DIEGO COLLADO'S
GRAMMAR OF THE
JAPANESE
LANGUAGE

Edited and Translated
by
Richard L. Spear

International Studies, East Asian Series
Research Publication, Number Nine

Center for East Asian Studies.
The University of Kansas.

Dedicated
TO
The Memory of
Joseph K. Yamagiwa

Table of Contents

Preface

I Introduction [1]

The Grammatical Framework [3]

The Phonological System [6]

The Morphological System [8]

The Structure of Collado's and Rodriguez' Descriptions

Contrasted [11]

Bibliography [26]

Editorial Conventions [28]

II Ars Grammaticae Iaponicae Linguae

III A Grammar of the Japanese Language [105]

Prologue to the Reader [107]

The noun—Its Declension and its Gender [111]

Pronouns [118]

First Person Pronouns—Ego, etc. [118]

Second Person Pronouns—Tu, tui, tibi, etc. [119]

Third Person Pronouns—Ille, illa, illud. [120]

Relative Pronouns [122]

The Formation of the Verb and its Conjugation [123]

The Preterit, Perfect, Imperfect, and Pluperfect [124]

The Future of the First Conjugation [125]

The Imperative of the First Conjugation [125]

The Optative of the First Conjugation [126]

The Subjunctive of the First Affirmative Conjugation [127]

The Infinitive [129]

The First Negative Conjugation [131]

The Second Affirmative Conjugation [134]

The Second Negative Conjugation [135]

The Third Affirmative Conjugation [135]

The Third Negative Conjugation [136]

The Conjugation of the Negative Substantive Verb [137]

The Conditional Particles [139]

The Potential Verb [140]

The Conjugation of Irregular Verbs [141]

The Aforementioned Verbs—Their Formation and Diversity [143]

Certain Verbs Which of Themselves Indicate Honor [147]

Cautionary Remarks on the Conjugations of the Verb [148]

The Adverbs: First Section [156]

Adverbs of Place [156]

Adverbs of Interrogation and Response [159]

Adverbs of Time [159]

Adverbs of Negation [160]

Adverbs of Affirmation [160]

Comparative Adverbs [161]

Superlative Adverbs [162]

Adverbs of Intensity and Exaggeration [162]

Accumulative Adverbs [162]

Adverbs that Conclude and Claim Attention [163]

The Case Prepositions [164]

Conjugation and Separation [166]

Interjections [167]

The Syntax and the Cases that are Governed by the Verbs [168]

Japanese Arithmetic and Numerical Matters Concerning Which

Much Painful Labor Is Required [174]

Some Rules on the Conjugation of the Verb in the Written

Language [182]

IV Works Consulted [185]

V Index to Grammatical Categories [187]

VI Index to Grammatical Elements [189]

Preface

The purpose of this translation of Collado's Ars Grammaticae Iaponicae Linguae of 1632 is to make more readily available to the scholarly community an annotated version of this significant document in the history of both Japanese language study and grammatical description in general.

Collado's work, derived in all its significant features from the Arte da lingoa de Iapam completed in 1608 by João Rodriguez, is in a strict, scholarly sense less valuable than its precursor. However, if used with the Arte as a simplified restatement of the basic structure of the language, Collado's Grammar offers to the student of the Japanese language an invaluable ancillary tool for the study of the colloquial language of the early 17th Century.

While less extensive and less carefully edited than the Arte, Collado's Grammar has much to recommend it as a document in the history of grammatical description. It is an orthodox description attempting to fit simple Japanese sentences into the framework established for Latin by the great Spanish humanist Antonio Lebrija. Thus, as an application of pre-Cartecian grammatical theory to the structure of a non-Indo-European language, the Ars Grammaticae is an important document worthy of careful examination by those wishing insight into the origins of what three centuries later was to become the purview of descriptive linguistics.

The present translation was begun with the able assistance of Ms. Roberta Galli whose contribution to my understanding of the Latin text is most gratefully acknowledged. For his continued encouragement in this undertaking I am grateful to Professor Roy Andrew Miller. Thanks are also due to the Graduate School of the University of Kansas for its support in the preparation of the manuscript and to Ms. Sue Schumock whose capable typing turned a scribbled, multi-lingual draft into a legible manuscript. The imperfections are my own.

R.L.S.

Lawrence, Kansas

May, 1975

Introduction

In 1632, as the Christian Century in Japan was drawing swiftly to a close, three works pertaining to the Japanese language were being published at Rome by the Congregation for the Propagation of the Faith. These works were by the Spanish Dominican Father, Diego Collado (d. 1638), who had spent the years from 1619 to 1622 in Japan. Their publication clearly reflects the vitality of the missionary spirit in that age as well as the important place reserved for language study in the propagation of the faith.

The first two works, whose manuscripts had been prepared in Madrid the year before, were a grammar and a dictionary of Japanese. The third, prepared in 1631, while the larger works were being seen through the press, was a guide to the taking of confession written in both Latin and Japanese.[^[1]] The grammar, drafted in Spanish, was published in Latin in 1632 under the title Ars Grammaticae Iaponicae Linguae. It is this work that is translated here. The dictionary, only at the last moment supplied with Latin glosses to supplement those in Spanish, was published in the same year with the title Dictionarium sive Thesauri Linguae Iaponicae Compendium.[^[2]] Taken together these three works by Collado constitute the final extant efforts of those who studied the Japanese language first hand during the Christian Century.[^[3]]

Two other grammatical works must be mentioned here as central to the proper assessment of Collado's Grammar. They are both by the great Jesuit scholar, Father João Rodnguez (1561-1634);[^[4]] the Arte da Lingoa de Iapam (Nagasaki, 1604-8, hereafter the Arte), and the Arte Breve da Lingoa Iapoa (Macao, 1620, hereafter Arte Breve). The first

is by any standards the greatest grammatical study of Japanese made during the Christian Century. It is further, as we shall see, the primary source for Collado's Grammar. The Arte Breve, on the other hand, is not directly related to Collado's work. Indeed it is clear that Rodriguez' 1620 Macao publication was unknown to Collado. Nevertheless, since the Arte Breve is an abbreviated version of the Arte with a purpose similar to the Ars Grammaticae, a comparison of these two books with respect to the way they systematize the material from the Arte is included in this introduction to contribute some insight into the treatment of the Japanese language at the beginning of the Tokugawa Period.

