rǝ ‘To Be’ 492

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

442

Contents of Chapter 6

2.4.4

Defective Verb tǝ ‘To Say’ 494 2.4.4.1 Converb Form tǝ 494 2.4.4.2 Final Form tǝ 498 2.4.4.3 Subordinative Converb Form tǝ-te 502 3 Verbal Affixes 504 3.1 Verbal Prefixes 504 3.1.1 Prefix i- 505 3.1.2 Prefix na- and Circumfix na-…-sǝ 512 3.1.3 Prefix ka- 515 3.1.4 Prefix ta- 517 3.1.5 Traditional Prefix sǝ- 520 3.1.6 Prefix ari- 520 3.1.7 Prefix uti- 522 3.1.8 Prefix kaki- 526 3.1.9 Prefix api- 530 3.1.10 Prefix e- 533 3.2 Verbal Suffixes  534 3.2.1 Sentence-Final Verbal Suffixes 535 3.2.1.1 Final Predication Suffix -u ~ -i 535 3.2.1.2 Attributive -uru ~ -u ~ -ru ~ -ǝ 550 3.2.1.2.1 Attributive as a Modifier 551 3.2.1.2.2 Attributive as a Nominalized Form 553 3.2.1.2.3 Attributive as a Final Predicate 555 3.2.1.2.4 Attributive as a Final Predicate without kakari-musubi 561 3.2.1.3 Evidential -ure ~ -e ~ -re 572 3.2.1.3.1 Special Constructions -(a)m-ɛ ya and -(u) ram-ɛ ya  579 3.2.1.4 Imperative -e ~ -ǝ 583 3.2.1.5 Zero Imperative and Its Extended Form -yǝ 590 3.2.1.6 Negative Imperative -una 595 3.2.1.7 Desiderative -ana ~ -na 599 3.2.1.7.1 Special Contracted Form -an- ~ -n- 602 3.2.1.8 Subjunctive -amasi ~ -masi 606 3.2.1.9 Suppositional -urasi ~ -asi 611 3.2.1.9.1 Special Form -asi 614 3.2.1.10 Negative Tentative -aⁿzi ~ -ⁿzi 617 3.2.1.11 Negative Potential -umasiⁿzi 622 3.2.1.12 Exclamative -umǝ ~ -mǝ 626

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

Contents of Chapter 6

3.2.2

3.2.3

443

Sentence-Non-Final Verbal Suffixes 630 3.2.2.1 Converbs 631 3.2.2.1.1 Converb -i 631 3.2.2.1.1.1 Special Construction -i-wor- 641 3.2.2.1.2 Converb -u 645 3.2.2.1.2.1 Special Construction -(a)ⁿz-u pa 649 3.2.2.1.2.2 Special Construction -(a)ⁿz-u ar- 650 3.2.2.1.2.3 Special Contracted Form -(a)ⁿz-ar- 652 3.2.2.1.3 Conditional Converb -amba ~ -mba 654 3.2.2.1.4 Conjunctive Converb -mba 663 3.2.2.1.5 Concessive Converb -ⁿdǝ[mǝ] 672 3.2.2.2 Nominalizers 679 3.2.2.2.1 Nominalizer -i 679 3.2.2.2.1.1 Special Uses of Nominalized Verbs 682 3.2.2.2.2 Nominalizer -u 685 3.2.2.2.3 Nominalizer -aku ~ -ku 687 3.2.2.2.3.1 Special Construction -(a)ku n-i  692 3.2.2.2.3.2 Special Constructions -(a) m- aku por- and -(a)m-aku posi 694 3.2.2.2.4 Nominalizer -usa 698 Word-Non-Final Verbal Suffixes 701 3.2.3.1 Negative -an- ~ -aⁿz- ~ -n- ~ -nz- 701 3.2.3.2 Tentative -am- ~ -m- 713 3.2.3.2.1 Special Form -kem- 723 3.2.3.2.2 Special Construction -(a)m-u tǝ s-u 725 3.2.3.3 Tentative -uram- ~ -ram- 730 3.2.3.4 Iterative -ap- ~ -ǝp- 738 3.2.3.5 Passive -aye- ~ -raye- ~ -ye- 745 3.2.3.6 Passive -are- 755 3.2.3.7 Honorific -as- 758 3.2.3.8 Causative -asimɛ- ~ -simɛ- 771 3.2.3.9 Causative -as- ~ -(a)se- 776 3.2.3.10 Debitive -umbɛ- ~ -mbɛ- 784 3.2.3.11 Progressive -er- 792 Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

444

Contents of Chapter 6

4 Auxiliaries  804 4.1 Bound Auxiliaries 804 4.1.1 Word-Final Bound Auxiliaries 804 4.1.1.1 Subordinative Converb -te 804 4.1.1.1.1 Special Construction -te pa 812 4.1.1.1.2 Special Construction -te mǝ 813 4.1.1.1.3 Special Construction -te wor- 815 4.1.1.1.4 Special Construction -te ok- 815 4.1.1.2 Coordinative Converb -tutu 819 4.1.1.3 Coordinative Converb -naŋgara 825 4.1.1.4 Coordinative Converb -katera 826 4.1.1.5 Past -ki ~ -si ~ -sika 827 4.1.2 Word-Non-Final Bound Auxiliaries  844 4.1.2.1 Perfective -n- 844 4.1.2.2 Perfective -te- 855 4.1.2.3 Perfective-Progressive -tar- and Its Uncontracted Form -te ar- 866 4.1.2.3.1 Uncontracted Form -te ar- 872 4.1.2.4 Retrospective -ker- 879 4.1.2.5 Potential -kate- ~ -ŋgate- 888 4.1.2.6 Negative Potential -kane- 893 4.1.2.7 Benefactive -kǝse- 899 4.2 Lexical Auxiliaries 903 4.2.1 Honorific and Humble Auxiliaries 903 4.2.1.1 Honorific Auxiliaries 903 4.2.1.1.1 Honorific Auxiliary tamap- 903 4.2.1.1.1.1 Contracted Form tamb- 908 4.2.1.1.2 Honorific Auxiliary imas- 910 4.2.1.1.3 Honorific Auxiliary wos- 916 4.2.1.1.4 Honorific Auxiliary kikǝs- 918 4.2.1.1.5 Honorific Auxiliary mes- 921 4.2.1.1.6 Honorific Auxiliary sirasimes- 924 4.2.1.1.7 Honorific Auxiliary nǝtamap- ~ nǝritamb~ nǝtamb- 926 4.2.1.1.7.1 Contracted Form nǝritamb- 926 4.2.1.1.7.2 Contracted Form nǝtamb- 927 4.2.1.1.7.3 Contracted Form nǝtamap- 927

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

Contents of Chapter 6

4.2.2

445

4.2.1.2 Humble Auxiliaries 929 4.2.1.2.1 Humble Auxiliary matur- 929 4.2.1.2.2 Humble Auxiliary tatematur- 933 4.2.1.2.3 Humble Auxiliary tamapɛ- 935 4.2.1.2.4 Humble Auxiliary tamapar- ~ tambar- 938 4.2.1.2.5 Humble Auxiliary mawos- ~ mawus- 940 4.2.1.2.6 Humble Auxiliary mawi- 944 Other Auxiliaries 946 4.2.2.1 Reported Action Auxiliary nar- 946 4.2.2.2 Cooperative Auxiliary ap- 951 4.2.2.3 Auxiliary ar- 953 4.2.2.4 Auxiliary wor- 958 4.2.2.5 Directive Auxiliaries 962 4.2.2.5.1 Directive Auxiliary kǝ- 962 4.2.2.5.2 Directive Auxiliary yuk- 966 4.2.2.5.3 Directive Auxiliaries iⁿde- and iⁿdas- 970 4.2.2.5.4 Directive Auxiliaries ir- and ire- 973 4.2.2.5.5 Directive Auxiliaries aŋgɛ- and aŋgar- 975 4.2.2.5.6 Directive Auxiliary yar- 977 4.2.2.5.7 Directive Auxiliaries watar- and watas- 979 4.2.2.5.8 Directive Auxiliary yǝr- 983 4.2.2.5.9 Directive Auxiliary kakɛ- 984 4.2.2.5.10 Directive Auxiliaries tuk- and tukɛ- 985 4.2.2.6 Resultative Auxiliary ok- 988 4.2.2.7 Auxiliary -ŋ-kata- 990 4.2.2.8 Potential Auxiliary e- 991

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

Section 1

Verbal Grammatical Categories Before a detailed discussion of the Western Old Japanese verbal morphology, I would like to briefly outline the major grammatical categories of the verb in Western Old Japanese. Since this book is built on the principle ‘from the marker to the category’ rather than vice versa, I consider it necessary to crossreference verbal markers and the categories they express at the very beginning, so that a reader who is interested in a description of a certain category in Western Old Japanese may easily find the necessary section using the Index of grammatical markers or the Contents. I list the following categories together with their markers. 1.1

Mode

There are two modes in Western Old Japanese: affirmative and negative. The affirmative mode does not have a special marker. The negative mode is marked by the suffixes -an- ~ -n-, -aⁿz- ~ -ⁿz-, in the indicative mood and by the suffix -una, the prefix na-, and the circumfix na-…-sǝ in the imperative mood. There are special negative mode forms for the negative tentative (-aⁿzi ~ -ⁿzi), and negative potential mood (-umasiⁿzi). 1.2

Aspect

There are five aspects in Western Old Japanese: imperfective with a zero marker, perfective marked by auxiliaries -n- and -te-, progressive marked by the suffix -er-,1 perfective progressive marked by the analytical construction subordinative converb -te + auxiliary ar- ‘to exist,’ the auxiliary -tar- (that represents a contraction of the -te + ar- construction), and continuous (analytical, converb + auxiliary wor- ‘to exist’).

1  Historically -er- comes from a converb -i followed by the auxiliary verb ar- ‘exist.’

© Alexander Vovin, 2020 | doi:10.1163/9789004422810_007

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

447

Verbs

1.3

Tense

There are two tenses in Western Old Japanese: present-future, with a zero marker, and past, marked by the auxiliary -ki (that also has other suppletive forms). 1.4

Mood

Mood is the pride and beauty of Western Old Japanese. There are fourteen moods in the language: indicative (no marker), imperative (no marker or suffixes -e and -yǝ), tentative (suffix -am- ~ -m-), negative tentative (suffix -aⁿzi ~ -ⁿzi), second tentative (suffix -uram- ~ -ram-), debitive (suffix -umbɛ-), potential (auxiliary -kate-), negative potential (suffix -umasi ⁿzi-, auxiliary kane-), subjunctive (suffix -amasi ~ -masi), suppositional (suffix -urasi), optative (analytical, with the verb por- ‘to want’ or the adjective posi ‘to be wanted’), desiderative (suffixes -ana ~ -na), benefactive (auxiliary kǝse-), and exclamative (suffix -umǝ ~ -mǝ). 1.5

Voice

There are four voices in Western Old Japanese: active (no marker), passive (suffixes -aye- ~ -ye- and -are-), causative (suffixes -ase- ~ -se- and -asimɛ- ~ -simɛ-), and reciprocal-cooperative (prefix api- or analytical, converb + ap- ‘to combine, to meet’). 1.6

Retrospective

The retrospective form in Western Old Japanese is rendered by the auxiliary -ker-. 1.7

Reported Action

The reported action form in Western Old Japanese is rendered by the auxiliary nar-.

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

448 1.8

Chapter 6

Iterative

The iterative form in Western Old Japanese is rendered by the suffix -ap- ~ -ǝpor by the prefix ari-. 1.9

Predication

There is a sharp distinction between the forms of final and non-final predication. The former include the final predicative form (suffixes -u ~ -i), the attributive form (suffixes -uru ~ -ru ~ -u), the evidential form (suffixes -ure ~ -re ~ -ɛ), and others. The latter are represented by the converb (suffixes -i ~ -u), the subordinative converb (auxiliary -te), the coordinative converbs (auxiliaries -tutu and -naŋgara), the coordinative converb (auxiliary -katera), the conditional converb (suffix -amba ~ -mba), the conjunctive converb (suffix -mba), and the concessive converb (suffix -ⁿdǝ[mǝ]). 1.10

Honorification

There are three basic degrees of honorification in Western Old Japanese. A verbal form may be honorific, humble, or neutral. There are several subdegrees of honorific and humble expressions within each of these two subsections. Honorification is expressed by a number of auxiliary verbs that are either honorific or humble. They follow the converb of the main verb. 1.11

Politeness

In contrast to Middle (Classical) Japanese, there is no category of politeness in Western Old Japanese.

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

Section 2

Verbal Classes The majority of Western Old Japanese verbs fall into two major classes: consonant verbs, with a root ending in a consonant, and vowel verbs, with a root ending in the vowels /ɛ/ or /ï/. In traditional Japanese terminology, consonant verbs are called quadrigrade (四段 yodan), vowel verbs ending in /ɛ/ are classified as lower bi-grade (下二段 simonidan), and vowel verbs ending in /ï/ as upper bi-grade (上二段 kaminidan). Following Frellesvig’s proposal, I now think that traditional ‘regular’ upper mono-grade verbs (上一段 kamiichidan) ending in the high vowels /i/ or /ï/ are better classified as an irregular class (Frellesvig 2007: 224–225). In addition, there are four other irregular classes of verbs, known in Japanese tradition as (か変 ka-hen, さ変 sa-hen, ら変 ra-hen, and な変 na-hen). Finally, there are several defective verbs that do not have full paradigms, and they are not recognized in the traditional Japanese grammar. It is necessary to keep in mind that a number of Western Old Japanese verbs can belong to a different verbal class vis-à-vis the same verbs in Middle (Classical) Japanese. Thus, cf. WOJ kakur- ~ kakure- ‘to hide’ (consonant or vowel verb) vs. MJ kakure- ‘id.’ (vowel verb), WOJ wasur- ~ wasure- ‘to forget’ (consonant or vowel verb) vs. MJ wasure- ‘id.’ (vowel verb), WOJ pur- ~ pure- ‘to touch’ (consonant or vowel verb) vs. MJ fure- ‘id.’ (vowel verb),2 WOJ ik- ‘to live’ (consonant verb) vs. Late MJ iki- ‘id.’ (vowel verb), WOJ omb- ‘to wear at the waist’ (consonant verb) vs. MJ obi- ‘id.’ (vowel verb), WOJ momit- ‘to turn red or yellow (of leaves or grass)’ (consonant verb) vs. MJ momiti- ‘id.’ (vowel verb), WOJ wak- ‘to divide’ (consonant verb) vs. MJ wak- ~ wake-‘id.’ (consonant or vowel verb), WOJ yǝrǝkǝmbï- ‘to rejoice’ (vowel verb) vs. MJ yorokob- ‘id.’ (consonant verb), WOJ kanasimbï- ‘to grieve’ (vowel verb) vs. MJ kanasib- ‘id.’ (consonant verb), WOJ yǝkï- ‘to avoid’ (vowel verb) vs. MJ yok- ~ yoki- ‘id.’ (consonant 2  Shirafuji considers the WOJ consonant verbs kakur-, pur-, and wasur- to be earlier forms, while the WOJ vowel verbs kakure-, pure, and wasure- are later forms (Shirafuji 1987: 132). It is true that the former appear to belong to Early Old Japanese, and the latter to Late Old Japanese, but we should not forget that these two varieties in all probability reflect two different dialects (Asuka and Nara) rather than two chronological stages (Early WOJ and Late WOJ), because the temporal distance between them does not exceed one century. Also, there is a semantic difference in one case: while WOJ wasur- means ‘to forget intentionally’, WOJ wasure- is ‘to forget unintentionally’.

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

450

Chapter 6

or vowel verb), WOJ mawi- ‘to come/go (HUM)’ (vowel verb) vs. MJ mawir- ‘id.’ (consonant verb), WOJ osori- (upper bigrade type vowel verb) vs. MJ osore- ‘id.’ (lower bigrade vowel verb), etc. There are also several cases when Western Old Japanese irregular verbs have regular counterparts in Middle Japanese and vice versa: WOJ isati- ‘to sob’ (irregular verb) vs. MJ isati- (upper bigrade vowel verb), WOJ arambï- ‘to be in a wild mood’ (irregular or vowel verb) vs. MJ. arabi- ‘id.’ (vowel verb), WOJ pï- ‘to dry’ (vowel verb) vs. MJ fi- (irregular verb), WOJ pï- ‘to blow nose’ (vowel verb) vs. MJ fi- ‘id.’ (irregular verb), WOJ kuwe- ‘to kick’ (vowel verb) vs. MJ ke- ‘id.’ (irregular verb). Since Western Old Japanese irregular verbs that became regular vowel verbs in Middle Japanese are all polysyllabic, and, to the contrary, all Western Old Japanese regular verbs that became irregular vowel verbs in Middle Japanese are monosyllabic, I think that here we deal with the simple case of a structural pressure: regular vowel verbs are mostly polysyllabic, while the overwhelming majority of the monosyllabic vowel verbs are irregular. Consequently, irregular polysyllabic verbs were under structural pressure to become regular, and in quite a similar way, regular monosyllabic vowel verbs were under structural pressure to become irregular. The only exception to this rule is WOJ kuwe- ‘to kick,’ which is polysyllabic and regular, but turns up as an irregular MJ ke-.3 But this unique case is easily explained as a result of a secondary, albeit not quite regular, contraction -uwe- > -e-. 2.1

Consonant Verbs

Consonant verbs are by and large the most populated verbal class in Western Old Japanese, probably comprising at least 70% of all verbs attested in the Western Old Japanese corpus. As mentioned above, all roots of consonant verbs end in a consonant. Most roots have the shape VC-, CVC-, VCVC-, or CVCVC-. Although longer shapes also do occur, historically they represent compounds. In Chart 23 below I provide the main inflectional forms of consonant verbs.

3  M J ke- ‘to kick’ constitutes by itself its own verbal class, traditional lower mono-grade (下一段 shimoichidan).

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

451

Verbs chart 23 Main inflectional forms of consonant verbs

form

top- ‘ask’

tǝr- ‘take’

tǝk- ‘untie’

converb final attributive evidential imperative negative tentative

top-i top-u top-u top-ɛ top-e top-antop-am-

tǝr-i tǝr-u tǝr-u tǝr-e tǝr-e tǝr-antǝr-am-

tǝk-i tǝk-u tǝk-u tǝk-ɛ tǝk-e tǝk-antǝk-am-

As one can see from the chart above, consonant verbs have the following peculiarities: (1) There are no morphophonemic changes with vowel-initial suffixes. (2) -r- in the attributive (-u < *-uru) and evidential (-ɛ < *-urɛ) forms is lost according to the rule of medial -r- loss (Whitman’s law, see chapter 2, section 2.5.4). This results in the lack of a differentiation between final and attributive forms on the morphological level. (3) Since there is no contrast between /i/ ~ /ï/ and /e/ ~ /ɛ/ after coronals (see chapter 2, sections 2.2.3–2.2.4), there is no phonetic contrast between evidential and imperative forms after the roots ending in coronals. 2.2

Vowel Verbs

Vowel verbs represent the second largest verbal class in Western Old Japanese, probably comprising at least 28% percent of all verbs. As mentioned above, all vowel verb roots end in the vowels /ï/ or /ɛ/. Most roots have the shape VCV- or CVCV-. Monosyllabic V- and CV- forms are very rare. The longer forms VCVCV- and CVCVCV- are not infrequent, although most of them can probably be traced to compounds or complex morphological derivations. In Chart 24 below I provide the main inflectional forms of vowel verbs.

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

452

Chapter 6

chart 24 Main inflectional forms of vowel verbs

form

okï- ‘rise’

tǝⁿdǝmɛ- ‘stop’

ɛ- ‘get’

converb final attributive evidential imperative negative tentative

okï-∅ ok-u ok-uru ok-ure okï(-yǝ) okï-nokï-m-

tǝⁿdǝmɛ-∅ tǝⁿdǝm-u tǝⁿdǝm-uru tǝⁿdǝm-ure tǝⁿdǝmɛ(-yǝ) tǝⁿdǝmɛ-ntǝⁿdǝmɛ-m-

ɛ-∅ ∅-u ∅-uru ∅-ure – ɛ-nɛ-m-

Vowel verbs have a number of peculiarities: (1) When a vowel-initial suffix is added to the root of a vowel verb, the vowel of the suffix is generally dropped, but the stem itself remains unaffected. If a suffix consists just of a single vowel, this vowel disappears altogether, like in the case of the converb -i, which is not overtly expressed. In the chart above I marked this converb as a zero morpheme -∅, but in the description itself the converbs that are not overtly expressed will not be marked and in the glosses the converb forms of vowel verbs are denoted as VERB(CONV). (2) There are exceptions to rule (1) above that concerns combinations of vowel verb roots with following final -u, attributive -uru and evidential -ure. In these cases the vowel of the suffix remains intact, but the last vowel of the verbal root is dropped. Since regular vowel verbs can lose their last vowel in a paradigmatic form, I call them “weak vowel verbs,” in contrast to irregular vowel verbs (see 2.3.1) that never lose their last vowel of the root, and which I call “strong vowel verbs.” (3) As a sequence of rule (2), monosyllabic vowel verbs of V- shape like ɛ‘to get’ in the chart above completely lose their root before the final -u, the attributive -uru and the evidential -ure, becoming in a certain way ‘empty’ roots. (4) The imperative form in -yǝ, so typical for vowel verbs in Middle Japanese texts (MJ -yo), does not occur frequently after vowel verbs in Western Old Japanese texts. This is particularly true for upper bi-grade verbs ending in the vowel /ï/. Thus, the verb okï- ‘to rise’ seems to be the only one among the upper bi-grade class that is attested in this form (twice in the same poem, MYS 16.3873). The imperative form in -yǝ is more frequent

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

453

Verbs

after lower bi-grade verbs ending in the vowel /ɛ/ (Shirafuji 1987: 141), but vowel verbs can also occur in their root forms as imperatives without any following -yǝ. 2.3

Irregular Verbs

There are five classes of irregular verbs in Western Old Japanese. Since each of them exhibits its own idiosyncrasies, it is better to discuss them one by one. All irregular verbs combined together do not exceed 2% of the total number of verbs attested in Western Old Japanese, although among them one finds some of the most high-frequency verbs. I must emphasize that below I describe only irregularities occurring in basic paradigmatic forms. Irregular verbs demonstrate a number of other irregularities found throughout the verbal system. These will be described below in the sections dedicated to relevant suffixes and auxiliaries when found in combination with irregular verbs. 2.3.1 Strong Vowel Verbs All strong vowel verbs end in high vowels /i/ or /ï/.4 The majority of them are monosyllabic with a CV- structure, like mi- ‘to see’ and mï-5 ‘to go around,’ although there are some exceptions, like WOJ isati- ‘to sob’ or arambï- ‘to be in a wild mood.’ In Chart 25 below I provide the main inflectional forms of strong vowel verbs.

4  Some Japanese linguists indicate in their grammars that strong vowel verbs always ended in the kō-rui /i/ vowel (Shirafuji 1987: 127). This is probably not true, as it becomes clear from the following chart and the following footnote, although the majority of strong vowel verbs indeed have kō-rui /i/ and not otsu-rui /ï/. 5  Most of the surviving forms of mï- ‘go around’ indicate that it is a strong vowel verb, although there is one example of the attributive form m-uru (MYS 6.942), not mï-ru, that is believed to point to a regular vowel verb paradigm (Omodaka et al. 1967: 720, 719). In addition to being just a hapax legomenon, I believe that we have here a problem of morphological segmentation, as the context in question is 許伎多武流浦 kǝŋg-i-tam-uru URA row-CONV-go. around-ATTR bay ‘bays that [we] are rowing around’ (MYS 6.942), where we actually have the verb tamï- ‘to go around,’ ‘to turn.’ Since it is not possible to segment ta- in this verb, its relationship to mï- remains unproven. Thus, I prefer to view mï- ‘to go around’ as a strong vowel verb.

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

454

Chapter 6

chart 25 Main inflectional forms of strong vowel verbs

form

mi- ‘see’

mï- ‘go around’

converb final attributive evidential imperative negative tentative

mi-∅ mi mi-ru mi-re mi(-yǝ) mi-nmi-m-

mï-∅ mï mï-ru mï-re – mï-n–

Strong vowel verbs have the following peculiarities: (1) Being essentially vowel verbs, they never lose the final vowel of the root before any vowel-initial suffixes, such as the attributive -uru or the evidential -ure. On the contrary, initial vowels of any suffixes following their roots are always deleted. Since the strong vowel verbs always preserve their root vowels, I prefer to call them ‘strong.’ (2) The final form of strong vowel verbs is identical in form to their converbs, but it is never found in isolation.6 See 3.2.1.1 for details. The replacement of the final form with the attributive form apparently had occurred only in the Middle Japanese period. (3) Similar to regular vowel verbs, the imperative form can occur in both the -yǝ form and as a root, see, e.g., mi ‘see(IMP)’ in MYS 1.27. 2.3.2 Verb kǝ- ‘To Come’ The verb kǝ- ‘to come’ has a paradigm highly reminiscent of regular vowel verbs, but it also has some differences. In Chart 26 below I provide the main inflectional forms of the verb kǝ-:

6  Cf. WOJ mi tǝmo ‘see CONJ’ ‘even if [one] sees’ (the conjunction tǝmo follows the final form of a verb).

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

455

Verbs chart 26 Main inflectional forms of kǝ- ‘to come’

form

kǝ- ‘come’

converb final attributive evidential imperative negative tentative

k-i k-u k-uru k-ure kǝ kǝ-nkǝ-m-

The verb kǝ- ‘to come’ has the following peculiarities: (1) In most cases it behaves like a regular vowel verb: it loses the final vowel of its stem before the final -u, the attributive -uru, and the evidential -ure, but retains it before other vowel-initial suffixes that lose their initial vowels upon suffixation. There is, however, one exception: in contrast to regular vowel verbs that retain their final vowels before the converb -i that is regularly deleted, kǝ- deletes its final root vowel /ǝ/ but retains the following converb -i. (2) Unlike regular vowel verbs and strong vowel verbs, kǝ- ‘to come’ never has an imperative form *kǝ-yǝ, and the root form kǝ is the only possible imperative form for this verb in Western Old Japanese.7 2.3.3 Verb se- ~ -sǝ ‘To Do’ The verb se- ~ -sǝ ‘to do’ also has a paradigm highly reminiscent of regular vowel verbs and especially of kǝ- ‘to come,’ but it has its own peculiarities as well. In Chart 27 below I provide the main inflectional forms of the verb se- ~ -sǝ:

7  Middle Japanese, in contrast, only has the imperative form ko-yo ‘come!’ and the root form is never used.

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

456

Chapter 6

chart 27 Main inflectional forms of se- ~ -sǝ ‘to do’

form

se- ~ -sǝ ‘do’

converb final attributive evidential imperative negative tentative

s-i s-u s-uru s-ure se(-yǝ), -sǝ se-nse-m-

The verb se- ~ -sǝ ‘to do’ has the following peculiarities: (1) Unlike any other verb in Western Old Japanese, it has two alternating roots: se- and -sǝ,8 where the alternation cannot be explained by morphophonological rules. The variant -sǝ occurs only as a part of the negative imperative circumfix na-V-CONV-sǝ ‘do not do V.’ The variant se- is found in most paradigmatic forms. (2) In most cases se- behaves like a regular vowel verb: it loses the final vowel of its stem before the final -u, the attributive -uru, and the evidential -ure, but retains it before other vowel-initial suffixes that lose their initial vowels upon suffixation. Similar to kǝ- ‘to come,’ there is an exception: in contrast to regular vowel verbs that retain their final vowels before the converb -i that is regularly deleted, se- deletes its final root vowel /e/ but retains the converb -i. (3) Similar to regular and strong vowel verbs, and unlike kǝ- ‘to come,’ the -yǝ imperative form is optional for se- ‘to do’: in other words, there are examples of both se and se-yǝ imperatives in Western Old Japanese. 2.3.4 R-irregular Verbs So far we have seen irregular verbs that are essentially vowel verbs that have certain idiosyncrasies in their paradigms. R-irregular verbs, on the other hand, are essentially consonant verbs with roots ending in the consonant /r/ that exhibit one irregularity. There are three primary (non-derived) r-irregular verbs in Western Old Japanese: ar- ‘to exist,’ wor- ‘to exist, to stay, to sit’ and por- ‘to want.’ The last one is usually treated as a consonant verb (Omodaka et al. 1967: 8  Historically, most likely, the root sǝ- and the extended stem se- < *sǝ-i, although this speculation may be difficult to prove.

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

457

Verbs

662), but there is some evidence that it is actually an r-irregular verb (see section 3.2.2.2.3.3 ‘Special Constructions -(a)m-aku posi and -(a)m-aku por-i’ for a detailed discussion). In Chart 28 below I provide the main inflectional forms of r-irregular verbs: chart 28 Main inflectional forms of r-irregular verbs

form

ar- ‘exist’

wor- ‘exist, stay’

converb final attributive evidential imperative negative tentative

ar-i ar-i ar-u ar-e ar-e ar-anar-am-

wor-i wor-i wor-u wor-e wor-e wor-anwor-am-

The only irregularity that distinguishes r-irregular verbs from their regular consonant counterparts is that they have the final form suffix -i, and not -u. Thus, the converb and final forms are phonologically indistinguishable. Similar to other consonant verbs ending in coronals, there is no phonological distinction between evidential and imperative forms due to the merger of /e/ and /ɛ/ after coronals. One must keep in mind that Western Old Japanese auxiliaries that historically include ar- ‘to exist’ as their last component, such as -tar-, -ker-, and -erfollow the paradigm of r-irregular verbs. The same applies to the derived verbs nar- ‘to be’ < n-i ar- and nar- ‘to be located at’ < -ni ar- (see 2.4.1.4). 2.3.5 N-irregular Verbs The irregularity of n-irregular verbs manifests itself in the fact that these verbs mix consonant and vowel verb paradigms, although their stems end in consonants. There are only two verbs in this class: sin- ‘to die’ and in- ‘to go away.’ Because there are no regular consonant verbs in Western Old Japanese that end in /n/,9 it seems likely that these two verbs originally belonged to a 9  A puzzle itself for the internal reconstruction: /n/ and /y/ and /w/ are the only consonants that do not appear as a final consonant in Western Old Japanese consonant verbs. While there might be a cogent explanation why /y/ and /w/ behave oddly in this respect, the absence of /n/ is baffling in spite of several explanations that have been offered.

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

458

Chapter 6

vowel verb paradigm but later acquired parts of a consonant verb paradigm. In Chart 29 below I provide the main inflectional forms of n-irregular verbs: chart 29 Main inflectional forms of n-irregular verbs

form

sin- ‘die’

in- ‘go away’

converb final attributive evidential imperative negative tentative

sin-i sin-u sin-uru sin-ure sin-e sin-ansin-am-

in-i in-u in-uru in-ure in-e in-anin-am-

On the basis of the chart above it is possible to see that the converb, final, imperative, negative, and tentative forms mirror those of a consonant verb paradigm, while the attributive and evidential forms look exactly like forms of a vowel verb paradigm. 2.4

Defective Verbs

All defective verbs are irregular, but in contrast to other irregular verbs they have very few paradigmatic forms. There are four defective verbs in Western Old Japanese: n- ‘to be,’ tǝ ‘to be,’ rǝ ‘to be,’ and tǝ ‘to say,’ which, as I will demonstrate below, have different origins.10 The necessity of a separate and detailed description of defective verbs is dictated by the fact that the Japanese linguistic tradition does not recognize them as verbs, treating different forms of these verbs as various particles. In addition, they have a number of irregularities not found in other irregular verbs. In Chart 30 below I provide all the existing inflectional forms of defective verbs, followed by a detailed discussion with textual examples of these forms.

10  Cf. Frellesvig’s attempt to demonstrate that n- and both tǝ lexemes ultimately have the same origin (Frellesvig 1999).

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

459

Verbs chart 30 Inflectional forms of defective verbs

form

n- ‘be’

tǝ ‘be’

rǝ ‘be’

tǝ ‘say’

converb final attributive subordinative converb -ar- form

n-i – n-ǝ n-i-te

tǝ – t-u –

– rǝ rǝ –

tǝ tǝ tǝ (?) tǝ-te

nar- < n-i ar-

–a





a Starting from the Heian period, a copula tar- < t-ar- form also appears, but it is predominantly used in kanbun kundoku texts, which are foreign in origin. It hardly ever appears in prose, and never in poetry. This finds a good explanation in the fact that, as I will show below, tǝ ‘to be’ has a foreign origin, being borrowed from Korean. The derivation of t-ar- is similar to that of n-ar-: converb tǝ + ar- ‘exist.’

Among these defective verbs, n- ‘to be’ behaves basically as a consonant verb, although it has an irregular attributive form in -ǝ, not -u. On the other hand, tǝ ‘to be’ and tǝ ‘to say’ are likely to be vowel verbs, although tǝ ‘to be’ has an attributive form in -u, like a consonant verb, and both have converbs ending in /ǝ/, which is not typical for any other vowel verb. The special form nar- is a combination of the converb n-i of n- ‘to be’ and verb ar- ‘exist’: nar- < n-i ar-. 2.4.1 Defective Verb n- ‘To Be’ As indicated in Chart 30 above, the defective verb n- ‘to be’ has four paradigmatic forms: the infinite n-i, the attributive n-ǝ, the subordinative converb n-i-te, and the special form nar- < *n-i-ar-. Below I provide a description of their functions in Western Old Japanese along with examples. 2.4.1.1 Converb Form n-i The converb form n-i ‘being’ is used in a variety of functions. It may occur after both nominals and verbs, the latter normally being in their nominalized or attributive forms.

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

460

Chapter 6

(1) The converb n-i can be used as a copula in a nominal predicate: 那許曾波遠迩伊麻世婆

na kǝsǝ pa wo n-i imas-e-mba you FP TOP man DV-CONV exist(HON)-EV-CON since you are a man (KK 5) Note that n-i imas- in this and the immediately following example represents an honorific variant of the construction n-i ar- discussed in 2.4.1.4 below. 伊可奈留夜比止爾伊麻世可

ika nar-u ya pitǝ n-i imas-e ka how be-ATTR IP person DV-CONV be(HON)-EV IP Oh, what kind of person is [he]? (BS 5) 迦賀那倍弖用迩波許許能用比迩波登袁加袁

ka-ŋga nambɛ-te yo n-i pa kǝkǝnǝ yo pi n-i pa tǝwo-ka-wo day-day line.up(CONV)-SUB night DV-CONV TOP nine night day DV-CONV TOP ten-CL-ACC counting all the days, as for nights [it is] nine nights [and] as for days [it is] ten days (KK 26) 伊毛袁斯多那岐爾

imo-wo sita nak-i n-i beloved-ABS secretly weep-NML DV-CONV [My] beloved secretly wept, and … (KK 78) 美那許袁呂許袁呂爾

mi-na kǝworǝ kǝworǝ n-i HON-water churning churning DV-CONV water churning, churning (KK 100) 汝多知方貞仁明伎心乎以天

IMASI-tati pa SAⁿDAKA n-i AKA-ki KƏKƏRƏ-wo MƏT-I-te you-PLUR TOP loyal DV-CONV bright-ATTR heart-ACC hold-CONV-SUB you, with loyal and clear hearts … (SM 37)

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

461

Verbs 清麻呂其我姉法均止甚大尓悪久奸流妄語乎作弖

KIYOMARƏ SI-ŋga ANE POPUKUN-tǝ ITƏ OPO-KI n-i ASI-ku KAⁿDAM-Er-u ITUPAR-I-ŋ-GƏTƏ-wo TUKUR-I-te Kiyomarǝ he-POSS elder.sister Popukun-COM very big-ATTR DV-CONV badCONV be.insincere-PROG-ATTR lie-NML-GEN- word-ACC make-CONV-SUB Kiyomarǝ with his elder sister Popukun created an extremely big, bad and insincere lie … (SM 44) 吾哉難二加還而将成

WARE YA nani n-i ka KAPER-I-TE NAR-AM-U I EP what DV-CONV IP return-CONV-SUB become-TENT-ATTR Why should I return? (MYS 13.3265) 伊米爾毛伊母我美延射良奈久尓

imɛ-ni mo imǝ-ŋga mi-ye-ⁿz-ar-an-aku n-i dream-LOC FP beloved-POSS see-PASS-NEG(CONV)-exist-NEG-NML DV-CONV [my] beloved [always] appears in [my] dreams (MYS 15.3735) 奈泥之故我波奈爾毛我母奈

naⁿdesiko-ŋga pana n-i moŋgamǝ na carnation-POSS flower DV-CONV DP EP [I] wish [you] were a carnation flower! (MYS 17.4010) 老人毛女童兒毛之我願心太良比爾

OYI-PITƏ mo WOMINA WARAPA mo si-ŋga NE ŋG-AP-U KƏKƏRƏ-N-tar-ap-i n-i old(NML)-person FP woman child FP they-POSS desire-ITER-ATTR heartLOC-be.enough-ITER-NML DV-CONV old people, women, and children, all [of them get] enough of what they desire to [their] hearts’ [content] (MYS 18.4094) (2) The converb n-i is frequently used after nouns and adjectives for adverbializations: 久路岐美祁斯遠麻都夫佐爾登理與曾比 … 阿遠岐美祁斯遠麻都夫佐迩登理 與曾比

kuro-ki mi-kes-i-wo ma-tumbusa n-i tǝr-i-yǝsǝp-i … awo-ki mi-kes-i-wo matumbusa n-i tǝr-i-yǝsǝp-i black-ATTR HON-wear(HON)-NML-ACC INT-without fail DV-CONV takeCONV-dress-CONV … blue-ATTR HON-wear(HON)-NML-ACC INT-without fail DV-CONV take-CONV-dress-CONV neatly wearing a black garment … neatly wearing a blue garment (KK 4) Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

462

Chapter 6

袁登賣爾多陀爾阿波牟登

wotǝme-ni taⁿda n-i ap-am-u tǝ maiden-DAT direct DV-CONV meet-TENT-FIN DV thinking to meet maidens face to face (KK 18) 麻許曾迩斗比多麻閇

ma kǝsǝ n-i top-i-tamap-ɛ truth FP DV-CONV ask-CONV-HON-EV [It] is right that [you] asked [me] (KK 72) In this example the focus particle kǝsǝ separates the noun ma ‘truth’ from the adverbializing converb n-i. 於夜那斯爾奈礼奈理鶏迷夜

oya na-si n-i nare nar-i-k-em-ɛ ya parent no-FIN DV-CONV you be.born-CONV-PAST/FIN-TENT-EV IP Were you possibly born without parents? (NK 104) Note that n-i follows not the attributive form na-ki, but the final form na-si of the adjective na- ‘not to exist’ as demonstrated by this example. 意比久留母能波毛毛久佐爾勢米余利伎多流

op-i-k-uru mǝnǝ pa momo kusa n-i semɛ-yǝr-i-k-i-tar-u pursue-CONV-come-ATTR thing TOP hundred kind DV-CONV assault(CONV)approach-CONV-come-CONV-PERF/PROG-ATTR the things that pursue [us], come assaulting [us] in a hundred varieties (MYS 5.804) 横風乃爾布敷可爾覆來礼婆

YƏKƏSIMA-KAⁿZE-nǝ nipumbuka n-i OPOP-I-K-I-TAr-e-mba cross.wind-GEN sudden DV-CONV cover-CONV-come-CONV-PERF/ PROG-EV-CON a cross wind suddenly came to cover [us] (MYS 5.904) 都可倍麻都良米伊夜等保奈我尓

tukapɛ-matur-am-ɛ iya tǝpo naŋga n-i serve(CONV)-HUM-TENT-EV plentifully long long DV-CONV [I] will serve [you] plentifully and for a long, long time (MYS 18.4098) 都麻母古騰母毛乎知己知爾左波爾可久美為

tuma mǝ ko-ⁿdǝmǝ mo woti kǝti-ni sapa n-i kakum-i wi spouse FP child-PLUR FP there here-LOC many DV-CONV surround-CONV exist(CONV) Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

463

Verbs

both my spouse and children are around [me] in great numbers here and there (MYS 20.4408) 波都由伎波知敞爾布里之家

patu yuki pa ti-pe n-i pur-i-sik-e first snow TOP thousand-CL DV-CONV fall-CONV-cover-IMP First snow, fall in a thousand layers! (MYS 20.4475) 阿止乎美都都志乃波牟多太爾阿布麻弖爾麻佐爾阿布麻弖爾

atǝ-wo mi-tutu sinǝp-am-u taⁿda n-i ap-u-maⁿde-ni masa n-i ap-u-maⁿde-ni footstep-ACC see(CONV)-COOR yearn-TENT-FIN direct DV-CONV meet-ATTRTERM-LOC real DV-CONV meet-ATTR-TERM-LOC looking at [Buddha’s] footstep, [I] will yearn [for him], until [I] meet [him] directly, until [I] really meet [him] (BS 6) (3) Another usage of n-i is after the quasi-postpositions tamɛ ‘for,’ and yuwe ‘because,’ ‘for the sake of’ (lit. ‘reason,’ ‘cause,’ ‘sake’), which are historically bound nouns. Only the construction yuwe n-i is attested with relatively high frequency (50 cases in the Man’yōshū), and is found in both Early and Late Western Old Japanese. The construction tamɛ n-i is very rare (only 8 cases in the whole Man’yōshū), and it is attested reliably only in Late Western Old Japanese.11 耶麼能謎能故思麼古喩衛爾

yamanǝmbɛ-nǝ ko-sima ko yuwe n-i Yamanǝmbɛ-GEN DIM-island child sake DV-CONV for the sake of the girl Ko-sima (lit.: Small island) from Yamanǝmbɛ (NK 79) 人嬬故尓吾戀目八方

PITƏ-ⁿ-DUMA YUWE n-i ARE KOPÏ-m-ɛ ya mo person-GEN-spouse reason DV-CONV I love-TENT-EV IP EP because [she] is the wife of [another] person, should I love [her]? [Certainly not!] (MYS 1.21) 和礼由惠尓於毛比和夫良牟伊母我可奈思佐

ware yuwe n-i omop-i-wamb-uram-u imǝ-ŋga kanasi-sa I reason DV-CONV think-CONV-worry-TENT2-ATTR beloved-POSS dear-NML [feeling of] the endearment for [my] beloved who probably worries because of me (MYS 15.3727) 11  In contrast to Middle Japanese bakari n-i (Vovin 2003: 173), there is no construction *mbakari n-i ‘being about’ in Western Old Japanese. Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

464

Chapter 6

曽己由惠尓情奈具也

sǝkǝ yuwe n-i KƏKƏRƏ naŋg-u ya there reason DV-CONV heart calm.down-FIN IP will [my] heart calm down due to those circumstances? (MYS 19.4154) 之可流可由惠尓序礼宇気牟比止良

sik[a]-ar-u-ŋga yuwe n-i sǝre ukɛ-m-u pitǝ-ra thus-exist-ATTR-POSS reason DV-CONV that receive-TENT-ATTR person-PLUR because of this, the people who will receive that (SSI)12 淺茅原後見多米尓

ASAⁿDI PARA NƏTI MI-M-U tamɛ n-i Asaⁿdi field after see-TENT-ATTR for DV-CONV in order to see Asandi field later (MYS 7.1342) 多我多米爾奈礼

ta-ŋga tamɛ n-i nare who-POSS for DV-CONV you for whom? [For] you! (MYS 17.4031) 和藝毛故尓美勢牟我多米尓母美知等里氐牟

wa-ŋg-imo-ko-ni mi-se-m-u-ŋga tamɛ n-i mǝmit-i tǝr-i-te-m-u I-POSS-beloved-DIM-DAT see-CAUS-TENT-ATTR-POSS for DV-CONV leaves. turn.red/yellow-NML take-CONV-PERF-TENT-FIN [I] want to take red leaves in order to show [them] to my beloved (MYS 19.4222) 知知波波賀多米爾毛呂比止乃多米爾

titi papa-ŋga tamɛ n-i morǝ pitǝ-nǝ tamɛ n-i father mother-POSS for DV-CONV many person-GEN for DV-CONV for father and mother, for many people (BS 1) (4) The converb n-i can be also followed by the verbs nar- ‘to become,’ se- ‘to do,’ and the adjective posi ‘be desirable.’ These combinations produce special constructions: – X n-i nar- ‘to become X’ – X n-i posi ‘to want as X,’ ‘to want to be X’ 12  I cite this example according to Omodaka et al. (1967: 407), who unfortunately does not indicate what exactly the document is.

