The pronunciation of Old Japanese follows a very simple scheme. There are the five vowels, a, e, i, o, u, and no diphthongs, and the following consonants and digraphs b, ch, d, dz, f, g, h, j, k, m, n, r, s, sh, t, ts, w, y, z, zh. The syllables are all open, and consist of

Sixty-four open syllables in all, out of 105 possible ones, of which and their various agglutinations the whole language is made up.

The vowels are pronounced as in Italian, a being the accented a, e and o the open sounds. The vowel u is English u in put, never as in rut or lute. The vowels a, e, o are half-long, i and u are shorter, and u shortest of all. Value is given to each syllable, subject as above, with scarcely any ictus (as in French), but the last syllable of a word, especially in u, is always weakest, and the penultimate rather the strongest.

The consonants are pronounced as in English, h well aspirated, but rather forward, even between vowels, z as in zany, zh as the s in pleasure, f may have descended from an original p, with h and w.

The scheme, according to that of the Oxford Dictionary, would be—

The Chinese and Japanese languages differ from Aryan and Semitic forms of speech in the total absence of all concords dependent upon number, person, case, and gender, in the like default of relative words, and (from Aryan speech) in the absence of narratio obliqua. They further differ in the almost complete absence of any morphological moods or tenses. Chinese has, broadly speaking, no accidence at all; the grammar is a syntax teaching the right order and use of vocables (used as words), double vocables (constituting words), and the few form-words (empty words the Chinese call them) which serve as variously connective particles. In Japanese there is a very scanty accidence of the few adjectives the language possesses, and a more complete one of the verb; but the verbal forms are all (with an exception or two) resolvable into locutions, more or less agglutinated, scarcely constituting true inflexions. Hence, morphologically, even in Japanese there is neither mood nor tense, but there is an approach to both of the highest value to the language, giving it, in conjunction with a goodly number of particles, an articulation and plasticity wanting to Chinese. From the above considerations it will be readily understood that the power of expression even in Japanese is far inferior to that of Western speech. The imagery is, of necessity, extremely limited in range and flat in tone. It seems to me, however, of greater range and higher quality than in Chinese. But neither Chinese nor Japanese possesses a tithe of the capacity of Aryan and Semitic languages to express human thought and feeling, and describe the works of man or the appearances of nature.

? d??? a????, &c., thought and words, are absolutely unintelligible to the whole Far East.

Nevertheless the poetry of the two great Far Eastern languages has its charm, especially the early poetry of Japan, but that charm, depending as it does largely upon suggestiveness rather than definite statement, and upon characteristic form and decoration rather than content, can only be felt by those who are able to read the texts. The Japanese texts, apart from the labour of decipherment, in themselves present no difficulty, once their simple grammar and construction understood, and in their romanized dress, with the brief grammar that is now subjoined, and the other aids offered in the present volume and its accompanying volume of translations are, it is hoped, made accessible to the English reader who cannot give several years to the acquisition of the complicated scripts which Dai Nippon has taken over, by necessity rather than choice, from the Middle Kingdom.

The following sketch of the grammar of Old Japanese is intended merely to elucidate the texts, principally those of the ManyÔshiu and the Taketori.[1] By Old Japanese is meant the unsinicized language of the Kojiki and Nihongi (as read japonicÉ) and particularly of the uta quoted in those works, of the norito or rituals, and of the texts above mentioned. There are no texts illustrating earlier stages of the language, and all etymologies are doubtful, both on that account and because the elements of Old Japanese are mere agglutinations (more or less contracted) of only sixty-four open syllables.

Japanese (by which expression Old Japanese is here always intended) has no affinity with Chinese, a language consisting of disconnected elements (simple or compound), for in it agglutination has to some extent contracted into inflexion or quasi-inflexion, and it possesses in addition a large number of particles which give it a plasticity not found in Chinese. In my opinion, had the development of Japanese not been arrested by Chinese influences about the middle of the first millennium, it might have won a far higher place than it occupies in the hierarchy of human speech.

This agglutinative inflexion is practically confined to the verb, and—to a slight extent—to nouns adjective, simple or verbalized. What may be called the stem of the verb shows a pure vocalic inflexion in a, e, i, o or u. The stem in u is the form usually found in native and foreign dictionaries—in LemarÉchal’s it is the stem in e or i. To the stem the various locutions are suffixed to construct the forms representing, logically, moods and tenses as well as negative, causative, and potential-passive ‘voices’.

The stems in e, i and o are used as imperatives; koge, row!, mi, look!, ko come!, often with addition of the vocative particle yo (or ro)—miyo, koyo. They are also used as ‘indefinite’ (Chamberlain) forms, mood and tense being determined by that of the principal verb in the sentence. Before proceeding further it is well to state that in the regular Japanese sentence the principal verb is always at the end, the object (if expressed) in the middle, and the subject (if expressed) at the beginning of the sentence, while words of qualification precede the words qualified. In poetry however, especially in the ManyÔshiu, inversions are common.