In presenting this translation two potential audiences are envisioned. The first, and more restricted, group is that having an interest in the history of the Japanese language. It is hoped that an English version of this work will make more readily available this significant material pertaining to the Japanese language as spoken in the early modern period. I use the word significant here to avoid granting excessive value to a work which derives such a large portion of its material and insight from Rodriguez' Arte.

The second, and wider group for whom this translation is intended is that which has a need for an edited edition of an important document in the history of grammatical description. In this area of scholarship Collado's work is of more than moderate significance. It was accepted for publication by the prestigious Propaganda Press; and, even if those more familiar with Japanese than the editorial board of that Press might have had serious reservations concerning the linguistic accuracy of the text, it is reasonable to assume that the Press judged it to be a good example of grammatical description. It thus represents a grammar of a non-European language which suited the requirements of the day for publication at Rome.[^[5]]

In order to permit this translation of the Ars Grammaticae to be of use in both these areas of scholarship I have made an effort to reduce to a minimum those places where a knowledge of either Japanese or Latin is required for the comprehension of the translation. It is sincerely hoped that the result is not an effort that is satisfying to neither, and thus to no one.

Because of the derivative nature of the text, this translation has put aside a number of important philological problems as better dealt with within the context of Rodriguez' grammars. This decision has its most obvious consequences in the section on the arithmetic, where innumerable data require exposition. However, since a basic purpose of this translation is within the context of the history of descriptive grammar, these tantalizing side roads have been left unexplored. It is, nevertheless, hoped that this translation will serve as a convenient tool for those wishing to make a more detailed investigation into the philological questions raised by the text. But I must caution those who would undertake such an inquiry that they had best begin with a careful study of the works of Father Rodriguez.

With its limitations acknowledged, the Ars Grammaticae Iaponicae Linguae remains a document worthy of our interest, and I offer this translation in order that Collado's work may more easily find its proper place in the history of descriptive grammar.

The Grammatical Framework

Collado perceived his task to be the presentation of a grammar of Japanese which would have sufficient scope to equip those dedicated to the propagation of the faith with a knowledge of the proper spoken language of his time. While he concludes his grammar with a brief, and rather presumptuous, statement concerning the written language, his purpose is clearly to train his students in the fundamentals of colloquial speech. His sensitivity to this point is demonstrated by his carefully transforming those examples presented by Rodriguez in the written language in the Arte into correct colloquial expressions in his own grammar.

The description is, of course, prescriptive. But given its age and its purpose this ought not to be construed in the contemporary, pejorative

sense. Collado, as Rodriguez and indeed all the grammarians of the period, felt obligated to train their students in those patterns of speech which were appropriate to the most polite elements of society. Particularly as they addressed themselves to missionaries, they wished to warn them away from such illiteracies as might undermine their capacities to propagate the faith.

The description further reflects the traditional process conceptualization of language. This is particularly obvious in the treatment of the verb. Thus:

Praesens subiunctiui fit ex praesenti indicatiui mutato u in quo finitur in eba.... (The present subjunctive is formed from the present indicative by changing the u in which it ends to eba....) [p. [23]].

In general each of the verbal forms is conceived to be the result of a specified alteration of a basic form. Likewise the nouns are treated within the framework of the declension of cases.

The treatment of Japanese forms is based upon a semantic framework within which the formal characteristics of the language are organized. For example, given the construction aguru coto aró (p. [31]) and its gloss 'Erit hoc quod ist offere: idest offeret (It will be that he is to offer, or he will offer),' it is clear that the aguru coto is classified as an infinitive because of its semantic equivalence to offere. The same is true of the latter supine. If the form in Latin is closely associated with such constructions as 'easy to,' or 'difficult to,' the semantically similar form which appears as the element iomi in iominicui 'difficult to read,' must be classed as the latter supine. Rodriguez in his Arte Breve of 1620—unknown to Collado—makes an attempt to classify the structural units of Japanese along more formal lines; but in Collado's treatment the semantic, and for him logical and true, classes established by the formal structure of Latin constitute the theoretical framework through which the Japanese language is to be described.

Collado makes reference to two specific sources of influence upon his grammar. The first is included in the title to the first section of the grammar, Antonius Nebrissensis. It is to this great Spanish humanist,

better known as Antonio Lebrija (1444-1522), that Collado turns for the model of his description.

An examination of Lebrija's grammar, the Introductiones Latinae (Salamanca, 1481), shows that from the basic outline of his presentation, to the organization of subsections and the selection of terminology, there is little departure by Collado from his predecessor.

Even in such stylistic devices as introducing the interrogatives by giving the form, following it with "to which one responds," and then listing a number of characteristic answers; Collado is faithful to the Introductiones.

But it is from his Jesuit colleague, Father João Rodriguez, that Collado receives his most significant influence. There is no section of his grammar that does not reflect Rodriguez' interpretation of the raw linguistic data of Japanese. On the basis of the innumerable examples taken from Rodriguez—most of the substantive sentences are directly quoted from the Arte—as well as the parallel listing of forms and identical descriptions of certain grammatical phenomena, it is clear that the writing of the Ars Grammaticae Iaponicae Linguae consisted to no small degree of abridging the exhaustive material contained in Rodriguez' grammar and arranging it within the framework of Lebrija's Introductiones.

To say that Collado followed Lebrija in the general structure of his description is not to imply that he fell heir to all of his precursor's virtues. The Salamanca grammar of 1481 is a masterpiece of orderly presentation. Printed in lettera formata with carefully indented subdivisions, it offers the student a clear display of the conjugational system as well as long columns of Latin examples of a given grammatical structure, accompanied on the right side of the page with Spanish equivalents. Collado makes little effort at copying this orderly display. There are in his presentation no paradigms, but instead only loosely connected sentences that talk the student through the various forms of the conjugation; and there is no orderly array of examples. Add to this the innumerable factual and typographical errors, and one is left with a presentation that lacks most of the basic scholarly virtues of its precursor.

