107
THE CHARACTER OF THE INDO-EUROPEAN MOODS 811.1/.2 GONDAJ characte 196147 WITH SPECIAL REGARD TO GREEK AND SANSKRIT By J. Gonda 1956 aTTO HARRASSOWITZ . WIESBADEN

Gonda.1956.The Character of the Indo-European Moods.pdf

Embed Size (px)

Citation preview

Page 1: Gonda.1956.The Character of the Indo-European Moods.pdf

THE CHARACTER

OF THE INDO-EUROPEAN MOODS 811.1/.2

GONDAJ characte

196147

WITH SPECIAL REGARD TO GREEK AND SANSKRIT 'i¡'isT:>~

By

J. Gonda

1956

aTTO HARRASSOWITZ . WIESBADEN

Page 2: Gonda.1956.The Character of the Indo-European Moods.pdf

JI., l'

Copyright 1956 by Otto Harrassowitz, Wiesbaden

AlIe Rechte vorbehalten

PllOtomechanische oder Photographische Wiedergaben jeder Art nur mit ausdrücklicher Genehmigung des Verlages

Printed in Germany

Gesamtherstellung: Hnbert & Co., Gottingen

~------------------""".l"'"

Contents

Abbreviations ...................................................... :. VII

1. Introduetion .................................................... 1

II. Preliminary observations on some verbal forms in non l.-E. languages .... " 10

III. Tense and primitive thought ....................................... 23

IV. The Injunetive . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 33

V. The essential funetion of the Optative . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 47

VI. The funetion of the aneient l.-E. Subjunetive ......................... 68 Exeursua on the form of the Vedie subjunetive . . . . . . . . . . . . . . . . . . . . . .. 109

VII. Additional eritieism ofMiss Hahn's views of the optative and subjunetive moods 117

VIII. The funetion of the so-ealled modal partieles .......................... 132 Exeursus 1 .................................................. " 147 Exenrsus 11 ................................................... 147

IX. The eonjunetion sl and the eonditional sentenees ....................... 149

X. The partiele *me ................................................. 197

Seleet Bibliography .................................................... 205

Index ...... '" ...... " ..... " ................................... " " 207 Index of snbjeets ................................................. " 207 Seleet glossary ....•................................................ 209 Index of names . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 210 Seleeted list of passages quoted ........................................ 210

Page 3: Gonda.1956.The Character of the Indo-European Moods.pdf

p , .

Abbreviations

Afr. Stud. = Mrican Studies, Johannes-burgo

Ahd. = Althochdeutsch. Ai. A. = Aitareya-ara:r;tyaka. Ait. Br. = Aitareya-brahma:r;ta. A. J. P. = American J ournal of Philo­

logy, Baltimore. Anc. = ancient. Aun. Bhand. Oro Res. Inst. = Annals of

the Bhandarkar Oriental Research In­stitute, Poona.

Aor. = Aorist. A. Or. = Acta Orientalia, Leiden. ApMB. = Apastambiya-mantra-brah-

ma:r;ta. ApS. = Apastambiya-Srauta-sütra. Arab. = Arabic. Aram. = Aramaic. Arc. = Arcadian. Arj. Wiw. = (Old Javanese) Arjunawi-

waha. Ass. = Assyrian. Asv. Bc. = ASvaghoi¡la, Buddhacarita. Asv. G. S. = Asvalayana-grhyasütra. ASv. Saund. = ASvagho¡;¡a Saundara-

nanda. Ath. = Atharva (-Veda). Att. = Attic. Av. = Avestan. BarU. or BrhArUp. = Brhadara:r;tyaka­

upanii¡lad. Bezz. Beitr. = Bezzenberger's Beitrage

zur Kunde der indogermanischen Sprachen, Gottingen.

Bhag. Puro = Bhagavata-pura:r;ta. BhY. = (Old Javanese) Bharatayuddha. Bim. = Bimanese. Boeot. = Boeotian. Brhatk. = Brhatkatha (slokasarpgraha). B. S. L. = BuIletin de la Société de Lin-

guistique, Paris. Chino = Chinese. Ch. SI. = Church Slavonic. ChUpo = Chandogya-upanif¡!ad. Class. Quart. = Classical Quarterly, Lon­

don. Comm. = commentary.

Cypr. = Cyprian. Dar. Bis. A. = Darius, Bisitun inscrip-

tion A. Dasakum. = Dasakumaracarita. Divyav. = Divyavadana. Dor. = Doric. El. = Elean. EngI. = English. Eur. H. F. = Euripides, Hercules Fu.

rens. Fr. = fragmento Fr. = French. Fut. = future. Ga. = Gatha. Gaut. = Gautama-srnrti. Germ. = German. Gop. Br. = Gopatha-brahma:r;ta. Goth. = Gothic. Gr. or G. = Greek. Hariv. = Harivamsa. Harv. Stud. = Harvard Studies, Cam­

bridge Mass. Hebr. = Hebrew. HirGS. or H. G. = Hira:r;tyakeSi-grhya-

sütra. Hitt. = Hittite. Idg. = indogermanisch. 1. E. = Indo-European. l. F. = Indogermanische Forschungen,

Berlin. 1. G. = Inscriptiones Graecae. Imper. = imperative. Impf. = imperfecto IN. = Indonesian. Ind. = Indian. Ind. = indicative. Inf. = infinitive. Inj. = injunctive. Ion. = Ionic. Ita!' = Italian. J ahrb. f. Phil. = J ahrbücher für Philo­

logie, München. J. Am. Oro Soco = Journal of the Ame­

rican Oriental Society, New Haven, Conn.

Jaim. Br. or JBr. = JaimÍlliya-brah­ma:r;ta.

j

Page 4: Gonda.1956.The Character of the Indo-European Moods.pdf

VIII Abbreviations

J. R. A. S. = J ournal Royal Asiatic So­ciety, London.

J1JpBr. = J aiminiya -upani¡;;ad -brah-maI,la.

KiU. = Kalidása. Kál. Kum. = Kumarasambhava. Kal. Malav. = Málavikág~limitra. Kál. Sak. = Sakuntalá. Kal. Urv. = Urvasi. Kathás. = Kathasaritsi1gara. Kau¡;;. Br. 01' KBr. = Kau¡;;itaki-bráh­

maI,la. KS. = Kathaka-samhita. KZ. = K.;mn's Zeitschrift, Zeitschrift

für vergleichende Sprachwissenschaft, Gottingen.

Lat. = Latin. Lesb. = Lesbian. lit. = literary. Lith. = Lithuanian. Mac. = Macassar. Maitr. S. 01' MS. = MaitrayaI,la-sa:q:iliitá. Mal. = Malay. Mbh. = Mahábhi1rata. Med. = medieval. Mém. Soco Ling. = Mémoires de la So-

ciété de Linguistique de Paris. Mod. = moderno Mrcch. = Mrcchakatika. Mt. PUl'. = Matsya-puraI,la. Mus. Helv. = Museum Helveticum, Ba-

sel. N. H. G. = New High German. N. W. Gr. = Northwest Greek. O. Ch. SI. = Old Church Slavonic. O. E. 01' O. Engl. = Old English. O. H. G. = Old High German. O. Lat. = Old Latin. O.N. = OldNorse. O. Pers. = Old Persian. O. Sax. = Old Saxon. PaI,l. = PaI,lini. Pañc. = Pañcatantra. PBr. = Pañcavimsa.brahmana Pers. = Persian. . . . Petr. Dict. = Petrograd Dictionary, i. e.

O. Bohtlingk-R. Roth, Sanskrit.Wor. terbuch, Petrograd.

PoI. = Polish. Preso = presento Ps. Aug. Serm. = Pseudo-Augustinus,

Sermones. Ram. = RámayaI,la. l;tg-l¡tV. = l;tg(-Veda),l;tg-veda.saIl1hita. Roczn. Orjent. = Rocznik Orjentali-

styczny, Lwów, Poland. Russ. = Russian. Scand. = Scandinavian. S. Cr. = Serbo-Croatian. Sem. = Semitic. Skt. = Sanskrit. Slav. = Slavonic. Sog. = Sogdian. SV. = Samaveda. Syr. = Syrian. TA. = Taittirlya-araI,lyaka. TaittS. 01' TS. = Taittirlya.sa:q:iliitá. TB. 01' TBr. = Taittiriya-bri1hmaI,la. Thes. Ling. Lat. = Thesaurus Linguae

Latinae, München. T. L. V. = Bijdragen tot de Taal-, Land­

en Volkenkunde, Den Haag. Univ. of Cal. Cl. Phil. = University of

California, Classical Philology. V. 01' Vendo = Vendidi1d. VádhS. = Vadhüla sütra. Vaikh. Sm. S. = Vaikhanasa.smi1rta.

sütra. Varah. Yog. = Vari1hamihira, Yoga.

sütra. Ved. = Vedic. Vino = Vinaya. Vol. subj. = voluntative subjunctive. VS. = Vajasaneyi-sa:q:iliita. Vulgo = Vulgata. Xen. Aponm. = Xenophon, Apomnemo­

neumata. Xen. R. Eg. = Xenophon, De re eque.

stri. Y. = YaSna. Yt. = Yast. ZDMG. = Zeitschriftderdeutschenmor­

genlandischen Gesellschaft, Leipzig. Wiesbaden.

l. Introduction

Although the theory of the Indo-European moods as established by Delbrück1 ) was in substance subscribed to by many scholars, dissentient views of the fundamental idea expressed by the modal categories or of the interrelation of their functions were always pronounced. Whitney 2) for instance questioned the correctness of Delbrück's thesis that the difference between subjunctive and optative originally had anything to do with a distinction between willing and wishing as was held by the German author. Goodwin 3) maintained that there is no element of will in the subjunctive expressing futurity, nor an element of wish in the potential optative, futurity and potentiality being according to Delbrück only secondary functions of these moods. The American scholar felt justified in regarding the subjunctive as originally and essentially a 'future', and in considering the optative to have been a weaker sub­junctive. A later period of Indo-European studies saw the opposition of Hirt 4), who attempted to demonstrate that the prevailing theories of the subjunctive which essentially are founded on Delbrück's publi­cations are "entschieden falsch" and should be replaced by the hypo­thesis that the subjunctive originally was a future and that the original function of the opto was to express "das Zeitlose und die Gegenwart". Holger Pedersen 5 ) even went so far as to deny the existence of the former mood in Original Indo-European. His compatriot Jespersen6 )

presented the view that all attempts at finding out the fundamental notions attached to these categories were failures.

1) See B. Delbrück, Syntaktische Forschungen, I, Halle 1871, p. 13; IV, Halle 1879, p. 117; V, Halle 1888, p. 302; Vergleichende Syntax, II, Straf3burg 1879, p. 302; Neue Jahrb. f. d. klass. Altertum 1871, p.326.

2) W. D. Whitney, American Journal of Philology, 13, p.294. Cf. also A. Bergaigne, De coniunctivi et optativi ... vi antiquissima, Paris 1877, p. 41ff.; 57ff.

3) W. W. Goodwin, Syntax of the moods and tenses of the Greek verb, London 1889, p. 371ff.

4) H. Hirt, Indogerm. Gramm., VI, Heidelberg 1934, p. 271; 278ff.; VII (1937), p.147ff. ,

5) H. Pedersen, Tocharisch voro Gesichtspunkt del' idg. Sprachvergleichung, Copenhagen 1941, p. 192f.

6) O. Jespersen, Language, ch. XIX, § 11. 1 Gonda, Indo·European moods

Page 5: Gonda.1956.The Character of the Indo-European Moods.pdf

2 The Indo-European moods

Recently, Miss Hahn 1 ) devoted a new book to this intricate subject, developing the thesis that the l.-E. subjunctive and optative actually were, both of them, futures, "and nothing else", and thus often inter­changeable, but with the general distinction that the former tended to be used for more vivid futurity, and the optative for more remote futurity 01' potentiality. By writing this book Miss Hahn has no doubt done useful work, showing the untenability of sorne standpoints defended by former scholars and forcing the reader to decide on his position towards sorne of the most vexing problems of Indo-European linguistics. Despite the critical acumen displayed by her and despite the many critical observations made on many pages, the learned authoress has, however, not succeeded in convincing me of the correctness and finality ofher expositions 2 ). Her book has, on the contrary, strongly fortified me in the view that the 'orthodox' opinions on these moods are, in sub­stance, nearer to the truth than the new theory advocated by her, part of which was already proposed by her compatriot Goodwin. On the other hand, a confrontation of her views with those which in many respects may be called traditional convinced me that the latter are in urgent need of radical revision.

Perusing sorne of the more 01' les s recent leading publications on tenses and moods in the early stages of l.-E. languages, one becomes, indeed, much impressed by the traditionalistic way in which many authors often continue discussing this subject. The moods are for instance still defined as categories intimating the state of mind, 'mood' 01' disposition of the person speaking 3). The distinction made by the Ancients, on which this definition is founded, is still subscribed to and praised: the subj. and opto make the Ota-&éa8Lr; 'ljJvxijr; of the speaker known to the hearer; in contradistinction to the indicative which enables the speaker to make objective statements, they are, it is often held, subjective in

1) E. Adelaide Hahn, Subjunctive and Optative. Their origin as futures, New York 1953. In chapters I and II of this book the theories of other scholars are considered.

2) I refer to my review in the Museum, vol. 59, Leiden 1954, 185ft'. 3) Cf. e.g. W. Streitberg, Gotisches Elementarbuch5, Heidelberg 1920, p.204.

'Die Modi bezeichnen einen Gemütszustand des Sprechenden'; J. B. Hofmann, in Stolz-Schmalz, Lateinische Gramm.5, München 1928, p.565 'geistige Haltung des Sprechenden zum Verbalvorgang'; W. Vondrák, Vgl. Slav. Gramm., II, Gottingen 1928, p.394; M. Schonfeld, Historische grammatika van het Neder­lands4, Zutfen 1947, p. 144; O. Jespersen, Modern Engl. Grammar, VII, Copen­hagen 1949, p. 623. Cf. also Hahn, o.c., p. 4, n. 10; Ch. Bally, in 'Cahiers Ferd. de Saussure', II, Geneva 1942, p. 3.-0ther definitions are not always better: that given by L. H. Gray, Foundations of language, New York 1933, p.208 ("denotes the manner in which the action or state is performed or e:x;ists") is, for instance, unsatisfactory.

Introduction 3

character 1). That this opinion is due to a misunderstanding may appear from the following considerations. If the subj. and the opto are described as being subjective in character 01' as expressing a subjective disposition, then it must be said that the function of the indicative is similar. The indo expresses that the person speaking visualizes the process-i. e. the act, action, event, in short any'idea' expressed by a verb-as real 01' actual. As a rule the speaker is not conscious of this 'subjective> character of his statements expressed by means of verbs in the indicative. Very often his visualization is in accordance with objective facts: this man is caUedPeter; this book is written by X.; the dor;¡ barks. But this is not always the case. If 1 say "It is cold today" it is always possible to meet a person who answers "Do you think so 1" 01' who pronounces the opposite opinion. A statement can be objectively incorrect, a fact not a:lways realized in former centuries when logically correct speech was consldered to be the main, 01' even the sole, object of linguistic science. A wish, a desire, an appreciation are, moreover, often couched in the words of an objective statement. Thus Schwyzer-Debrunner are in my opinion right in observing that "auch del' gewohnliche Indikativ insofern sub­jektiv (ist), als el' nur angibt, was del' Sprechende als wirklich aufgefaBt wissen will" 2).

There seemsalso to be room for the observation that those scholars who adhere to the view that the choice of a mood is determined by the attitude of the mind of the speaker towards the contents of the sentence are forced to admit that in sorne cases this choice is determined by the character of the clause itself and its relation to the main clause on which it depends. 3 )

Other points which do not fail to strike the reader of the extensive literature on the moods in the ancient l.-E. languages concern the inclination to attach considerable value to Delbrück's distinction between will and wish as central 'notions' expressed by subj. and opt.-a topic to which we have to revert-, and the almost stubborn search for a chronological relation between the various functions assumed for both

1) Cf. J. Humbert, Synta:x:e grecque, Paris 1954, p. 113; P. Chantraine, Gram­maire homérique, II, Paris 1953, p. 205; A. Meillet-J. Vendryes, Traité de gramo maire comparée des langues classiques, Paris 1948, p. 19If. - It may be paren­thetically observed that not only authors belonging to earlier periods, like Gott­fried Hermann (according to whom the indicative was the 'image' of reality, the subjunctive and optative that of possibility) were e:x:posed to the danger of ~eco~­ing mixed up with metaphysics; cf. e.g. F. Thomas, Recherches sur le subJonctlf latin, Paris 1938, p. XIV.

2) E. Schwyzer-A. Debrunner, Griechische Grammatik, II, München 1940, p. 303. Cf. also R. Kühner-B. Gerth, Ausf. Grammatik der griechischen Sprache, Hanno­ver-Leipzig, II, 1, p. 201.

3) Cf. e.g. Jespersen, The philosophy of grammar, London 1924, p. 313.

1*

Page 6: Gonda.1956.The Character of the Indo-European Moods.pdf

4 The Indo-European moods

subjunctive and optative. After studying the various uses of these moods and in classifyingthem in such sub-groups as cupitive and potential optative, voluntative and prospective subjunctive, many students of l.-E. linguistics attempted to transpose theproblem offered by the simultaneous existence of these °classes' into a question of priority. It is, however, clear that in such cases as the one under discussion, in which the main uses of the categories are already known to aH ancient docu­ments, the investigation of the °historical development' of these uses becomes a question of ° glottogonic speculation'. The comparative method proper does not allow us to go beyond the conclusion that a variety of functions like that shown by the ancient texts was already a feature of pre-historic, and common, Indo-European 1). The answers given to the question which function must be regarded as more original, which as younger, derived, 01' secondary, were accordingly often of a general, philo­sophicaJ, 01' psychological rather than a linguistic nature. Whereas for instance those scholars who advocated the priority of the cupitive optative argued, like Mutzbauer 2) that "eine WeiterentwickIung del' Wunschvorstellung nur davon ausgehen (kann), ob das Individuum seinen Wunsch als mehr oder minder erfüllbar fühlt. 1st es sich dabei klar bewuBt, daB derselbe erfüllbar ... ist, so tritt die Vorstellung del' M6glichkeit in seinem BewuBtsein in den Vordergrund. Dadurch wandelt sich die Wunschvorstellung in die Vorstellung del' M6glichkeit", those who preferred the opposite view maintained, like Hammerschmidt3), that so rare a conception as wish would hardly have needed a special formo Humbert 4

) who in his recent book on Greek syntax subscribes, "for various reasons", to the former opinion, attempts to uphold this by contending that "le voeu est quelque chose de primitif et de tout concret, de vraiment spontané", whereas, he adds, "il faut une raison déja exercée pour distinguer, parmi les souhaits, ceux qui sont possibles de ceux qui ne le sont pas". My comment would be similar to that made by Miss Hahn 5): 1 fail to see how such speculations on the assumed psychology of primitive man can help us any further.

The linguistic arguments adduced to solve the problem of the chrono­logical order of the various uses of the moods do not seem to be conclusive

1) Cf. also Delbrück, Altind. Syntax, p. 302. 2) C. Mutzbauer, Die Grundbedeutung des Konjunktiv und Optativ und ihre

Entwicklung im Griechischen, Leipzig-Berlin 1908, p. 144. See also Delbrück, S.F. I, p. 12f.

3) H. Hammerschmidt, Über die Grundbedeutung von Konjunktiv und Optativ, Erlangen 1892, p. 25.

4) Humbert, O.C., p. 117. 5) Who, however, rejects the good with the bad in refusing to learn anything

from a careful and well-founded examination of non-l.-E. phenomena and of the results of ethno-psychological research.

Introduction 5

either. Brugmann 1), for instance, drew attention to such semantic shifts as that of the Germ. miJgen 1) "imstande sein, k6nnen", 2) "may, like to" and of Fr. pouvoir in puisse le ciel donner ... , but as long as we do not know whether the °potentiality' expressed by the prehistoric optative was similar to the idea originally implied in these verbs-and 1 for one llave grave doubts about this similarity-, such random illustrations have little value, the less so as such ideas as are expressed by ° can' , expressing ability, and °may', implying permission, possibility, incer­tainty, wish, 01' hope, are very commo~:r, i;lterminl?led in their.li~l?~istic expression. Humbert's argument ~hat .Sl l~ ~otlOn ,d~ 'p0sslbl~~te est secondaire, on s'explique assez qu elle art du etre preClsee par ay (~8), a la différence de l'optatif de voeu" may be refuted by the observatlOn that there are historicaHy attested instances of the opposite development. Even those authors who, like Brugmann and Schwyzer 2 ), are inclined to a certain scepticism, believing that either of the two main functions of tlle optative may represent the starting-point, express themselves in terms of alternatives 3).

Another point concerns the vague terminology used in discussing the categories and phenomena at issue 4). It is not our intention to ex?ati~te upon the time-honoured practice of ~uthors of .gran::mars of a~tnbutmg to foreign languages a variety of Latm categones wlthout askmg them­selves whether these °moods' and °tenses' really correspond to the subjunctive future, etc. ofthat ancient l.-E. language. Without insisting once more on the necessity to distinguish between real moods (cpéeOl) and combinations with auxiliary verbs (he may bear) used in modern languages to translate the ancient moods, we should like to make a few remarks in connection with the application of the term moods to entities other than the indicatives, subjunctives etc. in those languages in which real moods are living categories. "Die Beziehungen, die die Modi ausdrücken, k6nnen auch durch andere Mittel, und zwar in gr6Berer Fülle und genauer ausgedrückt werden, durch Hilfs,:"erba wie nh~. wolle~, sollen, müssen, konnen, dürfen, oder durch Adverbra und Partlkeln Wle

nhd. wohl, vielleicht ... Auch Sprachen, die formal mehrere Modi unter­scheiden, k6nnen solche Hilfsmittel benutzen, z.B. griech. Xe1], OVl'aa{}al,

1) K. Brugmann, Grundri13 der vergl. Gramm. d. indogerm. Sprachen, Ir, 3, StraI3burg 1916, p.857.

2) Brugmann, o.c., p.835ff.; 857f.; Schwyzer-Debrunner, o.c., p.309f. and 319f.

3) See e. g. also the recent Homeric Syntax by Chantraine (see aboye), p.212f. ,

4) The following remarks may not be superfluous, although Jespersen's criticism (Philosophy of Grammar, p.315) already pointed out part of the mistake (cf. however p. 320 !). .

Page 7: Gonda.1956.The Character of the Indo-European Moods.pdf

6 The Indo-European moods

oú, lO'W¡; . .. " In these words Schwyzer's Greek Syntax 1) express es an opinion which seems to be shared by many other authors: the auxiliaries are equivalent to the modal verb forms. Rence also the frequent state­ment that in a specified language or document a subjunctive or optative is periphrastically formed: "in Medieval Dutch the subjunctive is often expressed by a circumlocutional moeten, mogen or willen"2); in Sangirese, an Indonesian idiom, the optative-which is explicitly stated to be alien to the language-is said to be paraphrased by a word malowo n:e~~~n~ :'being inc!i~ed ,~o";~) such vocables as, for instance, the Malay ben glvmg, permlttmg ; btar "permitting; allowing, letting"; harus "fitting, right, proper"; patut "fit, proper"; baik "good, well"; mari "hither, come", ta dapat tiada "inevitably" are described as modal auxiliaries 4

). It would appear to me that the peculiarity of certain so­called auxiliary verbs in modern Western languages to fulfill a modal function and to cover, in translating, the ancient Greek, Sanskrit, or Latin moods has led many scholars to overlook the original difference in function between moods proper and 'auxiliary verbs. Making a, distinction between morphological moods-thus in English the oppo­sition 'real' (he is here): 'unreal' (il he W61'e here)-and syntactic moods­such as he will write, shall write, can write, must write, may write _5) does not add very much to our understanding of the mutual relations between both phenomena.

If we may describe the verbal category of mood (such as it appears in Greek or Sanskrit) as a means of intimating the speaker's view or conception of the relation of the process expressed by the verb to reality, it will be clear that the main distinction made is between what the speaker puts forward as fact (whether it be true or not) and what he does not regard as such 6). If he wishes to mention a process which in his opinion is a fact or actuality, he uses the indicative, if he wishes to put forward a process as a contingency the ancient Greek used the optative etc. On t~e other hand, the 'auxiliaries'-or semi-auxiliaries as they are sometlmes called-a considerable variety of which is often available in ancient and modern languages, help to express the speaker's visual­ization of the relation of the subject to a specified process as far as the subject's qualifications or disposition with regard to the process are

1) Schwyzer-Debrunner, o.c., p.304. 2) Thus F. A. Stoett, Middelnederlandsche Spraakkunst, Syntaxis, The Hague

1923, p. 210. ., 3) N. Adriani, Sangireesche Spraakkunst, Leiden 1893, p. 185. 4) See e.g. C. Spat, Maleische Taal4, Breda 1920, p.212. I also refer to my

paper 'De zg. hulppraedicaatswoorden in Maleis en Javaans', Bijdr. Taal-, Land­en Volkenkunde, 102, The Hague 1943, p. 95ff.

5) Thus L. Bloornfield, Language, London 1935, p. 273. 6) Cf. also W. J. Entwistle, Aspects of language, London (1953), p. 218.­

In the aboye definition 'mood' is taken in a narrow sense.

Introduction 7

concerned. These verbs can themselves occur in any mood. They can also be used absolutely, i. e. without a verbal complement: P 489 el O'V ys -&v,up I 0'0 1:{}éAst¡; "if you in your heart are willing". As, moreover, the meanings and connotations expressed by these 'semi-auxiliaries' are as a rule rather especial' the tir. U}éAsw for instance covering such English terms and phrases as "to be willing, to consent, to wish, to care, to be naturally disposed, etc."-,the difference in function between this class of verbs and the moods is considerable.

Rowever, the use of a subjunctive or optative resulting from the speaker's 'intention' not to put forward the process as fact is often determined by his 'conviction', belief, or assumption that the realization of the process is desirable, possible, necessary. In P 151 Ilare6xAq> fiero;; x6/-tr¡v onáO'at/-tt rpée8O'<&at the function of the optative might be explained as amounting to something like this: "1 am of the opinion that it would be well if 1 give this lock to P .... " But although it has rightly been translated by "1 would fain give unto the warrior P. this lock to fare with him" (Murray), it is not equivalent to a sentence containing the phrase S<&éAW onáO'at. Whereas T 187 ravra o' sywv S<&éAW o/-t60'at means "this oath 1 am ready to swear", the attention being drawn to the explicit statement of the subject's willingness and readiness, the optative in the former quotation only indicates that the speaker visualizes the process as contingent, the context and the situation enabling the hearer to make inferences about the speaker's psychical motives for expressing himself in this way. It is even possible to combine a form of an 'auxiliary' and a real mood: Lat. volo lacias; Plaut. Caso 559 veniat velim. So the absolute equalization of the moods and these 'auxiliaries' seems to resolve itself into a confusion between linguistics and psychology.

Rowever, the phrase 'auxiliary' + infinitive, being explicit, personal, vigorous, sometimes deferential or specialized in connotation, and, generally, stating facts rather than referring to the subtle distinctions involved in the modal forms, was not long in entering into competition with the moods proper 1). In ORGerm scal and mag are said to serve as a 'periphrasis' 2) of the so-called subjunctive: Nib. 506, 5 ez mac wollüge wesen; Otfr. Lud. 25 thes scal 61' gote thankón: cf.26 thes thankó. But there is oftEm a difference in meaning, and scal can also verge towards a temporal function. Cf. also Mod. Germ. wer mag das sein? 'insteadof' wer ist das? and similar constructions. As a verb of complete predication the Eng. may in ancient texts meant "to have power or influence, to

1) Occasionally the identity of a subjunctive and an indicative form has been an important factor in the spread of the 'periphrastic construction' : Germ. mogen wir glücklicli heimkehren.

2) It is of course incorrect to speak of 'Umschreibungen' of a mood (thus e. g. W. Wilmanns, Deutsche Grammatik, lII, 1, Stra13burg 1906, p. 216).

Page 8: Gonda.1956.The Character of the Indo-European Moods.pdf

8 The Indo-European moods

be strong, to prevail"; it could also express possibility, opportunity, absence of prohibitive conditions etc. Hence it carne to be an element in the well-known periphrasis, expressing, in final clauses, in dependen ce on verbs oí wishing, fearing etc., etc. the idea formerly expressed by the subjunctive which apart from such more or less established phrases as God bless you is now of a distinctly literary character. The original sense of the periphrastic construction which is not lost is, however, until the present time often recognizable. The Dutch kunnen, which as a °Vollverb' means "can, be able, capable, competent, skilful, have the power to, be in a position to, be possible", can as an ° auxiliary' correspond to certain uses of the Engl. may: 't kan zijn "it may be" ; hij kan de trein gemist hebben "he may have missed the train"; ik kan wellaat thuiskomen "1 may come home late" (mark the adverb wel); 't kan vriezen, 't kan dooien, lit. "it might freeze, it might thaw", i. e. "every °may be' has a °may not be' ". But there are cases in which 1 am scarcely able to make out in which function kunnen is used: als jij deze opmerking maakt, dan kan hij zeggen can answer to "if you make this remark, then he could say 01', might say: he is in a position to say" as well as" ... , he may say". It is from those cases in which the original sense of the auxiliary and the idea of possibility, probability, uncertainty, hypothesis, expectation etc. involved in the periphrastic construction are fairly equivalent that the latter must have started1).-It may be remembered that in the Modern Indo-Aryan languages the ancient modal distinctions are, generally speaking, likewise lost 2). Whereas the Middle-Indian optative for instance is a continuation of the ancient category, Bengalí etc. have replaced it by special phrases and constructions containing an indicative 01' consisting of periphrases with an auxiliary.

An attempt briefly to characterize these so-called auxiliaries cannot omit the well-known fact that they often develop into a temporal function rather than a modal. Cases are not rare in which an initial inclination towards modality ends in the fulfilment of a temporal task. The dicere habeo occurring in some of the writings of Cicero (S. Rose. 100; Nat. Deor. 3, 93; Ep. Att. 2, 22, 6 de re publica nihil habeo ad te scribere) expresses possibility; in later Latin it serves as a °periphrastic future': Aug. Tract. in Jo. 4, 1, 2 tempestas illa tollere habet totam paleam de area, although such connotations as are expressed by our can 01' shall are long in disappearing 3). The history of the Engl. will and shall, of Fr. vouloir,

1) Cf. O. Behaghel, Deutsche Syntax, n, Heidelberg 1924, p.243, quoting Schupp, Streits. 2, 54 nooh wolt ioh lieber in Dorn und Disteln baden, als mit falsohen Zungen seyn beladen. '

2) See e.g. S. K. Chatterji, The Origin and development of the Bengali language, Calcutta 1926, p.901.

3) The reader may be referred to Ph. Thielmann, Archiv f. lat. Lex., n, p. 48ff.; 157ff.

Introduction 9

devoir, of Gr. (B)f)éJ...w and Mod. Gr. {}á etc. is too well known to require special commentl).

1) See e. g. J. Wackernagel, Vorlesungen über Syntax:, 12, Basel .1926, p. 193f:f.­Wackernagel's definition ofmood (Vorl., I, p. 210) may lead to mlsunderstandmg: see Hirt, Indog. Gramm., VI, p. ~51. .. . _ _.

It would in my opinion be advlsable more clearly to dlstmgulsh, m descrlptlOn and terminology, between moods-being verb forms like the Gr. epif!str;, epéf!ur;, epéf!otr; and intimating the speaker's view ~f the ~elation be~ween the ,process and reality-, and what is often caIled modahty (thlS term bemg taken m a broader sense). It may be true that the Gr. epéf!otr; can be paraphrased by such sentences as "I (the person speaking) wish, consi~er it possi?le, ~dvisa~l~ e:c. that ?,OU ~~ar (will bear etc.)", and logic may distingm~h,the,;arlOus ~,odahtl~s o~ qualificatl?nS of the assertion expressed in the proposltlOn you bear , quahficatlOns accordmg as it is, in the opinion of the speaker, actual, possible, necessary, the 'modality' (in the strict sense ofthe term) is often not explicit-epéf!OIr;-; on the other hand such judgments of the qualifications 01' dispositions of the subject of the 'dictum' as are expressed by the person speaking by means of the °semi-auxiliaries' (w~en used as verbs of complete predication) have no bearing on the aboye relat~on between the process contained in the °dictum' and reality. As long as verbs hke "be able, will, shall" etc. are 'Vollverb' they can like any other verb form part of a dictum qualified as possible, necessary etc. by a person judging and spea~ing. As to the Sanskrit desiderative formations (Piil).. 3, 1, 5ff.) and those synthetlC forms primarily ex:pressing the will, wish, intention of the speaker-for in~tance, the desiderative, promissive, benedictive, cohortative, propositive, ~pt~tlve etc. of Korean-described by G. J. Ramstedt, Korean Grammar, HelsinkI 1939, p.81 under the heading 'volitive forms'-, they are no moods in the proper sense either. One could in the latter case use the term verbal combinations (cf. e.g. G. W. Gilbertson, The Balochi language, Hertford 1923, p. 132).