All syllables in Japanese are open, and the terminal syllables of the stems are the consonants b, f, g, k, m, n, r, s, the digraph ts, w, y, and z followed by the vowels a, e, i, o, u (also u without consonant as in suu, e stem suwe, to place). Before i ts becomes ch, and before e and a t, before i s becomes sh and z zh; before every vowel except u f becomes h. All verbs have the stem in u, most verbs have also the stem in a and i, some have also the stem in e only, some in i only. There are a very few irregular verbs. The following list of verbs and their stems will illustrate the above remarks:—

The Japanese verb has neither number nor person, the forms that follow therefore are indefinite as to number and person. The locutions—or more strictly the complements that with the stem make up the locution—are given below as suffixed to the stem.

Two moods may be distinguished. One I call Independent, in which form the verb is unconnected with any other verb, the other Dependent, in which the verb is connected with some verb in the Independent mood.[2] Mr. Chamberlain calls them Indicative and Oblique respectively.

There are in each mood three forms or quasi-tenses—a present, temporally indefinite; a past denoting action or state completed at or before the present or some indicated past time; and a future denoting action, &c., that may or will be completed at a future (or past-future) time. Each of these tenses in the Independent mood has a predicative or simpler form and a relative form—as relating to some noun. It is the simpler form, identical with the stem in u, which is found in the dictionaries (in LemarÉchal’s the stem in a, i, or e is found). The relative form ends in uru (modern iru or eru). Thus toki sugu, time passes; suguru toki, tempus quod fugit (some verbs in uru are transitive as tatsu, stand, tatsuru, make stand, set up). Only those verbs which have e or i stems throughout appear to possess this relative form;[3] in verbs with i and a stems this morphological distinction does not exist.

The Past has four forms, each double (P. predicative, R. relative).

(1) P. ki, R. -shi.

(2) P. tari (te-ari), R. -taru (te-aru).

(3) P. -tariki, R. -tarishi (composed of (2) and (1)).

(4) P. -nu, R. -nuru.

[ki (kuru) = come (cp. Ital. vien fatto); shi (suru) = do; te = stem of tsu, tsuru, continue; ari is an irr. P. form of aru, be; nu is an old verb = be.] In i a verbs there is a past in P. -eri, R. -eru (see below).

The Future is formed by suffixing mu (or namu) to stem or naramu (ni [nu] aramu [aru]) to R. form if any.

Schematic Examples.

kogu, row (with oars or sculls).

tomu, stop, and sugu, pass, are conjugated like kogu, except that tomu replaces i and a of kogu forms by e, and sugu by i.

In the Dependent mood the forms are constructed with the help of -ba (the particle ha, voiced). Shikaba (etym. of shika?), tareba, taraba (te areba, -araba), naba (nu) and naraba (ni araba).

Schematic Examples.

Of tomu, sugu, the Dependent Mood is similarly conjugated preserving the e and i vowels, and for tomeba using tomureba for tomaba tomeba, for sugiba sugureba, for sugaba sugiba.

The negative voice (there are no negative words = not, no, none). The Present Independent is formed by changing the -mu of future into -zu or -zaru (zu aru), P. and -nu, R. Thus kogu (kogamu), kogazu, kogazaru, koganu; sugu, sugimu, sugizu, &c.

In the past -zu becomes zariki, P., and -zarishi, R.; in the future -zaramu or -zhi, P. and R. (kogazaramu, kogazhi).

There is a second future with sense of obligation exemplified as follows:—kogu (not koga) -mazhiku (indefinite), -mazhi, P., -mazhiki, R., -mazhikariki, P., past fut. -mazhikarishi, R., past fut. -mazhikereba, conditional, when as, &c., and -mazhikuba, hypothetical, if, &c.

In the Dependent Mood we have:—

The Imperatives are:—

Kogazare (kogazu are); kogu nakare (naku are); koguna, kogu koto nakare (row-thing do-not), na kogi so, yume kogu na (emphatic imperative). But na with i or e stem is an affirmative imperative kogi na, do row!

There is another past in tsu[4], tsuru as kogitsu, kogitsuru, and a continuative form in tsutsu, kogitsutsu, while rowing. Also a form in -keri, P. and -keru, R. with a dependent -kereba as kogikeru, kogikereba. This keri form seems equivalent to a perfect definite, have rowed. Its etymology must be ki -ari; of tsu the stem would be te, which added to stem in i or e makes forms such as kogite, tomete, which are continuative indefinites almost equal to ‘whilst’, kogite, yuku, whilst-row-go, i.e. row to (a place).

The negative form of -te is -de, kogade, or -zu, -zu ni, -zu shite, as kogazu, not-row, kogazu ni, in, or by not-row, kogazu shite, not-row-do, all meaning essentially whilst-not-row.