A similar criticism may be leveled against the work from the point

of view of Rodriguez' influence. Without matching the Introductiones in orderliness, the Arte more than compensates for its casual format by containing a mass of exhaustively collected and scrupulously presented linguistic data.[^[6]] There was available no better source than the Arte from which Collado might have culled his examples of Japanese.

One doubt that remains in assessing Collado's use of Rodriguez' material is that perhaps his presentation of the most readily understandable material in the Arte is not so much an effort on his part to simplify the learning of Japanese for his students, as it is a reflection of his lack of adequate familiarity with the language he was teaching.

The Phonological System

A study of the phonological data reveals the Ars Grammaticae Iaponicae Linguae to be of minimal historical value. Any student of the phonology of early modern Japanese should turn to the far more reliable work of Father Rodriguez. Nevertheless, certain aspects of Collado's transcription require our attention.

The most obvious innovation in the representation of the language is Collado's transcription with an i of the palatal consonant which all his contemporaries record with a y. Thus in the text we find iomi and coie (terms for native words and Chinese borrowings) where Rodriguez writes yomi and coye. This change was affected while the text was being translated from the Spanish manuscript which uses y; and Collado himself must have felt the innovation to be of dubious value since he retained y for the spellings in the Dictionarium.[^[7]]

Collado's handling of the nasal sounds is too inconsistent to be a reliable source for phonological data. Given his rather awkward specification that nasalization is predictable before what we must assume he means to be the voiced stops and affricates,[^[8]] his grammar presents an uncomfortably irregular pattern in the transcription of the phenomena. Thus, on page [39] we find vo mõdori aró ca? as well as

modori aró ca?. Again, what he presents as the ending zũba in his description of the formation of the negative conditional (p. [34]) appears in tovazunba in its only occurrence in a sample sentence (p. [62]). To further confound the issue such forms as tovazunba and qinpen occur in contrast to sambiacu, varambe, and varãbe.

In Chart 1 the traditional pattern of the gojūonzu (chart of 50 sounds) is followed as a convenient framework in which to display the transcriptional system employed by Collado.

Chart 1

COLLADO'S TRANSCRIPTION SYSTEM

In this chart the phonemic grid is presented in a broad phonetic

notation while the underlined entries are in the form used by the text. Dashes indicate sequences which do not occur in the Christian material; while the forms in parentheses are sequences which do not occur in the text but have been reconstructed on the basis of the overall system from sequences attested to elsewhere. The forms ie, vo, vó, and vô have been placed in brackets to indicate that neither /e/, /o/, /oo/, or /au/ occur in the syllable initial position; and, where in the modern language they do, the text regularly spells that with an initial i or v. The forms in eô at the foot of the chart represent sequences that are phonetically identical to the forms above them, but which are transcribed differently to reflect morphological considerations; e.g., the form agueô from the stem ague. The phonetic values of /au/, /uu/, and /ou/ are [ɔ:], [u:], and [o:].

Two aspects of the usage of q should be noticed. First, as in the Arte, c is changed to q before o and u, when the sequence occurs at a morphological juncture; e.g., ioqu 'well,' and iqó 'I shall go.' (This rule does not extend to a in such contexts; cf., iocatta 'was good.') Second, in contrast to the system used by Rodriguez, Collado does not feel compelled to follow q with u in all contexts. Thus what Rodriguez spells as queredomo Collado spells as qeredomo. Finally, the text records one usage of the letter h in the exclamation ha.

The Morphological System

Collado's treatment of the morphology contains one quite obvious difference from those of his predecessors: he isolates the particles of the language as separate elements of the structure. While his effort is more or less carelessly maintained by the type setter, his attempt to establish a division between the semantemes (shi) and the morphemes (ji) of Japanese by establishing formal distance between his verba and particula, reflects his consciousness that the morphological elements in Japanese are of a different order than those in Latin. At times, such as when he describes the preterit subjunctive as agueta raba, his divisions fly in the face of derivational history. But he can claim a reasonable justification for his decision by citing Rodriguez' rule for the formation of this form; "add raba to the preterit of the verb" (Arte, 18v). Perhaps it is a prejudice founded upon familiarity with

contemporary romanizations, but I cannot help but consider this attempt to give greater independence to the particles as an improvement in the representation of the morphological system.

In all other significant facets of the morphology Collado follows the principles established by Rodriguez with the one exception that in the over-all systematization of the verbal formation and conjugation he follows the classifications established in Lebrija's Introductiones rather than those which Rodriguez inherited from the Institutiones of Alverez. The most significant difference between the two systems is the use by Lebrija of the term subjunctive in his description of the moods where Rodriguez gives independent status to the conjunctive, conditional, concessive, and potential. As we shall see, after presenting the conjugational system of the verb within the framework of Lebrija, Collado breaks the expected sequence of his description of the verb to interject a section on conditional constructions and another on those of the potential.

In the treatment of the tenses Collado breaks with Rodriguez in not attempting to establish an imperfect for Japanese, but he does follow him in the overall classification of the conjugations. Thus:[^[9]]

To the description of this general system Collado adds the treatment of the substantive verbs. This section in many respects is the weakest in his grammar with a portion of his description lost in composing the final text.

Since Collado does not, as Rodriguez, present the conjugations in paradigmatic form, I have extracted from his presentation the most representative forms of the verb ague, uru for each of the categories of the system, and presented them in Chart 2 for reference.

Chart 2

THE CONJUGATIONAL SYSTEM

The Structure of Collado's and Rodriguez' Descriptions Contrasted

In every section of his description, Collado is indebted to the material presented by Rodriguez in his Arte da Lingoa de Iapam. The structure of the Ars Grammaticae, however, follows a much more simplistic design than that of the Arte. As a consequence Collado found it necessary to assemble his data from various sections of Rodriguez' description. In the paragraphs which follow we will briefly sketch the structural relation between these two grammars.

As he clearly states in his title to the main portion of the grammar Collado bases his description on the Introductiones of Antonio Lebriya, and more specifically upon that portion of the great Latin grammar which dealt with the parts of speech. Further, he limits himself to the spoken language rather than attempting, as does Rodriguez, an integrated treatment of both the spoken and written grammars.