Page 9: Gonda.1956.The Character of the Indo-European Moods.pdf

~ ~------ --~-------------------.. ---------~--~-----------......... _-_ ...

n. Preliminary observations on sorne verbal forms in non-l.-E. languages

Although the descriptions of the verbal categories in many languages and their peculiarities leave much to be desired it may safely be said that tense is far from being common to any form of human speech. A closer investigation into many idioms which have hitherto been inade­quately recorded will no doubt considerably enhance the number of those in which tense is not indicated by special forms or where it is far from being a well-marked category. It may for instance be doubted whether Strehlow1) is right in ascribing to the Australian Aranda four tenses, present, future, preterit, and perfect, and fourteen modes 2).

Often distinctions which really are aspectual, or sometimes also modal in character-or even modal in the improper sense-are understood to be tenses and described as such by the authors of grammars 3). It is true that in many cases European tense forms are the only adequate means of translating such forins in an intelligible manner. Thus the so-called present or future of the Sudano-Guinean idioms really is an imperfective aspect, indicating that a process has not reached completion. According to Delafosse who calls it, 'faute de mieux', an aorist, this formation can be translated by a narrative or habitual present, by a future, although in this case it properly signifies an inevitable or ap­proaching realization of the state or action; sometimes it can be rendered also by a past tense. Another aspect, the perfect, may for practical purposes, be equalized to the past tenses of verbs which express action and the present of verbó which express a state. Often however, a present or future will be needed in translating. A third aspect indicates that an activity is in progress or a state is being entered upon, under the influence of another activity or state. This formation can be rendered by sub­junctives, optatives and other categories and aH three tenses. In describing

1) T. G. H. Strehlow, Aranda gra=ar, Oceania, 13 (1943), cf. p.194ff. and 31 Off. ; see also A. Capell, The structure of Australian languages, Oceania, 8, p. 27ff., esp. p. 40ff.

2) See also J. Guiart, in A. Meillet-M. Cohen, Les langues du monde, nouvelle édition, Paris 1952, p. 705: "Il n'y a malheureusement pas encore d'étude qui analyse la valeur exacte des termes européens usuels employés, termes qui semblent recouvrir parfois des nuances d'aspect"; A. So=erfelt, La langue et la société, Oslo 1938, p.68.

3) See also E. Cassirer, Philosophie der symbolischen Formen, 1. Die Sprache, Berlin 1923, p. 176.

Verbal forms in non l.-E. languages 11

the grammatical system of Berber André Basset observed: "Les appel­lations précédentes (prétérit, aorist, forme d'habitude etc.) sont traditio­nelles et ne préjugent en rien de la valeur de ces parties du verbe. Les valeurs restent d'ailleurs le plus souvent a préciser. L'emploi temporel des deux formes personelles essentielles, le prétérit pour le passé et l'aoriste pour le futur est évidemment secondaire"l).

In Tibetan, an opposition between a translatif and an intranslatif, and an aspectual distinction, which is, morphologically speaking, mixed up with the former opposition, may be discerned. Tense categories do not exist, in spite of the native grammarians who under the influ~ce of Sanskrit grammar attempted to find a present, past, and future in their mother-tongue 2). With regard to the American Maya idioms Seler 3 ) stated that "eine Zeit-differenz bei weitem nicht die Rolle in der Sprache spielt, die man nach den ausgeführten Konjugationsschematen der verschiedenen Grammatiker vermuten sollte". But this statement was contradicted by another author describing the same language, Alfred Tozzer 4 ), who established the view that Maya has very definite tense signs for a present, two pasts and two futures.

In the language of the Chiricahua Apache 5) five tense-modes are recognized, each of which may theoreticaHy be varied for aspect as well. But in no c¡¡,se does a verb have a distinct form for each aspect and tense-mode. According to the author of the structural description the five principal 'modes' are: imperfective, perfective, progressive (from

1) A. Basset, In the 'Conférences de l'Inst. de Ling. de l'Univ. de Paris', Ir (1934), p. 15. I also refer to the survey given by M. Delafosse and A. Caquot, in Les langues du monde2, p. 742ff. It may be added that A. von Duisburg in his description of the Sudano-Guinean Mbum (Mitt. Sem. f. Oriento Spr., Berlin 28 (1925), p. 146ff.) distinguishes an aorist or 'Grundform'; a progressive (a particle is added), an habitual (a suffixlike element is added), a perfect (a word ma, probably meaning "to be ready") is added; "eine Form für das Futurum laJ3t sich mit Bestimmtheit nicht nachweisen".

2) See H. Maspéro, in Les langues du monde2, p. 542. For the e:x:pression of an 'état acquis' and 'subsidiairement la nuance de passé' see also J. A. Durr, Morpho­logie du verbe tibétain, Heidelberg 1950, p.48.

3) E. Seler, Das Konjugationssystem der Maya-Sprachen, Berlin 1887, p.30. 4) A. M. Tozzer, A Maya Gra=ar, Cambridge Mass. 1921, p. 80; 84. This

author denies the existence of any sharp distinction between "the different modes" in Maya. "The Spanish grammarians in their endeavor to find correspondin~ forms for everything in the Latin gra=ar give forms for the different modes WhICh are, in many cases, most artificial. A subjunctive mood for instance is lacking. The idea of a future possibility is e:x:pressed by certain forms of the future. Besides there is a defective verb utsak which has the meaning "it is possible, perhaps". What formerly was called an opt., properly is a formation containing a root qat, which e:x:presses the idea of "desire'~. A conditional statement is usually introduced by the particle ,wa; the verb in a wa clause does not differ from the verb in the main part of the sentence.

5) See H. Hoyer, Chiricahua Apache, Linguistic Structures of native Amerioa, New York 1946, p. 76ff.

Page 10: Gonda.1956.The Character of the Indo-European Moods.pdf

---~-- -------------------------------. ...... ---------------------------...... ----iIIIIIIIiIJIIl

12 The lndo-European moods

which the future is formed), iterative (also the base for the customary), and the optative. The three categories mentioned last are, from the structural point of view, more closely related. The 'optative' is more like an l.-E. desiderative: gddóya? "would that (the ceremony) begin". Reference to time other than the future is made by means of enclitics, not by special verb forms. The verbal category called 'assertions'l) "resemble tenses but refer to realms of validity rather than of time". One of them declares a general 01' customary truth, another an expect­ancy. In the language of the Tunica Indians 2 ) there are so me thirty enclitic particles expressing "the future tense", various modes, negation and a variety of other notions, the paradigmatic categories of the verb consisting of the semelfactive, habitual, and repetitive aspects, and a conditional. According to Li Fang-Kuei, who prepared a sketch of the Chipewyan language spoken in Canada3), the verb of that idiom, which is a complicated structure, has three 'aspects', to wit the imperfective, referring to a process which is going on; the perfective, referring to a process which has been performed; and the future, referring to a process which is to take place. The future, however, is often identical with the imperfective, sometimes identical with the perfective.

The J apanese verb is indifferent with regard to person and number and, to a certain extent, also with regard to tense. In any case the relevant suffixes "do not serve to define such relatively precise time­relations as can be expressed in English". Thus, the suffix mu is generally described as a future suffix, but yukamu can be translated both 'he will go' and 'he will probably go" and even 'he probably goes'4). The suffixes tsu and nu, though usually described as forming a past tense could­at least from the historical point of view-more accurately be considered affirmative in character. Only incidentally they are temporal.

Passing now to literary Chinese 5) it must be stated that the verb by itself only refers to a state or process; neither the 'agens' (subject) nor the time when, or the way in which, the process is performed is implied in the verbo Notions of time are generally speaking left to begathered from the context, from 'temporal adverbs' or particles which are added if an explicit reference to a definite time is wanted. The verb is also in­different with regard to modal categories. It is interesting to notice that certain 'verbs', meaning "must" etc. can serve as indices of futurity.

1) B. L. Whorf, The Hopilanguage, Toreva dialect, Ling. Struct. ofnat.Am., p.176. 2) Mary R. Haas, A grammatical sketch of Tunica, Ling. Struct. of nato Am.,

p. 346; 363f. 3) Li Fang-Kuei, Chipewyan, ibidem, p. 404ff. 4) G. Sansom, An historical grammar of .J apanese, Oxford 1928, p. 173. 6) See J. L. M. Mullie, Grondbeginselen van de Chinese letterkundige taal,

Leuven, n, p. 12ff.; the same, The structural principIes of the Chinese language (An introduction to the spoken language), l, Peiping 1937, p. 62ff.; n, p.3ff.; R. A. D. Forrest, The Chinese language, London, p. 63ff.

Verbal forms in non l.-E. languages 13

Words for "intending, desiring, being inclined, being about to" when preceding another 'verb' can also do the same duty. In modern spoken Chinese the future, when sufficiently indicated by the context and situation, is likewise left unexpressed; on the other hand, imperative forms, verbs for "must" and the forms oI' possibility are often used to refer to futurity. The past tense, expressed by -la placed after the verb, or after the object (if there is any), is, with intransitive verbs often used where we would resort to a present, because the Chinese fix their atten­tion upon the event 01' action which has already happened and is finished, "leaving it to us to ascertain the result of it", whereas we rather express the result in words.

The use of the so-called 'modal particle' k'~ in literary Chinese is illustrative l). Being placed between subject and verb it must be trans­lated differently, according as circumstances may require, because the process can be situated in the past, the present, or the future. In address­ing other people and speaking about future actions k'~ helps to formulate a wish or a guess. In addressing oneself the k'~ clause expresses an in­tention. In narrative style it expresses a presumption, or expectation on the part of the speaker that the process shall take place. If the process is situated in the past 01' the present, a person addressing others can by means 01' a k'~ clause utter a supposition that a certain event 01' action will have occurred in a certain manner: e.g. "S. must have loved (k'í + verb) his parents very much as he longed to see them when he was fifty", 01' "in all pro bability S. loved ... ".

In Mbum, a Sudano-Guinean idiom, the verb has a 'Grundform', a so-called aorist, which can be used irrespective of time; adverbs are added to refer to a particular moment 01' periodo Besides, there are various 'tense-forms', the so-called future expressing, however, the idea of "ich will essen, ich beabsichtige jetzt zu essen" and the elements serving to form a 'preterit' occurring also as separate adverbs 2 ). In another idiom spoken in the same region, the Mbaka-Limba, the 'con­jugation' is "hochst unvollkommen"3): "Wie im Baja wird die Grund­form, del' Aorist, so viel wie moglich angewendet, auch für die Vergangen­heit und Zukunft. Falls notig, wird dann die V. ode:r Z. durch Adverbien angedeutet. In Nebensatzen, bei denen aus dem Zeitwort des Haupt­satzes hervorgeht, um welche Zeit es sich handelt, wird immer der Aorist angewendet." Besides, there is a particular form to denote completed action. We are reminded of the ancient Indo-European state of affairs at the period when the injunctive was -a living category4).

1) See Mullie, Grondbeginselen, n, p. 34ff. 2) For particulars see G. Tessmann, Die Sprachen der Mbaka-Limba ... ,

Mitt. Sem. f. Oriento Spr., Berlin, 33 (1930), p.70f. 3) The same, ibidem, p. 61. 4) For similar phenomena in Lakka see Tessmann, o.c., p.79f.

Page 11: Gonda.1956.The Character of the Indo-European Moods.pdf

--------------------14 The lndo-European moods

It might be remembered that so-called 'periphrastic constructions' are not rareo In the majority of the Sudan languages a verb for "coming or going" can serve to form phrases answering to a certain extent to future categories in other languages.

With regard to the Bantu idioms in general and Kirundi in particular van der Burgtl) observed that the structure of the verb in itself is simple and 'logical', but completely different from ours, especially as far as the category of tense is concerned. "La modification des verbes est attribuée surtout a cinq circonstances: 1) la forme du radical meme des verbes ... ; 2) la classe, la personne et quelquefois l'objet; 3) la différence de mode­l'indicatif, exprimant un fait actuel; le subjonctif exprimant un fait encore renfermé dans l'esprit, l'impératif qui peut etre rapporté a l'un ou a l'autre des modes précédants comme quasi-indic. et comme subj. (impér.) ... ; 4) la durée dans le mode indicative, dans laquelle il faut distinguer deux stages: a) le stage transitif et non-permanent; b) le stage permanent ou parfait; 5) la différence d'actualité, d'apn3s qu'une phrase est affirmative ou négative." There are also auxiliaries: "Les auxiliaires sont des particules verbales mises devant le verbe principal pour en déterminer le temps (see, however, further on), le mode de pensée ou d' autres notions . . . Beaucoup de ces auxiliaires expriment la meme idée que certains de nos adverbes et conjonctions... «a l'instant, précisément, déja, encore, pas encore, ne point, quand, jusqu'a ce que etc. etc.». L'auxiliaire a ... exprime un mouvement ou une action déja passée mais qui reste encore en pensée sans aucune relation a la durée. (Thus na-wona indicates a vague present: involving habitual action: «l have seen, 1 saw, 1 see», and also implying the idea of being able to do so). L'élément ra exprime une continuation soit dans l'action du présent soit dans l'action du passé. Les formes avec cya donnent aux verbes le sens de durer encore ou pas encore. Zo est l'auxiliaire essentiel pour le temps futur et dérivé indubitablement de ku-zo «venir». L'auxi­liaire ka forme le temps historique, espece de passé narratif, enga indique un temps hypothétique, conditionnel."

In Kikami, a Bantu idiom spoken in the East oí Africa 2), the verbal category described as a present is at the same time used to express futurity, although a special form for the latter may be employed. There is a perfect category, and a subjunctive which, being in frequent use, is neutral with regard to the category of tense; in translating the latter such verbs as "may, shall, will (sollen, mogen, dürfen, wollen)" must be resorted too Besides, the subj. can be used "instead oí" the imper.; the negative imp. is more often than not expressed in this way.

1) J. M. M. van der Burgt, Eléments d'une grammaire Kirundi, Mitt. Sem. f. Oriento Spr., Berlín, 5 (1902), Afr. Stud., p.32ft'.

2) See C. Velten, Kikami, Mitt. Sem. f. Oriento Spr., Berlin, 3 (1900), Afr. Stud., p.15ft'.

Verbal forms in non l.-E. languages 15

In Kikerewe, another Bantu language, the use of the verbal forms and phrases answering, to a certain extent, to modi and tempora of other languages is "une des questions les plus difficiles comme aussi la plus intéressante de toute la grammaire" 1). Whereas the so-called indicative properly speaking is an habitual andat the same time (preceded by the particle na) a narrative, the language distinguishes between an assertive future and an optative 01' imperative 'future'. There is a 'yesterday past' which is also used to express a present state (he is ill, red, big, where do you live?).There are other forms referring, in various ways, to the past: "le 'passé défini' exprime le passé avec notation de l'espece de temps passé, tandis que le passé indéfini exprime simplement le passé, sans déterminer de quel passé il s'agit".

It has already been observed that it is often the context 01' situation, linguistic or non-linguistic, which makes clear what the time of the predication is and what are the other implications of the forms used 2). In Vietnamese such a phrase as anh léim (ti choi apart from a context, is ambiguous as to time and modality; it may be: "Mr. Lam is going; M. L. will go; M. L. went for a walk" 01' even "Go for á walk, M. L." 3).

The verb forms themselves do not carry the categories of tense and mode. These are left to extra-grammatical context, and to some extent also carried by the sentence construction. Besides, a substantive 01' a verb phrase as a subordinate member of a predication can indicate a particular time: "when 1 went (1 didn't inform my wife)". Sometimes ambiguity can be avoided by using a verb meaning "be (act) already", "will, shall" etc. In the other languages oí Further India; aspectual distinctions are as a rule much more developed than temporal. In Thai, for instance, action which is being accomplished is distinguished from action ac­complished, and in the latter case the result is distinguished from the accomplishment as a fact by itself.

Similar phenomena recur in many other languages oí the South-East. The same Tonga (Polynesian) sentence can be translated into English by: "as soon as people know ... , as soon as p. knew, as soon as p. will know". In Ai:nu, in the north-east corner of Asia, the mere root bears the force of a 'tempus indefinitum'; it depends on the context whether we have to translate by a present, a past, 01' a future tense 4). Whereas, in the so-called Palaeo-Siberian idioms, the opposition between a durative

1) E. Hurel, La langue kikerewe, ibid.12 (1909), p.47ft'. 2) Ethnologists stressed the interesting fact that the speech of primitive men

often is nearly incomprehensible, if the situation in which it was spoken or the gestures accompanying it are unknown. See e.g. J. R. Firth, The tongues of Men, London 1937, p. 126ft'.; B. Malinowski, Coral Gardens and their Magic, JI, London 1935, p. 23ft'.'

3) See M. B. Emeneau, Studies in Vietnamese (Annamese) grammar, Berkeley and Los Angeles 1951, p. 54; 63; 72.

4) l refer to Ch. Haguenauer, in Les langues du monde2, p.481.

Page 12: Gonda.1956.The Character of the Indo-European Moods.pdf

16 The Indo-European moods

and a perfectif-or a continuative and a cessative-aspect is essential -the use of suffixes being possible to express iterative, intensive, and other modifications 1)-, temporal concepts are not explicitly signified. The same languages are, on the other hand, rich in modal distinctio~s: Chukchi has different prefixes for the indicative, optative, prospectIve or projective and conditional modes, Yukagir distinguishes bet:veen possibility, probalibity, evidence and necessity of the process. GIlyak distinguishes 'moods' that do not admit a subordinate verb, 'moods' which require, and 'moods' which permit it without requiring it. T~is does not mean that no temporal relation whatever can be expressed m these languages: Alnu can, by means of suffixes, suggest the existence of a chronological order or a causal relation between two occurrences. Gilyak possesses forms indicating the relation of a sub~rdiJ:~ated pr?cess to the main state or action; these forms at the same tIme Imply sImul­taneousness, priority etc. as well as concession or interruption 2). In Chukchi temporal concepts may be expressed by nominal derivatives; a category which may be considered equivalent to a future in other languages is derived from the verb theme in a manner analogous to the formation of modes; temporal subordination is often expressed by syntactic forms of the verbal noun 3).

In Eskimo a 'modal' category, the so-called conjunctive, expresses various actual circumstances belonging to the past or to the present time; the subjunctive refers to a condition in the futu:e or to a ~ypo­thetical condition; other modes help to refer to precedmg, or to sImul­taneous, events 4). Aspect rather than tense is, further, characteristic of Manchu; here the speaker indicates the state in which the action ~s at the moment of speaking, the so-called imperfect for instance connotmg that an action is in process of being performed, without any reference to a particular time 5). Another form (that in -mbi), though corresponding to a West-European present and described as a present indicative or future, really is neither aspectual nor temporal;. it symbolizes t?e mere v~rbal idea, the process as opposed to the nomInal representat~on of the Idea.

In most grammars of Indonesian and other AustronesIan 6 ) languages the chapters dealing with the various forms of the verb and the func-

1) See also R . .Takobson, The Paleosiberian languages, American Anthropologist, 44 (1942), p. 616f.

2) R . .T akobson, in Les, Langues du monde2, p. 421. . . 3) I refer to W. Bogoras, in the Randbook of AmerlCan IndIan languages, II,

Washington 1922, p. 683. . .,' . 4) I refer to M. Swadesh, in 'Linguistic structures of natlVe Amenca, VIkmg

Fund Publ. in Anthr., VI, NewYork 1946, p. 30ff.; W. Thalbitzer, Eskimo, Rand. book of American Indian Languages I, Washington 1911, p. 1031ff.

5) D. Sinor,Le verbe mandjou, Bull. de la Soco de Lin!?,45 (1949), p. l~?ff. 6) For particulars I refer to my paper 'Tense in Indoneslan languages, BIJdr.

Taal-, Land- en Volk-, 110, The Rague 1954, p.240ff.

Verbal forms in non l.-E. languages 17

tions fulfilled by these contain incomplete and often confused state­ments, explications, and interpretations of the relations of these forms to the categories under consideration. Up to recently authors following the beaten track usually discussed tense, overloolring the fact that aspect and mode are as a rule of greater importance. Thus affixes and auxiliaries which can now be said essentially to bear upon aspect or mode were very often explained as temporal in character. Often the temporal, modal, or aspectual functions are closely bound up one with the other: if a specified element may be regarded as referring to past time, there is nearly always an idea of completion or another aspectual function implied in it. Moreover, these elements are not infrequently optional. Some authors, however, were right in observing' that their temporal force is, general1y speaking, hardly anything, that it is, therefore, better to dwell as little as possible upon their temporal character, that we had better describe them as suggesting a 'narrative tense', not the past, as laying emphasis upon the reality of the events or objects heard or seen by the speaker. In Leenhardt'sl) short descriptions of the Austro­Melanesian idioms two sentences regularly recur: "Le temps n' est pas exprimé. Des morphemes nuancent les aspects de l'action ou de l'état". Such morphemes as the Ajie (Mel.) ma do not convey the idea of tense: it expresses duration, continuity, unbroken course of action, but by indicating "la distance parcourue depuis que le sujet l'accomplie, ou a parcourir pour que le sujet l'accomplisse" this auxiliary can in a definite situation refer to the past, in another situation to the future.

In the IN. languages the idea of time very often becomes clear from the context or the situation described, from the mutual relation of clauses or from auxiliaries suggesting completeness, will, intention, etc. The verbal forms by themselves do, as a rule, not imply any particular tense. In those langnages which possess a variety of verbal categories the main distinction sometimes seems to be between contingent actions, actions taking place not exactly now and actual actions which are now in progress (Sangirese, north of Celebes). In Tagalog 2 ), as far as the transient formations are concerned, a modal distinction prevails: there are two modes, one denoting actuality and the other contingency. Each ofthese has two aspects, punctual and durative. The actual mode envisages the occurrence as actually having taken place or taking place; in the punctual aspect the occurrence is viewed in its entirety, without regard to duration, and hence always as past; in the durative aspect it is viewed as a process going on in time, past or presento The púnctual aspect of the contingent mode envisages the occurrence as possible, hypothetical, or commanded;

1) M. Leenhar~t, Langues et dialectes de l'Austro-Mélanésie, Paris 1946, p. 20; 28; 32; 37 etc.

2) I refer to L. Bloomfield, Tagalog te:x:ts with grammatical analysis, Univ. of Illinois, Urbana 1917, II, p.217; 226ff.; 23lf. 2 Gonda, Indo-European moods

Page 13: Gonda.1956.The Character of the Indo-European Moods.pdf

18 The Indo-European moods

the durative views it as future from the point of view of the past or presento In another class o~ verb~ which express. the act~r in a simrle action or process without mvolvmg a comparatlvely dehberate actlOn and a considerable effect on external objects the distinction between punctual and durative categories prevails over the opposition actual: contingento Temporal concepts or relations are,. generally speak~ng, expressed otherwise. Affixes added to, or partlCles accompany~g; expressions of time or other words enable the speaker to refer to tIme without using verbal forms expressing temporal relations.

It may be added that a comparison of modern and ancient stages of the same language-if the latter are known to us-shows that the influence of modern civilization in some cases led to a more extensive use of temporal categories. Whereas, in classical Malay (XVIIth century) futurity is very frequently left unexpressed, the modern language (Bahasa Indonesia) almost always indicates it explicitly. And according to Leenhardt the younger generation of some Melanesian peoples prefers to use temporal auxiliaries rather than the aspectual elements, still in use in the conversation 'of the aged.

In the Khmer languages of the mainland of S. E. Asia the opposition of infective and perfective aspects is regularly expressed; there is also a category which might be caIle.d a future or an intentional. The aspec~~al idea is conveyed either by partlCles after the verb, or by a sort of auxlha- ..... ries before the verbo As the languages belonging to this group resort to different elements, this device, if not liable to chance with regard to these elements, must be comparatively young.

Other cases are far from rare in which similar or even identical elements combine temporal value with another function. While, in Korean, part of the suffixes are indifferent as to aspect and tense, others are temporal in character. However, part of these have also a determinative function. Other elements, being mainIy aspectual, can imply time, e.g. the inten­tional or < quasi future' 1). In the Altaic languages in general (Turkish, Mongol Tungus, Korean) a deverbative. adjective ending in -r and having 'the value of a future or present-future can easily be disting-

uished 2). In Taos, a language spoken by a smaIl number of Indians in north~rn

New Mexico, the categories of temporallocation (tense) and extenslOn (aspect) occur only in certain combinations with each other. Present is present-durative, with the aspectual idea more prominent than that of time. Past is durative, and, moreover, non-present and non-future.

1) 1 refer to G. J. Ramstedt, A Korean grammar, Hel~inki 1939, p. ,14ff. 2) For particulars see Ramstedt, Einführung in die altalsche SprachwIssenschaft

n, Helsinki 1952, p. 86ff., where the reader may also be informed of the originally nominal character of the Altaic verb forms, with the exception of the impero and

the opto

Verbal forms in non l.-E. languages 19

Future is a momentary future 1). Although the Turkish grammarians usually distinguish eight zamans, i. e. tempora, these basic forms of the Turkish finite verb cannot be identified with l.-E. tenses or moods. "S~e stellen im wesentlichen acht verschiedene modale Betrachtungs­WeIsen des Geschehens dar, und zwar vom gegenwartigen Standpunkt des Subjekts aus gesehen" 2): "das <Prasens' stellt eigentlich augenblicklich sich vollziehende Handlungen dar; der < Aorist' ist eine zeitlose Form' er drückt eine Gewohnheit, regelma13ige Wiederkehr, fortdauernd~ Handlung aus; die <bestimmte Vergangenheit' bezeichnet eine ab­geschlossene Handlung, ein abgeschlossenes Geschehen das man selbst . ' bezeugen kann; das <Futur' stellt ein Geschehen, eine Handlung dar, die in der Zukunft bestimmt eintritt; der <Necessitativ' bezeichnet eine no~w~ndig~, bisw~ilen auch eine wahrscheinliche Handlung; der <Op­tatlv bezelChnet 1m Hauptsatz den Wunsch, im Nebensatz nach ki die Folge oder die Absicht" etc.

In Chitimacha (spoken in southern Louisiana) a future, an aorist, a continuative, a usitative, and otherverbalcategories may be distinguished. Whereas, however, the future is used to state that a process will take place in the future, the aorist makes a simple reference to a process occurring in the present-past time period 3).

The history of the analogous phenomena in the Chamito-Semitic languages is worthy of a somewhat longer recapitulation4). In the western Sem. languages the unaccomplished action (imperfective: atelic aspect) is mainly indicated by prefixes, the accomplished (perfective: telic aspect) by suffixes. Eastern Sem., and the Berber and Kushitic divisions, which onIy possess prefixed forms, these two aspects are either expressed otherwise or wanting. From these facts it has been ?oncluded that the dis~inction between the two aspects is not <original'; m that case the Chamlto-Semitic once had an indicative which referred to a state or process without aspectual or other modifications, and perhaps. al~o. a durative. In any case, tense formations were wanting. As the mdlvlduallanguages, especially Akkadian on the one hand and the Western Sem. idioms on the other, pursued their own course in developing <temporal' categories and in adapting aspectual distinctions to suit ~ore complicated functions, this process may be regarded as comparatlvely recento Aspect rather than tense remained characteristic

1) G. L. Trager, in Ling. Str. of Nat. Am., p.203. 2) L. Peters, Grammatik der türkischen Sprache, Berlín 1947, p.67ff. 3) M. Swadesh, in Ling. Str. of nato Am., p. 322f. 4) For pa~ticulars see: M. Cohen, Le systeme verbal sémitique et l'expression du

temps, Parls 1924; C. Brockelmann, Grundri13 der vergl. Gramm. der semit. Sprachen, n, Berlin 1913, p. 144f.; G. Bergstrasser, Einführung in die semitischen ~prachen, München 1928, p. 10ff. etc.; H. Bauer, Die Tempora im Semitischen, lhre Entstehung und ihre Ausgestaltung in den Einzelsprachen, Thesis Berlin 1910. 2·

Page 14: Gonda.1956.The Character of the Indo-European Moods.pdf

20 The Indo-European moods

of the whole family. In Akkadian the so-calIed preterit (type: iprus)­originalIy the only category-was mainly. ~eserved . for punct~al­narrative functions. The younger 'present' (~paras), bemg a fie~tlVe­durative referred to continuance in time 01' to repeated processes m the present ~r the future, but also in the pasto In the beginning t~e temporal distinctions were mainly objective 01' relative and determmed by t~e contexto In the course of time the so-calIed present came, to a certam extent to be used to indicate, 'subjectively', future processes and durati~n in the present time. Another category (iptaras) came into use to express the 'punctual' presento That the 'imperfect' .was originally used on a very large scale appears in Hebrew from certam phrases a:nd constructions in which it refers to the past-e. g. after the consecutlve ua in narrative style, 01' to indicate progressive duration ?r repetit.ion­~r even to the future, 01' where it is used irrespective of tIme. But m the main it was reserved for references to present 01' future occurr.ences; these are, in the first case, either permansive in character 01' typlOal 01'

recurrent· in the second case an imperfect can also answer to our con­struction; "to be able to ... may ... must ... will ... etc.". The perfect, expressing accomplishment of actual 01' imaginative existe~ce, had developed into the main formation referring ~o the past, that IS to say: to occurrences which now have reached theIr close; to (non-recurrent) events described as really belonging to the past; it can also correspond to oul' pluperfect. However, in other cases the perfect must be translate.d by a present form; e. g. when it corresponds to the t;ype ol~a 01' to a gnorolO aoristo Besides, it can, for instance in connectlOn wlth expected 01'

promised occurrences, refer to the future ~i~e. The asp~ctual character of these forros and their non-temporal ongm are unmlstakable.

In Arabic 1) both categories can likewise occur in connection w~th any time. The imperfect generalIy expres~~s action.s, ~vents etc. ,;hlOh -irrespective of time-are in progress ( m yerwI~~lOhung .. be~riffe~e Handlungen"), 01' the repetition of which contmues, sofern. ~amhch dIe sich wiederholenden Akte zu einer Gesamtanschauung VereIlllgt werden und als deren Ablauf erscheinen"2). With regard to the perfect the essential point is. reference made to completeness 01' ~cco.mplishment of phenomena, events, actions. It is, inter alia, resultatlVe m the pr~~;~, the past, the future; it can also denote repeated processes, etc.· Ir sollen Verba finita begreifen, die zeitlos sind und zumal unsere Perfekt~ und Imperfekta zur Ubersetzung verwenden, ohne dabei etwas Praeterl­tales zu denken."3) However, the perfect was also in this language used to represent the subject in relation to a process to which one could look

1) See also H. Reekendorf, Die syntaktisehen VerhaItnisse des Arabisehen, Leiden 1898, p. 52ff.; the same, Arabisehe Syntax, Heidelberg 1921, p. 12ff.

2) Reekendorf, Verhaltnisse, p. 57. 8) Reekendorf, Verhaltnisse, p. 52.

Verbal forms in non l.-E. languages 21

back because it had reached completion. Hence its use in connection with experienced occurrences, in narratives, and under other circum­stances which could contribute to its assuming temporal functions.

Some words must also be said on the Dravidian languages which are generally described as being in possession of the "three natural tenses" 1). On closer examination the relevant formations turn out, however, to be no pure tense-categories. Especially the future is not very distinctive. When used at all it denotes the future more commonly than other time but it also serves to signify that an action 01' occurrence is habitual o; possible, 01' to express eventuality or contingency. There even e~ist 'future' forms which it would be better to regard as indeterminate tenses. Telugu grammarians called these aorists, because their reference to the f~ture is st~llless distinct. The so-called present has also wider applica­tlOns than ltS name would suggest. "Un signe zéro lui suffit en principe par opposition aux autres temps: mais du fait que ce 'présent' combine les sens d'éventualité et de réalité, chaque langue aura été obligée de se créer un moyen de noter plus particulierement le présent réel s'opposant au passé réel, et dans une mesure moindre au futur."2) The hypothesis has been defended 3) that the suffixes which bear the temporal force originally were nominal suffixes, serving to form cases 01' derivatives.

A brief reference may also be made to other cases in which temporal categories are, 01' may be considered to be, of a comparatively recent date. In the Finno-Ugric íamily of languages tenses and modes are far from being homogeneous in character; originally, they must have been implied in verbal forms which mainly fulfilled other functions (the expression of aspect and orientation). In Samoyed verb forms still express a 'present', 01' a 'past', 01' a 'future' according to the sense of the theme contained in them 4

). In those members ofthis family which have under­gone a cultural development, distinct categories of tense have come into existence.

In the Indian Munda languages various aspects and 'modalities' can from ~f old be ex~ressed; e.g. accomplished and unaccomplished aotion; there IS a resultatlve, a non-resultative, an 'aorist', a durative etc. They are formed by means of suffixes. The possibility of expressing temporal ?oncepts by special means; to wit auxiliaries, is ascribed to the powerful mfluence oí the surroundmg Indo-Aryan and Dravidian tongues5). In Santali the particle kan, referring to the present time, joins only the

1) J. Vinson, Manuel de la langue tamoule, Paris 1903, p.110f. 2) J. Bloeh, Strueture grammatieale des langues dravidiennes, Paris 1946, p. 59. 8) Bloeh, o.e., p. 60: "Le passé reposerait done respeetivement sur l'expression

de l'appartenanee au sujet, et le futur exprimerait un mouvement au earaetere eentrifuge partant du sujet."