The particles mo, too, to mo, that too (with variant domo) keredomo (ki-ari-to-mo), following or suffixed to verb-forms, give a concessive force, although, even though, &c., beku, with sense of must, can, will, ought, is suffixed to the u stem, thus kogubeku is must, or ought to row, or will row—the forms of beku are beshi, P.; beki, R.; bemi, bekariki, P.; -shi, R.; bekereba, bekuba, bekarazu, &c.; beku mazhiku almost = bekarazu.

Causative verbs (often used as transitives or as honour-forms) are easily recognized. Thus nageku, sigh, lament, nagekasu; or forms derived from su (suru), do, are added—shimuru, seshimuru, sasuru. Thus yaku, burn, yakasu, yakashimuru; suru (se), do, sesasuru, make-do.

Nu (ni) is an old verb ‘be’. Ni + aru = naru, become, be at, of, in. The future namu is a common auxiliary suffix. Namu may become namashi, by composition with mashi, more, a suffixed optative particle. Another optative is formed by adding the slight interrogative particle ya to the form in -aba as yukabaya, should like to go; sometimes the future form is in -ramu (aramu?) added to the u-stem, as aruramu for aramu, kofuramu for kohamu (kohimu), will love. Aru, be, is peculiar in that its P.-form is ari not aru.

The verb miru, see, keeps the mi throughout, miru, miki, mishi, mimu, miba, mitsu (past), mizu, minu, mizhi, mide, &c.; but mireba, miredomo, miru or miheku, &c. So hi(ru), dry, i(ru), shoot, (w)i(ru), be in or at, ki(ru), wear ni(ru), be like, ni(ru), boil. See Chamberlain, p. 66. By adding u, uru (get) with or without aru (be) verbs passive and potential are obtained. Thus yaku (burn) + ari + uru = yakaruru, so we have homeraruru (homu, praise), sugiraruru (sugu, pass), miraruru (miru, see). Or we have miyuru, see-get = seem, so omoyuru, omoheru (from omofu, think), or omoharuru (see Glossary, omofu, &c.), iheru (ihi-uru), is said, &c. From aru (be) lengthened to arafu, we get araharu, araharuru, arahaseru, be manifest or make manifest.

Other passive-like forms are iyu (i), shot, oyu (oi), grown old, omohoyu (omofu), thought. There is a form in -aku, thus ifu, say, ihaku—even the future has it, kakemu, will utter, kakemaku, and the past omoheraku, the morphological origin of which is unknown. It seems to give a substantival character to the verb; ihaku, the saying (is); omoheraku, what was thought (is): also a frequentative (rare) in mi, wemi-mi, wemazu mo, smiling and not smiling.

Under koso will be found the verbal forms used with that particle.

Of the verb su suru, do, the principal forms are:—

Independent Mood.

Dependent Mood.

Negative Voice.

Of the verb ku kuru, come, the principal forms are:—

Independent Mood.

Dependent Mood.

Negative Voice.

The other forms of kuru are easily formed from those of suru, replacing se and su by ko and ku.

There are not many true adjectives. Such have a sort of positional inflexion—a predicative form in -shi, akashi, be red, a form used chiefly with verbs (but also indefinitely) in -ku, akaku naru, become red, and a form in -ki, used mainly with nouns, akaki mono, red thing. There is also a form in -mu, -mi, which has a verbal force denoting a state or condition.

Adjectives may also be verbalized by the addition of aru, be, and many of its forms with obvious contractions.

Lastly, adjectival expressions are formed by adding naru, be or become, to adjectival stems yaharaka, yaharaka-naru, soft, gentle.

Nouns substantive (and pronouns) are absolutely devoid of gender or case, and only occasionally have a plural suffix ra or tachi or nodo or domo.

Pronouns are few, and—the personal pronouns especially—are little used, the subject of the verb being generally unexpressed and left to be gathered from the context.

Interrogative Pronouns.

Other pronouns are shi, onore, mi, self, onore and mi, often I myself, or I, ono-ono, every, mina, all. Other pronominal forms, chiefly indefinite, are explained in the glossary.

Only the principal particles need here be noticed. Others are explained in the glossary. The following are of special importance, and are found mainly in connexion with nouns as postpositions. They are wa (ha), ga, no, ni, he (e, ye) and wo.

ha (wa, ba) isolates and emphasizes the noun rather as apart from the verb—chichi haha ha, uwe-samukaramu, father and mother, they will be hungry and cold; kono toki ha, ika ni shitsutsuka, this time then while doing what. This emphatic force explains its occurrence after verbs and sentences or clauses—ihazu-ba, not say indeed = should we not say …; Yamato ni ha, it is in Yamato that …; yuki furu yo ha, a night indeed on which snow falls: a wo koso se to ha, it is I indeed whom (you should call) husband.

ga is a possessive particle, wa-ga (warega) ohokimi, my great lord; Wazami ga hara, plain of Wazami. It may be used before verbs, tori ga naku, the bird cries; or after, nabiku ga gotoku, like bending before. Sometimes it resembles ha kimi ga agari-ki-masamu, my lord, he will embark.