Under these influences Collado's grammar takes on the following form:

A Prologue (including the phonology) ... 3-5

The Body of the Grammar (by parts of speech) ... 6-61

A Brief Syntax ... 61-66

A Treatment of the Arithmetic ... 66-74

A Note on the Written Language ... 74-75

In contrast Rodriguez' Arte, prepared under the influence of Alvarez' Institutiones, develops its description over the span of three books which treat both the spoken and written grammar in progressively greater detail. Thus:

The Introduction ... iii-v

Book I

The Declensions ... 1-2v

The Conjugations ... 2v-54

The Parts of Speech (Rudimenta) ... 55-80v

Book II

The Syntax of the Parts of Speech ... 83-168

Styles, Pronunciation, Poetics, etc. ... 168-184

Book III

The Written Language ... 184v-206v

Names, Titles, etc. ... 206v-212v

The Arithmetic ... 212v-239

Given these differing formats[^[10]] it is clear that Collado is unable to cope adequately with the more complex aspects of the grammar, specifically those syntactic constructions to which Rodriguez devotes almost an entire book.

An analysis of Collado's description and a listing of the portions of Rodriguez' grammar from which material was taken yields the following:

Two aspects of Japanese were not able to be described with any degree of satisfaction by Collado; the adjectives (adjectiva) and the prepositions (praepositio). His difficulties, attributable to the basic structural difference between Latin and Japanese, were compounded by the fact that Rodriguez too was unable to find a satisfactory solution to their description.

With respect to the adjectives, Collado attempts to deal with their functions in the manner appropriate to Latin, that is as a sub-class of

nouns (pp. [9]-11). He also recognizes their formal similarity to the verb and treats them briefly as a sub-class of the substantive verb (pp. [32]-33), but his heavy reliance upon the semantic categories of Latin does not permit him to follow Rodriguez who is able more clearly to recognize their formal as well as their functional distinctiveness.

Concerning prepositions, Collado was confronted with an all but insurmountable taxonomic problem. Here too Rodriguez was unable to develop a completely satisfactory descriptive framework. In the Arte the term posposição is used for those particles which function in a manner similar to the Latin prepositions; e.g., tameni, taixite, and tomoni (cf. 73-73v and 140-148v); the term artigo is used for those particles having the functions of the inflectional endings of Latin; e.g., ga, ye, and ni (cf. 1-2, 78, and 137-140); and the general term particula is used to cover the broad spectrum of particles that include adverbs, conjunctions, and exclamations, as well as those otherwise unaccounted for elements which end phrases, clauses, and sentences; e.g., no, nite, and yo (cf. 77-78 and 144-154v).

Collado, rather than attempting to refine the system suggested by Rodriguez, follows the Arte in listing as praepositio those elements which translate the Latin prepositions (pp. [57]-59) but uses the term particula to cover all the other particles of the language.

This tendency of Collado's to retreat from the challenging problems left unresolved by Rodriguez constitutes the greatest weakness of his description. Given concise grammatical descriptions on the one hand and over-simplified versions of previous works on the other, the Ars Grammaticae Iaponicae Linguae unfortunately falls among the latter.

In his shorter work, the Arte Breve of 1620, Rodriguez retains the same general format, but makes every effort to reduce the description to its barest essentials. Thus:

Book I

A General Note on the Language ... 1-2

An Essay on How to Learn the Language ... 2v-6

The Orthography ... 6-8

Composition of the Syllables ... 8v-9v

The Way to Write and Pronounce the Letters ... 10-12v

The Declension of Nouns ... 13-18

The Conjugation and Formation of Verbs ... 18-52

Book II

The Rudimenta ... 52-59v

The Syntax ... 59v-66v

Book III

The Written Language ... 67-75

The Various Kinds of Names ... 75v-98v

Of particular interest in the context of Collado's grammar is the manner in which Rodriguez displays the verbal system. While the Ars Grammaticae presents the verbal system as a series of alterational rules to be applied to the base forms, the Arte Breve goes even further than the Arte to differentiate the formational rules from the conjugational displays. Rodriguez tries several devices to elucidate his material. For example, Charts A and B below represent very early attempts to use a bordered format for linguistic description.

In order to indicate the differences to be found between the descriptions presented by Rodriguez and Collado, I have extracted the formational rules from the Arte Breve and, setting aside only two short appendices dealing with variant forms, present them here in their entirety.

THE CONJUGATION AND FORMATION OF THE TENSES AND MOODS OF THE VERBS

All the verbs of this language may be reduced (se reduzem) to four affirmative and three negative conjugations. This is because the negative conjugation of the adjectival verb, which we discussed before,[^[11]] agrees with the second of the three conjugations; and the conjugation of the substantive verb Sǒrai, Sǒrǒ, or soro, which is an abbreviated form of Samburai, samburǒ[^[12]] both in the affirmative and the negative is reduced to the third conjugation. At this point we will treat the three affirmative and three negative ordinary conjugations of the regular personal verbs.[^[13]] Following this, and on account of its particular usage and formation, we will discuss the conjugation of the adjectival verb.

The verbs of this language do not change (naõ fẽ variedade) to show person and number as do those of Latin; rather, one form (voz)

is used for all persons, singular and plural. Number and person are understood according to the subject (Naminativo [sic]), or pronoun, which is joined to the verb. The moods of the verb, which in this language have distinct forms for the tenses, are indicative, imperative, conjunctive, conditional, and preterit participle. The remaining moods are made up of these forms joined to certain particles. Each mood has but three tenses which have distinct forms; these are preterit, present, and future. These forms are signified by the Japanese terms (vocabulos) Quaco, ghenzai, mirai. The preterit imperfect and pluperfect are made up of the present, preterit, and preterit participle together with the substantive verb, as will be seen below in the conjugations.