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

465

Verbs

– Y(-wo) X n-i se- ‘to make Y being X,’ ‘to treat Y as X’ Examples: 由布弊爾奈礼婆伊射祢余登

yupu-m-be n-i nar-e-mba iⁿza ne-yǝ tǝ evening-GEN-side DV-CONV become-EV-CON INTER sleep-IMP DV When [it] became evening, and [we] told [him]: ‘[Go to] sleep!’ (MYS 5.904) 阿佐奈佐奈安我流比婆理爾奈里弖之可

asa-na [a]sa-na aŋgar-u pimbari n-i nar-i-te-si ka morning-PLUR morning-PLUR rise-ATTR skylark DV-CONV become-CONVPERF(CONV)-PAST/ATTR EP every morning, [I] want to have become a skylark, flying up (MYS 20.4433) 朕高御座爾坐始由理今年尓至麻低六年尓成奴

WARE TAKA MI-KURA-ni IMAS-I-SƏMƐS-U-yuri KƏ TƏSI-ni ITAR-U-maⁿde MU TƏSI n-i NAR-I-n-u I high HON-seat-LOC be(HON)-CONV-begin-ATTR-ABL this year-LOC reachATTR-TERM six year DV-CONV become-CONV-PERF-FIN [It] has been six years this year since I have been on the high throne (SM 7) 企許斯遠周久爾能麻保良叙可爾迦久爾保志伎麻爾麻爾斯可爾波阿羅慈迦

kikǝs-i-wos-u kuni-nǝ ma-po-ra ⁿzǝ ka n-i ka-ku n-i posi-ki manima n-i sika n-i pa ar-aⁿzi ka rule(HON)-CONV-HON-ATTR country-GEN INT-top-LOC FP thus DV-CONV thus-CONV DV-CONV desire-ATTR according DV-CONV thus DV-CONV TOP exist-NEG/TENT IP in the highest place of the country, where [the emperor] rules, [it] would not be thus according to what [you] wish to be this way and that way, [would it]? (MYS 5.800) 奈禮乎曾與咩爾保師登多禮

nare-wo sǝ yǝme n-i posi tǝ tare you-ACC FP bride DV-CONV be.desirable DV who who wants you as [his] bride? (NR 2.33) 家尓之弖吾者将戀名

ipe n-i s-i-te WARE pa KOPÏ-M-U na home DV-CONV do-CONV-SUB I TOP long-TENT-FIN EP After [I] come home, I will long for [it]! (MYS 7.1179)

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

466

Chapter 6

戀之家婆形見尓将為

KOPÏsi-k-emba KATAMI n-i SE-M-U miss-ATTR-COND farewell present DV-CONV do-TENT-FIN If [you] miss [me], [I] want to make [it] to be a farewell present [for you] (MYS 8.1471) 挿頭尓将為跡我念之櫻花

KAⁿZASI n-i SE-M-U tǝ A-ŋGA OMƏP-I-si SAKURA-NƏ PANA head.ornament DV-CONV do-TENT-FIN DV I-POSS think-CONV-PAST/ATTR sakura-GEN blossom [I] think [I] want to make the sakura blossoms about which I thought into a head ornament (MYS 16.3786) 2.4.1.1.1 Special Compressed Form -nThe converb form n-i of the defective verb n- ‘to be’ may have a special compressed form -n- that surfaces as a prenasalization (realized as -m-, -ⁿ-, -ŋ-) if the underlying form of a following head nominal starts with a voiceless consonant /p/, /t/, /k/, or /s/. However, to the best of my knowledge, there is only one uncontroversial example of the contracted form -ŋ- in Western Old Japanese: 曾能阿牟袁阿岐豆波夜具比

sǝnǝ amu-wo akiⁿdu paya-ŋ-kup-i that horsefly-ACC dragonfly quick-DV(CONV)-eat-CONV a dragonfly quickly ate that horsefly (KK 97) Comparative Data Level A: Other Japonic A1: Eastern Old Japanese The converb form n-i is attested in Eastern Old Japanese in the same functions as in Western Old Japanese with the exception of the constructions yuwe n-i and tamɛ n-i. (1) Nominal predicate: 夜麻邊能之牙可久尓伊毛呂乎多弖天左祢度波良布母

yama-m-BE-nǝ siŋge-k-aku n-i imo-rǝ-wo tate-te sa-ne-ⁿ-do parap-umǝ mountain-GEN-side-GEN thick-ATTR-NML DV-CONV beloved-DIM-ACC make.stand(CONV)-SUB PREF-sleep(NML)-DV(ATTR)-place clean-EXCL as the mountain (side) is overgrown [with bush], [I] let my beloved stand, and [I] am clearing a place [for us] to sleep! (MYS 14.3489)

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

467

Verbs 比登祢呂爾伊波流毛能可良

pitǝ ne-rǝ n-i ip-ar-u monǝkara one peak-DIM DV-CONV say-PROG-ATTR CONJ Although [they] say that [we] are one peak … (MYS 14.3512) 多可伎祢爾久毛能都久能須和礼左倍爾伎美爾都吉奈那

taka-ki ne-ni kumo-nǝ tuk-u-nǝsu ware sapɛ n-i kimi-ni tuk-i-n-ana high-ATTR peak-LOC cloud-GEN attach-ATTR-COMP I RP DV-CONV lord-DAT attach-CONV-PERF-DES Even I would like to cling to [my] lord like clouds cling to a high peak (MYS 14.3514) (2) Adverbialization: 麻左弖尓毛乃良奴伎美我名宇良尓弖尓家里

masaⁿde n-i mo nǝr-an-u kimi-ŋga NA ura-ni [i] ⁿde-n-i-ker-i clear DV-CONV FP say-NEG-ATTR lord-POSS name divination-LOC go.out (CONV)-PERF-CONV-RETR-FIN The name of [my beloved] lord which [I] did not tell [to anyone] became clearly known through divination (MYS 14.3374) 安夜尓可奈思母

aya n-i kanasi-mǝ extreme DV-CONV be.dear-EXCL [she] is extremely dear [to me]! (MYS 14.3408) 志保悲乃由多尓於毛敝良婆

sipo pï-nǝ yuta n-i omop-er-amba tide dry(NML)-COMP carefree DV-CONV think-PROG-COND If I would be thinking [about you], as carefree as an ebbing tide (MYS 14.3503) (3) The constructions yuwe n-i ‘because of’ and tamɛ n-i ‘for’ are not attested in Eastern Old Japanese. (4) Before the verbs nar- ‘to become’ and se- ‘to do’:13

13  The construction n-i posi ‘to want as’ is not attested in Eastern Old Japanese.

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

468

Chapter 6

安比豆祢能久爾乎佐杼抱美安波奈波婆斯努比尓勢毛等比毛牟須婆佐祢

Apiⁿdune-nǝ kuni-wo saⁿ-tǝpo-mi ap-an-ap-amba sinop-i n-i se-m-wo tǝ pimo musumb-as-an-e Apiⁿdune-GEN land-ABS PREF-far-GER meet-NEG-ITER-COND long.for-NML DV-CONV do-TENT-ATTR DV cord tie-HON-DES-IMP If [we] continue not to meet, because the land of Apindune is far, [I] want [you] to tie [your garment] cords, as if longing for [me] (MYS 14.3426) 麻多妣爾奈理奴

ma-tambi n-i nar-i-n-u INT-journey DV-CONV become-CONV-PERF-FIN [it] became a really [long] journey (MYS 20.4388) A2: Ryukyuan The converb n-i of the defective verb n- ‘to be’ is attested in Old Ryukyuan as well as in modern dialects, including Southern Ryukyuan. Examples: Old Ryukyuan 主里もりあせはつちぎりにきらせ

SIYORI mori ase fa tuti-gir-i n-i kir-as-e Shuri castle warrior TOP ground-(DV)cut-NML DV-CONV cut-HON-IMP Warriors [of] the Shuri castle, cut [the enemy] as cutting the ground (OS 1.33) Classical Ryukyuan いろいろに言ちもいかなしも行かぬ

iro-iro n-i I-ti-mo ika nas-i-mo Ik-an-u different DV-CONV say(CONV)-SUB-PT how do-CONV-PT go-NEG-FIN whatever [you] say, and whatever [you] do, [I] will not go (RK 725) Yaeyama

daa katti n-i śaa you one’s own way DV-CONV do(IMP) Do as you like (Nohara 1998: 474) 2.4.1.2 Attributive Form n-ǝ The function of the attributive form n-ǝ is the same as that of its Middle Japanese or Modern Japanese counterpart n-o in such examples as MJ Taketori n-o okina ‘old man Bamboo-cutter,’ aruzi n-o otoko ‘a man who is the host,’ Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

469

Verbs

Isitukuri n-o miko ‘prince Isitukuri’ and MdJ tomodati no gakusei ‘a student who is [my] friend,’ mei no Sumiko ‘Sumiko who is [my] niece,’ etc. The attributive ending -ǝ, rather than -u that might have been expected, is possibly paralleled by the Eastern Old Japanese attributive suffix -o, e.g.: ar-o ‘exist-ATTR,’ pik-o ‘pull ATTR,’ tat-o ‘stand-ATTR,’ yuk-o ‘go-ATTR,’ which will be discussed in more detail in the section on the attributive. A similar attributive form -o is preserved in the modern Hachijō dialect. The identification, however, is not without its problems, because there is some evidence that this final attributive -o in Eastern Old Japanese goes back to kō-rui /o/ rather than to otsu-rui /ǝ/. 比佐迦多能阿米能迦具夜麻

pisa kata n-ǝ amɛ-nǝ kaŋgu-yama eternal hard DV-ATTR heaven-GEN Kaŋgu(p. n.)-mountain eternal and strong Amɛ-no Kaŋguyama (lit.: Heavenly Kaŋgu mountain) (KK 27) 都流岐能多知

turuki n-ǝ tati double-edged sword DV-ATTR long.sword a long sword that is a double-edged sword (KK 33) 佐斯夫能紀

sasimbu n-ǝ kï sasimbu DV-ATTR tree sasimbu tree (KK 57) 賣杼理能和賀意富岐美能淤呂須波多他賀多泥呂迦母

meⁿdǝri n-ǝ wa-ŋga opǝ kimi-nǝ or-ǝs-u pata ta-ŋga tane rǝ kamǝ Meⁿdǝri DV-ATTR I-POSS great lady-GEN weave-HON-ATTR fabric who-POSS material DV EP The fabric my lady Meⁿdǝri weaves, I wonder for whom (lit. whose) the material is? (KK 66) 奈爾能都底擧騰

nani n-ǝ tute-kǝtǝ what DV-ATTR report(NML)-word What message [do you have]? (NK 128) 等保乃朝庭

tǝpo n-ǝ MIKAⁿDO

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

470

Chapter 6

far DV-ATTR Imperial Court distant Palace (MYS 5.794) 多那礼乃美巨騰

ta-nare n-ǝ mi-kǝtǝ hand-accustom(NML) DV-ATTR HON-koto favorite koto (MYS 5.812) 麻多麻奈須布多都能伊斯乎

ma-tama-nasu puta-tu n-ǝ isi-wo INT-jewel-COMP two-CL DV-ATTR stone-ACC to stones like real jewels (MYS 5.813) 等保能久尓

tǝpo n-ǝ kuni distant DV-ATTR country distant country (MYS 15.3688) 可多美能許呂母

katami n-ǝ kǝrǝmǝ keepsake DV-ATTR garment a garment that is a keepsake (MYS 15.3733) 伊豆礼能日麻弖安礼古非乎良牟

iⁿdure n-ǝ PI-maⁿde are kopï-wor-am-u which DV-ATTR day-TERM I long.for(CONV)-exist-TENT-FIN until what day should I be longing for [you]? (MYS 15.3742) 兄乃君

SE n-ǝ KIMI beloved DV-ATTR lord beloved lord (MYS 16.3885) 弟乃美許等

OTƏ n-ǝ mi-kǝtǝ younger brother DV-ATTR HON-thing darling who is [my] younger brother (MYS 17.3957) 佐夜麻太乃乎治我其日爾母等米安波受家牟

sa yamaⁿda n-ǝ woⁿdi-ŋga SƏNƏ PI-ni mǝtǝmɛ ap-aⁿz-u-k-em-u

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

471

Verbs

so Yamaⁿda DV-ATTR old.man-POSS that day-LOC search(CONV) meet-NEGCONV-PAST/FIN-TENT-FIN So, old man Yamaⁿda searched for [him] on that day, [but] did not find [him] (MYS 17.4014) 此橘乎等伎自久能可久能木實等名附家良之母

KƏNƏ TATImBANA-wo tǝkiⁿzi-ku n-ǝ kaŋg-u n-ǝ KƏ-NƏ MÏ tǝ NA-ⁿ-DUKƐ-kerasi-mǝ this mandarin.orange-ACC be off season-CONV DV-ATTR smell-ATTR DVATTR tree-GEN fruit DV name-LOC-attach(CONV)-RETR-SUP-EXCL [we] should call these mandarin oranges fragrant tree fruits that are off season (MYS 18.4111) Note that in this example the first n-ǝ follows the converb form -ku. The usage of the second attributive n-ǝ in this example is idiosyncratic and even possibly ungrammatical, since the attributive modifies the following noun by itself, and n-ǝ is not necessary. 己礼乃与波宇都利佐留止毛

kǝre n-ǝ yǝ pa utur-i sar-u tǝmo this DV-ATTR world TOP change-CONV go.away-FIN CONJ Even though this world changes and goes away … (BS 10) 久須理師波都祢乃母阿礼等麻良比止乃伊麻乃久須理師多布止可理家利米 太志加利鶏利

kusurisi pa tune n-ǝ mǝ ar-e-ⁿdǝ marapitǝ n-ǝ ima-nǝ kusurisi taputǝ-k-ar-i-ker-i mɛNdasi-k-ar-i-ker-i medicine.man TOP usual DV-ATTR FP exist-EV-CONC guest DV-ATTR now-GEN medicine.man revered-CONV-exist-CONV-RETR-FIN praiseworthy-CONVexist-CONV-RETR-FIN Although there are usual medicine men, too, the present Guest Medicine Man is [indeed] revered. [He] is praiseworthy (BS 15) Note the ellipsis of the word kusurisi ‘medicine man’ after the first n-ǝ. As a result, tune n-ǝ ‘usual’ functions by itself as a noun phrase without a following head noun, and consequently it is directly followed by the focus particle mǝ. 与都乃閇美伊都都乃毛乃乃阿都麻礼流伎多奈伎微乎婆

yǝ-tu n-ǝ pɛmi itu-tu n-ǝ monǝ-nǝ atumar-er-u kitana-ki mï-womba four-CL DV-ATTR snake five-CL DV-ATTR demon-GEN gather-PROG-ATTR dirty-ATTR body-ACC(EMPH) the dirty body where four snakes and five demons have accumulated (BS 19) Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

472

Chapter 6

2.4.1.2.1 Special Compressed Form -nThe attributive form n-ǝ of the defective verb n- ‘to be’ may have a special compressed form -N- that surfaces as a prenasalization (realized as -m-, -ⁿ-, -ŋ-) if the underlying form of a following head nominal starts with a voiceless consonant /p/, /t/, /k/, or /s/. 夜弊賀岐都久流

ya-pe-ŋ-gaki tukur-u eight-fold-DV(ATTR)-fence make-FIN [I] am making an eight-folded fence (KK 1) 阿加陀麻

aka-ⁿ-dama red-DV(ATTR)-jewel red jewel (KK 7) 美都具理能曾能那迦都迩

mi-tu-ŋ-guri-nǝ sǝnǝ naka-tu ni three-CL-DV(ATTR)-chestnut-COMP that middle-GEN/LOC clay that clay from the middle that is like three chestnuts (KK 42) 加久能碁登那爾淤波牟登

ka-ku-nǝ ŋgǝtǝ na-ni op-am-u tǝ thus-CONV-GEN like name-LOC carry-TENT-FIN DV in order to perpetuate that [it] was like that (KK 97) 古陀加流伊知

ko-ⁿ-daka-[a]r-u iti DIM-DV(ATTR)-high-exist-ATTR meeting.place slightly elevated meeting.place (KK 101) 柂我佐基泥佐基泥曾母野倭我底騰羅須謀野

ta-ŋga sakï-ⁿ-de sakï-ⁿ-de sǝ mǝ ya wa- ŋga te tǝr-as-umo ya who-POSS chap(NML)-DV(ATTR)-hand chap(NML)-DV(ATTR)-hand FP EP EP I-POSS hand take-HON-EXCL IP whose chapped hand, chapped hand will take my hand?! (NK 108)

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

473

Verbs

Comparative Data Level A: Other Japonic A1: Eastern Old Japanese The attributive form n-ǝ functions in Eastern Old Japanese in the same way as it does in Western Old Japanese: 西乃古何夜蘇志麻加久理

se n-ǝ ko-ŋga yaso sima kakur-i beloved DV-ATTR child-POSS eighty island hide-CONV beloved girl hides [behind] eighty islands … (FK 8) 勢能古

se n-ǝ ko beloved DV-ATTR child beloved man (MYS 14.3458) 麻等保久能野尓毛安波奈牟

ma-tǝpo-ku n-ǝ NO-ni mo ap-ana-m-u INT-distant-CONV DV-ATTR field-LOC FP meet-DES-TENT-FIN [I] would like to meet [you] even in a distant field (MYS 14.3463) 意富伎美乃之許乃美多弖等伊埿多都和例波

opo kimi-nǝ sikǝ n-ǝ mi-tate tǝ iⁿde-tat-u ware pa great lord-GEN unworthy DV-ATTR HON-shield DV exit(CONV)-leave-ATTR I TOP I leave today to be an unworthy shield of the sovereign (MYS 20.4373) 比毛多要婆安我弖等都気呂許礼乃波流母志

pimo taye-mba a-ŋga te-tǝ tukɛ-rǝ kǝre n-ǝ paru mǝs-i cord tear-COND your.own-POSS hand-COM attach-IMP this DV-ATTR needle hold-CONV if the cords [of your garment] tear, attach them with your own hand, holding this needle (MYS 20.4420) Special Compressed Form -n-

Eastern Old Japanese also has the special compressed -n- form: 波故祢能祢呂乃爾古具佐

pakone-nǝ ne-rǝ-nǝ niko-ŋ-kusa Pakone-GEN peak-DIM-GEN niko-DV(ATTR)-grass niko grass of the little peak in Pakone (MYS 14.3370) Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

474

Chapter 6

A2: Ryukyuan The attributive form n-o of the defective verb n- ‘to be’ is well attested in Old Ryukyuan. The attributive form n-u is attested in many modern dialects, including South Ryukyuan, but in the available descriptions this form is mostly found in the attributive form of demonstrative pronouns, like ʔunu ‘that’ which is historically from *ʔu n-u ‘that DV-ATTR.’ Examples: Old Ryukyuan あけのみあおり

ake n-o mi-aori beautiful DV-ATTR HON-shade.umbrella beautiful shade umbrella (OS 12.657) Yaeyama

patuũãã n-u śima Hatoma DV-ATTR island Hatoma island (Nohara 1998: 445) ʔunu pïtu that(ATTR) person that person (Nohara 1998: 474) 2.4.1.3 Subordinative Converb Form n-i-te The subordinative converb form n-i-te is a rare form in Western Old Japanese. It is attested only in the Man’yōshū, where it occurs five times. N-i-te has only two functions: (1) copula in a nominal predicate, (2) marker of a location of an action or a state. Examples: (1) As a copula in a nominal predicate: 常丹毛冀名常處女畠手

TUNE n-i moŋga na TUNE WOTƏME n-i-te eternal DV-CONV DP EP eternal maiden DV-CONV-SUB [I] want to be eternal, being an eternal maiden! (MYS 1.22) 都祢比等能故布登伊敷欲利波安麻里爾弖和礼波之奴倍久奈里尓多良受也

tune pitǝ-nǝ kop-u tǝ ip-u-yori pa amari n-i-te ware pa sin-umbɛ-ku nar-i-n-itar-aⁿz-u ya

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

475

Verbs

ordinary person-GEN long.for-FIN DV say-ATTR-ABL TOP excess DV-CONV-SUB I TOP die-DEB-CONV become-CONV-PERF-CONV-PERF/PROG-NEG-FIN IP Did not [it] become so that I should die, feeling (lit.: being) much more than what ordinary people call ‘longing’? (MYS 18.4080) (2) As a marker of a location of an action or a state: 京師尓而誰手本乎可吾将枕

MIYAKO n-i-TE TA-ŋGA TAMƏTƏ-wo ka WA-ŋGA MAKURAK-AM-U capital DV-CONV-SUB who-POSS sleeve-ACC IP I-POSS use.as.a.pillow-TENTATTR Whose sleeves will I use as a pillow at the capital? (MYS 3.439) 多婢尓弖毛母奈久波也許登

tambi n-i-te mo mǝ na-ku paya kǝ tǝ journey DV-CONV-SUB FP misfortune no-CONV quick come(IMP) DV [my beloved] said: “Come [back] quickly without any misfortune on [your] journey!” (MYS 15.3717) 家尓底母多由多敷命

IPE n-i-te mǝ tayutap-u INƏTI home DV-CONV-SUB FP be.unstable-ATTR life [my] life which is uncertain even at home (MYS 17.3896) Comparative Data Level A: Other Japonic A2: Ryukyuan The subordinative converb n-i-te is attested only in Classical Ryukyuan (in Ryūka) (Hokama 1995: 508), and does not reveal itself in either Old Ryukyuan or modern dialects. Thus, it is certainly a loan from Classical Japanese. The absence of the n-i-te form in both Eastern Old Japanese and Ryukyuan suggests that this form is a result of an internal development in Central Japanese. 2.4.1.4 Special Form nar- < n-i arThere is a special form nar-, which resulted from contraction of the converb form n-i with the following verb ar- ‘to exist.’ While in Middle Japanese this form is the only one used, except in cases where the focus particles mo and fa, the emphatic particle si, or the interrogative particle ka are placed between the converb n-i and ar-, in Western Old Japanese the non-contracted form with

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

476

Chapter 6

ar- directly following n-i is used much more frequently than the contracted form. The contracted form nar- is quite rare, and the list of examples below includes almost all attested examples where the contracted form is written phonographically or partially phonographically. The verb ar- ‘to exist’ is used after n-i as a ‘dummy’ auxiliary that allows the attachment of other auxiliaries or suffixes that cannot directly follow the defective verb n-. Care must be taken to distinguish three confusing verbs in Western Old Japanese: nar- ‘to become’ (consonant verb), nar- ‘to be’ (contraction of the converb form n-i of the defective verb n- ‘to be’ and ar- ‘to exist,’ an r-irregular verb), and nar- ‘to exist at,’ ‘to be located at’ (a contraction of the locative case marker -ni and ar- ‘to exist,’ an r-irregular verb). For examples of the last one, see chapter 4, section 1.2.2.4.4. 許能美岐波和賀美岐那良受

kǝnǝ mi-ki pa wa-ŋga mi-ki nar-aⁿz-u this HON-rice.wine TOP I-POSS HON-rice.wine be-NEG-FIN This rice wine is not my rice wine (KK 39) 柂摩儺羅磨婀我裒屢柂摩

tama nar-amba a-ŋga por-u tama jewel be-COND I-POSS desire-ATTR jewel if [my beloved] were a jewel, [she would be] a jewel I desire (NK 92) 蘇餓能古羅破宇摩奈羅麼譬武伽能古摩

Soŋga-nǝ ko-ra pa uma nar-amba pimuka-nǝ ko-[u]ma Soŋga-GEN son-PLUR TOP horse be-COND Pimuka-GEN DIM-horse If the sons of Soŋga were horses, [they would be] the stallions of Pimuka … (NK 103) 迦久能尾奈良志

ka-ku nǝmï nar-asi thus-CONV RP be-SUP [It] is likely to be just this way (MYS 5.804) 安布毛能奈良婆

ap-u monǝ nar-amba meet-ATTR thing be-COND if [we] would meet (MYS 15.3731) 加武賀良奈良之

kamu-ŋ-gara nar-asi Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

477

Verbs

deity-DV(ATTR)-nature be-SUP [It] is likely to be [its] divine nature (MYS 17.4001) 秋風尓比毛等伎安氣奈多太奈良受等母

AKI KAⁿZE-ni pimo tǝk-i-akɛ-na taⁿda nar-aⁿz-u tǝmǝ autumn wind-LOC cord untie-CONV-open-DES direct be-NEG-FIN CONJ [I] wish that the autumn wind would untie the cords, even if [it] is not directly (MYS 20.4295) 伎美奈良奈久尓

kimi nar-an-aku n-i lord be-NEG-NML DV-CONV as [it] is not [my] lord (MYS 20.4447) 逆在流人止母在而

SAKASIMA NAr-u PITƏ-ⁿdǝmǝ AR-I-TE rebellious be-ATTR person-PLUR exist-CONV-SUB there were people who were rebellious (SM 16) 汝多知諸者吾近姪奈利

IMASI-tati MƏRƏ pa WA-ŋGA TIKA-KI WOPI nar-i you-PLUR all TOP I-POSS close-ATTR nephew be-FIN All [of] you are my close nephews (SM 17) 伊可奈留夜比止爾伊麻世可

ika nar-u ya pitǝ n-i imas-e ka how be-ATTR IP person DV-CONV be(HON)-EV IP Oh, what kind of person is [he]? (BS 5) 2.4.1.4.1

Non-contracted Form n-i ar-

伊麻許曾婆和杼理迩阿良米能知波那杼理爾阿良牟遠

ima kǝsǝ pa wa-ⁿ-dǝri n-i ar-am-ɛ nǝti pa na-ⁿ-dǝri n-i ar-am-u-wo now FP TOP I-OSM-bird DV-CONV exist-TENT-EV after TOP you-OSM-bird DV-CONV exist-TENT-ATTR-ACC Now [I] am my bird, later [I] will be your bird, so … (KK 3) 比登都麻都比登迩阿理勢婆

pitǝ-tu matu pitǝ n-i ar-i-s-emba one-CL pine person DV-CONV exist-CONV-PAST/ATTR-COND if a lone pine were a person (KK 29) Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

478

Chapter 6

伊可爾安良武日能等伎爾可母

ika n-i ar-am-u PI-nǝ tǝki-ni kamǝ how DV-CONV exist-TENT-ATTR day-GEN time-LOC EP I wonder, in what time of the day … (MYS 5.810) 大船尓伊母能流母能尓安良麻勢婆

OPO PUNE-ni imǝ nǝr-u mǝnǝ n-i ar-amas-emba big boat-LOC beloved board-ATTR thing DV-CONV exist-SUBJ-COND If [it] were the case that [my] beloved boarded [my] big boat … (MYS 15.3579) 伊毛尓安礼也夜須伊毛祢受弖安我故非和多流

imo n-i ar-e ya yasu i mo ne-ⁿz-u-te a-ŋga kopï-watar-u beloved DV-CONV exist-EV IP easy sleep FP sleep-NEG-CONV-SUB I-POSS long.for(CONV)-cross-ATTR Is [she] my beloved? I do not sleep easily, and continue to long for [her] (MYS 15.3633) 安米都知能可未奈伎毛能尓安良婆

amɛ tuti-nǝ kamï na-ki monǝ n-i ar-amba heaven earth-GEN deity no-ATTR thing DV-CONV exist-COND If heaven and earth were without deities … (MYS 15.3740) 烏梅乃花美夜万等之美尓安里登母

uMƐ-nǝ PANA mi-yama tǝ sim-i n-i ar-i tǝmǝ plum-GEN blossom HON-mountain DV grow.thick-NML DV-CONV exist-FIN CONJ Even though plum blossoms are blooming densely as a mountain (MYS 17.3902) 伊米尓波母等奈安比見礼騰多太尓安良祢婆孤悲夜麻受家里

imɛ-ni pa mǝtǝna api-MI-re-ⁿdǝ taⁿda n-i ar-an-e-mpa kopï yam-aⁿz-u-ker-i dream-LOC TOP in.vain REC-see-EV-CONC direct DV-CONV exist-NEG-EVCON love(NML) stop-NEG-CONV-RETR-FIN although [we] see each other in vain in dreams, because [our meetings] are not direct, [our] longing does not stop (MYS 17.3980) 伊可爾安流布勢能宇良曾毛

ika n-i ar-u puse-nǝ ura sǝ mo how DV-CONV exist-ATTR Puse-GEN bay it FP The bay of Puse, how incredibly [beautiful] it is! (MYS 18.4036)

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

479

Verbs 安佐之保美知尓与流許都美可比尓安里世婆都刀尓勢麻之乎

asa sipo mit-i-ni yǝr-u kǝtumi kapi n-i ar-i-s-empa tuto n-i se-masi-wo morning tide full-NML-LOC approach-ATTR trash shellfish DV-CONV existCONV-PAST/ATTR-COND souvenir DV-CONV do-SUBJ-ACC if the trash that is brought up by the full morning tide were shellfish, [I] would bring [it] as a souvenir, but … (MYS 20.4396) The non-contracted form n-i ar- is also found with the intervening focus particles mǝ, pa, kǝsǝ, and sǝ, the emphatic particle si and the interrogative particle ka in the constructions n-i mǝ ar-, n-i pa ar-, n-i kǝsǝ ar-, n-i sǝ ar-, n-i si ar- and n-i ka ar-. Among these, n-i kǝsǝ ar- is attested only once, but the others occur several times each, although none of them is very frequent. Interestingly, neither the emphatic particle namo, nor the interrogative particle ya occur between n-i and ar-. 阿波母與賣迩斯阿禮婆那遠岐弖遠波那志

a pa mǝ yǝ me n-i si ar-e-mba na-wo [o]k-i-te wo pa na-si I TOP EP EP woman DV-CONV EP exist-EV-CON you-ACC leave-CONV-SUB man TOP no-FIN Because I am a woman, [I] have no [other] man, besides you (KK 5) 企許斯遠周久爾能麻保良叙可爾迦久爾保志伎麻爾麻爾斯可爾波阿羅慈迦

kikǝs-i-wos-u kuni-nǝ ma-po-ra ⁿzǝ ka n-i ka-ku n-i posi-ki manima n-i sika n-i pa ar-aⁿzi ka rule(HON)-CONV-HON-ATTR country-GEN INT-top-LOC FP thus DV-CONV thus-CONV DV-CONV desire-ATTR according DV-CONV thus DV-CONV TOP exist-NEG/TENT IP in the highest place of the country, where [the emperor] rules, [it] would not be thus according to what [you] wish to be this way and that way, [would it]? (MYS 5.800) 飛立可祢都鳥尓之安良祢婆

TƏmB-I-TAT-I-kane-t-u TƏRI n-i si ar-an-e-mba fly-CONV-depart-CONV-NEG/POT(CONV)-PERF-FIN bird DV-CONV EP existNEG-EV-CON [I] could not fly away because [I] am not a bird (MYS 5.893)

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

480

Chapter 6

妹等安里之時者安礼杼毛和可礼弖波許呂母弖佐牟伎母能尓曽安里家流

IMO-tǝ ar-i-si TƏKI pa ar-e-ⁿdǝmo wakare-te pa kǝrǝmǝⁿde samu-ki mǝnǝ n-i sǝ ar-i-ker-u beloved-COM exist-CONV-PAST/ATTR time TOP exist-EV-CONC separate (CONV)-SUB TOP sleeve cold-ATTR thing DV-CONV FP exist-CONV-RETRATTR Although there was a time when [I] was with [my] beloved, since [we] separated, [my] sleeves are cold (MYS 15.3591) 秋夜乎奈我美爾可安良武

AKI-NƏ YO-wo naŋga-mi n-i ka ar-am-u autumn-GEN night-ABS long-GER DV-CONV IP exist-TENT-ATTR Is [it] probably because the autumn night is long? (MYS 15.3684) 伊敝之麻波奈尓許曽安里家礼

ipe sima pa na n-i kǝsǝ ar-i-ker-e Home island TOP name DV-CONV FP exist-CONV-RETR-EV “Home island” turned out to be just a name (MYS 15.3718) 知里比治能可受尓母安良奴和礼由惠尓於毛比和夫良牟伊母

tiri piⁿdi-nǝ kaⁿzu n-i mǝ ar-an-u ware yuwe n-i omop-i-wamb-uram-u imǝ dust dirt-COMP number DV-CONV FP exist-NEG-ATTR I reason DV-CONV think-CONV-worry-TENT-ATTR beloved [My] beloved, who was worrying on behalf of me who is not even worth counting like dust and dirt (MYS 15.3727) 許己呂奈伎登里尓曽安利家流保登等藝須

kǝkǝrǝ na-ki tǝri n-i sǝ ar-i-ker-u potǝtǝŋgisu heart no-ATTR bird DV-CONV FP exist-CONV-RETR-ATTR cuckoo cuckoo, [you] are indeed a bird that has no heart (MYS 15.3784) 伊毛我多可々々尓麻都良牟許己呂之可尓波安良司可

imo-ŋga taka taka n-i mat-uram-u kǝkǝrǝ sika n-i pa ar-aⁿzi ka beloved-POSS high high DV-CONV wait-TENT2-ATTR heart thus DV-CONV TOP exist-NEG/TENT IP the heart of [my] beloved who probably waits for [me] eagerly would not [it] be this way? (MYS 18.4107) 伊弊婢等乃伊波倍尓可安良牟

ipe-m-bitǝ-nǝ ipapɛ n-i ka ar-am-u

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

481

Verbs

home-GEN-person-GEN purify(NML) DV-CONV IP exist-TENT-ATTR Is [it] probably because of [my] home folks’ purification rites? (MYS 20.4409) 美麻久能富之伎吉美尓母安流加母

mi-m-aku-nǝ posi-ki kimi n-i mǝ ar-u kamǝ see-TENT-NML-GEN desire-ATTR lord DV-CONV FP exist-ATTR EP I wonder whether [it] is also [my] lord whom [I] want to see (MYS 20.4449) Comparative Data Level A: Other Japonic A1: Eastern Old Japanese Both the contracted form nar- and the non-contracted form n-i ar- are attested in Eastern Old Japanese, but their relative frequency is the opposite to that of Western Old Japanese: the contracted form nar- is attested much more frequently than the non-contracted form n-i ar-, which appears only two times. This probably should come as no surprise, given the fact that Eastern Old Japanese has a clear tendency to various kinds of contractions. 波奈都豆麻奈礼也

pana t-u tuma nar-e ya flower DV-ATTR spouse be-EV IP Are you a flower-wife? (MYS 14.3370) 伊可奈流勢奈可和我理許武等伊布

ika nar-u se-na ka wa-ŋgari kǝ-m-u tǝ ip-u how be-ATTR beloved-DIM IP I-DIR come-TENT-FIN DV say-ATTR what kind of beloved is [he], who says that [he] will come to me? (MYS 14.3536) 阿志氣比等奈里

asi-kɛ pitǝ nar-i bad-ATTR person be-FIN [he] is a bad person (MYS 20.4382) 久毛為尓美由流志麻奈良奈久尓

kumowi-ni mi-y-uru sima nar-an-aku n-i distance-LOC see-PASS-ATTR island be-NEG-NML DV-CONV although [it] is not an island that is seen in the distance (MYS 20.4355)

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

482

Chapter 6

Non-Contracted Form -n-i ar哭乎曽奈伎都流手兒尓安良奈久尓

NE-wo sǝ nak-i-t-uru teŋGO n-i ar-an-aku n-i voice-ACC FP cry-CONV-PERF-ATTR baby DV-CONV exist-NEG-NML DV-CONV [I] sobbed loudly, although [I] am not a baby (MYS 14.3485) 佐弁奈弁奴美許登尓阿礼婆

sape-n-ape-n-u mi-kǝtǝ n-i ar-e-mba refuse(NML)-LOC-match-NEG-ATTR HON-word DV-CONV exist-EV-CON Because [it] is [my sovereign’s] order that [I] cannot refuse … (MYS 20.4432) A2: Ryukyuan It seems that only the contracted form nar- ‘to be’ is present in Ryukyuan, and in addition it is limited to Old and Classical Ryukyuan. Thus, it is more than likely that here we are dealing with a loanword from mainland Middle Japanese into Old and Classical Ryukyuan. Old Ryukyuan いによはのおきてもちなる

Iniyofa-no okite-moti nar-u Iniyofa-GEN rule-holder be-ATTR [the one] who is a governor of Iniyofa (OS 8.456) Classical Ryukyuan 北京お主てだやずまにそなれゆが

FICIN o-SHU-teda ya zuma-ni so nar-e-yu ga Ficin-GEN HON-lord-sun TOP where-LOC next be-CONV-FIN IP Where does the emperor of Beijing live? (RK 3888) Level B: External Comparisons On the basis of all data presented above, it is clear that the defective verb n- ‘to be’ is present in all major branches of Japonic: Central Japanese, Eastern Old Japanese, and Ryukyuan. This allows us to project this verb back to the protoJaponic level and to reconstruct PJ *n- ‘to be.’ None of the ‘Altaic’ languages have any copulas that are comparable with PJ *n-. However, the copula in Ainu is ne, e.g.:

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

483

Verbs

Yoshiko ku-ne ruwe ne Yoshiko 1ps-be AFFIR be I am Yoshiko (Nakagawa and Nakamoto 1997: 22) Whether PJ *n- and Ainu ne represent a chance similarity, or evidence for old contacts between these two languages remains to be seen. However, my intuitive inclination is to view it as the latter possibility, since linguistic contacts between Ainu and Japonic unfortunately remains a poorly researched area, and there are quite a number of striking parallels between these two languages. 2.4.2 Defective Verb tǝ ‘To Be’ The defective verb tǝ ‘to be’ has only two paradigmatic forms: the converb tǝ and the attributive t-u. Presumably tǝ is also the root of this verb, and the converb form tǝ can be explained as a result of the loss of the converb -i after the vowel final root that can be observed with vowel verbs as well. Note that the defective verbs n- and tǝ are essentially doublets, although tǝ does not have as many functions or paradigmatic forms as n- does. Tǝ also occurs much more rarely than n-. In other words, it has a narrower scope of application and distribution. I will provide an explanation of these phenomena at the end of this section after discussing the comparative data from Eastern Old Japanese and Ryukyuan. 2.4.2.1 Converb Form tǝ The converb form tǝ ‘being’ is used in the same functions as the converb n-i, although in contrast to n-i it is found only once after the quasi-postposition tamɛ ‘for.’ It occurs predominantly after nominals, although there is one example when it appears after the attributive form of a verb (see MYS 4.780 below). The almost total absence of tǝ after the attributive forms of verbs and the total lack of it after nominalized verbal forms is significant, as I will demonstrate later. (1) The converb tǝ can be used as a copula in a nominal predicate: 那爾騰柯母于都倶之伊母我磨陀左枳涅渠農

nani tǝ kamǝ utukusi imǝ-ŋga mata sak-i-[i]ⁿde-kǝ-n-u what DV EP beautiful beloved-POSS again bloom-CONV-exit(CONV)-comeNEG-ATTR for (lit: being) what [reason], I wonder, does not [my] beautiful beloved bloom again? (NK 114) 布流由岐得比得能美流麻提烏梅能波奈知流

pur-u yuki tǝ pitǝ-nǝ mi-ru-maⁿde uMƐ-nǝ pana tir-u Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

484

Chapter 6

fall-ATTR snow DV person-GEN see-ATTR-TERM plum-GEN flower fall-FIN plum blossoms fall to such an extent that people will perceive them as falling snow (MYS 5.839) 吾勢子尓令見常念之梅花其十方不所見

WA-ŋGA se-KO-ni MI-SE-M-U tǝ OMƏP-I-si UMƐ-NƏ PANA SƏRE tǝ mo MI-YE-ⁿZ-U I-POSS beloved-DIM-DAT see-CAUS-TENT-FIN DV think-CONV-PAST/ATTR plum-GEN flower that DV FP see-PASS-NEG-CONV The plum blossoms that [I] was going to show to my beloved, do not look like those (i.e., like plum blossoms) (MYS 8.1426) 烏梅乃花美夜万等之美尓安里登母

uMƐ-nǝ PANA mi-yama tǝ sim-i n-i ar-i tǝmǝ plum-GEN blossom HON-mountain DV grow.thick-NML DV-CONV exist-FIN CONJ Even though plum blossoms are blooming densely as a mountain (MYS 17.3902) 大伴乃遠都神祖乃其名乎婆大來目主等於比母知弖

OPƏTƏMƏ-nǝ TƏPO t-u KAMU-OYA-nǝ SƏNƏ NA-womba OPƏKUMƐNUSI tǝ op-i-mǝt-i-te Opǝtǝmǝ-GEN distant DV-ATTR deity-ancestor-GEN that name-ACC(EMPH) Opokumɛnusi DV bear-CONV-hold-CONV-SUB Opǝtǝmǝ clan, bearing the name of [our] distant divine ancestor, as Opokumɛnusi (MYS 18.4094) 多礼乎可伎美等弥都都志努波牟

tare-wo ka kimi tǝ mi-tutu sinop-am-u who-ACC IP lord DV see(CONV)-COOR long.for-TENT-ATTR whom shall [I] long for, viewing [him] as [my] lord? (MYS 20.4440) 現御神止大八嶋国所知天皇大命

AK-I-TU MI-KAMÏ tǝ OPƏ-YA-SIMA-ŋ-GUNI SIRASIMES-U SUMERA-ŋGA OPƏMI-KƏTƏ open-NML-GEN/LOC HON-deity DV great-eight-island-GEN-country rule (HON)-ATTR emperor-POSS great-HON-deity emperor—Great Deity who rules the Great Country of Eight Islands as a Manifested Deity (SM 1) 授留人乎波一日二日止択比

SAⁿDUK-Uru PITƏ-womba PITƏ PI PUTU-KA tǝ ERAmb-i

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

485

Verbs

approve-ATTR person-ACC(EMPH) one-day two-CL DV select-CONV do [they] select an approved person just for one or two days? (SM 7) Similar to the converb n-i, the converb tǝ can be followed by the dummy auxiliary ar- ‘to exist’ (or its honorific form imas-), although, in contrast to the n-i ar- construction that can contract to nar-, there is no contracted form tar- in Western Old Japanese texts. The examples of tǝ ar- in Western Old Japanese are extremely rare: 志可登阿良農比宜可伎撫而

sika tǝ ar-an-u piŋgɛ kaki-NAⁿDE-TE thus DV exist-NEG-ATTR beard PREF-caress(CONV)-SUB stroking my thinning (lit.: not being thus) beard (MYS 5.892) 伊加登伊可等有吾屋前尓

ika tǝ ika tǝ AR-U WA-ŋGA YAⁿDO-ni spacious DV spacious DV exist-ATTR I-POSS house-LOC In my [extremely] spacious house … (MYS 8.1507) 人跡不在者桑子尓毛成益物乎

PITƏ tǝ AR-AⁿZ-U pa KUWA-KO n-i mo NAR-Amasi MƏNƏwo person DV exist-NEG-CONV TOP mulberry-child DV-CONV FP become-SUBJ CONJ if [I] was not a human being, [I] would [like to] become a silkworm, but … (MYS 12.3086) 美麻斯乃父止坐天皇乃美麻斯尓賜志天下

mimasi-nǝ TITI tǝ IMAS-U SUMERA-MIKƏTƏ-nǝ mimasi-ni TAMAP-I-si AMƐ-NƏ SITA you-GEN father DV exist(HON)-ATTR emperor-deity-GEN you-DAT give(HON)CONV-PAST/ATTR heaven-GEN bottom the land under the Heaven that the emperor-deity, who is your father, gave to you (SM 5) The non-contracted form tǝ ar- is also found with the intervening focus particles mǝ and pa, although these examples are extremely rare: there is just one example of tǝ mǝ ar- and two examples of tǝ pa ar-. Again, in sharp contrast to n-i ar-, the intervening focus particles kǝsǝ and sǝ, the emphatic particle si and the interrogative particle ka are not attested in the constructions *tǝ kǝsǝ ar-, *tǝ sǝ ar-, and *tǝ si ar-. Similar to n-i, neither the emphatic particle namo, nor the interrogative particle ya occur between tǝ and ar-. Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

486

Chapter 6

勤和氣登将譽十方不有

ISOSI-KI wakɛ tǝ POMƐ-M-U tǝ mǝ AR-AⁿZ-U hard-working-ATTR fellow DV praise-TENT-ATTR DV FP exist-NEG-FIN [he] is not even to be praised as a hard-working fellow (MYS 4.780) 比等等波安流乎比等奈美爾安礼母作乎

pitǝ tǝ pa ar-u-wo pitǝ nami-ni are mǝ TUKUR-U-wo person DV TOP exist-ATTR-ACC person usual-COMP I FP make-ATTR-ACC although [I] am a human being, [and] although I also make [it] like people usually [make it] (MYS 5.892) 不戀時等者不有友

KOPÏ-N-U TƏKI tǝ pa AR-AN-E-Ndǝmǝ long.for(CONV)-NEG-ATTR time DV TOP exist-NEG-EV-CONC although [it] is not a time when [I] do not long for [him] (MYS 13.3329) 弥蘇知阿麻利布多都乃加多知夜蘇久佐等曾太礼留比止

miso-ti amar-i puta-tu n-ǝ katati yaso kusa tǝ sǝⁿdar-er-u pitǝ thirty-CL exceed-CONV two-CL DV-ATTR mark eighty type DV be completePROG-ATTR person a person, who is endowed with thirty-two marks and eighty [lesser sign] types (BS 2) (2) Like the converb n-i, the converb tǝ is used after nouns and adjectives for adverbializations, but this usage in Western Old Japanese appears to be very infrequent: 佐佐那美遅袁須久須久登和賀伊麻勢婆夜

sasanami-ⁿ-di-wo suku-suku tǝ wa-ŋga imas-e-mba ya Sasanami-GEN-road-ACC rapidly DV I-POSS go(HON)-EV-CON EP as I went rapidly along the Sasanami road, yeah (KK 42) 都祢斯良農道乃長手袁久礼々々等伊可尓可由迦牟

tune sir-an-u MITI-nǝ NA ŋGA te-wo kure-kure tǝ ika n-i ka yuk-am-u usual know-NEG-ATTR road-GEN long place-ACC dark-dark DV how DV-CONV IP go-TENT-ATTR How would [I] go along the full length of the road that [I] normally do not know, being in a dark [mood]? (MYS 5.888)

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

487

Verbs 久礼久礼登獨曽我來

kure-kure tǝ PITƏ-RI WA-ŋGA K-URU dark-dark DV one-CL I-POSS come-ATTR I come alone in a dark [mood] (MYS 13.3237) (3) There is only one example in the Western Old Japanese corpus when tǝ follows the quasi-postposition tamɛ ‘for, in order to’: 和礼爾於止礼留比止乎於保美和多佐牟多米止宇都志麻都礼利

ware-ni otǝr-er-u pitǝ-wo opo-mi watas-am-u tamɛ tǝ utus-i-matur-er-i I-DAT be worse-PROG-ATTR person-ABS many-GER lead across-TENT-ATTR in.order.to DV carve-CONV-HUM-PROG-FIN because there are many people who have been worse than me, [I] have carved [Buddha’s footprints] in order to save [them] (BS 13) (4) The converb tǝ can also be followed by the verb nar- ‘to become,’ and the subordinative converb s-i-te of the verb se- ‘to do.’ In contrast to the converb n-i, it is not attested with the adjective posi ‘be desirable.’ The combinations of tǝ with nar- ‘to become’ and s-i-te ‘do-CONV-SUB’ produce special constructions: – X tǝ nar- ‘to become X’ – X tǝ s-i-te ‘doing as X,’ ‘in the capacity of X’ Both constructions occur quite seldom, especially tǝ s-i-te, which mostly appears in the Senmyō. Examples: 何時可毛比等等奈理伊弖天

ITU SI kamo pitǝ tǝ nar-i-iⁿde-te when EP EP person DV become-CONV-exit(CONV)-SUB when [he] becomes a [grown-up] person (MYS 5.904) 古京跡成者

PURU-KI MIYAKO tǝ NAR-I-NUR-E-mba be.old-ATTR capital DV become-CONV-PERF-EV-CON because [it] became an old capital (MYS 6.1048) 今者春部登成尓鷄類鴨

IMA pa PARU-pɛ tǝ nar-i-n-i-ker-u kamo now TOP spring-? DV become-CONV-PERF-CONV-RETR-ATTR EP [It] turned out that now [it] is (lit.: has become) spring! (MYS 8.1433)

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

488

Chapter 6

豊乃登之思流須登奈良思雪能敷礼流波

TƏYƏ n-ǝ tǝsi sirus-u tǝ nar-asi YUKI-nǝ pur-er-u pa abundant DV-ATTR year show.a.sign-ATTR DV become-SUP snow-GEN fallPROG-ATTR TOP A snowfall seems to become a good omen for an abundant year (MYS 17.3925) 真木乃伊多度乎等杼登之弖和我比良可武尓

MA-KÏ-nǝ ita-ⁿ-do-wo tǝⁿdǝ tǝ s-i-te wa-ŋga pirak-am-u-ni INT-tree-GEN board-GEN-door-ACC rap DV do-CONV-SUB I-POSS openTENT-ATTR-LOC when I open a wooden door made of a real tree with a rap (MYS 14.3467) Although this poem is found in the Azuma-uta volume, it does not have any Eastern Old Japanese features, therefore I treat it as a Western Old Japanese text. 令文所載多流乎跡止為而

NƏRI-NƏ PUMI-NI NƏSE-tar-u-wo ATƏ tǝ S-I-TE law-GEN scripture-LOC place(CONV)-PERF/PROG-ATTR-ACC FOOTSTEP DV DO-CONV-SUB taking what has been written in the law scriptures as a precedent (SM 2) 仲末呂伊忠臣止之天侍都

Nakamarǝ-i TAⁿDASI-KI OMI tǝ s-i-te PAmBER-I-t-u Nakamarǝ-ACT loyal-ATTR noble DV do-CONV-SUB serve-CONV-PERF-FIN Nakamarǝ has served as a loyal noble (SM 34) Comparative Data Level A: Other Japonic A1: Eastern Old Japanese The converb tǝ is also attested in Eastern Old Japanese in the same functions as in Western Old Japanese, but there is only one example of tǝ in the construction tǝ pa nar-, where tǝ is followed by the intervening topic particle pa. Most often tǝ is found in Eastern Old Japanese after the interrogative pronoun aⁿ- ‘what, why.’ 阿杼可多延世武

aⁿ-dǝ ka taye se-m-u why-DV IP break(NML) do-TENT-ATTR why should [we] break up? (MYS 14.3397)

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

489

Verbs 安加奴乎安杼加安我世牟

ak-an-u-wo aⁿ-dǝ ka a-ŋga se-m-u satisfy-NEG-ATTR-ACC what-DV IP I-POSS do-TENT-ATTR since [it] was not enough [for me], what should I do? (MYS 14.3404) 奴流我倍爾安杼世呂登可母

n-uru-ŋga [u]pɛ-ni aⁿ-dǝ se-rǝ tǝ kamǝ sleep-ATTR-POSS top-LOC what-DV do-IMP DV EP besides sleeping [with her], I wonder what [else do I] do? (MYS 14.3465) 波姑祢乃夜麻爾安波麻吉弖実登波奈礼留乎阿波奈久毛安夜思

Pakone-nǝ yama-ni apa mak-i-te MÏ tǝ pa nar-er-u-wo ap-an-aku/apa na-ku mo ayasi Pakone-GEN mountain-LOC millet sow-CONV-SUB fruit DV TOP becomePROG-ATTR-ACC meet-NEG-NML/millet no-NML FP be.strange(FIN) Although [I] have sown millet on the Pakone mountain and [it] ripened (lit.: became fruits), it is strange that [we] do not meet/there is no millet (MYS 14.3364) 安素乃河泊良欲伊之布麻受蘇良由登伎奴与

Aso-nǝ KApara-yo isi pum-aⁿz-u sora-yu tǝ k-i-n-u yǝ Aso-GEN river-bed-ABL stone tread-NEG-CONV sky-ABL DV come-CONVPERF-FIN EP [I] came from the river-bed of Aso, as from the sky, without treading on stones! (MYS 14.3425) 意富伎美乃之許乃美多弖等伊埿多都和例波

opo kimi-nǝ sikǝ n-ǝ mi-tate tǝ iⁿde-tat-u ware pa great lord-GEN unworthy DV-ATTR HON-shield DV exit(CONV)-leave-ATTR I TOP I leave today to be an unworthy shield of the sovereign (MYS 20.4373) A2: Ryukyuan In Old Ryukyuan, as far as I can tell, the copula to is attested only once: in OS 13.854. All other examples come from Classical Ryukyuan Ryūka and Kumiodori texts that were influenced much more by the mainland Japanese language and poetry patterns than the Old Ryukyuan texts in the Omoro sōshi. A comprehensive study by Nohara Mitsuyoshi (1998) reveals that Ryukyuan tu as a copula appears only in the Northern and Central Ryukyus. There are no traces of it

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

490

Chapter 6

to the south of Kumejima and the Okinawan islands. Even in the majority of the Northern and Central Ryukyuan languages surveyed by Nohara, the only occurrence of the copula tu seems to be limited to the usage after adjectival stems, and if this limitation was not enough, in most cases it is limited to reduplicated adjectival stems, like Nase magimagii tu ‘being extremely big’ (Nohara 1998: 65), see the data on other dialects in Nohara 1998: 207, 275, 297. 2.4.2.2 Attributive Form t-u The attributive form t-u is very rare in Western Old Japanese. To the best of my knowledge, the following list of examples is exhaustive (excluding controversial cases): 波毘呂由都麻都婆岐

pa-m-birǝ yu t-u ma-tumbaki leaf-GEN-broad sacred DV-ATTR INT-camellia a true sacred camellia with broad leaves (KK 57) 意富美夜能袁登都波多傅須美加多夫祁理

opǝ-miya-nǝ wotǝ t-u pataⁿde sumi katambuk-er-i great-place-GEN that DV-ATTR edge(?) corner incline-PROG-FIN The edge corners of that side of the great palace are falling apart (KK 105) 等富都比等末都良能加波尓和可由都流

tǝpo t-u pitǝ matu-[u]ra-nǝ kapa-ni waka [a]yu tur-u distant DV-ATTR person Matu-ura-GEN river-LOC young sweetfish angle-FIN people from far away (lit.: distant people) angle young sweetfish at the Matu-ura river (MYS 5.857) 乎登都日毛昨日毛今日毛由吉能布礼礼婆

wotǝ t-u PI mo KINƏPU mo KEPU mo yuki-nǝ pur-er-e-mba that DV-ATTR day FP yesterday FP today FP snow-GEN fall-PROG-EV-CON when the snow has been falling the day before yesterday, yesterday, and today (MYS 17.3924) 登保都比等

tǝpo t-u pitǝ distant DV-ATTR person people from far away (lit.: distant people) (MYS 17.3947)

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

Verbs

491

之許都於吉奈

sikǝ t-u okina stupid DV-ATTR old man stupid old man (MYS 17.4011) 大伴乃遠都神祖乃其名乎婆大來目主等於比母知弖

OPƏTƏMƏ-nǝ TƏPO t-u KAMU-OYA-nǝ SƏNƏ NA-womba OPƏKUMƐNUSI tǝ op-i-mǝt-i-te Opǝtǝmǝ-GEN distant DV-ATTR deity-ancestor-GEN that name-ACC(EMPH) Opǝkumɛnusi DV bear-CONV-hold-CONV-SUB Opǝtǝmǝ clan, bearing the name of [our] distant divine ancestor, as Opǝkumɛnusi (MYS 18.4094) 高津神乃災高津鳥乃災

TAKA t-u KAMÏ-nǝ WAⁿZAPAPI TAKA t-u TƏRI-nǝ WAⁿZAPAPI high DV-ATTR deity-GEN calamity high DV-ATTR bird-GEN calamity calamities from deities high [in Heaven], calamities from birds high [in the sky] (NT 10) Comparative Data Level A: Other Japonic A1: Eastern Old Japanese There is only one example of the attributive form t-u in Eastern Old Japanese: 波奈都豆麻

pana t-u tuma14 flower DV-ATTR wife flowery wife (MYS 14.3370) Level B: External Comparisons The fact that the defective verb tǝ has a limited usage in Western Old Japanese, both qualitatively and especially quantitatively, strongly suggests that it is a kind of anomaly. This is further supported by the fact that it is essentially a doublet of the defective verb n- ‘to be.’ The existence of doublets in a language normally indicates that one of them must be a loan, and it seems that, out of

14  Although the character 豆 /ⁿdu/ is used in most manuscripts, making this example somewhat questionable, the Hanawa-bon manuscript has pana t-u tuma with a clear voiceless initial /t-/ (Mizushima 1983: 470).