4) See A. Sauvageot, in Les langues du monde2, p. 302f. 5) Cf. e.g. J. Hoffmann, Mundari Grammar, Caleutta 1903, p. 112ff.

I

Page 15: Gonda.1956.The Character of the Indo-European Moods.pdf

22 The Indo-European moods I Verbal forms in non l.-E. languages

infective aspects, but tah~kan, which refers to the l?ast, can .also follow a perfect: then it places an accomplished ac~ion and ltS result In the pasto

In other languages tenses exist, but not m the threefold s~h~me ?f our school grammars. Sometimes <present' and <future' are not dIstmgmshed, a present-future existing beside a past tense.. .

From the aboye survey, which could easily be amphfied, the c~n- .

l. may be drawn that there is no cogent reason to regard, wlth

c USlOn h' t . 1 E t ystem Hirt and Miss Hahn, the existence of a pre lS 01'10 .- : ense s . involving a special future tense, as self-evident .. The eVlde~c~ derlved from a study of various languages belonging to dIfferent familJes seems, on the contrary, to prove that the categories oí tense are, on a ~arge scale either absent, 01' <impure', 01' a phenomenon of comparatlvely late

appearance.

IIl. Tense and primitive thought

Those linguistic phenomena which occur in every part oí the world cannot be kept apart from 'ethno-psychological'phenomena, that is to say: from certain characteristics of human thought. "Wie del' einzelne Laut (i. e. das einzelne Wort) zwischen den Gegenstand und den Men­schen, so tritt die ganze Sprache zwischen ihn und die innerlich und auBerlich auf ihn einwirkende Natur."l) Although the connections between thought 01' mental structure of the so-called primitive, 01' pre­scientific, peoples, groups, and individuals on the one hand and their language 01' speech on the other, have often been commented upon by various authors, the particular points with which we are concerned here seem to have escaped their attention. So it may be useful to look more closely at <primitive man's' conceptions of time and of events and occur­rences taking place in time, and to enlarge upon the question whether the concepts with which our grammarians and philosophers usually operate in attempting to explain the aspectual, modal, and temporal categories under consideration, are in all respects usable in studying those stages of languages in which these categories carne into existence 01' were, at least, in their heyday.

Primitive, i. e. pre-scientific 01' natural, man 2) is mainly interested in concrete objects of reality and in those occurrences and activities which have a bearing upon his own existence 3). His thoughts are, generally

1) W. von Humboldt, Kawisprache, Einleitung: Über die Verschiedenheit des menschlichen Sprachbaues ... , ed. 1907, 1, p.60.

2) For the term <primitive' see e.g. G. van del' Leeuw, La structure de la men­talité primitive, Paris 1928; G. S. Overdiep, in: Volkstaal en dialektstudie, Ant­werpen 1947, p.56f.; ;ro Gonda, Stilistische studie over Atharvaveda I-VII, Wageningen 1938, p. 19.

3) See e.g. R. Thurnwald, Die Psychologie des primitiven Menschen, München 1924; the same, Primo Denken, in M. Ebert, Realle:x:ikon der Vorgeschichte, Berlin 1924-29; the same, in the Anthropos, vol. 12-15 (1917-1920); the same, Mitt. Anthr. Ges., Wien 1921; K. Th. PreuI3, Die geistige Kultur del' Naturv5lker, Leipzig 1914; F. Boas, The mind of primitive man, New York 1927; the same, Handbook of American Indian Languages 1, Introduction; L. Lévy-Brühl, Les fonctions mentales dans les sociétés inférieures, Paris 1912; the same, La mentalité primitive, Paris 1925 (and the critiques of these books); T. G. Strehlow, Aranda graromar (see above, ch. II); M. Leenhardt, Langues et dialectes de l'Austro­Mélanésie, Paris 1946, Introduction; the same, Le temps et la personnalité chez les Canaques, Revue philosophique 1937; ;ro Murphy, Primitive man, Oxford 1927, passim; P. Radin, Primitive man as philosopher, NewYork-London 1927; E. Cas­sirer, Philosophie der symbolischen Formen, 1, Berlin 1923, p. 166ff.; B. Mali-

Page 16: Gonda.1956.The Character of the Indo-European Moods.pdf

24 The Indo-European moods

speaking, tied d.own t.o his immediate interests, t.o his daily .occupati.o~s and t.o actual situati.ons. His interests in .objects and phen.omena 18

subjective, his judgement influenced by his affects, ~is vie,:"s .of thin~s eg.ocentric. When c.onversing with his fell.ow-men he lS n.ot m the hablt .of discussing abstract ideas. F.or him, language is n.ot a means ?f re a­s.oning, S.olving pr.oblems' .01' expressing scientific th.ought, bu~ an mstru­ment facilitating c.ompani.onship and daily interc.ou~se. wl~h men .of similar cultural and intellectual standards. In c.ontradIstmctlOn t.o edu­cated and m.odern pe.ople wh.o tend t.o be analytical, argumentative; and intr.ospective, and are bent .on hyp.othetical th.ought, supp.ositi.ons, and abstract ideas, pre-scientific man dpes n.ot like t.o br.o.od .over what might have been, .over what c.ould .01' sh.ould hav~ been, .01' .over what w.ould .01' might have happened .01' have been d.one m the past .01' at any. .other m.oment. His mind is mainly captivated by c.omplex phen.omena, n.ot by attempts t.o analyse these. His c.oncepts are likewise c.omparatively c.omplex and vague, and largely determined by en~.oti.onal fact.ors. In thinking and explaining he .often res.orts t.o timeless Images, myths, and .other c.om~lete structures. In his. <literature' time. in .our ~.odern se:r;s,e is n.ot an lmp.ortant fact.or l ). HlS mental .operatlOns are asyntactlC. He has n.o c.ommand .over c.oherent units .of great c.ompass. Realityand imaginati.on are n.ot always kept aparto Oa~sal c.onnecti.ons and c.onstruc­ti.ons, if they interest him at aH, are .often different fr.om .ourS. Tw.o even~s which in .our .opini.on are cause and effect can present themselves t.o hIS mind as identical, f.or like pr.oduces like, and the effect resembles its cause. Things which have .once been in c.ontact with each .other c.ontinue t.o act .on each .other after that c.ontact has been severed. These <principIes .of th.ought' d.o n.ot urge pe.ople t.o attain t.o m.ore than vague n.oti.ons .of chr.on.ol.ogical relati.ons. What we w.ould caH a symb.ol f.orms part .of th.e .object which it represents; signs .01' .omens are apt t.o. be n.o~ .only ~ntl­cipati.ons, but preliminaries .01' even causes. Oaus~s bemg easlly ascnbed t.o the invisible are .often c.onsidered extra-spatlal and extra-temp.oral. N.ot being clearly distinguished fr.om their effects .01' manifestati.ons the same .occurrences can (as willed by the invisible) bel.ong t.o the past (as causes) and t.o the future (as manifestati.ons). In many acts .of <pr.ov.o­cative' magic this identity was even systematized. Dreams which fr.om

nowski in Science, Religion and Reality, ed. by J. Needham, London 1925, p. 19ff.'; G. van der Leeuw, Religion in essence and manifestation, Lon~on 1?38; T. S. G. Moelia, Het primitieve denken in de moderne wetenschap, Thesls LCld~n 1933; G. J. Held, Papoea's van Waropen, Leiden 1947, p. ,265ff.; S. Schayer, DIe Struktur der magischen Weltanschauung, Zs. für Buddhlsmus, ?, .p. 259ff.; M. Gluckman, The logic of African science and witchcraft, Rhodes-LlVmgstone Inst. Journal, I (1944), p. 61ff.; B. Malinowski, Magic, Science and Religion, NewYork 1954; J. Gonda, Inleidingtot het Indische denken, Antwerpen-Nymegen 1949, ch. lo

1) In folk- and fairy-tales chronology is, aS,a rule, a ~atte~ of n~ cons~~uence, the hero and heroine remaining young and tlille sometImes standmg stlll .

Tense and primitive thought 25

the ?bjec~ive P?int .of view bel.ong t.o the past are c.onsidered t.o be real ando Iden~lCal Wlt~ .occurrences which bel.ong, .01' may bel.ong, t.o the future. Besldes, m mythlCal th.ought chr.on.ol.ogy is unessential. Mythical realit is timeless; it is, .01' was, bef.ore aH times, "in the beginning", as well Zs actual, and presento Th.ough timeless, it can always be repeated and represented here and n.ow.

Tw.o principIes .of th.ought .on which a magic view .of the w.orld and its phen.omena is based can .often be distinguished: the c.onvicti.on that lik~ pr.oduc~s like, that the effect resembles the cause, and the belief that .obJects WhlCh have .once been in c.ontact with each .other c.ontinue t.o act .on each .other at a distance after the physical c.ontact has been severed. Whe:e. these princi~le~ pr~vail they certainly tend t.o prevent man fr.om acqmrmg an exact mSlght mt.o the true character .of the relati.on between pr.ocesses and the time during which they take place, fr.om acquiring a m.odern c.ommand .of the fundamental c.oncepti.on .of time inv.olving :ec.ogniti.on .of the ideas .of bef.ore and after, past, present, ~nd future. m the sequence .of events. '

M~re.over, the c.onten~ .of the idea .of time is, f.or the <primitive', largely pr.ovlded by the m.ost Imp.ortant peri.odical events in nature .01' in the life . .of the tribe .01' the individual. Whereas science and a highly .organized S.oCl~ty le~d. man t.o. attach value t.o accuracy with regard t.o time, the ma~I~.o-rehglO~s belIefs ~nd . .observances .of primitive man, laying stress ?n hlgh:days and speClallmp.ortant m.oments, and sub.ordinating the I:r;tervenmg ep.ochs t.o ~hese may have had the .opp.osite effect. The primi­tIVe ~.ows t~at ther.e IS a lap.se .of time between s.owing and reaping, but he mamly dlrects hls attentlOn t.o these tw.o imp.ortant events, n.ot t.o tw.o dates each .of which, .objectively speaking, is a measurable fracti.on .of that lapse .of time. Every imp.ortant event has in a sense its .own time, f.or it is n.ot <da.te~'; n.ot.see~ as.happening at a special chr.on.ol.ogical m.oment. F.or the prlmltlve tIme IS, mdeed, n.ot an <.order .of successi.ons' .01' a h.om.ogene.ous quantum. It is sensed qualitatively - characterized by the .eve~ts taking place in it-rather than represented intellectually and .obJectlvely. Imp.ortant .occurrences have a character .of their .own and mI the time in which they take place l ). "Le temps, au sens primitif du m.ot, est le c.ourant des événements, la .oU ce c.ourant est le plus puis­sant." 2) Wher~as we. regard ~ime as ~ stream .01' straight line, as a regular

~uccesslOn .of smgl~, umque, a~d Irrev.ocable moments, with.out beginning m the past and WIth.out end m the future, the primitive experiences it

:), See for, instance also H. Hubert (et M. Mauss), Mélanges d'histoire des rellglOns, ~arls 1909, p .. 197ff.; Hubert, La représentation du temps, Ecole pratique des hautes étud~s, Parls 1905, p. 6ff.; Lévy-Brühl, Mentalité primitive, p.88f.; 126; 219; P. Samtyves, in the Revue Hist. Rel., 79, p.83ff.

2) G. van der Leeuw, L'homme primitif et la religion, Paris 1940, p.99.

J

Page 17: Gonda.1956.The Character of the Indo-European Moods.pdf

26 The lndo-European moods

as duration 01' as periodical recurren ce, conceiving it as a cycle, as some­thing that can return, and does return periodically, as something that can be renewed.

Rence, no doubt, the remarkable fact that many ancient words for "time" in Indo-European languages 1) do not cover our modern concept of time, the fundamental conception involving recognition of the ideas of before and after, past, present, and future, in the sequence of events. Some of the ancient words for "time" are from roots for "stretch" 01'

"measure", others are used especially for definite moments 01' special times. The Greek w(!a refers to any period, fixed by natural laws and revolutions, whether of the year, month, 01' day, and more particularIy

" ""." R O 584 -Q' JI r¡ to the seasons: Z 148 we0C; ... roer¡ sprmg; es. p. 'uS(!eoc; we "summer"; in a especial senso' it is, inter alia, used of the ca:mpaigning season; besides, wea refers to morning, noon, evening, and mght, the ever-returning important 'periods' 01' times of the day; very often, however the word is applied to the season 01' fitting time for an act 01'

undertaking, especially when it recurs at more 01' less regular intervals: thus Romer has the phrases wer¡ 1-toh:olO, vnvov, ()6enOlO "time for bed, for sleep, for the evening meal", Resiodus has wer¡ a(!6rov and a{tf¡rov (Op. 460; 575) "time 01' season of tillage, harvest-time", yá{tov wer¡ is <ethe time of marriage" and a girl can reach the "time for a husband": Plato, Critias 113 D sic; av()eOc; weav -!í1-tovl1a; wea 01' wea 811rív + inf. means "~t is time to do a thing", r~v w(!av is "at the right time" etc. etc. The SlavIC jara which is etymologically related to wea means "spring". The Avestan yar, yar'J and the German O. R. G. jar etc., which are also of the ~an:e origin, denote the great periodical period, "the year", b~t O. En? gea'/' lS

also "spring". The Greek ua~e6c;, though usually referrmg to tIme, can also mean "due measure, fitness", 01' "vital part" in a 'local> sense-cí. Uaí(!lOC; "in 01' at the right place", uaíelOV "a vital part of the body" , uat(!lOC; of a wound: "mortal": when applied to time it means "the exact 01' critical time, the opportunity": Pind. P. 4,286 uweoc; neoc; av&edmw.v {leaxv (tSreov eX8~ "time and tide wait for no man"; uweoc; rijc; revyr¡c; lS

"harvest-time, time for the vintage"; ol uweoí are "the critical times". The word xe6voc; often refers to a definite time 01' period: the time of youth, an age 01' lifetime, a season 01' portion of ayear, a date 01' term, ayear. The Irish amm, am means "time, occasion, and point of time". The Gothic peihs, used to translate 1-tw(!6c; as well as xe6voc;, is usually considered to be etymologically related to O.N., O.E. ping "judicial as­sembly", O. E. pingan "to conclude a treaty"; O. E. ping-gem~arc "meas­ured time", O. Sax. firthingian "to promise" 2); as the word p~ng may be assumed to have derived its name from its periodical character, this Gothic

1) See also O. D. Buck, A dictionary of selected synonyms in the principal r.-E. languages, Ohicago 1949, p.953ff. . .

2) l refer to S. Feist, Vergl. W6rterbuch del' gotIschen Spr., Leiden 1939, p. 494.

Tense and primitive thought 27

term for "time", whatever the root to which it ultimately belongs 1), seems to have originally referred to important and recurring periods rather than time in the modern sense of the word. The Goth. lveila must origin­al1y have denoted the special moment of taking a rest, an important act likewise occurring at regular intervals : ga-lveilains (cf. Germ. verweilen) renders avS(J'~C; "relaxation, recreation", O. N. hvila means "resting place, bed", O.R.G. (h)wila "time" and "Weile": the Lat. quies "rest" and O. Oh. SI. po-cUi "to rest" are cognates. The Old Norse timi (cf. EngI. time) often denoted "an appointed time, 01' period of time", in mod. Scand. the word has come to mean "hour", in German dialects zime is "time" as well as "opportunity". The Dutch tijd and Germ. Zeit are etymologically identical with the Engl. tide, the original meaning of which (' 'period, season") survives in eventide, Whitsuntide, Ohristmastide ; the English word also denotes "the alternate rise and fall of the surface of the sea" which in Dutch is called tij: pI. getijden < Med. Dutch ghetíde "space of time, fixed time, church festival,the proper time for doing a thing"; the EngI. tide is also used in connection with circum­stances and affairs: the high tide 01 lortune; the tum 01 the tide, etc. An interesting group.ofwords are the O. R. G. etc. O. N. etc., mal (cf. Dutch maal) , which meant "Zeitpunkt, point of time", hence "the fixed time for having one's meal", and "meal"; cf. also the O. EngI. mael "measure, fixed time, meal had at a fixed moment, meal", also used of "opportunities etc.". The Germ. Mahlzeit, "früher auch für ein festliches Mahl gebraucht, hat sich als W ort del' gew6hnlichen Rede für das regelmaBig eingenommene Essen festgesetzt" 2). The Germ. mal, Dutch maal in zweimal, tweemaal etc. which refers to "Wiederholung del' gleichen Situation in verschiedenen Zeitpunkten"3) is the identical word. The PoI. and S.Or. daba, used for "time" in special senses 01' phrases, originally referred to "fitting time": cf. Oh. SI. podaba jestU "is fitting" etc. 4 ). The Oh. SI. vrémr¿, Russ. vremja and other Slav. words must have developed on the basis of the idea of "turn": cf. Skt. vartman- "track of a wheel, course", Lat. ve1'tere "to turn": the idea of "time" developed from that of "a cycle oí seasons, fixed 01' important times". The Skt. kala- 5) likewise often ap­plies to a "fixed 01' right point oí time", a "space of time", "the proper time 01' season for doing (periodical) acts"; it can stand for "time to eat, meal-time", for "time of death etc.". The word 1'tu- which is more usual in the ancient documents, means "any settied' point of time, fixed time, time appointed for sacrifices and other regular worship, 01'

1) See e.g. Walde-Pokorny, Vergl. Wtb. d. idg. Spr., l, Berlin-Leipzig 1930, p.724f.

2) H. Paul, Deutsches W6rterbuch2, Halle a. S. 1908, p.342. 3) Paul, O.C., p. 343. 4) E. Berneker, Slavisches etymol. Wtb., Heidelberg 1924, p.203. 5) See also F. B. J. Kuiper, Zs. für lndologie u. lranistik, 8, p.241ff.

i

I

1 I

j

Page 18: Gonda.1956.The Character of the Indo-European Moods.pdf

28 The lndo-European moods

for any action in general" ; apart from "season, right time" it is also used for "the time after a woman's periods". Similar observations might be made in connection with many non-l.-E. words: the Rebr. 'et for instance, though translated by "time" especially applies to "fixed, right, and naturally proper moments" such as the right time for eating, for coming, for rounding up cattle, for a woman's periods, for the fixed moments of God's rule etc. etc. The plural can be used for "important occurrences". This sense likewise belongs to the plural of the Arabic words for "day" and "year". In that language, waqtu "time, hour, moment, season" belongs to a group of words meaning "to fix, determine (the moment or proper time for an action)". In the Indonesian Bare'e (Oelebes) tU1'O, translated by "time", properly means "the right time for an action, opportunity"; ta'u, in the same language, indicates "the course of time in which the crop ripens", hence "a rice-season", and "a year"; the widespread Austronesian tahun "year" is the identical word, which in cognate idioms is likewise used for other "yeq,rs", i. e. periods or seasons of less than twelve months: Mal. tahun padi "seasonal from rice-crop to rice-crop". In Bolaang-Mongondow (Oelebes) toñgantag refers to the appointed time as well as a fixed and right place.

If, however, time is experienced as periodical recurrence the past can return and then be the future. What does not return is not important; only the periodical 'high-days' attract man's attention so as to occupy a place in his 'chronological system', A past or future event is, moreover, not, generally speaking, represented as occupying a definitely determined place on a straight lineó it is rather sensed as being distant 1). From various idioms it appears that, temporally as well as spatially2), the main distinction often is between the near and the far, between the here-and­now, or here or now, and the not-here, there, or not-now. One might compare the 'double meaning' still inherent in such a comparatively recent vocable as the Dutch straks 3), a temporal adverb meaning "pre­sently" and "just now". In Sanskrit, tat1'a "there" when used in a tem­poral sense, can refer to the past (e. g. AitBr. 7,28) and the future (e. g. Mbh. 3, 65, 6). Of. also G. n07:e "at some time or other", 7:67:8 "at that time, then" which are used in reference to the past as well as the future; the Engl. then "at some former time", but also "at that time in the future"; the Rebr. az "then", likewise referring to both past and future and other such temporal adverbs in other languages. Do we err greatly ifwe consider these words to reflect an ancient distinction: now-here: not­now-here1 The same phenomenon occurs in other languages: in the IN.

1) See e.g. also Murphy, o.c., p. 133ff.; v. d. Leeuw, Homme prim., p.95ff.; Religion, p. 384ff.

2) E. Cassirer, Philosophie der symbolischen Formen, l, p.168. 3) See Franck-Van Wijk, Etymol. Wdb. d. Nederl. Taa12, The Hague 1929,

p.674.

Tense and primitive thought 29

Letti anki and nanki mean "the other day" and "presently", and the Ohin. chiu a~d various words in Polynes~an idioms show the same peculiarity. We mIght recall here the at first sIght curious phenomenon occurring in American and other languages (e. g. Ewe), where the same word serves to express the idea of "yesterday" and that of "to-morrow1)"; the Poly­nesian usage, described by Leenhardt 2), to indicate bythe same particle­which does not refer to any particular time-the space of time passed away after performing an act as well as the space of time to be passed in order to accomplish a work in the future; a particle such as the Tlingit gu- or gA- which can help to refer to both past and future events 3).

The opposition past-future is, in the aboye 'division' of time, much less clearly distinguished "'). Especially the future is not infrequently vague­ly represented. In many languages, also in modern Europe 5), future pr6cesses are far from being always indicated by special verb forms. One of the factors contributing to this usage is the following: future events can be expected and regarded as certain, and by the strong em?tions which.they arouse be experienced as present 6). They can, psycho-10gICall~ speakmg, belong to the psychological now. We might in this connectlOn not only remember the well-known present instead of a future in the speech of diviners and oracles: Pind. O. 8,42 IIéeyaflor; aflCPt 7:8aír;, fíewc;, xeedr; Beyaatatr; áAtax,e7:at; Aesch. Ag. 126 xeÓYW flBY ayeei IIetáflOV nó).t'I' (jjje x,é).em'}or;, but also such passages as Aes~hines 2,183 fltx,ea o' en ebrwy rjor¡ x,u7:a{3atYw 7).

The aboye is not to say that the primitive have no idea of time. The only inference we may make seems to be this: they do not attach much importance to many temporal distinctions which playa role in 'modern life', and therefore do not refer to them in speaking. So the lack of tem­P?ral categories, like the absence of certain classes of words (for instance hIgher numerals), must be mainly considered to be due to the lack of their

. 1) See .also Cassire~, O.C., p. 173. This is probably the key to the puzzling Goth. f~8tradagUJ for aiJQwv ro Mt. 6,30: "wie Uppstrom bereits bemerkt, (bedeutet) auch 1m Ahd. ~-gester 'übermorgen'" (F. L. Stamm-M. Heyne Ulfilas Paderborn 1878 p.7). '"

:) M. Leenhardt, Langues et dialectes del'Austro-Melanésie, Paris 1946, p. XLI. ) See J. R. Swanton, in F. Boas, Handbook of American lndian languages

l, Washington 1911, p. 176. ' 4) It may be observed that the aboye "now" is not a mathematical abstraction

or an indivisible moment, but a psychological "now". 5) See also Ch. Bally, Linguistique généralE;l et linguistique fran\(aise2, Berne

1944, p. 355. . 6) "The past has been a fact, whereas the future is a mere. notion associated

variously with desire, will, obligation, emotion, and incompletion", W. J. Entwistle, Aspects of language, p. 184. . . 7) Cf. also G. S. Overdiep, Stilistische grammatica van het moderne Neder­landsch, Zwolle 1937, p. 311: if the speaker is convinced of the occurrence of a process in the near future, many verbs can be used in the present.

Page 19: Gonda.1956.The Character of the Indo-European Moods.pdf

30 The Indo-European moods

need. The mode of life and thought of the primitive pre-histol'ic milieu apparently was such that they wel'e not, 01' only in part, requil'ed. But just as peoples living undel' the classificatory system of kinship nomen­clature, calling whole gl'oups of l'elated people by such tel'ms as "mother, brother, etc.", clearly distinguish theil' individual parents from the 'classificatory ones, so peoples looking at the events in nature and society in the aboye way can and do, e. g. for practical purposes and in daily life, distinguish what is earlier from what is later, what happens 01' exists now from what happens or exists at another time. They are in possession of the faculty of arranging events and of visualizing them in the order in which they have occurred in the past and are expected to occur in the future. But, comparatively speaking, this faculty is not much applied in their lives. It remains undeveloped. The idea of time is not completely abstracted from the situation, occurrence, 01'

activity, which takes place at a particular moment 01' during a parti­cular period 1).

It should also be noticed that momentary events, moments of time, and the relations which order them in a series are all known only after a long process of reflection, abstraction, and intellectual construction. The same remark holds good with regard to the faculty of assigning experien­ces to a particular moment or space of time and of determining their chronological relation to other experiences 2). The relations of before, after, and simultaneous with, as given in experience, are not mutually exclusive. Protensive events may very well overlap, and therefore we must recognize that the most general relation between them is that of partial precedence or consequence. So man was, and is, confronted by the problem of defining momentary events, moments, and serial relations of before and after in terms of protensive events and partial precedence.

It takes time to accustom oneself to operate with temporal concepts and to fill them with the content of one's own experience. Statements representing an occurrence as taking place in time, as 'protensive' or as simply occurring bear upon a single fact without presupposing com­parison or combination. Judgements placing occurrences in time presup­pose the speaker's ability to relate these occurrences to other states, events, 01' actions, or to grasp the temporal relations between two 01'

more occurrences. In: the day before yesterday the lightning struok the

1) Cf. N. E. Parry, The Lakhers, London 1932, p. 192ff.: "If it is desired to nx the date of an occurrence, the ordinary way to do so is to refer it to some dennite stage ín the growth of the crop." See also F. Boas, in the Handbook of American Indian Languages, I, p. 35; 42ff.

2) See H. Maier, Psychologie des emotionalen Denkens, Tübíngen 1908, p. 231ff.; p. 234 "Die temporalen Lagebestimmungen sínd durc~weg zeitliche Re~ationen eínes Objekts (z.B. eínes Vorgangs oder Zustands) zu emem anderen ObJekt ... , oder aber temporale Beziehungen mehrerer Objekte zueinander" (no "einfache Elementarurteile, sondern Relationsvorstellungen").

Tense and primitive thought 31

house 01' in: when 1 oame here the lightning struok the house a relation between two past events is formulated. Even in a sentence: the lightning struok the house such a relation is implied, for it means: "at a moment before now the lightning struck the house".

'Primitive' people on the other hand have a dislike of analysis and abstract thought. Where modern scientific man analyzes, they operate with complex ideas and see totalities. Myths, images, visual 01' other sensorial representations help them to understand the phenomena and their interrelations. In qualities without connection with the objects to which they belong, in states disconnected from the objects which are in these states they take no great interest. A percept like river for them primarily is an integration of visual and other sensorial sensations in association with former sensations and experiences of a mass of running water, bathing, crossing to the opposite bank, watering cattle etc.; a 'tree' is a similar integration of sensations in association with former sensations in association with mass, height, shade, and the like, with the impulse near at hand to movement in relation to it,e.g. to climb it, 01' to stand in its shade 1). Even for us the things and phenomena themselves, the objects by themselves, are often of minor importance to the way in which they present themselves to us, 01' in which we see them 2). Primitive man, considering himself to be an integrant part of a whole the other components of which are like himself, attempts to understand these other elements of creation and to get an insight into their essence and character, into their power and manner of being, into their situation and configuration, by attempting to grasp them mentally as wholes in accordance with the way in which they confront and impress him. Like substances, events and occurrences are not, or scarcely, ana­lysed 01' chronologically determined. They too are preferably visualized as complex wholes. Hence, it might be surmised, the frequent occurrence of aspectual rather than temporal categories in languages of ancient and 'primitive' cultures, showing how difficult it is for primitive man to abstract the chronological sequence 01' moment from the events themselves. "Ces morphemes, anciens verbes, ou autres mots amenuisés par l'usage, montrent l'impossibilité ou est le Canaque de saisir le temps; ils révelent son effort pour recouper quelques moments de la durée, et les garder, a cause de la qualité qu'ils conservent a ses yeux. Si le Cana que pouvait tracer le graphique de ces temps de qualité qu'il retient (le calendrier saisonnier: temps de la récolte ... ), ille tracerait par des points, ou, au mieux, il l'écrirait a la fayon des mesures musicales, chaque période figurée par une mesure avec une note ... Et entre les mesures remplies, des pauses. Il marquerait ainsi le rythme de sa vie et de ses

1) See also Murphy, o.c., p. 122ff. For ancient India see various passages in the well-known books by H. Oldenberg.

2) See also H. Gunther, in.the Archiv für Volkerkunde, n, Vienna 1947, p. 69ff.

Page 20: Gonda.1956.The Character of the Indo-European Moods.pdf

32 The Indo-European moods / Tense and primitive thought

valeurs, mais nullement la continuité d'un temps qu'il ne vit pas, et ne soup90nne pas ... Il n'y a pas idée d'écoulé ou de futur, paree que tout cet ensemble reste embrassé dans un seul regard; ce qui le dépasse n'est pas retenu, et demeure hors de l'espace et du temps. On pourrait dire que le sujet ignore la distinction présent et passé, et qu'il emploie done un procédé unique, qui reIeverait des lors du theme qu'on appelle aoriste 1)".

Generally speaking, the primitive visualize-and it is not immaterial to them to do so-events, actions, and occurrences as 'complete wholes', visualize them as. taking place, happening, going on, 01' developing and coming into existence, 01' as completed, finished, having reached a definite ~oal, and so on. What is viewed, in speaking and in formulating thoughts, lS the 'configuration' ofthe events, the way in which they, as totalities or complex wholes, present themselves to the speaker (and the hearer) and the way in which they are important to them. Hence, 1 would suppose, the prevalen ce of aspectual categories in many languages. Hence also the transition of their aspectual function into a more or less temporal role when temporal distinctions are introduced in a more 'modern' way.

1) Leenhardt, Langues et dialectes ... , p. XLVf.-Cf. e.g. also G. Morgen­stierne, in the Norsk Tidsskr. f. Sprogvid. 12, p. 110.

IV. The Injunctive

The 'injunctive' 1) has given rise to much controversial discussion among scholars. After having gradually been distinguished from other forms 2) this "past indicative without augment which serves as an imperative or subjunctive" was claimed for Orig. Indo-European by Delbrück3) and disconnected from the imperfect by Brugmann 4), who was followed by most scholars in regarding it as a separate mode beside indicative, subjunctive and optative. Brugmann's thesis that *bhé1'et "er trug" and *bhéret "el' trage" were not identical was justly attacked by Delbrück5), whose opinion with regard to this point has, as far as 1 am aware, prevailed. Occasionally, and rightly, his sup­position that it was co-ordinated with the other moods was called in question 6). The conclusion drawn by the Father of l.-E. syntax, "daB der Injunktiv ursprünglich eine gegen die Modusbedeutung noch gleichgültige Form gewesen sei" does, however, not always seem to have attracted the attention which it deserved, although it underlies the view expressed by various scholars that the "injunctive' represents a remainder of an undifferentiated category which provided the foundation of the indicative (the additional i in the personal endings), of the Indo-Iranian imperative in -tu (bharat: bharatu, addition of the particle -u etc.), and

1) Strictly speaking, this term which originated with Brugmann (Morphol. Unters., Leipzig 1878ff., nI, p. 2) and was applauded by Delbrück (Vergl. Synt., n, p. 354), should, for reasons implied in the following argumentation, be avoided. The name proposed by Delbrück "improper subjunctive" ("unechter Konjunktiv" ; see Synt. Forsch., I, Halle 1871, p. 5), though adopted by A. Thumb, Hdb. d. Sanskrit, Heidelberg 1905, p.301 and other authors, is equally incorrecto The term 'primitive' (see Schwyzer-Debrunner, Griech. Gramm., I, p. 645) has, for those who are convinced of the comparatively original character of this category, much to recommend it, but for practical reasons the generally adopted 'injunc­tive' may, preceded by the warning 'so-called' be retained. See also V. Georgiev, in the Annuaire de l'université de Sofia 1934, 31,4, p. 62f.; 82ff.

2) The reader may be referred to the survey given by Delbrück, Vergl. Synt., n, p. 352ff., and to BrugmalUl, Vergl. Gramm.2, n, 3, p. 517ff.-Miss Hahn is (o. c., p. 38, § 55) not completely correct in saying that the 'injunctive' is a creation of Delbrück's. .

3) See Delbrück, Die Grundlagen der griech. Synta:x: (= Synt. Forsch., IV), Halle 1879, p. 68. Historically speaking, the inj. is no augmentless past tense.

4) Brugmann, Morph. Unters., nI, p. lff. 6) Delbrück, Vergl. Synt., n, p. 354. A serious disadvantage ofBrugmann's view

is the implication that *ébheret was more original than *bheret "el' trug". 6) See C. D. Buck, Comp. Grammar of Greek and Latin, Chicago 1933, p. 238.