Concerning the formation of the tenses and moods of the verbs in general, one is reminded that to understand the actual root (raiz) and the natural formation of all the tenses and moods, both affirmative and negative, it is extremely important to take notice of the usage of the Goyn,[^[14]] which are the five vowels (cinco letras vogaes) in the syllables which are below each aforementioned formation; and that it is also important to understand Canadzucai,[^[15]] which is the way to write with Firagana as well as the way one joins together syllables, or letters, to form other words (palauras), while noticing which syllable is changed by which, what constitutes long, short, or diphthongal syllables, which combinations cause contraction (sincope), which cause augmentation (incremento) of the verb, whether one makes a syllable liquid (liquescit)[^[16]] or not, and how the tenses of the moods are written with the same Cana.[^[17]] The term Goyn, not only indicates the syllables, or Cana, which are transformed to others, such as Fa, Fe, Fi, Fo, Fu, which are changed to the closely related sounds Ba, Be, Bi, Bo, Bu and Pa, Pe, Pi, Po, Pu; but it also indicates another kind of change from one sound to another in the same order (ordem), as happens among the syllables Fa, Fe, Fi, Fo, Fu. Ba, Be, Bi, Bo, Bu. Ma, Me, Mi, Mo, Mu,

etc. where often by rule (regna) Ma is changed to Mi; or to the contrary Bu to Ba and Bi to Ba, and likewise for others. The greater part of the formation of the tenses of each mood is confined to such changes, as is clearly seen in the way one writes the tense forms with Cana. It is to this that another change belongs. That which exists among those syllables having a certain relationship and rapport between them, as Ma, Fa, Ba, Pa; Me, Fe, Be, Pe; Mi, Fi, Bi, Pi; Mo, Fo, Bo, Po; Mu, Fu, Bu, Pu; with Mu and V. Thus, what is written Vma in Cana is written Muma, and Mume written for Vme in order to conform more closely to its pronunciation.[^[18]] Also Mu is written for Bu[^[19]] so that all the harmony (armonia) in the formations of this language are contained in the rules for Goyn and Canadzucai. Those who are informed see, as native speakers, how the tenses are formed for any mood, and which letter, or syllable, must be changed to another to affect a formation. Concerning this matter there is a booklet[^[20]] which teaches Canadzucai, and the general rules on the subject. Teachers should have this booklet to teach more easily and advantageously those students who are learning Cana. Lacking a knowledge of Goyn and Canadzucai, some of the rules which until now have been used in the formation of verbs (some of which I have let remain as they were), are not the original and natural rules as are the Goyn.[^[21]] They are rather devices, some forming affirmative tenses and moods from negative forms and others forming them from yet other more remote sources, which appear to correspond to formational rules, but for which the proper rules are not known. The fact is that the affirmative as well as negative are formed from the affirmative, beginning with the root, as will be seen below.

Speaking in general of the formation of the verb, the forms of the indicative and imperative moods of all three conjugations are formed from the root of the verb. The rest of the tenses in the other affirmative moods are formed from either the indicative or imperative forms. In the same way, the negative indicative present is formed from the root of the verb and the other tenses of the indicative are formed from

the present form. The other negative moods are formed from the indicative forms.

Formation of the Tenses for the Indicative and Imperative Moods of the Verbs of the First Affirmative Conjugation

The final syllables of the roots of the first affirmative conjugation, by which the verbs conjugated here are known, and from which the tenses of the indicative will be formed, end in E, with the exception of the verb "to do," Xi, or Ii, with its compounds and certain other verbs which end in I. The verbs which belong to the first conjugation, are as follows [in Charts A & B].

The verb Xi "to do," with its compounds ending in Xi or Ii, follows the formation of the verbs of the first conjugation. Ii is Xi which has been changed (alterado) to Ii because it follows the letter N. Xi conforms to the rules for the syllables which are changed (se mudam) to others. Thus:

Many of these verbs have another, less used, form made by adding Ru to the root; e.g., Abi, abiru; Mochiy, mochiyru; xiy, xiyru. Among these are some that have only this second form and lack the first; e.g., Mi, miru; Ni, niru; Fi, firu; Cagammi, cagammiru; Ki, kiru "to dress," as distinct from Ki, kuru "to come"; and y, yru.

Chart A

[The Formation of First Conjugation Verbs Ending in E]

Chart B

Formation of the Optative, Conjunctive, and Conditional Moods, and the Participle

The optative mood does not have forms of its own but compensates for this in part by adding to the imperative certain particles which indicate desire, in part by adding to the future indicative particles which show regret for not doing something, and in part by circumlocutions with the conditional mood and certain particles, as will be seen in the conjugations.

The conjunctive mood has two sorts of proper forms. The first is the common and ordinary form ending in Eba, corresponding to the Latin cum. The other ends in Domo, corresponding to the particle "although (posto que)." The other verbs of this mood do not have their own forms, but are expressed by circumlocutions as we shall see.[^[24]]

The present tense of the first conjunctive is formed from the present indicative by changing the final Ru to Reba; e.g., Motomureba. For the preterit Reba is added to the preterit indicative; e.g., Motometareba. For the future the final Ru of the third form of the future indicative is changed to Reba; e.g., Motomeôzureba. For a second form of the future the syllable Rǒ is added to the indicative preterit perfect; e.g., Motometarǒ. This particle is Ran in the written language; e.g., Motometaran.[^[25]] An utterance (oraçam) does not end in this form, but must be followed by a noun.[^[26]]

The present tense of the second conjunctive is formed by changing the final Ru of the present indicative to Redomo; e.g., Motomuredomo. For the preterit Redomo is added to the indicative preterit perfect; e.g., Motometaredomo. Strictly speaking this form is Motomete aredomo, losing the E of the participle. Furthermore, Motometa, together with the other preterit forms in Ta is from Motometearu which is first elided to Motometaru and then by common usage (pratica) to Motometa. All of which is seen in its Canadzucai. For the future, the final Ru of the future indicative is changed to Redomo; e.g., Motomeôzuredomo.

The conditional mood, for the present tense, is formed by adding the syllable Ba to the root of the verb and Naraba or Ni voiteua to the

present tense form; e.g., Motomeba, motomuru naraba, and motomuruni voiteua. For the preterit, Raba, Naraba, or Ni voiteua are added to the indicative preterit; e.g., Motometaraba, which is in reality Motomete araba, motometa naraba, and motometani voiteua. For the future Naraba or Ni voiteua are added to the future forms; e.g., Motomeô naraba and motomeôni voiteua. The present tense forms are also used for the future.