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

492

Chapter 6

the two, tǝ would be the prime candidate for a loan, especially given its even more limited distribution in Eastern Old Japanese and Ryukyuan. I have surveyed the case of tǝ elsewhere in greater detail (Vovin 2010: 73–77), demonstrating that it is an early loan from some variety of Old Korean. The borrowing possibly occurred in the Kofun period (prior to the lenition *-t- > -l- [-r-] in Korean) into the predecessor of Western Old Japanese: MK ilwo- ‘be’ < OK *ito > WOJ tǝ. 2.4.3 Defective Verb rǝ ‘To Be’ The defective verb rǝ ‘to be’ is most frequently attested in combination with the following emphatic particle kamǝ (or once with its variant kamu), which prompted some Japanese scholars to treat it as a single exclamatory particle rǝkamǝ (Takagi et al. 1957: 39). However, there are two examples when rǝ is attested without the following kamǝ (see MYS 4.654 and MYS 8.1548 below). In addition, the emphatic particle kamǝ is well attested by itself, therefore it can be segmented from the sequence rǝkamǝ. The defective verb rǝ is attested after the uninflected adjective wosǝ ‘hasty, rush,’ the attributive form -ki of inflected adjectives, and after nouns. It is quite clear from the latter usage that the function of rǝ must be that of a copula; otherwise we end up with a sentence that lacks a copula part of its nominal predicate. Since the particle kamǝ always triggers the change of the preceding final form into an attributive, we should treat the defective form rǝ as an attributive form in most cases. However, there is one clear case (in MYS 4.654) when it can only be analyzed as a final form. 宇斯呂傅波袁陀弖呂迦母

usirǝ-ⁿ-te pa woⁿ-tate rǝ kamǝ back-GEN-place TOP DIM-shield DV(ATTR) EP [Her] back is [like] a small shield! (KK 42) 芝賀波能比呂理伊麻須波淤富岐美呂迦母

si-ŋga pa-nǝ pirǝr-i-imas-u opǝ kimi rǝ kamǝ it-POSS leaf-GEN be.broad-CONV-HON-FIN great lord DV(ATTR) EP its leaves are broad, as the great lord! (KK 57) 賣杼理能和賀意富岐美能淤呂須波多他賀多泥呂迦母

Meⁿdǝri n-ǝ wa-ŋga opǝ kimi-nǝ or-ǝs-u pata ta-ŋga tane rǝ kamǝ Meⁿdǝri DV-ATTR I-POSS great lady-GEN weave-HON-ATTR fabric who-POSS material DV(ATTR) EP The fabric my lady MeNdǝri weaves, I wonder for whom (lit. whose) the material is? (KK 66)

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

493

Verbs 微能佐加理毘登登母志岐呂加母

mï-nǝ sakar-i-m-bitǝ tǝmǝsi-ki rǝ kamǝ body-GEN bloom-NML-DV(ATTR)-person envious-ATTR DV(ATTR) EP How [I] am envious of the people who are in their prime! (KK 95) 耆瀰破介辞古耆呂介茂

kimi pa kasiko-ki rǝ kamo lord TOP awesome-ATTR DV(ATTR) EP [my] lord is awesome! (NK 47) 處女之友者乏吉呂賀聞

WOTƏME-ŋGA TƏMƏ TƏMƏSI-ki rǝ kamo maiden-POSS companion envious-ATTR DV(ATTR) EP [I] am envious of maidens’ companions! (MYS 1.53) 更經見者悲呂可聞

KAPAR-AP-U MI-RE-mba KANASI-KI rǝ kamo change-ITER-ATTR see-EV-CON sad-ATTR DV(ATTR) EP when [I] see how [they] change, [it] is sad! (MYS 3.478) 相見者月毛不經尓戀云者乎曽呂登吾乎於毛保寒毳

API-MI-TE pa TUKÏ mo PƐ-N-AKU n-i KOP-U TƏ IP-Amba wosǝ rǝ tǝ WARE-wo omop-as-am-u kamǝ COOP-see(CONV)-SUB TOP month FP pass-NEG-NML DV-CONV long.for-FIN DV say-COND hasty DV(FIN) DV I-ACC think-HON-TENT-ATTR EP If [I] tell [you] that [I] long for [you] without even a month passing since [we] met each other, I wonder [whether you] would think that [I] am hasty (MYS 4.654) 久志美多麻伊麻能遠都豆尓多布刀伎呂可儛

kusi mi-tama ima-nǝ wotutu-ni taputo-ki rǝ kamu mysterious HON-stone now-GEN reality-LOC awesome-ATTR DV(ATTR) EP [these] mysterious stones are awesome in the present [day’s] reality! (MYS 5.813) 咲花毛乎曽呂波恃

SAK-U PANA mo wosǝ rǝ pa itǝpasi bloom-ATTR flower FP hasty DV(ATTR) TOP be.unpleasant(FIN) The fact that blooming flowers are in haste [to fall], too, is unpleasant (MYS 8.1548)

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

494

Chapter 6

Comparative Data Level B: External Comparisons Limited distribution in Japonic as well as limited functionality in Western Old Japanese alongside with the fact that no native Japonic word can start with an initial /r-/ strongly suggest that WOJ rǝ is a short-lived loan. Its source is probably the same as the defective verb tǝ ‘to be,’ described above: OK *ito‘to be’ > MK ilwo-. Certainly, WOJ rǝ must be a later loan from Korean than WOJ tǝ, which occurred after the lenition *-t- > -l- [-r-] took place in the history of the Korean language. It can probably be pinned down to the last great wave of Korean loanwords which should be associated with a large influx of immigrants from the Korean peninsula after the fall of Paekche in 660 AD and Koguryo in 668 AD. This later nature of borrowing probably also explains the fact why WOJ rǝ has never been fully assimilated in the language and why it turned out to be short-lived, while WOJ tǝ actually turned out to be a successful survivor that remains in the Japanese language even today. 2.4.4 Defective Verb tǝ ‘To Say’ The defective verb tǝ ‘to say’ is attested in three forms: the converb tǝ, the final tǝ, and the subordinative converb tǝ-te. The latter form is attested only in one example, so it may be questionable. There is also one example when tǝ can be treated as an attributive because it is found after the focus particle sǝ (see the example from NR 2.33 in section 2.4.4.2). 2.4.4.1 Converb Form tǝ The converb form tǝ in Western Old Japanese is predominantly found in the position introducing quotation clauses preceding verbs of verbal or mental activity, such as ip- ‘to say,’ nǝr- ‘to say, to name,’ kik- ‘to hear,’ omǝp- ‘to think,’ tǝp- ‘to ask,’ etc., although there are also cases when it is found without any following verbs of verbal or mental activity, which will be surveyed below. This structural peculiarity is well supported typologically by other languages of the ‘Altaic’ Sprachbund, where the verbs of verbal and mental activity are preceded by rudimentary verbs introducing quotations, e.g., in Manchu: ume fudara-ka hūlha-de aisila-ra se-me ulhi-bu-me NEG/IMP rebel-PERF bandit-DAT help-TENT/ATTR say-CONV understandCAUS-CONV [Jaohūi] made [the people] understand that [they] should not help the rebels and in Khalkha Mongolian:

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

495

Verbs

margaash yaw-an ge-j xel-sen tomorrow go-PRES say-SUB speak-PAST [He] said that he would go tomorrow There is no difference in Western Old Japanese between direct and indirect speech. Contrary to the claim of Frellesvig, who maintains that tǝ ‘to be’ and tǝ ‘to say’ etymologically represent the same word, namely the copula (Frellesvig 1999), the syntactic difference between those two appears to be fundamental: while the copula tǝ ‘to be’ follows nominals and nominalized forms of verbs, the defective verb tǝ ‘to say’ is almost always found after the final form of verbs (examples of the exceptions are provided below). Below I provide the examples when the defective verb tǝ ‘to say’ is followed by verbs of verbal or mental activity. 故志能久邇邇佐加志売遠阿理登岐加志弖久波志売遠阿理登岐許志弖

Kosi-nǝ kuni-ni sakasi me-wo ar-i tǝ kik-as-i-te kupasi me-wo ar-i tǝ kik-ǝs-i-te Kosi-GEN province-LOC wise woman-ABS exist-FIN DV hear-HON-CONV-SUB beautiful woman-ABS exist-FIN DV hear-HON-CONV-SUB [Opo kuni nusi] heard that there is a wise woman in the Kosi province, heard that there is a beautiful woman (KK 2) 那迦士登波那波伊布登母

nak-aⁿzi tǝ pa na pa ip-u tǝmǝ weep-NEG/TENT DV TOP you TOP say-FIN CONJ Even though you say that [you] would not weep (KK 4) 佐泥牟登波阿禮波意母閇杼

sa ne-m-u tǝ pa are pa omǝp-ɛ-ⁿdǝ thus sleep-TENT-FIN DV TOP I TOP long.for-EV-CONC Although I long so much to sleep [with you] (KK 27) 阿軻娜磨廼比訶利播阿利登比登播伊珮耐

aka-ⁿ-tama-nǝ pikari pa ar-i tǝ pitǝ pa ip-ɛ-ⁿdǝ red-DV(ATTR)-jewel-GEN light TOP exist-FIN DV person TOP say-EV-CONC Although people say that the red jewel has light (NK 6)

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

496

Chapter 6

不聴跡雖謂話礼話礼常詔許曾志斐伊波奏強話登言

INA tǝ IP-Ɛ-ⁿDƏ KATAr-e KATAr-e tǝ NƏR-AS-E kǝsǝ Sipï-i pa MAWOS-E SIPÏ-ŋ-GATAR-I tǝ NƏR-U no DV say-EV-CONC speak-IMP speak-IMP DV say-HON-EV FP Sipï-ACT TOP say(HUM)-EV forced-DV(ATTR)-say-NML DV say-FIN Though [I] say: ‘No,’ [you] command [me]: ‘Speak, speak!,’ [but the things that] Sipï says, [you] call a forced speech (MYS 3.237) In this example the last tǝ follows the nominalized form katar-i of the verb katar- ‘to speak.’ This is a rare usage and could be explained by the ellipsis of the final form of the copula after the nominalized form of the verb. 余能奈可波牟奈之伎母乃等志流等伎子伊与余麻須万須加奈之可利家理

yǝ-nǝ naka pa munasi-ki mǝnǝ tǝ sir-u tǝki si iyǝyǝ masu-masu kanasi-k-ar-i-ker-i world-GEN inside TOP empty-ATTR thing DV know-ATTR time EP more.and. more more.and.more sad-CONV-exist-CONV-RETR-FIN When [I] realized that the world is empty, [it] turned out to be more and more sad (MYS 5.793) In this example tǝ follows the noun mǝnǝ ‘thing,’ but similar to the example above, it could be explained by the ellipsis of the final form of the copula after the noun. 等伎波奈周迦久斯母何母等意母閇騰母

tǝk[ǝ]-ipa-nasu ka-ku si mǝŋgamǝ tǝ omǝp-ɛ-ⁿdǝmǝ eternal-rock-COMP thus-CONV EP DP DV think-EV-CONC Although [I] think that [I] want to be (thus) like an eternal rock … (MYS 5.805) In this example tǝ follows the desiderative particle mǝŋgamǝ that itself functions syntactically as a predicate (“I want, I wish”). 安礼乎於伎弖人者安良自等富己呂倍騰

are-wo ok-i-te PITƏ pa ar-aⁿzi tǝ pokǝr-ǝp-ɛ-ⁿdǝ I-ACC leave-CONV-SUB person TOP exist-NEG/TENT DV boast-ITER-EV-CONC although [I] repeatedly boast that there are probably no other persons besides me (MYS 5.892) 伊等乃伎提短物乎端伎流等云之如

itǝ nǝkite MIⁿZIKA-KI MƏNƏ-wo PASI kir-u tǝ IP-ER-U-ŋGA ŋGƏTƏ-KU very exceptionally short-ATTR thing-ACC end cut-FIN DV say-PROG-ATTRPOSS like-CONV like [the proverb] says: ‘to cut the end of an already very short thing’ (MYS 5.892)

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

497

Verbs 出波之利伊奈奈等思騰許良爾佐夜利奴

IⁿDE-pasir-i in-ana tǝ OMƏP-Ɛ-ⁿdǝ kǝ-ra-ni sayar-i-n-u exit(CONV)-run-CONV go.away-DES DV think-EV-CONC child-PLUR-DAT be.prevented-CONV-PERF-FIN although [I] think that [I] would like to run out and go away, [I] am prevented by [my] children (MYS 5.899) 伊波多野爾夜杼里須流伎美伊敞妣等乃伊豆良等和礼乎等波婆伊可爾伊 波牟

Ipata-NO-ni yaⁿdǝr-i s-uru kimi ipe-m-bitǝ-nǝ iⁿdu-ra tǝ ware-wo tǝp-amba ika n-i ip-am-u Ipata-field-LOC lodge-NML do-ATTR lord home-GEN-person-GEN where-LOC DV I-ACC ask-COND how DV-CONV say-TENT-FIN [Oh, my] lord who lodged at the Ipata field. If people from [your] home ask me (saying) where [are you], what should [I] answer? (MYS 15.3689) 麻佐吉久登伊比低之物能乎

ma-saki-ku tǝ ip-i-te-si MƏnǝwo INT-safe-CONV DV say-CONV-PERF(CONV)-PAST/ATTR CONJ Although [I] have said that [I will return] safely (MYS 17.3958) 伊久欲布等余美都追伊毛波和礼麻都良牟曾

iku yo p-u tǝ yǝm-i-tutu imo pa ware mat-uram-u sǝ how.many night pass-FIN DV count-CONV-COOR beloved TOP I wait-TENT2ATTR FP [My] beloved will probably wait for me, counting: ‘How many nights have passed?’ (MYS 18.4072) 都祢比等能故布登伊敷欲利波安麻里爾弖

tune pitǝ-nǝ kop-u tǝ ip-u-yori pa amari n-i-te ordinary person-GEN long.for-FIN DV say-ATTR-ABL TOP excess DV-CONV-SUB [feeling] much more than what ordinary people call ‘longing’ (MYS 18.4080) 伊爾志加多知与乃都美佐閇保呂夫止曾伊布

in-i-si kata ti yǝ-nǝ tumi sapɛ porǝmb-u tǝ sǝ ip-u go-CONV-PAST/ATTR side thousand life-GEN sin RP disappear-FIN DV FP say-ATTR [they] say that even the sins of one thousand former lives will disappear (BS 17)

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

498

Chapter 6

藤原朝臣麻呂等伊負図亀一頭献止奏賜不爾

PUⁿDIPARA-NƏ ASƏMI MARƏ-RA-i PUMI-WO OP-ER-U KAMƐ-WO PITƏ-TU TATEMATUR-AKU tǝ MAWOS-I-TAMAp-u-ni … Puⁿdipara-GEN retainer Marǝ-PLUR-ACT writing-ACC bear-PROG-ATTR tortoiseACC one-CL offer(HUM)-NML DV say(HUM)-CONV-HON-ATTR-LOC [They] said that the retainer Puⁿdipara Marǝ and others had offered a tortoise bearing writing [on its back] … (SM 6) 国王伊王位仁坐時方菩薩乃浄戒乎受与止勅天在

KOKU-WAU-i WAU-WI-ni IMAS-U TƏKI pa BOSATU-nǝ ZYAUKAI-wo UKƐ-yǝ tǝ NƏTAMAP-I-te AR-I country-king-ACT king-position-LOC be(HON)-ATTR time TOP bodhisattvaGEN commandment-ACC receive-IMP DV say(HON)-CONV-SUB exist-FIN [Buddha] said that a king of a country, when [he] is on the throne, [should] accept commandments of the Bodhisattva (SM 28) 2.4.4.2 Final Form tǝ As mentioned above there are also cases when tǝ ‘to say’ can occur by itself without the following verbs of verbal or mental activity. Unless these cases can be accounted for by ellipsis, we probably should differentiate between the converb form tǝ preceding verbs of verbal and mental activity and the isomorphous final and/or attributive form tǝ that can be used by itself. Some of the following examples are translated with an English gerund form for the sake of the text’s flow, but this should not be taken as proof that the Western Old Japanese forms are really not final. In addition, tǝ ‘to say’ can be followed by the reported action auxiliary nar- (see section 4.2.2.1) that follows final forms of verbs, so this should be additional proof for the existence of the final form tǝ in Western Old Japanese. 阿波志斯袁美那迦母賀登和賀美斯古良迦久母賀登

ap-as-i-si womina ka mǝŋga tǝ wa-ŋga mi-si ko-ra ka-ku mǝŋga tǝ meet-HON-CONV-PAST/ATTR woman that DP DV I-POSS see(CONV)-PAST/ ATTR girl-DIM thus-CONV DP DV [I] think [I] wish that [much] the woman [I] met; [I] think [I] wish this [much] the girl I saw (KK 42) 久良波斯夜麻袁佐賀志美登伊波迦伎加泥弖和賀弖登良須母

kurapasi-yama-wo saŋgasi-mi tǝ ipa kak-i-kane-te wa-ŋga te tǝr-as-umǝ Kurapasi-mountain-ABS steep-GER DV rock hang-CONV-NEG/POT(CONV)SUB take-HON-EXCL

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

499

Verbs

[I] think that Mount Kurapasi is steep. Being unable to cling to the rocks, [I wish you] take my hand! (KK 69) 阿礼乎婆母伊可爾世与等可

are-womba mǝ ika n-i se-yǝ tǝ ka I-ACC(EMPH) FP how DV-CONV do-IMP DV IP What do [you] think I [should] do? (MYS 5.794) 遠等咩良何遠等咩佐備周等可羅多麻乎多母等爾麻可志余知古良等手多豆 佐波利提阿蘇比家武

wotǝme-ra-ŋga wotǝme sambï s-u tǝ kara tama-wo tamǝtǝ-ni mak-as-i yǝti ko-ra-tǝ TE taⁿdusapar-i-te asomb-i-k-em-u maiden-PLUR-POSS maiden like do-FIN DV China jewel-ACC wrist-LOC wrapHON-CONV same.age child-PLUR-COM hand.hold-CONV-SUB play-CONVPAST/FIN-TENT-FIN The maidens, thinking to behave like maidens, wrap their wrists with [bracelets made of] Chinese jewels, and would play holding hands with girls of the same age (MYS 5.804) 伊射祢余登手乎多豆佐波里

iⁿza ne-yǝ tǝ TE-wo taⁿdusapar-i INTER sleep-IMP DV hand-ACC hand.hold-CONV [we] told [him]: ‘Go to sleep!’ [He], taking [our] hands … (MYS 5.904) 如今将相跡奈良婆此篋開勿勤

IMA ŋGƏTƏ AP-AM-U tǝ-nar-amba KƏNƏ KUSI ŋGƐ PIRAK-UNA YUMƐ now like meet-TENT-FIN DV(FIN)-RA-COND this comb.box open-NEG/IMP at.all if [you] say that [you] want to meet me [again] like now, do not open this comb box at all (MYS 9.1740) 和伎毛故波伊都登加和礼乎伊波比麻都良牟

wa-ŋg-imo-ko pa itu tǝ ka ware-wo ipap-i mat-uram-u I-POSS-beloved-DIM TOP when DV IP I-ACC pray-CONV wait-TENT2-ATTR My beloved will probably wait for me, praying [to the deities], and thinking: ‘When [will he return]?’ (MYS 15.3659)

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

500

Chapter 6

伊都之加登奈気可須良牟曾

itu si ka tǝ naŋgɛk-as-uram-u sǝ when EP IP DV lament-HON-TENT2-ATTR FP [she] probably laments, saying: ‘When [will he return]?’ (MYS 17.3962) 子乃去禍蒙服麻久欲為流事波為親爾止奈利

KO-nǝ WAⁿZAPAPI-WO SAR-I SAKIPAPI-WO KA ŋGAPUR-Am-aku POR-I S-Uru KƏTƏ pa OYA-nǝ tamɛ n-i tǝ-nar-i child-GEN misfortune-ACC go.away-CONV happiness-ACC receive-TENT-NML want-NML do-ATTR matter TOP parent-GEN for DV-CONV DV(FIN)-RA-FIN [They] say that the fact that children want to avoid misfortune and obtain happiness is for the sake of [their] parents (SM 25) 奈禮乎曾與咩爾保師登多禮

nare-wo sǝ yǝme n-i posi tǝ tare you-ACC FP bride DV-CONV be.desirable DV who who says [he] wants you as [his] bride? (NR 2.33) In this example tǝ can only be attributive, as it is preceded by the particle sǝ. Comparative Data Level A: Other Japonic A1: Eastern Old Japanese Similar to Western Old Japanese, the defective verb tǝ in Eastern Old Japanese can also be used before the verbs of verbal or mental activity (converb form) or independently (final or attributive form). Examples of tǝ followed by verbs of verbal and mental activity: 麻許登可聞和礼爾余須等布

ma-kǝtǝ kamo ware-ni yǝs-u tǝ [i]p-u INT-thing EP I-DAT bring.close-FIN DV say-ATTR I wonder [whether it is] true that [people] say that [she] has an intimate relationship with me (MYS 14.3384) 伊可奈流勢奈可和我理許武等伊布

ika nar-u se-na ka wa-ŋgari kǝ-m-u tǝ ip-u how be-ATTR beloved-DIM IP I-DIR come-TENT-FIN DV say-ATTR what kind of beloved is [he], who says that [he] will come to me? (MYS 14.3536)

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

501

Verbs 和呂多比波多比等於米保等

warǝ tambi pa tambi tǝ omɛp-o-ⁿdǝ I journey TOP journey DV think-EV-CONC Although I think that [my] journey is [just] a journey … (MYS 20.4343) Examples of the independent usage of tǝ: 古奈宜可久古非牟等夜

ko-naŋgï ka-ku kopï-m-u tǝ ya DIM-water.hollyhock thus-CONV long.for-TENT-FIN DV IP Do [you] think [I] will long so [strongly] for a small water-hollyhock? (MYS 14.3415) 奴流我倍爾安杼世呂登可母

n-uru-ŋga [u]pɛ-ni aⁿ-tǝ se-rǝ tǝ kamǝ sleep-ATTR-POSS top-LOC what-DV do-IMP DV EP besides sleeping [with her], I wonder what [else] do [you] say [I] should do? (MYS 14.3465) Since this tǝ precedes kamǝ it can only be treated as an attributive. 古非爾思奈武乎伊可爾世余等曾

kopï n-i sin-am-u-wo ika n-i se-yǝ tǝ sǝ long.for(NML) DV-CONV die-TENT-ATTR-ACC how DV-CONV do-IMP DV FP although [I] will die from longing, what (lit.: how) will [you] tell [me] to do? (MYS 14.3491) There is also a special Eastern Old Japanese form te of this defective verb that is attested only once: 知々波々我可之良加伎奈弖佐久安礼天伊比之氣等婆是和須礼加祢豆流

titi papa-ŋga kasira kaki-naⁿde sa-ku ar-e te ip-i-si kɛtǝmba ⁿze wasure-kane-t-uru father mother-POSS head PREF-stroke(CONV) safe-CONV exist-IMP DV sayCONV-PAST/ATTR word FP forget(CONV)-NEG/POT(CONV)-PERF-ATTR [I] cannot forget the words: “Be safe!” that [my] father and mother said, stroking [my] head (MYS 20.4346)

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

502

Chapter 6

A2: Ryukyuan The reflexes of the defective verb tǝ are well presented in all Ryukyuan dialects. The modern Shuri form is said to be Ndi, but this form apparently represents a reanalysis of verb final -N + defective verb *ti as a single morpheme. After the final verb ending in -N one normally finds just di, not Ndi. The form di is probably from *ti that underwent voicing assimilation under the influence of the preceding /N/. Examples: Shuri

nuu Ndi ’yu-ta-ga what DV say-PAST-Q What (DV) did [you] say? (RKJ 435) ’ic-uN di ’yu-ta-N go-FIN DV say-PAST-FIN [He] said that [he] would go (RKJ 435) 2.4.4.3 Subordinative Converb Form tǝ-te The subordinative converb form tǝ-te is a kind of puzzle. In the whole Western Old Japanese corpus it is attested only once: 此七日爾波不足弖隠坐事奇止弖見所行須時

KƏNƏ NAN-UKA-ni pa TAR-AⁿZ-U-te KAKUR-I-[I]MAS-U KƏTƏ AYA-SI tǝ-te MI-SONAP-As-u TƏKI this seven-CL-LOC TOP be.enough-NEG-CONV-SUB hide-CONV-HON-ATTR matter strange-FIN DV(CONV)-SUB look(CONV)-offer(HON)-HON-ATTR time When these seven days did not completely pass, [he] thought that [it] is strange that [she] secluded herself, and when [he] looked … (NT 12) Yamada Yoshio believed that this single example of tǝ-te is due either to the fact that this Norito is itself a later text, or to a later corruption of the text (Yamada 1954: 452). While we obviously cannot exclude the possibility of text corruption, this Norito certainly represents an Early Western Old Japanese text, as demonstrated by Bentley (2001: 25–26). Japanese linguists normally explain tǝ-te as a contraction from tǝ ip-i-te ‘DV say-CONV-SUB,’ but as Bentley correctly notes, the explanation should be rejected, because there was no intervocalic -ploss even in Late Western Old Japanese (Bentley 2001: 137). I should also add that the form to-te is frequently observed in Middle Japanese texts as well, and

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

Verbs

503

although -p- > -w- occurred by this time, the further shift -w- > -∅- before /i/ did not occur before the Kamakura period except in the verb ma(w)ir- ‘to come/ go(HUM)’ (Tsukishima 1969: 360–363). Thus, even MJ to-te < to iwite would be speculative. Therefore, although I normally would not accept a hapax legomenon as evidence, I am inclined to view the rarity of WOJ tǝ-te as a trick played on us by the existing corpus. Ryukyuan data that I briefly discuss below also indicates that the subordinative converb form tǝ-te may go back to Proto-Japonic. Comparative Data Level A: Other Japonic A2: Ryukyuan There are many attestations of the subordinative converb form tete (= tote) appearing in Old and Classical Ryukyuan texts in various spelling forms. These different spellings exhibit a bewildering variety: tete, teti, dete, reiti (= tote) (Hokama 1995: 441). The absence of a variant *tote suggests that they are not loans from mainland Middle Japanese. Therefore, it is likely that WOJ tǝ-te and these Old Ryukyuan forms go back to a proto-Japonic formation, although one must support it with data from Southern Ryukyuan that so far I have been unable to find.

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

Section 3

Verbal Affixes Verbal affixes in Western Old Japanese comprise prefixes and suffixes. There is also one circumfix na-…-sǝ, but functionally it represents a variant of the prefix na-, so both will be treated together below in section 3.1.2. The majority of affixes are suffixes. 3.1

Verbal Prefixes

One of the most striking differences between Western Old Japanese and Middle (Classical) Japanese is that the former has a comparatively rich system of verbal prefixes. While there are verbal prefixes in Middle Japanese as well, they all are preverbs either synchronically or diachronically, in other words they all have a more or less transparent verbal origin. Not so in Western Old Japanese, where half of the prefixes cannot be traced back to any verbal forms. The existence of prefixes in a SOV language is, of course, an anomaly, as SOV languages normally have only suffixes. The presence of even a rudimentary prefixation system, therefore, may point to the fact that the SOV word order in Old Japanese may be a comparatively recent phenomenon due to the ‘Altaicization’ and ‘Koreanization’ of Japonic. In most cases, a verbal form contains only one prefix, but there are three exceptions where two prefixes can be found in a verbal form. In two out of the three exceptions the prefix ta- occupies the second slot, and the prefix i- occupies the first slot. Examples: 朝奈藝尓伊可伎渡

ASA naŋgi-ni i-kaki-WATAR-I morning calm-LOC DLF-PREF-cross.over-CONV crossing over there in the morning calm (MYS 8.1520) 麻佐吉久毛安里多母等保利

ma-saki-ku mo ari-ta-mǝtǝpor-i INT-safe-CONV FP ITER-PREF-wander.around-CONV so that [you] will be wandering around safely, and … (MYS 17.4008)

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

505

Verbs 乎可乃佐伎伊多牟流其等尓

woka n-ǝ saki i-ta-m-uru ŋgǝtǝ n-i hill DV-ATTR cape DLF-PREF-turn.around-ATTR every DV-CONV every time [I] turn around a hilly cape (MYS 20.4408) 3.1.1 Prefix iThe prefix i- indicates directive-locative focus for a verb. In the previous literature this prefix is only briefly mentioned and virtually left undescribed. Thus, for example, Yamada Yoshio briefly states that this and other prefixes have only rhythmic function, and provides a list of examples which is very far from being complete, since all examples that Yamada Yoshio cites are from the Man’yōshū (Yamada 1954: 530–532). None of the numerous examples listed below from the Kojiki kayō or the Nihonshoki kayō are included in his grammar. The only studies that dedicate some attention to this prefix are Hino 1997, Russell (2006: 140–142), and Yanagida and Whitman 2009. Hino argues that the prefix i- is an agentive marker (1997: 2–5) while Russell defines this prefix as a focus marker indicating the goal of the predicate (Russell 2006: 142). She lists a few (but not all) attestations, but gives only two examples (Russell 2006: 141–142). I trust that the definition of the prefix i- as a goal focus marker is too vague. Yanagida and Whitman 2009 argue that i- indicates ergative alignment. Their argument is certainly the most solid among all three hypotheses that are alternatives to the one advanced here, and although I do not accept it, it would take a separate article to refute their claim, which is not in the scope of this grammar. I provide below most of the examples of this prefix that I was able to locate in Western Old Japanese texts, excluding the textual variants of the Nihonshoki kayō songs if they are also found in the Kojiki kayō, as well as some examples from the Man’yōshū that include contexts close to other already cited examples. I believe that the following examples allow narrowing the definition of this prefix to a focus marker indicating the direction or location of an action. The prefix i- occurs only with a limited number of Western Old Japanese verbs: kakur- kaper- kǝŋg- kǝⁿz(i)- kum- kir- mak- mure- pap-

‘to be hidden’ ‘to return’ ‘to row’ ‘to dig out’ ‘to entwine’ ‘to cut’ ‘to whirl, to roll’ ‘to gather’ ‘to crawl’

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

506 pate- puk- sak- sik- sop- ta-mï-15 taⁿdor- tukus- tumor- tuŋg- tor- watar- yǝr- yuk-

Chapter 6

‘to anchor’ ‘to blow’ ‘to bloom’ ‘to reach’ ‘to snuggle’ ‘to turn’ ‘to seek’ ‘to exhaust’ ‘to pile up’ ‘to continue’ ‘to hold’ ‘to cross’ ‘to lean at’ ‘to go’

I tried to convey the directional-locative meaning of the prefix i- by inserting English ‘here’ or ‘there’ wherever it was possible in the translations of the examples below. 伊勢能宇美能意斐志爾波比母登富呂布志多陀美能伊波比母登富理宇知弖 志夜麻牟

ise-nǝ umi-nǝ opï-[i]si-ni pap-i-mǝtǝpǝr-ǝp-u sitaⁿdami16-nǝ i-pap-i-mǝtǝpǝr-i ut-i-te si yam-am-u Ise-GEN sea-GEN grow(CONV)-stone-LOC crawl-CONV-go.around-ITER-ATTR seashell-COMP DLF-crawl-CONV-go.around-CONV hit-CONV-SUB EP stopTENT-FIN like the shellfish that are constantly crawling around on the growing rocks of the Ise sea, [we] will crawl around [them] there, smite and stop [them] (KK 13) 伊那佐能夜麻能許能麻用母伊由岐麻毛良比

Inasa-nǝ yama-nǝ kǝ-nǝ ma-yo mǝ i-yuk-i mamor-ap-i Inasa-GEN mountain-GEN tree-GEN interval-ABL FP DLF-go-CONV watchITER-CONV going there from between the trees of Mount Inasa and watching [out for enemies] constantly (KK 14)

15  With the verb mï- ‘to turn’ the prefix i- is found only preceding the prefix ta-. 16  志多陀美 /sitaⁿdami/, a kind of an edible seashell (MdJ kisago). Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

507

Verbs 斯理都斗用伊由岐多賀比麻弊都斗用伊由岐多賀比

siri-tu to-yo i-yuk-i-taŋgap-i mape-tu to-yo i-yuk-i-taŋgap-i back-GEN/LOC door-ABL DLF-go-CONV-differ-CONV front-GEN/LOC doorABL DLF-go-CONV-differ-CONV going there from different [directions], from the front door, [and] from the back door (KK 22) 牟迦比袁流迦母伊蘇比袁流迦母

mukap-i-wor-u kamǝ i-sop-i-wor-u kamǝ face-CONV-exist-ATTR EP DLF-snuggle-CONV-exist-ATTR EP Oh, [she] is facing [me]! Oh, [she] is snuggling here with [me]! (KK 42) 麻由美伊岐良牟登許許呂波母閇杼伊斗良牟登許許呂波母閇杼

mayumi i-kir-am-u tǝ kǝkǝrǝ pa [o]mǝp-ɛ-ⁿdǝ i-tor-am-u tǝ kǝkǝrǝ pa [o] mǝp-ɛ-ⁿdǝ mayumi DLF-cut-TENT-FIN DV heart TOP think-EV-CONC DLF-take-TENT-FIN DV heart TOP think-EV-CONC although [I] wish [in my] heart to cut the mayumi trees here, although [I] wish [in my] heart to take [them] here (KK 51) 伊岐良受曾久流

i-kir-aⁿz-u sǝ k-uru DLF-cut-NEG-CONV FP come-ATTR [I] return without cutting [them] there (KK 51) 夜麻斯呂迩伊斯祁登理夜麻伊斯祁伊斯祁阿賀波斯豆摩迩伊斯岐阿波牟 迦母

Yamasirǝ-ni i-sik-e Tǝriyama i-sik-e i-sik-e a-ŋga pasi-ⁿ-duma-ni i-sik-i ap-am-u kamǝ Yamasirǝ-LOC DLF-follow-IMP Tǝriyama DLF-follow-IMP DLF-follow-IMP I-POSS beloved-DV(ATTR)-spouse-DAT DLF-follow-CONV meet-TENT-ATTR EP To Yamasirǝ—go there, Tǝriyama! Follow [her] there, follow [her] there! Follow my beloved spouse there and meet [her]! (KK 59) 意富岐美袁斯麻爾波夫良婆布那阿麻理伊賀弊理許牟叙

opǝ kimi-wo sima-ni pambur-amba puna-amar-i i-ŋgaper-i-kǝ-m-u ⁿzǝ great lord-ACC island-LOC exile-COND boat-exceed-CONV DLF-returnCONV-come-TENT-ATTR FP if [they] exile [my] great lord to an island, there are many boats, and [he] would return here (KK 86) Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

508

Chapter 6

母登爾波伊久美陀気淤斐須惠幣爾波多斯美陀気淤斐伊久美陀気伊久美波 泥受

mǝtǝ-ni pa i-kum-i-ⁿ-dakɛ opï suwe-pe-ni pa ta-sim-i-n-dakɛ opï i-kum-i-n-dakɛ i-kum-i pa ne- nz-u root-LOC TOP DLF-entwine-NML-DV(ATTR)-bamboo grow(CONV) top-sideLOC TOP PREF-grow.densely-NML-DV(ATTR)-bamboo grow(CONV) DLFentwine-NML-DV(ATTR)-bamboo DLF-entwine-CONV TOP sleep-NEGCONV at [their] roots grows bamboo entwined there, at [their] tips grows a dense bamboo; we did not sleep entwined there as the bamboo entwined there (KK 91) 袁登賣能伊加久流袁加

wotǝme-nǝ i-kakur-u woka maiden-GEN DLF-hide-ATTR hill the hill where a maiden is hiding (KK 99) 夜本爾余志伊岐豆岐能美夜

yapo ni yǝ-si i-kiⁿduk-i n-ǝ miya eight.hundred ground good-FIN DLF-build-NML DV-ATTR palace a palace built there on an eight hundred [times] good soil (KK 100) 和賀淤富岐美能阿佐斗爾波伊余理陀多志由布斗爾波伊余理陀多須和岐 豆紀

wa-ŋga opǝ kimi-nǝ asa-to-ni pa i-yǝr-i-ⁿdat-as-i yupu-to-ni pa i-yǝr-i-ⁿdat-as-u wakiⁿdukï I-POSS great lord-GEN morning-place-LOC TOP DLF-lean-CONV-standHON-CONV evening-place-LOC TOP lean-CONV-stand-HON-ATTR arm-rest the arm-rest that my sovereign leans on in the morning and leans on in the evening (KK 104) 避奈莵謎廼以和多邏素西渡

pina-tu me-nǝ i-watar-as-u se to countryside-GEN/LOC woman-GEN DLF-cross-HON-ATTR narrow passage narrow passage that a woman from the countryside crosses here (NK 3) 瀰既能佐烏麼志魔幣莵耆瀰伊和哆羅秀暮

mi-kɛ-nǝ sawo-m-basi mapetukimi i-watar-as-umo HON-tree-GEN pole-GEN-bridge minister DLF-cross-HON-EXCL [I wish] that the ministers would cross here the bridge [made] of the poles of the sacred trees! (NK 24) Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

509

Verbs 柯彼能矩盧古磨矩羅枳制播伊志歌孺阿羅磨志

Kapï-nǝ kuro koma kura ki-s-emba i-sik-aⁿz-u ar-amasi Kapï-GEN black stallion saddle put.on(CONV)-PAST/ATTR-COND DLF-reachNEG-CONV exist-SUBJ if [he] would put a saddle on the black stallion from Kapï, [he] would not reach here [on time] (NK 81a) 以矩美娜開余嚢開謨等陛嗚麼莒等爾都倶唎

i-kum-i-ⁿ-dakɛ yǝ-ⁿ-dakɛ motǝ-pe-womba kǝtǝ-ni tukur-i DLF-entwine-NML-DV(ATTR)-bamboo section-DV(ATTR)-bamboo root-sideACC(EMPH) koto DV-CONV make-CONV [they] made roots of an entwined bamboo, of section bamboo into a koto, and … (NK 97) 阿箇悟馬能以喩企波波箇屢麻矩儒播羅

aka-ŋ-goma-nǝ i-yuk-i-pambakar-u ma-kuⁿzu-para red-DV(ATTR)-stallion-GEN DLF-go-CONV-be hesitant-ATTR INT-vine-field A red stallion is hesitant to go through the field covered with thick vines (NK 128) 山際伊隠萬代道隈伊積流萬代尓

YAMA-NƏ MA-NI i-KAKUR-U-maⁿde MITI-NƏ KUMA i-TUMOr-u-maⁿde-ni mountain-GEN space-LOC DLF-hide-ATTR-TERM road-GEN bend DLF-pile. up-ATTR-TERM-LOC until [Mt. Miwa] will hide there between the mountains, until the road bends will pile up there … (MYS 1.17) 冬乃林尓飄可毛伊巻渡等念麻弖

PUYU-nǝ PAYASI-ni TUMUⁿZI kamo i-MAK-I-WATAR-U tǝ OMƏP-U-maⁿde winder-GEN forest-LOC whirlwind EP DLF-whirl-CONV-cross-FIN DV thinkATTR-TERM to the point that [we] thought: “Could [it] be a whirlwind that whirls there across the winder forest?” (MYS 2.199) 神風尓伊吹或之

KAMU-KAⁿZE-ni i-PUK-I-MATOP-As-i divine-wind-LOC DLF-blow-CONV-be.confused-CAUS-CONV making the divine wind to blow [them] away in confusion (MYS 2.199)

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

510

Chapter 6

鹿自物伊波比伏管 … 鶉成伊波比廻 SISI ⁿzimǝnǝ i-pap-i-PUS-I-tutu … UⁿDURA-nasu i-pap-i-MƏTƏPOR-I dear like DLF-crawl-CONV-lie.down-CONV-COOR … quail-COMP DLF-crawlCONV-go.around-CONV constantly crawling and lying down like a deer … crawling around like a quail (MYS 2.199) 遠等咩良何佐那周伊多斗乎意斯比良伎伊多度利与利提

wotǝme-ra-ŋga sa-n-as-u ita-to-wo os-i-pirak-i i-taⁿdor-i-yǝr-i-te maiden-PLUR-POSS PREF-sleep-HON-ATTR board-door-ACC push-CONVopen-CONV DLF-pursue-CONV-approach-CONV-SUB [gentlemen] push open wooden doors where maidens sleep, and [they] pursue [maidens] there (MYS 5.804) 伊刀良斯弖伊波比多麻比斯麻多麻奈須布多都能伊斯

i-tor-as-i-te ipap-i-tamap-i-si ma-tama-nasu puta-tu n-ǝ isi DLF-hold-HON-CONV-SUB pray-CONV-HON-CONV-PAST/ATTR INT-jewelCOMP two-CL DV-ATTR stone holding two stones like real jewels that [she] was praying to … (MYS 5.813) 四良名美乃五十開廻有住吉能濱

sira nami-nǝ i-SAK-I-MƐ ŋGUR-ER-U SUMINƏYE-nǝ PAMA white wave-GEN DLF-bloom-CONV-go.around-PROG-ATTR Suminǝye-GEN beach Suminǝye beach, where white waves go around blooming [like flowers] (MYS 6.931) 去年春伊許自而

KƏⁿZƏ-NƏ PARU i-kǝⁿz-i-TE17 last.year-GEN spring DLF-dig.out-CONV-SUB having dug [them] out there in the spring of the last year (MYS 8.1423) 朝奈藝尓伊可伎渡

ASA naŋgi-ni i-kaki-WATAR-I morning calm-LOC DLF-PREF-cross.over-CONV crossing over there in the morning calm (MYS 8.1520) 17  It is not clear whether we are dealing here with a consonant verb kǝⁿz- or a vowel verb kǝⁿzi-, because among its basic paradigmatic forms only the converb is attested (Omodaka et al. 1967: 294; Takagi et al. 1959: 283). Since there are more consonant verbs than vowel verbs, I tentatively adopt the interpretation that this verb is a consonant one. Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

511

Verbs 左小舟乃伊行而将泊河津

sa-WOⁿ-BUNE-nǝ i-YUK-I-TE PATE-M-U KAPA-ⁿ-DU PREF-DIM-boat-GEN DLF-go-CONV-SUB anchor-TENT-ATTR river-GEN-harbor the river harbor, where [his] small boat will go and anchor (MYS 10.2091) 左壮鹿之聲伊續伊継

sa-WO-SIKA-NƏ KƏWE i-TU ŋG-I i-TU ŋG-I PREF-male-deer-GEN voice DLF-continue-CONV DLF-continue-CONV a call from the male deer continues there, continues there (MYS 10.2145) 左夫流其兒尓比毛能緒能移都我利安比弖

samburu SƏNƏ KO-ni pimo-nǝ wo-nǝ i-tuŋgar-i-ap-i-te whore that girl-DAT cord-GEN cord-COMP DLF-tie-CONV-join-CONV-SUB [you] tied [yourself] together with that whore girl like a cord of cords (MYS 18.4106) 四方能美知尓波宇麻乃都米伊都久須伎波美布奈乃倍能伊波都流麻泥尓

YƏ MO-nǝ miti-ni pa uma-nǝ tumɛ i-tukus-u kipami puna-nǝ pɛ-nǝ i-paturu-maⁿde-ni four direction-GEN road-LOC TOP horse-GEN hoof DLF-exhaust-ATTR limit boat-GEN prow-GEN DLF-anchor-ATTR-TERM-LOC on the roads to four directions, [to the] limit where horses’ hooves reach, to the point where a boat’s prow [could be] anchored (MYS 18.4122) 新年始尓思共伊牟礼氐乎礼婆

ARATASI-KI TƏSI-NƏ PA ᷠZIMƐ-ni OMƏP-U-ⁿ-DOTI i-mure-te wor-e-mba new-ATTR year-GEN beginning-LOC think-ATTR-GEN-companion DLF-gather (CONV)-SUB exist-EV-CON When the friends who think [in the same way] are gathering at the beginning of the year … (MYS 19.4284) 乎可乃佐伎伊多牟流其等尓

woka n-ǝ saki i-ta-m-uru ŋgǝtǝ n-i hill DV-ATTR cape DLF-PREF-turn.around-ATTR every DV-CONV every time [I] turn around a hilly cape (MYS 20.4408) 之麻豆多比伊己藝和多利弖

sima-ⁿ-dutap-i i-kǝŋg-i-watar-i-te island-LOC-pass.along-CONV DLF-row-CONV-cross-CONV-SUB [I] row across [the ocean] passing along islands (MYS 20.4408)

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

512

Chapter 6

Comparative Data Level A: Other Japonic A1: Eastern Old Japanese There is only one example of the prefix i- in Eastern Old Japanese: 伊波能倍尓伊可賀流久毛

ipa-nǝ [u]pɛ-ni i-kakar-u kumo rock-GEN top-LOC DLF-hang-ATTR cloud clouds, hanging over the rocks (MYS 14.3518) A2: Ryukyuan To the best of my knowledge the prefix i- is attested only in Old and Classical Ryukyuan. In spite of the fact that there is no support from modern Ryukyuan languages, it is unlikely that OR i- represents a loan from WOJ i-, as there were no contacts between Western Old Japanese and Old Ryukyuan. Hokama claims that in Old and Classical Ryukyuan i- could be attached to adjectives and nouns as well (Hokama 1995: 53), but his examples with nouns are likely to include historically a different kind of i: namely, an adjective i ‘sacred, tabooed’ also attested in Western Old Japanese. There is only one example of i- combined with an adjective: i-duyo-ku ~ i-diyo-ku ‘PREF-strong-CONV’ attested exclusively in Classical Ryukyuan (Hokama 1995: 74–75). This limited attestation is unlikely to have an old provenance. Old Ryukyuan としが三年いきよてとしが四年いきよて

tosi-ga SAN-NEN i-kiyo-te tosi-ga YO-NEN i-kiyo-te year-POSS three-year DLF-invite-SUB year-POSS four-year DLF-invite-SUB inviting [the goddess] here for three years, inviting [the goddess] here for four years (OS 12.658) There is also another example cited by Hokama from Classical Ryukyuan with the verb fasir- ‘to run’ (Hokama 1995: 82), but since I do not have access to the text it is attested in, I do not cite this example here. 3.1.2 Prefix na- and Circumfix na-…-sǝ In contrast to Middle Japanese, where only the circumfix na-…-sǝ is found (Vovin 2003: 195), in Western Old Japanese there are examples containing only the first element na- that can be consequently treated as a prefix. Both the prefix na- and the circumfix na-…-sǝ function as markers of the negative imperative.