tsu is possessive, amatsu kami, gods of heaven, kuni tsu kami, gods of the land. It seems to follow names of things only.

no (originally nu, be?) is used as a genitive particle, like ‘of’. It is more general than ga, which is special; no sometimes almost equals ha (wa). Mi-torashi no adzusa no yumi no hazu no oto, sound of the notch of the bow of whitewood of his royal grasp; Uchi no ohonu, the great moor of Uchi. This particle may connect other words than nouns, ari no kotogoto, all one has, or, there is (ari); miru no goto, like miru (seaweed); ame tsuchi no wakareshi toki, heaven—earth’s separated time (time of separation of heaven and earth). This connexion by no of parts of a sentence—often wide apart—other than nouns must be carefully kept in mind. Thus read hito no [mono wo omofu], a man’s thinking of things (regretfully): not [hito no mono] wo omofu, to think of a man-thing (or man’s things, &c.)[5].

ni (perhaps a stem of nu, be) = in, to, at, for, with, by, on, near; Yamato ni, in Yamato; toki ni, at time = when; miya ni amori, descend from heaven to the palace; te ni tori, take in the hand; ashita ni, on the morrow; asagari ni tatasu, start for morning-hunt; tokoro ni yuku, go to a place; also adverbially, yasukaranaku ni, in a not-restful way; aya ni, strangely; kogi-yuku ni, in or while rowing on; tokoshihe ni, everlastingly. Ni sometimes transfers the action of a causative or transitive verb to the noun it follows.

wo, after a noun indicates it as object of action or, sometimes, as subject of state or condition; unasaka wo sugite, passing beyond bounds of ocean; miyako wo tohomi, miyako (as to), be distant (Aston); mikado wo sadame, settle on a site for a palace; kuni wo sadame, terram debellare. In this sense wo is often omitted, oi mo sezu, old-age even not-doing = never growing old. It may follow a verb as an emphatic particle, or even a noun as such, but in these cases there is probably an ellipsis of suru (do) omofu, (think) or the like, or again it may follow a particle, to bakari, &c. (Aston). Where it appears to have an adversative force there is probably ellipsis of omoheba, omohite, &c. Wo ba = wo ha, and singles out the noun as specially emphasized. Wo sometimes almost equals ni.

he (lit. quarter, direction, locality, tract in space or time) = towards (ni = to). Rare in the texts in this volume, yori (lit. approach, or be close to, or stop at), shortened often to yo, yu, means by extension ‘from’—in later language also ‘than’.

ya, yo, is a vocative or exclamative interjection.

to, that, is a connective particle = and, also after a phrase marks it as quotative; toko miya to sadame, establish as an everlasting shrine (or palace), sugimu to omoheba, thinking (intending) that it should outlast. Hikohoshi ha Tanabata tsu me to, Hikohoshi and the Weaving Woman; se to ha norame (see under ha); kamusabu to, in a divinely awful manner; hito wo yahase to … kuni wo osame to … makitamaheba, as he was charged to subdue the people and pacify the land (people subdue that … land pacify that—as he was charged with). The different uses of to can always be made out by taking it as that.

Other particles are:—

ya, slight dubitative and interrogative, also exclamative. With negative it gives an affirmative sense.

ka, stronger than ya.

ya ka with ha (yaha kaha) imply a certainty.

ka mo, final interjectional expression = is it not even so? i.e. emphatically or admiringly, it is so!

kana (gana, mo gana), final interjection of emotion suited to what precedes.

mo, also, too, even; mo … mo, both … and.

so (zo) perhaps, sore, this! emphasizes the preceding word.

koso (ko[re], so[re]) this-that (Aston), more emphatic than so (zo), commonly precedes the verb, expressed or understood. There is a quasi-conjugation with koso—

koso koge;

koso kogere;

koso kogishika,—kogitare,—kogitarishika,—koginure,—kogitsure,—kogame,—koginame,—kogurame, &c.

a wo koso se to ha norame—in prose, a wo se to ha koso norame, do call me husband!

namu (nan) is very like koso, but less emphatic.

goto = gotoku, like also sometimes kotogotoku, generally, all; goto ni (after the noun), every, each.

dani, even, at all events; sura, even (unexpected); sahe, even (additional).

shi, nomi, bakari, just, only, precisely—in ascending degree of certainty. All of these, especially shi, are often not more than slightly emphatic expletives.

made, up to; gari, direction of; kara, from; kara (gara) = nagara (naru karada or naru kara), just as (applied to preceding noun); after verbs, whilst, although.

kaku, thus, to mo kaku mo, that way and this way, anyhow, somehow.

kashi, be it thus, so be it, may it be so! (Aston).

mashi (mase, &c.), mahoshiki, verbal terminations expressing desire or contingency.

meru, verbal termination (derived from fut. in mu, mu or mi aru?), indicates some degree of likelihood.

ramu, for aramu, hana chiruramu = hana chiru aramu = hana chiramu, the flowers will, wilt.

rashi = ramu, nearly; natsu kitarurashi, seemeth the summer to have come (Aston). There is an adjectival ending, rashiki, of similar import.