Verbs of the First Conjugation That End in I

There are some irregular verbs ending in I which follow the formational rules of the first conjugation, both affirmative and negative. There are a precise number of them. Those which have been found to date are shown below. They are formed for the present indicative by changing I to Uru, for the preterit by adding Ta to the root of the verb, and for the future by adding long û, ûzu, or ûzuru to the same root. For the present conditional Ba is added to the root, for the preterit Raba is added to the preterit indicative, and for the future Naraba is added to the future indicative. For the present conjunctive the Ru of the present indicative is changed to Reba, for the preterit Reba is added to the same preterit indicative, and for the future the final Ru of the future is changed to Reba. All the other forms are formed as has been stated for the formation of the first conjugation. Thus:[^[27]]

Formation of the Verbs of the Second Conjugation

All the roots of second conjugation verbs end in I. There are eight final syllables for these verbs; i.e., Bi, Chi, Ghi, Ki, Mi, Ni, Ri, Xi. It is by these syllables that the verbs of the second conjugation (except for those mentioned above as being in the first conjugation) are recognized, and from which the tenses are formed.

The roots ending in the syllables Bi, Ghi, Ki, Mi, and Ri change the I to V for the present tense; e.g., Tobi, tobu; Coghi, coghu; Caki, caku; Yomi, yomu; Kiri, kiru.

Those ending in Chi change to Tçu for the present; e.g., Mochi, motçu; Cachi, catçu; Tachi, tatçu.

Those ending in Ni change to Nuru for the present; e.g., Xini, xinuru; Yni, ynuru.[^[28]]

Those ending in Xi change to Su for the present; e.g., Fanaxi, fanasu; Cudaxi, cudasu; Taraxi, tarasu.

For the preterit those ending Obi and Omi change to ôda; e.g., Yomi, yôda; Tobi, tôda; Yobi, yôda; Yorocobi, yorocôda. Tomi becomes tonda.

Those ending in Abi and Ami change to ǒda; e.g., Yerabi, yerǒda; Vogami, vogǒda; Yami, yǒda.

Those ending in Imi change to ûda; e.g., Najimi, najǔda; Nijimi, nijǔda; Ximi, xûda.[^[29]]

Those ending in Umi and Ubi change their endings to Vnda or in some instances ûda. While some have two forms others have only one form which is seen in use, the more general is Vnda; e.g., Musubi, musunda; Susumi, susunda or susûda; Nusumi, nusunda or nusûda; Sumi, sunda or sûda; Cumi, cunda only.

Those ending in Ebi and Emi change to Eôda; e.g., Sakebi, sakeôda; Sonemi, soneôda.

Those ending in Ghi change to Ida; e.g., Auoghi, auoida; Voyoghi, voyoida; Coghi, coida.

Those ending in Ni change to Inda; e.g., Xini, xinda; Yni, ynda.

Those ending in Chi and Ri change to Tta; e.g., Machi, matta; Cachi, catta; Tachi, tatta; Kiri, kitta; Chiri, chitta; Cari, catta.

Those ending in Ki and Xi change to Ita; e.g., Caki, caita; Faki, faita; Nuki, nuita; Todoki, todoita; Sosoki, sosoita; Saxi, saita; Fataxi, fataita; Maxi, maita or maxita; Coxi, coita or oxita. The following add Ta to the root; e.g., Moxi, moxita; Muxi, muxita; Fuxi, fuxita; Mexi, mexita.

The future can be formed in two ways. The first and more common way is to change I to ǒ, ǒzu, or ǒzuru; e.g., Yomi, yomǒ, yomǒzu, yomǒzuru; Yerabi, yerabǒ, etc.; Kiri, kirǒ; Xini, xinǒ; Auoghi, auogǒ.[^[30]] Those ending in Chi change to Tǒ; e.g., Cachi, catǒ, etc.; machi, matǒ. Those ending in Xi change to Sǒ; e.g., Mǒxi, mǒsǒ; Nagaxi, nagasǒ; Mexi, mesǒ; Coxi, cosǒ, etc. The other way, which is easy too, is to change the final V of the present indicative to ǒ; e.g., Yomu, yomǒ; Kiku, kikǒ; Mǒsu, mǒsǒ; Mesu, mesǒ. Those ending in

çu change to Tǒ; e.g., Tatçu, tatǒ; Catçu, catǒ. Those ending in Nuru change to Nǒ; e.g., Xinuru, xinǒ; Ynuru, ynǒ. This second rule seems to be more naturally in accord with the rules for the Japanese language.

The imperative changes the final I of the root to E. Those ending in Chi change to Te; e.g., Yome; Kike; Tamochi, tamote; Vchi, ute; Machi, mate.

The present conjunctive is formed by adding Ba to the imperative; e.g., Yomeba; Tateba.[^[31]] For the preterit, Reba is added to the preterit indicative; e.g., Yôdarebe.[^[32]] For the future the final Ru of the future indicative is changed to Reba; e.g., Yomǒzureba. The conjunctive in Domo is formed in the same manner; e.g., Yomedomo, yôdaredomo, yomǒzuredomo.

The conditional is formed from the future indicative by changing the ǒ to Aba; e.g., Yomaba; Tataba.[^[33]] The preterit is formed by adding Raba to the indicative preterit; e.g., Yôdaraba; Tattaraba.[^[34]]

The preterit participle is formed from the preterit by changing the A to E; e.g., Yôde; Kite; Tatte. The present participle, in Te, is formed by adding Te ['hand'] to the root of any verb. This is properly a substantive and thus governs the genitive as do the other substantives. It does not indicate tense; e.g., Yomite; Cakite; Machite, etc.