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

513

Verbs

The circumfix occurs much more frequently than the prefix. Martin’s observation that -sǝ in the circumfix na-…-sǝ represents the root of the verb se- ~ sǝ- ‘to do’ (Martin 1988: 967) is undoubtedly correct. At the same time, I find his definition of na- as a prohibitive adverb (Martin 1987: 489, 920; 1988: 967) much less convincing, because as the reader will see below, WOJ na- is never separated from a verb by any other word, including particles. Thus, I will treat it as a prefix. Examples: 伊波那佐牟遠阿夜爾那古斐岐許志

i pa n-as-am-u-wo aya n-i na-kopï-kikǝs-i sleep TOP sleep-HON-TENT-ATTR-ACC extremely DV-CONV NEG-yearn (CONV)-HON-CONV because [you] will sleep (a sleep), [do] not yearn too much (KK 3) 吾大王物莫御念

WA-ŋG-OPO KIMI MƏNƏ NA-OMƏP-OS-I I-POSS-great lord thing NEG-think-HON-CONV Oh, my sovereign, please do not be concerned … (MYS 1.77) 父母毛表者奈佐我利

TITI-PAPA mǝ UPƐ pa na-saŋgar-i father-mother FP top TOP NEG-go.down-CONV Father [and] mother! [Do] not leave [me] (MYS 5.904) 龍田彦勤此花乎風尓莫落

TATUTA PIKO YUMƐ KƏNƏ PANA-wo KAⁿZE-ni NA-TIR-AS-I Tatsuta male[deity] at.all this flower-ACC wind-LOC NEG-scatter-HON-CONV Male deity of Tatsuta! Please do not scatter these flowers at all by the wind (MYS 9.1748) 雲莫田名引

kumo NA-tanamBIK-I cloud NEG-trail-CONV clouds, do not trail [over the moon] (MYS 11.2669) 安礼奈之等奈和備和我勢故

are na-si tǝ na-wambï wa-ŋga se-ko I no-FIN DV NEG-be.disheartened(CONV) I-POSS elder.brother-DIM Do not be disheartened, my elder brother that I am not [here] (MYS 17.3997)

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

514

Chapter 6

和我世兒乎安宿勿令寐

wa-ŋga se-ko-wo YASU I NA-NE-SIMƐ I-POSS beloved-DIM-ACC easy sleep NEG-sleep-CAUS(CONV) Do not let my beloved sleep an easy sleep (MYS 19.4179) The circumfix na-…-sǝ encircles the converb form of verb, but in the case of the irregular verb se- ~ -sǝ ‘to do’ it encircles the root form se-, see the example from MYS 20.4487 below. Examples of the circumfix na-…-sǝ: 能知波那杼理爾阿良牟遠伊能知波那志勢婆多麻比曾

nǝti pa na-ⁿ-dǝri n-i ar-am-u-wo inǝti pa na-si-se-tamap-i-sǝ later TOP you-OSM-bird DV-CONV exist-TENT-ATTR-ACC life TOP NEG-dieCAUS(CONV)-HON-CONV-do because [I] will be your bird later, do not kill [your] life [with desire] (KK 3) 宇梅能半奈半也久奈知利曽

uMƐ-nǝ pana paya-ku na-tir-i-sǝ plum-GEN blossom early-CONV NEG-fall-CONV-do Plum blossoms! Do not fall early (MYS 5.849) 於伎都風伊多久奈布吉曽

oki-tu KAⁿZE ita-ku na-puk-i-sǝ offing-GEN/LOC wind painful-CONV NEG-blow-CONV-do wind of the offing, do not blow strongly (MYS 15.3592) 許能之具礼伊多久奈布里曽

kǝnǝ siŋgure ita-ku na-pur-i-sǝ this drizzling.rain painful-CONV NEG-fall-CONV-do This drizzling rain! Do not rain hard (MYS 19.4222) 伊射子等毛多波和射奈世曽

iⁿza KO-ⁿdǝmo tapa waⁿza na-se-sǝ INTER child-PLUR stupid deed NEG-do-do Hey, children, do not do stupid things (MYS 20.4487) A desiderative-imperative form -(a)n-e can be used after the negative imperative circumfix na-…-sǝ, but this usage is not very frequent in Western Old Japanese:

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

515

Verbs 雪奈布美曽祢

YUKI na-pum-i-sǝ-n-e snow NEG-step.on-CONV-do-DES-IMP do not step on the snow (MYS 19.4228) 久佐奈加利曽祢

kusa na-kar-i-sǝ-n-e grass NEG-cut-CONV-do-DES-IMP do not cut the grass (MYS 20.4457) Comparative Data Level A: Other Japonic A1: Eastern Old Japanese Both the prefix na- and the circumfix na-…-sǝ are attested in Eastern Old Japanese, but the prefix na- occurs only twice and the circumfix na-…-sǝ only once, being followed by the desiderative-imperative form -(a)n-e. Examples of the prefix na-: 安乎許等奈多延

a-wo kǝtǝ na-taye I-ACC word NEG-break(CONV) Do not break [exchanging] messages with me (MYS 14.3501) 由古作枳尓奈美奈等惠良比

yuk-o saki-ni nami na-tǝwerap-i go-ATTR destination-LOC wave NEG-rise-CONV Waves, do not rise on my way (MYS 20.4385) Example of the circumfix na-…-sǝ: 和爾奈多要曾祢

wa-ni na-taye-sǝ-n-e I-DAT NEG-break(CONV)-do-DES-IMP Do not become estranged from me (MYS 14.3378) 3.1.3 Prefix kaThe prefix ka- is a rare example of a primary morphological marker shared by both adjectives and verbs. Although all but one example of its usage are attested with adjectives, I consider that it is more appropriate to describe the prefix ka- in the chapter dedicated to verbs, because it occurs only with inflected, i.e., verbal forms of adjectives. The underlying form of this prefix may be *kaⁿ-, as Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

516

Chapter 6

witnessed by a form kaⁿ-guro- ‘pitch black’ below. Although this is the only form when ka- occurs before an obstruent, kaⁿ-guro- and not *ka-kuro- is confirmed twice in the texts, making a scribal mistake unlikely. In all other cases which do not involve an initial obstruent of a root we would normally expect a development of *kaⁿ- > ka-. No definition of its function is given in Japanese scholarship (Yamada 1954: 532). As far as I can tell on the basis of the limited examples that are attested in the Western Old Japanese texts, the prefix ka- is an intensifier, indicating that an action or a state is extreme or full/real. Thus, the intensive prefix ka- for verbs and inflected adjectives probably represents an analogue for the intensive prefix ma- used with nouns and uninflected adjectives (see chapter 4, section 1.1.2). The prefix ka- with a verb: 香縁相者彼所毛加人之吾乎事将成

ka-YƏR-I-AP-Amba SƏKƏ mo ka PITƏ WA-wo kǝtǝ NAS-AM-U INT-approach-CONV-meet-COND there FP IP person I-ACC word do(HON)TENT-ATTR if [we] indeed get together, would the people spread rumors about me and you as well? (MYS 4.512) The prefix ka- with inflected adjectives: 荒礒乃上尓香青生玉藻

AR-ISO-nǝ UPƐ-ni ka-AWO-KU OP-URU TAMA MO rough-rock-GEN top-LOC INT-green-CONV grow-ATTR jewel seaweed jewel seaweeds that grow very green on rough rocks (MYS 2.131) 迦具漏伎可美爾伊都乃麻可斯毛乃布利家武

kaⁿ-guro-ki kami-ni itu-nǝ ma ka simo-nǝ pur-i-k-em-u INT-black-ATTR hair-LOC when-CEN interval IP frost-GEN fall-CONV-PAST/ FIN-TENT-ATTR at what point in time, would the frost have fallen on [their] pitch-black hair? (MYS 5.804) 可具呂伎可美尓都由曽於伎尓家類

kaⁿ-gurǝ18-ki kami-ni tuyu sǝ ok-i-n-i-ker-u INT-black-ATTR hair-LOC dew FP put-CONV-PERF-CONV-RETR-ATTR [it] turned out that the dew fell on the pitch-black hair (MYS 15.3649) 18  The character 呂 transcribing otsu-rui /rǝ/ is apparently a scribal mistake for kō-rui /ro/.

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

517

Verbs 蚊黒為髪尾信櫛持於是蚊寸垂

kaⁿ-GURO-si KAMI-wo MA-KUSI MƏT-I kaki-TARE INT-black-FIN hair-ACC INT-comb hold-CONV PREF-make.hang.down(CONV) making pitch black hair hang down with a comb (MYS 16.3791) 手放毛乎知母可夜須伎

TA-ⁿ-BANARE mo woti mǝ ka-yasu-ki hand-LOC-separate(NML) FP there FP INT-easy-ATTR both leaving the hand and [returning] there were very easy (MYS 17.4011) 3.1.4 Prefix taThe prefix ta-, like the prefix ka- discussed above, can precede both verbs and inflected adjectives, although, in contrast to ka-, it is attested more frequently with verbs than with inflected adjectives. Similar to the prefix ka- < *kaⁿ-, the underlying form of the prefix ta- is also likely to be *taⁿ-, since this is the form that surfaces consistently before the first obstruent of the following root, see tam-basir- ‘to run’ and taⁿ-dǝpo- ‘to be far’ below.19 The prefix ta- also has two other peculiarities: 1) it can occur after another prefix (examples are limited to the combinations i-ta- and ari-ta- with the prefixes i- and ari-) or after the first verb in a verbal compound (examples are limited to pap-i-ta-mǝtǝpor- ‘to crawl around,’ kǝŋg-i-ta-mǝtǝpor- ‘to row around,’ and tǝmb-i-ta-mǝtǝpor- ‘to fly around’; 2) it is predominantly attested with the verb mǝtǝpor- ‘to go/wander around,’ although it also occurs with other verbs. No definition of its function is given in Japanese scholarship (Yamada 1954: 534). Pierson defined ta- as an intensifying prefix (Pierson 1963: 8). He might be right, although the intensifying meaning is unlikely to fit into most of the examples involving the verb mǝtǝpor- below. It may not be possible to find the solution at all, since examples are not numerous and most of them are limited to usage with the verb mǝtǝpor-, as I have already mentioned above. The prefix ta- occurs only with a limited number of Western Old Japanese verbs: mǝtǝpor- mï- pasir-

‘to go around’ ‘to turn’ ‘to run’

19  Two apparent exceptions are ta-sim- ‘PREF-grow.densely’ in KK 91 below, and ta-sar‘PREF-go.away’ in NK 40 below, but the line with ta-sar- is considered to have problems in interpretation (Tsuchihashi 1957: 150).

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

518 sar- sim- wasure-

Chapter 6

‘to go away’ ‘to grow densely’ ‘to forget’

The prefix ta- with verbs: 伊久美陀気伊久美波泥受多斯美陀気多斯爾波韋泥受能知母久美泥牟

i-kum-i-ⁿ-dakɛ i-kum-i pa ne-ⁿz-u ta-sim-i-ⁿ-dakɛ tasi n-i pa wi-ne-ⁿz-u nǝti mǝ kum-i-ne-m-u DLF-entwine-NML-DV(ATTR)-bamboo DLF-entwine-CONV TOP sleep-NEGCONV PREF-grow.densely-NML-DV(ATTR)-bamboo thorough DV-CONV TOP exist(CONV)-sleep-NEG-CONV later FP entwine-CONV-sleep-TENT-FIN [we] did not sleep entwined there [as] the bamboo entwined there, and [we] did not sleep closely [like] densely growing bamboo, [but] later [we] will sleep entwined (KK 91) 嚢伽多佐例

ta ka ta-sar-e who IP PREF-go.away-EV Who goes away? (NK 40) 其夜乃梅乎手忘而

SƏNƏ YO-nǝ UMƐ-wo ta-WASURE-TE that night-GEN plum-ACC PREF-forget(CONV)-SUB [I] forgot about the plum [blossoms] of that night, and … (MYS 3.392) 若子乃匍匐多毛登保里

MIⁿDƏRI KO-nǝ PAP-I-ta-mǝtǝpor-i young child-COMP crawl-CONV-PREF-go.around-CONV crawling around like a baby (MYS 3.458) 之夫多尓能佐吉多母登保理

Simbutani-nǝ saki ta-mǝtǝpor-i Simbutani -GEN cape PREF-go.around-CONV going around Simbutani cape (MYS 17.3991) 麻佐吉久毛安里多母等保利

ma-saki-ku mo ari-ta-mǝtǝpor-i INT-safe-CONV FP ITER-PREF-wander.around-CONV so that [you] will be wandering around safely, and … (MYS 17.4008)

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

519

Verbs 多古能之麻等妣多毛登保里

Tako-nǝ sima tǝmb-i-ta-motǝpor-i Tako-GEN island fly-CONV-PREF-go.around-CONV [he] flew around Tako island (MYS 17.4011) 乎敷乃佐吉許藝多母等保里

Wopu-nǝ saki kǝŋg-i-ta-mǝtǝpor-i Wopu-GEN cape row-CONV-PREF-go.around-CONV rowing around cape Wopu (MYS 18.4037) 霜上尓安良礼多婆之里

SIMO-NƏ UPƐ-ni arare tam-basir-i frost-GEN top-LOC hail PREF-run-CONV Hail falls on the frost … (MYS 20.4298) 乎可乃佐伎伊多牟流其等尓

woka n-ǝ saki i-ta-m-uru ŋgǝtǝ n-i hill DV-ATTR cape DLF-PREF-turn.around-ATTR every DV-CONV every time [I] turn around a hilly cape (MYS 20.4408) The prefix ta- with inflected adjectives: 道乎多遠見思空安莫國嘆虚不安物乎

MITI-wo taⁿ-DƏPƏ-mi OMƏP-U sora YASUKƐ NA-ku n-i NA ŋGƐK-U SORA YASU-K-AR-AN-U MƏNƏwo way-ABS PREF-far-GER think-ATTR RP easy no-CONV DV-CONV lament-ATTR RP easy-CONV-exist-NEG-ATTR CONJ although it is not easy to lament and to love because the way is far (MYS 4.534) 言云者三々二田八酢四

KƏTƏ-NI IP-Ɛ-mba mimi-ni ta-yasu-si word-LOC say-EV-CON ear-LOC PREF-easy-FIN when [I] say [it] in words, [it] is insignificant (lit.: easy) for [your] ears (MYS 11.2581) 道乎多騰保美山河能敝奈里氐安礼婆

MITI-wo taⁿ-dǝpo-mi YAMA KAPA-nǝ penar-i-te ar-e-mba way-ABS PREF-far-GER mountain river-GEN be.separated-CONV-SUB existEV-CON because the way is far, and because [we] were separated from mountains and rivers (MYS 17.3957) Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

520

Chapter 6

美知乎多騰保弥間使毛夜流余之母奈之

miti-wo taⁿ-dǝpo-mi MA-ⁿ-DUKAPI mo yar-u yǝsi mǝ na-si way-ABS PREF-far-GER interval-GEN-messenger FP send-ATTR chance FP no-FIN because the way is far, there is not even a chance to send a messenger (MYS 17.3962) 3.1.5 Traditional Prefix sǝYamada Yoshio as well as Omodaka et al. also mention a prefix sǝ- that is carved by them from three Western Old Japanese verbs: sǝⁿdatak- ‘to caress,’ sǝⁿdar- ‘be plentifully endowed,’ and sǝnapɛ- ‘to prepare’ (Yamada 1954: 535; Omodaka et al. 1967: 402–403). However, we have several problems here. The first of these verbs, sǝⁿdatak- ‘to caress’ is a hapax legomenon attested only in KK 3. Moreover, tatak- means ‘to hit,’ not ‘to caress,’ and even if we follow the speculation that sǝⁿdatak- ‘to caress’ actually means ‘hit slightly,’ with a prefix sǝ(ⁿ)- conveying the meaning ‘slightly,’ that will bring us into contradiction with the suggested meaning of the sǝ(ⁿ)- in the verb sǝⁿdar- ‘be plentifully endowed,’ because in this case sǝ(ⁿ)- must mean ‘plentifully.’ In addition, sǝⁿdar‘be plentifully endowed’ is also a hapax legomenon attested in BS 2. The last of these verbs, sǝnapɛ- ‘to prepare,’ is attested in several texts, but it does not have a complete phonographic attestation,20 and what is even more important is that the verb *napɛ- ‘to prepare’ does not present itself, thus there is no justification for segmenting any ‘prefix’ here.21 Therefore, I have a strong suspicion that here we are dealing with a case of over-segmentation, and the ‘prefix’ sǝsimply does not exist. 3.1.6 Prefix ariThe prefix ari- is a marker of the iterative. It is defined as such in Omodaka et al. 1967: 57, but more often than not it is not even mentioned in the existing grammatical descriptions of Old Japanese, as, for example in Yamada Yoshio’s seminal grammar (Yamada 1954). There is a possibility that historically it is a preverb rather than a prefix, being a converb form of the verb ar- ‘to exist.’ Even if this is the case, the presence of an auxiliary verb before the main verb once again manifests a strong contradiction to the SOV word order typology, where auxiliaries are supposed to follow main verbs. In spite of several cases

20  Only its derived form sǝnapar- ‘be provided’ is attested phonetically. 21  Apparently WOJ napɛ- ‘to weaken, to become numb’ (Omodaka et al. 1967: 530) does not belong here.

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

521

Verbs

of sporadic prenasalization of a following initial voiceless obstruent k- in the verb kayop- ‘to set out, to go back and forth,’ the underlying form must be *ari-, since this prenasalization fails to occur not only before other initial voiceless obstruents, but also before kayop- in another example (KK 2). The prefix arioccurs only with a limited number of Western Old Japanese verbs: kayop- mɛŋgur- mǝtǝpor- naŋgusamɛ- sar- tat- tat- watar-

‘to set out, to go back and forth’ ‘to go around’ ‘to go around’ ‘to console’ ‘to go away’ ‘to depart’ ‘to stand’ ‘to cross’

Examples: 佐用婆比爾阿理多多斯用婆比迩阿理加用婆勢

sa-yomb-ap-i-ni ari-tat-as-i yomb-ap-i-ni ari-kayop-as-e PREF-call-ITER-NML-LOC ITER-set.out-HON-CONV call-ITER-NML-LOC ITER-set.out-HON-EV [I] set out (repeatedly) to woo [her] there, [I] set out (repeatedly) to woo [her] (KK 2) 嶋之埼耶伎安利立有花橘

SIMA-NƏ SAKI-ⁿzaki ari-TAT-ER-U PANA TATImBANA island-GEN cape-cape ITER-stand-PROG flower mandarin.orange flowering mandarin orange [trees] that are standing all the time at the capes of the islands (MYS 13.3239) 阿里佐利氐能知毛相牟等於母倍許曽

ari-sar-i-te nǝti mo AP-Am-u tǝ omǝp-ɛ kǝsǝ ITER-go.away-CONV-SUB after FP meet-TENT-FIN DV think-EV FP [Time] constantly goes away, and [I] hope that [we] will meet later, too (MYS 17.3933) 麻佐吉久毛安里多母等保利

ma-saki-ku mo ari-ta-mǝtǝpor-i INT-safe-CONV FP ITER-PREF-wander.around-CONV so that [you] will be wandering around safely, and … (MYS 17.4008)

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

522

Chapter 6

由久敝奈久安里和多流登毛

yuk-u pe na-ku ari-watar-u tǝmo go-ATTR side no-CONV ITER-cross-FIN CONJ Even though [a cuckoo] constantly comes over, without having a direction to go (MYS 18.4090) 和期於保伎美余思努乃美夜乎安里我欲比賣須

wa-ŋgǝ opo kimi Yǝsino-nǝ miya-wo ari-ŋgayop-i mes-u I-POSS great lord Yǝsino-GEN palace-ACC ITER-go.back.and.forth-CONV look(HON)-FIN My emperor constantly visits the palace in Yǝsino, and looks [around] (MYS 18.4099) 之麻豆多比伊己藝和多利弖安里米具利

sima-ⁿ-dutap-i i-kǝŋg-i-watar-i-te ari-mɛŋgur-i island-LOC-pass.along-CONV DLF-row-CONV-cross-CONV-SUB ITER-go. around-CONV [I] go rowing in a boat from island to island, and [I] constantly go around [these islands] (MYS 20.4408) Comparative Data Level A: Other Japonic A1: Eastern Old Japanese There is one example of the prefix ari- attested in Eastern Old Japanese: 久自我波々佐氣久阿利麻弖

Kuⁿzi-N-kapa pa sakɛ-ku ari-mat-e Kuⁿzi-GEN-river TOP safe-CONV ITER-wait-IMP Wait for [me] (all this time) safely at Kunzi river! (MYS 20.4368) 3.1.7 Prefix utiThere is a possibility that historically uti- is a preverb rather than a prefix, being the converb form of the verb ut- ‘to hit, to strike.’ This etymology may be supported by numerous logographic spellings of uti- as 打 or, less frequently, 敲 ‘to hit, to strike.’ The difficulty of accepting this etymology is ultimately connected with the fact that it is very difficult if not plainly impossible to trace the meaning of any ‘hitting’ or ‘striking’ in most if not all of the examples given below. Omodaka et al. indicate the same problem (Omodaka et al. 1967: 119), but more often than not the prefix uti- is not even mentioned in the existing grammatical descriptions of Old Japanese, as, for example in Yamada Yoshio’s

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

523

Verbs

seminal grammar (Yamada 1954). It is nevertheless possible that this prefix or preverb is historically connected with ut- ‘to strike, to hit,’ as most of the examples cited below, with a possible exception of MYS 5.892, indicate some kind of momentary or punctuated action. Whether uti- represents a prefix or an auxiliary, the presence of either of them before the main verb once again manifests a strong contradiction to the SOV word order typology, where auxiliaries are supposed to follow main verbs and prefixes are very rare. The prefix uti- occurs only with a limited number of Western Old Japanese verbs: iⁿde- kakɛ- kǝyi- koye- ki- kiras- mure- mi- naŋgɛk- nambik- naⁿde- nas- nǝmbǝr- ok- pamɛ- panap- panat- papɛ- papuk- parap- pur- pure- saras- sinap- sinop- susurǝp- tukɛ- watas- wor- yuk-

‘to go out’ ‘to hang, to place’ ‘to lie down’ ‘to cross over, to pass over’ ‘to wear’ ‘to cover in fog, to make cloudy’ ‘to gather’ ‘to see’ ‘to lament, to sigh’ ‘to stretch’ ‘to caress, to stroke’ ‘to make sound’ ‘to climb’ ‘to put, to place’ ‘to insert’ ‘to sneeze’ ‘to release’ ‘to stretch’ ‘to flutter wings’ ‘to clean’ ‘to fall (of precipitations)’ ‘to touch’ ‘to bleach’ ‘to bend’ ‘to long for, to yearn’ ‘to sip noisily’ ‘to attach’ ‘to pass, to take over’ ‘to break’ ‘to go’

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

524

Chapter 6

Examples: 許能登理母宇知夜米許世泥

kǝnǝ tǝri mǝ uti-yamɛ-kǝse-n-e this bird FP PREF-stop(CONV)-BEN-DES-IMP [I] wish [you] would stop [the singing of] these birds (KK 2) 清瀬乎馬打和多思

KIYO-KI SE-wo UMA UTI-watas-i clear-ATTR rapids-ACC horse PREF-carry.across-CONV [I] am taking [my] horse across the clear rapids (MYS 4.715) 許許呂由母於母波奴阿比陀爾宇知那毘枳許夜斯努礼

kǝkǝrǝ-yu mǝ omǝp-an-u apiⁿda-ni uti-nambik-i kǝy-as-i-n-ure heart-ABL FP think-NEG-ATTR interval-LOC PREF-stretched.out-CONV lie. down-HON-CONV-PERF-EV while even in [my] heart [I] did not think, stretched out [she] was lying (MYS 5.794) 阿迦胡麻尓志都久良宇知意伎

aka-ŋ-goma-ni situ kura uti-ok-i red-DV(ATTR)-stallion-LOC pattern saddle PREF-place-CONV [they] put adorned saddles on red stallions, and … (MYS 5.804) 久佐太袁利志婆刀利志伎提等許自母能宇知許伊布志提

kusa-n-da-wor-i simba tor-i sik-i-te tǝkǝ ⁿzimǝnǝ uti-kǝyi-pus-i-te grass-?-hand-break-CONV road.side.grass hold-CONV spread-CONV-SUB bed like PREF-lie.down-CONV-lie.prone-CONV-SUB [I] broke off some herbs, and taking road side grass, [I] spread [it] and lied down on a bed-like [thing] (MYS 5.886) 糟湯酒宇知須々呂比弖

KASU-YU-ⁿ-ZAKƐ uti-susurǝp-i-te dreg-hot.water-DV(ATTR)-rice.wine PREF-sip.noisily-CONV-SUB [I] sip noisily the hot water [containing] rice wine dregs, and … (MYS 5.892) 波祢左之可倍弖宇知波良比

pane sas-i-kapɛ-te uti-parap-i feather insert-CONV-cross.over(CONV)-SUB PREF-clean-CONV [they] mingle their feathers and clean [frost from them] (MYS 15.3625)

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

525

Verbs 鸎能奈久々良多尓々宇知波米氐

U ŋGUPISU-nǝ nak-u kura tani-ni uti-pamɛ-te bush.warbler-GEN sing-ATTR dark valley-LOC PREF-insert(CONV)-SUB Throwing [myself] into a dark valley where a bush warbler sings (MYS 17.3941) 宇知奈妣伎登許尓己伊布之

uti-nambik-i tǝkǝ-ni kǝyi-pus-i PREF-stretch.out-CONV bed-LOC lie.down(CONV)-lie.prone-CONV [I] lie down, stretched out on [my] bed (MYS 17.3969) 近在者加敝利尓太仁母宇知由吉氐

TIKA-KU AR-Amba kaper-i-ni ⁿdani mǝ uti-yuk-i-te close-CONV exist-COND return-NML-LOC RP FP PREF-go-CONV-SUB if [you] were close, [I] [would] go [to you] just on [my] way back, and … (MYS 17.3978) 宇麻宇知牟礼弖

uma uti-mure-te horse PREF-gather(CONV)-SUB gathering horses [together] (MYS 17.3993) 惠美々惠末須毛宇知奈氣支可多里家末久

wem-i mi wem-aⁿz-u mo uti-naŋgɛk-i katar-i-k-em-aku smile-CONV ? smile-NEG-CONV FP PREF-sigh-CONV talk-CONV-PAST/FINTENT-NML the fact that [both of you] were probably talking, and sighing, smiling and not smiling (MYS 18.4106) Comparative Data Level A: Other Japonic A1: Eastern Old Japanese There is only one example of the prefix uti- in Eastern Old Japanese: 多麻母乃宇知奈婢伎比登里夜宿良牟

tama mǝ-nǝ uti-nambik-i pitǝ-ri ya NE-ram-u jewel seaweed-COMP PREF-stretch-CONV one-CL IP sleep-TENT2-ATTR will [you] sleep alone, stretched like a jewel seaweed? (MYS 14.3562)

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

526

Chapter 6

A2: Ryukyuan To the best of my knowledge the prefix uti- is attested only in Old and Classical Ryukyuan. Since there is no support from modern Ryukyuan languages, there is a possibility that OR and CR uti- represents a loan from MJ uti-, since the prefix uti- is attested not only in Old Japanese, but also in Middle Japanese. Nevertheless, there is also a chance that OR and WOJ uti- are genuine cognates. Examples from Old Ryukyuan: なりとよみうちあげてなりきよらはうちあげて

nar-i-toyom-i uti-age-te nar-i-kiyora fa uti-age-te sound-CONV-resound-NML PREF-raise(CONV)-SUB sound-CONV-beautiful TOP PREF-raise(CONV)-SUB raising the “Resounding” [drum], raising the “Beautifully sounding” [drum] (OS 1.37) しよりもりうちあよでまたまもりうちあよで

Siyori mori uti-ayode ma-tama mori uti-ayode Shuri shrine PREF-walk/SUB INT-jewel shrine PREF-walk/SUB walking through the Shuri shrine, walking through the true jewel shrine (OS 1.40) 3.1.8 Prefix kakiThere is a possibility that historically kaki- is a preverb rather than a prefix, being a converb form of the verb kak- ‘to scratch.’ Omodaka et al. indicate that the prefix kaki- precedes verbs that indicate some action done by the fingertips (Omodaka et al. 1967: 176), but more often than not this prefix is not even mentioned in the existing grammatical descriptions of Old Japanese, as, for example, in Yamada Yoshio’s seminal grammar (Yamada 1954). This etymology may be supported by numerous logographic spellings of kaki- as 掻. However, the difficulty of accepting this etymology is ultimately connected with the fact that it is very difficult if not outright impossible to trace the meaning of any ‘scratching’ in most if not all of the examples given below. It is nevertheless possible that this prefix or preverb is historically connected with kak- ‘to scratch,’ as most of the examples cited below, with a possible exception of those from KK 5, MYS 8.1520 and MYS 19.4191, indicate some kind of physical contact done by a hand, although not necessarily by fingertips. Whether kakirepresents a prefix or an auxiliary, the presence of either of them before the main verb once again manifests a strong contradiction to the SOV word order, where auxiliaries are supposed to follow main verbs and prefixes are nonexistent in most SOV languages.

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

527

Verbs

The prefix kaki- occurs only with a limited number of Western Old Japanese verbs: kaⁿzopɛ- kiras- saŋgur- sute- tare- tat- tukɛ- naⁿde- pak- pik- mï- mukɛ- musumb- tar- tuk- wakɛ- watar- yǝse-

‘to count’ ‘to cover in fog, to make cloudy’ ‘to seek’ ‘to discard’ ‘to make hang down’ ‘to depart’ ‘to attach’ ‘to caress, to stroke’ ‘to sweep’ ‘to pull, to play (instruments)’ ‘to turn, go around’ ‘to face’ ‘to tie, to bind’ ‘to hang down’ ‘to be attached’ ‘to separate, to divide’ ‘to cross’ ‘to approach’

Examples: 加岐微流伊蘇能佐岐

kaki-mï-ru iso-nǝ saki PREF-go.around-ATTR rocky.shore-GEN cape the cape on the rocky shore that [you] are going around (KK 5) 之餓阿摩離虚等珥莵句離訶枳譬句椰

si-ŋga amari kǝtǝ n-i tukur-i kaki-pik-u ya it-POSS remainder koto DV-CONV make-CONV PREF-play-FIN IP [they] made a koto out of its remainders, will [they] play? (NK 41) 陁倶符羅爾阿武柯枳都枳

ta-kumbura-ni amu kaki-tuk-i arm-fleshy.part-LOC gadfly PREF-attach-CONV a gadfly sat on [the sovereign’s] upper arm (NK 75)

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

528

Chapter 6

比宜可伎撫而

piŋgɛ kaki-naⁿde beard PREF-stroke(CONV) stroking [my] beard (MYS 5.892) 御津松原可吉掃弖

Mitu-nə MATU- ͫ-BARA kaki-PAK-I-te Mitu-GEN pine-GEN-field PREF-sweep-CONV-SUB having swept the pine field of Mitu (MYS 5.895) 菅根乎衣尓書付

SU ŋGA-NƏ NE-wo KƏRƏMƏ-ni kaki-TUKƐ sedge-GEN root-ACC garment-LOC PREF-attach(CONV) attaching a root of sedge to [her] garment (MYS 7.1344) 朝奈藝尓伊可伎渡

ASA naŋgi-ni i-kaki-WATAR-I morning calm-LOC DLF-PREF-cross.over-CONV crossing over there in the morning calm (MYS 8.1520) 咲有花乎指折可伎數者

SAK-I-TAR-U PANA-wo OYOmBI WOR-I kaki-KAⁿZOP-URE-mba bloom-CONV-PERF/PROG-ATTR flower-ACC finger bend-CONV PREF-countEV-CON when [I] counted blooming flowers on my fingers (MYS 8.1537) 加吉結常代尓至

kaki-MUSUmB-I TƏKƏ YƏ-ni ITAR-I PREF-bind-CONV eternal life-LOC reach-CONV concluding [the agreement of marriage] and reaching [the land of] eternal life (MYS 9.1740) 葦垣之末掻別而

ASI-ŋ-GAKI-NƏ SUWE KAKI-WAKƐ-TE reed-GEN-fence-GEN top PREF-divide(CONV)-SUB dividing the top of the reed fence (MYS 13.3279) 蚊黒為髪尾信櫛持於是蚊寸垂

kaŋ-GURO-si KAMI-wo MA-KUSI MƏT-I kaki-TARE INT-black-FIN hair-ACC INT-comb hold-CONV PREF-make.hang.down(CONV) making pitch black hair hang down with a comb (MYS 16.3791) Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

529

Verbs 吾等尓可伎无氣念之念婆

WARE-ni kaki-mukɛ OMƏP-I si OMƏP-Amba I-DAT PREF-turn(IMP) love-NML EP love-COND if [you] indeed love [me], turn to me (MYS 19.4191) 美母乃須蘇都美安氣可伎奈埿

mi-mǝ-nǝ suso tum-i-aŋgɛ kaki-naⁿde HON-skirt-GEN hem pick-CONV-raise(CONV) PREF-caress(CONV) [my mother] picked up [her] skirt hems and caressed [me] (MYS 20.4408) 手肱爾水沫書垂向股爾泥書寄弖

TA-NA PIⁿDI-ni MI-N[A] AWA kaki-TAR-I MUKA MOMO-ni PIⁿDI kaki-YƏSE-te arm-PLUR elbow-LOC water-PLUR foam PREF-hang.down-CONV front thighLOC dirt PREF-approach(CONV)-SUB the water foam was dripping at the elbows, the dirt was sticking up to the thighs (NT 1) Comparative Data Level A: Other Japonic A1: Eastern Old Japanese There are two examples of the prefix kaki- attested in the Eastern Old Japanese texts: 可伎武太伎奴礼杼安加奴乎安杼加安我世牟

kaki-muⁿdak-i n-ure-ⁿdǝ ak-an-u-wo aⁿ-tǝ ka a-ŋga se-m-u PREF-embrace-CONV sleep-EV-CONC satisfy-NEG-ATTR-ACC what-DV IP I-POSS do-TENT-ATTR although [I] slept [with her] keeping [her] in my arms, since it was not enough [for me], what should I do? (MYS 14.3404) 知々波々我可之良加伎奈弖佐久安礼天伊比之氣等婆是和須礼加祢豆流

titi papa-ŋga kasira kaki-naⁿde sa-ku ar-e te ip-i-si kɛtǝmba ⁿze wasure-kane-t-uru father mother-POSS head PREF-stroke(CONV) safe-CONV exist-IMP DV sayCONV-PAST/ATTR word FP forget(CONV)-NEG/POT(CONV)-PERF-ATTR [I] cannot forget the words: “Be safe!” that [my] father and mother said, stroking [my] head (MYS 20.4346) A2: Ryukyuan The WOJ prefix kaki- can be compared to the prefix kai- ~ kaki- found in Classical Ryukyuan (Hokama 1995: 177, 181). However, since this prefix is not

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

530

Chapter 6

attested in Old Ryukyuan and/or modern Ryukyuan languages, and because it is also present in Middle Japanese in the form kai- ~ kaki- with the variant kai- that does not occur in Western Old Japanese, it is likely that CR kai- ~ kakirepresents a loan from MJ kai- ~ kaki-. 3.1.9 Prefix apiThe prefix api- is a marker of the reciprocal-cooperative voice, indicating that an action is either reciprocal or is performed together. It is defined as such in (Omodaka et al. 1967: 57), but more often than not the prefix api- is not even mentioned in the existing grammatical descriptions of Old Japanese, as, for example in Yamada Yoshio’s seminal grammar (Yamada 1954). It is quite likely that historically the prefix api- represents the converb form of the verb ap- ‘to meet, to join.’ Whether api- is historically a prefix or an auxiliary, the presence of either of them before the main verb once again manifests a strong contradiction to the SOV word order, where auxiliaries are supposed to follow main verbs and prefixes are non-existent. It is also interesting that the reciprocalcooperative voice can also be expressed in Western Old Japanese analytically by the verb ap- ‘to meet, to join’ that follows the converb form as an auxiliary. The latter form is, of course, much more consistent with the SOV typology. Thus, it is likely that the coexistence of both forms in Western Old Japanese demonstrates the last stages of transition from a SVO to a SOV language. The prefix api- occurs only with a limited number of Western Old Japanese verbs: arasop- ‘to fight’ ip- ‘to say’ kipop- ‘to compete’ mak- ‘to roll’ makuramak- ‘to pillow’ muk- ‘to face’ mi- ‘to see, to look’ ne- ‘to sleep’ nǝm- ‘to drink’ omǝp- ‘to think’ tomburap- ‘to visit’ tǝyǝm- ‘resound’ uⁿdunap- ‘to treat with care’ wakare- ‘to separate’ yǝr- ‘to approach’ yombap- ‘to marry’

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

531

Verbs

Examples: 古波陀袁登売袁迦微能碁登岐許延斯迦杼母阿比麻久良麻久

Kopaⁿda wotǝme-wo kamï-nǝ ŋgǝtǝ kik-ǝye-sika-ⁿdǝmǝ api-makuramak-u Kopaⁿda maiden-ABS deity-GEN like hear-PASS(CONV)-PAST/EV-CONC RECpillow-FIN Although [it] is rumored that the maiden from Kopaⁿda is like a goddess, [we] slept together (lit.: pillowed each other) (KK 45) 許許呂袁陀迩迦阿比淤母波受阿良牟

kǝkǝrǝ-wo ⁿdani ka api-omǝp-aⁿz-u ar-am-u heart-ACC RP IP REC-think-NEG-CONV exist-TENT-ATTR will [we] not think about each other at least in our hearts? (KK 60) 吉備那流伊慕塢阿比瀰莵流慕能

Kimbï-n-ar-u imo-wo api-mi-t-uru monǝ Kimbï-LOC-exist-ATTR beloved-ACC REC-see(CONV)-PERF-ATTR CONJ [My] beloved who is in Kimbï [and I], have seen each other, but … (NK 40) 陁黎耶始比登謀阿避於謀婆儺倶爾

tare ya si pitǝ mo api-omop-an-aku n-i who EP EP person FP REC-love-NEG-NML DV-CONV because [she] is not mutually in love with anyone else (NK 93) 空氣衝之相別去者

ANA IKIⁿDUK-Asi API-WAKARE-n-amba INTER lament-ADJ(FIN) REC-part(CONV)-PERF-COND Oh, how lamentable! If [we] part with each other … (MYS 8.1454) 安比於毛波奴君尓安礼也母

api-omop-an-u KIMI n-i ar-e ya mǝ REC-think-NEG-ATTR lord DV-CONV exist-EV IP EP is [it my] lord, who no [longer] thinks [about lamenting of the people of this world] in return?! (MYS 15.3691) 相見婆登許波都波奈爾

API-MI-RE-mba tǝkǝ patu pana n-i REC-look-EV-CON eternal first flower DV-CONV when [we] looked at each other, it was [always] like eternal first flowers (MYS 17.3978)

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

532

Chapter 6

天地乃神安比宇豆奈比

amɛ tuti-nǝ KAMÏ api-uⁿdunap-i heaven earth-GEN deity COOP-treat.with.care-CONV deities of Heaven and Earth all together treated [us] with care … (MYS 18.4094) 安比見流毛乃乎須久奈久母年月經礼波古非之家礼夜母

api-MI-ru monǝwo sukuna-ku mǝ TƏSI TUKÏ P-Ure-mba kopïsi-kere ya mǝ REC-see-ATTR CONJ few-CONV FP year month pass-EV-CON be.longing-EV IP EP although [we] saw each other, as the time went by, were [we still] longing for [each other] just a bit?! [—No, quite a lot!] (MYS 18.4118) 相飲酒曽斯豊御酒者

API-NƏM-AM-U SAKƐ sǝ KƏNƏ TƏYƏ MI-KI pa COOP-drink-TENT-ATTR rice.wine FP this eternal HON-rice.wine TOP the rice.wine that [we] will drink together, this eternal rice wine (MYS 19.4264) 天地与相左可延牟等

AMƐ TUTI-TƏ API-sakaye-m-u tǝ heaven earth-COM COOP-flourish-TENT-FIN DV [I] wish that [you] would flourish together with Heaven and Earth … (MYS 19.4273) 悪奴止母止相結弖

ASI-KI YATU-ⁿdǝmǝ-tǝ API-MUSUmB-I-te bad-ATTR scoundrel-PLUR-COM COOP-tie-CONV-SUB [they] tied [themselves] together with bad scoundrels, and … (SM 43) Comparative Data Level A: Other Japonic A1: Eastern Old Japanese The reciprocal-cooperative prefix api- occurs in Eastern Old Japanese as well: 和賀西奈尓阿比与流等可毛

wa-ŋga se-na-ni api-yǝr-u tǝ kamo I-POSS beloved-DIM-DAT COOP-approach-FIN DV EP I wonder, does [it] say that my beloved [and I] will meet together? (MYS 14.3483)

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

533

Verbs

A2: Ryukyuan The cooperative prefix ai- is possibly attested only in one example in Old Ryukyuan. There are no other attestations in Classical Ryukyuan or modern Ryukyuan languages. In spite of the fact that we are dealing here with a unique attestation, it is still possible that this example reflects an Old Ryukyuan cognate of WOJ api-, because the prefix afi- is not found in Middle Japanese, and direct loans from Western Old Japanese to Old Ryukyuan do not exist. さいわたるのさくらしけしけとおりさちへけおよりあいいでらむ

sa-i-watar-u n-o sakura sike-sike to or-i-sat-ife keo-yori ai-ide-ram-u bloom-CONV-cross-ATTR DV-ATTR sakura denze-denze DV bend-CONVstretch-CONV today-ABL COOP-go.out-TENT2-FIN [Sailors!] From today [you] should go out together bundling together [like] blooming sakura [trees] (OS 10.531) 3.1.10 Prefix eThe prefix e- is a marker of the potential. It is defined as a potential adverb in Omodaka et al. (1967: 57), but with the exception of a dubious case of e in MYS 18.4078 that is likely to have a different explanation, e- is invariably found immediately before verbal roots, so it is more appropriate to view it as a prefix, at least on the synchronic level. Historically the prefix e- represents the converb form of the verb e- ‘to get.’ The prefix e- occurs only with a limited number of Western Old Japanese verbs: ar- ip- mi- se- yuk-

‘to exist’ ‘to say’ ‘to see’ ‘to do’ ‘to go’

There is only one example where e- is written phonographically in the Bussoku seki-no uta. In all other cases it is spelled logographically with the character 得. Examples: 打乍二波更毛不得言

ututu-ni pa SARA N-I mo E-IP-AⁿZI reality-LOC TOP again DV-CONV FP POT-say-NEG/TENT [I] would not be able to say [it] again in reality (MYS 4.784)

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

534

Chapter 6

面忘太尓毛得為也

OMO WASURE ⁿdani mo E-S-U ya face forget(NML) RP FP POT-do-FIN IP Could [I] just forget [his] face? (MYS 11.2574) 美阿止須良乎和礼波衣美須弖伊波爾惠利都久多麻爾惠利都久

mi-atǝ-sura-wo ware pa e-mi-ⁿz-u-te ipa-ni wer-i-tuk-u tama-ni wer-i-tuk-u HON-footprint-RP-ACC I TOP POT-see-NEG-CONV-SUB rock-LOC carve-CONVattach-FIN jewel-LOC carve-CONV-attach-FIN I was not able to see even the footprints of the Buddha, so [I] carve [them] on the rock, carve [them] on the jewel (BS 3) 3.2

Verbal Suffixes

If we follow a strict linguistic analysis of Western Old Japanese, we have to divide all verbal suffixes into two major groups: sentence-final suffixes which normally occur at the end of the last verbal form in a sentence or a clause, and sentence non-final suffixes that cannot occur in sentence-final position unless ellipsis occurs. Sentence-non-final suffixes, in their turn, can be further subdivided into word-final and word-non-final suffixes. Word-final suffixes can conclude a verbal form, while word-non-final suffixes cannot be final in a verbal form by themselves: they must always be followed either by sentence-final or word-final suffixes. The markers of final predicates, such as the suffixes of final predication, attributives, imperatives, most markers of mood, and a few other markers, are sentence-final suffixes. The converbs, which are markers of non-final predicates, are sentence-non-final word-final suffixes. The markers of voice, negation, and aspect are all word-non-final suffixes. A big watershed that divides Western Old Japanese from Middle (Classical) Japanese is that a number of auxiliaries in the former became suffixes in the latter by the process of losing the morphemic boundary between the converb -i and a following auxiliary. Thus, for example, the Classical Japanese objective retrospective auxiliary -iker- goes back to a Western Old Japanese analytical construction involving the converb -i of the main verb and the auxiliary -ker(that itself is likely to be a contraction of the converb k-i of the verb kǝ- ‘to come’ and the auxiliary verb ar- ‘to exist’). While the form -iker- is the only one that occurs in Middle Japanese, which prevents us to subdivide it morphemically as -i-ker-, in Western Old Japanese -ker- can follow other converb forms,

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

535

Verbs

such as an converb -u, found after the negative marker -aⁿz-, for example: kopï yam-aⁿz-u-ker-i love(NML) stop-NEG-CONV-RETR-FIN ‘[our] love does not stop’ (MYS 17.3980). This and other examples convincingly demonstrate that at the time of Western Old Japanese there was a morphological boundary between the converb -i and the retrospective -ker-. 3.2.1 Sentence-Final Verbal Suffixes Western Old Japanese contains a number of sentence-final verbal suffixes: the markers of final predication -u ~ -i; the attributive suffixes -uru ~ -u ~ -ru; the evidential marker -ure ~ -ɛ ~ -e ~ re; the imperative suffixes -e ~ -yǝ ~ -∅; the negative imperative -una; the mood markers: the negative tentative -aⁿzi ~ -ⁿzi, the desiderative -ana ~ -na, negative potential -umasiⁿzi, and the subjunctive -amasi ~ -masi. 3.2.1.1 Final Predication Suffix -u ~ -i The final predication suffix has two allomorphs -u and -i. The allomorph -u follows the stems of all verbs except r-irregular verbs which have the allomorph -i as their final predication suffix. The main descriptive problem concerns final predication forms of strong vowel verbs. In Middle Japanese strong vowel verbs clearly take the special allomorph -ru as their final predication suffix, which is identical to the attributive suffix -ru that they also have. The situation is not that obvious in Western Old Japanese, in spite of the fact that most existing Japanese grammars promptly list -ru as a suffix of final predication for strong vowel verbs (Yamada 1954: 155; Shirafuji 1987: 127). The opposite point of view is expressed by Iwai Yoshio, who believes that the final predication form of strong vowel verbs in -ru is not attested in Old Japanese (Iwai 1970: 52). I think that he is right, because there is only one example in the texts where we can possibly have a glimpse of the final predication form in -ru: 麻須羅遠能佐都夜多波佐美牟加比多知伊流夜麻度加多波麻乃佐夜氣佐

masura wo-nǝ satu-ya ta-m-basam-i mukap-i tat-i i-ru ya mato-kata pama-nǝ sayakɛ-sa excellent man-GEN hunt-arrow hand-LOC-squeeze-CONV face-CONV standCONV shoot-FIN(?)/ATTR(?) EP Mato-kata (lit.: Target-shape) beach-GEN bright-NML The brightness of the beach at Matokata [that is like] a target (mato) which gentlemen facing [it] shoot at while standing, squeezing hunting arrows in [their] hands (FK 20)

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

536

Chapter 6

However, the interpretation of this poem and consequently the establishment of i-ru ‘shoot’ as the final predication form here faces several obstacles. First, the first part of the poem up to the place name Matokata is believed to be a jo-type makura-kotoba, which leaves room for doubting its grammaticality. Second, although the form i-ru ‘shoot’ should represent a final predication form whether we interpret the particle ya after i-ru as an emphatic particle or as an interrogative particle, the interpretation of the whole text with a final predication form rather than an attributive form modifying Matokata hardly makes any sense. Third, this poem comes from a fragment of the Ise Fudoki, which means that it may represent a dialect different from Western Old Japanese. Fourth, it is not inconceivable that the word ya after i-ru is not a particle, but the noun ya ‘arrow,’ which then will represent another play on words in this poem.22 In this case i-ru can only be an attributive form. Finally, it is difficult to argue for the existence of the final predication form -ru on the basis of a single example, which, in addition, is not perfect. There is, however, evidence for a different form of final predication for strong vowel verbs. It was already observed by Yamada Yoshio (who did not make any conclusions on the basis of this observation) that the form of the irregular vowel verb mi- ‘to see’ preceding the conjunction tǝmǝ ‘even if’ is just mi in Western Old Japanese (Yamada 1954: 154). It is well known that the conjunction tǝmǝ is preceded by a form of final predication, therefore Iwai Yoshio came to the correct conclusion that mi in mi tǝmǝ represents a Western Old Japanese final predication form (Iwai 1970: 53).23 Examples:

22  I am grateful to Matthew McNicoll who drew my attention to this possibility during a seminar on Western Old Japanese that I was teaching at the University of Hawai’i of Mānoa in the spring of 2007. 23  Both Yamada and Iwai also mention such specific Western Old Japanese forms as mi-ramu ‘see-TENT2-FIN,’ and mi-mbɛ-si ‘see-DEB-FIN,’ and ni-rasi ‘cook-SUP’ (cf. corresponding MJ mi-r-uram-u, mi-r-ube-si, and ni-r-urasi) (Yamada 1954: 153–154; Iwai 1970: 52–53). From the viewpoint of a traditional Japanese analysis, where -ramu, -besi, and -rasi are supposed to be auxiliaries following a final predication form, these forms may represent additional valuable evidence. However, these forms are not as valuable under a structural analysis of the Western Old Japanese verbal system. First, -mbɛ in mi-mbɛ-si certainly goes back to the adverb umbɛ ‘must, duly’; and, second -ramu and -rasi are likely to historically include the stative form -ur-, and not the final predication marker -u. Anyway, on the synchronic level both are analyzed better as -(u)ram- and -(u)rasi.

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

Verbs

537

萬代見友将飽八

YƏRƏⁿDU YƏ N-I MI tǝmǝ AK-AM-Ɛ ya ten thousand generation DV-CONV look(FIN) CONJ lose.interest-TENT-EV IP even if [I] look [at the palace] for ten thousand generations, would [I] lose interest [in looking at it?—No, I would not!] (MYS 6.921) 比祢毛須尓美等母安久倍伎宇良尓安良奈久尓

pinemosu n-i mi tǝmǝ ak-umbɛ-ki ura n-i ar-an-aku n-i all.day DV-CONV look(FIN) CONJ lose.interest-DEB-ATTR bay DV-CONV existNEG-NML DV-CONV even if [I] look all day, it is not a bay that [one] could lose interest [in looking at] (MYS 18.4037) 都婆吉都良々々尓美等母安可米也

tumbaki tura-tura n-i mi tǝmǝ ak-am-ɛ ya camellia intently DV-CONV look(FIN) CONJ lose.interest-TENT-EV IP even if [I] look intently [at] the camellia, would [I] lose interest? [No, I would not!] (MYS 20.4481) 之婆之婆美等母安加無伎弥加毛

simba-simba mi tǝmǝ ak-am-u kimi kamo often look(FIN) CONJ get.enough-TENT-ATTR lord EP even if [I] look [at you] often, would [I] get enough of [my] lord, I wonder? (MYS 20.4503) A couple of additional comments are in order. The underlying form of the final predication form for the strong vowel verb mi- ‘to see, to look’ is probably *mi-u, but this form has to be simplified since it contradicts the rules of Western Old Japanese phonotactics that prohibit vowel clusters. Because strong vowel verbs never lose their root vowel, it is the suffix of the final predication that has to go. The surface form in Western Old Japanese, is therefore, just mi, which formally looks exactly the same as the verbal root. Consequently, the Middle Japanese final predication form in -ru, which is identical to the attributive form, in all probability just represents a replacement of the final form by an attributive. It is well known that the end result of this replacement was the complete loss of final predication forms: a process that took centuries to complete. But it is quite apparent that strong vowel verbs alongside with consonant verbs was the first verbal class to be affected by this process in the history of Japanese.