3 Gonda, Indo-Euxopean rooods .

Page 21: Gonda.1956.The Character of the Indo-European Moods.pdf

34 The Indo-European moods

of the imperfect (e-, the augment)1). However, Delbrück himself 2 )

made the suggestion that the 'injunctive' may have been "eine Abspaltung des Indikativs", because it seems to have arisen there "wo die Pro­hibitivnegation mit einer Prateritalform von punktueller Bedeutung sich verbunden hat": "punktuelle Bedeutung" since he was inclined to subscribe to Miller's3) view that the oldest aorist injunctives constituted the oldest mass of this category. But this conclusion 4) is not final: the fact that in the ~gveda5) and in the Gathas 6) ma only occurs in connection with the 'injunctive' does not mean that this formation originally did not occur unless it was accompanied by this particle. Delbrück' s suggestion was indeed dismissed by Streitberg whose remarkable opinions were unfortunately never published in full 7). Emphasizing the fact that in the J;tgveda about 56 % of the injunctive forms are distributed among the several stems of the aorist, of which the strong aorists, esp. the root-aorist, stand foremost S), he nevertheless rejected the viewthat the use of the category under consideration as an 'indefinite present' is the youngest stage in the development of its functions. Without enter­ing, as far as 1 am able to see, into many details as to the difference be­tween present and aorist, he held that as an indo preso the inj. indicated the moment "des Eintrittes bzw. del' Vollendung einer Handlung". "Wenn dieser Moment vom Subjekt erwartet wird, so entwickelt sich die modale Bedeutung des Inj." The' modal', i. e. non-indicative, function was also regarded as secondary by Hirt 9) who, though surprisingly denying any difference between an inj. and a subj .10), attempts to derive the 'modal' function (its use as a subj. and an imper.) from its use as a

1) See e. g. R. Thurneysen, KZ. 27, p. 172ff.; Brugmann, Grundri13l, StraI3burg II, p. 1276; ibid. 2 II, 3, p. 579; 593; Delbrück, Altind. Syntax (= Synt. Forsch., V), p. 360; Hahn, o.c., p. 39. In addition to these see also L. Renou, Les formes dites d'injonctif dans le ~gveda, in the Etrennes de linguistique offertes a E. Benveniste, Paris 1928, p. 63ff.

2) Delbrück, Vergl. Synt., II, p. 355f. See also E. Olafiin, in Language, 12, p. 34. 3) O. W. E. Miller, in the Amer. Journal of Phil., 13 (1892), p.420. Of. also

H. Reichelt, Awest. Elementarb., Heidelberg 1909, p.320. 4) Which was adopted by Brugmann, Kurze vergl. Gramm., StraI3burg 1903,

p.551. 5) See H. GraI3maml, Wtb. zum Rig-veda, 1026f. 6) See Ohr. Bartholomae, Arische Forschungen, Halle 1882ff., II, p. 29 Anm.;

Altiran. Wtb. 1095ff.; Reichelt, Awest. Elementarb., p. 317; 320; 322f. 7) W. Streitberg, Indog. Forsch., 9, Anz., p.170; Verh. d. 44. Versammlung

deutscher Philol. und Schulm. 1897, p. 165f. 8) Theseparticulars are borrowed from J.Avery, The unaugmented verb forms

of the Rig- and Atharva-vedas, J. Am. 01'. Soc., 11 (1885), p. 326ff., esp. p. 329. 9) H. Hirt, Indog. Forsch., 12, p. 212ff. The exposition given in his Indog.

Gramm., IV, Heidelberg 1928, p. 291ff. is substantially similar. Of. vol. VI (1934), p.266ff.

10) Of. also Meillet, Introduction a l'étude comp. des 1. i.-eur.8, p.247, and especially J. S. Speyer, Vedische und Sanskrit-Syntax, StraI3burg 1896, p. 56.

The Injunctive 35

future 1), maintaining that "Formen punktueller Bedeutung in zweierlei Weise verwendet werden konnen, entweder durch Bezeichnung del' Ver­gangenheit als Aoriste, oder als Futura". But here he incorrectly takes no account of psychological time, considering only the chronological sequence of moments in which the 'present' is an infinitesimal point. In a later publication 2) the same scholar greatly errs in holding that the Indian inj. is "einfach futurisch". He alternatively proposed 3) to explain, the origin of this formation from forms which had been infinitives 01' j 'verbalia': "Finden wir doch den Inf. als Imp., aber auch als Inf. hist.,! und das sind eigentlich die beiden Bedeutungen, die wir zur Erklarung des Inj. brauchen". The last clause is, in my opinion, incorrecto In both explanations Hirt widely differs from those who like Macdonell4 ) were inclined to maintain that the inj. originally expressed an injunction, an opinion which, however, cannot be substantiated by a reference to the use of the "2nd and 3rd persons imperfect" as "regular imperatives".

The works of other scholars writing after Delbrück and before Miss Hahn can for the moment be passed ayer in silence5).

1 now proceed to recapitulate the radical views of the American scholar6 ).

While denying the existence of the inj. as a separate unity she dismisses as "really quite unworthy of comment" any discussion of the 'non­modal'-and in her opinion only original-use of an augmentless for­mation to refer to past time. Most scholars will agree with her when she adds that the augment originally was a separate adverbial particle, but, whilst supposing the process of adding this element to have taken place "at a period when the difference between 'present' and 'past' tenses was a matter of aspect rather than of time", she does not explain what was at that period the difference between the forms with primary and those with secondary endings. LikeWhitney 7) Miss Hahn seems, further, too positive in ascribing a past, a present, 01' another function to this category, not asking heTself whether part of the forms must not be regarded as neither past nor present in the sense of our terminology. Re­serving this point fol' consideration till the following pages we only express our dOl1jltS with regard to the argument derived from the comparatively small frequency of the present function: does this mean that this func­tion-if function it be-was less original? Do the occasional cases of the

1) See also Reichelt, O. c., p. 320. 2) Hirt, Indog. Gramm., IV, p. 292; cf. also VI, p. 266ff. 3) Hirt, Indog. Gramm., IY, p. 293. 4) A. A. Macdonell, Vedic Grammar, StraI3burg 1910, p. 316. Of. also Brugmann,

KV.G., p. 551. 5) Some authors of comparative grammars do not enter into any discussion 01'

even omit mentio:r¡.ing the inj.: V. Pisani, Glottologia indeuropea, Torino 1949, p. 203; H. Krahe, Indog. Sprachw., Berlin 1943, p. 119ff.

6) Of. Hahn, o.c., p. 38ff.; 142. 7) W. D. Whitney, A Sanskrit gramma1'5 (Leipzig 1924), p.221, § 587a.

3*

Page 22: Gonda.1956.The Character of the Indo-European Moods.pdf

The Indo-European moods

inj. without ma In the brahmaI).as prove to be a recent phenomenon because they are less in number than the instan ces of an inj. with mil? Apart from this, the relation between the 'present' and the 'past' func­tions and the interchange of present and past tenses (adduced by the authoress as an argument for her thesis that the entire tense system of the Vedic period was in a highly fluid state) cannot be discussed in abstracto, that is to say: without studying the texts in which these forms occur 1). The interchange of 'past' and 'present' in documents deal­ing with mythical events, which are timeless but may be believed to belong to the past, is quite another thing than the same phenomenon when found in a scientific history-book. According to Miss Hahn the use of the inj. to refer to a present time is the result of the fluid tense-system of Vedic. But if we should be able to explain this fluidity, to show that there is some sense in it, this function of the inj., could also be viewed in another light.

1 can on the other hand subscribe to Miss Hahn's view that in the phrase mil + inj. the 'modal force'-that is to say: a force which often is incorrectly called <modal'-lies-or, rather, originally and in ancient texts lied-in the particle. However, her subsequent reflections on a past

I tied.up with the matter of aspect can, in my opinion, be dropped, since l the InJ. was no pasto Her thesis that the use of the inj. as the equivalent of a subj. 01' an opto may be due to a misinterpretation of its use in prohibitions and also to a common confusion between past indicatives and optatives can, it would appear to me, be replaced by a more con­vincing hypothesis which is not founded on the assumption of confusion.

It now remains for to expound my own views. In contradistinction I to Miss Hahn 1 am convinced that the <injunctive' is an old formation, I which already belonged to the Orig. l.-E. period, that it originally was

n_ot ~~_aJlgmentless past indicative. It would appear to me that the vagtieness in meaning and the great, and in the eyes of modern man astonishing, variety of itsfunctions-thatis tosay: thelack of modal and,temporal eJiaracterization-did notonly exist <from th~ beginning',

'1 but'was also typical of forms of this description in different families of languages. Kurylowicz 2) established the indifferent character of the A vestan injunctive, which must sometimes be translated by a pasto tense (e. g. Ga. 29,1), sometimesby a present (e. g. Ga. 32, Il; 50,10) 61' future(e. g. Ga.34,13),-sometimes by a wish orcommand (Ga. 33,7;

1) We shall return to the interchanges between the Vedic verb forms as far as they concern us here in an appendix.

2) J. Kurylowicz, Injonctif et subjonctif dans les Giithiis de l'Avesta Roznik Orjentalistyczny, III (1925/7), p. 164ff. 1 have my doubts about the corre~tness of the author's conclusion (p. 171): "Sémantiquement l'injonctif équivaut a pronom personnel + participe actif ou nom d'agent. Il admet done toutes les nuances d'une phrase nominale."

The Injunctive 37

34, 7). Renou 1) was, in my opinion, perfectly right in concluding from an examination of the facts that intheR.gveda the forms 01 the so-called 'infunctive donot refer totimeor modality, ("indéterminées quant au temps et au mode"), and in adding, in a recent book 2), tb._at."il s'agit d'uneJorlllation indifférenciée, gui en son fond doit appartenir a la couche la .phw.ailciep.ned.tlS mantra". 1 ask, however, myself whether w:e should follow him in believrng that this category the function of which was, he says, "le présent ou peut-etre (in my opinion: «no doubt») plus exactement une notion verbale élémentaire, étrangere a toute catégorie", "s'est t6t imprégnée de valeurs complexes lors de la normalisation du type primaire" 3). It would appear to me that the very possibility of occasional <complex' (so-called modal etc.) value was an original and natural characteristic of this formation.

It has often been noticed4 ) that after the mantra-period the injunctive has rapidly declined. This is in my opinion not to be regarded as a purely linguistic process, but rather as an event of socio-linguistic order: in the mantras so much of the <old world', of the <primitive' mental structure which expresses itself withoutthe help of temporally and mod­ally determined verbal forms survived that the injunctive could be largely retained. In the meantime the 'progress of civilization', 01' the needs of practicallife, had accustomed the speakers of the living language to the use of tense forms and to a greater employment of those modal categories which contributed to the replacement of the ancient injunctive. The ancient Vedic prose was, notwithstanding its many stilistic peculiarities, founded on the spoken language in a later stage of its development than that which must be considered to have been the foundation of the idiom represented by the mantras. 01' perhaps we should 'hther say: it was to a greater extent founded on the language of daily life. That is why the <decline' of the injunctive is, or rather seems to be, so sudden. This formation had been part and parcel of an ancient culture- 01' to express myself more exactly: of a stage in the history of the Aryans, in which a particular <primitive; trait had prevailed-which, in the <brahmaI).a period' had come to an end.

If the aboye views be, in substance, right and if the injunctive can 1I II be considered a,sort of <primitive' comparable to similarcategories in other languages 5), itfüllows thát any attemptexactly totranslate them into a modern Western idiom must turn out a failure. "Die interpreta-

1) L. Renou, Les formes dites d'injonctif cÍans le ~gveda, p. 63ff. 2) Renou, Grammaire de la langue védiqne, Paris 1952, p. 368f. 3) Renon, Les formes ... , p. 79. 4) See e. g. Whitney, 1. c.; Renon, 1. c.; Renon, Grammaire sanscrite, Paris 1930,

p. 409; 414, Speyer, o.c., p. 56, § 187; Avery, o.c., p. 326. 5) See aboye, p. 3lf.-Cf. also P. Kretschmer, in the Glotta, 20 (1931), p.246;

M. Bréal, in the Mém. Soc .. Ling. Paris, 11 (1900), p.268ff.

Page 23: Gonda.1956.The Character of the Indo-European Moods.pdf

--------,-----~--_ ...... - ..... :.,.", .... -------------~---------"--~---"-----------:'.-.-~-=--------

~ 38 The Indo-European moods

torische Schwierigkeit", Delbrück 1 ) observed, "beruht darin, daB man recht oft nicht genau weiB, wie eine Stelle, in welcher ein Injunktiv vor­liegt, zu übersetzen ist. Wer die Injunctive des ~V. durchsieht, erhaIt den Eindruck, als habe el' eine Sammlung del' Schwierigkeiten des Veda VOl' sich ... " So geht denn auch die Auffassung del' Übersetzer oft aus­einander, z.B.: prá roci 1,121,6 «erglanzte» GraBmann, «es strahle hervor» Ludwig ... " The disagreement of the translators has in the 66 years which have elapsed after these words were written, not diminished: whereas for instance arcan 3,14,4 meant according to Gr. "sangen", and to L. "sollen singen", Geldner prefers: "singen"; similarly cetayat 4,1,9: "trieb an" Gr., "leihe Verstandniss" L., "erleuchtet" Ge. and these instances could easily be multiplied. Nevertheless Delbrück attempted to distinguish between an "indicativische Masse" and a "nicht-indicati­vis che Masse", and other authors followed him. Even Renou 2 ), though emphasizing the weakness of the modal and temporal force perceptible­"tantót c'est l'équivalent d'un présent ('général' plutOt qu'actuel), tantOt c'est une forme semi-modale exprimant l'éventuel: intermédiaire entre indicatif et subjonctif ... "; "en tant qu'imparfait ou aoriste inaugmenté, l'injonctif équivaut a un prétérit faible, de type narratif ... "-did not explicitly notice that in many passages any translation is apt to add to the original, to modify it in a more 01' less arbitrary manner. The very 'syntactic instability of this formation results from its being funda­mentally different from our modal and temporal forms. 01', rather, this instability only exists if we try to equalize the inj. forms to ours 01' to those of classical Sanskrit. We must avoid applying this qualification to the prehistoric structure of Indo-lranian.

{

It would even appear to me that translators in clinging to the conviction that an inj. must be equivalent to either a 'subjunctive' (01' future) 01' a

, past (01' present) tense and in forcing their choice upon a particular text have often misunderstood the character of the passages dealt with. Often

¡ the trains of thought of which these texts give evidence do not admit of any modal 01' temporal distinction in the more modern way. And although the use of the inj. in the ~gveda is no doubt not in conformity with its 'original' spread and function, Jllis categ()ry S(3elUso to be r~tained ~specially in th()~e p.I1SSl1ges in which it hadTts. ancient yalue. tt is, inter

) alia~often used in mystic passages, in the so-called cosm.ic enigmas 3),

( references to myths, refrains containing timeless acts and occurrences, a

: universal rule 01' truth etc. 4).

1) Delbrück, Altindische Syntax, p.354. 2) Renou, Gramm. véd., Le. 3) Fo!' the character of which see Renou, in the Journal asiatiqÍle, 237 (1949),

esp. p. 14ff.; J. Gonda, Notes on brahman, Utrecht 1950, p.58ff. 4) See Renou, Les formes ... , p. 76f. For Avestan, Reichelt, o.c., § 660 ("in

Sentenzen oder in sentenzahnlichen Satzen").

1 The Injunctive 39

It may first be remembered that in speechwhich like the Vedic mantras is intended to achieve supranormal ends, the mere reference to mighty powers 'often is an instrument rousing their energy and activity, the mere al1usion to the desired results a means of obtaining them. In ancient societies such as the Vedic and the prehistoric Indo-European the power of words which affirm 01' describe what the operator wants to happen is great. By merely stating a fact, by calling a fact. what is a wish 01' potentiality, 01' by asserting that a desired result has already occurred, primitive man expects to exert influence upon the unseen powers. Without explicitly couching his thoughts and desires in suppli­catory 01' imperative stylíl, in terms of wish, prayer, 01' incantation, he attempts to attain his ends by identifying objects of actual existence with what they"itre supposed 01' desired to be, by reciting stories relating to the fulfilment of a wish, by enumerating the deeds 01' epithets of a mighty being, by praising, i. e. invigorating him. In short he resorts to the power concealed in names and words 1). Thus when it reads AthV. 6,56,1 sárp,yatarfl,1 ná2 ví3 §parad4 vyáttarp,5 ná6 sám7 yamat8

••• the meaning may simply be: "what1 (is) shut1 together1 does2 not2 unclose3

,4

(01': "willnot be unclosed", ideal existence); what5 (is) open5 does6 not6

shut8 together7" 2). At any rate a differentiationbetween "does not ... " and "may not ... " would not be required. Cf. also AthV. 1,8,1 idárf¿l havú,2 yatudhánan3 nadí4 phénam5 ivá6 vahat7 "this1 oblation2 brings7

(01': sha11 bring7 ?) the sorcerers3, as6 a stream4 (does) the foam5"-notice the timeless process referred to in the simile-; 1,3,1 téna1 te2 tanve3 Sárp,4 karam5 "with1 thatl J5 make5 (01': will 1 make5) wea14 for3 your2 body3"­

in stanza 7 the indo pI'. is used: pm . . . bhinaclmi "1 split", and not "1 will split" -and many similar passages. Description of mighty and import­ant entities often suffices to strengthen their power and to enable, 01'

induce, them to be active for the benefit of those praying 01' reciting: the well-knoW11 Indian 'praising' 3).

In other places translations waver between a present and a past tense. However, in the sphere of thought of the ancient Indian poets the diffel'-

1) The reader may be referred to: G. Appel, De romanorum precationibus, GieJ3en 1909, passim; G. Foucart, in Hastings' Encyc1. of Religion and Ethics, IX (1917), p.130ff. R. Thurnwald, in Ebert's Reallexikon del' Vorgeschichte, VIII (19~7), p. 432ff.; H. Webster, Magic, Stanford Cal. 1948, p.92ff.; Gonda, The meanmg of Vedic bhü/?ati, Wageningen 1939, p. 10ff.; M. Müller, Die Stilform del' altdeut­schen Zaubersprüche, Thesis Kiel 1901; A. Wuttke-E. H. Meyer, Del' deutsche Volksaberglaube3, 1900, § 226ff. See also G. vl;tn del' Leeuw, Wegen en Grenzen2

,

Amsterdam 1948, p. 155ff. 2) Aeeording to the eommentary the mouth of a snake is meant: the line belongs

to a text recited for protection against serpents. Whitney and Lanman translate: "may it not unclose" ete.-For the subj. see ch. VI.

3) See also my Remarks on similes, Leiden 1949, p. 75ff.; C. M. Bowra, Heroic poetry, London 1952, p. 276ff. One might also eonsider the timeless eharaetel' of fairy-tales.-The verb forms eharacterized by -a- will be discussed further on.

Page 24: Gonda.1956.The Character of the Indo-European Moods.pdf

40 The lndo-Europeal' moods

ence between present and past time often was not essential; it could even be absent. A mythical event, although it is localized in the past when presented in the form of a narrative, is always renewed and repeated; it really is actual. In myth the past and the present coincide, or the events take place in timeless space. Moreover, the mythical occurrence is repeated and re-actualized by the ritual acts which belong to the present or are described as such 1). Thus the horses of the Maruts ná ... srathayanta2 "d02 notl slacken2, never slacken" ~V. 5, 54, 10; the god Soma always assists Indra, a1though the poets like to refer to particular occasions of his support; hence e. g. 2, 22, 1-3 (refrain) 8aína'f(l- 8ascad (inj.) devó devám; 1;tV. 1,167,4 (after a perfect form) nál roda8í2 ápa3

nudanta4 ghorá5 ju§ánta6 vfdha'f(l-7 8akhyáyaB deválJ9, a1though translated by Geldner: "Die Grausigen5 stieBen4 die Rodasi2 nichtl ab3, die Gotter9

lieBen sich6 die Starke7 zur GesellschaftB gern6 gefallen6", must rather, with Saya:r;ta 2), be taken as referring to a 'timeless' occurrence, Rodasi being always closely associated with the Maruts (cf. 5, 56, 8; 6, 66, 6 where the perfect is used); 4, 43, 5 urúl Vá'f(l-2 ráthalJ3 pári4 nak§ati5

dyám6, where both Saya:r;ta and Geldner prefer the indicative: "weit1

überholt4,5 euer2 Wagen3 den HimmeI6". Although such great mythic events as Indra's combat with Vrtra are as a rule referred to by past tenses (cf. e. g. 1;tV. 1,32,5 etc.) passages are not wanting in which other verb forms are used: the preso indico e. g. 2, 19, 4; 6,68,3; 7,22,2 (cf. also 6,56,2; 57,3; 73,2 etc.), the imperative 1,23,9; 8,89,4, and 6, 44, 15. From the fact that the god is not infrequently described as slaying the demon or similar enemies in the present or is invoked to do so, it appears that he was regarded as constantly renewing the combat (cf. also 10, 124, 6, where the god himself uses the subjunctive). It is therefore not surprising to see Geldner doubt whether the numerous injunctives in a text like 1;t V. 1, 174 are to be taken 'konjunktivisch' or 'praeterital' 3). This document, in which injunctives and imperatives alternate, can indeed be regarded as giving a fair idea of what the pre­historic use of verbal forms may have been: Indra is, in nominal con­structions, said to be king and protector, he is by means of imperatives requested to save those who pray, but his deeds and achievements are referred to by injunctives: rrJór apálJ "thou causest the waters to move", vrtrám . .. randhilJ "thou delivered V. ayer to ... ", etc. The events described are, in fact, beyond time, yet actual. "Mythos und Gegenwart

1) Cf. e.g. also 1,73,6, where the mythical cows ofVala and the cows required in the soma sacrifice are ine:x:tricably identified; see also K. F. Geldner, Der Rig-Veda, l, Harvard 1951, p. 97; Renou, Les formes ... , p. 76f. ,

2) SayaI}.a's commentary in e:x:plaining an inj. not seldom has an indicative. 3) K. F. Geldner, Der Rig-veda, l, Cambridge Mass. 1951, p.252.-For the

nuances e:x:pressed by an inj. cf. e.g. ~V. 1, 167,4 (after a perf.), and 4, 16, 10 (after an imper.).

The lnjunctive 41

flieBen mehrfach ineinander".l», is also Geldner's comment on 1, 121: cí. e. g. sto ll.

In other cases the poets make mention of ritual acts or processes, the occurrence or efficacy of which likewise does not pertain to any fixed time: e. g. ~V. 4, 58, 8 ghrtá8ya dMrálJ 8amídho na8anta, where Geldner, in accordance with Saya:r;ta, prefers a German preso ind.: "die Strome des Ghrta be;rühren die Brennholzer"; cf. also 1,148,5; 2,23,12; 3,54,14; 7; 1,4; 43,4; and AthV. 2, 17, 1-7; 3,4,1; 7,10,1; 56,4; 57, 1 etc. EIsewhere it is a mystical, mythical, or religious truth which is expressed in'the same manner: ~V. 1,61,6; 10,80,2; AthV. 5, 1, 3; 6; 2, 2 etc. The mere reference to mighty beings or events which are con­sidered to be sources of fortune and prosperity, the mere invocation of divine favour, suffices to set power in motion: AthV. 4, 8, 2 túbhya'f(l-l devá2 ádhi3 bruvan4, though translated by: "may the gods2 bless3

,4 youl",

may be more like "the gods bless you!" Cf. also 1;tV. 2, 2, 7; 3, 36, 10 etc. Besides, reference is not infrequently made to timeless facts, to states

which will eternally remain unaltered: ~V. 1, 152,3 the morning-sun rtám1 píparty2 ánrtalJ'ft3 ní4 tárit5 "promotes2 (present) truthl, dispels4,5

(inj.) falsehood3"; to general situations 01' conditions, etc.: 1;tV. 4, 21, 8; 10,29,5; 01' to processes which we could denote by various modal or temporal forms without any substantial difference in meaning, e. g. the frequent formula (pra) vocam (cf. ~V. 1, 129, 3; 154, 1; 136, 6; 5, 31, 6; 41, 14); "1 (will, shall) proclaim".

From the extensive use of the inj. in the Gathas of the Avesta it may be inferred that the contents of these texts often did not require a more exact and specialized expression of tense and mood. In many cases the old form was maintained. A more accurate reference to time would, in these passages alluding to creation and revelation, apparently be a superfluity, and in a sense inconsistent with the mythically timeless and eternal character of the facts referred to and the visionary way in which they are presented and described. Apart from the context, sometimes other words help to mark the time.

It must however be remembered that in a large number of cases the injunctive alternates with other forms 2 ). In the 1;tg- and Atharvasa:rp,­hitas taken as a whole, the flowering-age of the injunctive, which may have existed at a prehistoric period, already belonged to the pasto The conclusion that this form may, 01' must, be taken as having the same value as an imperative or other forms with which it is co-ordinated has indeed often been drawn. Reasoils of rhythm 01' versification, a predilection for traditional phrases and other stylistic factors may have induced poets to compose such lines as Ath V. 6, 131, 2 ánumate,1 'nv2

idá'f(l-3 manya8va4 áküte5 8ám6 idá'f(l-7 namalJB "O Anumati1, assent2

,4

1) Geldner, a.c., l, p.,164. 2) Cf. Delbrück, Altind. Syntax, p. 355.

Page 25: Gonda.1956.The Character of the Indo-European Moods.pdf

42 The Indo-European moods

(imp.) to this3 ; O design5, constrain6,8 (inj.) this7"1). Yet the question might arise, whether there was not a subtle difference between the inj. in J;tV. 5, 70, 1 ... váñsi1 va/rr¡,2 sumatím3 and 2 ... va'l(/,l samyág2 ... í$am3 asyama4 dháyase5-according to Geldner: "mochte1 ich1 eure2 Gunst3

gewinnen1 (inj.)" and "mochten4 wir4 vollstandig2 euer1 LabsaJ3 zu genieBen5 bekommen4 (opt.)". Has the poet in sto 1 avoided using the opt., because this form could leave room for doubt about the realization of the wish1 Be this as it may, we cannot help being under the im­pression that often a single reference to a particular tense or mood could suffice2): Ath V. 3, 12,5 ... sumána1 asas2 tvám3 áthiJhmábhyarp,5 ... rayí'l(/,6 dalp7 "(We) expect2 (subj.) thee3 t02 be2 well-willingl (hope that etc.), then4 thou7 givest7 us5 wealth6". Cf. also such passages as J;tV. 3, 17,5 prá yaja (imp.) ... atha no dha (inj.) adhvarám; 6,44,18 lcah .. . krry.,uhi; 3,54,13 srry.,avan .. . dháta (imper.); 4,34,5 yata (imper.) .. . gman; 1, 186, 5 and 6; 3, 2, 10 yati . .. didharat. -Avestan instances are: Ga. 46, 2; 48,7.

This is not to contend that such modifications as are expressed by those who possess a variety of modal and temporal forms are entirely missing in the speech of people who must 'manage with' injunctives 01' similar undifferentiated categories. As we have seen an indication of what is usually called the 'modal 01' temporal' force often lies in particles 01' other elements of the clause, in the order of words-an inj. expressing a command occupies the initial position e. g. Ga. 30, 2; 31,19; J;tV. 7, 82, 8-, in the modulation of the sentence or in extra-grammatical mea~s of conveying these modifications. The aboye formula pra vocam, for lllstance, is often accompanied by the particle nu which inter alia expresses the sense of "now, at once, now then" or similar 'stimulative or hortative indeclinabilia' and is also often found in connection with a subjunctive or an imperative3): cf. e. g. J;tV. 1, 32, 1; 59, 6; 154, 1; 2, 15, l. A similar force is borne by the emphatical u, u $U etc. which are sometimes rendered by "now, well, at once, etc.": cf. e. g. J;tV. 1,27,4; 164,26; the emphatical sma: 6,44,18. Words denoting a particular moment or space of time may also do duty: J;tV. 5, 82, 4 adya ... saviJ:¿ where the god's bestowing (s.) is said to take place today (a.), i. e. "instantly"; Ath V. 5, 1,4 prá1 yád2 eté3 pratará'l(/,4 pürvyám5 gÚJ:¡,6 "when2 these3 formerly5 went6 further4 forthl ". Similarly such particles as atha which serves to intro-

1) There is abundant evidence of interchange of the inj. with the imp., and wi~h the ~ubj. 01' opt.: see M. Bloomfield and F. Edgerton, Vedic Variants, 1, PhiladelphIa 1930, p. 99ff. and ~05ff.; 109 respectively.

2) For the tendency to avoid 'Übercharakterisierung' see W. Horn, Sprachk6rper und Sprachfunktion, Berlín 1921, p. 54; E. Fraenkel, I.F.Anz. 41, p. 16; Anz. 43, p. 46; W. Havers, Glotta 16 (1927), p. 105ff.

3) 1 subjoin Delbrück's attempt to describe the function of this particle (Ved. Synt., p. 515): "Vielleicht darf man sagen, da13 nú einen Entschlu13 del' sprechenden Person begleitet, welcher sich aus del' gegenwartigen Lage ergibt."

~ The Injunctive 43

duce a new element: AthV. 2, 6, {) "overcome (imp.) all difficulties, thenl

thou4 givest4 us2 wealth3" (átha1smábhya'l(/,2 ... rayí'l(/,3 da?¿4). The pronoun ta- (saJ:¡, etc.) when bearing, in conjunction with another pronoun, a conclusive force seems to have a similar function: J;tV. 6, 33,4 sá l

tvám2 na3 ... avitá4 vrdhé5 bhüh6 "sei6 du2 uns3 ••• ein Helfer4 zum

Gedeihen5" (Geldner):· this phra~e is often found in hortative clauses 1);

cf. 6,4,4; sa. alone can also be used so: ~V. 5, 33, 6. As already observed, an inj. which seems to be translatable by an imp.

sometimes opens the olause: J;t V. 2, 2, 7 001 n02 agne3 brható4 dáJ:¡,5 sahas-1"íry.,aJ:¡,6 "give1 us2, O Agni3, solid4, give5 thousand-fold6

••• "; 6,19,10: as is well known 2) imperatives often occupy this position. But it can also occur elsewhere, e. g. at the end, where it is not uncommon, e. g. RV. 2, 4, 9; 10,47,1. Could the often monosyIlabic injunctive in the final position convey a certain emphasis 13)-

In this connection the use of the adverbal 'prepositions' is also inter­esting. This class of words, though (generaIly speaking) mainly used to convey spatial ideas and to vivify or emphasize, as gesture-words, the conception of direction, are often, by their graphic and suggestive character, also employed so asto suggest, at the same time, 'aspectual, modal or temporal' notions 4). In more or less exclamative clauses a single adverbial preposition can serve as an 'hortative' element: Z 331 aAA; aya "up! arise!"; J;tV. 5, 39,4 índram1 úpa2 prásastaye3 "(1 invite, will invite, letme invite) Indral near2 to glorification3"; 7, 66, 5 prál nú2

yáman3 . .. yé4 ... "on the journey3 (or: this turn) those4 (must, shall, may be) before1 wh04 ... " (for nu see above); cf. also 1, 10, 11 á1 tú,2 ... indra3 . .. piba4 "near1 (here1), 1 pray2, O Indra3 ... drink4

", i. e. "come here ... and drink"5). It seems evident that the prepositions do not only fulfill these functions if "a verb of motion is to be supplied", but also in those cases in which they are accompanied by a neutral and in­determinate verbal formo J;tV. 4, 16, 9 ácha1 kaví'l(/,2 nrmary.,03 gaJ:¿4 lit. "towardsI, O manly-minded one3, (do) you g04 to the sage2"; 7, 62, 2 prá1 mitráya2 várury.,aya3 VOcáJ:¿4 "forthl (do) thou announce(st)4 to M.2 and V.3", i. e. "announce to M. and V.". Those 'old injunctives' which

1) See e.g. the examples adduced by Speyer, Sanskrit Syntax, Leyden 1886, p.344f.

2) 1 refer to my book 'Remarques sur la place du verbe', Utrecht 1952, p. 67. 3) See my article 'Monosyllaba am Satz- und Versschlu13 im Altindischen':

Acta Orientalía, 17, p. 123ff.; esp. p. 126 and 134f.; and M. Bloomfield, Indog. Forsch., 31, p. 170ff.

4) See e.g. Schwyzer-Debrunner, Griech. Gramm., lI, p. "266ff.; Gonda, L'emploi du verbe simple 'au líen d'un composé' etc., Acta 01'., 20, p. 167ff.; F.H.Parigger, Aanschonwelíjkheidsdrang als factor bij de beteekenis-ontwikkelíng del' Latijnsche praepositie, Thesis Utrecht 1941, p. lff.

5) 1 refer to Delbrück, Vergl. Syntax, 1, p. 652; .J. Wackernagel, Vorlesungen über Syntax, 112, Basel 1928, p. 165f.

Page 26: Gonda.1956.The Character of the Indo-European Moods.pdf

--~-" ~""----

44 The Indo-European moods

in the absence of original imperative forms supply, in the 2nd and 3rd

duo and 2nd pI., the place of the latter, are like the genuine imperatives very often accompanied by these adverbial elements: l;{V. 1, 92, 17 á l

na2 úrja1?L3 vahatam4 asvina5 yuvám6 , which originally may have approx­imately meant: "towardsl bring4 vigour3 t02 us2, O Asvins5, ye6 !", i. e. "O A., bring us v." (cf. also 6,59,9 á na ihá prá yaehata1?L rayím). It may parenthetically be remarked that the personal pronoun by adding an element of insistence, may in cases like the one under consider­ation have contributed to the development of the specialized functions of the injunctive. However, the <imperative' or other <modal' use of general verb forms is a widespread phenomenon: it may, in a way, be regarded as being on a par with the same use of the present indicative in classical Sanskrit (e. g. Mbh. 3,53,25 gaeehamol vayam2 apy3 uta4

instead of gaeehama "let! us2 gol als03") and other languages: Dutch je zegt zo iet8 niet "you ought not to say so"; "you must not say things like that"; Lat. vide8 instead of vide, etc. l ).