The negative present can be formed in two ways. The first, and that which accords with the rules for Canadzucai, is formed by changing I of the root to Anu or Azu; e.g., Corobi, corobanu, corobazu; Yomi, yomanu, etc.; Coghi, coghanu; Caki, cacanu;[^[35]] Kiri, kiranu; ini, inanu. Those ending in Chi change to Tanu; e.g., Tachi, tatanu. Those ending in Xi change to Sanu; e.g., Fanasanu. Another formation common to all is made with the future indicative by changing ǒ to Anu or Azu; e.g., Corobǒ, corobanu, corobazu; Yomǒ, yomanu, etc.; Coghǒ, coghanu; Cakǒ, cakanu; Kirǒ, kiranu; Inǒ, inanu; Tatǒ, tatanu; Matǒ, matanu; Fanasǒ, fanasanu. This rule is common to all three conjugations by changing the affirmative future indicative ǒ to Anu and the ô

and û to Nu or Zu;[^[36]] e.g., Todomeô, todomenu, todomezu; Saxeô, saxenu, etc.; Tateô, tatenu; Miû, minu; Yomǒ, yomanu; Tatǒ, tatanu; Fanasǒ, fanasanu; Narauǒ, narananu; Vomouǒ, vomouanu; Furuuo, furuuanu. For the second conjugation preterit, those in Nu are changed to Nanda: e.g., Yomananda. For the preterit participle Da is changed to De; e.g., Yomanande. For the second form of the negative participle, the Nu is changed to Ide; e.g., Yomaide, Corobaide, Tataide, Totonouaide. For the future the particle majij[^[37]] or mai is added to the affirmative present indicative; e.g., Yomumajij, yomumai; Matçumajij, matçumai.

Formation of the Third Conjugation and the Roots from Which the Tenses Are Formed

The final syllables of the third conjugation are the diphthongs Ai, Oi, Vi. By these syllables the verbs are known to belong to this conjugation, and from them the tenses are formed. The present indicative is formed by changing Ai to ǒ, Oi to ô, and Vi to û; e.g., Narai, narǒ; Vomoi, vomô; furui, furû. The preterit is formed by adding the syllable Ta to the present; e.g., Narǒta, Vomôta, Furûta. The future is formed by changing the final I of the root to Vǒ, vǒzu, vǒzuru; e.g., Narauǒ, narauǒzu, etc.; Vomouǒ, vomouǒzu, etc.; Furuuǒ, furuuǒzu, etc. The present imperative is formed by changing the final I to Ye; e.g., Naraye, Vomoye, Furuye.

For the present conjunctive Ba or Domo is added to the imperative; e.g., Narayeba, narayedomo; Vomoyeba, vomoyedomo; Furuyeba, furuyedomo. For the preterit Reba or Redamo[^[38]] is added to the indicative preterit; e.g., Narǒtareba, narǒtaredomo; Vomôtareba, vomôtaredomo; Furûtareba, furûtaredomo.

The present conditional is formed by changing ǒ of the future to Aba; e.g., Narauaba, Vomouaba, Furuuaba. The preterit is formed by adding Raba to the indicative preterit; e.g., Narǒtaraba, Vomôtaraba, Furûtaraba.

The negative present is formed by changing the I of the root to Vanu or vazu; e.g., Narai, narauanu, narauazu; Vomoi, vomouanu, etc.; Furui, Furuuana, etc. This form can also be formed from the future by changing the ǒ to Anu or azu; e.g., Narauǒ, narananu, etc. For the preterit the Nu is changed to Nanda; e.g., Narauananda. For the preterit participle the Da is changed to De; e.g., Narauanande. For the second form the Nu of the present is changed to Ide; e.g., Narauaide, Vomouaide, Furuuaide. For the future the particle Majii,[^[39]] maji, or mai is added to the affirmative present indicative; e.g., Narǒmajii, narǒmaji, narǒmai; Vomômajii, ji, or mai; Furûmajii, ji, or mai.

The verb Yy 'to speak' becomes Yû, yûta, yuǒ, yye, yuanu. Yei or yoi 'to become sick' becomes Yô, yôta, youǒ, yoye, yonanu. The substantive verb Saburai, which also belongs to this conjugation, becomes Saburǒ, saburauanu; and Sǒrai becomes Sǒrǒ, soro, sǒraite, sǒraye, sorouanu.

Rodriguez follows these formational rules with a full display of all the forms of the three conjugations. In his display he, like Alvarez before him, recapitulates the appropriate rules for each form. Collado nowhere presents his conjugational system as a paradigm but does, as we shall see, include a full complement of example sentences in his description, something which Rodriguez does not do in the Arte Breve.

Bibliography

In the examination of any portion of the Christian materials certain works are indispensable. Father Johannes Laures, S.J., Kirishitan Bunko (Tokyo, 1957) remains the basic bibliographic source for the study of all sources of the Christian Century, while Hashimoto Shinkichi, Kirishitan kyōgi no kenkyū (Tokyo, 1929) and Doi Tadao, Kirishitan gogaku no kenkyū (Tokyo, 1942) serve as indespensible guides to our understanding of the linguistic aspects of the field. A later contribution to the general bibliography has been made by Fukushima Kunimichi, Kirishitan shiryō to kokugo kenkyū (Tokyo, 1973).

The basic grammatical study of the period, based upon the shōmono materials, is Yuzawa Kōkichirō, Muromachi jidai gengo no kenkyū

(Tokyo, 1958). More closely related to the language reflected in the text is his "Amakusabon Heike monogatari no gohō," in Kyōiku ronbunshū (no. 539, Jan. 1929). An English treatment of the grammatical system of the period is to be found in R. L. Spear, "A Grammatical Study of Esopo no Fabulas," an unpublished doctoral thesis (Michigan, 1966). The phonology has been carefully analyzed by Ōtomo Shin'ichi, Muromachi jidai no kokugo onsei no kenkyū (Tokyo, 1963), with a valuable contribution made in English by J. F. Moran, "A Commentary on the Arte Breve da Lingoa Iapao of João Rodriguez, S.J., with Particular Reference to Pronunciation," an unpublished doctoral thesis (Oxford, 1971). This latter work presents an exhaustive examination of the phonological system reflected in the Arte Breve of 1620 within the framework of Berhard Bloch's phonemic theory.