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

538 chart 31

Chapter 6 Distribution of the allomorphs of the final predication suffix

verb class

allomorph

consonant verbs regular vowel verbs k-irregular verbs s-irregular verbs n-irregular verbs strong vowel verbs r-irregular verbs

-u -u -u -u -u -∅ -i

Almost the same picture is applicable to suffixes and bound auxiliaries: auxiliaries that historically incorporate the r-irregular verb ar- ‘to exist’ have the final predication suffix -i; all other auxiliaries and the majority of wordnon-final suffixes (with the exception of those that take predication markers identical to inflected adjectives) are followed by the predication suffix -u.24 Thus, the predication markers -u and -i are in complimentary distribution, which can be seen in Chart 32 below. It must be emphasized that a distinction between -u and -i in Western Old Japanese is formal rather than functional, since when an auxiliary that requires subsequent -i is followed by a suffix that requires -u in its turn, then the final predication suffix will be -u, and not -i, for example, a combination of -ker-, RETR + -aⁿz-, NEG + -u or -i, FIN, results in -ker-aⁿz-u RETR-NEG-FIN, and not in *-ker-aⁿz-i. Nevertheless, since -i follows essentially the stative verbs ar- ‘to exist,’ wor- ‘to exist, to stay, to sit,’ por- ‘to want’ and the derivatives of ar-, while -u predominantly is associated with action verbs, we cannot exclude the possibility that at the proto-Japonic level a distinction between -u and -i was of a functional nature. I will return to this problem below in the comparative section.

24  The allomorph -ru is not found after suffixes and auxiliaries.

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

539

Verbs

chart 32 Combinations of the final predication suffixes -u and -i with preceding suffixes and bound auxiliaries

suffixes and auxiliaries

-u

-i

tentative -(a)mtentative2 -(u)ramnegative -(a)ⁿzpassive -(a)ye-, -rayecausative -(a)simɛiterative -aphonorific -asperfective -teperfective -nretrospective -kerprogressive -erperfective-progressive -tar-

-am-u -(u)ram-u -(a)ⁿz-u -(a)y-u -(a)sim-u -ap-u -as-u -t-u -n-u – – –

– – – – – – – – – -ker-i -er-i -tar-i

The final predication suffix does not have to be the last morpheme in the sentence. It can be followed by various particles or conjunctions, such as the interrogative particle ya or the conjunction tǝmǝ ‘even though, even if.’ In the case of direct or reported speech it is always followed by forms of the defective verb tǝ- ‘to say’ and the appropriate verb of verbal or mental activity, if any. In most existing grammars of Old Japanese there is no definition of the function of final predication markers (Yamada 1954; Saeki 1959; Shirafuji 1987). It is likely that it is implied there that the final predication is the function. Iwai Yoshio, however, defines the main function of final predication markers as final predication (Iwai 1970: 11; 1981: 22), following the tradition of general histories of the Japanese language (Kobayashi 1936: 83–86; Yuzawa 1943: 45–46).25 Examples of the final predication suffix -u: 夜弊賀岐都久流

ya-pe-ŋ-gaki tukur-u eight-fold-DV(ATTR)-fence make-FIN [I] am making an eight-folded fence (KK 1) 25  On the controversy in defining a function of final predication markers in Middle Japanese see Vovin 2003: 197–198.

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

540

Chapter 6

那迦士登波那波伊布登母

nak-aⁿzi tǝ pa na pa ip-u tǝmǝ weep-NEG/TENT DV TOP you TOP say-FIN CONJ Even though you say that [you] would not weep (KK 4) 袁登賣爾多陀爾阿波牟登

wotǝme-ni taⁿda n-i ap-am-u tǝ maiden-DAT straight DV-CONV meet-TENT-FIN DV thinking to meet maidens face to face (KK 18) 佐泥牟登波阿禮波意母閇杼

sa ne-m-u tǝ pa are pa omǝp-ɛ-ⁿdǝ thus sleep-TENT-FIN DV TOP I TOP long-EV-CONC Although I long so much to sleep [with you] (KK 27) 許能美岐波和賀美岐那良受

kǝnǝ mi-ki pa wa-ŋga mi-ki nar-aⁿz-u this HON-rice.wine TOP I-POSS HON-rice.wine be-NEG-FIN This rice wine is not my rice wine (KK 39) 毛毛知陀流夜迩波母美由

momo-ti-ⁿ-tar-u ya nipa mǝ mi-y-u hundred-thousand-GEN-be.enough-ATTR house garden FP see-PASS-FIN [I] can see flourishing houses [and] gardens (lit: plentiful with hundreds and thousands houses and gardens are seen) (KK 41) 都奴賀能迦迩余許佐良布伊豆久迩伊多流

tunuŋga-nǝ kani yǝkǝ sar-ap-u iⁿduku-ni itar-u Tunuŋga-GEN crab side go.away-ITER-FIN where-LOC reach-FIN The crab from Tunuŋga goes all the time along the side[way]. [To] where will [it] arrive? (KK 42) 摩佐豆古和芸毛玖迩弊玖陀良須

Masaⁿduko wa-ŋg-imo kuni-pe kuⁿdar-as-u Masaⁿduko I-POSS-beloved province-DIR descend-HON-FIN Masaⁿduko, my beloved, goes towards [her] province (KK 52) 芝賀波能比呂理伊麻須波淤富岐美呂迦母

si-ŋga pa-nǝ pirǝr-i-imas-u opǝ kimi rǝ kamǝ it-POSS leaf-GEN are broad-CONV-HON-FIN great lord DV EP its leaves are broad, as the great lord (KK 57) Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

541

Verbs 加久能碁登那爾淤波牟登

ka-ku-nǝ ŋgǝtǝ na-ni op-am-u tǝ thus-CONV-GEN like name-LOC carry-TENT-FIN DV in order to perpetuate that [it] was like that (KK 97) 宇利波米婆胡藤母意母保由

uri pam-ɛ-mba ko-ⁿdǝmǝ omǝp-oy-u melon eat-EV-CON child-PLUR think-PASS-FIN When [I] eat melon, [I] cannot help thinking of [my] children (lit.: I suddenly think of my children) (MYS 5.802) 烏梅能波奈佐吉多留僧能能阿遠也疑波可豆良爾須倍久奈利爾家良受夜

uMƐ-nǝ pana sak-i-tar-u sǝnǝ-nǝ awo yaŋgï pa kaⁿdura n-i s-umbɛ-ku nar-in-i-ker-aⁿz-u ya plum-GEN blossom bloom-CONV-PERF/PROG-ATTR garden-GEN green willow TOP wig DV-CONV do-DEB-CONV become-CONV-PERF-CONV-RETR-NEGFIN IP Did not [it] become so that [we] should make [our] wigs out of the green willows in the garden where the plum blossoms have bloomed? (MYS 5.817) 布流由岐得比得能美流麻提烏梅能波奈知流

pur-u yuki tǝ pitǝ-nǝ mi-ru-maⁿde uMƐ-nǝ pana tir-u fall-ATTR snow DV person-GEN see-ATTR-TERM plum-GEN flower fall-FIN plum blossoms fall to such an extent that people will perceive them as falling snow (MYS 5.839) 等富都比等末都良能加波尓和可由都流

tǝpo t-u pitǝ matu[u]ra-nǝ kapa-ni waka [a]yu tur-u distant DV-ATTR person Matu-ura-GEN river-LOC young sweetfish angle-FIN people from far away (lit.: distant people) angle young sweetfish at the Matu-ura river (MYS 5.857) 飛立可祢都鳥尓之安良祢婆

TƏmB-I-TAT-I-kane-t-u TƏRI n-i si ar-an-e-mba fly-CONV-depart-CONV-NEG/POT(CONV)-PERF-FIN bird DV-CONV EP existNEG-EV-CON [I] could not fly away because [I] am not a bird (MYS 5.893)

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

542

Chapter 6

出波之利伊奈奈等思騰許良爾佐夜利奴

IⁿDE-pasir-i in-ana tǝ OMƏP-Ɛ-ⁿdǝ kǝ-ra-ni sayar-i-n-u exit(CONV)-run-CONV go-DES DV think-EV-CONC child-PLUR-DAT be prevented-CONV-PERF-FIN although [I] think that [I] would like to run away, [I] am prevented by [my] children (MYS 5.899) 安麻乎等女等母思麻我久流見由

ama wotǝME-ⁿdǝmǝ sima-ŋ-kakur-u MI-y-u fisher maiden-PLUR island-LOC-hide-ATTR see-PASS-FIN [I] see fisher maidens hiding in [the shadow of] the island (MYS 15.3597) 安伎乃野尓草乎思香奈伎都

aki-nǝ NO-ni sa-wo-sika nak-i-t-u autumn-GEN field-LOC PREF-male-deer cry-CONV-PERF-FIN male deer cried in the autumn field (MYS 15.3678) 伊波多野爾夜杼里須流伎美伊敞妣等乃伊豆良等和礼乎等波婆伊可爾 伊波牟

Ipata-NO-ni yaⁿdǝr-i s-uru kimi ipe-m-bitǝ-nǝ iⁿdu-ra tǝ ware-wo tǝp-amba ika n-i ip-am-u Ipata-field-LOC lodge-NML do-ATTR lord home-GEN-person-GEN where-LOC DV I-ACC ask-COND how DV-CONV say-TENT-FIN [Oh, my] lord who lodged at the Ipata field. If people from [your] home ask me (saying) where [are you], what should [I] answer? (MYS 15.3689) 伊豆礼能日麻弖安礼古非乎良牟

iⁿdure n-ǝ PI-maⁿde are kopï-wor-am-u which DV-ATTR day-TERM I long.for(CONV)-exist-TENT-FIN until what day should I be longing for [you]? (MYS 15.3742) 和可伎兒等毛波乎知許知爾佐和吉奈久良牟

waka-ki KO-ⁿdǝmo pa woti kǝti-ni sawak-i-nak-uram-u young-ATTR child-PLUR TOP there here-LOC make.noise-CONV-cry-TENT2-FIN young children will probably cry loudly here [and] there (MYS 17.3962)

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

543

Verbs 佐夜麻太乃乎治我其日爾母等米安波受家牟

sa Yamaⁿda n-ǝ woⁿdi-ŋga SƏNƏ PI-ni mǝtǝmɛ ap-aⁿz-u-k-em-u so Yamaⁿda DV-ATTR old man-POSS that day-LOC search(CONV) meet-NEGCONV-PAST/FIN-TENT-FIN So, old man Yamaⁿda searched for [him] on that day, but did not find [him] (MYS 17.4014) 伊久欲布等余美都追伊毛波和礼麻都良牟曾

iku yo p-u tǝ yǝm-i-tutu imo pa ware mat-uram-u sǝ how many night pass-FIN DV count-CONV-COOR beloved TOP I wait-TENT2ATTR FP [My] beloved will probably wait for me, counting: ‘How many nights have passed?’ (MYS 18.4072) 和期於保伎美余思努乃美夜乎安里我欲比賣須

wa-ŋgǝ opo kimi Yǝsino-nǝ miya-wo ari-ŋgayop-i mes-u I-POSS great lord Yǝsino-GEN palace-ACC ITER-go.back.and.forth-CONV look (HON)-FIN My emperor constantly visits the palace in Yǝsino, and looks [around] (MYS 18.4099) 曽己由惠尓情奈具也

sǝkǝ yuwe n-i KƏKƏRƏ naŋg-u ya there reason DV-CONV heart calm.down-FIN IP will [my] heart calm down due to those circumstances? (MYS 19.4154) 秋風尓比毛等伎安氣奈多太奈良受等母

AKI KAⁿZE-ni pimo tǝk-i-akɛ-na taⁿda nar-aⁿz-u tǝmǝ autumn wind-LOC cord untie-CONV-open-DES direct be-NEG-FIN CONJ [I] wish that the autumn wind would untie the cords, even if [it] is not directly (MYS 20.4295) 阿止乎美都都志乃波牟

atǝ-wo mi-tutu sinǝp-am-u footstep-ACC see(CONV)-COOR yearn-TENT-FIN looking at [Buddha’s] footstep, [I] will yearn [for him] (BS 6) 己礼乃与波宇都利佐留止毛

kǝre n-ǝ yǝ pa utur-i sar-u tǝmo this DV-ATTR world TOP change-CONV go away-FIN CONJ Even though this world changes and goes away … (BS 10) Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

544

Chapter 6

朕高御座爾坐始由理今年尓至麻低六年尓成奴

WARE TAKA MI-KURA-ni IMAS-I-SƏMƐS-U-yuri KƏ TƏSI-ni ITAR-U-maⁿde MU TƏSI n-i NAR-I-n-u I high HON-seat-LOC be(HON)-CONV-begin-ATTR-ABL this year-LOC reachATTR-TERM six year DV-CONV become-CONV-PERF-FIN [It] has been six years this year since I have been on the high throne (SM 7) 兵発之武

IKUSA OKƏS-Asim-u army raise-CAUS-FIN [Nakamarǝ] made armies to raise (SM 28) 仲末呂伊忠臣止之天侍都

Nakamarǝ-i TAⁿDASI-KI OMI tǝ s-i-te PAmBER-I-t-u Nakamarǝ-ACT loyal-ATTR noble DV do-CONV-SUB serve-CONV-PERF-FIN Nakamarǝ served as a loyal noble (SM 34) 和己於保支美波多比良気久那何久伊末之弖等与美岐麻都流

wa-ŋgǝ opo kimi pa tapirakɛ-ku naŋga-ku imas-i-te tǝyǝ mi-ki matur-u I-POSS great lord TOP safe-CONV long-CONV exist(HON)-CONV-SUB abundant HON-rice.wine present(HUM)-FIN [I] present the abundant rice wine so that my sovereign (lit.: great lord) [would] live safely and long (SNK 4) Examples of the final predication suffix -i: 故志能久邇邇佐加志売遠阿理登岐加志弖久波志売遠阿理登岐許志弖

Kosi-nǝ kuni-ni sakasi me-wo ar-i tǝ kik-as-i-te kupasi me-wo ar-i tǝ kik-ǝs-i-te Kosi-GEN province-LOC wise woman-ABS exist-FIN DV hear-HON-CONV-SUB beautiful woman-ABS exist-FIN DV hear-HON-CONV-SUB [Opo kuni nusi] heard that there is a wise woman in the Kosi province, heard that there is a beautiful woman (KK 2) 阿加陀麻波袁佐閇比迦禮杼斯良多麻能岐美何余曾比斯多布斗久阿理祁理

aka-ⁿ-dama pa wo sapɛ pikar-e-ⁿdǝ sira tama-nǝ kimi-ŋga yǝsǝpi si taputo-ku ar-i-ker-i red-DV(ATTR)-jewel TOP cord RP shine-EV-CONC white jewel-COMP lordPOSS adorned.appearance EP revered-CONV exist-CONV-RETR-FIN Although even the cord of red jewels shines, [I] realized [that I] feel reverence [for my] lord’s adorned appearance, which is like a white jewel (KK 7)

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

545

Verbs

本都延波阿米袁淤弊理那加都延波阿豆麻袁淤弊理志豆延波比那袁淤弊理

po-tu ye pa amɛ-wo op-er-i naka-tu ye pa Aⁿduma-wo op-er-i siⁿ-tu ye pa pina-wo op-er-i top-GEN/LOC branch TOP heaven-ACC cover-PROG-FIN middle-GEN/LOC branch TOP lands.in.the.east-ACC cover-PROG-FIN bottom-GEN/LOC branch TOP rural.region-ACC cover-PROG-FIN [Its] top branches are covering the Heaven, [its] middle branches are covering the lands in the East, and [its] lower branches are covering the rural regions (KK 100) 意富美夜能袁登都波多傅須美加多夫祁理

opǝ-miya-nǝ wotǝ t-u pataⁿde sumi katambuk-er-i great-place-GEN that DV-ATTR edge(?) corner incline-PROG-FIN The edge corners of that side of the great palace are falling apart (KK 105) 阿軻娜磨廼比訶利播阿利登比登播伊珮耐

aka-ⁿ-dama-nǝ pikari pa ar-i tǝ pitǝ pa ip-ɛ-ⁿdǝ red-DV(ATTR)-jewel-GEN light TOP exist-FIN DV person TOP say-EV-CONC Although people say that the red jewel has light (NK 6) 余能奈可波牟奈之伎母乃等志流等伎子伊与余麻須万須加奈之可利家理

yǝ-nǝ naka pa munasi-ki mǝnǝ tǝ sir-u tǝki si iyǝyǝ masu-masu kanasi-k-ar-i-ker-i world-GEN inside TOP empty-ATTR thing DV know-ATTR time EP more.and. more more.and.more sad-CONV-exist-CONV-RETR-FIN When [I] realized that the world is empty, [it] turned out to be more and more sad (MYS 5.793) 比左可多能月者弖利多里

pisa kata n-ǝ TUKÏ pa ter-i-tar-i long hard DV-ATTR moon TOP shine-CONV-PERF/PROG-FIN The eternal and hard moon is shining (MYS 15.3672) 烏梅乃花美夜万等之美尓安里登母

uMƐ-nǝ PANA mi-yama tǝ sim-i n-i ar-i tǝmǝ plum-GEN blossom HON-mountain DV grow.thick-NML DV-CONV exist-FIN CONJ Even though plum blossoms are blooming densely as a mountain (MYS 17.3902)

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

546

Chapter 6

伊米尓波母等奈安比見礼騰多太尓安良祢婆孤悲夜麻受家里

imɛ-ni pa mǝtǝna api-mi-re-ⁿdǝ taⁿda n-i ar-an-e-mba kopï yam-aⁿz-u-ker-i dream-LOC TOP in.vain REC-see-EV-CONC direct DV-CONV exist-NEG-EV-CON love(NML) stop-NEG-CONV-RETR-FIN although [we] see each other in vain in dreams, because [our meetings] are not direct, [our] longing does not stop (MYS 17.3980) 和礼爾於止礼留比止乎於保美和多佐牟多米止宇都志麻都礼利

ware-ni otǝr-er-u pitǝ-wo opo-mi watas-am-u tamɛ tǝ utus-i-matur-er-i I-DAT be.inferior-PROG-ATTR person-ABS many-GER lead across-TENT-ATTR in.order.to DV carve-CONV-HUM-PROG-FIN because there are many people who have been inferior than me, [I] have carved [Buddha’s footprints] in order to save [them] (BS 13) 久須理師波都祢乃母阿礼等麻良比止乃伊麻乃久須理師多布止可理家利米 太志加利鶏利

kusurisi pa tune n-ǝ mǝ ar-e-ⁿdǝ marapitǝ n-ǝ ima-nǝ kusurisi taputǝ-k-ar-i-ker-i mɛⁿdasi-k-ar-i-ker-i medicine man TOP usual DV-ATTR FP exist-EV-CONC guest DV-ATTR nowGEN medicine man revered-CONV-exist-CONV-RETR-FIN praiseworthyCONV-exist-CONV-RETR-FIN Although there are usual medicine men, too, the present Guest Medicine Man is [indeed] revered. [He] is praiseworthy (BS 15) 汝多知諸者吾近姪奈利

IMASI-tati MƏRƏ pa WA-ŋGA TIKA-KI WOPI nar-i you-PLUR all TOP I-POSS close-ATTR nephew be-FIN All [of] you are my close nephews (SM 17) Comparative Data Level A: Other Japonic A1: Eastern Old Japanese The final predication markers -u and -i are also found in Eastern Old Japanese. I cannot observe any functional difference between their usage in Western and Eastern Old Japanese. Examples of the final predication suffix -u: 安素乃河泊良欲伊之布麻受蘇良由登伎奴与

Aso-nǝ KApara-yo isi pum-aⁿz-u sora-yu tǝ k-i-n-u yǝ

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

547

Verbs

Aso-GEN river-bed-ABL stone tread-NEG-CONV sky-ABL DV come-CONVPERF-FIN EP [I] came from the river-bed of Aso, as from the sky, without treading on stones! (MYS 14.3425) 麻等保久能野尓毛安波奈牟

ma-tǝpo-ku n-ǝ NO-ni mo ap-ana-m-u INT-distant-CONV DV-ATTR field-LOC FP meet-DES-TENT-FIN [I] would like to meet [you] even in a distant field (MYS 14.3463) 多知和可礼伊爾之与比欲利世呂爾安波奈布与

tat-i-wakare in-i-si yǝpi-yori se-rǝ-ni ap-an-ap-u yǝ depart-CONV-part(CONV) go.away-CONV-PAST/ATTR night-ABL beloved-DIMDAT meet-NEG-ITER-FIN EP from the night when [we] parted and [he] went away, [I] never met [my] beloved! (MYS 14.3375) 麻登保久於毛保由

ma-tǝpo-ku omop-oy-u INT-far-CONV think-PASS-FIN [it] suddenly seems to be very far away (MYS 14.3522) 麻多妣爾奈理奴

ma-tambi n-i nar-i-n-u INT-journey DV-CONV become-CONV-PERF-FIN [it] became a really [long] journey (MYS 20.4388) Examples of the final predication suffix -i: 阿志氣比等奈里

asi-kɛ pitǝ nar-i bad-ATTR person be-FIN [he] is a bad person (MYS 20.4382) 以弊乃母加枳世之己呂母尓阿加都枳尓迦理

ipe-nǝ [i]mǝ-ŋga ki-se-si kǝrǝmǝ-ni aka tuk-i-n-i-kar-i home-GEN beloved-POSS wear-CAUS(CONV)-PAST/ATTR garment-LOC dirt attach-CONV-PERF-CONV-RETR-FIN dirt stuck to the garment that my beloved at home made [me] wear (MYS 20.4388)

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

548

Chapter 6

A2: Ryukyuan Ryukyuan comparative data are crucial in three respects. First, there is no distinction between the allomorphs -u and -i, which we observed above in both Western and Eastern Old Japanese. Thus, Ryukyuan does not offer any evidence for the stative / active distinction between -i and -u found in Western Old Japanese. It is equally possible that either of these two languages underwent an innovation: either Ryukyuan losing it or Western Old Japanese acquiring it. The second solution may be more viable if we view this development in Western Old Japanese as a structural change under the influence of Korean. Second, although the final predication suffix is mostly spelled as -u ~ -yu in Old and Classical Ryukyuan, modern dialects clearly indicate the PR form *-um, with a final *-m which can be safely reconstructed both on the basis of the reflexes in dialects, cf., e.g., Koniya ’iky-um ‘go-FIN’ vs. Shuri ’ich-uN ‘id.’ as well as on Shuri morphophonemic rules, cf. the Shuri interrogative form ’ich-um-i go-FIN-QS ‘will you go?’ with the interrogative suffix -i before which we can see the underlying form *-um. Third, and most importantly, the palatalization k > ch in Shuri, the presence of -y- in the Koniya form, and, finally the peculiar spellings in Old and Classical Ryukyuan, for example in RK 3888 below, all indicate that historically the final predication suffix *-um is an auxiliary that followed the converb form in -i-. In all probability, we can reconstruct the following development from proto-Japonic to Western Old Japanese and Shuri: PJ → Shuri →

*yik-i-um → WOJ go-CONV-FIN → ’ich-uN go-FIN

yuk-u go-FIN

Examples: Old Ryukyuan さいわたるのさくらしけしけとおりさちへけおよりあいいでらむ

sa-i-watar-u n-o sakura sike-sike to or-i-sat-ife keo-yori ai-ide-ram-u bloom-CONV-cross-ATTR DV-ATTR sakura dense-dense DV bend-CONVstretch-CONV today-ABL COOP-go.out-TENT2-FIN [Sailors!] From today [you] should go out together bundling together [like] blooming sakura [trees] (OS 10.531) Classical Ryukyuan いろいろに言ちもいかなしも行かぬ Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

549

Verbs

iro-iro n-i I-ti-mo ika nas-i-mo Ik-an-u different DV-CONV say(CONV)-SUB-PT how do-CONV-PT go-NEG-FIN whatever [you] say, and whatever [you] do, [I] will not go (RK 725) 北京お主てだやずまにそなれゆが

FICIN o-SHU-teda ya zuma-ni so nar-e-yu ga Ficin-GEN HON-lord-sun TOP where-LOC next be-CONV-FIN IP Where does the emperor of Beijing live? (RK 3888) Shuri

’ich-uN di ’yu-ta-N go-FIN DV say-PAST-FIN [He] said that [he] would go (RKJ 435) husi-nu chura-sa-N star-NOM beautiful-NML-FIN The star is beautiful (Nishioka and Nakahara 2000: 10) kunu maNgaa umu-sa-N doo this cartoon interesting-NML-FIN FP this cartoon is interesting (Nishioka and Nakahara 2000: 18) Kumejima

yaa-naa wu-N yoo house-LOC exist-FIN EP [I] am at home (Nohara 1986: 126) Miyako

taka-k-a-m ko:na high-CONV-exist-FIN buy/NEG/IMP [It] is expensive. Do not buy [it] (Karimata 1997a: 400) One of the etymological explanations that goes back to Hattori Shirō, connects this *-um with the verb wor- ‘to exist.’ If we take into consideration the Ryukyuan final forms of this verb, such as, for example, Shuri wuN, the analogy may seem perfect. But unfortunately this etymology has too many problems. First, the Shuri form itself goes back to *wor-i-um with expected and regular development *wor-i-um > *wu-i-um > wuN. Second, the loss of intervocalic -wAlexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

550

Chapter 6

does not occur at all in Western Old Japanese, and the reliable cases of the loss of -r- are found in Western Old Japanese or pre-Western Old Japanese only in some verbal morphological markers, such as the attributive or evidential (see chapter 2, section 2.5.4). Finally, and most importantly, Shuri wuN < *wor-i-um still includes the same morpheme of final predication -N < *-um, and, therefore, the whole explanation becomes completely circular. Thus, at the present stage of our knowledge we can only conclude that *-um is likely to be an obsolete auxiliary, but we do not know anything about its origin or etymology. 3.2.1.2 Attributive -uru ~ -u ~ -ru ~ -ǝ The attributive suffix’s main allomorph is -uru, which is found after all vowel verbs (except strong vowel verbs, where it becomes just -ru), and all irregular verbs (except r-irregular verbs). After all consonant verbs and r-irregular verbs the allomorph -u is used, according to the morphonological rule of -rloss mentioned in chapter 2, section 2.5.4. The defective verb n- has a special attributive ending in -ǝ: n-ǝ, and the defective verb t- has the attributive -u like consonant verbs. chart 33 Distribution of the allomorphs of the attributive suffix

verb class

allomorph

consonant verbs regular vowel verbs k-irregular verbs s-irregular verbs n-irregular verbs strong vowel verbs r-irregular verbs defective verb tdefective verb n-

-u -uru -uru -uru -uru -ru -u -u -ǝ

A similar picture is observed if -uru is used not after a verbal stem but after another verbal suffix or an auxiliary: if a suffix or an auxiliary ends with a vowel, the main allomorph -uru is used, and the final vowel of the preceding suffix or an auxiliary is apocopated. If a suffix or an auxiliary ends in a consonant, the allomorph -u is used.26 The only exception is the perfective auxiliary -n- that has the attributive form -n-uru. The causative suffix -as- ~ -(a)se-, which may end either in a consonant or in a vowel, has accordingly either -as-u or -as-uru. 26  The allomorph -ru is not used after suffixes and auxiliaries. Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

551

Verbs chart 34 Combinations of the attributive suffixes -uru and -u with preceding suffixes and bound auxiliaries

suffixes and auxiliaries

-uru

-u

tentative -(a)mtentative2 -(u)ramnegative -(a)npassive -(a)ye-, -rayecausative -(a)simɛcausative -as- ~ -(a)seiterative -aphonorific -asperfective -teperfective -nretrospective -kerprogressive -erperfective-progressive -tar-

– – – -(a)y-uru -(a)sim-uru -as-u – – -t-uru -n-uru – – –

-(a)m-u -(u)ram-u -(a)n-u – – -as-uru -ap-u -as-u – – -ker-u -er-u -tar-u

The attributive form has three main functions: (1) a modifying function; (2) a verbal noun function (nominalized form); and (3) a final predicate function. 3.2.1.2.1 Attributive as a Modifier There are two types of modifiers with an attributive function: simple, when an attributive form constitutes a sentence of its own and modifies the following nominal; and extended, when the attributive represents a predicate in an extended sentence which in this case as a whole becomes a modifier of the nominal following the attributive. 都流岐能多知

turuki n-ǝ tati double-edged sword DV-ATTR long.sword a long sword that is a double-edged sword (KK 33) 賣杼理能和賀意富岐美能淤呂須波多

Meⁿdǝri n-ǝ wa-ŋga opǝ kimi-nǝ or-ǝs-u pata Meⁿdǝri DV-ATTR I-POSS great lady-GEN weave-HON-ATTR fabric The fabric that my lady Meⁿdǝri weaves (KK 66)

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

552

Chapter 6

柂摩儺羅磨婀我裒屢柂摩

tama nar-amba a-ŋga por-u tama jewel be-COND I-POSS desire-ATTR jewel if [my beloved] were a jewel, [she would be] a jewel I desire (NK 92) 企許斯遠周久爾能麻保良叙

kikǝs-i-wos-u kuni-nǝ ma-po-ra ⁿzǝ rule(HON)-CONV-HON-ATTR country-GEN INT-top-LOC FP in the highest place of the country, where [the emperor] rules (MYS 5.800) 意比久留母能波毛毛久佐爾勢米余利伎多流

op-i-k-uru mǝnǝ pa momo kusa n-i semɛ-yǝr-i-k-i-tar-u pursue-CONV-come-ATTR thing TOP hundred kind DV-CONV assault (CONV)-approach-CONV-come-CONV-PERF/PROG-ATTR the things that pursue [us], come assaulting [us] in a hundred varieties (MYS 5.804) 和礼由惠尓於毛比和夫良牟伊母我可奈思佐

ware yuwe n-i omop-i-wamb-uram-u imǝ-ŋga kanasi-sa I reason DV-CONV think-CONV-worry-TENT2-ATTR beloved-POSS be.dear-NML [feeling of] endearment for [my] beloved who probably worries because of me (MYS 15.3727) 家尓底母多由多敷命

IPE n-i-te mǝ tayutap-u INƏTI home DV-CONV-SUB FP be.unstable-ATTR life [my] life which is uncertain even at home (MYS 17.3896) 此橘乎等伎自久能可久能木實等名附家良之母

KƏNƏ TATImBANA-wo tǝkiⁿzi-ku n-ǝ kaŋg-u n-ǝ KƏ-NƏ MÏ tǝ NA-ⁿ-DUKƐker-asi-mǝ this mandarin.orange-ACC be off season-CONV DV-ATTR smell-ATTR DV-ATTR tree-GEN fruit DV name-LOC-attach(CONV)-RETR-SUP-EXCL [we] should call these mandarin oranges fragrant tree fruits that are off season! (MYS 18.4111) Note that in this example the first n-ǝ follows the converb form -ku. The usage of the second attributive n-ǝ in this example is idiosyncratic and even possibly ungrammatical, since the attributive modifies the following noun by itself, and n-ǝ is unnecessary.

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

553

Verbs 阿佐奈佐奈安我流比婆理爾奈里弖之可

asa-na [a]sa-na aŋgar-u pimbari n-i nar-i-te-si ka morning-PLUR morning-PLUR rise-ATTR skylark DV-CONV become-CONVPERF-PAST/ATTR EP every morning, [I] want to have become a skylark, flying up (MYS 20.4433) 佐久波奈波宇都呂布等伎安里

sak-u pana pa utur-ǝp-u tǝki ar-i bloom-ATTR flower TOP wither-ITER-ATTR time exist-FIN There is a time when blooming flowers will be withering (MYS 20.4484) 与都乃閇美伊都都乃毛乃乃阿都麻礼流伎多奈伎微乎婆

yǝ-tu n-ǝ pɛmi itu-tu n-ǝ monǝ-nǝ atumar-er-u kitana-ki mï-womba four-CL DV-ATTR snake five-CL DV-ATTR demon-GEN gather-PROG-ATTR be.dirty-ATTR body-ACC(EMPH) the dirty body where four snakes and five demons have accumulated (BS 19) 逆在流人止母在而

SAKASIMA NAr-u PITƏ-ⁿdǝmǝ AR-I-TE rebellious be-ATTR person-PLUR exist-CONV-SUB there were people who were rebellious (SM 16) 清麻呂其我姉法均止甚大尓悪久奸流妄語乎作弖

KIYOMARƏ SI-ŋga ANE POPUKUN-tǝ ITƏ OPO-KI n-i ASI-ku KAⁿDAM-YEr-u ITUPAR-I-ŋ-GƏTƏ-wo TUKUR-I-te Kiyomarǝ he-POSS elder.sister Popukun-COM very be.big-ATTR DV-CONV badCONV be.insincere-PROG-ATTR lie-NML-GEN-word-ACC make-CONV-SUB Kiyomarǝ with his elder sister Popukun created an extremely big, bad and insincere lie … (SM 44) 3.2.1.2.2 Attributive as a Nominalized Form Either the attributive itself functions as a verbal noun or the whole sentence where the attributive serves as the final predicate can function as one nominalized form. In this function the attributive can be followed by various case markers: the focus particles pa and mǝ, conjunctions, and the defective verb n-.

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

554

Chapter 6

伊麻許曾婆和杼理迩阿良米能知波那杼理爾阿良牟遠

ima kǝsǝ pa wa-ⁿ-tǝri n-i ar-am-ɛ nǝti pa na-ⁿ-tǝri n-i ar-am-u-wo now FP TOP I-OSM-bird DV-CONV exist-TENT-EV after TOP you-OSM-bird DV-CONV exist-TENT-ATTR-ACC Now [I] am my bird, later [I] will be your bird, so … (KK 3) 多久夫須麻佐夜具賀斯多爾

taku-m-busuma sayaŋg-u-ŋga sita-ni taku-GEN-cover rustle-ATTR-POSS bottom-LOC under the rustling of the taku covers (KK 5) 道乎多遠見思空安莫國嘆虚不安物乎

MITI-wo taⁿ-DƏPƏ-mi OMƏP-U sora YASUKƐ NA-ku n-i NA ŋGƐK-U SORA YASUK-AR-AN-U MƏNƏwo way-ABS PREF-far-GER think-ATTR RP easy no-CONV DV-CONV lament-ATTR RP easy-CONV-exist-NEG-ATTR CONJ although it is not easy even to lament and to love because the way is far (MYS 4.534) 布流由岐得比得能美流麻提烏梅能波奈知流

pur-u yuki tǝ pitǝ-nǝ mi-ru-maⁿde uMƐ-nǝ pana tir-u fall-ATTR snow DV person-GEN see-ATTR-TERM plum-GEN flower fall-FIN plum blossoms fall to such an extent that people will perceive [them] as falling snow (MYS 5.839) 毛呂比登能阿蘇夫遠美礼婆

morǝ-pitǝ-nǝ asomb-u-wo mi-re-mba all person-GEN enjoy-ATTR-ACC see-EV-CON when [I] see all the people enjoying [themselves] (MYS 5.843) 豊乃登之思流須登奈良思雪能敷礼流波

TƏYƏ n-ǝ tǝsi sirus-u tǝ nar-asi YUKI-nǝ pur-er-u pa abundant DV-ATTR year show.a.sign-ATTR DV become-SUP snow-GEN fallPROG-ATTR TOP A snowfall seems to become a good omen for an abundant year (MYS 17.3925)

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

555

Verbs 和藝毛故尓美勢牟我多米尓母美知等里氐牟

wa-ŋg-imo-ko-ni mi-se-m-u-ŋga tamɛ n-i mǝmit-i tǝr-i-te-m-u I-POSS-beloved-DIM-DAT see-CAUS-TENT-ATTR-POSS for DV-CONV leaves. turn.red/yellow-NML take-CONV-PERF-TENT-FIN [I] want to take red leaves in order to show [them] to my beloved (MYS 19.4222) 阿止乎美都都志乃波牟多太爾阿布麻弖爾麻佐爾阿布麻弖爾

atǝ-wo mi-tutu sinǝp-am-u taⁿda n-i ap-u-maⁿde-ni masa n-i ap-u-maⁿde-ni footstep-ACC see(CONV)-COOR yearn-TENT-FIN direct DV-CONV meet-ATTRTERM-LOC real DV-CONV meet-ATTR-TERM-LOC looking at [Buddha’s] footstep, [I] will yearn [for him], until [I] meet [him] directly, until [I] really meet [him] (BS 6) 令文所載多流乎跡止為而

NƏRI-NƏ PUMI-NI NƏSE-tar-u-wo ATƏ tǝ S-I-TE law-GEN scripture-LOC place(CONV)-PERF/PROG-ATTR-ACC FOOTSTEP DV DO-CONV-SUB taking what is written in the law scriptures as a precedent (SM 2) 朕高御座爾坐始由理今年尓至麻低六年尓成奴

WARE TAKA MI-KURA-ni IMAS-I-SƏMƐS-U-yuri KƏ TƏSI-ni ITAR-U-maⁿde MU TƏSI n-i NAR-I-n-u I high HON-seat-LOC be(HON)-CONV-begin-ATTR-ABL this year-LOC reachATTR-TERM six year DV-CONV become-CONV-PERF-FIN It has been six years (reaching to) this year since I have been on the high throne (SM 7) 3.2.1.2.3 Attributive as a Final Predicate If the particles sǝ ~ ⁿzǝ, ya, ka, namo27 are found anywhere in a sentence before the final verb, then the final predication suffix -u is automatically replaced with the attributive suffix -uru (or other allomorphs of the attributive suffix). The same rule applies to other attributive forms that do not involve -uru or its allomorphs, e.g., the past attributive -si or the adjectival attributive -ki. This rule is known in the traditional grammar as 係り結び kakari-musubi ‘the rule of linking.’ It is necessary to note that the particle namo practically does not appear in 27  Since the MJ cognate is namu, and the otsu-rui vowel /ǝ/ does not raise to /u/, we can conjecture that the pre-OJ form was *namo with kō-rui vowel /wo/.

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

556

Chapter 6

the Western Old Japanese poetic texts.28 All examples of its usage are limited to prose texts, and more exactly only to the Senmyō. Strangely enough, although the Norito is also a prose text, namo is not attested there. One more important observation is that both the particles ka and kamǝ trigger the change of the final predicative form to an attributive when they are found both before and after a verb, although this change which occurs with ka and kamǝ following the verb is not traditionally classified as a kakari-musubi. 牟迦比袁流迦母伊蘇比袁流迦母

mukap-i-wor-u kamǝ i-sop-i-wor-u kamǝ face-CONV-exist-ATTR EP DLF-snuggle-CONV-exist-ATTR EP Oh, [she] is facing [me]! Oh, [she] is snuggling with [me]! (KK 42) 伊岐良受曾久流

i-kir-aⁿz-u sǝ k-uru DLF-cut-NEG-CONV FP come-ATTR [I] come back without cutting [them] there (KK 51) 阿袁那母岐備比登登等母迩斯都米婆多怒斯久母阿流迦

awo na mǝ Kimbï pitǝ-tǝ tǝmǝ n-i si tum-ɛ-mba tanosi-ku mǝ ar-u ka green vegetables FP Kimbï person-COM together DV-CONV EP pick-EV-CON delightful-CONV FP exist-ATTR IP Is [not] it delightful when [I] pick green vegetables together with the girl from Kimbï? (KK 54) 那爾騰柯母于都倶之伊母我磨陀左枳涅渠農

nani tǝ kamǝ utukusi imǝ-ŋga mata sak-i-[i]ⁿde-kǝ-n-u what DV EP beautiful beloved-POSS again bloom-CONV-exit(CONV)-comeNEG-ATTR for (lit: being) what [reason], does [my] beautiful beloved not bloom again? (NK 114) 京師尓而誰手本乎可吾将枕

MIYAKO n-i-TE TA-ŋGA TAMƏTƏ-wo ka WA-ŋGA MAKURAK-AM-U capital DV-CONV-SUB who-POSS sleeve-ACC IP I-POSS use.as.a.pillowTENT-ATTR Whose sleeves will I use as a pillow at the capital? (MYS 3.439) 28  With the exception of a single attestation in MYS 12.2877 cited below. MJ namu does not appear in poetic texts of the Heian period either (Iwai 1981: 207).

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

557

Verbs 伊豆久由可斯和何伎多利斯

iⁿduku-yu ka siwa-ŋga k-i-tar-i-si where-ABL IP wrinkle-POSS come-CONV-PERF/PROG-CONV-PAST/ATTR Where did the wrinkles come from? (MYS 5.804) In this example we have the past attributive -si that is used as a form of final predication after the particle ka. 今者春部登成尓鷄類鴨

IMA pa PARU-pɛ tǝ nar-i-n-i-ker-u kamo now TOP spring-? DV become-CONV-PERF-CONV-RETR-ATTR EP It turned out that now [it] is (lit.: has become) spring (MYS 8.1433) 戀友何如妹尓相時毛名寸

KOP-URE-ⁿdǝmǝ NANI si ka IMO-ni AP-U TƏKI mo na-ki love-EV-CONC what EP IP beloved-DAT meet-ATTR time FP no-ATTR although [I] love [her], why is there no time at all to meet [my] beloved? (MYS 12.2994) In this example we have the adjectival attributive -ki that is used as a form of final predication after the preceding interrogative particle ka. 妹等安里之時者安礼杼毛和可礼弖波許呂母弖佐牟伎母能尓曽安里家流

IMO-tǝ ar-i-si TƏKI pa ar-e-ⁿdǝmo wakare-te pa kǝrǝmǝⁿde samu-ki mǝnǝ n-i sǝ ar-i-ker-u beloved-COM exist-CONV-PAST/ATTR time TOP exist-EV-CONC separate(CONV)SUB TOP sleeve cold-ATTR thing DV-CONV FP exist-CONV-RETR-ATTR Although there was a time when [I] was with [my] beloved, since [we] separated, [my] sleeves are cold (MYS 15.3591) Note that although in this example focus particle sǝ is inside the noncontracted form of the copula, it still triggers the change of the final form to the attributive one. 伊毛尓安礼也夜須伊毛祢受弖安我故非和多流

imo n-i ar-e ya yasu i mo ne-ⁿz-u-te a-ŋga kopï-watar-u beloved DV-CONV exist-EV IP easy sleep IP sleep-NEG-CONV-SUB I-POSS love(CONV)-cross-ATTR Is [she] my beloved? I do not sleep easily, and continue to love [her] (MYS 15.3633)

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

558

Chapter 6

和伎毛故波伊都登加和礼乎伊波比麻都良牟

wa-ŋg-imo-ko pa itu tǝ ka ware-wo ipap-i mat-uram-u I-POSS-beloved-DIM TOP when DV IP I-ACC pray-CONV wait-TENT2-ATTR My beloved will probably wait for me, praying [to the gods], and thinking: ‘When [will he return]?’ (MYS 15.3659) 秋夜乎奈我美爾可安良武

AKI-NƏ YO-wo naŋga-mi n-i ka ar-am-u autumn-GEN night-ABS be.long-GER DV-CONV IP exist-TENT-ATTR Is [it] probably because the autumn night is long (MYS 15.3684) Note that although in this example the interrogative particle ka is inside the non-contracted form of the copula, it still triggers the change of the final form to the attributive one. 之路髪麻泥尓大皇尓都可倍麻都礼婆貴久母安流香

siro KAMI-maⁿde-ni OPO KIMI-ni tukapɛ-matur-e-mba TAPUTO-ku mǝ ar-u ka white hair-TERM-LOC great lord-DAT serve(CONV)-HUM-EV-CON be.awesomeCONV FP exist-ATTR IP When [one] serves [his] sovereign until grey hair, is [it not] awesome? (MYS 17.3922) 布流雪乃比加里乎見礼婆多敷刀久母安流香

pur-u YUKI-nǝ pikari-wo MI-re-mba taputo-ku mǝ ar-u ka fall-ATTR snow-GEN light-ACC see-EV-CON be.awesome-CONV FP exist-ATTR IP When [one] sees the light of the falling snow, is [it not] awesome? (MYS 17.3923) 伊都之加登奈気可須良牟曾

itu si ka tǝ naŋgɛk-as-uram-u sǝ when EP IP DV lament-HON-TENT2-ATTR FP [she] probably laments, saying: ‘When [will he return]?’ (MYS 17.3962) 石乎毛珠等曽吾見流

ISI-wo mo TAMA tǝ sǝ WA-ŋGA MI-ru stone-ACC FP jewel DV FP I-POSS see-ATTR I view stones as jewels, too (MYS 19.4199)

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

559

Verbs 多礼乎可伎美等弥都都志努波牟

tare-wo ka kimi tǝ mi-tutu sinop-am-u who-ACC IP lord DV see(CONV)-COOR long.for-TENT-ATTR whom shall [I] long for, viewing [him] as [my] lord? (MYS 20.4440) 美麻久能富之伎吉美尓母安流加母

mi-m-aku-nǝ posi-ki kimi n-i mǝ ar-u kamǝ see-TENT-NML-GEN desire-ATTR lord DV-CONV FP exist-ATTR EP I wonder whether [it] is also my lord whom [I] want to see (MYS 20.4449) Note that in this example, the change from final to attributive form is triggered by the particle kamǝ that follows the affected verb ar- ‘to exist.’ 伊爾志加多知与乃都美佐閇保呂夫止曾伊布

in-i-si kata ti yǝ-nǝ tumi sapɛ porǝmb-u tǝ sǝ ip-u go-CONV-PAST/ATTR side thousand life-GEN sin RP disappear-FIN DV FP say-ATTR [they] say that even the sins of one thousand former lives will disappear (BS 17) As I mentioned above, other than a single attestation from the Man’yōshū cited below, the usage of the focus particle namo (spelt as 奈母 or 奈毛) is limited to the Senmyō text where it occurs ninety-three times. This is quite a spectacular number compared to the focus particle sǝ ~ ⁿzǝ (twenty-two times), the interrogative particles ya (twenty times) and ka (thirteen times), and the emphatic particle kamǝ (nineteen times). Examples: 何時奈毛不戀有登者雖不有得田直比來戀之繁母

ITU PA namo KOPÏ-ⁿZ-U AR-I tǝ pa AR-AN-E-ⁿDƏ utate KƏNƏ KƏRƏ KOPÏ si SI ŋGƐ-KI29 mǝ when TOP FP long.for-NEG-CONV exist-FIN DV TOP exist-NEG-EV-CONC unusually this time long.for(NML) EP thick-ATTR EP Although there is no [time] when [I] say that [I] do not long for [you] this time [my] longing is unusually strong (MYS 12.2877) 29  Most of the Man’yōshū commentators read this as SIⁿGƐ-SI ‘thick-FIN’ (Takagi et al. 1960: 265), etc., but this reading goes against the kakari-musubi rule that requires a final predicate to be in its attributive form after the particle namo. Certainly, there are no other cases of the attributive form in -ki attested after namo, but on the other hand there are no final forms in -si attested after namo either. Given that verbal final forms become attributives after namo, I believe that it is more consistent to interpret this form as siŋgɛ-ki rather than siŋgɛ-si.

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

560

Chapter 6

此食国天下之政事者平長将在止奈母所念坐

KƏNƏ WOS-U KUNI AMƐ-NƏ SITA-NƏ MATURiŋgƏTƏ pa TAPIRAKƐ-KU NA ŋGA-KU AR-AM-U tǝ namǝ OMƏP-OS-I-[I]MAS-U this rule-ATTR country heaven-GEN under-GEN governance TOP safe-CONV be. long-CONV exist-TENT-FIN DV FP think-HON-CONV-HON-ATTR [I] deign to think that the governance would be safe and last long in this country under the Heaven that [I] rule (SM 3) 治賜比慈賜來業止奈母随神所念行湏

WOSAMƐ-TAMAp-i UTUKUSImBÏ-TAMAP-I-K-URU WAⁿZA tǝ namǝ KAMUNA-ŋ-GARA OMƏP-OS-I-MEs-u rule(CONV)-HON-CONV show.benevolence (CONV)-HON-CONV-come-ATTR deed DV FP deity-PLUR-GEN-nature think-HON-CONV-HON-ATTR [I], as a deity, deign to think that [they are] the deeds that [I] deign to administer with benevolence (SM 3) 食国天下乎婆撫賜惠賜夫止奈母神奈我良母念坐湏

WOS-U KUNI AMƐ-NƏ SITA-womba NAⁿDE-TAMAP-I UTUKUSImBÏ-TAMAp-u tǝ namǝ KAMU-na-ŋ-gara mǝ OMƏP-OS-I-[I]MAs-u rule-ATTR country heaven-GEN under-ACC(EMPH) cherish(CONV)-HONCONV show.benevolence (CONV)-HON-FIN DV FP deity-PLUR-GEN-nature FP think-HON-CONV-HON-ATTR [I], as a deity, deign to think that [I] show benevolence and cherish the country under the Heaven that I rule (SM 13) 成奴礼波歓美貴美奈毛念食流

NAR-I-n-ure-mba KƏKƏRƏmBOSI-mi TAPUTO-mi namo OMƏP-I-TAMAP-Uru become-CONV-PERF-EV-CON be.glad-GER be.awesome-GER FP thinkCONV-HUM-ATTR because [it] became [as the deity said], [the sovereign] thought that [it] was joyful and awesome (SM 15) 伊豫国与利白祥鹿乎献奉天在礼方有礼志与呂許保志止奈毛見流

Iyǝ-NƏ KUNI-yǝri SIRO-KI SIRUSI N-Ə SIKA-wo TATEMATUR-I-te Ar-e-mba uresi yǝrǝkǝmb-ǝsi tǝ namo MI-ru Iyǝ-GEN province-ABL white-ATTR mark DV-ATTR deer-ACC present(HUM)CONV-SUB exist-EV-CON glad be.joyful-ADJ DV FP see-ATTR when [they] had presented [us] with a deer with white marks from the province of Iyǝ, [we] regarded [this] as [a] joyful and auspicious [event] (SM 46)

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

Verbs

561

臣等止共仁異奇久麗白伎形乎奈毛見喜流

OMI-TATI-tǝ TƏMƏ n-i KƏTƏ N-I AYASI-ku URUPASI-KI SIRO-ki KATATI-wo namo MI YƏRƏKƏmB-Uru noble-PLUR-COM together DV-CONV different DV-CONV be.strange-CONV be.beautiful-ATTR be.white-ATTR shape-ACC FP see(CONV) rejoice-ATTR [we] rejoice together with nobles looking at this unusual, strange, and beautiful white shape [of the deer skin] (SM 46) 3.2.1.2.4 Attributive as a Final Predicate without kakari-musubi There are cases when the attributive is used as a final predicate even when the particles sǝ ~ ⁿzǝ, ya, ka, or namo are not used in the sentence before the final verb.30 Saeki maintains that only very few examples of this usage are attested in the Old Japanese of the eighth century, and he provides eight of them for Western Old Japanese (Saeki 1959: 134).31 Below I present all examples that I was able to find in Western Old Japanese texts. As the reader will see, there are altogether more examples than are cited in the literature, but in general I have to agree with Saeki that this usage in Western Old Japanese is still rare. Saeki further notes that when the subject is present in a sentence that ends in an attributive without kakari-musubi, this subject is always marked by -nǝ or -ŋga. As several examples listed below demonstrate, this is not a necessary condition. Examples: 袁夜迩須賀多多美伊夜佐夜斯岐弖和賀布多理泥斯

wo-ya-ni suŋga-tatami iya-saya sik-i-te wa-ŋga puta-ri ne-si DIM-house-LOC sedge-mat rustling spread-CONV-SUB we-POSS two-CL sleep(CONV)-PAST/ATTR in a little hut two of us slept [together] spreading rustling sedge mats (KK 19) In this example we have the past attributive -si that is used as a form of final predication without a preceding particle sǝ ~ ⁿzǝ, ya, or ka.