Of the syntax of Old Japanese little need be said. It is simple owing to the absence of almost the whole apparatus of Western grammar. The order of words has already been mentioned, and in prose is rigid—in poetry inversions are common. In large measure it is the opposite of English order, and this fact, together with the relegation of the verb to the end of the sentence, and the absence of expressed subject, constitutes the initial and principal difficulty of Japanese, apart from that of the vocabulary, the elements of which have usually a connotation different from that of their nearest representatives in any Western language. The absence of relatives and paucity of pronouns are additional difficulties, and the reader has to grow familiar by practice with the modes in which the more definite thought of the West is represented in the vaguer and looser language of Old Japan, where the visual aids of later Sinico-Japanese are not present. Nevertheless if the real meanings of the words be attained, the logical subject kept in mind as gathered from the context, the relations of words and phrases in apposition rather than in accidental or strict syntactical connexion be observed, the influence of the particles in edifying the sentence into a construction understood, and some facility gained in keeping the mind in suspense until the principal verb—read with the relations to it of the subordinate verbs—be reached, there is no great difficulty, apart from the inevitable one of difference of circumstance and allusion, in arriving at a comprehension of the texts. And these will be found, especially the Lays, to have preserved a peculiar beauty and charm, if not of the highest order, of their own, which no version can hope to convey.

The opening of the first lay sufficiently exemplifies the reversed order of the Japanese sentence:—

Read almost directly backwards, we get:—

Ko (maid) tsumasu ([who] plucks) na (salads) ni (on) kono woka (this knoll), mochi (holding) fukushi (truel), yo mo (oh the) fukushi (truel), mochi (holding) mi ko (fair basket), mo yo (oh the) ko (basket).

Maid who pluckest salads on this knoll holding a truel—oh the truel!—holding a fair basket—oh the basket!… a wo se to norame, me for thy husband name!

But even rendered almost in Japanese order, the lay is quite intelligible, and perhaps better to be appreciated than in any imitative version:—

In dealing with the less easy texts of the ManyÔshiu the following considerations must be kept in mind:—

The honour-forms (see below) indicate a (logical) second or third person, or something connected with such a person, directly or indirectly.

The particle wo may often be rendered as if ni, or as an exclamation, or as if followed by some form of omofu. I do not regard it as in itself a quasi-conjunction. The various functions of no must be kept in mind.

The frequency of inversion in order of words must not be forgotten. The makura kotoba may be ruled out, as well as prefaces, exordiums, introductions, and the like—thus the kernel of meaning may be got at, and the decoration then added.

It must be recollected that of the elements of the vocabulary very few can be exactly rendered by a single English word, their connotation for the most part is more or less different from that of any possible English equivalent. In the Glossary an exhaustive definition of the meaning is not attempted, enough only is given to suggest the translation of the sentence in which the word appears. Hence since meanings came to change in the course of time, those given being only what are required for these texts sometimes differ from the meanings found in the foreign dictionaries, all of which are very imperfect—very seldom, however, from those offered by that excellent native dictionary the Kotoba no Izumi.

THE LANGUAGE OF THE MANYÔSHIU

I found the following remarks upon the somewhat lengthy, and—to me—not always very clear, observations of Masazumi concerning the language of the ManyÔshiu contained in his sÔron or General Introduction.

Phonetic Contractions. These are not uncommon, and are principally vocalic. Such are wagimo (my sister) for waga imo; ariso (wild shore) for ara iso; kafuchi (within the rivers) for kaha uchi; amori (descent from heaven) for ame ori; kurenawi (a shade of red) kure no awi, lit. an indigo (i.e. a dye) brought from China; konure (tree-top) ko no ure; ke (come, pass) ki he—ke nagaku kohishi, long time loved; ke no kono goro ha, a time lately gone by; arumi (wild sea) ara umi; futsuma (stout horse) futo uma. The above are substantival, others are adjectival or verbal, as:—

naru (be in, at, of), ni aru, as in Yamato naru, ihe naru, &c.

nari—ni ari, as oto su nari = oto suru de ari.

naru—no aru, as hana naru toki = hana no aru toki.

sareba—shi areba, as haru sareba.

chifu—to ifu (they call, ’tis called).

ku, ki-yu; ke, ki-he; mesage, meshiage; kakage, kaki age; motage, mote age, &c.

Phonetic Extensions. Such are:—

nagaru—nagarafu, nagarahi, flow, flow in quantity, &c.

yobu—yobafu, yobahi, call, invite, pay court to (girl).

nageku—nagekafu, nagekahi, draw deep breath, lament, &c.