Another particle of special sense and function was the <prohibitive' má < l.-E. me. Ancient Indo-European was not alone in possessing a special particle expressive of objection or prohibition: in the Polynesian Tongan we find <oua beside <ikai "not" in non-prohibitive sentences; in the Indonesian Malay jañanl takut2 "don't bel afraid2", beside tiada takut "(he, she) is not afraid"; cf. Arab. mii, lam "ov": la "pl¡"; Hungarian nem "ov": ne "P17", etc. In literary and modern Chinese prohibitions are formulated by means of special negatives. In Georgian a fact is denied by ara, a possibility by vera; nu expresses a prohibition. In ,Algonquian an <imperative order' is distinguished from a <prohibitive order'. It seems therefore to be a wide-spread tendency to distinguish between negation and <prohibition', 01' rather: rejection of realization.

~ Now in Vedic ma joins present injunctives as well as aorists, the /j difference between both constructions being mainly a matter of aspect: l;{V 1, 84, 20 etc. má dabhan "let them not harm" (pr.): AV. 5, 30, 1 má nú ga7y, "go not now". Wackernagel in his able discussion of the expressions for negative orders and prohibitions 2 ) seems to have been under a misapprehension in writing that in Anc. Indian "bei V erboten in del' Regel nur Aoristformen des Injunktivs verwendet sind". It is perfectly true that especially in the Atharvaveda the instances of mii + aor. inj. are more numerous, but in both sa:rp.hitas about 50 occurrences of ma + preso inj. can be counted. rhe prOIlOuncedpreference 3 ) fO:r'the _aorist lJlay be understood" from the predilection for the punctual aspectin forml;llatlng prohibitions, but the existence of the other construction

_." -

1) Cf, e.g . .J. B. HofmmID, in the Indogerm . .Jahrb., 2 (1915), p. 93; L. Spitzer, Aufsatze zur romano Syntax und Stilistik, Halle 1918, p. 344, A. 1.

2) Wackernagel, Vor!. Synt., 12, Basel 1926, p.216. 3) For particulars (about 5: 1) see Whitney, Sanskrit Grammar, § 579a.

·"1"-··~ ¿'

1 j

1 ,

~-~-""---.......... .....,..------~-

The Injlmctive 45

~annot be disregarded l). In the ancient _Irania~ di~lects the phrase with t preso inj. even outnumbers the other ~ comb~.nat~ons .. It d?,es .therefore not seem warranted to conclude that es ursprunghch mcht ubhch (war), Verbote im Prasens auszudrücken" so that the Greeks when they com­menced to connect P17 with the preso imp. were not prevented from so doing by the traditional phrase pI¡ + <non-imp~rative': Th~ expl~~ation must, in my (¡pinion, be modified 2): the partlCle *me bemg ongmally usual in connection with all forms of the preso inj., including those which in Sanskrit and Greek survived as part of the imperative paradigm 3), spread (1 would suppose), starting from the latter, 0:ver the genuine imperative forms with which these came to be systematlCally connected: pi} fl.tPYS and pi} fHpY8-r:W were modelled upon pi} ptpYS"íS etc.; ma Mara upon mii bhamtam, má bhamta etc. 4). It would furtherappear to me that these comparatively frequent má constructions contributed to the survival of the corresponding affirmative forms: Alnd. bhamta "bear" (plur.); bharatam "bear" (du.); cf. in Greek cpé(J.s"ís, cpé(J.s"íOy 5

); in Latin lege, legite. These forms were short and therefore apt to express a command;

1) Curiously enough ma + preso inj. do es not occur in the Mahabharata ~here this construction with the aorist, which prevails in other documents (WhItney, Sanskrit Grammar, § 579), is not rare (see A. Holtzmann, Grammatische,s aus dem Mahabharata, Leipzig 1884, p. 20). In later texts the preso becomes very mfrequent (e.g. Ram. 2, 9, 23; grammarians continue incorrectly to call.these forms <a:ug-mentless imperfects'). Although for an explanation ~f th~s development a deta~led description of the relevant facts would be needed, 1t mIght perhaps be surllllsed that a double set of forms carne to be superfluous as soon as the difference in function between present and aorist had become less pronounced (for this point see Speyer, Sanskrit Syntax, § 334); the category which, in point of frequency, was by far the strongest, survived. . . .

2) It may be recalled that in Greek the usual constructlOn to express a prohIbl-tion 01' dissuasion of an individual character is p.~ + subj. of the aor.; p.~ + preso imp.tlts also regular; but the preso subj. does not occur before the Septuagint, and the imp. aor. is very infrequent.

3) Thus, for Greek: Brugmann, Grundri132, II, 3, p. 564; Meillet, Introductio~8, p.236; Meillet-Vendryes, Grammaire comparée2

, p.333; Buck, ComparatIve Grammar, p. 302, etc. Otherwise Hirt, Indogerm. Gramm., IV, p. 119; 140f., who was foIlowed by Schwyzer-Debrunner, O.C., I, p. 799. Hirt neither considered the dual forms nor such relícs of the 'injo' as Gr. Ll 410 p.i} ••• lv{ho "do not put" (see also Wackernagel, o.c., I, p. 214f.). He is, moreover, inclined to regard the element -te as having arisen from a particle (ibid., p. 119). The form.of the typ~s 8tota, juhota(na), which he adduces as an argument, can be explamed otherwlse: see Bloornfield Amer . .Journ. of Phil., 5, p.16ff. and Renou, Gramm. véd., p.258.

4) In A~c. Indian mi'i + imp. does not occur in the Vedas: I?elbrück, Altind. Syntax, p. 361; Whitney, Sanskrit Grammar, § 579~. In the epIcs and .la:t~r the construction begins to appear and becomes an ordmary form of prohlbltlOn.­Moreover it is not surprising that occasionally affirmative and negative commands are mixed: see, also E. H. Sturtevant, A comp. GraIllmar of the Hittite Language, Philadelphia 1933, p. 251, n. 70.

5) One might compare the remarks made by H . .Jacobsohn, Kuhn's Zs., 45, p.342f. .

'IIi

\ \ r

I I

• I

Page 27: Gonda.1956.The Character of the Indo-European Moods.pdf

46 The Indo-European moods / The Injunctive

in many cases their outward appearance was in keeping with that of other much used imperatives with which they were associated: bharata: bhara (2 nd sg.), bharatu (3 rd sg.); 8unuta: 8unu, etc.; f/JB(!B : f/Jé(!BTB -TB

?onveyi?g the idea of the plur. 2nd pers. Last but not least: ori~inal lmperatlves ofthe 2nd pl., the 2nd and 3rd dUo were missing, therewereno competitors 1).

In other functions-dismissing the augmentless imperfect--competitors were ~ot absent: the vague and indefinite injunctive had to abandon the field m favour of the more precise and differentiated indicative sub­junct!ve~ optative and g~n~i~e i~perative: Although some rare instances ?f ma wIth the opt.-thlS ldlOm IS noted Just once in the Veda and very mfrequently later 2)-are found, the negative construction continued to be formed with the old injunctive 3).

,1) Oth~rwise: Schwyzer-Debrunner, O.C., I, p. 799, according to whom rpé(!Bre = rpé(!B (sg. Imp.). + the sec. ending -7:6. They do not, however, explain why the secondary en~~g ,,:as added, nor do they consider the Sanskrit parallel.

I~ contradlstlnctlOn to Miss Rahn (o.c., p. 45) I believe we should consider the Latm type fuam, dulís (cf. eram) and its relatives a sort of injunctives: see Leumann, Lat. Gramm., p. 325.

2) See Whitney, o.c., § 579b; Renou, Gramm. sanscr., p. 412; E. D. Kulkarni, Annals of the Bhandarkar Oro Res. Inst. Poona, 24, p. 83ft'.; Bloomfield-Edgerton, o.c., p. 109f., an~ Rahn, o.c., p. 41, n. 65. As to má + subj. see Whitney, o.c., § 579d.-For *me see also chapter X.

3) The te~acit~ ofthis construction may perhaps be compared to the remarkáble :pref~r~nce, m varlOUS languages, for expressing a negative wish or prohibition by an m:fin~tlv~: P 501; Ital. non fumare; Fr. ne pas se pencher au dehors; "L'anc. fr. exprlI~alt souvent la défense par l'infinitif, surtout la défense négative": Rol 1113 s~re cumpainz, amis, ne l' dire fa (F. Brunot, Lapensée et la langue Paris 1936' p. 562). Of. also L. Spitzer, Italienische Umgangssprache, BOllll-L~ipzig 1922' p. 265f.; (Leumallll-)Rofmallll, Lat. Gramm., p. 590f. '

V. The esselltial fUllCtioll of the Optative

We first come to a discussion of those endings which are characteristic of the <injunctive', to wit the secondary series. Although the verbal endings obviously serve to indicate person and diathesis and although Brugm~nn 1) left the original nature of the distinction between primary and secon~ary series undecided, some authors, among whom is Miss Rahn 2), seem to suggest that there exists an essential connection be­tween the latter group and the conception of past time. At any rate the American author do es not hesitate to associate the secondary endings of the optative-"though whether as cause 01' as effect it would be hard to say"-with the supposed "remoteness" ofthat category3), which in her view served to express "remote futurity, 01' futurity thrown into the past" 4). The curious distribution of the endings-the optative going with the injunctive and the augmented tenses, and the primary endings being typical of the present indicative, with fÍuctuation in the sub­junctive-, was made a problem by Whitney5) and dismissed as insoluble by Delbrück6 ). 1 would venture the hypothesis that the optative has the secondary endings because at the period "in which it came into being" they were the normal 01' at least "neutral" markers of person and diathesis. The distinct temporal value conveyed especiaHy by the augmented forms with secondary endings (and then also by the unaug­mented past tenses) could only exist in opposition to another temporal value borne by the forms with primary endings 7). At aH events it was the present indicative which was conspicuous by its personal endings, not the optative 8 ). We may suppose the primary endings to have had

1) Brugmann, GrundriJ32, II, 3, p. 587. Of. e.g. also Meillet, IntroductionB, p.246f. 2) Hahn, O.C., p. 4lf. " ... past tenses are common in both orders and prohibi­

tions. Several forms of the imperative seem to have secondary endings ... ". See also D. B. Monro, A Grammar of the Homeric Dialect2, Oxford 1891, p. 290 (the optative in its origin was a mood of past time); P. Ohantraine, Grammaire homéri­que, II, Paris 1953, p. 213, and E. Benveniste, Prétérit et optatif en indo-européen, Bull. Soc. Ling., 47 (1951), p. 11ft'.

3) Hahn, O.C., p. 70f. and 150f. See also W. Vondrák, Slavische Grammatik, II, Gi:ittingen 1928, p. 396.

4) Hahn, O.C., p.74 and elsewhere. 6) Whitney, in the Amer. Journal of Phil., 13, p.294. 6) Delbl'ück, Vel'gl. Syntax, II, p. 350: "Ich glaube aber nicht, daJ3 es gelingen

wird, das zu ermitteln". 7) We need nO,t discuss here the problem of the priority of origin: according to

Thurneysen, Kuhn's Zeitschr. and other scholars (see Schwyzer-Debrullllel', o.c., I, p.658) the pl'imary endings were younger.

B) We shall have to discuss the subjunctive on p. 109ft'.

Page 28: Gonda.1956.The Character of the Indo-European Moods.pdf

L

48 The Indo-European moods

something to do with the 'idea' conveyed by the indicative: actuality and the psychological present time.

A large number of views and observations have been published on the subject of the original function of the optative. The main crucial point concerned the chronological relation between the optative of wish and the potentiaP): Delbrück 2), admitting that he was not able convin­cingly to explain the development of the latter from the former, pre­ferred to treat both functions as originally lndo-European, although the essential and primary task of this mode must, in his view, have been to express wish, the subjunctive expressing the notion of "will". For a definition of "will" and "wish" he referred to the philosopher Herbart 3

),

who asserted that "Wunsch ist del' gelinderte Ausdruck für dasjenige Streben, was wir ... mit del' allgemeinen Benennung des Begehrens be­legten ... Was man verlangt, das glaubt man aus irgendeinem Grunde erreichen zu konnen, was man will, dessen Erreichung setzt man bestimmt voraus. Napoleon wollte als Kaiser, und begehrte auf St. Helena ... Wer da spricht: «ich will» hat sich des Zukünftigen schon im Gedanken bemachtigt; er sieht sich schon vollbringend, besitzend, genieBend."

Later scholars who usually were under Delbrück's influence, often differed from him in particulars. Thus Brugmann was ofthe opinion that either of "the two original functions" of the optative may represent the starting-point 4), and one of the most recent grammars, the Greek syntax of Schwyzer-Debrunner5), shares this scepticism 6). The opinion

1) For a summary of views see Hahn, o.c., p. 6ff! I cannot agree with F. W. Householder who in reviewing Miss Hahn's book (Language, 30, p. 399) contends that to ask questions about 'original meanings' does not directly contribute to the advancement of knowledge.

2) I refer to the quotations given by Miss Hahn, o.c., p. 6. See especially Del­brück, Altind. Syntax, p. 302; Die Grundlagen der griech. Syntax ( = Synt. Forsch., IV), Halle 1879, p. 116f.; Vergl. Syntax, n, p.369ff.

3) Delbrück, Vergl. Synt., n, p. 349. As D. omitted giving references I subjoin the following bibliographical note for which I am indebted to my colleague K. Kuypers: J. F. Herbart, Psychologie, n, § 151; Lehrbuch zur Psychologie3 (1887), p. 154f. (see also: Allgemeine Padagogik, p. 330); cf. e. g. G. WeiJ3, Herbart und seine Schule (1928). I wonder whether the many linguists who up to Miss Hahn, following in Delbrück's footsteps, discussed the concepts of "wish" and "will" have ever looked up in Herbart what the father of l.-E. syntax meant by these terms.

4) Brugmann-Thumb, Griechische Grammatik4, München 1913, p. 578; Grund­rill2, n, 3, p.857.

ó) Schwyzer-Debrunner, o.c., n, p. 320, where sorne other opinions are quoted. Other scholars gave up any attempt to find a 'Grundbedeutung', preferring such general formulas as: "der Opto ist der Modus der Vorstellung" (Kühner-Gerth, Ausf. Gramm. d. griech. Spr., n, I, Hannover 1898, p.201).

6) There is no use in referring to the name of this mode - Ohantraine, O. C., n, p. 212f.: "L'optatif, comme l'indique son nom (eVXTtX~ l!yxAH1tC;), e:x:primait essen­tiellement le souhait" -, since this term only gives evidence of the otherwise well­founded conviction of the ancient Greek scholars with regard to the function of this mode in their language.

The Optative 49

held by Hirtl) cannot convine e me: the opto which originally was "timeless" and expressive of the present, could-which i's correct-also refer to past and future prOct3SSes; the future being often uncertain a potential could in the case mentioned last, easily develop. But what about the potential use in connection with past and present processes ~ According to the same scholar the cupitive represented only a parti­cular development of the "imperative" use of the opto But what was the relation between this 'imper.opt.' and the aboye timeless 01' present opto ~ And in what respect did this opto differ from the imperative proper? l cannot help wondering if we were wise to continue in discussing this 'alternative', and ask myself whether the question is well posed if it is formulated in terms of an historical alternative 2).

The key to the problem of the original character and essence of the l.-E. moods should however be sought in the structure of 'pl'imitive' or if one would prefel' anothel' tel'm, pl'escientific thought and speech rathe; than in time-honoured theories ultimately based on an inadequate ana­lysis of the classicallanguages and on European philosophical systems.

Primitive man tries to find an explanation fol' all the happenings in his environment without knowing the facts and relations discovered by modern science. To a certain extent belief in powers and what is often called 'magiC> serves him as a substitute for science. 'Magic' is based on the belief in supranol'mal forces and powers intervening in the lives of men and determining their fate. Although these powers can be in­fluenced and set in motion by man, they are vel'y often supposed to be controlled by gods, demons 01' othel' beings which can be influenced by ritual acts 3 ). The style and the tenor of the wol'ds and utterances pro­nounced on these occasions are not dissimilar to those of the uttel'ances resorted to in imploring and entreating fellow-men. As primitive man is deeply convinced of the gl'eat and often creative power of the spoken word his word can in his opinion influence the unseen powel's just as it can affect the thought 01' action of human beings 4).

Drawing no shal'p dividing line between himself, other beings, and in­animate objects, between natural and supl'anatural phenomena, knowing no boundal'ies between the possible and the impossible, primitive man is apt to project himself, his emotions and aspirations into his surroundings

1) Hirt, Indog. Gramm., VI, p. 277ff. 2) Scholars living in the age of evolutionism and interested mai:rily, 01' even

e:x:clusively, in the investigation of historical developments, were always apt to transpose problems like that under discussion irito questions of priority. See also Lingua, IV, p. 2.

3) See e.g. Van der Leeuw, Religion, p. 23ff.; Thurnwald, in. Ebert, Real1. d. Vorgesch.; H. Webster, Magic, Stanford Oal. 1948; E. E. Evans-Pritchard, Witch­craft, Oracles, and Magic among the Ozande, Oxford 1937, passim; J. Gonda, Notes on brahman, Utrecht 1950, passim.

4) See also Gonda, RemarksonsimilesinSanskrit literature, Leiden 1949, p. 75ff. 4 Gonda, Indo-European moods

Page 29: Gonda.1956.The Character of the Indo-European Moods.pdf

50 The Indo-European moods

and to attribute to animals, inanimate beings and nature will and reason, because they are conceived as powerful. Whether he feels himself depend­ent on foreign-human or other-power or whether it stands confro~t- . ing him, 01' is counteracted by his own will, he always has to reckon wIth powers which have, or are believed to have, a will. And he knows that also in daily life there is power in wishes, threats and commands, even unuttered. In uttering a wish or in opposing his will to another's our distinction between realizable and unrealizable wishes or commands is unessential. Or rather, his formulations about the desirable, about what is to occur is to be avoided, or to be expected do not transgress the range of ~hat is according to his ·primitive· logic, possible, without taking into account what would be really possible. .

So the belief in unseen powers, which may however, mamfest them­selves, and ·magic· and its corollaries, witchcraft, certain forms of mysticism etc., are the very texture of ·primitive man's' thinking. They are lived rather than reasoned about. Experiences of every kind are often results of a coincidence which science must leave unexplained, or which we ascribe to providence, chance or bad luck. The occurrence of unexpected, undesired, unexplicable events, the realization ofpossibilities of any kind depends, according to the ·philosophi of the primit~ve, .on the wishes and intentions of powers and powerful meno The reahzatlOn of a possibility is willed or wished. Re can counteract it by opposing his own will to the other's, just as he opposes, in daily life, his will to that of his fellow-man.

But, being unable to test his ·theol'Íes· by scientific experiments he believes power or powerful beings to be omnipotent. The will of the man who knOWS how to realize it can, in principIe, effect anything. Any wish is within reasonable limits determined by tradition and common sense1

)

a possibility. While believing so, ·primitive man' does not as a rule engage in arguments about abstractions and in disinterested discussions about remote possibilities. Apart from a few exceptional persons, who so to ~ay represent science in pre-scientific milieus, he does not often need speClal­ized conditional, hypothetical, potential and other verb forms and clauses. The Australian Aranda has no real equivalents of "to hope" 01"

"to despair"; he only "expects" 2). Even the Greek BA'lt{~ is not only translatable by our "hope", but also by "expectation" (cf. Plato, Leg. 6440: o6;u fl8AA6'J!'twv) and "anxious thought on the future"; the Lith. viltis "hope" belongs to pavelti "wish, permit", Lat. velle "to will"; the Skt. asa is "wish, desire, hope", but also "expectation 01' prospect"; the identical German word means "BA'lt{~, hope" in Gothic (wens) , "expect­ation, opinion, hope" in O.N. (van), O.E. wén, "supposition, fancy, false

1) He does as a rule not ask for rain in those months of the year in whioh rain usually fails to oome. Cf. also R. Allier, Le non-oivilisé et nous, Paris 1927, p. 84.

2) Strehlow, 1.0. (see oh. Ir).

The Optative 51

opinion, hape, thought" in M.R.G. and N.R.G. (wahn), "expectation, hope, opinion, doubt, erroneous idea" in Med. Dutch (waen), "to think, believe" in English (ween); the Irish doehus "hope, expectation, supposi­tion" is del'Íved froID doig, doieh "likely" 1). The words for "will" and "wish"-to return once more to Delbrück's famous distinction between the subj. and the opt.-are "mixed together" to such a degree that Buck's Dictionary of Synonyms has listed them under one and the same heading 2 ). The majority of the ancient l.-E. words for these ideas, indeed, cover both of tb-em without distinction between the simple "wish" and the notions of "volition" in a more specialized 01' modern technical sense. Words for "possible", properly referring to what is "practicable by those who are able and competent", rather than "what may or may not happen" 01' "what is capable of existing", are in a significant way not infrequently related to, 01' identical with, terms for "being powerful, able, mighty": Skt. sakya- "possible, capable of being", also meaning "able" belongs to sak- "to be powerful, able, competent"; Gr. ovvu't6~ "strong; able; powerful, influential; practicable, what can be done", etc. to OVVUflut; Lat. possibilis to posse.

So, if a prehistoric 01' protohistoric Indo-European, whether he ad­dressed men 01' the unseen, expressed himself by means of an indicative, the process referred to was to his mind actual, even if, from an objective point of view it was not: by using the indicative the ancients, like men living under modern circumstances when their emotion and phantasy have a considerable influence upon their speech 3) could visualize what is not actual as if it were really existent at the moment of speaking. When in the Atharvaveda it, e.g., reads (3, ll, 1) "1 releas e (muñeami) you by oblation", the priest who recites the text is convinced that by the power inherent in the words, together with the sacramental force borne by the rite, the diseased person is actually released from his lllness. By resorting to an imperative he pronounces a command: (st.4) satá'f!b jiva sal'ádaJ:¿ "live a hundred autumns (years)", not (as is often found in translations): "may you live". The subjunctive, to which we shall have to revert, served him to what may broadly speaking be called visualization. The optative, it would appear to me, enables the speaker to introduce the elements of visualization and contingency, the latter being, in my opinion, the main character of this mood. In using this form the ancient Indo-European took, with regard to the process referred to and which existed in his mind, the possibility of non-occurrence into account; he visualized this process as non -actual: it is possible, 01' it is wished for, or desirable, 01' generally advisable 01' recommended and

1) See H. P¡¡dersen, Vergl. Gramm. d. keltisohen Spr., Ir, Gottingen 1913, p.666f.

2) Buok, 0.0., p. 1160f., to whom 1 refer. 3) W. Havers, Handbuc\h del' erklarenden Syntax, Heidelberg 1931, p. 41ff.

4*

Page 30: Gonda.1956.The Character of the Indo-European Moods.pdf

I

l

52 The lndo-European moods

therefore individually problematic; it may be probable, supposed, hypothetical, or even imaginary, its realization is dependent on a con­dition or on some event that may or may not happen. This condition or other event may be expressed, be implicit or even be vaguely or generally inherent in the situation. If this hypothesis be correct it becomes also clear why the opto of wish (cupitive) originally referred to realizable as well as unrealizable wishes. Being the mode of eventuality the opta­tive also renders useful services to those who want to be guarded in what they sayo Whether in a particular case an optative is 'potentiar, 'general', expressive of some wish or other, depends, to a considerable extent, on the situation or the context, and if such should be resorted to, on other syntactical means (order of words, particles, conjunctions etc.).

Although part of these functions or 'nuances' of the opto are too common to need much illustration, it may be of some use to give a some­what circumstantial account, since we break with the traditional presen­tation of the relevant facts 1). We shall however confine our examples to a limited number 2 ). According to Pal).ini 3, 3, 161 liñ, i. e. the so­called optative is used if it is intended to express an injunction or precept, an invitation, a courteous address implying permission (Fr. "auto­risation"), a wish, an inquiry, a request or solicitation (vidhi-nimantra1}a­iimantra1}a-adhi#a-saf!Lprasna-prarthane~tb Uñ): danaf!L dadyat "qu'il fasse un cadeau"; iha bhuñjithiiJ:¿ "mange ici"; yathiiruci bhuñjita "qu'il mange a son gré"; adhyapayet "je souhaite qu'il enseigne" (vyakara'l}am adhiyiya "ich mochte gern die Grammatik studieren") (cf. also Pal).. 3, 3, 153; 157); kif!L nu tarkam adhiyiya "apprendrai-je la logique?"; labheya bhik~am "je voudrais obtenir une aumone"3). It is clear that in all these cases the process is non-actual, but contingent4). Thus we might translate TS. 2,1,2,3 pratheya1 pasubhiJ:¿2 pra3 prajaya4 jayeya5 by "m ay 1 be extended1

with cattle2, with offspring4 be propagated3,5" or "1 wish to be ... ", 01'

"1 should like to be"; ChUpo 7, 3, 1 mantran adhiyiya "1 should like to learn the mantras" or "1 have in mind to learn the m."; Ait. Br. 2, 25, 1 "The gods could not agree on the precedence in drinking Soma; they

1) For Sanskrit see also Renou, Grammaire sanscrite, p. 411f.; W. D. Whitney. A Sanskrit Grammar, § 573, and Speyer's weIl lmown books; for Latine. g. Leumann­Hofmann, 0.0., p.568ff.

2) The order of the instanoes is rather arbitrary. 3) The examples are borrowed from lndian commentaries; the translations are

Renou's (L. Renou, La grammaire de Piiil;lÍni, Paris 1948-1954) and Boehtlingk's (O. Boehtlingk, Pa:r;lÍni's Grammatik, Leipzig 1887).-This enumeration is not meant to be limitative.-In Anc. lnd. the cupitive opto is often aocompanied by the pers. pronoun: ~V. 1, 157,2; 4,50,6.

4) For the ourious preference for the opto of verbs of wishing (Lat. velim, Goth. wiljau eto.) see Wackernagel, Vorlesungen über Syntax, 12, p. 60f. It may be added that this idiom is also found in Sanskrit (cf. e.g. Mbh. 3, 65, 23; 296,24); it was even explicitly mentioned by Pa:Q.ini 3,3, 160 (who adds: vartamane "in the present").

:;a .u

The Optative 53

desired (akiimayanta): "!1 wóuld, will, shall .. ., let me . . . drink3 first 2 " :

aham1 prathamaJ:¿2 pibeyam3 aham prathamaJ:¿ pibeyarn ity (only one of them shall actually drink first). Cf., in Greek lJI151 IIaTe6uAq> ifQw;; u6p,r¡v o'Jiáaalp,t cpéew{)m "1 would fain give unto the warrior P. this lock to fare withhim" (Murray), "1 should like to ... ", "it would be a pleasure to ... " (neither "may" .nor "could" woUld do) and O 45 ustvq> erro 'Jiaea-p,v{)r¡aatp,r¡v I Ttí rp,ev "1 would oounsel ... "1), whereas 0506 vppw ó(JOt-'Ji6ewv 'Jiaea{)stp,r¡v could rather be rendered by: "1 intend to ... , 1 think 1 shall ... ,1 will set before you as wages for your journey ... ". In other passages, always dependent on context and situation, the use of the mood may rather be taken to be determined by a suggestion on the part of the speaker: A 42 TtaStaV Llavaot epa (Jáueva aoí,at {3éAW(JtV. Instances like cp 402 a~ rae MI TO(JaOVTOV oV1}aw~ aVTláastsjl I w~ ... express irony and derision, not a wish in the proper sense of the word. A request, stimulation, order is expressed by sentences like the following: A 791 Tavr sY'JiOt~ 11xtAijt "you could speak thus to A.", which comes to: "please, say" ("du bist vielleicht so gut und sagst das dem A.", Havers); KIlI aAA' sr Tl~ uat T015a(Js ... uaUastsv "1 would that one should go and summon these also"; Q 149 ufjev~ Tt~ oí fl'JiOtTO "a herald could, might attend him" (Dutch: "zou met hem mee kunnen gaan; moest maar met hem meegaan") 2).

The 'difference' between this 'adhortative' opt., used in addressing equals 01' inferiors and the cupitive again depends on context and situa­tion; the opto is the same, expressing visualization and contingency. It is more polite to express oneself in such a way that the person addressed can, at least to the letter of the request, omit performing it. An imperative is, however, not rarely preferred when there is no time to lose, when the person addressed should react immediately: cf. in Dutch pas op as against je mar¡ wel oppassen. As fiu as the opto is concerned 1 recall such frequent constructions as Germ. Sophie, man kann jetzt die Kartoffeln schülen; Fr. pourrai-je vous demander?; Eng. would you mind waitinr¡ ... ? etc., discussed by Havers 3) under the heading "liebenswürdiger Ausdruck einer Aufforderung". In German this use of the opto has led to its be­coming a "courteous imperative", in Slavonic it has even replaced the

1) For this voluntative opto see F. Slotty, Del' Gebrauch des Konjunktivs und Optativs in den griech. Dialekten, Gottingen 1915, p. 76; Reichelt, Awest. Elem., p.318.

2) Cf. also Wackernagel, Vorlesungen, l, p.232f., and Hahn, 0.0., p. 86f. and n. 198. Sorne Avestan instances may be found in Reichelt, O.C., p. 318f. This use of the opto seems to be mostly taken as cupitive in character (e.g. Schwyzer­Debrunner, 0.0., p. 322; Chantraine, o.c., p. 216; Humbert, o.c., p. 119). However, the 'other standpoint' might also be defended (cf. Kühner-Gerth, 0.0., p.230). Compare also suoh phrases as the Engl. could you tell me, Fr. pourriez-voUB, Germ. konntest duo

3) Havers, 0.0., p. 87ff,'

Page 31: Gonda.1956.The Character of the Indo-European Moods.pdf

I

54 The Indo-European moods

imp.I). A kind request in an interrogative form: Ll 9317 eá yv poi n n{{}olo "will you now listen to me 1", cf. also r¡ 22 OVX ay ... i¡y1](JalO ; There is,

1) Irefer to Vondrák, Slav. Gramm., Il, p. 394.-0f. also, in the Romanoe languages, suoh forms as the Fr. veitille, sache whioh originalIy are 'subjunotives', not imperatives.-This does not mean, that the opto was always used where modern politeness-whioh for the rest is no invariable oonoept-would require oiroumlooutional phrases. In addressing the unseen powers for instanoe Vedio man often used the imp., relying on his magio teohnique and knowledge by whioh he oould influenoe them. In Homer the imp. is very often used in pronounoing prayers eto. (Iliad 54, Od. 29, as against 15 and 53 oases of the opto respeotively): J. Th. Beokmann, Das Gebet bei Homer, Thesis Würzburg 1932, p. 49ff.; but we may disagree with the author with regard to the interpretation of these numbers. Various explanations have been proposed, alI of them, however, being one-sided: K. Ziegler, De preoationum apud Graeoos formis quaestiones seleotae, Thesis Breslau 1905 (before the 5th oent. B. O. the Greeks being less "modest" and "reli­gious" would prefer the imp.; he is inoorreot in oontending that the opto in prayers is a Greek idiom, whioh appeared in Pindar's times; see e.g. Waokernagel, K. Z. 33 (1895), p. 28ff.); Beokmann 1.0. (the preferenoe for the imp. in the !liad is due to the oomparatively savage state of oivilisation depioted in it); Havers, in the Anzeiger für Altertumswissensohaft, 4, Vienna 1951, p. 134 (the use of the imp. is rooted in the "do ut des" attitude in praying); M. van del' Valk, Museum, 59, Leyden 1954, 150 ("the person who prays tries to infiuenoe the gods in a magioal way and therefore makes use of the imperative" ; but the attitude of Homerio man towards the anthropomorphio god oannot be oalled magio). T.he imp. is more direot, not impolite; it is assertive and deoided; unlike the opto it does not leave room for any oontingenoy, so it oan be more and more entreating and insistent (of. e.g. A 393ff.; 451ff.; of. 504ff.;r 280; 351; E 117f.; 348; Z 306f.; 477; H 203ff.; K 278ff.; 464; e 445ff.; e 324ff.). Its forms are often stereotype: ó6~, Mre, U(I'IÍr¡vov, uJ.i5fh, auovaov, etc. Signifioantly enough suoh typioal absolute 'prayer' oonoepts whioh do not admit of oontingenoy are always expressed by imperatives: eUw(}e and eUr¡aov "show meroy"; u(}1jr¡vov "fulfill"; uAfHh "hear" eto. The verb *ZJ.r¡pt even ooours only in this form: lJ.r¡fh "be graoious". Of. also the Lat. igno8ce and miserere (see e.g. J. B. Hofmann, Lat. Umgangsspraohe2, Heidelberg 1936, p. 132) and the more 01' less stereotyped but not impolite blessings: xaí(}e, Dutoh vaarwel; eto. In appeals and supplioations the imp. is retained: Soph. Phi1. 501 av pe awam', av p' eUr¡aov. The opto is often (but not always: r¡ 22 etc.) used when a god in human shape oonverses with aman (Beokmann, 1.0.). In oases suoh as B 418; t 534 the opto expresses the aboye nualloes; it may sometimes be due to a tendenoy to avoid the-psyohologioalIy rather than magioalIy?-"dangerous" impero (r 300f. as against 322; v 62ff.; of. also r¡ 148ff. In a prayer beginning with Mre the speaker sometimes adds more speoified and partioular wishes by means of an opt.: Z 476ff. ("give in any oase ... ; please fulfilI also this desire"); v 115ff. The opto is often not the 2nd person and the question may be posed whether a direot address in the 2nd pers. of the imp. would be a oommon alternative (of. e.g. t 534f.) 01' whether the· 3rd person would be syntaotioally equivalent: it expresses a oommand 01' exhor­tation to be exeouted by a person who is not direotly addressed (of. e.g. A 189; see e. g. Ohantraine, O. o., Il, p. 229); so, in oalling upon a god to aohieve a man's belated home-ooming (t 534f.; JI 238ff.; (} 240), it would be somewhat 'illogioal' to use an (aotive) impero in oonneotion with the absent man. We are for the rest reminded of the suppletive relation between impero and subj.-opt. in other langu. ages; see e.g. Dutoh leef "Jife", but zy leve; Meillet-Benveniste, Gramm. du vieux­perse2, Paris 1931, p. 142f.; Oh. BalIy, Linguistique générale et ling. fran9aise2, Berne 1944, p. 179.