30  The same is applicable to other attributive forms in Western Old Japanese, such as the past attributive -si and the adjectival attributive -ki. 31  Yamada Yoshio provides only four examples, one overlapping with Saeki’s list, and one being in fact an attributive functioning as a verbal noun and not as a form of final predication (Yamada 1954: 167).

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

562

Chapter 6

我二人宿之

WA-ŋGA PUTA-RI NE-si we-POSS two-CL sleep(CONV)-PAST/ATTR two of us slept [together] (MYS 2.109) In this example we have the past attributive -si that is used as a form of final predication without a preceding particle sǝ ~ ⁿzǝ, ya, or ka. 常丹跡君之所念有計類

TUNE n-i tǝ KIMI-ŋGA OMƏP-I-TAR-I-ker-u eternal DV-CONV DV lord-POSS think-CONV-PERF/PROG-CONV-RETR-ATTR [my] lord wished [to live] eternally (MYS 2.206) 玉緒乃不絶射妹跡結而石

TAMA-NƏ WO-nǝ TAYE-ⁿZI-i IMO-tǝ MUSUmB-I-TE-si pearl-GEN cord-GEN tear-NEG/TENT-ACT beloved-COM tie-CONV-PERF (CONV)-PAST/ATTR Insolubility of the pearly cord tied [me] with [my] beloved (MYS 3.481) In this example we have the past attributive form -si that is used as a form of final predication without a preceding particle sǝ ~ ⁿzǝ, ya, or ka. 意比久留母能波毛毛久佐爾勢米余利伎多流

op-i-k-uru mǝnǝ pa momo kusa n-i semɛ-yǝr-i-k-i-tar-u pursue-CONV-come-ATTR thing TOP hundred kind DV-CONV assault (CONV)-approach-CONV-come-CONV-PERF/PROG-ATTR the things that pursue [us], come assaulting [us] in a hundred varieties (MYS 5.804) 我衣手乃干時毛名寸

WA-ŋGA KƏRƏMƏⁿDE-nǝ POR-U TƏKI mo na-ki I-POSS sleeve-GEN dry-ATTR time FP no-ATTR there is no time at all for my sleeves to dry (MYS 10.1994) In this example we have the adjectival attributive form -ki that is used as a form of final predication without a preceding particle sǝ ~ ⁿzǝ, ya, or ka. 我衣袖之干時毛奈吉

WA-ŋGA KƏRƏMƏⁿDE-nǝ POR-U TƏKI mo na-ki I-POSS sleeve-GEN dry-ATTR time FP no-ATTR there is no time at all for my sleeves to dry (MYS 12.2954) In this example we have the adjectival attributive form -ki that is used as a form of final predication without a preceding particle sǝ ~ ⁿzǝ, ya, or ka.

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

563

Verbs 君乎衣尓有者下毛将著跡吾念有家留

KIMI-wo KƏRƏMƏ n-i AR-Amba SITA mo KI-M-U tǝ WA-ŋGA OMƏP-ER-I-ker-u lord-ACC garment DV-CONV exist-COND beneath FP wear-TENT-FIN DV I-POSS think-PROG-CONV-RETR-ATTR I came to think that if [you] were a garment, [I] would wear you beneath [my clothes] (MYS 12.2964) 安礼爾都気都流

are-ni tuŋgɛ-t-uru I-DAT report(CONV)-PERF-ATTR [thus he] reported to me (MYS 17.3957) 伊可尓安流布勢能宇良曽毛許己太久尓吉民我弥世武等和礼乎等登牟流

ika n-i ar-u puse-nǝ ura sǝ mo kǝkǝⁿdaku n-i kimi-ŋga mi-se-m-u tǝ ware-wo tǝⁿdǝm-uru how DV-CONV exist-ATTR Puse-GEN bay it FP so.much DV-CONV lord-POSS see-CAUS-TENT-FIN DV I-ACC stop-ATTR The bay of Puse, how incredibly [beautiful] it [is], [so that my] lord will stop [me] wishing to show [it to me] (MYS 18.4036) Note that in this example sǝ is not a focus particle, but a demonstrative pronoun. 比登母等能奈泥之故宇惠之

pitǝ-mǝtǝ n-ǝ naⁿdesiko uwe-si one-CL DV-ATTR carnation plant(CONV)-PAST/ATTR [I] planted one carnation (MYS 18.4070) In this example we have the past attributive form -si that is used as a form of final predication without a preceding particle sǝ ~ ⁿzǝ, ya, or ka. 保登等藝須伊登祢多家口波橘乃播奈治流等吉尓伎奈吉登余牟流

potǝtǝŋgisu itǝ neta-k-eku pa TATImBANA-nǝ pana-ⁿ-tir-u tǝki-ni k-i nak-itǝyǝm-uru cuckoo very distasteful-ATTR-NML TOP mandarin.orange-GEN flower-GENfall-ATTR time-LOC come-CONV sing-CONV-sound-ATTR Cuckoo, [you] are very distasteful! [You] come and sing at the time when mandarin orange flowers are falling (MYS 18.4092)

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

564

Chapter 6

夜度乃烏梅能知利須具流麻埿美之米受安利家流

yaⁿdo-nǝ uMƐ-nǝ tir-i-suŋg-uru-maⁿde mi-simɛ-ⁿz-u ar-i-ker-u house-GEN plum-GEN fall-CONV-pass-ATTR-TERM see-CAUS-NEG-CONV exist-CONV-RETR-ATTR [it] turned out that [you] did not let [me] see the plum [blossoms] in [your] house until they had completely fallen (MYS 20.4496) 宇梅乃波奈知利須具流麻弖伎美我伎麻左奴

uMƐ-nǝ pana tir-i-suŋg-uru-maⁿde kimi-ŋga k-i-[i]mas-an-u plum-GEN blossom fall-CONV-pass-ATTR-TERM lord-POSS come-CONV-HONNEG-ATTR [you] did not come, [my] lord, until the plum blossoms had completely fallen (MYS 20.4497) 加良須止伊布於保乎蘇止利能去止乎能米等母邇止伊比天佐岐陀智伊奴留

karasu tǝ ip-u opo woso tǝri-nǝ kǝtǝ-wo nǝmï32 tǝmǝ n-i tǝ ip-i-te saki-ⁿ-tat-i in-uru crow DV say-ATTR big hasty bird-GEN word-ACC RP together DV-CONV DV sayCONV-SUB ahead-LOC-depart-CONV go.away-ATTR Crows, big hasty birds, only cry together—[you] departed [from this world ahead [of me] (NR 2.2) Practically all examples cited above and especially the two examples from MYS 20.4496 and 4497, which represent a poetic exchange, seem to indicate that the usage of an attributive as a form of final predication without kakari-musubi functionally represents a strong confirmatory or explanatory statement, very similar to the modern Japanese verb + no da construction. Comparative Data Level A: Other Japonic A1: Eastern Old Japanese Besides the Western Old Japanese-looking attributives in -u and -uru, Eastern Old Japanese has a special attributive -o (sometimes misspelled as -ǝ) after consonant and r-irregular verbs as well as after suffixes ending in a consonant and r-irregular auxiliaries. This -o attributive represents an archaic feature: namely, in pre-Old Japanese, the attributive form had an original vowel

32  The original text has 能米 /nǝmɛ/, but this is likely to be a scribal mistake for 能未 /nǝmï/.

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

565

Verbs

/o/ that was raised to /u/ in Western Old Japanese.33 In one case there is also an attributive -a which is difficult to explain (see MYS 14.3526 below). The functions of the attributive in Eastern Old Japanese are the same as in Western Old Japanese. The attributive as a modifier: 奴麻布多都可欲波等里賀栖安我己許呂布多由久奈母等奈与母波里曽祢

numa puta-tu kayop-a tǝri-ŋga su a-ŋga kǝkǝrǝ puta yuk-unam-ǝ tǝ na-y-ǝmǝpar-i-sǝ-n-e marsh two-CL go.over-ATTR bird-POSS nest I-POSS heart two go-TENT2-ATTR DV NEG-?34-think-PROG-CONV-do-DES-IMP [I] wish [you] are not thinking that my heart would go [to] two [different places like] nests of birds that go over to marshes (MYS 14.3526) 久毛為尓美由流志麻奈良奈久尓

kumowi-ni mi-y-uru sima nar-an-aku n-i distance-LOC see-PASS-ATTR island be-NEG-NML DV-CONV although [it] is not an island that is seen in the distance (MYS 20.4355) 奈苦古良乎意伎弖

nak-u ko-ra-wo ok-i-te cry-ATTR child-PLUR-ACC leave-CONV-SUB leaving [my] crying children (MYS 20.4401) 佐弁奈弁奴美許登尓阿礼婆

sape-n-ape-n-u mi-kǝtǝ n-i ar-e-mba refuse(NML)-LOC-match-NEG-ATTR HON-word DV-CONV exist-EV-CON Because [it] is [my sovereign’s] order that [I] cannot refuse … (MYS 20.4432)

33  The problem is difficult, however, and Frellesvig argues against the archaic nature of the attributive -o in Eastern Old Japanese (Frellesvig 2008: 190, footnote 17). Russell, on the other hand, believes that the primary form is -ǝ as preserved in the attributive form n-ǝ of the defective verb n-, while the -ur- in -uru may be explained as a stative extension (Russell 2005: 641–644). Since it is not the goal of the present book to go into the details of the Proto-Japanese reconstruction, I will not elaborate further on this problem here. 34   This -y- most likely reflects Ainu 3rd person singular object marker i- (Vovin 2012: 2014).

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

566

Chapter 6

The attributive as a nominalized form: 奴流我倍爾安杼世呂登可母

n-uru-ŋga [u]pɛ-ni aⁿ-dǝ se-rǝ tǝ kamǝ sleep-ATTR-POSS top-LOC what-DV do-IMP DV EP besides sleeping [with her], what [else do I] do? (MYS 14.3465) 之良夜麻可是能宿奈敝杼母古呂賀於曽伎能安路許曽要志母

sira yama kaⁿze-nǝ NE-n-ap-e-ⁿdǝmǝ ko-rǝ-ŋga osǝki-nǝ ar-o kǝsǝ ye-si-mǝ white mountain wind-GEN sleep-NEG-ITER-CONC girl-DIM-POSS garmentGEN exist-ATTR FP good-FIN-EXCL although [I] continue not to sleep at the [cold] wind from the White Mountain, [it] is good to have my girl’s garment! (MYS 14.3509) 比登祢呂爾伊波流毛能可良

pitǝ ne-rǝ n-i ip-ar-u monǝkara one peak-DIM DV-CONV say-PROG-ATTR CONJ Although [they] say that [we] are one peak … (MYS 14.3512) 多可伎祢爾久毛能都久能須和礼左倍爾伎美爾都吉奈那

taka-ki ne-ni kumo-nǝ tuk-u-nǝsu ware sapɛ n-i kimi-ni tuk-i-n-ana high-ATTR peak-LOC cloud-GEN attach-ATTR-COMP I RP DV-CONV lord-DAT attach-CONV-PERF-DES Even I would like to cling to [my] lord like clouds cling to a high peak (MYS 14.3514) 波波乎波奈例弖由久我加奈之佐

papa-wo panare-te yuk-u-ŋga kanasi-sa mother-ACC separate(CONV)-SUB go-ATTR-POSS sad-NML sadness of going, leaving [my] mother behind (MYS 20.4338) The attributive as a final predicate appears after the interrogative particles ya and ka, and the focus particle sǝ ~ⁿzǝ ~ ⁿze when these particles are found before the final verb in a sentence. Similar to Western Old Japanese, the particle kamo ~ kamu can trigger the change from a final form to an attributive irrespective of its position before or after a final verb.

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

567

Verbs 麻許登可聞和礼爾余須等布

ma-kǝtǝ kamo ware-ni yǝs-u tǝ [i]p-u INT-thing EP I-DAT bring.close-FIN DV say-ATTR I wonder [whether it is] true that [people] say that [she] has an intimate relationship with me (MYS 14.3384) 阿杼可多延世武

aⁿ-dǝ ka taye se-m-u why-DV IP break(NML) do-TENT-ATTR why should [we] break up? (MYS 14.3397) 安加奴乎安杼加安我世牟

ak-an-u-wo aⁿ-dǝ ka a-ŋga se-m-u satisfy-NEG-ATTR-ACC what-DV IP I-POSS do-TENT-ATTR since it was not enough [for me], what should I do? (MYS 14.3404) 哭乎曽奈伎都流手兒尓安良奈久尓

NE-wo sǝ nak-i-t-uru teŋGO n-i ar-an-aku n-i voice-ACC FP cry-CONV-PERF-ATTR baby DV-CONV exist-NEG-NML DV-CONV [I] sobbed loudly, although [I] am not a baby (MYS 14.3485) 伊可奈流勢奈可和我理許武等伊布

ika nar-u se-na ka wa-ŋgari kǝ-m-u tǝ ip-u how be-ATTR beloved-DIM IP I-DIR come-TENT-FIN DV say-ATTR what kind of beloved is [he], who says that [he] will come to me? (MYS 14.3536) 多麻母乃宇知奈婢伎比登里夜宿良牟

tama mǝ-nǝ uti-nambik-i pitǝ-ri ya ne-ram-u jewel seaweed-COMP PREF-stretch-CONV one-CL IP sleep-TENT2-ATTR will [you] sleep alone, stretched like a jewel seaweed? (MYS 14.3562) 知々波々我可之良加伎奈弖佐久安礼天伊比之氣等婆是和須礼加祢豆流

titi papa-ŋga kasira kaki-naⁿde sa-ku ar-e te ip-i-si kɛtǝmba ⁿze wasure-kane-t-uru father mother-POSS head PREF-stroke(CONV) safe-CONV exist-IMP DV sayCONV-PAST/ATTR word FP forget(CONV)-NEG/POT(CONV)-PERF-ATTR [I] cannot forget the words: “Be safe!” that [my] father and mother said, stroking [my] head (MYS 20.4346)

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

568

Chapter 6

意保枳美能美己等可之古美阿乎久牟乃等能妣久夜麻乎古与弖伎怒加牟

opo kimi-nǝ mi-kǝtǝ kasiko-mi awo kumu-nǝ tǝnǝmbik-u yama-wo koyǝ-te k-i-n-o kamu great lord-GEN HON-word awesome-GER blue cloud-GEN trail-ATTR mountainACC cross(CONV)-SUB come-CONV-PERF-ATTR EP Since the sovereign’s (lit.: great lord’s) order is awesome, [I] came [here] crossing mountains where dark clouds trail! (MYS 20.4403) The attributive as a final predicate without kakari-musubi: 和乎布利弥由母阿是古志麻波母

wa-wo pur-i-mi-y-umǝ Aⁿze ko si map-am-ǝ I-ACC swing-CONV-look-PASS-EXCL Aⁿze girl EP dance-TENT-ATTR the girl from Aⁿze is going to dance, suddenly looking back at me! (FK 7) 兒良波安波奈毛比等理能未思弖

KO-ra pa ap-ana-m-o pitǝ-ri nǝmï s-i-te girl-DAT TOP meet-DES-TENT-ATTR one-CL RP do-CONV-SUB being absolutely alone, [I] wish to meet [this] girl (MYS 14.3405) 安比豆祢能久爾乎佐杼抱美安波奈波婆斯努比尓勢毛等比毛牟須婆佐祢

Apiⁿdune-nǝ kuni-wo saⁿ-tǝpo-mi ap-an-ap-amba sinop-i n-i se-m-o tǝ pimo musumb-as-an-e Apiⁿdune-GEN land-ABS PREF-far-GER meet-NEG-ITER-COND long.for-NML DV-CONV do-TENT-ATTR DV cord tie-HON-DES-IMP If [we] continue not to meet, because the land of Apiⁿdune is far, [I] wish [you would] tie [your garment] cords, as if longing for [me] (MYS 14.3426) 和奴爾故布奈毛

wanu-ni kop-unam-o I-DAT long.for-TENT2-ATTR [you] will probably long for me (MYS 14.3476) 和可加敞流弖能毛美都麻弖宿毛等和波毛布

waka kaperute-nǝ momit-u-maⁿde NE-m-o tǝ wa pa [o]mop-u young maple-GEN leaves.turn.red/yellow-ATTR-TERM sleep-TENT-ATTR DV I TOP think-FIN I think that [we] should sleep [together] until the young maple becomes red (MYS 14.3494)

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

569

Verbs 夜麻敞呂能思之奈須於母敞流

yama pe-rǝ-nǝ sisi-nasu omǝp-er-u mountain side-DIM-GEN deer-COMP think-PROG-ATTR [they] are thinking [of me] like [of] a deer on the [little] mountain side (MYS 14.3531) 須流河乃祢良波苦不志久米阿流可

Suruŋga-nǝ ne-ra pa kupusi-ku mɛ ar-u ka Suruŋga-GEN summit-PLUR TOP lovely-CONV FP exist-ATTR IP are [not] the summits of Suruŋga [mountains] lovely? (MYS 20.4345) A2: Ryukyuan Most modern Ryukyuan dialects have an allomorph -uru that follows both consonant and vowel verbs alike and an allomorph -ru that is used after the copula ya- ‘to be’ and also may be found as an alternative form after vowel verbs. Sometimes -ru is used as an alternative form after consonant verbs with a root ending in -r. In Southern Ryukyuan a form -ï is also found (< *-ïrï < *-uru). Let us look at the formation of the attributive forms in Ryukyuan taking Shuri as an example: chart 35 Attributive forms in Shuri

gloss

verbal root

attributive form

write go stand hold die fly read take laugh boil forget be

kak’iktatmutshintubyumturwararniwashiya-

kach-uru ’ich-uru tach-uru much-uru shin-uru tub-uru yum-uru tu-y-uru, tu-i-ru wara-y-uru, wara-i-ru ni-y-uru, ni-i-ru washi-y-uru ya-ru

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

570

Chapter 6

We can see on the basis of the palatalizations k > ch and t > ch found in the attributive forms of consonant verbs ending in the consonants -k or -t, as well as on the basis of such forms as tu-y-uru, tu-i-ru, wara-y-uru, wara-i-ru, ni-y-uru, ni-i-ru, and washi-y-uru, that historically the attributive suffix -uru is likely to be an auxiliary that followed the converb form -i. Thus, the picture we discover here is quite similar to the final predication form *-um. Similar to *-um, which is believed to be a final predication form of wor-, -uru is also believed historically to be an attributive form wor-u of the verb wor- ‘to exist.’ This traditional point of view that dates back to Hattori Shirō seems to be quite widespread among Japanese linguists today as well (Nishioka and Nakahara 2000: 31). Although the phonetic equation is better than in the case of the final predication form *-um, we still face some of the same problems when we try to trace back the attributive form not only in Ryukyuan, but in Japonic in general to the attributive wor-u of the verb wor- ‘to exist.’ First, medial -w- is not lost in Old Japanese in intervocalic position. Second, we are still faced with the problem of circularity: under this explanation it appears that only wor- could have an independent attributive form -u < *-o, while all other verbs could not add the attributive directly to their stems. Possibly, Russell’s analysis of the attributive -uru as bimorphemic, consisting of *-ur-, a stative extension plus the attributive -ǝ (preserved as such in the attributive form n-ǝ of the defective verb n-), with *ur-ǝ undergoing assimilation to *-ur-o with a consequent raising to -ur-u (Russell 2005: 641–644) is on a right track. The problem of explaining the basic function of the stative *-ur-, and why it is added to vowel verbs, but not to consonant verbs, certainly remains to be solved. One additional serious problem that prevents us from considering the Shuri attributive -uru, which appears after all classes of verbs, consonant and vowel alike, to be an archaism is that in Old Ryukyuan the situation with consonant verbs is quite similar to the one found in Eastern Old Japanese: in most cases we find the attributive form -u and in some more rare cases the attributive form -o. If the spellings with -o found in the Omoro sōshi do indeed reflect a pre-raising situation as have been suggested before (Serafim, p.c.), the parallelism with Eastern Old Japanese is striking and cannot be easily ignored. Most vowel verbs in Old Ryukyuan with the exception of the irregular verbs ko- ‘to come’ and se- ‘to do’ have already shifted to a monograde conjugation, and therefore only the allomorph -ru is found directly after the final vowel of a verb root. Consider the following chart that is based on the data presented in (Torigoe 1968: 218–249):

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

571

Verbs chart 36 Attributive forms in Old Ryukyuan

gloss

verbal root

attributive form

hear bloom stand hold attack scoop sound pray rise take protect exist see descend grow old attach raise caress do come

kiksaktatmotoso(w)kumtoyominoragartormaburarmiori-/oreowetukeagenadeseko-

kik-u, kik-o sak-u tat-u mot-u oso-u, oso-o kum-u, kum-o toyom-u, toyom-o inor-u, inor-o agar-u, agar-o tor-u mabur-u, mabur-o ar-u mi-ru ore-ru owe-ru tuke-ru age-ru nade-ru s-uru, shi-y-uru k-uru, k-uro

Selected examples from Old Ryukyuan and Shuri are presented below: Old Ryukyuan いによはのおきてもちなる

Iniyofa-no okite-moti nar-u Iniyofa-GEN rule-holder be-ATTR [the one] who is a governor of Iniyofa (OS 8.456) せたかこかみまふろすへまさるわうにせ

se taka ko-ga mi-mabur-o sufe masar-u wau-nise spirit high girl-POSS see(CONV)-protect-ATTR destination excel-ATTR kingHON the honorable king who [goes to] an exhausted destination is protected by the high priestess (OS 12.741) Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

572

Chapter 6

おしぢへたるゑつかさくどゑ

os-i-dife-tar-u we tukasa-gu do we push-CONV-exit(CONV)-PERF/PROG-ATTR hey tukasa-DIM FP hey hey, [we] pushed [the boat] ahead; hey, [priestess] Tukasa (OS 13.747) ふてつたむななのおとぢや

fute-tsu ta mu nana n-o otodiya one-CL two six seven DV-ATTR brother one, two, six, seven brothers (OS 13.898) きみかいのろもりに

kimi-ga inor-o mori-ni lady-POSS pray-ATTR shrine-LOC in the shrine where the lady [priestess] prays (OS 22.1526) Shuri

’ari-ga ch-uuru-madi maQco-ok-ee he-NOM come-ATTR-TERM wait(CONV)-put-HORT Let us wait until he comes (RKJ 1983: 357) taruu-ga hwich-uru sanzhin-oo chichigutu du ya-ru Tarō-POSS play-ATTR shamisen-TOP pleasant.to.hear FP be-ATTR The shamisen that Tarō plays is pleasant to hear (Nishioka & Nakahara 2000: 26) In this example the last attributive form ya-ru of the copula ya- is triggered by the kakari-musubi rule due to the presence of the focus particle du in the sentence. 3.2.1.3 Evidential -ure ~ -e ~ -re The evidential suffix’s main allomorph is -ure, which is found after all vowel verbs (except strong vowel verbs, where it becomes just -re), and all irregular verbs (except r-irregular verbs). After all consonant verbs and r-irregular verbs the allomorph -ɛ ~ -e- is used, according to the morphonological rule of -r- loss mentioned in chapter 2, section 2.5.4.

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

573

Verbs chart 37 Distribution of the allomorphs of the evidential suffix

verb class

allomorph

consonant verbs regular vowel verbs k-irregular verbs s-irregular verbs n-irregular verbs strong vowel verbs r-irregular verbs

-ɛ ~ -e -ure -ure -ure -ure -re -e

A similar picture is observed if -ure is used not after a verbal stem but after another verbal suffix or an auxiliary: if a suffix or an auxiliary ends with a vowel, the main allomorph -ure is used, and the final vowel of the preceding suffix or an auxiliary is apocopated. If a suffix or an auxiliary ends in a consonant, the allomorph -ɛ ~ -e is used.35 The only exception is the perfective auxiliary -nthat has the evidential form -n-ure. chart 38 Combinations of the evidential suffixes -ure and -ɛ ~ -e with preceding suffixes and bound auxiliaries

suffixes and auxiliaries

-ure

-ɛ ~ -e

tentative -(a)mtentative2 -(u)ramnegative -(a)npassive -(a)ye-, -rayecausative -(a)simɛiterative -aphonorific -asperfective -teperfective -nretrospective -kerprogressive -erperfective-progressive -tar-

– – – -(a)y-ure -(a)sim-ure – – -t-ure -n-ure – – –

-(a)m-ɛ -(u)ram-ɛ -(a)n-e – – -ap-ɛ -as-e – – -ker-e -er-e -tar-e

35  The allomorph -re is not used after suffixes and auxiliaries.

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

574

Chapter 6

The evidential is used by itself as a final form. In most cases it appears when the particle kǝsǝ precedes it in the sentence, replacing the final predication suffix.36 This usage is quite similar to the replacement of the final predication form by the attributive after the particles ya, sǝ ~ ⁿzǝ, ka, kamǝ and namo, described above in section 3.2.1.2.3. Similar to this rule, the rule of replacement of the final form by the evidential is also known in the traditional grammar as 係り結び kakari-musubi ‘the rule of linking.’ Nevertheless, in contrast to Middle Japanese, where koso always triggers the change of the final predication form to the evidential, there are many cases in Western Old Japanese when the evidential can be used as a final predication form without the preceding kǝsǝ. Although it is difficult to say exactly what the evidential form means, I believe that used in isolation, its function is close to that of the exclamation point in English; in any case, it obviously indicates some kind of emphatic statement. The fact that it is used together with kǝsǝ, a strong emphatic particle, further supports this proposal. I call it evidential because frequently it emphasizes fact(s) that should be evident to the speaker or his/her addressee.37 Otherwise, the evidential is usually followed by the conjunctive converb -mba or the concessive converb -ⁿdǝ/-ⁿdǝmǝ, usages that I survey below in the sections dedicated to them. The evidential as a final predicate with kakari-musubi: 伊麻許曾婆和杼理迩阿良米

ima kǝsǝ pa wa-ⁿ-dǝri n-i ar-am-ɛ now FP TOP I-OSM-bird DV-CONV exist-TENT-EV Now [I] am my bird (KK 3) 佐和佐和爾那賀伊弊勢許曾宇知和多須夜賀波延那須岐伊理麻韋久禮

sawa-sawa n-i na-ŋga ip-es-e kǝsǝ uti-watas-u ya-ŋga-paye-nasu k-i-ir-i-mawik-ure noisily-noisily DV-CONV you-POSS speak(CONV)-HON-EV FP PREF-carry.acrossATTR ?-POSS-?-COMP come-CONV-enter-CONV-HUM(CONV)-come-EV you spoke noisily and [you] came inside like ? that crossed from afar (KK 63)

36  In contrast to Middle Japanese, if the predicate after kǝsǝ is an inflected adjective, in Western Old Japanese the adjectival final predication form -si is replaced by the attributive -ki, and not by the evidential -kere. See chapter 5, section 2.3 for details and examples. Needless to say, auxiliaries that have an adjectival paradigm follow this rule as well in Western Old Japanese. 37  This terminological usage is different from modern language typology, where ‘evidential’ is used in a sense of experiential.

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

575

Verbs 宇倍志許曾斗比多麻閇麻許曾迩斗比多麻閇

umbɛ-si kǝsǝ top-i-tamap-ɛ ma kǝsǝ n-i top-i-tamap-ɛ be.proper-FIN FP ask-CONV-HON-EV truth FP DV-CONV ask-CONV-HON-EV [It] is proper that [you] asked [me], [it] is right that [you] asked [me] (KK 72) 可久斯己曽烏梅乎加射之弖多努志久能麻米

ka-ku si kǝsǝ uMƐ-wo kaⁿzas-i-te tanosi-ku nǝm-am-ɛ thus-CONV EP FP plum-ACC decorate-CONV-SUB be.merry-CONV drinkTENT-EV decorating [our hair] with plum [blossoms] in this way, [we] should drink merrily (MYS 5.833) 伊毛我多毛等乎和礼許曽末加米

imo-ŋga tamotǝ ware kǝsǝ mak-am-ɛ beloved-POSS sleeve I FP use.as.a.pillow-TENT-EV I will use as a pillow the sleeves of [my] beloved (MYS 5.857) 伊敝之麻波奈尓許曽安里家礼

Ipe sima pa na n-i kǝsǝ ar-i-ker-e Home island TOP name DV-CONV FP exist-CONV-RETR-EV “Home island” turned out to be just a name (MYS 15.3718) Note that although in this example the focus particle kǝsǝ is inside the non-contracted form of the copula, it still triggers the change of the final form to the evidential. 安我未許曾世伎夜麻故要弖許己爾安良米許己呂波伊毛爾与里爾之母能乎

a-ŋga mï kǝsǝ seki yama koye-te kǝkǝ-ni ar-am-ɛ kǝkǝrǝ pa imo-ni yǝr-i-n-i-si mǝnǝwo I-POSS body FP barrier mountain cross(CONV)-SUB here-LOC exist-TENT-EV heart TOP beloved-DAT approach-CONV-PERF-CONV-PAST/ATTR CONJ My body has crossed barriers and mountains, and is probably here. But [my] heart stayed near [my] beloved! (MYS 15.3757) 昨日許曽敷奈弖婆勢之可

KINƏPU kǝsǝ puna-[i]ⁿde se-sika yesterday FP boat-exit(NML) do(CONV)-PAST/EV Yesterday [we] sailed out (MYS 17.3893)

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

576

Chapter 6

故之能吉美良等可久之許曾楊奈疑可豆良枳多努之久安蘇婆米

kosi-nǝ kimi-ra-tǝ ka-ku si kǝsǝ YAnaŋgï kaⁿdurak-i tanosi-ku asomb-am-ɛ Kosi-GEN lord-PLUR-COM thus-CONV EP FP willow wear.as.a.laurel-CONV pleasant-CONV amuse-TENT-EV [I] will amuse [myself] with lords from Kosi by putting willow [branches] as laurels in our hair in this way (MYS 18.4071) 可久之許曽都可倍麻都良米伊夜等保奈我尓

ka-ku si kǝsǝ tukapɛ-matur-am-ɛ iya tǝpo naŋga n-i thus-CONV EP FP serve(CONV)-HUM-TENT-EV plentifully long long DV-CONV Thus, [I] will serve [you] plentifully and for a long, long time (MYS 18.4098) The evidential as a final predicate without kakari-musubi: 佐用婆比爾阿理多多斯用婆比迩阿理加用婆勢

sa-yomb-ap-i-ni ari-tat-as-i yomb-ap-i-ni ari-kayop-as-e PREF-call-ITER-NML-LOC ITER-set.out-HON-CONV call-ITER-NML-LOC ITER-set.out-HON-EV [I] set out (repeatedly) to woo [her] there, [I] set out (repeatedly) to woo [her] (KK 2) 佐和佐和爾那賀伊弊勢許曾

sawa-sawa n-i na-ŋga ip-es-e kǝsǝ noisily-noisily DV-CONV you-POSS speak(CONV)-HON-EV FP you spoke noisily (KK 63) 嚢伽多佐例

ta ka ta-sar-e who IP PREF-go.away-EV Who goes away? (NK 40) 大雪乃乱而來礼

OPO YUKI-nǝ MIⁿDARE-TE K-I-TAr-e big snow-GEN be.chaotic(CONV)-SUB come-CONV-PERF/PROG-EV big snow was falling down chaotically (MYS 2.199)

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

577

Verbs 許許呂由母於母波奴阿比陀爾宇知那毘枳許夜斯努礼

kǝkǝrǝ-yu mǝ omǝp-an-u apiⁿda-ni uti-nambik-i kǝy-as-i-n-ure heart-ABL FP think-NEG-ATTR interval-LOC PREF-stretched out-CONV lie. down-HON-CONV-PERF-EV while even in [my] heart [I] did not think, stretched out [she] was lying (MYS 5.794) 伊布許等夜美霊剋伊乃知多延奴礼

ip-u kǝtǝ yam-i TAMA-KIPARU inǝti taye-n-ure say-ATTR thing stop-CONV jewel-? life cease-PERF-EV [he] stopped to speak, and [his] jewel-like life ended (MYS 5.904) 比登比母伊毛乎和須礼弖於毛倍也

pitǝ pi mǝ imo-wo wasure-te omop-ɛ ya one day FP beloved-ACC forget(CONV)-SUB think-EV IP would [I] imagine forgetting [my] beloved even [for] one day? (MYS 15.3604) 伊毛尓安礼也夜須伊毛祢受弖安我故非和多流

imo n-i ar-e ya yasu i mo ne-ⁿz-u-te a-ŋga kopï-watar-u beloved DV-CONV exist-EV IP easy sleep FP sleep-NEG-CONV-SUB I-POSS long.for(CONV)-cross-ATTR Is [she] my beloved? I do not sleep easily, and continue to long for [her] (MYS 15.3633) 安杼毛倍香許己呂我奈之久伊米爾美要都流

aⁿ-dǝ [o]mop-ɛ ka kǝkǝrǝ-ŋ-ganasi-ku imɛ-ni mi-ye-t-uru what-DV think-EV IP heart-GEN-sad-CONV dream-LOC see-PASS(CONV)PERF-ATTR what was [I] thinking [about]? As [I] was sad in [my] heart, [she] suddenly has appeared in [my] dream (MYS 15.3639) 安比於毛波奴君尓安礼也母

api-omop-an-u KIMI n-i ar-e ya mǝ REC-think-NEG-ATTR lord DV-CONV exist-EV IP EP is [it my] lord, who no [longer] thinks [about lamenting of the people of this world] in return?! (MYS 15.3691)

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

578

Chapter 6

和伎毛故我可多美能許呂母奈可里世婆奈爾毛能母弖加伊能知都我麻之

wa-ŋg-imo-ko-ŋga katami n-ǝ kǝrǝmǝ na-k-ar-i-s-emba nani monǝ mǝt-e ka inǝti tuŋg-amasi I-POSS-beloved-DIM-POSS keepsake DV-ATTR garment no-CONV-exist-CONVPAST/ATTR-COND what thing hold-EV IP life continue-SUBJ If [I] did not have a garment of my beloved as a keepsake, would having anything keep [me] alive? (MYS 15.3733) Omodaka et al. consider mǝt-e in the example above to be a contraction of mǝt-i-te (1967: 741), however, this kind of contraction is not attested phonographically until Middle Japanese. In addition, the above example is the only example in the whole Western Old Japanese corpus where this alleged contraction appears in phonographic script. Cf. also three similar examples of evidential followed by the interrogative particle ka in MYS 15.3639 above and in Eastern Old Japanese cited below (MYS 14.3461, 14.3572). 何爲牟尓吾乎召良米夜

NANI SE-m-u-ni WA-wo MES-Uram-ɛ ya what do-TENT-ATTR-LOC I-ACC summon-TENT2-EV IP should [you] have summoned me in order to do something? [You should not!] (MYS 16.3886) 阿里佐利氐能知毛相牟等於母倍許曽

ari-sar-i-te nǝti mo AP-Am-u tǝ omǝp-ɛ kǝsǝ ITER-go.away-CONV-SUB after FP meet-TENT-FIN DV think-EV FP [Time] constantly goes away, and [I] hope that [we] will meet later, too (MYS 17.3933) 美阿止都久留伊志乃比鼻伎波阿米爾伊多利都知佐閇由須礼知知波波賀多 米爾毛呂比止乃多米爾

mi-atǝ tukur-u isi-nǝ pimbik-i pa amɛ-ni itar-i tuti sapɛ yusur-e titi papa-ŋga tamɛ n-i morǝ pitǝ-nǝ tamɛ n-i HON-footprint make-ATTR stone-GEN echo-NML TOP heaven-LOC reachCONV earth RP shake-EV father mother-POSS for DV-CONV all person for DV-CONV The echo of the stone, where [I] carved the footprints [of the Buddha], reaches Heaven, and shakes the earth as well, for father and mother, for all people (BS 1)

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

579

Verbs

3.2.1.3.1 Special Constructions -(a)m-ɛ ya and -(u)ram-ɛ ya There is a special usage of the evidential in combination with the preceding tentatives -am- ~ -m- or -uram- and the following interrogative particle ya that represents irony, or in other words it poses a question that implies an opposite answer (the negative answer to the question in the affirmative, and the positive answer to the question in the negative). Examples: 於夜那斯爾奈礼奈理鶏迷夜

oya na-si n-i nare nar-i-k-em-ɛ ya parent no-FIN DV-CONV you be.born-CONV-PAST/FIN-TENT-EV IP Were you possibly born without parents? [Certainly not!] (NK 104) 人嬬故尓吾戀目八方

PITƏ-ⁿ-DUMA YUWE n-i ARE KOPÏ-m-ɛ ya mo person-GEN-spouse reason DV-CONV I long.for-TENT-EV IP EP because [she] is the wife of [another] person, should I long for [her]? [Certainly not!] (MYS 1.21) 痛背乃河乎渡金目八

Anase-nǝ KAPA-wo WATAR-I-kane-m-ɛ ya Anase-GEN river-ACC cross-CONV-NEG/POT-TENT-EV IP would [I] be unable to cross the Anase river? [Certainly I would be able!] (MYS 4.643) 麻佐礼留多可良古爾斯迦米夜母

masar-er-u takara ko-ni sik-am-ɛ ya mǝ excel-PROG-ATTR treasure child-LOC reach-ATTR-EV IP EP Could the excellent treasures be equal to children?! [Certainly not!] (MYS 5.803) 和何世古我多那礼乃美巨騰都地爾意加米移母

wa-ŋga se-ko-ŋga ta-nare n-ǝ mi-kǝtǝ tuti-ni ok-am-ɛ ya mǝ I-POSS beloved-DIM-POSS hand-accustom(NML) DV-ATTR HON-koto groundLOC put-TENT-EV IP EP would anyone [dare to] put the favorite koto of my beloved on the ground?! [Certainly not!] (MYS 5.812) 佐夫志計米夜母吉美伊麻佐受斯弖

sambusi-k-em-ɛ ya mǝ kimi imas-aⁿz-u s-i-te be.sad-ATTR-TENT-EV IP FP lord come(HON)-NEG-NML do-CONV-SUB [if my] lord does not come, would [I] be sad?! [Certainly not!] (MYS 5.878) Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

580

Chapter 6

山櫻花日並而如是開有者甚戀目夜裳

YAMA SAKURA-m-BANA PI NARAmBƐ-TE KA-KU si SAK-ER-Amba PAⁿDA KOPÏ-m-ɛ ya mo mountain cherry-GEN-flower day line.up(CONV)-SUB thus-CONV EP bloomPROG-COND considerably long.for-TENT-EV IP EP If mountain cherry flowers were blooming thus every day, would [I] long for [them] considerably? [Certainly not!] (MYS 8.1425) 比故保思母和礼爾麻佐里弖於毛布良米也母

Pikoposi mǝ ware-ni masar-i-te omop-uram-ɛ ya mǝ Altair FP I-DAT surpass-CONV-SUB long.for-TENT2-EV IP EP Will Altair long for [his beloved] more than I [do]?! (lit.: surpassing me) [Certainly not!] (MYS 15.3657) 能知尓毛安波射良米也母

nǝti-ni mo ap-aⁿz-ar-am-ɛ ya mǝ later-LOC FP meet-NEG(CONV)-exist-TENT-EV IP EP will [we] not meet later?! [Of course we will!] (MYS 15.3741) 将若異子等丹所詈金目八

WAKA-k-eM-U KO-RA-ni NOR-AYE-kane-m-ɛ ya young-ATTR-TENT-ATTR girl-PLUR-LOC abuse-PASS(CONV)-NEG/POT-TENTEV IP would [you] be able not to be abused by girls who would be younger? [Certainly you would be abused!] (MYS 16.3793) 何爲牟尓吾乎召良米夜

NANI SE-m-u-ni WA-wo MES-Uram-ɛ ya what do-TENT-ATTR-LOC I-ACC summon-TENT2-EV IP should [you] have summoned me in order to do something? [You should not!] (MYS 16.3886) 保登等藝須伊麻奈可受之弖安須古要牟夜麻尓奈久等母之流思安良米夜母

potǝtǝŋgisu ima nak-aⁿz-u s-i-te asu koye-m-u yama-ni nak-u tǝmǝ sirusi ar-am-ɛ ya mǝ cuckoo now cry-NEG-NML do-CONV-SUB tomorrow cross-TENT-ATTR mountain-LOC cry-FIN CONJ sign exist-TENT-EV IP EP Cuckoo! [You] are not crying now: even if [you] cry in the mountains that [I] will cross tomorrow, would [there] be any sign?! [Certainly not!] (MYS 18.4052)

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

581

Verbs 都婆吉都良々々尓美等母安可米也

tumbaki tura-tura n-i mi tǝmǝ ak-am-ɛ ya camellia intently DV-CONV look(FIN) CONJ get.enough-TENT-EV IP even if [I] look intently at the camellia, would [I] get enough? [Certainly not!] (MYS 20.4481) Comparative Data Level A: Other Japonic A1: Eastern Old Japanese The evidential -ure ~ -e ~ -ɛ has the same functions in Eastern Old Japanese as in Western Old Japanese. The evidential as a final predicate with kakari-musubi: 刀奈布倍美許曽奈爾与曾利鶏米

tonap-umbɛ-mi kǝsǝ na-ni yǝsǝr-i-k-em-ɛ recite-DEB-GER FP you-DAT approach-CONV-PAST/FIN-TENT-EV as [I] had to recite [the magic formula], [I] approached you (MYS 14.3468) 奈爾己曾与佐礼

na-ni kǝsǝ yǝs-ar-e you-DAT FP approach-PROG-EV [I] approached you (MYS 14.3478) 於曾波夜母奈乎許曾麻多賣

osǝ paya mǝ na-wo kǝsǝ mat-am-e slow fast FP you-ACC FP wait-TENT-EV Whether [you come] quickly or slowly, [I] will wait for you (MYS 14.3493) The evidential -ure ~ -ɛ ~ -e as a final predicate without kakari-musubi: 安是登伊敞可

aⁿze tǝ ip-e ka why DV say-EV IP for what reason (lit.: saying: ‘Why?’) (MYS 14.3461) 安杼毛敞可

aⁿ-dǝ [o]mop-e ka what-DV think-EV IP Thinking what? (MYS 14.3572)

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

582

Chapter 6

波奈都豆麻奈礼也

pana t-u tuma nar-e ya flower DV-ATTR spouse be-EV IP Are you a flower-wife? (MYS 14.3370) 奈爾須礼曾波波登布波奈乃佐吉泥己受祁牟

nani s-ure sǝ papa tǝ [i]p-u pana-nǝ sak-i-[i]ⁿde-kǝ-ⁿz-u-k-em-u what do-EV FP mother DV say-ATTR flower-GEN bloom-CONV-exit(CONV)come-NEG-CONV-PAST/FIN-TENT-ATTR why (lit.: having done what) has the flower called ‘Mother’ not bloomed? (MYS 20.4323) Irony: 安左乎良乎遠家尓布須左尓宇麻受登毛安須伎西佐米也

asa-wo-ra-wo wo-ke-ni pususa n-i um-as-u tǝmo asu ki-se-sas-am-ɛ ya hemb-hemb.thread-PLUR-ACC hemp-container-LOC many DV-CONV spinHON-FIN CONJ tomorrow wear(NML)-do-HON-TENT-EV IP Even if [you] spin many hemb threads into a hemb container, would [you] wear [them] tomorrow [as a garment]? [Certainly, you would not!] (MYS 14.3484) A2: Ryukyuan It appears that in Old Ryukyuan the evidential form consistently replaces the final form after the focus particle siyo ~ su that is claimed to be related to WOJ kǝsǝ (Serafim and Shinzato 2005: 12–21). All examples below are borrowed from (Serafim and Shinzato 2005) with some changes in transliteration, glossing, and translation: しよりもりきみゝゝしよまふらめ

Siyori mori kimi-kimi siyo mabur-am-e Shuri shrine priestess-priestess FP protect-TENT-EV Priestesses from the Shuri Shrine will protect [it] (OS 13.853) はつにしやすまちよたれおきとばすまちよたれ

fatu nisi ya su mat-i-yo-tar-e okitoba su mat-i-yo-tar-e first north.wind TOP FP wait-CONV-exist(CONV)-PERF-EV north.wind FP wait-CONV-exist(CONV)-PERF-EV

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

Verbs

583

[We] waited for the first north wind. [We] waited for the north wind (OS 13.899) Similar to Western Old Japanese, there are also examples in Old Ryukyuan when the evidential form may be used as a form of final predication without the preceding focus particle siyo ~ su: こばおもりのきみゝゝまやゑておこらめ

Kobao mori-no kimi-kimi ma-y-aw-e-te okor-am-e Kobao shrine-GEN priestess-priestess dance-CONV-join-CONV-SUB sendTENT-EV Priestesses from the Kobao shrine will dance together and send [it] (OS 13.853) あやみやの大ころあまこあわちへもとらめ

aya miya-no ofo koro ama-ko awa-tife modor-am-e splendid palace-GEN big man eye-DIM lock(CONV)-SUB return-TENT-EV The elders at the splendid palace will lock [their] eyes and return (OS 21.1411) The modern Shuri dialect does not have any of the above usages, and, as far as I can tell, the grammars of other Ryukyuan dialects do not offer any information on the existence of these phenomena. Quite possibly, it is attested somewhere, but the brevity of most grammars of the dialects from islands outside Okinawa does not allow any conclusions at this time. 3.2.1.4 Imperative -e ~ -ǝ The imperative suffix -e is found after consonant verbs, including r-irregular verbs, and after n-irregular verbs. As its shape -e indicates, this imperative suffix probably goes back to monophthongization of the converb suffix -i + *-a, which should be the proper marker of the imperative. This imperative *-a, since it historically follows the converb, probably was originally an auxiliary, but due to the monophthongization *-i-a > -e, Western Old Japanese acquired a new suffix. The imperative suffix -e can follow either consonant verbal roots or the suffixes and auxiliaries that are listed below in Chart 39. There is also an aberrant imperative form -ǝ that occurs only after the benefactive auxiliary -kǝsewith a resulting form -kǝs-ǝ. It is probably due to the progressive assimilation of /e/ > /ǝ/ under the influence of the preceding /ǝ/. Since -kǝse- generally behaves like a vowel verb, we would expect such forms as *-kǝse-yǝ or *-kǝse, but they are not found in the texts.