So causative forms exist used as honour-forms, nagekasu for nageku, and a termination in aku as nagekaku, kakemaku, ihaku, for nageku, kakemu, ifu used substantively. Ihafu, bless, is perhaps an extension of ifu, speak, say.

Of yobahi (yobu extended) a humorous explanation is sometimes given, yo (by night) hahi (creep), visit stealthily by night. Etymology of this kind is extremely easy in Japanese.

Auxiliary Particles (tasuke kotoba):—

Such are shi (emphatic), kami-yo shi omohoyu, ihe shi shinubayu.

wo (emphatic), as in yatsu yo ni wo, tanushiku wo arana(mu), but in phrases like mitsutsu yukamu wo, wo = mono wo.

ya, ame shiru ya, ama tobu ya, kashikoki ya, Afumi no ya, naku ya uguhisu, &c.; but in kimi ya komu, ya is dubitative—interrogative. In kakusafubeshi ya, sugi nikerazu ya, ya = ya ha = Lat. num.

nane, term of address and endearment, must not be confused with nane of imonane, senane, which are terms of relationship.

i suffixed as in sekimori-i, Unahi-wotoki-i may be regarded as emphatic, or isolative (like ha, according to Chamberlain), or as etymologically = the Korean postposition i (Aston).

we, a terminal interjection as are ha sabushi we.

ro ka mo, tafutoki ro ka mo, here ro ka mo = is it not so even! ro, according to Mr. Chamberlain, is a post-position equivalent to te or nite.

mo, mo yo, mo ya, shi mo, ya shi, yo shi, are all interjectional expressions, not mere fill-gaps, variously emphasizing or calling attention to the meaning of the words, &c., they are attached to.

Honour-Words. Such are ohomi (exalted, grand), as ohomi kami, ohomi uta, ohomi ma; mi, as mikokoro, miko, mi-hakashi, mi-keshi, mi-nemasu (His godship the God, His Majesty’s song, His Majesty’s horse, my lord’s meaning, my lord the Son (Prince), girt by my lord, my lord’s garments, my lord sleeps).

Other honour- (or praise-) words are:—

oho (great), futo (stout), toyo (rich), o (for oho) as in obito, omi (court officials); mi, ma, mo, sa, oho also denote excellence (e?), wo may be antithetic to oho, or it may be a term of intimacy, a sort of diminutive of endearment.

The following honour-verbs should be here noted. Tamafu, tamaheru, tamahari, oboshimesu, kikoshimesu, shiroshimesu, takashiru, owasu, owashimasu, masu, mimakareru; verbs in causative form, -asu, -saseru, -seshimu; verbs in quasi-passive or potential, -uru, -ruru, raruru.

Other laudatory expressions are tama (precious), uma (delicious, fine), midzu (brilliant, fine), umashi (uma), kuhashi (lovely). A curious gradation of honour-forms is seen in the following: mi yomimaseru ohomi uta (where the lay is attributed to a mikado), ohomi uta, mi uta, yomitamaheru uta, yomeru uta (when the author is miko, ason, or official of lower rank).

Among terms of affection may be noted:—

waga (mine) or wago ohokimi, waga kimi, waga kuni, waga se, wagimo, a oto, ago, na, se na, na se nane, nanimo, imo nane (my great lord, my lord, our land, my husband, my younger sister (wife or mistress), my brother, my child, thou, thou elder brother, elder brother thou, thou elder sister, thou younger sister, &c.).

Words of exclamation, hope, entreaty, wonder, &c.:—

Ana, ana ni (ni is not a particle), ya (= yo!), ka (= ?), kana, gana, mo gana, kamo, mo ga mo, moga, ga mo na, ya mo, shika, sane, ne, kose, kosene, na-so na-yume, ari kosenu ka mo (ari-koso-ne). These are explained in the notes to the text or in the glossary—their meaning is obvious for the most part.

Repetitive or cumulative expressions (kasane kotoba) are common in the ManyÔshiu and contain, mostly, a quibble in meaning or rhyme in sound. Such are chichi no mi no chichi (father = chichi = maiden-hair tree—the quibble is on the two chichi, and the double meaning of mi, fruit and person). Hahasoba no haha, mother (haha), as of the oak (hahaso = Quercus dentata), &c. Shiga no karasaki sakiku araba (here the rhyme is Karasaki (Cape Kara in Shiga) and sakiku araba (if flourish)). Such expressions are explained in the notes or glossary.

Other repetitive expressions are simply emphatic or poetic: Tsudohi tsudohi imashite, they (the gods) assembling, kamu hafuri hafuri, burying—burying him (a mikado or miko) as a god; ake no sohobune sohobune, a red red-stained red-stained-ship, &c.