The Optative 55

however, genera11y speaking no use in attempting to distinguish between a cupitive or a potential function of the mood in these expressions; in translating sometimes "would", sometimes "could" might be preferred. Some authors considered this 'adhortative' as a special case ofthe poten­tial, others as developed from the cupitive; it would appear to me that this controversy is futile. If, in Sanskrit a word for "hope, expectation, desire, wish" (asarq¿sií) is added, the opto is obligatory (Pal).. 3, 3, 134) sal ced2 iígacchet3, kamaye4 yukto5 'dhiyiya6 "S'2 ill venait3, j'apprendrais6 ,

j'espere4, avec zele5" (Renou). Cf. also pal).. 3, 3, 159 1 ).

In sentences introduced by katham "how" and expressing disapproval the opto and indico are rivals (Pal).. 3, 3, 143): sal kathaYfl-2 südrarq¿3 ylijayet4 "comment2 se peut-il qu'iF fasse sacrifier4 par un südra3 1" In Greek nws ay (X8) + opto serves to express the nuance "how possib­ly ... 1": a 65, we also find the indic.: X 202; nws ,ay + opto frequently helps to express a wish 2). If, in Sanskrit, the sentence is introduced by the interrogative pronoun 01' one of its derivatives, lill (opt.) can also ex­'press disapproval: kol niíma2 südm'f(b3' yajayet4 "quil en vérité2 ferait sacrifier4 par un südra3 1", the future being in this case the alternative (Pal).. 3, 3, 144). The Greek phrase T{s ay (X8) + opto helps to formulate a wish, e. g. K 303. The Skt. fut. and opto can also alternate if the speaker does not believe the content of the predication to be possible 01' does not endure it: nal sraddadhe2 (na3 kf}ame4 ) durjanaly,5 pmtif}tharq¿6 labheta7 (lapsyate) "1 don'tl believe2 (1 don't3 endure4) that a villain5

sha11 find7 a high position6 " (Pal).. 3, 3, 145). A sentence introduced by jatu "at a11" 01' yad "that", the verb of

which is in the opt., expresses the same nuances: jatul v?,f}ala'f(b2 yiíjayen3

na4 marf}ayami5 "1 will not4 everl suffer5 it that he should cause3 an outcast2 to sacrifice3" (Pal).. 3, 3, 147; cf. also 148). The phrases yac ca 01' yatra + opto are used to express disapproval (Pal).. 3, 3, 149) 01' aston­ishment (150). The latter feeling can also be expressed by yadi + opt.: adbhutarq¿l yady2 andho3 girim4 iírohet5 "it would be surprising l if2 a blind man3 would climb5 a mountain4" (Pal).. 3, 3, 151). Such phrases as uta lcu1'yat mean "he will no doubt do" (Pal).. 3, 3, 152): as is well known, doubtless and its synonyms very often imply a certain uncertainty 01'

hesitation: the Engl. doubtless can also mean "probably". Like the other phl'ases mentioned in this paragraph the pl'edication uta 7curyat is con­tingent in chal'acter.

The same mood can help to give uttel'ance to doubt 01' delibel'ation: :?V. 10, 121, 1 kásmai1 deváya2 havíf}ti3 vidhema4 ? "to whichl god2 sha11 we worship4 with an oblation3~" 01' "who is the god whom we should (must, might, etc. wish to) wol'ship with oblations1". In cases like KaI. Sak. 5, 9 kim1 uddisya2 bhagavata3 ... Tf}ayalJ4 p1'eritiílJ5 syul.~6 "fol' what

1) See also Speyer, Sanskrit Syntax, p.263, § 343a. 2) 1 refer to Kühner~Gerth, o.c.,I, p. 235.

Page 32: Gonda.1956.The Character of the Indo-European Moods.pdf

56 The Indo-European moods

reason1,2 may6 the reverend3 have sent5,6 holy men4 ?" the person speak­ing gives also utterance to doubt or uncertainty. Cf. also Pañc. 1, 206 eka'rfb1 hanyad2 na3 va4 hanyéid5 i~U?¿6 "the arrow6 may kill2 one person1 or4

may not3 kiU5 (him)", and in Greek: Arist. PI. 374 no¿ Tt~ ovv TQánotTo; Theocr. 27, 24 ~at "ít .•• éé;atp,t, words spoken by a girl in reply to an un­expected proposal: the contingent opt., which does not pose the alternative as unequivocally as the subjunctive of deliberation, is very suitable to the occasion 1). Cf. the similar difference between Alnd. katha + subj. and katha + opto "how ... ". In Gothic: John 12, 27 ¡va qipau?: Ti eYnOJ;

In precepts and exhortations of general import Sanskrit very often prefers the opt.: Sato Br. ll, 5, 6, 2 ahar-aharl dadyat2 "one should give2 every dayl"; 3, 3, 1 tasmiidl brahmaciiri2 samidham3 a4 haret5 ... "therefore1 a student2 should fetch4,5 fuel3 ... " (the so-called optativus generalis: "ein ganz allgemeines, nicht auf eine bestimmte Handlung bezügliches Gebot") 2): it may be expected that in individual cases the prescription will not be carried out. Greek instances of this use are not frequent, cf. Xen. R. eq. 1, 8 and in a Cypr. inscr.: M)~ot vv {Ja(ftASV~ •.• "í()V XWQov (Cauer-Schwyzer, De.3 679,6, 16). One might recall here the rule prevailing in Gothic, according to which a general prescription or direction obtaining for all relevant cases is expressed by an opto a) Greek instances of the same idiom are {J 232, where after p,f¡ Tt~ e"ít ... rjnto~ e(JTOJ it reads: tIAA' alst xaAsnó~ T' s'h} ~at at(JvAa ééCot and the well-known line Arist. Vesp. 1431eQ(jot"ít~ fjv e~a(JTo~ sMstr¡ dxvr¡v. Compare in Avestan V. 5, 54 U8 tanüm snayaeta "he must (should) wash himself"4). This also brings us to such Skt. instances as Taitt.S. 2, 6, 4, 4 ya'rfb1 dvi~yéit2 ta'rfb3 dhyayet4 "he4 should4 (when occasion arises) think4 of anyonel,3 whom3 he hates2". The indicative phrase ya'rfb dv~ti, referring to an actual "him­who" (cf. SatBr. 1,5,4,12; 8,3,1; JBr. 1, 129; ChUpo 7, 5, 3) would not be translatable by: "(think of) anyone" which involves non-actual objects of the process of hating. In these cases the opto might also be translated by the Fr. "éventuel", or by "perhaps, by chance", cf. also Ath V. 15, ll, 1 yásyailvá'rfb2 vidván3 vrátyo4 'tithir5 grhán6 éigácheP ..• "t06 whose1 house6 a vratya4 Wh02,3 is familiar with this knowledge 2,3

1) Cf. the note by G. Hermann quoted in Th. Kiessling's edition, Leipzig 1819, p. 719.-See also Slotty, o.C., p.80ff.-The difference between katha dasema 1;tV. 5, 41,16 aud katha ... braváma st.11 seems to have been this: the opto enables the poet speaking to hint at the possibility of his failing in performing the act of worship; in using the subj. he emphasizes the uncertainty expressed by "how" leaving no doubt about the-for the present, it is true, notional-actuality of the process. Of course this difference may have become obscured in individual cases.

2) Delbrück, Der Gebrauch des Conj. u. Opto im Sanskrit U. Griech. (= Synt. Forsch.,I), Halle 1871, p.27.

3) See e.g. W. Streitberg, Gothisches Elementarbuch5, Heidelberg 1920, p. 205. 4) Reichelt, o.c., p.319.

The Optative 57

may come7 as guest5 ••• "; Mbh. 13, 35, 23 pratigraha'rfbl ye2 ne3ccheyus4 tebhyo5 rak~ya'rfb6 tvaya7 "you7 should be on your guard6 against those5

wh02 do nota desire4 to acceptl" 1); cf. also Kathas. 25, 24 janiyat1 sa2 vrddho3 jatu4 téi'rfb5 purim6 "that2 old man3, methinks4, will know1 that5

town6". - Passing mention may also be made of the optative of unspecified frequency, e. g. Dem. 30, 20 ... Aap,{JávOJv ~a{F ónO(Jovv (jéOtTo ...

The line B 340 is in my opinion incorrectly regarded as a weakened cupitive by Chantraine2): ev nVQt· MI {Jo'IJAat "íS ysvotaTo ... "one could as well cast all counsels into the fire". It rather gives utterance to the feeling or opinion on the .part of the speaker that he does not care one way or another: "if you behave like this, counsels and plans are worthless and nugatory". There is no sharp line between this use and the concessive opt.: (p 274 enStTa 158 ~at Tt 'JrIÍ1}Otp,t "thereafter let come upon me what may"; cf. 360. Although this nuance has often been considered a weak wish 3), I would prefer to regard it as an expression of indifference, acquiescence and surrender; the speaker gives up any positive attitude towards the content of the predication. The same nuance of "1 don't obje~t, 1 don't care, let one ... " is found in ~V. 7, 104, 15 adya,1 muriya2 yádia yéitudháno4 ásmi5 "noch heutel will ich sterben2, wenn3 ich5 ein Zauberer4 bin5" (Geldner) 4). Cf. also r¡ 224; Aesch. Prom. 1046; Plaut. Pseudo 478 iratus sit. We might subjoin here the case mentioned by Pal).. 3,3, 164: the opto can be used if the speaker wants to express a consent or to give an oppottunity, the process to take place after a short space of time: muhü1·tasyolpari2 kata'rfb3 kuryaJ.¿4 "you could make4 a straw mat3 after2 a momenV,2, tu devrais ... tu peux ... d'ici une heure". Such instances as <p 402 can in my opinion not be considered to be real wishes either: al yaQ (j~ TO(J(JOVTOV ovf¡(Jto~ tIv"ítá(Jstsv I w~ ... , conveying an iron­ical "would that the fellow might find profit in just such a measure as ... " : derision rather than a real wish. In Skt. the opto can bear the force of "to be able to ... " (Pal).. 3, 3, 172): sal bhara'rfb2 vahet3 "he1 is able to carry3 the 10ad2"; the predication does not involve the actual carrying by the person indicated by sa. Neither does it in the cases covered by Pal).. 3, 3, 169: the subject is worthy to perform a specified act: yuval kanya'rfb2 vahet 3 "the young manl is worthy to marry3 the girl2" and 168: kalo1 'ya'rfb2 yad3 bhuñjithaJ.¿4 "it is now2 timel for you4 to eat4" (lit. "this2 (now2) is timel that3 you eat4, will eat4").

This mood can also help to express a merely logical possibility which is not confronted with any reality. K 211 ("My friends, is there then no man who would go among the Trojans, if so be ... he might hear some

1) See also Renou, Monographies sanscrites, Paris 1937, p.34. 2) ChantraiQ.e, o.c., Ir, § 320. 3) See e.g. Kühner-Gerth, o.c., 1, p.238 (where other instances are given);

Chantraine, o.c., Ir, p.215f. 4) Cf. also Hahn, o.c.; p. 102f., n. 259; Vondrák, o.c., Ir, p.396.

Page 33: Gonda.1956.The Character of the Indo-European Moods.pdf

58 The lndo-European moods

rumour ... ), this might he learn (nv{}olTo), and come back (SA{}Ol) to us unhurt": these verbs together with the optatives preceding them con­stitute a merely notional chain of contingencies. Chantraine'sl) opinion that nv{}olTo and SA{}Ol represent an "emploi particulier" of the cupitive opto does not seem probable. Elsewhere the same mood is used when a state 01' action is supposed to set in if specified conditions are fulfilled 01' when another action has taken place: o 181 "so may Zeus grant ... , then will 1 (01': should 1) pray to thee (SVXSTOq!/lr¡y)". The latter predi­cation is contingent; whether it will actually set in, depends on other events. It is neither cupitive nor potential in the strict sense of the termo If one would like to apply one of the traditional terms, "conditional" could serve best. Cf. also Pind. O. 3,45 olJ /llY OlW;W' 'XSlYOr; etr¡y 2). In cases like the fo11owing a hypothesis is implied: Maitr. S. 2, 1,5 "(in this night one should not bring water into the house; verily, water meaning extinction,) one should (might) extinguish" (viz. "if one were to throw water into the house"): samayeyur eva.-As already observed, this mood can under certain conditions and interchanging with a future (Pal!. 3, 3, 9) enable the speaker to hint at fulfilment of a condition in a short space of time.

01' the opto bears the force of "assuming that, supposing that" by the sentence: Brhatk. 21, 36 kascinl mahat2 tapa7y,3 kuryat4 "supposez qu'onl se torture2-4". In Greek we find: ; 193 str¡ /ley yVY . .. sowo~ "suppose, we had food; if only we had food", 01' "would that ... we ... ". In Latin we might compare Cic. Off. 3, 75 at dares hanc vim Grasso . .. in foro, mihi crede, saltaret3). There are many parallels of this "Aufforderung (01':

supposition) statt einer Bedingung". Sentences expressing a supposition, proposal, suggestion, consent, concession, stimulation, 01' condition are often asyndetically fo11owed 01' sometimes preceded by another sentence with which they form a logical whole: Lat. Mart. 8, 56, 5 sint Maecencltes, non deerunt, Flacce, Marones; Germ. ich bin dir nah, du seist auch noch so fern. The verb is often in the imperative: Lat. Plaut. cras petito, dabitur; Dutch (Ieef hem een vin(ler, en hij neemt de he le hand (proverb) "give him an inch, and he will take an e11"; Eng. suppose we take a holiday next week; Fr. jetez-moi dans l'armée, je suis Achille (supposition); faites ce que vous voudrez, je m'en lave les mains (concession); Lat. serva me, servabo te; Gr. A 29 seo'· aTae 015 TOl . .. enalYso/lsy "do thou as thou wilt; but be sure we do not assent"; Germ. Schiller,Wallst. T. 2,3 sei im Besitze, und du wohnst im Recht. In Ancient Indian a condition is often implied in a sentence containing this mood: SatBr. 12, 6, 1,39

1) Chantraine, o.c., n, § 319. 2) Compare the scholion! 3) See e.g. H. Paul, Prinzipien der Sprachgeschichte5, p. 150; F. A. Stoett,

Middelnederlandsche Spraakkunst, Synta:x:is3, The Hague 1923, p. 217; Schwyzer. Debrunner, O.C., n, p.344; Hofmann, in Stolz-Schmalz, Lat. Gramm., p.570; O. Erdmann, Grundzüge der deutschen Synta:x: l, Stuttgart 1886, p. 120.

The Optative 59

kirp,l mama2 tata7y,3 syad4 iti5 "whatl would4 12 get4 1"5 (sc. "if 1 would comply with your request"); cf. also Par.l. 3, 3, 156 where the use of the opto in both the protasis and apodosis of a conditional period is taught: dak§i7y,arp,l ced2 yayan3 na4 sakatarp,5 paryabhavet6 "if2 he went3 to the rightl, the cart5 would not4 be upset6".

The opto can also occur to give utterance to a subjective supposition: Skt. (Pal!. 3, 3, 154) apil simsa2 girirp,3 bhindyat4 "il semble qu'il soit de taille a briserl , 4 la montagne3 avec se tete2" (action qui ne se réalise pas effectivement, Renou), and (3,3,155) sarp,bhavayamil sa2 bhuñjita3 "1 sup­posel that he2 is (will be) able to eat3"; T 321 ov ¡¡,By yete Tl 'Xa'XcfrrseoY UAAO ná{}ollu "naught more grievous than this could 1 suffer" (Murray), 01'

"1 cannot imagine that 1 shall suffer another greater evil"; Mbh. 3,61, 12 vindeta1pi2 sukharp,3 kvacit4 "she mayl somewhere4 attain tol happiness3".

The process can be imaginary: the imagined past is expressed in the Brhatkathasl. 28,41 süryol 'pi2 tad3 ahar4 manye5 bhagn(i6k§a7syandanos

bhavet9 "le chars du soleiP, j'imaginé, avait du9 rompre6 son essieu7 ce3

jour-la,4" (Renou). Of. also Oh Up. 6, 14, 1 yatha1 ••• puru§am2 ••• abhi­naddhak§am3 . .. visrjet4 "just as l ••• one might lead4 a person2 with his eyes bandaged3 ••• " ("imaginez qu'on ... " Senart); N 485 sl yae O/lr¡Al'Xtr¡ ys YSyotW-ÍJa "were we but of like age": the speaker figures to himself a state which cannot be realized. Completely fictitious processes are expressed in the same way: e 455 ovo' {fAa ootr¡r; "you would not give even a grain of salt": there has been no question of salt, but the speaker invents a situation in order to characterize his interlocutor. In Gothic (jabai +) the opto can render the Greek (sáy +) subj.; 1 COI'. 13, 3 jah jabai fmatjau allos aihtins meinos ('Xa¿ Say 1pW/ltrJw ... ).

In the Romeric sentences introduced by eSta "easily, lightly"-cf. y 231 esta {}s6r; y' S{}SAWY 'Xal Tr¡A6{}sy UYOea rJawrJal "easily might a god ... bring aman safe home" ("un dieu n'aurait aucune peine a" Chantraine l), Germ. "konnte")-the optative involves the idea of: the subject is in the speaker's view able to perform a specified deed; paying regard to a certain element of contingency, the speaker expresses a conviction; the utterance as sumes the character of an affirmation. The adverb eSta suggests the speaker's conviction that the positive issue must be empha­sized, is to be expected and, as soon as it has occurred, can be affirmed 2). Needless to say that the frequent occurrence of an adverb in the sentences under consideration can help us more exactly to determine the shade of meaning conveyed by the sentence in itsentirety. The word eeía emphasizes that according to the spéaker the predication can easily be imagined. In a similar way TáXa helps to express any contingency from

1) Chantrq,ine, o.c., p. 217. 2) J. M. Stahl, Kritisch.historÍsche Synta:x: des griechischen Verbums der

klassischen Zeit, Heidelberg 1907, p. 264ff.; 369ff.; B. L. Gildersleeve, Am. Journ. Phil., 30, p. 18.

Page 34: Gonda.1956.The Character of the Indo-European Moods.pdf

60 The Indo-European moods

a probability to bare possibility: Hes. Op. 401 (J¿e; p,sv yae ua¿ Telc; TáXa 7:8v;sa¿.

In passages such as Aesch. Ag. 1049 the utterance has the character of a polite supposition: ansl{}otr¡e; (J' Yawe;. The adverbs can however also be omitted: Epich. 21 K. neaTov p,sv aY u' ea{}ovT' Y(JOle; Vl'V, rJ.nO{}áVOle;: "potential als Ausdruck einer festen Überzeugung" 1); Alcaeus B.4 3, 83 aY u' dnole; Ta {}éASle;, aUoVaale; Tá u' OV {}éASle;. Compare, in Sanskrit, Kathá­sarits. 2, 37 sakrc1chrutamayarp,2 béilalp3 sarvarp,4 c{isdoorayed6 dhrdi7 "and5

this boy3 is able to retain6 by heart' all4 he has heard2 but once1".­EIsewhere the shade of meaning is: the subject may be considered to be able to achieve the impossible (often + api) apibeyarp,1 samudram2 "1 could drink up1 the ocean2"; 01' the opto may almost help to express the idea of "dare to": BárDp. 3, 1,2 katham1 ... bruvita2 "how1 does he dare to say?2, how1 can he say2?,' 01' the idea of "to be entitled, have a right to": Dasakum. 132, 12 caturthapañcamau1 sayita2 "he has a right to sleep 2 the fourth and fifth1 (night-watches)". In negative 01' inter­rogative sentences the opto may indicate improbability 01' impossibility: Mbh. 3, 60, 18 ye~arJ'b1 prakopad2 ... pra'rJayatrarp,3 na4 vindeyam5 "in con­sequence ofwhose1 wrath2 ... 1 canSnot4 procure5 mylivelihood3"; BarD. 3,4,2 na1 dr~ter2 dra~tararp,3 pasyelp4 "you could not1 see4 the seer3 of seeing2"; Ram. 2, 37, 32 loke1 na2 hi3 sa4 vidyeta5 ••• "in the world1 that man4 canSnot2 be found5 (who is not ... )".

Occasionally this mood helps to express what may hereafter prove to be true 01' to have been true: Plato Symp. 175 E rj yae ep,~ (sc. aoepta) epaVAr¡ Tle; I1v.sYr¡. "for it may tur~ out that my wisdom is of a mean kind" 2).

A potentlal In the more strlCt sense of the term is found in cases like ~ 838 nroe; T' &e' eOlTá(Js eeya; "how may these things be?" This "poten­tIaI' also covers undecided possibility and subjective uncertainty with regard to the actuality of a predication and similar nuances. Cf. also B 12 vvv yáe usv lAOl n6Al'V "since-as things stand now-he may take the city"; A 271 ustvolal (J' I1v oV Tle; I ... p,axéolTo "but with them could no man fight". For the "potential' value compare also, in Latin, Plaut. Truc. 221 stultus sit qui id miretur, and the use of such forms as sGÍas' videas etc. This use of the "unqualified" optative must be regarded a~ ancient, because it is also found in Sanskrit A vestan German and· , , , apart from Homer, in various Greek dialects; as it also occurs in the post-classical Greek of the Hellenistic period, it may be said to have been an element in the living language of the Greeks 3).

1) Slotty, o.c., p. 121. 2) See also Goodwin, o.C., p. 79. 3) See e.g. H. Vandaele, L'optatif grec, Paris 1897, 13; J. H. Moulton, Ein.

leitung in die Sprache des neuen Testaments, Heidelberg 1911, p. 312; F. Bla13. A. Debrunner, Grammatik des neutestamentlichen Griechisch6, G6ttingen 1931, § 384ff.; A. N. Jannaris, A historical Greek grammar, London 1897, § 2039; C. Harsing, De optativi in chartis aegyptiis usu, Bonn 1910, p. 28; 52; Slotty, o.C., p. 85f.; Schwyzer-Debrunner, o. C., Ir, p. 338.

The Optative 61

We now come to what has not incorrectly been called the "Optativ del' gemilderten Behauptung", which enables the person speaking to assume a cautious attitude, to make statements with reserve, to empha­size the subjective character of his pronouncements, to be guarded in his speech, to leave the actuality of the predication undecided. In Sans­krit this use is very frequent: Mbh. 3, 64, 66 api1 va2 svayam3 agacchet4 "or2 it may so happen that he shall come here4 ofhimself3" 52, 30 visi~tayal visi~tena2 sarp,gamo3 gU'rJavan4 bhavet5 "the union3 of the best1 with the best2 may be (is, 1 think)5 excellent4". (Cf. also 3,61, Uf. gacchet1 kadi'i­cit2 "she may come1 sometime2", but nilpsarp,sayam1 ... prapsyati2 "she will2 certainlyl reach2"). Thus this mood may be said to represent "a weak future 01' present indicative". In German translations "wohl" might be added: agacchet "er wird wohl kommen (bestimmt weiB ich es zwar nicht)": a typically contingent predication. This use of the mood, becoming more and more colourless, has largely contributed to its assuming the character of a variant of the indicative in later literature, preferred for metrical 01' other reasons1).-In Greek we come across such instances as o 24 avnle; entTebpslae; luam;a "you could yourself put all your possessions in the charge of ... "; and the "Optativ der gemilderten Behauptung": Plato, Crat. 402 A (J¿e; ee; nlv avnlv nOTap,ov ovu I1vep,{Jatr¡e;. By these constructions the speaker is enabled to desist from offering an opinion on the actuality of the process. In the Attic dialect this use led to the so-called optativus urbanitatis: in pronouncing an assertion 01'

opinion one prefers, for reasons of politeness, courtesy, deference, 01'

unpretentiousness the "potential" opt.: Soph. O. C. 647 p,éy' I1v Uyole; &óer¡p,a Tije; ;vvovatae;; cf. also Ai. 185; Hdt. 3, 23. The phrase Xen. Apomn. 3, 5, 7 wea I1v sYr¡ Aáysl'V though intimating that it is high time to speak, allows, in a courteous manner, by its contingent character, the interlocutor to be of different opinion, just like (J 193 nt{}ol6 p,Ol "listen to me" gives him the opportunity not to incline his ear to the speaker. Indian instan ces of the more polite and deferential (' 'rücksichts­voller") opto are not wanting: Mbh. 3, 74,2 prcchethalp "you could ask". This idiom seems to be determined by the same mood and considerations of courtesy which induce Frenchmen to say il peut avoir soixante ans "he may be sixty", and lead English people to prefer a polite I'm atraid 1 don't know to a blunt 1 don't know. The courteous opto is not foreign to the Slavonic languages either.

In Sanskrit this non-compromising opto is also used to refer, in a cau­tious way, to future acts 01' events: Mbh. 3, 65, 14 na1 tvarp,2 vidyur3

janalp4 "(1 have transformed your outward appearance,) people4 will noP recognize3 you2". This idiom is also "weakened" into signifying what

1) Already in a late Vedic text (Vaikh. Sm. S.) the opto and the indo are used (fuhuyát beside iuhoti) without any apparent difference: see W. Caland, Vai­khanasasmartasütram, Calcutta 1929, p. XIII.

Page 35: Gonda.1956.The Character of the Indo-European Moods.pdf

The lndo-European moods

may be, what is likely or usual1) and so even becomes a softened state­ment of what is. On the other hand, a phrase líke Kal. Sak. 2, 17 + kada­cid1 . .. kath.ayet2 may be rendered: "he will no doubt tel12 (it) at some time or other1".

The next point to be considered concerns a use of the optative which has given much trouble to those who cling to the time-honoured alter­native: cupitive or potential, to wit the optativus iterativus. The main difficu1ty found in this use of the mood was caused by the fact that it cannot be held to express a oláf}sal~ '1jJvX*. Thus Jespersen 2), while deeming it correct to say that the moods express certain attitudes of the mind of the speaker towards the contents of the sentence, continues: "in some cases the choice of the mood is determined not by the attitude of the actual speaker, but by the character of the clause itself and its relation to the main nexus on which it is dependent". It seems warranted to suggest the following explanation. The opto is often used in connection with processes which are not always, or not in all respects, actual, pro­cesses which take place intermittently. Recalling the aboye ahar-ahar dadyat "one should give every day" we might also draw attention to instances such as Asv. G. S. 1,3,1 yatra kva cal ho~yant2 syat3 "wheresoever he1 (one) has the intention to (Will)3 sacrifice2", which imply that there may be places where the (general) subject of the sentence does not intend to perform a sacrifice; there may be places where one shall not sacrifice. Cf. also ~V. 10,85,34 süryá1'fl-1 yó2 brahmá3 vidyát4 sá5 íd6

vádhüyam7 arhati8 "derjenige2 Priester3, welcher2 etwa4 die S.l kennt4,

der5 verdient8 das Brautgewand7" (Delbrück3)) or "if there is a priest who ... "4): this statement does not bear upon any priest, and, leaving considerable room for exceptions, implíes contingency: there may be priests who know, well they are worthy of the garment; there may be those who do not know, well they are not worthy of it. TS. 2, 1, 4, 6 SOl 'manyata2 y03 va4 imam5 alabheta6 mucyeta7smat8 papmana9 iti10 "he1

thought2: «he (i. e. anyone) wh03 shall offer6 him5 shall be freed7 from this8 evil9»"; 3,4,8,2 y01 ra~trad2 apabhütal},3 syat4 tasmai5 hotavyalJ8 "they should be offered6 for him5 who1 (if such should be the case, anywhere, at any time etc.) is4 expelled3 from his kingdom2"; Manu 8,358 striyacf[¿l sprsed2 ade8e3 yalJ4 sPNt05 va6 mar~ayet7 taya8 I ... sarvar[L9 Sar[Lgl'aharJar[LlO smrtamll "if one4 touches2 (or: if there is anyone who touches) a woman 1

in a place which ought not to be touched3 or6 allows7 oneselfto be touch­ed5 by her8 (in such a spot) ... a1l9 such acts are declaredll to be illicit intercourse10"5); Pañc. 1,271 ardharajyaharar[Ll bhrtyar[L2 yo3 na4

1) See also A. B. Keith, in the .Y.R.A.S. 1912, p. 774; Renou, Monogr. sanscr. l, p. 33.-Cf. Ramstedt, Einf. Altaische Sprachw., p.85.

2) .Yespersen, Philosophyofgrammar, p. 313. 3) Delbrück, Altind. Synt., p. 339. 4) See my paper on the l.-E. pronoun io-, Lingua 4, p. lff., esp. p. 33ff. 5) l also refer to Delbrück, Gebrauch d.Conj. U. Opt., p. 223; Altind. Synt., p. 340.

The Optative 63

hanyat5 sa6 hanyate7 "if one3 does not4 kil15 a servant (minister)2 who takes away half the kingdom\ one6 is killed oneself7" or "any ruler who does not kill . . .".

The same use of the "relative pronoun" and optative in "relative sentences of general import" is found in A vestan 1). In this connection the use of the opto in similes and comparisons of a general character is also worth mentioning: Ai. A. 3, 2, 1 yatha1 salava1'fl-se2 sarve3 'nye4 var[L-8CilJ5 samahitalJ6 syur7 evam8 "just asl a1l3 the other4 beams5 (whereever they may be, in any such case) are placed6,7 upon the chief part of the shed2, even S08 ... "; ChUpo 8, 3, 2 tad1 yathapi2 hirarJyanidhir[L3 nihitam4 alc~etrajña5 uparyupari6 sañcarant07 na8 vindeyur9 evamlO • •• "just as2

those who do not know the field5 walk7 again and again7,9 over6 the hidden4 treasure of gold3 without8 finding it9, even SOlO ... ".

It may, in brief, be said that the relative sentences of a general charac-ter have a predilection for the opt.: AV. 15, 10, 1 yásyal . .. rájñ02 'tithir3 grhán4 agáchet5 "to the house4 of whatever1 king2 a ... may come5 as guest3"; ¿; 508 Trp 06flsv, o~ ... otur¡v lf}vvTaTa StnOl "to be given to him whosoever... should utter the most righteous judgement"; K 489 8v Tlva T .... n}"f¡!;sls ... I T6v "whomsoever T. smote ... , him"; C 286 ua¿ o' u}"},,n vSflsaw, Tí Tl~ TOlavTá ys éSCOl; l 94. The opto also appears in descriptions of indefinite frequency or quantity, e. g. Dem. 30, 20 }"aflPávrov uaf)? onoaovovv oson:' J!cpopo~ narl afmov.

Here líe, it would appear to me, the roots of the so-called Greek optativus iterativus 2), referring to an indefinite number of identical processes in the past: O 22 8v 08 }"áPOlfll I étnwauov "whomsoever I caught, I would hurl"; 284 ayo(!ñ os B nav(!ot ... I vturov, onn6Ts UOV(!Ol s(!taastav ns(!i flv{}rov "in the place of assembly but few could surpass him, when (as often as) the young men were striving in debate"; in an si clause: Q 768; cf. also X 315. If, however, the person speaking regards "die Mehrheit gleichartiger Falle als ein einheitliches Ganzes ... dessen thatsachliches Vorhandensein betont wird" 3) the indicative is used: Xen. Anab. 1, 8, 1 naaw ol~ svsTvYXavsv sp6a.

It has repeatedly, and correctly, been observed that the iterative nuance is not conveyed by the optative, but by the verb of the principal clause, which in Homer is mostly a form in -aus, by other words such as na~, ast, BuáaToTS, no}"}"ául~, the general 8ans, on6Ts etc.; that it arises from the situation described. Cf. e. g. also such passages as Hdt. 3, 39

1) See also Reichelt, o.c., p. 369. 2) For particulars see e.g. Kühner-Gerth; o.c., l, p.254; .Y. T. Allen, Trans.