Two lexical works have been used as basic references in this translation. The Vocabulario de Lingoa de Iapam (hereafter the Vocabulario) produced by the Jesuit Mission Press at Nagasaki in the years 1603 and 04. In a carefully annotated version by Professor Doi, under the title Nippo jisho (Tokyo, 1960), this work is the most important single source for the vocabulary of the period. The second work is the Dictionarium sive Thesauri Linguae Iaponicae Compendium (hereafter the Dictionarium) which is the companion piece to the present text. This dictionary has been carefully edited and cross-referenced by Ōtsuka Mitsunobu, under the title Koriyaado Ra Su Nichi jiten (Tokyo, 1966). In this form it has served as a constant aid to the translator in the determination of the proper glosses for the lexical items in the text.

The aforementioned Arte of 1604-08 by Rodriguez, has been the single most frequently used tool in the preparation of this translation.[^[40]] As the most significant influence upon Collado's work and the source for most of his material, both theoretical and practical, I have related the two works at every point in the translation. In its Japanese version by Professor Doi, Rodorigesu Nihon daibunten (Tokyo, 1950), this work has been invaluable in gaining a clearer understanding of many of the passages which might have otherwise been obscure.

Rodriguez' Arte Breve of 1620, while having no influence upon the preparation of the Ars Grammaticae, is nevertheless of fundamental

importance as a work against which Collado's treatment of Japanese grammar is to be judged. This shorter grammar is as yet to be fully translated into English—Moran having limited his study to the treatment of the phonology.

With respect to the text itself I have made this translation on the basis of the facsimile edition published by the Tenri Central Library in 1972 as part of its Classica Japonica series. Ōtsuka Takanobu, Koiyaado-cho Nihongo bunten (Tokyo, 1934) and its revised edition under the title of Koriyaado Nihon bunten (Tokyo, 1957) have served as invaluable aids at every step of the translation.

Ōtsuka's second edition is of invaluable scholarly importance because it contains a cross-reference to the Spanish manuscript from which Collado prepared the printed Latin edition as well as a concordance to the Japanese vocabulary.[^[41]] This translation attempts to supplement Ōtsuka's invaluable contribution by relating the Latin text of this grammar with Rodriguez' Arte.

Editorial Conventions

The Latin matrix of the text is printed in italic letters while the Japanese is in roman. For this translation I have reversed the convention. (In footnotes where the text is quoted the style of the original is followed.) In making editorial corrections in the Japanese material the corrected version is presented in brackets with periods to indicate the general location;

e.g., mairu mai queredomo [... qeredomo]

(The only exception to this rule is the correcting of a missing open o, q.v.) Sentences that have been taken from the Arte are indicated by the parenthetical recording of the leaf number of the citation immediately after the sentence;

e.g., xô tame no chôqui gia (22) 'it is....

Shorter sentences and specific words that in all likelihood have been taken from the Arte are not listed if they are to be found in the section elsewhere noted as being the source of the material covered. Any

significant alteration in the form of the source is noted. Since the Arte is numbered by the leaf, v is added to the number to indicate the verso.

All the corrections made by the errata (on page 75 of the text) have been applied to the text without notation unless the correction is itself in error.

The punctuation follows the text with the following exceptions;

1. In translating from Latin the English follows modern rules of punctuation.

2. Single quotes have been introduced into the text to mark glosses and translations.

3. In transcribing the Japanese citations any alteration of the original punctuation is noted.

4. The spacing of words in Japanese—a relatively casual matter in the text—has been regularized on the basis of the predominant pattern.

5. Two specific rules, based upon Collado's more or less consistent usage, are followed in the citing of verb forms:

a. In the most frequent citation of verbs, where the root form is followed by the present indicative ending, a comma is used;

e.g., ari,u; ague,uru; mochi,tçu

b. In an alternate form of citation, where the two forms are given in their entirety, a colon is used;

e.g., ari:aru; ague:aguru; mochi:motçu

Spelling and accentuation are treated in the following manner:

1. The ſ in all instances is represented by s.

2. The usage of v and u has been regularized: the v serves as the consonant; and u as the vowel, semi-vowel, and orthographic symbol; e.g., vaga, uie, quan, and agueta.

3. The predictable nasalization—marked by a tilde in the text—has not been included in the translation unless the presence of nasalization is morphologically significant; e.g., tobu:tõda.

4. The accent grave—which appears in no discernible pattern—is not transcribed in the translation.

5. The accent acute is used in the translation to mark the long ú [u:] and the long, open ó [ɔ:], in those places where the length is marked by Collado. Since the most frequent typographical error in the text is the failure to mark the presence of these long syllables, I follow the convention of correcting the absence of this feature in the Latin text by using the inverted caret in the translation. Thus, the appearance in the translation of mósu indicates that Collado recorded the length of this word, either by an accent acute (e.g., mósu), or an inverted caret (e.g., mǒsu). The appearance of mǒsu indicates that he did not, and that its absence is being corrected. The form mǒsu in the translation is therefore the shorthand equivalent for what would more regularly be mosu [mósu].

6. The circumflex, which indicates the long, closed ô [o:], is corrected as other errors by placing the corrected version of the item in brackets; e.g., roppio [roppiô].

ARS
GRAMMATICAE
IAPONICAE
LINGVAE

IN GRATIAM ET ADIVTORIVM
eorum, qui prædicandi Euangelij causa ad
Iaponiæ Regnum se voluerint conferre.

Composita, & Sacræ de Propaganda Fide Congregationi
dicata à Fr. Didaco Collado Ordinis Prædicatorum
per aliquot annos in prædicto Regno
Fidei Catholicæ propagationis
Ministro.

ROMÆ,
Typis & impensis Sac. Congr. de Propag. Fide.
MDCXXXII.
SVPERIORVM PERMISSV.

A
Grammar
of the Japanese
Language

FOR THE SAKE AND HELP
of those who wish to go to the Kingdom of
Japan to preach the Gospel.

Composed and dedicated to the Blessed Congregation for the
Propagation of the Faith by Brother Didico Collado,
O.P., who was for many years in that Kingdom
as a Minister for the Propagation of
the Catholic Faith.

Printed by the Blessed Congregation for the Propagation of the Faith.
1632
BY PERMISSION OF THE SUPERIORS.

It may be published if it please the Most Reverend Father,

Master of the Holy Apostolic Palace.

For the Archbishop of Umbria. The Vicar General.