Words relating to the Sovran. Sumeramikoto (Supreme Majesty) is found only in the dai (arguments), not in the uta. The commonest term in the latter is Ohokimi (great Lord, grand seigneur), but this expression is not confined to the Sovran. Other terms are sumeragi (sumerogi, suberogi) and sumerami; -gi = prince, mi = princess (conf. Izanagi and Izanami, Inviting Male and Inviting Female (K.)); but see Aston, ShintÔ. What sume means is uncertain, it may be connected with sube (shiru), universally know, i.e. govern the land. Other expressions are kamuro kamu subera or sumera or sumero, kamu adding the notion of ‘divine’. Ohokimi, ‘chief’, I take to be the oldest; the other forms smack of China. As to ‘ro’, see above remarks on the particle ‘ro’. In the Kogi etymology ro = the re o of are oya in kamu are oya = god-born ancestry.

Mikado means grand gate or palace, and by metonymy came to signify its lord, just as at the present day miya (grand mansion) denotes an imperial prince. Mikado also means ‘sovran dominion’ as in toho-mikado, distant palace, that is, wide dominion, sometimes applied specially to the Tsukushi government, to Korea, or even to China, as on the confines of, but still within, the authority of the Ohokimi of Japan. The following phrases may be here noticed—mikoto kashikomi, dread majesty, Ohokimi no make (or hiki) no manimani, in obeisance to the Ohokimi’s will or appointment.

Mi has several meanings which must be distinguished. They are (1) an honour-prefix, grand, great; (2) prefix of praise, e? = ma, true, real; (3) self; (4) body, person; (5) a stem of miru, see (6) root of midzu; (7) fruit of tree or herb; (8) the numeral three; (9) as a termination se wo hayami, swift the stream; hiromi, broad-like; fukami, deep.

Of the intensive prefixes i, ka, and ta, no explanation has yet been given. They resemble Greek ?a. Examples are i-yuki, ka-guroki, ta-moto-horu.

For the purposes of the present work, Dr. Aston’s grammar of the written language (third edition) is much the most useful.

THE SCRIPT OF THE MANYÔSHIU

The Lays are written wholly in Chinese characters. But these are employed in several very peculiar ways, and the texts as they stand are completely unintelligible to a Chinese, even to a Japanese, who has not specially studied them. In all the editions, however, except the Riyakuge, the columns of text are accompanied by a kana transliteration—in the Riyakuge the kana (hira) transliteration is given separately from the text.

At the date of the compilation of the Anthology—the middle of the eighth century—neither of the existing Japanese syllabaries had been invented. Their creation is ascribed to the learned priest KÛkai (KÔbÔ Daishi), the Doctor Promulgator of the Law of Buddha, who died in 834, nearly a hundred years later than the date of the final tanka of the ManyÔshiu. The Chinese character had therefore, perforce, to be employed in writing down the lays collected in the manner set forth in the volume of translations. The ideographs were used in part phonetically, in part lexicographically, as they had already been used in writing the Kojiki and the poems cited in that history, and in the almost contemporaneous but very different Nihongi.

The ideographs used phonetically were not, however, always employed in the same way. The forty-seven sounds of the syllabary—

—were already recognized, and were represented by a sort of alphabet composed of several hundreds of Chinese ideographs, each pronounced exactly or approximately sinice—that is japonico-sinice, or according to on or Chinese sound. Thus a was represented by two ideographs, ? (a in Chinese) and ? (an in Chinese), shi by twenty-five characters, such as ? (chih in Chinese), ? (ssu in Chinese), &c. The other sounds were represented by varying numbers of characters. According to this system, ame (heaven or rain) would or might be written ??, tsuchi (soil) ?? and so forth. A complete list of these Chinese phonetic ideographs is given in the SÔron (Introductory) volume of the Kogi.

A second method of using the ideographs was to employ them according to their kun (reading i.e. in pure Japanese), thus utate (extremely) was represented by ???, chihahi (for sachihafi, bless) by ???. Sometimes two characters represented one sound, thus ?? for a, ?? (isozhi) for i, ??, ushi no naku, ‘moo’ of cow, for mu. Some sounds (ku, ri, ru, ro, wa) are not found thus symbolized, that is, japonicÉ. A curious double character is ?? for se, another is ?{?} (hitsuzhi no ashi) for shi. A third and very confused script is exemplified in ?? kaherikomu (return), where kaheri is kun and komu (kon) is on, and ?? shirasamu (shall know), where shira is kun modified grammatically, and samu (san = three) is on representing the inflexion.

Still more confusingly, an ideograph may be used with an on (Chinese) sound resembling a kun (native) word, and the kun word may be employed, not in its natural sense, but as it were punningly, though more often no quibble is intended. Thus ? of which the on is ken, may be used for the verbal termination kemu, as in ?? (karikemu, will have reaped), where ? is employed lexicographically as kemu, as just explained. So ??? does not mean kimi arazu (lord is not) but ari nakuni, as there is not. Similar examples are kaherikomu and shirasamu cited above. So ? nan (difficult) for nani, what?, {?} ten for -temu; ? toku for toko, and so forth.