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

584

Chapter 6

chart 39 Combinations of the imperative suffixes -e ~ -ǝ with preceding suffixes and bound auxiliaries

suffixes and auxiliaries

combination forms

desiderative -(a)na benefactive -kǝseprogressive -erhonorific -as-

-(a)n-e -kǝs-ǝ -er-e -as-e

As can be seen from the above chart, the imperative -e can combine only with the desiderative -(a)na, the progressive -er-, and the honorific -as-, besides, of course, plain consonant verb roots. The aberrant variant -ǝ is found only after the benefactive -kǝse-. Examples: 許能登理母宇知夜米許世泥

kǝnǝ tǝri mǝ uti-yamɛ-kǝse-n-e this bird FP PREF-stop(CONV)-BEN-DES-IMP [I] wish [you] would stop [the singing] of these birds (KK 2) 麻都理許斯美岐叙阿佐受袁勢

matur-i-kǝ-si mi-ki ⁿzǝ as-aⁿz-u wos-e present(HUM)-CONV-come(CONV)-PAST/ATTR HON-rice.wine FP shallowNEG-CONV drink(HON)-IMP Drink the presented rice wine deeply! (KK 39) 岐許志母知袁勢麻呂賀知 (KK script) 枳居之茂知塢勢摩呂餓智 (NK script)

kikǝs-i-mǝt-i-wos-e marǝ-ŋga ti drink(HON)-CONV-hold-CONV-HON-IMP I-POSS father Deign to drink [it], my father (KK 48, NK 39)

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

585

Verbs 夜麻斯呂迩伊斯祁登理夜麻伊斯祁伊斯祁阿賀波斯豆摩迩伊斯岐阿波牟 迦母

Yamasirǝ-ni i-sik-e Tǝriyama i-sik-e i-sik-e a-ŋga pasi-ⁿ-duma-ni i-sik-i ap-am-u kamǝ Yamasirǝ-LOC DLF-follow-IMP Tǝriyama DLF-follow-IMP DLF-follow-IMP I-POSS beloved-DV(ATTR)-spouse-DAT DLF-follow-CONV meet-TENTATTR EP To Yamasirǝ—go there, Tǝriyama! Follow [her] there, follow [her] there! Follow my beloved spouse there and meet [her]! (KK 59) 比能美古爾登余美岐多弖麻都良勢

pi-nǝ mi-ko-ni tǝyǝ mi-ki tatematur-as-e sun-GEN HON-child-DAT abundant HON-rice.wine present(HUM)-HON-IMP Present the abundant rice wine to the honorable child of the Sun! (KK 101) 以嗣箇播箇柂輔智箇柂輔智爾阿弥播利和柂嗣妹慮予嗣爾予嗣予利據祢

isi-kapa-kata-puti kata-puti-ni ami par-i-watas-i mɛ-rǝ yǝs-i-ni yǝs-i yǝr-i-kǝ-n-e stone-river-side-pool side-pool-LOC net spread-CONV-carry.across-CONV mesh-DIM bring.near-NML-COMP bring.near-NML approach-CONV-comeDES-IMP [Girls, I] wish [you] would come near, like [they] bring near the meshes, spreading nets across at the side pool of a stony river (NK 3) 不聴跡雖謂話礼話礼常詔許曾志斐伊波奏強話登言

INA tǝ IP-Ɛ-ⁿDƏ KATAr-e KATAr-e tǝ NƏR-AS-E kǝsǝ Sipï-i pa MAWOS-E SIPÏ-ŋ-GATAR-I tǝ NƏR-U no DV say-EV-CONC speak-IMP speak-IMP DV say-HON-EV FP Sipï-ACT TOP say(HUM)-EV forced-DV(ATTR)-say-NML DV say-FIN Though [I] say: ‘No,’ [you] command [me]: ‘Speak, speak!,’ [but the things that] Sipï says, [you] call a forced speech (MYS 3.237) 用流能伊昧仁越都伎提美延許曽

yoru-nǝ imɛ-ni wo tuŋg-i-te mi-ye-kǝs-ǝ night-GEN dream-LOC EP follow-CONV-SUB see-PASS(CONV)-BEN-IMP please appear [for me] continuously in [my] night dreams (MYS 5.807)

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

586

Chapter 6

烏梅能波奈伊米尓加多良久美也備多流波奈等阿例母布左氣尓于可倍許曽

uMƐ-nǝ pana imɛ-ni katar-aku miyambï-tar-u pana tǝ are [o]mǝp-u sakɛ-ni ukambɛ-kǝs-ǝ plum-GEN blossom dream-LOC tell-NML be.elegand(CONV)-PERF/PROG-ATTR blossom DV we think-FIN rice.wine-LOC let.float(CONV)-BEN-IMP The plum blossoms told [me] in [my] dream: “We think that [we] are elegant blossoms. Please let [us] float in the rice wine [cup]” (MYS 5.852) 此戸開為

KƏNƏ TO PIRAK-As-e this door open-HON-IMP please open this door (MYS 13.3310) 和須礼我比与世伎弖於家礼於伎都之良奈美

wasure-ŋ-gapi yǝse-k-i-te ok-er-e oki-tu sira nami forget(NML)-GEN-shell bring(CONV)-come-CONV-SUB put-PROG-IMP offingGEN/LOC white wave white waves of the offing, bring [to me] the shell of forgetfulness (MYS 15.3629) 都追牟許等奈久波也可敞里麻勢

tutum-u kǝtǝ na-ku paya kaper-i-[i]mas-e have.difficulty-ATTR no-CONV fast return-CONV-HON-IMP Return quickly, without having difficulties (MYS 15.3582) 美也故爾由加波伊毛爾安比弖許祢

miyako-ni yuk-amba imo-ni ap-i-te kǝ-n-e capital-LOC go-COND beloved-DAT meet-CONV-SUB come-DES-IMP if [you] go to the capital, meet [there my] beloved, and come [back] (MYS 15.3687) 之呂多倍能安我之多其呂母宇思奈波受毛弖礼和我世故多太尓安布麻弖尓

sirǝ tapɛ-nǝ a-ŋga sita-ŋ-gǝrǝmǝ usinap-aⁿz-u mot-er-e wa-ŋga se-ko taⁿda n-i ap-u-maⁿde-ni white mulberry.tree.bark.cloth-GEN I-POSS bottom-GEN-garment loseNEG-CONV carry-PROG-IMP I-POSS beloved-DIM direct DV-CONV meet-ATTR-TERM-LOC My beloved, carry [with you] my undergarment from white mulberry tree bark cloth without losing [it], until [we] meet directly (MYS 15.3751)

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

587

Verbs 伊波敞和我勢古多太爾安布末低爾

ipap-e wa-ŋga se-ko taⁿda n-i ap-u-maⁿde-ni pray-IMP I-POSS beloved-DIM direct DV-CONV meet-ATTR-TERM-LOC pray, my beloved, until [we] meet directly (MYS 15.3778) 古非之奈婆古非毛之祢等也

kopï-sin-amba kopï mo sin-e tǝ ya long.for(CONV)-die-COND long.for(CONV) FP die-IMP DV IP Do [you] tell [me]: ‘If [you] die longing, die from longing!’? (MYS 15.3780) 佐伎久安礼

saki-ku ar-e safe-CONV exist-IMP Be safe! (MYS 17.3927) 保等登芸須許欲奈枳和多礼

potǝtǝŋgisu kǝ-yo nak-i watar-e cuckoo this-ABL cry-CONV cross-IMP [Oh,] cuckoo, cross from here, crying (MYS 18.4054) 雪奈布美曽祢

YUKI na-pum-i-sǝ-n-e snow NEG-step.on-CONV-do-DES-IMP do not step on the snow (MYS 19.4228) 伊波敞神多智

ipap-e KAMÏ-tati protect-IMP deity-PLUR protect [him], oh deities! (MYS 19.4240) 多比良氣久於夜波伊麻佐祢都々美奈久都麻波麻多世

tapirakɛ-ku oya pa imas-an-e tutumi na-ku tuma pa mat-as-e safe-CONV parents TOP exist(HON)-DES-IMP obstacle no-CONV spouse TOP wait-HON-IMP Parents, please live safely! Wife, please wait [for me] without obstacles! (MYS 20.4408)

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

588

Chapter 6

和波己藝埿奴等伊弊尓都氣己曽

wa pa kǝŋg-i-[i]ⁿde-n-u tǝ ipe-ni tuŋgɛ-kǝs-ǝ I TOP row-CONV-go.out(CONV)-PERF-FIN DV home-DAT report(CONV)BEN-IMP Please tell [the folks at my] home that I have sailed out (MYS 20.4408) 久佐奈加利曽祢

kusa na-kar-i-sǝ-n-e grass NEG-cut-CONV-do-DES-IMP do not cut the grass (MYS 20.4457) 波都由伎波知敞爾布里之家

patu yuki pa ti-pe n-i pur-i-sik-e first snow TOP thousand-CL DV-CONV fall-CONV-cover-IMP First snow, fall in a thousand layers! (MYS 20.4475) 都止米毛呂毛呂須須賣毛呂母呂

tutǝmɛ morǝ-morǝ susum-e morǝ-mǝrǝ strive(IMP) all-all go forward-IMP all-all strive, everybody, go forward, everybody (BS 18) 己家々己門々祖名不失勤仕奉礼

ONƏ-ŋGA IPE-IPE ONƏ-ŋGA KAⁿDO-KAⁿDO OYA-NƏ NA USINAP-AⁿZ-U TUTOMƐ-TUKAPƐ-MATUr-e self-POSS house-house self-POSS gate-gate ancestor-GEN name lose-NEGCONV serve(CONV)-serve(CONV)-HUM-IMP [At] every house, [behind] every gate, serve without losing the names of your ancestors (SM 16) Comparative Data Level A: Other Japonic A1: Eastern Old Japanese The usage of the imperative suffix -e in Eastern Old Japanese is quite similar to Western Old Japanese with the difference that it is not attested after the progressive markers -ar- or -er-. The aberrant form -ǝ after the benefactive -kǝse- is also attested.

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

589

Verbs 和爾奈多要曾祢

wa-ni na-taye-sǝ-n-e I-DAT NEG-break(CONV)-do-DES-IMP Do not become estranged from me (MYS 14.3378) 奈我己許呂能礼

na-ŋga kǝkǝrǝ nǝr-e you-POSS heart tell-IMP Reveal [to me] your heart (MYS 14.3425) 知々波々我可之良加伎奈弖佐久安礼天伊比之氣等婆是和須礼加祢豆流

titi papa-ŋga kasira kaki-naⁿde sa-ku ar-e te ip-i-si kɛtǝmba ⁿze wasure-kane-t-uru father mother-POSS head PREF-stroke(CONV) safe-CONV exist-IMP DV sayCONV-PAST/ATTR word FP forget(CONV)-NEG/POT(CONV)-PERF-ATTR [I] cannot forget the words: “Be safe!” that [my] father and mother said, stroking [my] head (MYS 20.4346) 伊麻波許伎奴等伊母尓都氣許曽

ima pa kǝŋg-i-n-u tǝ imǝ-ni tuŋgɛ-kǝs-ǝ now TOP row-CONV-PERF-FIN DV beloved-DAT report(CONV)-BEN-IMP Please tell [my] beloved that [I] have sailed out now (MYS 20.4363) 久自我波々佐氣久阿利麻弖

Kuⁿzi-ŋ-gapa pa sakɛ-ku ari-mat-e Kuⁿzi-GEN-river TOP safe-CONV ITER-wait-IMP Wait for [me] (all this time) safely at Kuⁿzi River! (MYS 20.4368) A2: Ryukyuan Forms cognate to the WOJ imperative suffix -e < *-i-a are found both in Old Ryukyuan and modern dialects throughout the Ryukyuan archipelago. Examples: Old Ryukyuan 主里もりあせはつちぎりにきらせ

SIYORI mori ase fa tuti-gir-i n-i kir-as-e Shuri castle warrior TOP ground-(DV)cut-NML DV-CONV cut-HON-IMP Warriors of the Shuri castle, cut [the enemy] as cutting the ground (OS 1.33)

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

590

Chapter 6

いしとかねとあわちへすもせ

isi-to kane-to awa-tife-s-u mo[do]s-e stone-COM metal-COM combine-SUB-do-FIN retreat-IMP make [them] retreat using both stones and metal (OS 2.47) おもろよみおやせせるむよみおやせ

omoro-yo mioyas-e serumu-yo mioyas-e sacred song-ACC present(HUM)-IMP prayer-ACC present(HUM)-IMP Present the sacred song, present the prayer (OS 8.411) Naha

siriziti-ni an i-i sa know-LOC INTER say-IMP EP say since [you] know [it] (Nohara 1986: 74) Shuri

munoo maa-ku kam-ee thing(TOP) delicious-CONV eat-IMP Eat food with gusto! (RKJ 82) Miyako

t∫a:-yu num-e tea-ACC drink-IMP drink tea! Thus, we probably can reconstruct the Proto-Japonic imperative suffix as *-a. 3.2.1.5 Zero Imperative and Its Extended Form -yǝ In contrast to Middle Japanese, where vowel and vowel irregular verbs always have the imperative suffix -yo, there are examples in Western Old Japanese when these verbs just have their roots used as their imperative forms without any following -yǝ, although forms with -yǝ are considerably more frequent.38 Synchronically root imperatives should be treated as having a zero imperative marker. However, diachronically there might be a different explanation 38  I agree with Shirafuji that -yǝ forms are not original, and that -yǝ in all probability goes back to an exclamation particle yǝ (Shirafuji 1987: 140).

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

Verbs

591

that will allow us to solve the famous puzzle of the Old Japanese imperatives, namely why imperative markers are different for consonant and vowel verbs. My solution is that diachronically there is no distinction at all. We saw in section 3.2.1.4 that -e imperatives found after consonant verbs go back to the converb -i + *-a. By default the majority of vowel verbs have converbs that coincide with verbal roots, the only exception being k-irregular, and s-irregular verbs.39 By the rules of Old Japanese phonotactics, vowel clusters are prohibited, therefore, the addition of the imperative *-a would produce a vowel cluster that should be simplified. In the case of the majority of vowel verbs this simplification must have been achieved by the deletion of the imperative *-a, with the verbal root acquiring a function of the imperative. The only puzzling shapes under this scenario are the imperatives kǝ ‘come(IMP)’ and se ‘do(IMP).’ However, se ‘do’ can be treated as *s-i-a ‘do-CONV-IMP,’ so the only puzzling form is kǝ ‘come.’ But even this form can be explained if we suppose that a monphthongization *-i-a > -e took place prior to vowel contraction. Thus, *kǝ-i-a > *kǝ-e > kǝ. The verbal root as an imperative form: although in the following two examples the imperative form se is attested in Western Old Japanese in logographic writing, we can make an educated guess that it is se and not se-yǝ on the basis of the syllable count. Cf. also the example from the Senmyō below, where it is written phonographically, as well as the Eastern Old Japanese imperative form se, cited in the comparative section (see MYS 14.3369). 柳之蘰為吾妹

YANA ŋGÏ-NƏ KAⁿDURA SE WA- ŋG-IMO willow-GEN laurel do(IMP) I-POSS-beloved put on the willow laurel, my beloved (MYS 10.1924) 事計吉為吾兄子相有時谷

KƏTƏ PAKAR-I YƏ-KU SE WA- ŋGA SE-KO AP-ER-U TƏKI ⁿdani thing plan-NML good-CONV do(IMP) I-POSS beloved-DIM meet-PROG-ATTR time RP just [at] the time when [we] meet, do the planning well, my beloved (MYS 12.2949)

39  Although n-irregular verbs are essentially vowel verbs (at least historically), they have imperatives in -e, like consonant verbs, which can be explained along the same lines as the consonant verb imperative marker -e < *-i-a.

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

592

Chapter 6

伊敝妣等波可敝里波也許等

ipe-m-bitǝ pa kaper-i paya kǝ tǝ house-GEN-person TOP return-CONV quick come(CONV) DV [My] home folks say: ‘Come back quickly!’ (MYS 15.3636) 多婢尓弖毛母奈久波也許登

tambi n-i-te mo mǝ na-ku paya kǝ tǝ journey DV-CONV-SUB FP misfortune no-CONV quick come(IMP) DV [my beloved] said: “Come [back] quickly without any misfortune on [your] journey!” (MYS 15.3717) 之流久之米多弖比等能之流倍久

siru-ku simɛ tate pitǝ-nǝ sir-umbɛ-ku distinct-CONV sign erect(IMP) person-GEN know-DEB-CONV erect a distinct sign, so the people could know (MYS 18.4096) 吾等尓可伎无氣念之念婆

WARE-ni kaki-mukɛ OMƏP-I si OMƏP-Amba I-DAT PREF-turn(IMP) love-NML EP love-COND if you indeed love [me], turn to me (MYS 19.4191) 都止米毛呂毛呂須須賣毛呂母呂

tutǝmɛ morǝ-morǝ susum-e morǝ-mǝrǝ strive(IMP) all-all go forward-IMP all-all strive, everybody, go forward, everybody (BS 18) 此事倶仁西止伊射奈布尓

KƏNƏ KƏTƏ TƏMƏ n-i se tǝ iⁿzanap-u-ni this thing together DV-CONV do(IMP) DV entice-ATTR-LOC when [Wono-nǝ Andumapitǝ] enticed [him] saying: ‘Do this (thing) together [with us]’ (SM 19) An extended form with -yǝ is attested in Western Old Japanese with verbal roots and after the perfective and the causative, as can be seen from the following chart:

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

593

Verbs chart 40 Combinations of the extended imperative -yǝ with preceding suffixes and bound auxiliaries

suffixes and auxiliaries

combination forms

perfective -tecausative -(a)se-

-te-yǝ -se-yǝ

However, as far as I can tell, there is only one example of the causative with the following extended imperative -yǝ. Examples with the preceding perfective -teare also rare, so in the majority of cases -yǝ is found after a vowel verb root (if we follow a synchronic description), or after a vowel verb converb (if we follow a diachronic description). Examples: 愛寸事盡手四

URUPASI-ki KƏTƏ TUKUS-I-te-yǝ splendid-ATTR word exhaust-CONV-PERF-IMP exhaust [your] splendid words (MYS 4.661) 阿礼乎婆母伊可爾世与等可

are-womba mǝ ika n-i se-yǝ tǝ ka I-ACC(EMPH) FP how DV-CONV do-IMP DV IP What do [you] think I [should] do? (MYS 5.794) 由布弊爾奈礼婆伊射祢余登

yupu-m-be n-i nar-e-mba iⁿza ne-yǝ tǝ evening-GEN-side DV-CONV become-EV-CON INTER sleep-IMP DV When [it] became evening, and [we] told [him]: ‘[Go to] sleep!’ (MYS 5.904) 吾衣於君令服与

A- ŋGA KƏRƏMƏ KIMI-NI KI-SE-yǝ I-POSS garment lord-LOC wear-CAUS-IMP Put my garment on yourself! (MYS 10.1961) 保等登芸須許許爾知可久乎伎奈伎弖余

potǝtǝŋgisu kǝkǝ-ni tika-ku-wo k-i nak-i-te-yǝ cuckoo here-LOC be close-CONV-ACC come-CONV cry-CONV-PERF-IMP Cuckoo! Come close to here, and cry! (MYS 20.4438)

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

594

Chapter 6

国王伊王位仁坐時方菩薩乃浄戒乎受与止勅天在

KOKU-WAU-i WAU-WI-ni IMAS-U TƏKI pa BOSATU-nǝ ZYAUKAI-wo UKƐ-yǝ tǝ NƏTAMAP-I-te AR-I country-king-ACT king-position-LOC be(HON)-ATTR time TOP bodhisattvaGEN commandmend-ACC receive-IMP DV say(HON)-CONV-SUB exist-FIN [Buddha] said that a king of a country, when [he] is on the throne, [should] accept the commandmends of the Bodhisattva (SM 28) Comparative Data Level A: Other Japonic A1: Eastern Old Japanese Similar to Western Old Japanese, the extended imperative form -yǝ in Eastern Old Japanese is also more frequent than the zero imperative form. The verbal root as an imperative form: 須我麻久良安是加麻可左武許呂勢多麻久良

suŋga-makura aⁿze ka mak-as-am-u kǝ-rǝ se ta-makura sedge-pillow why IP use.as.a.pillow-HON-TENT-ATTR girl-DIM do(IMP) armpillow girl, why would [you] use the sedge pillow? Use [my] arms as a pillow! (MYS 14.3368) 多麻古須氣可利己和我西古

tama ko-suŋgɛ kar-i kǝ wa-ŋga se-ko jewel DIM-sedge cut-CONV come(IMP) I-POSS beloved-DIM my beloved, cut the jewel[-like] little sedge and come (MYS 14.3445) Eastern Old Japanese has extended imperatives ending in -yǝ and -rǝ. It seems that the form in -yǝ is a loan from Western Old Japanese, as it mostly occurs in poems that have no or very limited Eastern Old Japanese features. Meanwhile, -rǝ always occurs in poems that are clearly in Eastern Old Japanese, and therefore, it should be defined as an original Eastern Old Japanese form. Cf. the following examples: 奴流我倍爾安杼世呂登可母

n-uru-ŋga [u]pɛ-ni aⁿ-tǝ se-rǝ tǝ kamǝ sleep-ATTR-POSS top-LOC what-DV do-IMP DV EP besides sleeping [with her], what [else do you] tell [me to] do, I wonder? (MYS 14.3465)

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

Verbs

595

思良久毛能多要爾之伊毛乎阿是西呂等

sira kumo-nǝ taye-n-i-si imo-wo aⁿze se-rǝ tǝ white cloud-COMP break.off(CONV)-PERF-CONV-PAST/ATTR beloved-ACC how do-IMP DV what (lit.: how) [should I] do about [my] beloved who separated [from me] like a white cloud (MYS 14.3517) 比毛多要婆安我弖等都気呂許礼乃波流母志

pimo taye-mba a-ŋga te-tǝ tukɛ-rǝ kǝre n-ǝ paru mǝs-i cord tear-COND your.own-POSS hand-COM attach-IMP this DV-ATTR needle hold-CONV if the cords [of your garment] tear, attach [them] with your own hand, holding this needle (MYS 20.4420) 3.2.1.6 Negative Imperative -una The negative imperative suffix has just one allomorph -una that is found after consonant and vowel verbs alike. Vowel verbs lose the final vowel of their roots before -una. Japanese traditional linguistics treats this form as a particle na being added to the final form of all verbs except r-irregular verbs when it follows the attributive form (Omodaka et al. 1967: 513). Needless to say, this analysis does not make any sense either synchronically or diachronically. From the synchronic point of view it is not clear why the ‘particle’ na would follow different inflectional form of r-irregular verbs. It is quite clear that in both cases the alleged particle is preceded by the vowel /u/. For the diachronic evidence that this vowel /u/ cannot be a final predication form -u, see the discussion of the Ryukyuan comparative data below. The functional difference between this suffix on the one hand and the prefix na- and circumfix na-…-sǝ (see section 3.1.2) on the other is not clear: both can combine with honorifics and both can be used independently, so the distinction must be really subtle, if any exists at all. The existence of two negative imperatives, one expressed by a prefix, and another by a suffix, may again speak strongly in favor of the restructuring of the language, where the original SVO type morphology is gradually fazed out under the influence of the neighboring SOV languages.

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

596 chart 41

Chapter 6 Combinations of the negative imperative -una with preceding suffixes and bound auxiliaries

suffixes and auxiliaries

combination forms

passive -(a)ye-, -rayecausative -(a)simɛhonorific -asbenefactive kǝse-

-(a)y-una -(a)sim-unaa -as-una kǝs-una

a This form is attested only in Eastern Old Japanese, see MYS 14.3399 below.

Examples: 思寐能和倶吾嗚阿娑理逗那偉能古

simbi40-nǝ waku-ŋ-go-wo asar-i-[i]ⁿd-una wi-nǝ ko tuna-GEN young-DV(ATTR)-child-ACC fish-CONV-exit-NEG/IMP boar-GEN child child of the boar, do not fish out a young child of the tuna (NK 95) 吾為類和射乎害目賜名

WA-ŋGA S-Uru waⁿza-wo TƏ ŋGAmɛ-TAMAP-Una I-POSS do-ATTR deed-ACC reproach(CONV)-HON-NEG/IMP Do not reproach my behavior (MYS 4.721) 伊多豆良爾阿例乎知良須奈

itaⁿdura n-i are-wo tir-as-una useless DV-CONV we-ACC fall-CAUS-NEG/IMP Do not let us fall in vain (MYS 5.852a) 人尓所知名

PITƏ-ni SIR-AY-Una person-DAT know-PASS-NEG/IMP Do not [let it] be known by people (MYS 11.2762) 安乎忘為莫

a-wo WASUR-As-uNA 40  There is a play on words in this poem: simbi ‘tuna’ is homonymous with the name of the young noble Simbi.

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

597

Verbs

I-ACC forget-HON-NEG/IMP Do not forget me (MYS 12.3013) 毛美知和礼由伎弖可敝里久流末弖知里許須奈由米

momit-i ware yuk-i-te kaper-i-k-uru-maⁿde tir-i-kǝs-una yumɛ leaves.turn.red/yellow-NML I go-CONV-SUB return-CONV-come-ATTR-TERM fall-CONV-BEN-NEG/IMP at.all Red leaves! Please do not fall at all until I go and come back (MYS 15.3702) 和我世故我可反里吉麻佐武等伎能多米伊能知能己佐牟和須礼多麻布奈

wa-ŋga se-ko-ŋga kaper-i-k-i-[i]mas-am-u tǝki-nǝ tamɛ inǝti nǝkǝs-am-u wasure-tamap-una I-POSS beloved-DIM-POSS return-CONV-come-CONV-HON-TENT-ATTR timeGEN for life leave-TENT-ATTR forget(CONV)-HON-NEG/IMP Do not forget, my beloved, that [I] will stay alive until [you] return (MYS 15.3774) 牟奈許等母於夜乃名多都奈

muna kǝtǝ mǝ oya-nǝ NA tat-una empty word FP ancestor-GEN name break-NEG/IMP do not destroy the name of [your] ancestors [with] empty words (MYS 20.4465) 過无罪無有者捨麻湏奈忘麻湏奈

AYAMAT-I NA-KU TUMI NA-KU AR-Amba SUTE-[i]mas-una WASURE-[i]masuna make.mistake-NML no-CONV sin no-CONV exist-COND abandon(CONV)HON-NEG/IMP forget(CONV)-HON-NEG/IMP if [she] has no[t committed any] sin or mistake, do not abandon [her], do not forget [her] (SM 7) Comparative Data Level A: Other Japonic A1: Eastern Old Japanese The negative imperative suffix -una is also attested in Eastern Old Japanese: 安思布麻之牟奈久都波氣和我世

asi pum-asim-una kutu pak-ɛ wa-ŋga se foot step-CAUS-NEG/IMP shoe put.on-IMP I-POSS beloved Do not let [yourself] step [on it bare-]footed. Put on [your] shoes, my beloved (MYS 14.3399)

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

598

Chapter 6

夜蘇許登乃敝波思氣久等母安良蘇比可祢弖安乎許登奈須那

yaso kǝtǝ-nǝ pe pa siŋgɛ-ku tǝmǝ arasop-i-kane-te a-wo kǝtǝ nas-una eighty word-GEN leaf TOP thick-CONV CONJ resist-CONV-NEG/POT(CONV)-SUB I-ACC word make-NEG/IMP even though many rumors are growing thick [like a bush], do not talk about me, failing to resist (MYS 14.3456) A2: Ryukyuan Ryukyuan comparative data are important for establishing WOJ -una as a separate suffix rather than a sequence of the final predication suffix -u plus a negative particle na. Consider the data in the following chart: chart 42 Final predication and negative imperative forms in the Shuri dialect

gloss

verb root

final predication

negative imperative

write stand push read

kaktat’usyum-

kach-uN tach-uN ’us-uN yum-uN

kak-una tat-una ’us-una yum-una

As was noted above in section 3.2.1.1, a Ryukyuan palatalization of the final consonant of the root such as k > ch and t > ch before the final predication suffix indicates that this suffix was originally attached to the converb -i that triggered this palatalization. However, there are no traces of this palatalization in the negative imperative form. This discrepancy demonstrates that the vowel /u/ in the WOJ final predication suffix -u and the negative imperative -una have different origins, and that consequently the traditional theory of a ‘negative final particle’ na following the final predication form is completely untenable. Examples: Old Ryukyuan おぎもうちはなげくな

o-gimo uti fa nagek-una HON-liver inside TOP lament-NEG/IMP Do not lament inside your soul (OS 10.518)

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

599

Verbs Shuri

sumuchi yum-una book read-NEG/IMP Do not read the book Vowel correspondences in other dialects indicate PR *-ona rather than *-una (Serafim, p.c.), which consequently indicates PJ *-ona as well. It is more than likely that the WOJ negative imperative prefix na-, the Proto-Japonic negative imperative suffix *-ona and the PJ negative *-an- are all interrelated in some way, although at the present point we do not know how to properly analyze the negative imperative *-ona, because we do not know what the element *-o- in this marker could be. 3.2.1.7 Desiderative -ana ~ -na The desiderative suffix has two allomorphs: -ana after consonant verbs and -na after vowel verbs. It has three functions: (a) it expresses the desire of the speaker himself to perform an action (cf. Modern Japanese -tai), (b) it expresses the desire of the speaker for someone else to perform an action (cf. Modern Japanese -te hosii), (c) hortative, i.e., an invitation to do something together. The desiderative suffix -(a)na can combine with the following suffixes and bound auxiliaries: chart 43 Combinations of the desiderative -(a)na with other suffixes and bound auxiliaries

suffixes and auxiliaries

combination forms

imperative -[y]e tentative -(a)mexclamative -(u)mǝ perfective -nperfective -tebenefactive -kǝsenegative imperative na-…-sǝ

-(a)n-e -(a)na-m-(a)na-mǝ -n-ana -te-na -kǝse-n-ea na-…-sǝ-n-e

a When preceded by the benefactive auxiliary -kǝse-, the desiderative -ana must always be followed by the imperative -e. In other words, the form *-kǝse-na is not attested.

One can see that the desiderative suffix -(a)na shows an idiosyncratic behavior: it follows the perfective auxiliaries -n- and -te-, the benefactive auxiliary -kǝse-, and the negative imperative circumfix na-…-sǝ, but it precedes the tentative

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

600

Chapter 6

suffix -am- ~ -m-, the exclamative suffix -umǝ, and the imperative -e. In the last case with the imperative -e, it loses its final vowel before the following suffix, appearing in its special contracted form -an- ~ -n- that will be discussed below separately. Examples: 伊多低於破孺破珥倍廼利能介豆岐齊奈

ita te op-aⁿz-u pa nipo-ⁿ-dǝri-nǝ kaⁿduk-i se-na painful place carry-NEG-CONV TOP grebe-DV(ATTR)-bird-COMP dive-NML do-DES rather than to receive severe wounds, let us dive [together] like grebes (NK 29) 阿母儞擧曾枳擧曳儒阿羅毎矩儞儞播枳擧曳底那

amo-ni kǝsǝ kik-ǝye-ⁿz-u ar-am-ɛ kuni-ni pa kik-ǝye-te-na mother-DAT FP hear-PASS-NEG-CONV exist-TENT-EV land-DAT TOP hearPASS(CONV)-PERF-DES [I] probably will not be heard by [my] mother, but [I] wish to be heard by [the] land! (NK 82) 許智多鶏波乎婆頭勢夜麻能伊波帰爾母為弖許母郎奈牟

kǝt[ǝ]-ita-k-emba woⁿ-Batuse-yama-nǝ ipa kï-ni mǝ wi-te kǝmǝr-ana-m-u rumor-painful-ATTR-COND DIM-Patuse-mountain-GEN rock fortress-LOC FP lead(CONV)-SUB hide-DES-TENT-FIN if rumors are painful, [I] want to take [you] along to a rocky fortress on the Small Patuse mountain and hide away (FK 1) 今者許藝乞菜

IMA pa kǝŋg-i-iⁿde-na now TOP row-CONV-exit-DES Let [us] row out [to the sea] now (MYS 1.8) 雲谷裳情有南畝可苦佐布倍思哉

KUMO ⁿdani mo KƏKƏRƏ AR-Ana-mo kakus-ap-umbɛ-si YA cloud RP FP heart exist-DES-EXCL hide-ITER-DEB-FIN IP [I] wish at least the clouds [would] have feelings! Do [they] have to hide [Mt. Miwa] all the time? (MYS 1.18) 君尓因奈名

KIMI-ni YƏR-I-n-ana lord-DAT approach-CONV-PERF-DES [I] want to cling to [my] lord (MYS 2.114) Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

601

Verbs 烏梅能波奈 … 加射之尓斯弖奈

uMƐ-nǝ pana … kaⁿzas-i n-i s-i-te-na plum-GEN blossom … decorate-NML DV-CONV do-CONV-PERF-DES Let [us] decorate [our hair] with plum blossoms (MYS 5.820) 出波之利伊奈奈等思騰許良爾佐夜利奴

IⁿDE-pasir-i in-ana tǝ OMƏP-Ɛ-ⁿdǝ kǝ-ra-ni sayar-i-n-u exit(CONV)-run-CONV go-DES DV think-EV-CONC child-PLUR-DAT be prevented-CONV-PERF-FIN although [I] think that [I] would like to run away, [I] am prevented by [my] children (MYS 5.899) 妹許将遣黄葉手折奈

IMO-ŋgari YAR-AM-U MOMIT-I TA-WOR-Ana beloved-DIR send-TENT-ATTR leaves.turn.red/yellow-NML hand-break-DES [I] would like to break with [my] hand [branches with] maple leaves to send to [my] beloved (MYS 9.1758) 伊母爾見勢武爾和多都美乃於伎都白玉比利比弖由賀奈

imǝ-ni MI-se-m-u-ni wata-tu mi-nǝ oki-tu SIRA TAMA pirip-i-te yuk-ana beloved-DAT see-CAUS-TENT-ATTR-LOC sea-GEN/LOC dragon-GEN offingGEN/LOC white jewel pick up-CONV-SUB go-DES [I] want to go, picking up the white jewels from the offing of the sea dragon in order to show [them] to [my] beloved (MYS 15.3614) 秋風尓比毛等伎安氣奈多太奈良受等母

AKI KAⁿZE-ni pimo tǝk-i-akɛ-na taⁿda nar-aⁿz-u tǝmǝ autumn wind-LOC cord untie-CONV-open-DES direct be-NEG-FIN CONJ [I] wish that the autumn wind will untie the cords, even if [it] is not directly (MYS 20.4295) 美知乎多豆祢奈

miti-wo taⁿdune-na way-ACC seek-DES [I] want to seek the Way (MYS 20.4468) 安之婢乃波奈乎蘇弖尓古伎礼奈

asimbi-nǝ pana-wo soⁿde-ni kok-i-[i]re-na andromeda-GEN flower-ACC sleeve-LOC rub.through-CONV-insert-DES [I] want to rub andromeda flowers into [my] sleeves (MYS 20.4512)

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

602

Chapter 6

己乃美阿止夜与呂豆比賀利乎波奈知伊太志毛呂毛呂須久比和多志多麻波 奈須久比多麻波奈

kǝnǝ mi-atǝ ya-yǝrǝⁿdu pikari-wo panat-i-iⁿdas-i morǝ-morǝ sukup-i-watas-itamap-ana sukup-i-tamap-ana this HON-foot.print eight-ten.thousand light-ACC emanate-CONV-put.out-CONV all-all save-CONV-carry.across-CONV-HON-DES save-CONV-HON-DES These footprints emanate myriad lights, and [I] want [them] to save everybody, leading [them] across, [I] want [them] to save [them] (BS 4) 一二人乎治賜波奈止那毛所思行湏

PITƏ-RI PUTA-RI-wo WOSAMƐ-TAMAp-ana tǝ namo OMƏP-OS-I-MEs-u one-CL two-CL-ACC reward(CONV)-HON-DES DV FP think-HON-CONVHON-ATTR [I] deign to think that [I] want to reward one or two [of my subjects] (SM 10) 3.2.1.7.1 Special Contracted Form -an- ~ -nThe desiderative suffix has a special contracted form -an- ~ -n- when it is followed by the imperative suffix -e or by the tentative suffix -am-. In these cases the desiderative suffix functions not as a sentence-final, but as a word nonfinal suffix. The combination -(a)n-am- of the desiderative and the tentative will be described below in the section on the tentative -am-. The desiderativeimperative form -an-e ~ -n-e can occur after the stem of a verb, after the benefactive -kǝse-, and after -sǝ ‘do’ which occurs as a part of the negative circumfix na-…-sǝ. The desiderative-imperative form conveys a request or a command to the addressee that also implies a desire on the part of the speaker. Examples: 許能登理母宇知夜米許世泥

kǝnǝ tǝri mǝ uti-yamɛ-kǝse-n-e this bird FP PREF-stop(CONV)-BEN-DES-IMP [I] wish [you] would stop [the singing] of these birds (KK 2) 以嗣箇播箇柂輔智箇柂輔智爾阿弥播利和柂嗣妹慮予嗣爾予嗣予利據祢

isi-kapa-kata-puti kata-puti-ni ami par-i-watas-i mɛ-rǝ yǝs-i-ni yǝs-i yǝr-i-kǝ-n-e stone-river-side-pool side-pool-LOC net spread-CONV-carry.across-CONV mesh-DIM bring near-NML-COMP bring near-NML approach-CONV-comeDES-IMP [Girls, I] wish [you] would come near, like [they] bring near the meshes, spreading nets across at the side pool of a stony river (NK 3)

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

603

Verbs 於辞寐羅箇祢瀰和能等能渡烏

os-i-mbirak-an-e Miwa-nǝ tǝnǝ to-wo push-CONV-open-DES-IMP Miwa-GEN pavilion door-ACC push open the door of the Miwa pavilion (NK 17) 娑奘岐等羅佐泥

saⁿzaki tǝr-as-an-e wren grab-HON-DES-IMP [I wish you would] catch the wren! (NK 60) 余呂豆余尓伊麻志多麻比提阿米能志多麻乎志多麻波祢

yǝrǝⁿdu yǝ-ni imas-i-tamap-i-te amɛ-nǝ sita mawos-i-tamap-an-e ten.thousand age-LOC exist(HON)-CONV-HON-CONV-SUB heaven-GEN below report(HUM)-CONV-HON-DES-IMP May [you] live for ten thousand generations, and report [to the emperor about things in the country] under Heaven (MYS 5.879) 奈良能美夜故尓咩佐宜多麻波祢

Nara-nǝ miyako-ni mes-aŋgɛ-tamap-an-e Nara-GEN capital-LOC call(HON)(CONV)-raise(CONV)-HON-DES-IMP [I] want [you] to summon [me to come] up to the capital of Nara (MYS 5.882) 美也故爾由加波伊毛爾安比弖許祢

miyako-ni yuk-amba imo-ni ap-i-te kǝ-n-e capital-LOC go-COND beloved-DAT meet-CONV-SUB come-DES-IMP if [you] go to the capital, meet [there my] beloved, and come [back] (MYS 15.3687) 雪奈布美曽祢

YUKI na-pum-i-sǝ-n-e snow NEG-step.on-CONV-do-DES-IMP do not step on the snow (MYS 19.4228) 多比良氣久於夜波伊麻佐祢都々美奈久都麻波麻多世

tapirakɛ-ku oya pa imas-an-e tutumi na-ku tuma pa mat-as-e safe-CONV parents TOP exist(HON)-DES-IMP obstacle no-CONV spouse TOP wait-HON-IMP Parents, please live safely! Wife, please wait [for me] without obstacles! (MYS 20.4408)

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

604

Chapter 6

久佐奈加利曽祢

kusa na-kar-i-sǝ-n-e grass NEG-cut-CONV-do-DES-IMP do not cut the grass (MYS 20.4457) Comparative Data Level A: Other Japonic A1: Eastern Old Japanese The desiderative suffix -ana ~ -na is also attested in Eastern Old Japanese. Examples: 多可伎祢爾久毛能都久能須和礼左倍爾伎美爾都吉奈那

taka-ki ne-ni kumo-nǝ tuk-u-nǝsu ware sapɛ n-i kimi-ni tuk-i-n-ana high-ATTR peak-LOC cloud-GEN attach-ATTR-COMP I RP DV-CONV lord-DAT attach-CONV-PERF-DES Even I would like to cling to [my] lord like clouds cling to a high peak (MYS 14.3514) 和須礼波勢奈那伊夜母比麻須爾

wasure pa se-n-ana iya [o]mǝp-i-mas-u-ni forget(NML) TOP do-NEG-DES more.and.more think-CONV-increase-ATTR-LOC [I] wish [you] will not forget [me], because [my] desire [for you] increases more and more (MYS 14.3557) The desiderative-imperative form -an-e ~ -n-e is found quite frequently in Eastern Old Japanese: 和爾奈多要曾祢

wa-ni na-taye-sǝ-n-e I-DAT NEG-break(CONV)-do-DES-IMP Do not become estranged from me (MYS 14.3378) 加比利久麻弖尓已波比弖麻多祢

kapir-i-k-u-maⁿde-ni ipap-i-te mat-an-e return-CONV-come-ATTR41-TERM-LOC pray-CONV-SUB wait-DES-IMP pray and wait [for me] until [I] return (MYS 20.4339)

41  Formally the form k-u looks like final, but since it is followed by the case marker -maⁿde, functionally it must be attributive. In all probability, we simply have a raising -u < *-o here: k-u < *k-o. Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

605

Verbs 阿我母弖能和須例母之太波都久波尼乎布利佐氣美都々伊母波之奴波尼

a-ŋga [o]mǝte-nǝ wasure-m-ǝ siⁿda pa Tukumba ne-wo purisakɛ-mi-tutu imǝ pa sinop-an-e I-POSS face-GEN forget-TENT-ATTR time TOP Tukumba peak-ACC look. up(CONV)-look(CONV)-COOR beloved TOP yearn-DES-IMP When [you, my] beloved will be forgetting my face, please yearn for me, looking up at the Tukumba peak (MYS 20.4367) 安之我良乃美祢波保久毛乎美等登志努波祢

Asiŋgara-nǝ mi-ne pap-o kumo-wo mi-tǝtǝ sinop-an-e Asiŋgara-GEN HON-peak crawl-ATTR cloud-ACC see(CONV)-COOR yearnDES-IMP yearn [for me] while looking at the clouds crawling at the Asiŋgara peak (MYS 20.4421) A2: Ryukyuan There is also a desiderative suffix -ana in Ryukyuan attested both in Old Ryukyuan and in modern dialects. In modern Shuri this form also has a hortative meaning (RKJ 399), see the example below. Examples: Old Ryukyuan あやよりくせよりみらな

aya yor-i kuse yor-i mi-r-ana beautiful dance-NML rare dance-NML see-ATTR(?)-DES [I] want to see a beautiful dance, a rare dance (OS 14.1031) Classical Ryukyuan 恩納嶽あがた里が生まれ島森もおしのけてこがたなさな

UNNA-DAKI agata SATU-ga Umare-JIMA MUI mo os-i-noke-te kogata nas-ana Unna-peak over.there village-POSS born(CONV)-island mountain FP pushCONV-put away(CONV)-SUB here make-DES On the other side of the peak Unna [there is] the village on the island where [my beloved] was born. Pushing the mountain aside, [I] would like to place [it] here (RK 1243) Shuri

too sanruu ’isuj-i tuur-ana INTER Sanruu hurry-CONV pass-DES Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

606

Chapter 6

Hey, Sanruu, let us go fast (Nishioka and Nakahara 2000: 115) Since the desiderative suffix -(a)na is found in all branches of the Japonic language family, we can safely reconstruct PJ *-ana. However, there are no apparent external parallels. 3.2.1.8 Subjunctive -amasi ~ -masi The suffix -(a)masi is a marker of the subjunctive mood. It has two allomorphs: -amasi that is found after consonant, r-irregular, and n-irregular verbs, and -masi that is found after vowel verbs (including irregular vowel verbs). The same rule applies to the preceding suffixes and bound auxiliaries. The subjunctive suffix -amasi ~ -masi can combine with the following suffixes and bound auxiliaries: chart 44 Combinations of the subjunctive suffix -(a)masi with other suffixes and bound auxiliaries

suffixes and auxiliaries

combination forms

perfective -nperfective -tecausative -seprogressive -erconditional converb -amba

-n-amasi -te-masi -se-masia -er-amasi -(a)mas-emba

a This combination is attested only after strong vowel verbs mi- ‘to see’ and ki- ‘to wear.’

When -(a)masi combines with the conditional converb -amba, it becomes a word-non-final suffix. The resulting sequence -(a)mas-emba is due to the monophthongization of the vowel /i/ in -(a)masi and the vowel /a/ in -amba as a vowel /e/. Since the morphemic boundary is lost due to this fusion, the segmentation of this sequence as -(a)mas-emba is certainly artificial and represents only an approximation. Already Yamada Yoshio noticed that -(a)masi behaves like an attributive, because it is frequently found before the conjunction mǝnǝ ~ mǝnǝwo (Yamada 1954: 282–283).42 I should add that -(a)masi is also frequently found before the accusative -wo when it is used as a conjunction (see chapter 4, section 1.2.2.8 for details). 42  The form mǝnǝwo certainly derives historically from mǝnǝ ‘thing’ plus the accusative -wo. Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

607

Verbs

In many cases, -(a)masi appears in Western Old Japanese in parallel constructions like: -amba … -(a)masi, -amas-emba … -(a)masi, ‘if [it] were … then [it] would be.’ This usage is somewhat different from the one found in Middle Japanese where besides the construction -aba … -(a)masi, two different constructions are used frequently: -(a)masika-ba … -(a)masi and -(a)masi … -(a) masi, neither of which are attested in Western Old Japanese. The Western Old Japanese construction -amba … -(a)masi is predominantly attested in the form -s-emba … -(a)masi, with the past tense auxiliary -s- (< *-si, attributive form) preceding the conditional converb -amba.43 Examples: 比登都麻都比登迩阿理勢婆多知波気麻斯袁岐奴岐勢麻斯袁

pitǝ-tu matu pitǝ n-i ar-i-s-emba tati pakɛ-masi-wo kinu ki-se-masi-wo one-CL pine person DV-CONV exist-CONV-PAST/ATTR-COND long.sword make.wear-SUBJ-ACC garment wear-CAUS-SUBJ-ACC Oh, lone pine, if [you] were a human being, [I] would make [you] wear a long sword, [I] would make [you] wear a garment, but … (KK 29) 多遅比怒迩泥牟登斯理勢波多都碁母母母知弖許志母能

Taⁿdipi no-ni ne-m-u tǝ sir-i-s-emba tat-u-ŋ-gǝmǝ mǝ mǝt-i-te kǝ-masi mǝnǝ Taⁿdipi field-LOC sleep-TENT-FIN DV know-CONV-PAST/ATTR-COND standATTR-DV(ATTR)-rush.mat FP hold-CONV-SUB come-SUBJ CONJ If [I] would know that [I] would sleep at the Taⁿdipi field, [I] would also bring a standing rush mat (KK 75) 和加久閇爾韋泥弖麻斯母能淤伊爾祁流加母

waka-ku pɛ-n-i wi ne-te-masi mǝnǝ oyi-n-i-ker-u kamǝ young-CONV ?-DV-CONV bring(CONV) sleep(CONV)-PERF-SUBJ CONJ age (CONV)-PERF-CONV-RETR-ATTR EP [I] would have brought [her with me] and have slept [with her] if [she] were young, but it turned out that [she] has become old, alas! (KK 93) 柯彼能矩盧古磨矩羅枳制播伊志歌孺阿羅磨志

kapï-nǝ kuro koma kura ki-s-emba i-sik-aⁿz-u ar-amasi Kapï-GEN black stallion saddle put.on-PAST/ATTR-COND DLF-reach-NEGCONV exist-SUBJ if [he] would put a saddle on the black stallion from Kapï, [he] would not reach here [on time] (NK 81a) 43  Similar to the form -(a)mas-emba, described above, this past conditional form -s-emba also results from the monophthongization of *ia > e, producing fusion that obliterates the morphemic boundary. Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

608

Chapter 6

宇良志麻能古我多麻久志義阿気受阿理世波麻多母阿波麻志遠

urasima-nǝ ko-ŋga tama kusiŋgɛ akɛ-ⁿz-u ar-i-s-emba mata mǝ ap-amasi-wo Urasima-GEN child-POSS jewel box open-NEG-CONV exist-CONV-PAST/ ATTR-COND again FP meet-SUBJ-ACC If the boy Urasima did not open the jewel box, [he] would meet again [the dragon’s daughter], but [it did not happen] (FK 15) 速來而母見手益物乎山背高槻村散去奚留鴨

PAYA K-I-TE mǝ MI-te-masi MƏNƏwo YAMASIRƏ-NƏ TAKA TUKÏ mura TIR-I-n-i-ker-u kamo fast come-CONV-SUB FP see(CONV)-PERF-SUBJ CONJ Yamasirǝ-GEN high zelkova group fall-CONV-PERF-CONV-RETR-ATTR EP Although [I] would come fast and look [at them], most of the zelkova [flowers] in Yamasirǝ have fallen, alas! (MYS 3.277) 可久斯良摩世婆久奴知許等其等美世摩斯母乃乎

ka-ku sir-amas-emba kun[i]-uti kǝtǝ-ŋgǝtǝ mi-se-masi mǝnǝwo thus-CONV know-SUBJ-COND land-inside thing-thing see-CAUS-SUBJ CONJ if [I] would know that [I] would show [her] all things in the land, but … (MYS 5.797) 國尓阿良婆父刀利美麻之家尓阿良婆母刀利美麻志

KUNI-ni ar-amba TITI tor-i-mi-masi IPE-ni ar-amba PAPA tor-i-mi-masi province-LOC exist-COND father hold-CONV-see-SUBJ home-LOC existCOND mother hold-CONV-see-SUBJ if [I] were in [my] province, [my] father would take care [of me], if [I] were in [my] home, [my] mother would take care [of me] (MYS 5.886) 遠妻四高尓有世婆不知十方手綱乃濱能尋來名益

TƏPO TUMA si TAKA-ni AR-I-s-emba SIR-AⁿZ-U tǝmo TAⁿDUNA-nǝ PAMA-nǝ TAⁿDUNE-K-I-n-amasi distant wife EP Taka-LOC exist-CONV-PAST/ATTR-COND know-NEG-FIN CONJ [TaNduna-GEN beach-COMP—makura-kotoba] visit seek-(CONV)come-CONV-PERF-SUBJ If my wife, who is far away [from me], were in Taka, even if [I] did not know, [I] would come looking for [her] (MYS 9.1746)

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

609

Verbs 人跡不在者桑子尓毛成益物乎

PITƏ tǝ AR-AⁿZ-U pa KUWA-KO n-i mo NAR-Amasi MƏNƏwo person DV exist-NEG-CONV TOP mulberry-child DV-CONV FP become-SUBJ CONJ if [I] was not a human being, [I] would [like to] become a silkworm, but … (MYS 12.3086) 大船尓伊母能流母能尓安良麻勢婆羽具久美母知弖由可麻之母能乎

OPO PUNE-ni imǝ nǝr-u mǝnǝ n-i ar-amas-emba PA- ŋ-gukum-i mǝt-i-te yukamasi mǝnǝwo big boat-LOC beloved board-ATTR thing DV-CONV exist-SUBJ-COND wingLOC-cover-CONV hold-CONV-SUB go-SUBJ CONJ If [it] were the case that [my] beloved boarded [my] big boat, [I] would go holding her under [my] wings, but … (MYS 15.3579) 欲和多流月尓安良麻世婆伊敞奈流伊毛爾安比弖許麻之乎

yo watar-u TUKÏ n-i ar-amas-emba ipe-n-ar-u imo-ni ap-i-te kǝ-masi-wo night cross-ATTR moon DV-CONV exist-SUBJ-COND home-LOC-exist-ATTR beloved-DAT meet-CONV-SUB come-SUBJ-ACC if [I] were a moon that goes over through the night, [I] would come to meet my beloved, who is at home (MYS 15.3671) 君我牟多由可麻之毛能乎

KIMI-ŋga muta yuk-amasi monǝwo lord-POSS together go-SUBJ CONJ although [I] would go together with [my] lord … (MYS 15.3773) 家布毛可母美也故奈里世婆見麻久保里尓之能御馬屋乃刀尓多弖良麻之

kepu mo kamǝ miyako-n-ar-i-s-emba MI-m-aku por-i nisi-nǝ MI-MAYA-nǝ to-ni tat-er-amasi today FP EP capital-LOC-exist-CONV-PAST/ATTR-COND see-TENT-NML desire -CONV west-GEN HON-stable-GEN outside-LOC stand-PROG-SUBJ If [I] were in the capital today, too, [I] would be standing outside the Western Imperial Stables, wanting to see [you]! (MYS 15.3776) 保里江尓波多麻之可麻之乎

Pori-ye-ni pa tama sik-amasi-wo Pori-bay-LOC TOP jewel cover-SUBJ-ACC [I] would cover the Pori bay with jewels, but … (MYS 18.4056)