Amer. Phil. Ass., 33 (1902), p. 101ff., who was right in criticizing the opinion (W. W. Goodwin, Greek moods and tenses 2, 1889, p. 389; cf. also Brugmann, Griech. Gramm.3, 1899, § 562) that this opto only represents the corresponding subjunctive transferred to thepast); Schwyzer-Debrunner, o.c., H, p. 335f.; Chantraine,o.c., H, p.22M.

3) Kühner-Gerth, 1. C. ,

,

Page 36: Gonda.1956.The Character of the Indo-European Moods.pdf

64 The Indo-European moods

ouov yae levasts areareVB(J<&at, 1táyra oí eXÓJess svrvxéw~; Thuc. 2, 67, 4 1táyra~ ... oaov~ AáfJotsy (ltécp{}steoy. Whereas the "iterative value" is implied in these other words and the context, the opto express es contin-

1 414 " \ I '.Q I '.Q I " I gency. n X ov uya yae usauoy smX'u'ovtWV ay'U'eW1tWY ... OT:t~ acpsa~ slaacpíuotro "for they honoured no one of men upon the earth ... who­soever (in each case, occasionally) carne among them" the verb denotes the process of "honouring" viewed as a series of identical acts of shorter duration happening on various occasions, whereas the opto may be judged as the aboye contingent <generalis' occurring in <relative' clauses.

Nowthis Greek iterative refers to events 01' actions belonging to the past, although sporadic exceptions are not wanting: w 344 for instance may be interpreted in such a way that the opto emfJeíaswy follows the present llaat: llv<&a b' aya aracpvAal 1taywtat llaatY, o1t1t6u M¡ Lltd~ meat emfJeíaswy fJ1tSe<&Syl). The fact that the verb of the principal clause is as a rule in a past tense has no doubt conditioned the relation of these clauses to the past 2). Besides, alargepart ofthose events 01' actions which occur repeat­edly 01' usually, 01' which are qualified by adverbs meaning "over and ayer again", belong, in the experience of daily life, as a rule to the pasto It may be remembered that it was in Greek especially the imperfect which developed an "iterative" 01' "habitual" function 3): K 78 'wad¡e ... , é[> é' o yseatd~ ,ÓJYYv<&', or' e~ 1t6ASIWY ... <&wef¡aaotro "the girdle ... , with which the old man was wont to gird himself, whenever he put his har­ness on for battle". Verbal forms 01' phrases denoting habitual 01' repeated action are not seldom mainly 01' almost exclusively limited to the past4):

thus, in connection with pura "formerly" 01' sma an Ancient Indian present has rather more definitely the value of an habitual past tense: jayanti sma "they were wont to win"; the Engl. use to in the sense of "to be accustomed" is, apart from archaism only usual in the past formo

If expressing repetition, a sentence such as Mbh. 5, 181, 37 tan1 • ••

sarair2 . .. sa'f(tchidya3 bhümau4 ... patayeyam5 "cutting3 them1 asunder3

by my arrows2 1 made them fa1l5 on the earth4"5) may be considered an Indian instance of an optative of repetition. This idiom may have originated in contexts which like the aboye Greek passages were essen­tially <iterative' in character. We might for instance compare such passages as Jaim. Br. 1, 117 katha'f(tl nv2 ima3 aharp,4 prajas5 sTjeya6

ta7 mas sTfta9 na10pacayeyu(¿11 "(el' überlegte): woher kommt es1 doch2,

daS die3,7 Geschopfe5, die7 ich4 erschaffeG, michs, wenn sie erschaffen

1) Stahl, O.C., p.249; E. Hermann, Die Nebensatze in den griech. Dialekt-inschriften, Leipzig-Berlin 1912, p. 163f.

2) See also Vondrák, o.c., n, p.398. 3) Of. also Schwyzer-Debrunner, O. c., n, p. 278. 4) I also refer to Vondrák, o.c., n, p. 373. 5) Of. V. Pisani, Indogerm. Forsch., 50 (1932), p.21.

The Optative 65

sind9, nicht1° ehren11 r' (Caland); Gaut. 12, 1 añgam1 mocyo2 yeno3-

pahanyat4 "(a südra who has intentionally reviled twice-bom men must, shall) 10se2 the member\ with which3 he (i. e. the individual evil­doer who in each particular case is another südra) has every time when such a case arose offended4". Past and many present general (conditional) <relative' clauses are apt to imply repetition of the process.

The iterative-durative function of this mood was for Khotanese estab­lished by Baileyl). In Ossetic 2) we find such sentences as: az dar in éid zayina "1 used (01': made it a habit) to say him". Similar <habitual' functions are also found in other modern Indo-Iranian languages.

The process referred to by the optative, if realized, often belongs, from the objective point of view of time, to the future, but this mode itself is, as far as 1 am able to see, indifferent as to the concept of time. This essentially < achronic' character of the opto is especially evident in such passages as ChU. 4,14,2 ko1 nu2 tva3nusasase4ti5 :: koG nu7 maSnusifyad9

iti10 "WhOl has instructed4 yoU3 ?"5 :: "Wh06 should (01': would) have instructed9 me8 ?"lO; ~V. 8, 96,2 "the shooter has pierced ... ; in that no god ... would have (would, could, etc.) surpassed him (t'uturyat)?"; Ath V.12, 4, 24 "the gods asked the cow ... , that1 same one2 might3 N.4 have known3 (tám1 etárp,2 vidyan3 nárada(¡,4); he drove her away ... " 3). A similar mild assertion 01' measured pronouncement with regard to the past: Hdt. 9, 71 ravra pb ual cp<&6vC{J aY et1totSY "dieses mogen sie aus N eid gesagt haben" 4); in addition to these instances: a 79 YV)l ltey Itf¡r' etr¡~ ... ¡tf¡u yévow "better were it now ... that you were not living, nor had ever been bom"; T90; E85, etc.; Eur.Hel.I215 5). The process referred to by the opto is sometimes explicitly situated in the present time: B 12 vvv yáe usv sAot 1t6AW "now he may take the city"; A 271; cf. also Hdt. 3, 23; Soph. Ai. 186 (context). 01' the process is true of all periods:

1) See H. W. Bailey, Bull. School Oro and Afr. Stud., lO, p. 593; 907 (e.g. the Iast verb in a sentence meaning: "whoever were kings in .J., all blessed him and did his commands"); Trans. Phil. Soco 1945, Asica, p. 35.

2) W. Miller, Ossetisch, Grundr. Iran. Phil., I, Anh. p. 76, 80; Reichelt, O.C.,

p. 308, n. 1. 3) See also W. Neisser, Zs. f. Indo!. U. Iranistik, 5 (1927), p. 283; Speyer, Sansk.

Synt. § 345. 4) For other instances: Kühner-Gerth, O.C., I, p. 232f.-Of. also the interesting

idiom ASv. Saund. 4, 15 bhavec1 002 r~ta3 Jcila4 nama5 tasmai6 "and2 one would doubtless4 have supposed that she was1 angry3 with him6", in which nama conveys the nuance of "only in appearance". As was recognized by many scholars (litera­ture: Renou, Gramm. sanscr., p.412; Edgerton, O.C., p.160f.) texts written in "incorrect" Sanskrit use opto forms in the meaning of past indicatives (cf. e.g. Mbh. 2,76,6 nivavrte1 . .. 1 ••• ca2 ... dyuta'Yfb3 iyát4 punalj,5 "returned1 and2 carne back4 to sit at dice3 again5": see S. M. Katre, .Journal Amer. Oro Soc., 57, p. 316f.; New Indian Antiquary, I, p. 536). It rnay be asked whether this idiom has not arisen from the above use of the opto One may compare the use of the opto "instead of a mild indicative" which resulted in interchangeability.

5) See aIso Vondrák" o.c., n, p.396f. 5 Gonda, Indo-European ¡noods

Page 37: Gonda.1956.The Character of the Indo-European Moods.pdf

66 The Indo-European moods

t 241. A mild assertion with regard to the future: Mbh. 3, 66, 14, and in Sanskrit often with regard to the presento

It would therefore appear to me that in principIe all the aboye nuances of the opto could also occur when the process objectively speaking belongs to the pasto This hypothesis may, 1 think, be preferred to the assumption of special relations between this mood and past tenses 1), because this does not consider the frequent cases of the 'future' opt.; and certainly to Miss Hahn's view that this category originally indicated remote futurity.

There has been much discussion among scholars on a number of A vestan forms "die wie Optative aussehen, aber prateritale Bedeutung haben" 2). For instance: Y. 9, 5 yimahe1 xsa{}re2 ... noi~3 aot'iJm4 ti'iJha5 "in2 Yima'sl (the name of a mythical king) kingdom2 there was5 n03 coldness4 •••

(father and son went forwards 15 years of age,)" yavata1 xsayoi~ (opt.) yimo3 "as long as1 Y. 3 ruled 2 ". 1 would consider this use as follows: here the person speaking is not able, 01' does not wish, to envisage the process as actual; he has no personal knowledge of Yima's reign, and wishing to be cautious and guarded in his statements he leaves sorne room for contingencies. We might suppose the force of the opto to be: "(as long as Y.) is supposed, believed, traditionally held, said, to have ruled"3). The same yavata is followed by the indo preso Y. 43, 8 y.1 {}wa2 mazda3

staomi4 ujya5 éa6 "so long as1 14 praise4 thee2, O M.3, and6 sing ofthee5 ",

by the subj. Yt. 15,40 "(give us a husband who will take good care of us,) as long as we may expect to be (shall be, are) alive": y. gaya jvava. In the other contexts this Avestan opto likewise refers to mythical occurrences, which may have seemed to the poet to be of such a remote antiquity that he hesitated to use a past tense 4).

With regard to the OPers. avajaniya in Dar. Bis. A 1, 50 (§ 13) which is an opto of ava-jan- "to kill" opinions have diverged: sorne scholars following the parallel versions translated this form by a past tense (e.g. Benveniste5): "il tuait [habituellement]"), others as an opto (WeiB-

1) See ahoye, p.47. 2) Reichelt, Awest. Element. § 638, where the literature (as far as published

before the year 1909) is quoted. See also Benveniste, O. O., p. 17. Bartholomae's view that the relevant forms really are athematio aorists (Stud. zur indogerm. Sprachgesch., Ir, Halle 1891, p. 127; Grundr. der iran. Phi1., I, StraJ3burg 1901, p. 81) was rightly rejected by Caland (in his lectures); Tedesco (see further on n. 4); Benveniste (see n. 5); Pisani, Indog. Forsoh., 50, p. 21f.

3) I fail to see why these forms should generally indioate "répétition dans le passé" (Benveniste, B.S.L., 47, p. 17).

4) For other Indo-Iranian instanoes of optatives with preterite value see P. Tedesco, Zs. f. Indo1. U. Iran., 2 (1923), p.298ff. (Sogdian), H. W. Bailey, Bull. School Oro Lang., 10, p.593 (K.hotanese), and Transactions of the Philo1. Soc. 1945, p. 35, see also H. Pedersen, Tocharisoh, Kobenhavn 1941, p. 204ff.

6) Benveniste, in the Transactions of the Philo1. Society, London 1945, p. 50; B.S.L., 1. C.

The Optative 67

bach 1) "el' méichte téiten"). W ould we be correct in interpreting this passage as follows1: "Nobody took the risk of dethroning Gaumáta; the gentry felt much fear of him; he was believed to kill (to have killed, 01': to kill in the future; or: in their opinion he killed, had killed, 01': would kill etc.) many people of their rank ... ": at any rate, the author makes a non-committal statement, and the verbal form does not refer to a parti­cular space of time, although the past is by no means excluded 2). It is however evident that this meaning if it happens to be what the author intended could easily develop into an "habitual past".

1) F. H. Weillbach, Die Keilinschriften der Achameniden, Leipzig 1911, p. 19, of. n. f.

2) E. Benveniste, Transaotions of the Philo1. Soo. 1945, p. 50f.

5*

Page 38: Gonda.1956.The Character of the Indo-European Moods.pdf

VI. The function of the ancient l.-E. Subjunctive

Much has been written on the 'Grund- und Urbedeutung' of the l.-E. subjunctive. Whereas Brugmann, though distinguishing for practical reasons the voluntative, deliberative and prospective subjunctive 1),

was sceptical with regard to the possibility of finding an original function of this category2), other scholars tended to regard, with Delbrück3), the volitional use as the source of the others4). Many authorities however foster the opinion that the 'prospective' function must have been more original. Besides arguments which in themselves and as far as they go are valid-this function "has fallen into disuse" in the post-Romeric Greek main clauoe-, many reasonings have been adopted to convince opponents, which rest on an inadequate philological interpretation of the documents 01' are inspired by what would appear to be a persistent confusion of ideas 5). In contradistinction to Mutzbauer6), who held "die

1) Other authors prefer to distinguish between two main funotions: of. Delbruok, p. 374 volitiver und prospeotiver Konj.; Humbert, Syntaxe greoque, p. 113 "l'idée de volonté et oelle d'éventualité" ("l'év. se fonde sur une prévision de la réalité").

2) Brugrnann, Grundri.l32, In, p.835. Notioe, however, the observation: "daJ3 man auoh anders einteilen kann, ohne unrichtig einzuteilen, ist unbedingt zuzu­geben". Cf. also A. Thumb, in Brugrnann-Thumb, Gesoh. d. idg. Spraohw., I, 1916, p. 86f.; Schwyzer-Debrunner, o.c., I, p.309f. . .

3) Delbrück, Vergleichende Syntax, n, p.349; 374; 365ff.; ConJunctlv und Optativ, p. 16ff. (Conj. des Wollens und Conj. del' Erwartung); Neue Jahrbüoher 1902, p. 330ff. .

4) Cf. e.g. J. M. Stahl, Krit. hist. Syntax, p.23M. (all arguments adduced by this author in favour of this thesis suffer from the same defect, to wit the a priori assumption of the alternative: vol. 01' prosp. subj. Such remarks as "Beim Futurum ist die voluntative Bedeutung offenbar sekundar; dem entspricht es, wenn beim Konjunktiv dasselbe von del' futuralen gilt" do not help us very muoh; Stahl moreover disregards the fact that a prehistoric 'future' and the traditional 'future' of our school grammars are not identical.) Slotty, Gebrauoh des Konj. und Opt., p. 7ff.; F. Sommer, 1.F. 45, p. 60; Humbert, O. C., § 179: "en tout oa~, tout se passe comme si, a en juger par l'évolution ultérieure de la langue, la notlOn de volonté était la plus ancienne" .(?). It cannot be our task here to repeat all arguments of former soholars against those standpoints whioh may be regarded as modifioations of this point of view: whereas, for instanoe, J. Jolly, Ein Kapitel vergl. Syntax, München 1872, p. 37 derived the futural and the general funotions from an originally volitional subj., Mutzbauer (see n. 6), oonsidering the expression of 'Erwartung' to be the primary task of the subj., contends that "Verallgemeinerung dem Konj. überhaupt nicht eigen ist, ... sondern nur von uns hineingedaoht wird" (p. 7).

6) See ohapter 1. 6) C. Mutzbauer, Die Grundbed. des Konj. und Opto und fure Entw. im Griech.,

Leipzig-Berlin 1908, p. 8. Cf. alsoHumbert, o.c., p. 115 (§ 185) "le subj. qui s'attend a voir se produire un fait probable ou normal".

The Subjunctive 69

Vorstellung del' Erwartung" the "Grundbedeutung" of the subjunctive, deriving any Greek uses from this concept, most of these scholars con­sidered it a "future" 1) sorne of them even denying any difference between it and a real future 2). After the expositions given in the preceding chapters it may suffice to observe here that, given the fact that in the early documents the main functions of this mood, as we shall see here­after, are of wide distribution, this functional variety must decidedly be regarded as a prehistoric feature in Indo-European. Applying the commonly accepted rules of comparative reasoning it must be attributed to the Orig. l.-E. subjunctive, whatever its role and position in the structure of the Orig. Indo-European might have been. Any attempt to elucidate the chronological relations of these functions is therefore of a speculative character. Rence the psychological, 01' more 01' less philo­sophical arguments adduced to substantiate quasi-historical theories with regard to the priority of volition, "Erwartung", "éventualité" 01' other functions.

Under these circumstances, it would in the opinion of the present author be preferable to abstain, here also, from attempting to trace, in the traditional way, a process of prehistorical growth-including, inter alia, such 'problems' as the chronological relation between the ex­pression of "Aufforderung" and "Willenserldarung"3)-and to con­centrate our efforts upon the problem of the eS5ential character of this category, underlying all its particular functions, which then must be considered to be aspects of the central function, determined by context, situation, meaning of the verb, modulation of the sentence and similar factors.

Before entering into details it may be useful briefly to state what would be the view of the character of this mood defended in the following pages. Its general function may, if 1 am not mistaken, have been to indicate that the speaker views the process denoted by the verb as existing in his mind 01' before his mental eyes, 01' rather: as not yet having a higher degree of being than mental existence. The subjunctive,

1) See e.g. H. Hammerschmidt, Über die Grundbed. von Konj. U. Opt .•.. , Erlangen 1892, p.22 "del' Konj. ist ursprünglioh ein Tempus, das die Zukunft vom Standpunkt del' Gegenwart bezeiohnet" ; W. D. Whitney, A Sanskrit Grammar (1879), § 575 "there is nothing in the earliest employment of these modes (imper., subj., opt.) to prove that they might not all be specialized uses offorms originally equivalent-having, for instance, a general future meaning"; Goodwin, Synt. of the moods and tenses ... § 6. I also refer to Hahn, 0.0., ch. I and n. Cf. also Delbrüok, Synt. Forsch., IV, Halle 1879, p. 116f.: "Eine andere Mogliohkeit ware" (to regard the subj. as) "die Bezeiohnung del' nahen Zukunft"; Sommer, Vergl. Synt. d. Sohulspr., Leipzig 1921, p. 83; Hahn, 0.0.; of. Sohwyzer-Debnmner, 1.0.; Stahl, Krit.-hist. Synta:x:, p. 163; 228.

2) Hirt, Indog. Gramm., VI, p. 270ff.; A.Waltel', Die Grundbed. des Konj. im Griech., Heidelberg 1923.

3) See Slotty, O.C., versus Delbruck, Conj. und Opt., p. 20.

4 .•. J.

Page 39: Gonda.1956.The Character of the Indo-European Moods.pdf

~ .................. IE .. I5.&.a·· .................... --.... --__ ~ ______ ·o~r¡.;1~5 ____ ~

70 The Indo-European moods

in other words, expresses visualization. A process in the subj. represents a mental image on the part ofthe speaker which, in his opinion is capable of realization 1), 01' even awaits realization 2). There is, however, no question of contingency3). Whether the speaker expects this realization, desires it, fears it, orders 01' hopes it 01' whether he merely sees it before his mental eyes, is a matter of indifference 4). Any implication and specialization: wish, adhortation, deliberation, 'anticipation' 5) depends on circumstances: context, situation, intonation, meaning of the verb etc.6), and in spoken language also ongestures. We should not forget that

1) Kühner-Gerth, o.c., 1, p. 217 held that the subj., being the mood of "die zu erwartende Verwirklichung" was used if "nur die eventueUe Verwirklichung in Betracht kommt". 1 would avoid the terro "eventuell" in defining the character of the subjunctive, because for the person speaking the realization of the process was neither merely possible nor necessarily a result or a consequence of other events (a sense implied in the English use of "eventual").-One should not mistake this force of the subí. for imagination.

2) M. A. Pei and F. Gaynor, A dictionary of linguistics, New York 1954, p. 205 defining the subj. as foUows: "The action or state denoted by the verb is regarded as hypothetical ... ", disregard the difference between hypothesis (in a more 01'

less strict sense of the term) and 'mental existence'. B. F. C. Atkinson, The Greek language2, London 1952, p. 148f. describes this mood as denoting what is outside the reality of fact andwithin that of imagination and hypothesis. 1 would however also object to "imagination", the more so as it will appear in chapter IX that other verbal categories were used to express mental pictures of events not present and mental images of happenings not previously experienced, and not expected to be experienced.

3) Pa1).ini 3, 4, 7 puts the subj. in patati1 vidyut2 "may the lightning2 faU1" on a par with the opto in pated vidyut, which was to survive (cf. also Kühner-Gerth; o.c., 1, p. 225, 8). The words "there is no question of contingency" in the text must be taken to mean that the subj. as such does not imply any reference to contin­gency, not that it necessarily expresses what is opposite to contingency.

4) It is a matter for sorne astonishment that sorne authors should have ex­pressed very peremptory opinions on the necessarily temporal character of clauses containing a subjunctive. See e.g. Stahl, o.c., p. 166; 168, inter alía in connection with p, 96 au6neA.oV ne(J.tltatp,Ó)(J)aa, I ... el nofh p,eiCov l!A.rJat I uiÍ'¡;O, "eagerly searching around the promontory for ... whatever greater fish she-in the mental image called up by the narrator-ever catches" (in English: "she may catch"). Cf. e. g. a 94.-In defining the force of the subj. in the aboye way discussions on the 'vol.' or 'prosp.-fut.' sense of the verb in cases such as 1121; 'Ijl73 (cf. Stahl, o.c., p. 230) loose any importance.

5) G. Hale, The anticipatory subjunctive in Greek and Latin, Stud. Class. Phil., 1, Chicago 1894.

6) We shall do well to consider sorne factors which can also be recognized in studying corresponding verb categories in other languages. Whereas for instance such words as the Dutch zullen, the Engl. shall or will etc. have not everywhere and in every combination lost their original sense of obligation or volition, ancient subjunctives may, in definite circumstances, adopt, or impress us as adopting, a volitional or adhortative meaning. Complications may arise from the fact that questions do not always follow the same rules as assertions, that negative sentences do not always confo1'1n to positive statements, that modesty, emotions, the use of indirect speech often make their infiuence felt.

The Subjunctive 71

what we call the 'subjunctive' is the result of a generalization by which we endeavour to include in a single idea a11 that is common to a great variety of individual forms, each of which has its own meaning and usage. Each subj. stands in a sentence of a particular kind, may be limited by adverbs 01' particles and conditioned by the preceding course of thought1).

As long as the preponderatingly 'modal' and 'aspectual' view of pro­ces ses denoted by verb forms continued this character of the forms under consideration was maintained, but as soon as temporal distinctions increased in importance, the 'subjective' subjunctive denoting what the speaker saw in his mind's eye, was apt to develop into a more 'objective' future referring to what was to come after the moment of speaking2). Thus, it would appear to me, many subj. forms developed into regular futures: Lat. ero; laxo, etc. etc.; Gr. eCJofkut etc. \Vhere, however, the subj. was an integrant part of special constructions-e. g. in post-Homeric Greek the first pers. sing. in the cpé(Js phrases 01' in dubitative questions; the second person in echo-questions; the fixed phrase -d ná1lm etc.-it could preserve its original modal force. Also where it joined other forms to form a complete paradigm: thus the so-called first persons of the imperative in classical Sanskrit are nothing but a renll1.ant of the old subjunctive in its voluntative and adhortative uses; the other forms had become nearly extinct already at an early periodo Generally speaking, this later function involves no change of construction, but only restriction.

As a result of the inftuence oí various factors-context, person, meaning of the verb etc. etc.-a diversity of usages and idioms had, indeed, come into being, in which the use of the mood was more 01' less closely restricted to a particular function. Thus the decline of the subj. in subordinate and other uses in the comse of the Vedic period and its closer contact with the impero gradua11y led towards its specialization in the 'volitionaP function 3); becoming limited to direct speech it helped to express in the form of a quotation closed by iti what other languages would express by means of a subordinate final 01' consecutive clause.

1) For sn attempt to detel'lUine the infiuence of person, number, tense, voice, meaning of the verb, type of sentence (assertive, interrogative etc.) in ancient Latin see E.P.Morris, The subjunctive in Plautus, Amer. Journ. ofPhil., 18 (1897), p. 275ff.-The three verbs in Plaut. Amph. 928 valeas, tibi habeas Tes tuas, 1'eddas meas are, so far as their modal force is concerned alike, but the difference in effect is due to the difference in meaning of the verbs and to the particular way in which they are used.

2) It may be remembered that expressions fOl: the future in other languages also proceeded from forms bearing such a force as to imply futurity ("bei denen ... del' Hinblick auf die Zukunft eine notwendige Begleitvorstellung bildet", Behaghel, Deutsche Syntax, n, p. 257).

3) 1 refer to Renou, La décadence et la disparition du subj. (Mon. Sanskr., 1), Paris 1937, p. 16ff., and p. 18: this subj. "se sépare du souhait (opt.) ... , en ce que cette intention est prete a se manifester" (brahmaJ;las).

'1

Page 40: Gonda.1956.The Character of the Indo-European Moods.pdf

72 The Indo-European moods

So the picture presented by classical Greek: a surveyable number of definable uses in particular constructions does not give a reliable im­pression of the original state of affairs; it represents on the one hand a limitation and on the other a functional specialization, fixation and differentia tion.

There is-as we have already seen-no denying that, if we think in temporal concepts, the processes referred to by a subjunctive should they be realized, generally speaking, belong to a time subsequent to the moment of speakingl). But it does not follow that an ancient speaker primarily wished to stress this fact. Besides, instances are far from wanting in which the process does not belong to the future. It may be 'timeless', of general validity: b 69lf. ij r' B(17:¿ M'Xr¡ {}etO)'IJ (JuatJ..l¡o)'jJ· CiJ..J..ov 'X' ex{}u{r!'llat (J(!o-ewv 2 ), CiJ..J..ov 'XS rplJ..O{r¡ "one man they hate-a process existing in the speaker's mind, for the moment not attributed to any specified king, but apt to manifest itselfin particular persons-, and another they love-though this love may fail to turn up-" 3). Conditional and other uses of the mood are often timeless: Pind. 01. 6, 11 noJ..J..ol be [u3ftVUV-eUl, 'XuJ..dv el n novr¡{}ñ.

The subj. sometimes refers to a time preceding the moment of speaking: K97ff. "come, let us go to the sentinels, that we may look at them, lest ([di) ... they have fallen asleep ('XOlftnaWVTal) and have wholly forgotten (J..á{}WV-eUl) their watch". It would be incorrect to consider 'X. and J... cases of 'attraction', the verb "looking at" being in the same mood (ibwftsv), 01' to contend that the states of sleeping and forgetting are necessarily to be understood to begin in the period elapsing between the moment of speaking and that of 100king4). It may be suggested explaining these subjunctives also as indicative of 'mentalreality': "let us go (subj. so-called adhort.) and (subj.), lest they have fallen asleep etc. (subj.)": all these processes for the moment exist only in the speaker's mind. Similarly, and unambigu­ously, v 215f. "Come, I will ... go over (lbwftul) the goods, lest to my cost these men have carried off (OlXWV-eUl ... Ciyov-es!;) aught with them" (Murray): "have carried off": cf. sto 120ff. Compare also A 555 "now I fear, lest Th. have persuaded you ([tI¡ as nu(!s{nrJ)" where both the 'past' and the 'future' interpretation have been defended 5). Questioning in any case the correctness oí Goodwin's opinion 6) that the use oí aoristo subj. in a reíerence to the past must be attributed to "an earlier laxity of usage", I would deem both translations possible, but the former probable.

1) "Der Konjunktiv bezieht sich ... zunachst auf die Zukunft", Kühner-Gerth, O.C., I, p.217.

2) "AIs allgemeine Erfahrung", Slotty, o.c., p. 131, § 306. 3) D. B. Monro, A grammar of the Homeric dialect2, Oxford 1891 "a king will

(is sure to) hate one, he may love another" (see also Hahn, o.c., p. 84). 4) " ••• von der Besorgnis wegen eines bereits eingetretenen Factums" (Faesi­

Franke). 5) I refer to Hahn, o.c., p. 106f. 6) Goodwin, o.c., p. 27.

The Subjunctive 73

See also K 538. Yet, reference to the past is, on the whole, oí infrequent occurrence 1).

The 'timeless' use oí the subj. to which we may, for a moment, revert, is obvious in similes and comparisons which are instructive also in other respects. As is well-known, the verbs in an Homeric simile often are in the subj. 2): E 161 "even as a lion ... breaks (Ci~rJ) the neck ... of a cow ... , so the son oíT. thrust ((Jfías) them ... ". This subj. - which cannot reasonably be considered either voluntative 01' prospective­refers to neither past nor íuture; it corresponds to the process oí breaking in the mental image of the speaker which he wishes to create in the hearer 3). The verb of the governing clause refers to the past: K 486 W!; be /..BWV ftl¡J..OUJlV ... evo(!ovarJ, I w¡; ftev encpxs-eo, a feature noticed but not explained by StahI 4 ). It is, as far as I am able to see, used in connec­tion with more or les s incidental, though by no means fantastic, events and occurrences, not with normal situations or frequent happenings. In the latter case the poet often resorts to the 'actual' indicative: B 462 "even as tribes of wild geese etc. fiy (no-ewv-eat) this way and that"; 470 "even as the tribes of swarming fiies which buzz to and fro (~J..áa'XoV(Jw) throughout the herdsman's farmstead"; r 3ff.; 0620. Here the poet simply refers to well-known facts of everyday experience. When he prefers the subj, he calls up a mental picture of an event which has perhaps not been witnessed in reality by every hearer, but may occur in nature 01' human society, whenever the West Wind drives (a-evrpsJ..{~'[)) the clouds of the South Wind (A 305); whenever an ash-tree is cut down (N 180), a hawk darts over the plain to chose another bird (N 64), 01' blindness of heart comes upon aman (Q 480). Cf. also LI 141; K 183; 486; P no; 134.

That the subj. corresponds to mental 'reality', to visualization rather than to processes which "als nur unter gewissen Bedingungen, vor­kommenden Falls eintretend bezeichnet werden sollen" 5), may emerge

1) In Plautus, the subj. of the past is comparatively infrequent (84: 1366 pres. subj.). "It is only in a complicated style, in the complexity of the conditional sentence or the subordinate clause, that the plupf. finds a place (7 instances in Plautus)" (Morris, o.c., p.281).

2) For particulars see Kühner-Gerth, o.c., I, p. 25lf. ("es ist hier gleichgültig, ob im Hauptsatze ein Haupttempus oder eine historische Zeitform steht"); Stahl, o.c., p. 247f.; Schwyzer-Debrunner, o.c., Ir, p. 284; 312f. ("altertümlich").

3) How could Stahl's definition ofthe subj. (o. c., p. 241: "die Konj. bezeichnet Verwirklichung, sowohl erstrebte wie zukünftige") be of any use here? Nor can Gildersleeve's description of the mood do duty: "The subj. mood is the mood of anticipation. It anticipates the realization of the predicate, which anticipation appears chieflyas an act of the will" (B. L. Gildersleeve, Syntax of Classical Greek, New York 190~, p. 147).

4) Stahl, o.C., p. 247. I am not convinced by the explication proposed by this author (subj. "der individualisierenden Zusammenfassung").

5) Kühner-Gerth, o.c., I, p.251.

Page 41: Gonda.1956.The Character of the Indo-European Moods.pdf

74 The Indo-European moods

from such passages a¡; 1481 "he cherished me as a father cherishes (¡¡nA1}an) his only son"; 323 "even as a bird brings (neocpsenat) ... to her chicks whatever she may find (xs Aúf3nat)" ; M 167. Sometimes the poet, after a subj. in the opening line of asimile, passes on to an indicative, shifting from °the image conjured up in his mind to <reality', and in so doing enhancing the vividness of the simile and suggesting the actual character of the situation in his narrative: B 147f. "and even as some­times the West Wind stirs (xw1}an) a cornf1eld ... , and the ears bow (f¡/-rúst) under it ... "; LI 141ff.; 1 323f.; M 167ff.; 8 16ff.; O 579ff.; JI 298ff. etc. It cannot be maintained that this subj. is either "volitional" 01' "prospective"; it is, as far as 1 am able to see, the mood of "visuali­zation". In interpreting the words "1 imagine!" 01' "fancy!" might be (tacitly) added to the verbo

It is worthy of special notice that there exist also Vedic instances of the subj. in similes. Whereas in the yatMi ... evli (evam) similes, which express in the first place agreement in essence, resemblance in manner, nay, exact parallelism, and which are often didactic 01' ritualistic l ),

the indic. is the natural form of the verb, the subj. sporadically occurs in passages of different contents: ~V. 1, 51, 12 yátMil sutásome§u2

clikáno3 ••• slókam4 á5 rohase6 diví7 "asl thou art much pleased3 with those who press out the soma2, thou attainest5, 6 to renown4 in heaven7 ";

7,58, 3 gatól ná2dhvli3 ví4 tirliti 5 jantúm6 "as2 a path3 walked overl

leads4, 5 the men6 (thus he may lead us further)" ; cf. also ~V. 8, 66,4 "(Indra may accomplish) as he is determined to desire": lcrátvlil yáthli2

vásat3 (lit. "as2 he desires2 with determinationl ") 2).

Those scholars who deny the existence of any perceptible difference between a <prospective' subj. and a future 01' are at least convinced that these categories are "particularly close" 3) run the far from imaginary risk of blurring out subtle distinctions which may always be found to exist when a language possesses two apparently synonymous categories 4).

1) Sea my "Remarks on similes in Sanskrit literature", Leiden 1949, p. 102ff. 2) It may be remembered that the so-called 'possible future' (sambhávya

bhavif;yat) of Hindi which really is a sort of subj.-as far as it denotes futurities it does not represent them as realities-is likewise used in comparisons, denoting that with which the comparison is made, not as an objective reality, but mere1y "as a supposed case" (S. H. Kellogg, A grammar of the Hindi language3, London 1938, p. 458).