Or the Japanese reading of a character may be taken, but in a signification different from its true meaning, thus ? niha, a court or yard, for the particles ni ha, ? tama, jewel or pearl, for tamashii, soul, or even ?? yu-game = water-tortoise for yukame = will go, or the combination ?? = stone-two, which puzzled Shitagau so sorely, and finally turned out to mean made = until, to which may be added the commonest of all ? = kamo, a wild duck, used for ka mo, an expression of mingled entreaty and doubt.

The above devices were no doubt resorted to partly to supply the place of a syllabary, partly for purposes of abbreviation, the full writing of long Japanese words with a Chinese character for each syllable being found too laborious.

But characters are often also used in a punning or humorous way. Thus ?? = 2 × 2 are employed to represent the syllable shi (which means 4 in Japano-Chinese), but here is used phonetically to represent the emphatic particle shi or the syllable shi merely. So ????? = mountain ? upon mountain to represent ? idzuru, go forth, out, &c., ????? = three-prostrations-one-regard-night, i.e. a moonlight night (when one salutes the moon), ??? (81) as equivalent to ku ku, nine times nine, part of the word kukumeru (= fukumeru, hold in mouth, imply), and so forth. Characters used otherwise than to express their true meaning are called kariji, borrow-characters, i.e. characters borrowed to signify some other meaning than their own. Characters used in the orthodox manner are called manaji—true characters.

There are also guji, ?? double characters used, where one would suffice, thus ?? for ? nani. This is probably a mere embellishment.

Contracted script is not uncommon; thus we find ?? for ???? = arashi, a violent wind (rushing down from the mountains).

Lastly, there are abbreviated characters—a full list of which is given in the Kogi, together with a few characters peculiar to the Anthology and contracted expressions such as amori for ame ori, Yamato naru for Y. ni aru, haru sareba for haru shi areba, chifu for to ifu (modern tefu = cho of KyÔto), and so forth.

The above system of scripts is known as ManyÔgaki and is sometimes imitated in modern productions as in a recently published play now before me.

Texts preserved in such scripts could not fail to be corrupt or obscure in parts. Not only from the difficulty of reading the ideographs when these came to be translated into kana, but from the fact that the manuscripts were often in cursive character, easily misread and miswritten by the copyists. The text of the ManyÔshiu I have used is throughout that of the Kogi, but I have not omitted in the notes to the translations to give some attention to the various readings cited in that work. By far the best discussion of these seems to be that of Keichiu which with the Kogi is much superior in my opinion to the commentary of either Mabuchi or Motowori[6]—the latter especially appears to be lacking in critical acumen.

The following additional remarks on the script are important. According to Amano Nobukage (died 1734), in the ManyÔ script the Chinese characters are read japonicÉ in four ways:—

1. mana, as ? kokoro (heart, mind, &c.), where single characters are read as true Japanese words.

2. Where combined characters are read true, as ?? harukasumi (spring-mist); ?? aki-kaze (autumn-wind).

3. Where combined characters are read in full, but the meaning of the whole is not the sum of the meanings of the parts, as ??{?} kakitsubata (Iris laevigata)—the characters separately mean—kaki (fence), tsu (place), hata (flag).

4. Where combined characters are read together as a whole, as ?? uguhisu (Cettia cantans), lit. spring-bird; ??? mochidzuki (full-moon), lit. three-five-night = fifteenth night of a lunar month, when the moon is full.

To these categories must be added that of humorous combinations, thus ?? (sixteen) = shi shi (4 × 4), and is used for shishi (flesh); ??? kamo (wild-duck), lit. green-head-fowl.

In the Kogi text the characters are used in the following modes, all of which are fully illustrated by examples in a section of the soron or General Introduction to the Edition.

A. According to the on or Chinese sound (Japano-Chinese).

B. According to the kun (yomi) or Japanese reading.

In addition the Kogi gives the following categories:—

gikun, combinations not literally translated into Japanese—?{?} ametsuchi (heaven and earth), lit. dark-blue (of sky) and yellow-brown (of earth), ?? chichi-haha (parents, father and mother), lit. love-love. A great many of these are given; many would scarcely suggest the meaning of the combination; sometimes, as when ? futa (lid), is used to express ? futa (two), a word-play is involved.

kariji or borrowed characters which have been already exemplified. But one curious expression may be added, {?}?{?}?? masokagami (true-pure-mirror), the characters mean call (or bark?)-dog-follow-horse (ma) mirror; call-dog-follow is a humorous (?) description of ma, meaning horse, but here used for the homophon ma, true, e?, part of ma so, which has the same meaning.

Subjoined is the script of the text of Lay 118 which well exemplifies the peculiarities of ManyÔgaki. The columnar transliteration represents the similar transliteration into syllabic kana in the usual editions inclusive of the Kogi.

FOOTNOTES

[1] For a more complete account the grammars of Aston and Chamberlain must be consulted.