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

610

Chapter 6

安佐之保美知尓与流許都美可比尓安里世婆都刀尓勢麻之乎

asa sipo mit-i-ni yǝr-u kǝtumi kapi n-i ar-i-s-emba tuto n-i se-masi-wo morning tide full-NML-LOC approach-ATTR trash shellfish DV-CONV existCONV-PAST/ATTR-COND souvenir DV-CONV do-SUBJ-ACC if the trash that is brought up by the full morning tide were shellfish, [I] would bring [it] as a souvenir, but … (MYS 20.4396) Comparative Data Level A: Other Japonic A1: Eastern Old Japanese The subjunctive suffix -(a)masi is attested in Eastern Old Japanese as well. 伊利奈麻之母乃伊毛我乎杼許爾

ir-i-n-amasi mǝnǝ imo-ŋga woⁿ-dǝkǝ-ni enter-CONV-PERF-SUBJ CONJ beloved-POSS DIM-bed-LOC although [I] would [like to] enter the bed of [my] beloved (MYS 14.3354) 伊呂夫可久世奈我許呂母波曽米麻之乎

irǝ-m-buka-ku se-na-ŋga kǝrǝmǝ pa sǝmɛ-masi-wo color-LOC-deep-CONV beloved-DIM-POSS garment TOP dye-SUBJ-ACC Although [I] would dye deeply [my] beloved’s garment … (MYS 20.4424) A2: Ryukyuan The subjunctive suffix -(a)masi is attested only in Old Ryukyuan in ten examples in the Omoro sōshi (Torigoe 1968: 178–179). The absence of reflexes in the modern dialects strongly suggests that it is a loan from Middle Japanese. Old Ryukyuan みちへいちへいきぬはまし

mi-tife i-tife iki nup-amasi look(CONV)-SUB go(CONV)-SUB breath extend-SUBJ [by] going and looking [at it], [I] would extend [my] life (OS 11.557) うきおほちか世やてやもゝかめむすへまし

u-ki ofo ti-ka YO ya-te ya momo kame mu sufe-masi great-ATTR big father-POSS world be(CONV)-SUB TOP hundred jar FP place-SUBJ because it is the world of great ancestors, [we] would place as much as a hundred jars [of rice wine] (OS 11.559)

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

611

Verbs

3.2.1.9 Suppositional -urasi ~ -asi The suppositional suffix has reliable evidence for two allomorphs: -urasi and -asi. -Urasi occurs after all verbs with the possible exception of strong vowel verbs. The expected allomorph after strong vowel verbs is -rasi. However, there is only one example of the suppositional form of a strong vowel verb (ni- ‘to boil’) found in the Western Old Japanese texts: see the example from MYS 10.1879 below. Unfortunately, the part of the verb preceding the phonographically spelled -rasi is written logographically, so we will never know for certain, although it is most likely that the form is indeed *ni-rasi. The allomorph -asi is a special form that appears after certain paradigmatic forms. It will be discussed in detail below. As a rule -urasi is found as a sentence-final suffix, but in Western Old Japanese, in contrast to Middle Japanese, it can also be followed by the adjectival attributive form -ki. Thus, in this case it appears as a word-non-final suffix in the paradigmatic form -urasi-ki. chart 45 Combinations of the suppositional suffix -urasi ~ -asi with other suffixes and bound auxiliaries

suffixes and auxiliaries

combination forms

honorific -asperfective -nprogressive -erperfective-progressive -tarretrospective -kerexclamative -(u)mǝ attributive -ki

-as-urasi -n-urasi -er-asi -tar-urasi -ker-asi -urasi-mǝ -urasi-ki

Thus, the suppositional -urasi ~ -asi can combine with preceding markers of honorification, aspect, and retrospection, and it can be followed by the adjectival attributive -ki and the exclamative -(u)mǝ. The basic meaning of the suppositional is a conjecture, often associated with visual or other sensual perception. Examples: 那賀美古夜都毘迩斯良牟登加理波古牟良斯

na-ŋga miko ya tumbi n-i sir-am-u tǝ kari pa ko [u]m-urasi you-POSS prince EP ? DV-CONV rule-TENT-FIN DV wild goose TOP egg bear-SUP A wild goose probably laid an egg to show that your prince will rule ?-ly (KK 73) Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

612

Chapter 6

祁布母加母佐加美豆久良斯多加比加流比能美夜比登

kepu mǝ kamǝ saka-miⁿduk-urasi taka pikar-u pi-nǝ miya-pitǝ today FP EP rice.wine-soak-SUP high shine-ATTR sun-GEN palace-person Today, the courtiers from the palace of the High Shining Sun seem to be inebriated in wine, too! (KK 102) 奴弖由良久母淤岐米久良斯母

nute yurak-umǝ okimɛ k-urasi-mǝ bell sound-EXCL Okimɛ come-SUP-EXCL The bells are sounding! It seems that Okimɛ is coming! (KK 111) 蘇餓能古羅烏於朋枳瀰能莵伽破須羅志枳

Soŋga-nǝ ko-ra-wo opo kimi-nǝ tukap-as-urasi-ki Soŋga-GEN child-PLUR-ACC great lord-GEN use-HON-SUP-ATTR It looks like the sovereign [can] use the children of Soŋga (NK 103) 神代従如此尓有良之古昔母然尓有許曽虚蝉毛嬬乎相挌良思吉

KAMÏ YƏ-YORI KA-KU n-i AR-Urasi inisipe mǝ SIKA n-i AR-E kǝsǝ UTUSEMI mo TUMA-wo ARASOP-Urasi-ki deity age-ABL thus-CONV DV-CONV exist-SUP old.times FP thus DV-CONV exist-EV FP mortal FP spouse-ACC compete-SUP-ATTR [It] seems to be like that from the age of gods. Mortals seemed to compete for [their] spouses in the old times, too (MYS 1.13) 毛毛等利能己惠能古保志枳波流岐多流良斯

momo tǝri-nǝ kǝwe-nǝ koposi-ki paru k-i-tar-urasi hundred bird-GEN voice-GEN be.missing-ATTR spring come-CONV-PERF/ PROG-SUP It looks like the spring [with] voices of a hundred birds, that [I] missed, [finally] has come (MYS 5.834) 塩乾尓豆良志

SIPO PÏ-n-i-ker-asi tide dry(CONV)-PERF-CONV-RETR-SUP It seems that the tide has retreated (lit.: dried up) (MYS 9.1671) 春立奴良志

PARU TAT-I-n-urasi spring stand-CONV-PERF-SUP It seems that the spring has arrived (MYS 10.1819)

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

613

Verbs 鸎之春成良思

U ŋGUPISU-NƏ PARU N-I NAR-Urasi bush.warbler-GEN spring DV-CONV become-SUP [It] looks like [it] became a bush warbler’s spring (MYS 10.1845) 呎嬬等四春野之菟芽子採而畠良思文

WOTƏME-RA si PARU NO-NƏ UPA ŋGÏ TUM-I-TE NI-rasi-mo maiden-PLUR EP spring field-GEN upaŋgï pick.up-CONV-SUB boil-SUP-EXCL It seems that maidens pick up upaŋgï grass at the spring fields and boil [it]! (MYS 10.1879) 此夜等者沙夜深去良之

KƏNƏ YO-ra pa sa-YO PUKƐ-n-urasi this night-LOC TOP PREF-night deepen(CONV)-PERF-SUP It seems that (in this night) the night has grown deep (MYS 10.2224) 吾妹子之阿乎偲良志

WA- ŋG-IMO-KO si a-wo SINOP-Urasi I-POSS-beloved-DIM EP I-ACC long.for-SUP It seems that my beloved longs for me (MYS 12.3145) 欲波安氣奴良之

yo pa akɛ-n-urasi night TOP brighten(CONV)-PERF-SUP It seems that [it] has dawned (MYS 15.3598) 多奈波多之船乗須良之

Tanambata si PUNA-NƏR-I s-urasi Weaver EP boat-board-NML do-SUP It seems that the Weaver boards the boat (MYS 17.3900) 伊尓之敝乎於母保須良之母和期於保伎美

inisipe-wo omǝp-os-urasi-mǝ wa-ŋgǝ opo kimi past-ACC think-HON-SUP-EXCL I-POSS great lord My sovereign who seems to think about the past! (MYS 18.4099) 於保吉美乃都藝弖賣須良之多加麻刀能努敝

opo kimi-nǝ tuŋg-i-te mes-urasi Takamato-nǝ no-pe great lord-GEN follow-CONV-SUB look(HON)-SUP Takamato-GEN field-side Fields of Takamato that [my] sovereign seemed to look continuously at (MYS 20.4510) Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

614

Chapter 6

3.2.1.9.1 Special Form -asi After the verbs nar- ‘to become’ and ar- ‘to exist’ as well as after the derivatives based on the latter, such as nar- ‘to be,’ the auxiliary retrospective -ker-, and the progressive -er, -urasi appears in its special form -asi. Not all auxiliaries based on ar- ‘to exist’ follow this rule, e.g., the perfective-progressive -tar- and -urasi appear as -tar-urasi, and never as *-tar-asi (see the example from MYS 5.834 above). Furthermore, the parallel forms nar-urasi ‘become-SUP’44 and arurasi also occur in the texts (see the examples from MYS 1.13 and MYS 10.1845 above), although the contracted forms are much more frequent than the uncontracted ones. There are two possibilities to explain the form -ker-asi (there are no examples of *-ker-urasi) and the alternative forms ar-asi and nar-asi. It is possible that we are simply dealing here with a contraction due to the intervocalic *-rloss. But it is equally possible that -urasi is historically bimorphemic, consisting of a stative non-past marker *-ur- and a suppositional suffix -asi, an analysis suggested by Russell (2005: 652). In this case, which seems to be more realistic as a trisyllabic suffix is likely to have a complex morphemic origin, the forms with -asi like -ker-asi are archaic remainders, while the forms exhibiting -urasi are innovations. Russell’s proposal can be further supported by the fact that the negative tentative form in -(a)ⁿzi (discussed in section 3.2.1.10 below) is likely to go back to the contraction of the negative suffix -an- and the suppositional suffix -asi. It also might be significant that -asi is found after stative verbs which would be unlikely to include the stative marker *-ur-. Examples of -asi: 迦久能尾奈良志

ka-ku nǝmï nar-asi thus-CONV RP be-SUP [It] is likely to be just this way (MYS 5.804) 諾石社見人毎尓語嗣偲家良思吉

UmBƐ-si kǝsǝ MI-RU PITƏ ŋGƏTƏ n-i KATAR-I-TU ŋG-I SINOP-I-ker-asi-ki be.proper-FIN FP see-ATTR person every DV-CONV talk-CONV-pass-CONV yearn-CONV-RETR-SUP-ATTR [It] is proper that every person who sees [this beach], seems to yearn [for it] and tells others [about it] (MYS 6.1065)

44  But the uncontracted form *nar-urasi ‘be-SUP’ is not attested as far as I can tell. Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

Verbs

615

由槻我高仁雲居立有良志

yu tukï-ŋga TAKƐ-ni KUMOWI TAT-Er-asi sacred zelkova-POSS peak-LOC cloud rise-PROG-SUP [It] seems that the clouds are rising at the Sacred Zelkova peak (MYS 7.1087) This is the only example of -er-asi in the Western Old Japanese texts. Certainly, the form -er-asi here cannot be confirmed without doubts due to the partial logographic spelling. Commentators of the Man’yōshū almost invariably read it as tat-er-urasi here, cf. e.g., Takagi et al. 1959: 203, Kinoshita 2001 (CD-ROM edition). However, there are two powerful arguments for reading 立有良志 as TAT-Er-asi rather than as TAT-ER-Urasi here. First, tat-er-urasi brings the syllable count in the last line to eight, violating the meter. Second, we should not forget that the progressive suffix -er- is derived from the converb -i and ar- ‘to exist,’ and that the most frequent form is ar-asi and not ar-urasi. 和我多妣波比左思久安良思許能安我家流伊毛我許呂母能阿可都久見礼婆

wa-ŋga tambi pa pisasi-ku ar-asi kǝnǝ a-ŋga ker-u imo-ŋga kǝrǝmǝ-nǝ aka tuk-u MI-re-mba I-POSS journey TOP long-CONV exist-SUP this I-POSS wear(PROG)-ATTR beloved-POSS garment-GEN dirt attach-ATTR see-EV-CON It seems that my journey was long, when [I] see that this garment of [my] beloved which I am wearing became dirty (MYS 15.3667) 豊乃登之思流須登奈良思雪能敷礼流波

TƏYƏ n-ǝ tǝsi sirus-u tǝ nar-asi YUKI-nǝ pur-er-u pa abundand DV-ATTR year show.a.sign-ATTR DV become-SUP snow-GEN fallPROG-ATTR TOP A snowfall seems to become a good omen for an abundand year (MYS 17.3925) 加武賀良奈良之

kamu-ŋ-gara nar-asi deity-DV(ATTR)-nature be-SUP [It] is likely to be [its] divine nature (MYS 17.4001) 此橘乎等伎自久能可久能木實等名附家良之母

KƏNƏ TATImBANA-wo tǝkiⁿzi-ku n-ǝkaŋg-un-ǝKƏ-NƏ MÏ tǝ NA-ⁿ-DUKƐ-ker-asi-mǝ this mandarin.orange-ACC be.off.season-CONV DV-ATTR smell-ATTR DV-ATTR tree-GEN fruit DV name-LOC-attach(CONV)-RETR-SUP-EXCL [we] should call these mandarin oranges fragrand tree fruits that are off season! (MYS 18.4111) Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

616

Chapter 6

許己見礼婆宇倍之神代由波自米家良思母

kǝkǝ MI-re-mba umbɛ-si KAMÏ-YƏ-yu paⁿzimɛ-ker-asi-mǝ here look-EV-CON be.indeed-FIN deity-age-ABL begin(CONV)-RETR-SUP-EXCL when [you] look at this place, it indeed looks like [they] began [the building of the palaces] from the Age of Gods! (MYS 20.4360)45 Comparative Data Level A: Other Japonic A1: Eastern Old Japanese The suppositional suffix -urasi is also attested in Eastern Old Japanese: 布奈妣等佐和久奈美多都良思母

puna-m-bitǝ sawak-u nami tat-urasi-mǝ boat-GEN-person make.noise-FIN wave rise-SUP-EXCL boatmen are making noise. Waves seem to rise! (MYS 14.3349) 和我都麻波伊多久古非良之

wa-ŋga tuma pa ita-ku kopï-rasi I-POSS spouse TOP extreme-CONV long.for(CONV?)46-SUP It seems that my spouse longs for [me] dearly … (MYS 20.4322) 伊波乃伊毛呂和乎之乃布良之

ipa-nǝ imo-rǝ wa-wo sinǝp-urasi home-GEN beloved-DIM I-ACC yearn-SUP It seems that my beloved at home yearns for me (MYS 20.4427) A2: Ryukyuan There are very few examples of tentative cognates of WOJ -urasi appearing in the Omoro sōshi in the forms -urasi, -urasiya, -urase, and -asiyo (Torigoe 1968: 177–178).47 The absence of reflexes in modern dialects and the limited attestations even in the Omoro sōshi itself may suggest that we are dealing with a loan from Middle Japanese. However, the presence of the form -asiyo may indicate 45  Opinions differ, whether this poem belongs to a border-guard, or to Ōtomo-no Yakamochi himself. The latter seems more probable, since the poem is preceded by the line in Chinese: 陳私拙懐一首 ‘a poem stating my own humble thoughts,’ and also because it is written in perfect Western Old Japanese, without any elements typical for Eastern Old Japanese. Therefore, I treat it as a Western Old Japanese text. 46  The corresponding WOJ form kop-urasi ‘long.for-SUP’ with no converb form preceding -urasi is attested. 47  The form -urase is dubious, though, because it appears as a part of the proper name of a boat, Tamamedurase, also attested as Tamamedura (Hokama 2000.2: 14). Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

617

Verbs

that this is a genuine cognate, because the allomorph -asi was a relic morpheme already in Western Old Japanese, let alone Middle Japanese. Examples: Old Ryukyuan けおわのかしよらしよ

keo wa no ka s-i-yor-asiyo today TOP what IP do-CONV-exist-SUP What would [they] do today? (OS 7.376) たらもいやとくらしや

Tara moi ya tok-urasiya Tarō lord TOP arrive-SUP It looks like lord Tarō will arrive (OS 17.1157) わかまつかとくらし

Wakamatsu-ka tok-urasi Wakamatsu-NOM arrive-SUP It looks like Wakamatsu will arrive (OS 17.1207) 3.2.1.10 Negative Tentative -aⁿzi ~ -ⁿzi The negative tentative suffix has two allomorphs: -aⁿzi that appears after consonant verbs and -ⁿzi that is found after vowel verbs. Surprisingly enough, it combines only with the preceding passive -(a)ye- or progressive -er-. There is no reliable evidence for the combination -er-aⁿzi, as it is found only in logographic spelling and only with the verb ik- ‘to live,’ see the example from MYS 12.2905 below. chart 46 Combinations of the negative tentative suffix –aⁿzi ~ -ⁿzi with other suffixes and bound auxiliaries

suffixes and auxiliaries

combination forms

passive -(a)ye-, -rayeprogressive -er-

-(a)ye-ⁿzi -er-aⁿzi

It is likely that historically -(a)ⁿzi represents a contraction of the negative suffix -(a)n- and the suppositional -asi that was discussed above in section 3.2.1.9. The likelihood of such a development is strengthened by three simple observations: first, -(a)ⁿzi being a negative tentative suffix is expected to include a negative morpheme. Second, although a contraction of *naC to NC is Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

618

Chapter 6

not frequent, there are cases that demonstrate its existence, for example, kaⁿdo ‘gate’ < kana-to ‘metal door.’48 Third, -(a)ⁿzi being a negative tentative suffix is expected to include a modality morpheme that has a meaning compatible with a tentative. Certainly the suppositional -asi fits the bill here both phonetically and functionally. The negative tentative suffix has three functions in Western Old Japanese: negative presumption, negative intention, and mild prohibition. Examples: (1) negative presumption: 倶伊播阿羅珥茹

kuyi pa ar-aⁿzi ⁿzǝ regret(NML) TOP exist-NEG/TENT FP [You] would have no regret (NK 124) 企許斯遠周久爾能麻保良叙可爾迦久爾保志伎麻爾麻爾斯可爾波阿羅慈迦

kikǝs-i-wos-u kuni-nǝ ma-po-ra ⁿzǝ ka n-i ka-ku n-i posi-ki manima n-i sika n-i pa ar-aⁿzi ka rule(HON)-CONV-HON-ATTR country-GEN INT-top-LOC FP thus DV-CONV thus-CONV DV-CONV desire-ATTR according DV-CONV thus DV-CONV TOP exist-NEG/TENT IP in the highest place of the country, where [the emperor] rules, [it] would not be thus according to what [you] wish to be this way and that way, [would it]? (MYS 5.800) 安礼乎於伎弖人者安良自等富己呂倍騰

are-wo ok-i-te PITƏ pa ar-aⁿzi tǝ pokǝr-ǝp-ɛ-ⁿdǝ I-ACC leave-CONV-SUB person TOP exist-NEG/TENT DV boast-ITER-EV-CONC although [I] repeatedly boast that there are probably no other persons besides me (MYS 5.892) 幾不生有命乎

IKU-mBAKU MƏ IK-ER-AⁿZI INƏTI-wo how.much-extend FP live-PROG-NEG/TENT life-ACC [I] probably will not live much longer (MYS 12.2905)

48  Not literally a ‘door made of metal,’ but originally a kind of door that had metal parts in it.

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

619

Verbs 和礼乎於吉弖比等波安良自

ware-wo ok-i-te pitǝ pa ar-aⁿzi we-ACC leave-CONV-SUB person TOP exist-NEG/TENT except us, there would be no [other] men (MYS 18.4094) 伊毛我多可々々尓麻都良牟許己呂之可尓波安良司可

imo-ŋga taka taka n-i mat-uram-u kǝkǝrǝ sika n-i pa ar-aⁿzi ka beloved-POSS high high DV-CONV wait-TENT2-ATTR heart thus DV-CONV TOP exist-NEG/TENT IP the heart of [my] beloved who probably waits for [me] eagerly would not [it] be this way? (MYS 18.4107) 和我加度須疑自

wa-ŋga kaⁿdo suŋgï-ⁿzi I-POSS gate pass-NEG/TENT [the cuckoo] probably would not pass my gate (MYS 20.4463) 如是醜事者聞曳自

KA-KU N-Ə SIKƏ KƏTƏ pa KIK-Əye-ⁿzi thus-CONV DV-ATTR disgraceful thing TOP hear-PASS-NEG/TENT [I] would not be able to hear about such disgraceful things (SM 17) (2) negative intention: 那迦士登波那波伊布登母

nak-aⁿzi tǝ pa na pa ip-u tǝmǝ weep-NEG/TENT DV TOP you TOP say-FIN CONJ Even though you say that [you] would not weep (KK 4) 和賀韋泥斯伊毛波和須禮士

wa-ŋga wi ne-si imo pa wasure-ⁿzi I-POSS bring(CONV) sleep(CONV)-PAST/ATTR beloved TOP forget-NEG/TENT [I] would not forget [my] beloved, whom I brought with [me] and slept with (KK 8) Cf. the variant of the same poem below in NK 5 that has the consonant verb wasur-.

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

620

Chapter 6

和禮波和須禮士

ware pa wasure-ⁿzi I TOP forget-NEG/TENT I would not forget (KK 12) 麻都爾波麻多士

mat-u-ni pa mat-aⁿzi wait-ATTR-LOC TOP wait-NEG/TENT as [I] wait, [I] would not wait [any longer] (KK 88) 和我謂祢志伊茂播和素邏珥

wa-ŋga wi ne-si imo pa wasur-aⁿzi I-POSS bring(CONV) sleep(CONV)-PAST/ATTR beloved TOP forget-NEG/TENT [I] would not forget [my] beloved, whom I brought with [me] and slept with (NK 5) Cf. the variant of the same poem above in KK 8 that has the vowel verb wasure-. 夜都代爾母安礼波和須礼自許乃多知婆奈乎

ya-tu YƏ-ni mǝ are pa wasure-ⁿzi kǝnǝ tatimbana-wo eight-CL generation-LOC FP I TOP forget-NEG/TENT this mandarin. orange-ACC I would not forget these mandarin orange [flowers] even in eight (many?) generations (MYS 18.4058) 可敝里見波勢自

kaper-i-mi pa se-ⁿzi return-CONV-look(NML) TOP do-NEG/TENT [we] will not look back (MYS 18.4094) (3) mild prohibition: 比等爾波美要緇

pitǝ-ni pa mi-ye-ⁿzi person-DAT TOP see-PASS-NEG/TENT [I] should not be seen by people (MYS 15.3708) 宇梅乃花伊都波乎良自

uMƐ-no PANA itu pa wor-aⁿzi plum-GEN flower when TOP break-NEG/TENT when [one] should not pick plum blossoms? (MYS 17.3904) Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

621

Verbs 天地乃福毛不蒙自

AMƐ TUTI-nǝ SAKIPAP-I mo KA ŋGARAP-Aⁿzi heaven earth-HEN flourish-NML FP receive(HUM)-NEG/TENT [they] should not receive blessing from Heaven and Earth (SM 45) Comparative Data Level A: Other Japonic A1: Eastern Old Japanese In Eastern Old Japanese only the allomorph -aⁿzi is attested. It has the same functions as in Western Old Japanese: (1) negative presumption: 於曾波夜母奈乎許曾麻多賣牟可都乎能四比乃故夜提能安比波多我波自

osǝ paya mǝ na-wo kǝsǝ mat-am-e muka-tu wo-nǝ sipi-nǝ ko-yaⁿde-nǝ ap-i pa taŋgap-aⁿzi slow fast FP you-ACC FP wait-TENT-EV opposite.side-GEN/LOC peak-GEN chinquapin-GEN DIM-branch-GEN meet-NML TOP differ-NEG/TENT Whether [you come] quickly or slowly, [I] will wait for you. [It] is probably not different from the meeting of small branches of chinquapin trees at the peak on the opposite side (MYS 14.3493a) (2) negative intention: 伊刀尓奈流等毛和波等可自等余

ito n-i nar-u tǝmo wa pa tǝk-aⁿzi tǝ yǝ thread DV-CONV become-FIN CONJ I TOP untie-NEG/TENT DV EP [I] think that I would not untie [the cords of my garment] even if [they] become [thin] threads! (MYS 20.4405) (3) mild prohibition: 可麻久良乃美胡之能佐吉能伊波久叡乃伎美我久由倍伎己許呂波母多自

Kamakura-nǝ Miŋgosi-nǝ saki-nǝ ipa-kuye-nǝ kimi-ŋga kuy-umbɛ-ki kǝkǝrǝ pa mǝt-aⁿzi Kamakura-GEN Miŋgosi-GEN cape-GEN rock-slide-COMP lord-POSS regretDEB-ATTR heart TOP hold-NEG/TENT [My] lord should not have the heart, like the rockslide at the Miŋgosi cape in Kamakura, which [he] will have to regret [later] (MYS 14.3365)

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

622

Chapter 6

A2: Ryukyuan The negative tentative -(a)ⁿzi is attested only in Classical Ryukyuan and exclusively in the Ryūka (Hokama 1995: 315). The absence of its reflexes in both Old Ryukyuan and the modern dialects strongly implies that it is a relatively late loan from Middle Japanese. 3.2.1.11 Negative Potential -umasiⁿzi The negative potential suffix has only one allomorph –umasiⁿzi found after both consonant and vowel verbs alike. Although a four-syllable long verbal suffix must have a secondary origin, its internal etymology is unclear. As a rule -umasiⁿzi appears in Western Old Japanese as a sentence-final suffix, but like -umazi in Middle Japanese, it can also be followed by inflectional adjectival suffixes. However, in contrast to Middle Japanese where -umazi developed a full adjectival paradigm (Vovin 2003: 292), the only two adjectival inflectional forms that are found in Western Old Japanese after -umasiⁿzi are the attributive -ki and the adjectival gerund -mi. The latter is attested only once in SM 58. chart 47 Combinations of the negative potential suffix -umasiⁿzi with other suffixes and bound auxiliaries

suffixes and auxiliaries

combination forms

passive -(a)ye-,-rayepotential -kateattributive -ki gerund -mi

-(a)y-umasiⁿzi -kat-umasiⁿzi -umasiⁿzi-ki -umasiⁿzi-mi

The combination of negative potential -umasiⁿzi with the passive -(a)ye- is strange, since the passive can have a potential meaning by itself, and the main function of the negative potential -umasiⁿzi is negative potential. The same observation can be made about its combination with the bound potential auxiliary -kate-. But it is possible that in both cases here we have a similar semantic reinforcement that is observed in Middle Japanese when the negative potential -umazi combines with the potential preverb ye- in the same verbal form (Vovin 2003: 294). The negative potential suffix -umasiⁿzi is not attested at all in a number of the Western Old Japanese texts, such as the Kojiki kayō, the Jōgu teisetsu, the Bussoku seki-no uta, the Norito, the Fudoki kayō, and the Shoku nihongi kayō.

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

623

Verbs

Overall, it is a rare form, as it appears only twice in the Nihonsoki kayō, only fifteen times in the whole Man’yōshū (not all of the examples are spelled phonographically), and only four times in the Senmyō. In sharp contrast to Middle Japanese, where -umazi can have the functions of a negative debitive, a negative probability, and a negative potential, the WOJ suffix -umasiⁿzi is predominantly attested in the function of a negative potential. The function of the negative debitive is supported only by a single example in Senmyō 27, only partially written phonographically, and the function of the negative probability rests on two uncertain examples from the Man’yōshū (attested only in the logographic script in MYS 6.1053 and MYS 7.1385). Examples: (1) negative potential: 于羅遇破能紀豫屢麻志士枳箇破能区莽遇莽

ura-ŋ-gupa n-ǝ kï yǝr-umasiⁿzi-ki kapa-nǝ kuma-ŋguma back-DV-ATTR-mulberry DV-ATTR tree approach-NEG/POT-ATTR river-GEN bend-bend the bends of the river that the mulberry tree in the back cannot approach (NK 56) 耶麻古曳底于瀰倭柂留騰母於母之楼枳伊麻紀能禹知播倭須羅由麻旨珥

yama koye-te umi watar-u tǝmo omosiro-ki ima kï-nǝ uti pa wasur-ay-umasiⁿzi mountain cross(CONV)-SUB sea cross-FIN CONJ beautiful-ATTR Ima fortressGEN inside TOP forget-PASS-NEG/POT Even if [I] pass over the mountains and cross the seas, [I] cannot forget the inside of the beautiful Ima fortress (NK 119) 佐不寐者遂尓有勝麻之自

sa-NE-ⁿZ-U pa TUPI n-i AR-I-kat-umasiⁿzi PREF-sleep-NEG-CONV TOP final DV-CONV exist-CONV-POT-NEG/POT if [I] do not sleep [with you], [I] cannot live (MYS 2.94) 浮蓴邊毛奥毛依勝益士

UK-I NUNAPA PE-NI mo OKI-NI mo YƏR-I-kat-umasiⁿzi float-CONV water.shield shore-LOC FP offing-LOC FP approach-CONV-POT-NEG/ POT [my love for you is like] a water shield that cannot approach either a shore or an offing (MYS 7.1352)

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

624

Chapter 6

等保伎佐刀麻弖於久利家流伎美我許己呂波和須良由麻之自

tǝpo-ki sato-maⁿde okur-i-ker-u kimi-ŋga kǝkǝrǝ pa wasur-ay-umasiⁿzi distant-ATTR village-TERM see.off-CONV-RETR-ATTR lord-POSS heart TOP forget-PASS-NEG/POT [I] cannot forget the kindness of [my] lord who saw [me] off to [my] distant village (MYS 20.4482) 多能遍重天勅止毛敢末之時止為弖

AMATA n-ǝ TAmBI KASANE-te NƏTAMAP-Ɛ-ⁿdǝmo AP-Umasiⁿzi tǝ S-I-te many DV-CONV time pile.up(CONV)-SUB tell(HON)-EV-CONC be.ready.to.doNEG/POT DV do-CONV-SUB although [I] told him many times [to accept the job], [he] believed that [he] cannot be ready to do [it] (SM 26) 王等波己我得麻之字岐帝乃尊岐寶位乎望

OPO KIMI-TATI PA ONƏ- ŋGA ∅-Umasiⁿzi-ki MIKAⁿDO-nǝ TAPUTO-ki KURAWIwo NƏⁿZƏM-I great prince-PLUR TOP self-POSS get-NEG/POT-ATTR sovereign-GEN awesome-ATTR position-ACC desire-CONV the great princes desire the awesome position of the sovereign that [they] themselves cannot get, and … (SM 45) The root of the verb ɛ- ‘to get’ elides before the -umasiⁿzi. 汝乃志乎婆椈久乃間毛忘得末之自美奈毛

MIMASI-nǝ KƏKƏRƏⁿZASI-womba simasi-ku n-ǝ MA mo WASUR-Umasiⁿzi-mi namo you-GEN memorial.service-ACC(EMPH) be.a.little.while-CONV DV-ATTR interval FP forget-NEG/POT-GER FP because [I] cannot forget about memorial services for you even for a little while (SM 58) (2) negative debitive: 不言岐辞母言奴

IP-UMASIⁿZI-ki KƏTƏ mǝ IP-I-n-u say-NEG/POT-ATTR word FP say-CONV-PERF-FIN [he] also said words that [he] should not have said (SM 27)

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

625

Verbs

(3) negative probability: 百代尓母不可易大宮處

MOMO YƏ-ni mǝ KAPAR-UMASIⁿZI-KI OPO MIYA-ⁿ-DƏKƏRƏ hundred generation-LOC FP change-NEG/POT-ATTR great palace-GEN-place the place of the great palace that would not change even in [one] hundred generations (MYS 6.1053) 埋木之不可顕事尓不有君

UMORE-ŋ-GÏ-NƏ ARAPAR-UMASIⁿZI KƏTƏ n-i AR-AN-Aku n-i bury(NML)-DV(ATTR)-tree-GEN appear-NEG/POT matter DV-CONV existNEG-NML DV-CONV as it is not the case that the buried trees would not appear (MYS 7.1385) Comparative Data Level A: Other Japonic A1: Eastern Old Japanese There is only one example of the negative potential -umasiⁿzi in the Eastern Old Japanese corpus, but it occurs in a poem without any distinctive Eastern Old Japanese features. Thus, this poem should probably be treated as a Western Old Japanese text, and we should conclude that in all likelihood the negative potential –umasiⁿzi was not present in Eastern Old Japanese. Even if we are to take the following example at face value as an Eastern Old Japanese text, we should be aware of the fact that it represents the negative potential function like all the rest of the other reliable Western Old Japanese examples. 奈乎多弖天由吉可都麻思自

na-wo tate-te yuk-i-kat-umasiⁿzi you-ACC make.stand(CONV)-SUB go-CONV-POT-NEG/POT making you stand [and wait], [I] could not go away (MYS 14.3353) A2: Ryukyuan Old Ryukyuan has the negative debitive -umazi (Hokama 1995: 606), but its Middle Japanese-like form and absence of reflexes in modern dialects betray a loan from Middle Japanese. Thus, it is likely that -umasiⁿzi represents an Old Japanese innovation, and may even just be a Central Japanese one, given the dubious nature of EOJ -umasiⁿzi. An example from Old Ryukyuan:

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

626

Chapter 6

ともゝすへせいいくさよせるまじ

to momo sufe sei kusa yose-r-umazi ten hundred year force army approach-ATTR-NEG/POT for a thousand years an [enemy] army would not be able to approach (OS 13.763) 3.2.1.12 Exclamative -umǝ ~ -mǝ The Japanese linguistic tradition treats this suffix as a combination of the final predication form -u plus a final particle (終助詞 shūjoshi) mǝ. At first glance such an analysis may have a good logical ground, because the final particle mǝ also follows the adjectival final form in -si. There are, however, some problems with the traditional analysis. First, the final particle mǝ suspiciously looks like the homophonous focus particle mǝ, or emphatic particle mǝ and it seems that the main reason for maintaining a special ‘final’ particle is that it is found at the end of sentences. It is of course impossible to claim that this ‘final’ mǝ is a focus particle, because it is found only after noun phrases including nominalized verbs. It is also unlikely that we deal here with the emphatic particle mǝ, because it does not occur after the final form of verbs, and generally it is used as a sentence final only after the interrogative particle ya. Second, no other particles except the interrogative ya is ever found between the final predication -u and the ‘final’ particle mǝ, while certain particles, such as the emphatic particles ya and yǝ are invariably found after the ‘final’ mǝ, which in these cases ceases to be a ‘final’ particle in a strict sense. Third, since the adjectival final predication marker -si has not yet been completely established as a final predication marker and the whole adjectival inflectional system was not yet fully embedded in the language, as outlined in Chapter 5, it should come as no wonder that -(u)mǝ can be added to -si. Finally, and most importantly, -umǝ can function in Western Old Japanese as an independent suffix, which might have lost a synchronic connection with the final -u (< *-um). Thus, besides the expected negative exclamative form in -(a) ⁿz-umǝ, a negative exclamative form -(a)n-umǝ is also attested (see MYS 15.3684 below). Certainly, -(a)n-u cannot be a final negative form here, because it is an attributive form. Therefore, the analysis of -(a)n-umǝ as -(a)n-u mǝ ‘NEG-FIN FP’ turns out to be impossible. These facts make me suspect that we in fact deal here with a special sentence-final verbal exclamative suffix -umǝ, and not with a combination of the final predication suffix -u with the ‘final’ particle mǝ. Historically, of course, this suffix should be bi-morphemic, since the vowels /u/ and /ǝ/ cannot combine within the same morpheme. Remember now, that final predication suffix *-u goes back to PJ *-um (see section 3.2.1.1). Therefore, diachronically I divide

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

627

Verbs

this suffix as *-um-ǝ, where the remaining *-ǝ part represents some kind of an exclamative marker. However, there is no synchronic evidence for such analysis in Western Old Japanese; therefore I treat -umǝ here as a single exclamative suffix. The exclamative -(u)mǝ combines with the following preceding suffixes and auxiliaries: chart 48 Combinations of the exclamative -umǝ with preceding suffixes and bound auxiliaries

suffixes and auxiliaries

combination forms

negative -(a)ⁿz-, -(a)ntentative -(a)mdesiderative -(a)na honorific -ascausative -(a)sesuppositional -urasi adjectivizer -asi adjectival -si perfective -te-

-(a)ⁿz-umǝ, -(a)n-umǝ -(a)m-umǝ -(a)na-mǝ -as-umǝ -(a)s-umǝ -urasi-mǝ -asi-mǝ -si-mǝ -t-umǝ

On the basis of Chart 48 above one can see that the exclamative suffix has two allomorphs: -umǝ and -mǝ. The first occurs after the perfective auxiliary -teand most of the verbal suffixes with the major exception of the desiderative -(a)na. The second allomorph -mǝ appears after suffixes that have an adjectival paradigm and after the desiderative -(a)na. The exclamative -umǝ has only one function: that of exclamation. Examples: 和岐弊能迦多用久毛韋多知久母

wa-ŋg[a]-ipe-nǝ kata-yo kumowi tat-i-k-umǝ I-POSS-home-GEN side-ABL cloud rise-CONV-come-EXCL Clouds rise from the side of my home! (KK 34) 久良波斯夜麻袁佐賀志美登伊波迦伎加泥弖和賀弖登良須母

Kurapasi-yama-wo saŋgasi-mi tǝ ipa kak-i-kane-te wa-ŋga te tǝr-as-umǝ Kurapasi-mountain-ABS steep-GER DV rock hang-CONV-NEG/POT(CONV)SUB take-HON-EXCL [I] think that Mound Kurapasi is steep. Being unable to cling to the rocks, [I wish you would] take my hand! (KK 69)

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

628

Chapter 6

奴弖由良久母淤岐米久良斯母

nute yurak-umǝ okimɛ k-urasi-mǝ bell sound-EXCL Okimɛ come-SUP-EXCL The bells are sounding! It seems that Okimɛ is coming! (KK 111) 瀰既能佐烏麼志魔幣莵耆瀰伊和哆羅秀暮

mi-kɛ-nǝ sawo-m-basi mapetukimi i-watar-as-umo HON-tree-GEN pole-GEN-bridge minister DLF-cross-HON-EXCL [I wish] that the ministers [would] cross the bridge here [made] of poles of sacred trees! (NK 24) 柂我佐基泥佐基泥曾母野倭我底騰羅須謀野

ta-ŋga sakï-ⁿ-de sakï-ⁿ-de sǝ mǝ ya wa-ŋga te tǝr-as-umo ya who-POSS chap(NML)-DV(ATTR)-hand chap(NML)-DV(ATTR)-hand FP FP IP I-POSS hand take-HON-EXCL IP whose chapped hand, chapped hand will take my hand?! (NK 108) 雲谷裳情有南畝可苦佐布倍思哉

KUMO ⁿdani mo KƏKƏRƏ AR-Ana-mo kakus-ap-umbɛ-si YA cloud RP FP heart exist-DES-EXCL hide-ITER-DEB-FIN IP [I] wish at least the clouds [would] have feelings! Do [they] have to hide [Mt. Miwa] all the time? (MYS 1.18) 都摩夜佐夫斯久於母保由倍斯母

tuma-ya sambusi-ku omǝp-oy-umbɛ-si-mǝ spouse-room lonely-CONV think-PASS-DEB-FIN-EXCL [I] must suddenly think [how] lonely [our] bedroom [is]! (MYS 5.795) 母智騰利乃可可良波志母与

mǝti-ⁿ-dǝri-nǝ kakar-ap-asi-mǝ yǝ mochi-GEN-bird-COMP be.stuck-ITER-ADJ-EXCL EP [You] are stuck like a bird on a mochi [trap-stick]! (MYS 5.800) 等伎波奈周迦久斯母何母等意母閇騰母余能許等奈礼婆等登尾可祢都母

tǝk[ǝ]-ipa-nasu ka-ku si mǝŋgamǝ tǝ omǝp-ɛ-ⁿdǝmǝ yǝ-nǝ kǝtǝ nar-e-mba tǝⁿdǝmï-kane-t-umǝ eternal-rock-COMP thus-CONV EP DP DV think-EV-CONC world-GEN matter be-EV-CON stop(CONV)-NEG/POT(CONV)-PERF-EXCL Although [I] think that [I] want to be (thus) like the eternal rock, because [it] is a matter of this world, [I] cannot stop [life]! (MYS 5.805)

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

629

Verbs 烏梅乃波奈知良麻久怨之美和我曽乃々多氣乃波也之尓于具比須奈久母

uMƐ-nǝ pana tir-am-aku wosi-mi wa-ŋga sǝnǝ-nǝ takɛ-nǝ payasi-ni uŋgupisu nak-umǝ plum-GEN blossom fall-TENT-NML be.regretful-GER I-POSS garden-GEN bamboo-GEN grove-LOC bush.warbler sing-EXCL Because the bush warbler regrets that the plum blossoms will fall, [he] sings in the bamboo grove of my garden! (MYS 5.824) 古良何伊弊遅斯良受毛

ko-ra-ŋga ipe-ⁿ-di sir-aⁿz-umǝ girl-PLUR-POSS house-GEN-way know-NEG-EXCL [I] do not know the way to girls’ houses! (MYS 5.856) 山辺爾草乎思香奈久毛

YAMA-PE-ni sa-wo-sika nak-umo mountain-side-LOC PREF-male-deer cry-EXCL a male deer cries at the mountain side! (MYS 15.3674) 奈曽許己波伊能祢良要奴毛

naⁿzǝ kǝkǝmba i-nǝ ne-raye-n-umo why so sleep-GEN sleep-PASS-NEG-EXCL Why cannot [I] sleep at all?! (MYS 15.3684) 念意緒多礼賀思良牟母

OMƏP-U KƏKƏRƏ-wo tare ka sir-am-umǝ love-ATTR heart-ACC who IP know-TENT-EXCL who will know [my] loving heart?! (MYS 17.3950) 多知夜麻乃由吉之久良之毛波比都奇能可波能和多理瀬安夫美都加須毛

Tati yama-nǝ yuki si k-urasi-mǝ Papitukï-nǝ kapa-nǝ watar-i-ⁿ-SE ambumi tuk-as-umǝ Tati mountain-GEN snow EP melt-SUP-EXCL Papitukï-GEN river-GEN crossNML-DV(ATTR)-rapids stirrup soak-CAUS-EXCL It seems that the snow on Mound Tati has melted! [I] made my stirrups soak at the rapids that are the crossing of the river Papitukï! (MYS 17.4024)

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

630

Chapter 6

Comparative Data Level A: Other Japonic A1: Eastern Old Japanese The exclamative -(u)mǝ is also amply attested in Eastern Old Japanese: 和乎布利弥由母阿是古志麻波母

wa-wo pur-i-mi-y-umǝ Aⁿze ko si map-am-ǝ I-ACC swing-CONV-look-PASS-EXCL Aⁿze girl EP dance-TENT-ATTR the girl from Aⁿze is going to dance, suddenly looking back at me! (FK 7) 可奈師家兒良爾伊夜射可里久母

kanasi-ke KO-ra-ni iya-ⁿ-zakar-i k-umǝ beloved-ATTR girl-DIM-DAT more.and.more-DV(CONV)-become distant-CONV come-EXCL Oh, [I] came [here] growing more and more distant from [my] beloved girl! (MYS 14.3412) 夜麻邊能之牙可久尓伊毛呂乎多弖天左祢度波良布母

yama-m-BE-nǝ siŋge-k-aku n-i imo-rǝ-wo tate-te sa-ne-ⁿ-do parap-umǝ mountain-GEN-side-GEN thick-ATTR-NML DV-CONV beloved-DIM-ACC make stand(CONV)-SUB PREF-sleep(NML)-DV(ATTR)-place clean-EXCL as the mountain (side) is overgrown [with bush], [I] let my beloved stand, and [I] am clearing a place to sleep [for us]! (MYS 14.3489) 奈流世呂爾木都能余須奈須伊等能伎提可奈思家世呂爾比等佐敞余須母

nar-u se-rǝ-ni kǝtu-nǝ yǝs-u-nasu itǝ nǝkite kanasi-ke se-rǝ-ni pitǝ sape yǝs-umǝ sound-ATTR rapid-DIM-LOC debris-GEN approach-ATTR-COMP very specially beloved-ATTR beloved-DIM-DAT person RP relate-EXCL even [other] people make [their thoughts] approach [my] very special beloved, like the debris approaches each other in the singing rapids! (MYS 14.3548) We can probably reconstruct PJN *-um-ǝ, consisting of the final predication suffix *-um and the exclamative marker *-ǝ. 3.2.2 Sentence-Non-Final Verbal Suffixes Sentence-non-final verbal suffixes are the suffixes that occur at the end of a verbal form that itself is used as a non-final predicate, a verbal noun, or a

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

631

Verbs

non-final form of a verbal compound. There are two kinds of sentence-nonfinal verbal suffixes: converbs and nominalizers. 3.2.2.1 Converbs I replace Samuel E. Martin’s term ‘infinitive’ with ‘converb,’ because in spite of the wide popularity of the former in the field it is overall misleading: there are no infinitives in the Japonic languages in the same sense as in Indo-European. The converb is a special verbal form that is used in essentially two functions. In isolation it has the function of a non-final predicate, showing that the final predicate is yet to come, further on in the sentence. Some converbs are also used to build verbal compounds, including those that consist of a main verb and a following honorific verb. Every non-final component of a verbal compound must take a converb form:  … V-CONV-V-FIN. 3.2.2.1.1 Converb -i The converb has two allomorphs: an allomorph -i is used after consonant and irregular verbs, and an allomorph -∅ after vowel verbs (including strong vowel), with -i being suppressed after the final vowel of the stem, e.g.: *kopï-i > kopï-∅ ‘longs for and.’ In word-to-word translation I place such suppressed converbs in parentheses: (CONV). However, the converb -i itself suppresses the final vowel of the stem in irregular verbs, resulting in *kǝ-i > k-i ‘comes and’ and *se-i > s-i ‘does and.’ In isolation it has the function of a non-final predicate. The converb -i is also used to form verbal compounds, including those that consist of a main verb and an auxiliary. If several auxiliaries are used, the converb almost always appears between them, functioning as a kind of glue that keeps a verbal form together. Thus, the converb -i marks the non-final members of a verbal compound. Sometimes a non-final predicate in the converb form may be used as an adverbial modifier of a following predicate. The converb suffix -i can combine with various suffixes and bound auxiliaries. chart 49 Combinations of the converb -i with other suffixes and bound auxiliaries

suffixes and auxiliaries

combination forms

negative -(a)niterative -aphonorific -ascausative -(a)simɛ-

-(a)n-i -ap-i -as-i -(a)simɛ-∅-

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

632

Chapter 6

chart 49 Combinations of the converb -i with other suffixes (cont.)

suffixes and auxiliaries

combination forms

passive -(a)ye-, -rayesubordinative converb -te coordinative converb -tutu coordinative converb -naŋgara coordinative converb -katera past -ki, -si, -sika perfective -nperfective -teprogressive -erperfective-progressive -tarretrospective -ker-

-(a)ye-∅ -i-te -i-tutu -i-naŋgara -i-katera -i-ki, -i-si, -i-sika -n-i-, -i-n-te-∅-, -i-te-er-i-tar-i-, -i-tar-i-ker-

Note: A dash after combination forms indicates that a given combination form appears only as a word-non-final form. The lack of dash indicates that a combination form may be used as a word-final form.

As one can see from the chart above, the converb -i always follows suffixes, but it can either precede or follow bound auxiliaries. Some of the bound auxiliaries can be both preceded and followed by the converb -i. The interesting gap in distribution that we can observe on the basis of the above chart is that -i does not combine at all with mood markers. This gap means that non-final predicates with mood markers have a limited distribution: they cannot appear in coordinate clauses, and as the reader will see below, they can only occur as non-final predicates containing conditional and concessive converbs. Thus, modality markers are found only in the non-final predicates in subordinate clauses. There are also further limitations. As can be seen from the chart above, although the aspect markers unlike the mood markers can be followed by the converb -i, it is always used after them as a link to the following auxiliary, and never as a non-final predicate. Examples:

Alexander Vovin - 978-90-04-42281-0 Downloaded from Brill.com11/09/2020 11:16:03AM via free access

633

Verbs

(1) non-final predicate: 伊勢能宇美能意斐志爾波比母登富呂布志多陀美能伊波比母登富理宇知弖 志夜麻牟

Ise-nǝ umi-nǝ opï-[i]si-ni pap-i-mǝtǝpǝr-ǝp-u sitaⁿdami49-nǝ i-pap-i-mǝtǝpǝr-i ut-i-te si yam-am-u Ise-GEN sea-GEN grow(CONV)-stone-LOC crawl-CONV-go.around-ITER-ATTR seashell-COMP DLF-crawl-CONV-go.around-CONV hit-CONV-SUB EP stop-TENT-FIN like the shellfish that are constantly crawling around on the growing rocks of the Ise sea, [we] will crawl around [them] there, smite and stop [them] (KK 13) 伊那佐能夜麻能許能麻用母伊由岐麻毛良比多多加閇婆

Inasa-nǝ yama-nǝ kǝ-nǝ ma-yo mǝ i-yuk-i mamor-ap-i tatakap-ɛ-mba Inasa-GEN mountain-GEN tree-GEN interval-ABL FP DLF-go-CONV watchITER-CONV fight-EV-CON when [we] fought going there from between the trees of the mountain Inasa and watching constantly [out for enemies] (KK 14) 斯理都斗用伊由岐多賀比麻弊都斗用伊由岐多賀比宇迦迦波久

siri-tu to-yo i-yuk-i-taŋgap-i mape-tu to-yo i-yuk-i-taŋgap-i ukakap-aku back-GEN/LOC door-ABL DLF-go-CONV-differ-CONV front-GEN/LOC doorABL DLF-go-CONV-differ-CONV look(HUM)-NML going there from the front door, and going there from the back door, and looking [at you] (KK 22) 多古牟良爾阿牟加岐都岐曾能阿牟袁阿岐豆波夜具比加