3) Hahn, O. c., p. 80. This authoress views both cat egories as having a common originó in the course of time they differentiated- a process begun already in Indo­European-so that they could easily follow different lines of development, with the result that ultimately they would not necessarily correspond in every case (see also p. 61).-Otherwise: Humbert, o.c., p. 115 (§ 185).

4) Stahl's (o. c., p. 168) view of LI 191; l[F 112; <529 etc. ("del' Konj. des Aor. steht in posteriorischem Sinne statt des Futurums") do es not convince me: the physician in Ll 190f. shall lay (em{}f¡(m) simples on the wound which may be e:x:pected to allay the pain.-In concluding from the co-existence of a subj. and

The Subjunctive 75

The interpretations propounded by Walter undeniably suffer from this defecto Although this author is certain of the equivalence of both forms in such lines as n 437 ovx lJaf}' OV7:0~ d'/!~e ovo' lJaas7:a.t ovos ys'/!r¡7:m l ), there may be room for the observation that lJaaS7:al is, in Homer, often used as an emphatic form, preferred in positive and assertive speech: LI 164 etc. lJaaS7:al 1í/-tae 87:' iiv ... ; B 347; A 824, a nuance foreign to ys'/!r¡7:al 2 ).

Besides, Walter's translation: "El' wird auch nicht sein und auch nicht geboren werden", implies a tautology. An English " ... shall not live, nor can 1 imagine the birth of such aman" would, because of its length, be no adequate translation, but, 1 am sure, a tolerably satisfactory interpretation. In a similar way, A 262 ov yúe nw 7:o{ov~ roo'/! &.'/!sea~ ovos Yowl-lal is not exactly "und llicht werde ich sehen"3), but rather " ... 1 don't expect to see ... , 1 can't fancy them being seen again". In contra­distinction to the German author 1 would maintain that in Z 459ff. and H 87ff. xaf n07:s Tl~ dnnO'l ... W~ n07:s Tl~ 8eSsl the two verbs are not on a par4 ). Delbrück's interpretation of 8eSsl ("del' Gedanke ist ... wegen seiner Wichtigkeit als selbstandiger ausgedrückt") should be modified to: "the thought is, with some emphasis, repeated". The nearest approach to a correct translation might be this: "and someone will (1 expect, 1 am sure, 1 am convinced) say some day (01': so might some man speak in aftertime) ... } so will some man say", in Dutch: "en eens zegt dan iemand ... , zo zal eens iemand zeggen". This repetition in a different mood, which is more <emphatic' thall the former, can, 1 think, be best

a future in Homer that the "futurale Bedeutung für den Konj. nicht wesenhaft ist" (o. c., p. 235) the same author entirely overlooks the difference between these categoríes. In his subsequent argumentation in favour of the priority of the vol. subj.-"Der Ubergang zuro futuralen Sinne war dadurch gegeben, daJ3 im Willen das Streben nach zukÜllftiger Verwirklichung einbegriffen ist"-Stahl, moreover, disregards the methods of modern linguistics.

1) "Die Parallele zwischen l!aa8Tat und yéV1)Tat ist so auffallend, daJ3 es geradezu unmoglich ist, yéjJ1)Tat in anderer als futurischer Bedeutung zu fassen" (Walter, o.c., p. 10).

2) For the use of the subj. of yíYVOflat cf. e. g. fl 191 (where I doubt the correctness of Faesi-Hinríchs' "was eben, was gerade (jedesmal) geschieht"; in my opinion, Murray's translation of ¿(jfl8V (j' oaaa yéV1)Tal is better: "we lmow all things that come to pass"; 299; 465; Y 308. Miss Hahn (o. c., p. 80) is not right in supposing the difference between l!aa8Tat and yéV1)Tat to be one rather of vocabulary than of mood. As to e 201 (and A 262) I wonder whether the verdict: "la négation ov dé­nonce l'emploi du subj. en fonction de futur emphatique" (Chantraine, O.C., n, § 309) is completely correct: in ov(je yév1)Tat (O{jTO, dvf¡e) the speaker denies that he and other men are to be confronted with _such aman. - The words "foreign to yéV1)Tat" must not be taken to imply that yév1)Tat always expresses what is opposite to emphasis.

3) Walter, o.c., p. 11 (who errs in saying that Delbrück, Conj. U. Opto considera the meaning of, this subj. to be "rein futurisch").-Cf. e.g. also ]3,V, 6,12,2.

4) Walter, O.C., p. 10. See Delbrück, Le.; C. Hentze, Kuhn's Zs., 42, p. 132 is not correct in holding that the future means "lediglich als in Zukunft einmal ein­tretend".

i

~ I I

J

Page 42: Gonda.1956.The Character of the Indo-European Moods.pdf

76 The Indo-European moods

explained as the result of a tendency to repeat, in emotional speech, clauses or sentences in such a way that, under the influence of an inward urge to give fuller expression to emotions and intentions, the reiteration is also a reinforcement, or 'Steigerung'. This reinforcement can consist in a longer, but also in a more emphatic utterance. The phenomenon is very frequentl) in case of immediate succession of both utterances: in French (Duhamel) la rédaction en chef de ce noble canard, de ce tres puissant canard; (a young man in love) la est mon foyer, la est ma che1"e maison; iZ jetait, sur son domaine, SU1' son cher laboratoire; mon amie, mon cher amour; il était parfaitement gai, parfaitement a l'aise; in Dutch (an angry woman) daar gaat het niet om, daar gaat het absoluut niet om; in Greek a 169 aA).' ays p,ot -r6os sine "al a-res"éw~ "a-ráAs~oV; in Latin, (a double instance) Ter. Eun. 549 num quis hic est? nemost. num quis hinc me sequitur? nemo homost; Mart. 5, 44, 1 quid factum est, rogo, quid repente factum est?2). So far these passages may be considered to represent a particular case of the "Gesetz der wachsenden Glieder" (Behaghel's tendency)3). Examples can, however, also be cited of a similar repetition interrupted by one or several other sentences: Plaut. Amph. 904ff.: nam certo, si sis sanus aut sapias satis I ... I ... I ... nisi sis stultior stultissimo4 ); Trin. 954ff. qui mihi mille nummum crederet ... mihi concrederet.

A similar alternation of subj. and future is to be found in Euripides, Ion 756ff. -rí oewp,sv, 17áva-ro~ dív "s¿-rat néet; :: 757 :: smwp,sv, i} l1tywp,sv~ i} -rí 0eáaop,sv; cf. also Soph. Trach. 973. In the brahmal).as a future can join a subj. to state what will occur if once the mental process included in the subj. has been realized: AiBr. 2, 25, 1 hantiPjim2 ayama3 , sa4 yo5

na6 ujJ"e§yati7 sa8 prathamaJ:¿9 somasya10 pasyatillti12 "come!, let us run3

a race2 ; he4 of us6 wh05 wins7 shall drinkl1 first9 of the soma10 " 12.

It is precisely the more decided and-comparatively speaking­'emphatic' character of the future 5) which has eluded the attention of Mirt, Walter and other scholars who, generally speaking, did not perceive

1) It is my intention to revert to it in another publication.-Compare also the "Plural der affektischen Verallgemeinerung" (see Havers, Handbuch, p. 161 and 257), the use ofFr. canaille with reference to a single definite person (L. Spitzer, Zeitschrift f. franz. Spr., 50, p. 150), and similar phenomena.

2) Cf. also Hofmann, Lat. Umgangssprache, p. 92; Indogerm. Jahrb., 12, p. 182. -The French instances: G. Duhamel, Combat contre les ombres, Mere. de Fr., 2e éd., p. 54; 81; 244; 284; 304.

3) See Behaghel, in the Indog. Forsch., 25, p. 110ff. 4) See also J. B. Hofmann, I.ateinische Umgangssprache2, Heidelberg 1936,

p. 96f. ("die affektische Steigerung bei der Wiederholung (in Plaut. Amph. 904ff.) zeugt deutlich von der inneren Erregung ... "). Cf. also Plaut. Pseudo 373ff.

5) For the emphatic Use of the future in minatory and monitory conditions see B. L. Gildersleeve, Trans. Amer. Phi!. Ass. 1876, p. 13ff. (6áv and el clauses); Goodwin, O. C., p. 165 (future indico in protasis); Kühner-Gerth, O. c., I, p. 173 only states that "der Eintritt einer Handlung als bestimmt erwartet hingestellt wird".

The Subjunctive 77

any difference of sense between this tense and the subjunctive. This character is especially clear when, in utterances which imply a strong appeal to the feelings of the hearers, a threat, warning, command, assert­ion or emotional promise is expressed: A 431 "today you shall boast (BneV~sat) ... (433) or else you willlose ("sv oAéaarJ~) your life" 1); a 396ff. "one (of the other kings) haply may have ("sv ... 8XrJaw) this place, but 1 shall be lord (ava~ 8aOp,at)." A much-discussed text is A 184 2),

where the future nép,'ljJw seems to express, in anemphatic way, a promise or concession, the subj. ayw the subsequent, secondary process or sequel­the function of this mood which will be elucidated further on-in the sense of "but then 1 foresee that 1 shall ... ".

Interesting passages are, in addition to the aboye: 161 aAA' ay' ByWV ... I B~sínw "al návra (totality) Ott~op,at (climax; Behaghel, see above)3) ; Q 551 "you certainly will not bring him back to life (ava-r1}O'st~); before that, 1 see, you will suffer (ná17rJa17a) some other ill", "1 foresee the result will be, that ... " (in Dutch: "ik zie ervan komen, dat je ... ")4), cf. also Z 70f. the adhort. ,,-rsívwp,sv and the fut. O'VA1}as-rs; Eur. El. 967 Tí ofí-ra oewp,sv; p,r¡-rée' ij cpovsvaop,sv; ("shall we really kill our mother?"). Despite Walter's5) denial of any difference between 'IjJ 73 and A 126 1 would maintain that Eurycleia expresses an intention: aAA' ays -rOt "al afíp,a aetcpeaoe~ aAAo Tt stnw ('IjJ 73) "1 will tell you ... ", and the ruviner Teiresias, in speaking of events to come, refers to future facts: afíp,a oé rot Beéw p,áX (notice this adverb) aetCPeaoé~, ovoé as A1}O'S¿ (notice this second future and this additional litotes 6); A 126; cf. also 135; 137). Similarly, 'IjJ 326. As to H 29f. vvv p,ev navO'wp,sv n6Asp,ov ... I ... vaTeeOV anTe p,axi¡aov-rat (cf. H 290; E 129; p,291), the subj., to my mind, does not refer to a "consequence", nor does the future refer to a "simultaneous" act (Walter 7)): the speaker in winning over his auditory to his proposal

1) Like Miss Hahn (o. c., p. 85, n. 195) I fail to understand Monro, o.c., § 275, who holds that the greater 'stress' is laid on the alternative expressed by the subjunctive.

2) See Hahn, o.c., p. 81ff. 3) Expressions of totality, i. e. words for aU, every etc., like those of an absolute

degree ('superlatives' etc.) are well-known elements of emphatic expressions: see e.g. (Leumann-)Hofmann, Lat. Gramm.5, p. 390 (omnis); compare also the com­pounds of the type Dutch allerliefst "very dearest".

4) See also Slotty, Gebrauch des Konj. u. Opt., p. 130f.; Hahn, o.c., p.79, n. 176; Schwyzer-Debrunner, o.c., p. 310.

5) Walter, o.c., p. 19: "in genau demselben Zusammenhang"! 6) For the stilistic (emotive) value ofthe litotes see e.g. Wackernagel, VorI., n,

p. 298; Havers, Handbuch, p. 190f.; W. Krause, Kulm's Zs., 50, p. 87; M. Deutsch­bein, Handbuch der Englandkunde, I, p. 52; it is my intention to revert to this point in another publication.

7) I refer to Walter, o.c., p.14. That the future denotes a "zunachst zurück­gesteUte Handlung" (Ameis-Hentze, on E 129) does not seem to be the essential point.

Page 43: Gonda.1956.The Character of the Indo-European Moods.pdf

L

78 The Indo-European moods

to interrupt the battle, wishes to put beyond doubt his own resolution to resume fighting 1).

This difference between subj. and future 2)-which 1 would explain from the actual character of the latter rather than its temporal implications 3)_ is also found in Sanskrit4): ~V. 1,44, 5 stavi~yámil tvám2 ahám3 "P shall praise1 Thee2": notice the addition oí the "logically superfiuous", but emotive-emphatic pronoun aham, which can, inter alia, express the speaker's confidence in the realization of the process (cf. in Greek, 'YJ 28 , • , ~ , I ~,/: K 108 ,~ , , 1'" "') H To¿yae 8yro TO¿ ... UO/WV . .. u8¿"ro; ao¿ uS ¡.ta/'. S'1jJ0ft 8yro. ence the use of the future in promises: TS. 6, 5, 1, 1 "do not hurl it at me; there is this strength in me, that1 shall 1 bestow3, 4 on Thee2" 5 (tat1

te2 pra3 dasyamí4ti5); Bar Up. 6, 2, 8 ta'f(Ll tv2 ahar¡¡,3 tubhyar¡¡,4 vak~yami5 "but2 P shall teach5 it1 to you4 (for who can refuse you)"; in statements

1) Cf. also K 108, which is a clear instance of a future expressing more deter­mined thought: much of the comment made by Walter (o. O., p. 12) is superfluous.­Cf. also suoh more 01' less passionate passages as E 715ft'.; P 415ft'.; see also Henze, K.Z., 42, p. 138.

2) Attention may also be drawn to Plato, Rep. 376C rptAÓaorpo, 1¡p:iv 8a7:at Ó péAAWV uaAo, udya1}o, 8aea1}at gn5Aa~ " ... will prove to be a philosopher"; the emphatio 7:[ Aé~St, "what do you mean to say?" (often found in tragedy, e.g. Soph. Ph. 1233; Eur. Ion 1113), and the so-called gnomio future, e.g. Hdt. 5,56 005s1, dv1}l]dxltwv d5tuwv 7:[aw aVu dno7:[as¡; of. Plaut. Most. 1041 qui homo timidus erit in 1'ebus dubiis, nauci non erit "a man who is timid in a crisis is not worth a stiver"; in German: (Sohiller) nur der Starke wird das Sohioksal zwingen (Behaghel, Deutsche Syntax, n, p. 263 f.); cf. also suoh instanoes as: bist du zufrieden :: wie. werd' ioh denn zufrieden sein? and other oases in which the future, oomparatively speaking, has an emphatio, suggestive 01' positive value. 1 am not wholly convinced by the argument "that what is always true is true of both present and future, and, sinoe the future represents a period of far greater extent than the present it is natural enough to select a tense desoribing it for a repeated 01' habitual aotion belonging to the present.future sphere" (Hahn, 0.0., p.9). This explanation is, 1 believe, too intelleotual. In Dutoh instanoes of this future the implioation is, as far as 1 am aware, oertainty and positiveness 01' even inevitability (of. the Engl. boys wiU be boys) rather than the idea of "a period of great extent": dat zal de tegenpm'tij niet toegeven "it is not to be expected that opponents will admit that". This reminds us of Herodotus' predíleotion for the future in describing oustoms eto. (of. Kühner. Gerth, O. O., 1, p. 172), e. g. 1, 173 sll]opsVOV 56 é7:sl]ov 7:0V nAr¡a[ov 7:[, sir¡ ua7:aAé~st éwv7:oV pr¡7:l]ó1}sv ("inevitably, oertainly, regularly, on every ocoasion"); see also H. Stein, on Hdt. 1, 216. For Sanskrit, cf. also Renou, Gramm. sanscr., p. 462. Mention may also be made of the suggestive future: du wirst mioh dooh nioht aUein lassen (see alBo E. Lerch, Die Verwendung des romano Futurums ... , Leipzig 1919).

3) Cf. e.g. suoh passages as AiB. 7, 1, 1 athii}ta,ZL2 pasor3 vibhaktis4 taaya" vibhá· ga'l'(b6 vakmarmal/ "now1 ,2 the division4 of the animaJ3; we shall declare? the division6

of it5" (the explanation follows immediately). In pronounoing this sentenoe the author views the explication as a fact. Cf. also J¡tV. 10, 34,6 "the gambler goes to the assembly hall, asking himself «shall 1 conquer?»" (felJyámtti).

4) Notice also the juxtaposition of a future and an optative in cases like ChUpo 4, 4, 3 b1'ahmaoarya'l'(b bhagavati vatsyámi, upeyá'l'(b bhagavantam "1 have the firm intention to become a student of saored knowledge; may 1 become your pupil, Venerable Sir?".

The Subjunctive 79

of funetions: 1,5, 21 vadi~yamyl eva2ham3 iti4 vag5 dadhre6 "Speeeh5

resolved6 : P shall speak1 (1 am going to speak, my task is certainly to speak)"4; in peremptory refusals: JBr. 3, 193 na1har¡¡,2 tvar¡¡,3 sto~yami4 "1 (certainly)2 shall not1 praise4 Thee3"; ef. AiBr.l, 18, 1. Cí. also AiBr. 6, 1, 1 "the gods could not beat off evil ... A. K. said: «l will períorm that ... for you, then will you (fut.) beat off evil» " 1). In an emphatie exclamation: Mbh. 1, 157, 34 kuta eva parityalctt¿r¡¡, sutar¡¡, 8alc~yamy ahar¡¡, svayam? Removing any grounds oí objeetion: 3,54, II "how shall 1 be able to enter that palaee so well-guarded?" :: "you shall get access" (pravek~yasi, a god is speaking) 2). In a similar way we find in Latin: Plaut. Stich. 614 iam he7'cle ego per hortum ad amicam transibo meam; Capt.457 3). In modern Dutch, where a subj. no longer exists, a so-called future (the periphrastic construetion with zullen) is often used in positive statements, assertions, threats, commands, more 01' less emphatic expressions of willingness, convietion etc. 4). However, the original force of the auxiliary is still more 01' less evident in a great variety of cases. This speeial nuance of a future form may also be illustrated by passages in which it is aceomp­anied by a subordinate subj. of the same verb: Eur. Alc. 725 f}aveí, ys ftBVTO¿ ovauAds-, 8wv f}ávrJs-; Arist. PI. 600 ov yae 7Cs{as¿s-, ovo' i}v 7Cs{arJS-.

Another remarkable peculiarity may be eonnected with this difference in nuance between subj. and future. In the so-called adhortative use of the subj. the first person singular is, especially when employed without cpBeS 01' ays, less eommon than the plural-type Theognis 1047 vvv ftBV 7C{VOVUS- TSe7CWftsf}a-; the first pers. sing. of the future is on the other hand frequent. Compare, in Latin, the very common type eamus, the frequency of ibo, and the rarity of the type Plaut. Bacch. 1049 quod perdundumst pl'ope7'em perdere "let me . . ." 5).

1) The reader may also be invited to view the example given by Delbrück, Altind. Syntax, p. 290ft'. in this light. 1 also refer to my papel' on the pronouns of the 1Bt and 2nd person in the Acta Orientalia, 19, p. 2 11ft'. , esp. p. 225ft'.; 235, eto.

2) See also Speyer, Sanskrit Syntax, p. 266f. where the thesis is defended that the future in .syati is often employed so as to express a kind of subj. mood, ex· pressing exhortation, wish, doubt, probability, purpose. Speyer himself no doubt rightly translateB Pañc. 176 mmna1 yütha'l'(b2 tair3 ••• vyapádita'l'(b4 bhavÍijyati" : "myl fiook2 is sure to have been5 killed4 by them3 ". Yet the post·Vedio future, h'aving no subj. beside itself, should be examined separately; the study of the particles (api + fut., expressing a wish) should not be neglected.-Cf. ;ro Charpen. tier, Die Desidel'ativbildungen del' indo·iran. Spr., Uppsala 1912, p. 95ft'.; Hopkins, The Aryan future. Amer. ;ro of Phil., 13, 1ft'. (who can scaroely be right in assuming a voluntative·desiderative sense of the future to have been the most original); F. B. ;ro Kuiper, in the Acta Orientalia, 12, p. 286ft'.

3) See also H. Sj6gren, Zum Gebrauch des Futurums im Altlatein., SkI'. K. Hum. Vet.·Samf. Uppsala, 9,5 (1906), p. 144.

4) 1 alBO refer to G. S. Overdiep, Stílistische Grammatica van het moderne Nederlandsch, Zwolle 1937, p. 439ft'.

5) See A. Ernout.F. Thomas, Syntaxe latine2, Paris 1953, p.231.

Page 44: Gonda.1956.The Character of the Indo-European Moods.pdf

80 The Indo-European moods

It seems to be in harmony with the positive and assertive character of the future, and the idea of mental existence expressed by the subj. that the former often occurs in affirmative, the latter in negative sent­ences; in Homer the 'prospective' subj. is, in most cases, accompanied b ' H 97 ' , , f3' • \ , , 5:' "1 d 't y ov: 1 ov yae "tt~ fhs tri ys sxwv asxovra utr¡rat on see aman (1 cannot see any man) driving me ... ; no man can ... "; A 262; , 201; 1T; 437 etc.; and, in the same context: o 349ff. "1 shall devise (fhr¡rlcJOfhat) his death; his kinsmen ... will not give (ov ... ).sAáxwO't) him his due meed of fire in death, but the dogs shall rend (eevovO't) him ... ". In A 132 and Q 551 the future + ov is decidedly emphatic. Generally speaking, it is difficult to deny the possibility of an event in the future. We find it, however, ~V.1, 8i, 5 ná1 tvávG/rp,2 indra3 kás4 caná5 ná6 jat67 ná8 jani~yate9 "like Thee2, O Indra3, no1,6 one4,5 has been born7 nor8 shall be born9 ";

in peremptory and minatory speech and in predictions, cf. e.g. AiBr. 3, . r 1) 2) 30,2; SatBr. n, 6, 3, 11 .

The frequent Vedic instances of na + subj. serve to express "zuver­sichtliche Erwartung" 3), "auch bisweilen in dem Sinne, daB das Gegen­theil des Erwarteten geradezu als unméiglich erscheint": ~V. 1, 148, 5 andhá1 apasyá2 ná3 dabhann4 abhikhyá5 : "(it is not to be expected, it is even impossible)3,4 that they, being blindl, being unable to see2

, hurt4

(him) by casting a look upon him5". Cf. 2, 30, 74).

Whereas the combination of fhf¡ and the fut. indic. is, in Greek, of very rare occurrence 5), the constructions of the subj. with that particle are very common 6). The use of this mood after fhf¡ is self-evident. In A 251 rO'Xso fhr¡b' ovofhf¡vr¡~ "don't tell" the process of telling is not actual; s 356 fhf¡ rí~ fhOt vcpaívr¡O'w b6AOV ("Konj. der Besorgnis" Stahl) "1 fear ... may prove to be weaving"; the usa geof Attic prose illustrated by Plato, Gorg. 462 e Wy¡ ayemx6rseov i1 ro aAr¡{}8~ sl1T;BÍV implies a polite suggestion oí apprehension or hesitation; cf. also the type b 165 ép fh-Y} aAAOt aoO'O'r¡riíes~ 8WO'tV; the type A 522 a1T;60'"ttXs fhf¡ "tt vof¡O'r¡ "Her¡; fhf¡ after verbs ex­pressing apprehension, etc.-: the character of the subj. is always the same.

1) See also Delbrüok, Altind. Syn,tax, p.293. 2) The diffioult point of how far a subj. oould be an element in peremptory

speech may perhaps be eluoidated by a notioeable use of the subj. in the Veda. In menaces or indignation this mood "débouche dans la catégorie du futur" (L. Renou, Déo., § 17: "le subj. futurisant se préoise en menaoes"): ~V. 10, 10, 10 itI gha2 tit3

gachan4 úttara5 yugitni6 "viendrontI,4 sans doute2 des ages6 ultérieures3, (oh les

freres feront l'aote interdit aux freres)": the partiole gha, like "doubtless", does not exolude aIl doubt. Cf. also 10,108,4; 6. Here too the intonation may have been a factor of importanoe. In Dutoh a phrase like: ik zie le nog eens een ongeluk krijgen, whioh is not peremptory or emphatio in itself, may also assume that foroe.

3) GraJ3mann, Wi:irterb. zum Rigveda, 700. 4) See also Renou, Déo., p. 11; 15. 6) "Fut. instead of an imper.": Kühner-Gerth, 0.0., I, p. 176. 6) See ohapter X. The relevant Latin faots may be found in Leumann-Hofmann,

Lat. Gramm.5, p.573f.

The Subjunotive 81

If we adopt the aboye • general character' of the subjunctive some disputed questions no longer appear to be of great importance 1). Scholars, for instance, differ widely in opinion with regard to the 'position' of the so-called deliberative subj. 2): is it voluntative orprospective, or does it constitute a class by itself3) 1 This may, however, turn out to be one of these pseudo-problems which have arisen under the influence of a philo­sophical approach to the study of language. The argument advanced, with regard to -d 1T;á{}w and rí ysvWfhat, in Kühner-Gerth's grammar 4)

in favour of the second standpoint does not hold good, for it is not relevant whether or not "eine von auBen kommende Einwirkung in Frage geste11t wird". W e need not enter into a discussion of the questions as to how far there exists a difference between o 509 1T;ñ yae . . . rw and A 404 -r;{

1T;á{}w; whether -d 1T;á{}w ("was so11 ich anfangen", also A 404, Schwyzer­DebrUlll1-er) necessarily presupposes the effect of outward influences-cf. Arist. Nub. 798 Tí eyro 1T;á{}w; "what (else) am 1 to d01", "what can 1 do about it", Dutch "wat kan ik eraan doen1"-; and how far these locu­tions were stereotyped 5). The only point that, linguistically speaking, matters is: from the nature of the case this so-called deliberative subj. always occurs in an exclamatory interrogation 6). For the rest it is, like the subj. in general, the expression of 'visualization' on the part of the speaker 7). In contradistinction to many instances of th,e type aV{}t fhSVW 1}e {}sw (see further on) the interjectional Tí 1T;á{}w; is not fo11owed by an

1) See e.g. A. Waugh McWhorter, The delib. type of question, as found in Aesoh., Soph., and Eur., Trans. Am. Ph. Ass., 41 (1910), p. 157ff.; W. Gardner Hale, The 'extended deliberative' in Greek, Classical Review, 8 (1894), p. 27f.; the same, in the Trans. Amer. Ph. Ass. 1893; Sohwyzer-Debrunner, 0.0., II, p. 311.

2) MoWhorter, 0.0. is right in observing that the term deliberative is far from being wholly satisfaotory. We may distinguish, inter alia, genuine self-addressed deliberations with a practioal end in view; doubt (appeal to another for advioe 01' direotion); the real d1wllta (the speaker, in the midst of a difficulty, finds himself at a loss); the 'impersonal', speoulative type, without referenoe to any aot to be done by the speaker himself (-¡;[ ná{}w); the type -rt {JavASt olláaw,'

3) "Eine Abart des volunt. Konj.": Stahl, 0.0., p. 229 and 365; of. also Slotty, Festsohrift Hillebrandt, p. 177; Humbert, O. O., p. 114; "anticipatory": Hale, Stud. Class. Phil., 1, p.6; Delbrüok, Vergl. Synt., II, p.368; Brugmann-Thumb, Grieoh. Gramm.4, p. 573; of. also Waokernagel, Vorl., I, p. 236; a olass by itself: Brugmann, GrundriJ32, II, 3, p. 853; of. Sohwyzer-Debrunner, o.c., p. 318.

4) Kühner-Gerth, 0.0., I, p. 222f. (Anm.3). 6) A free use in various situations is proper to these more or less interjeotional

exolamations. 6) Although various olasses of interrogations oontaining a subj. are brought

roughly under the one head of the "deliberative subj.", not all these questions aotually express deliberation. ~V. 5,2,3 for instance is an emotive 'Selbstfrage', "what oan they do" (lcr1Javan).

7) Mention should also be made here of the various classes of deliberative subjunctives in Latín: see e. g. Bennett, O. c., I, p. 178:ff.: Plaut. Bacch. 406 sequere :: qua sequar?; Ter. Eun. 523 haec quar quaeritet?; Plaut. Bacoh. 634 quid faciam?; Ter. Eun. 822 quid ega dicam misera? 6 Gonda. Indo-European moods

Page 45: Gonda.1956.The Character of the Indo-European Moods.pdf

82 The Indo-European moods

answer 1). There are also Indian and Iranian instances 2), e. g. TS. 6, 5, 9, 1 tá~ viJ acikitsaj : juhávaní3 má hau¡¡á3m "with regard to it he reflected: «shall 1 offer, 01' shall 1 not offerh>"

The subj. was indeed often used when a speaker, in case of embar­rassment, irresolution, difficulties, etc. holds, in the form of a question, before his mental eyes (cf. JI 435 ueaotr¡ fl'sp,ove) what could be done: K 62 "What do you command me1 Sha11 1 abide there (ai'5{)¿ p,évw), 01'

run back (i]e {}éw) ... 1" :: "Abide there"; o 509ff.; X 167ff. and n 138 in a subordinate clause uaráAe~ov I i] ... eA{}w (the answer fo11ows in 150) 3); Plato, Gorg. 447 c JJ x., eeov avróv :: rí eeWp,at; In part of the instances the term "deliberative" (cf. JI 435ff.) is of dubious explication: Arist. Lys. 530 (J¿ima :: aoí y', JJ uaráeare, (J¿wnw 'yw is a case of the subj. 'indignantis' (cf. Ran. 1134) which is discussed in another section of this chapter; n 138 the speaker asks for instructions 4).

Sometimes the future is preferred in questions of doubt: Arist. Ach. 312 eh' eyÓJ aov CfJe{aop,a¿; where it is unmistakably emphatic (eira, eyÓJ!)5). When referring to a present intention the phrase rí Aé~8Le;; often found in tragedy (e.g. Eur. Rec. 1124) is decidedly o{ the same character: "what do you mean to say1" With regard to ambiguous forms (type ·d oeáaw;) we no doubt keep on the safe side by recognizing them as subjunctives when the questions with a form of that category admit of a completely satisfactory explanation.

Various views have been pronounced with regard to the so-ca11ed "extended" (indirect) deliberative in Greek-e. g. Soph. Ai. 514 e/w¿ yae ovuér' e(1UV ele; 8 iL {JAénw, I nAi}v aovo It would appear to me that (JAénw is a typical subj. in the sense in which it is taken in this chapte1'. Comparable instances are JI 435; n 74; Xen. Anab. 1, 7, 7 ofm exw 8 n ow "1 have nothing to give"; 2,4, 19f. In these Romeric instances the subordinate clause depends on expressions for "pondering", which

1) Cf. also Slotty, Konj. U. Opt., p.56ff. 2) See also Brugmann, Grundrill2, n, 3, p. 844.-For Greek, cf. also Slotty,

o.c., p. 45ff.; Kühner-Gerth, o.c., I, p. 221ff.-Mention may also be made ofth~ well-kno:wn Latin phrases Plaut. Men. 963 quid ego nunc faciam? "What shall I do now?"; Ter. Eun. 966 quid faciam? qU1'dne incipiam? dicam an non dicam? (cf. e.g. Ch. E. Bennett, Synta:x: of earIy Latin, I, Boston 1910, p. 184f.: "The nurober of subjunctives that can with accuracy be called deliberative is e:x:ceedingly smaU"). For Celtic: Brugmann, O.C., p. 851.

3) See also B. L. Gildersleeve, Synta:x: of Classical Greek, New York 1900, § 380; Monro, o.c., § 277. Whereas Chantraine (o. c., n, § 310) is, in.my opinion, right in stating that the subj. in specmed interrogative sentences "e:x:prIme l'attente du personnage qui parle", cases like K 62 may, I think, rather be regarded as ' alternatives present in the speaker's mind. E:x:pectation: A 150; e 465.

4) "Auch in Fragen des Unwillens, del' Verwunderung oder der Ironie wird dieser Konjunktiv gebraucht" (Kühner-Gerth, o.c., I, p.222). The repetition is characteristic of retorts.

5) Cf. Arist. Eq. 1183, etc.; Eur. Hipp. 1066.

The Subjunctive 83

ex elude any doubt as to the 'mental existence' of the processes mention­ed in that clause. The debate upon the question of the origin of these clauses (adhortative, voluntative, final) seems to be futile l ). In Anc. Indian we find Sat.Br. 10,4,3,7 ... tvaml eva2 tad3 brühi4 yatha5 te6 sarva1Ji7 rüpa1JYs upadadhame9tilO "please do thoul expound4 (to us) in what manne1'5 we shall (may, can) place9 thee6 in al17 formss" 10 2).

Following this train of reasoning we may say that the general sense of interrogations containing the subj. can be explained in the same way. The exclamatory questions of the type 1;tV. 1, 120, 1 are especially interesting: lcathál vidhaty2 ápmceta!b3 "wiel kann del' Thorichte3 (euch) vereh1'en2 !" (GraBmann 3

)); cf. also 8,86,2 lcathál nüná~2 va~n3 vímanCi4

úpa5 stavat6 "wiel sol1 euch3 jetzt2 del' Bet1'übte4 preisen5,61" (Geldne1'); 5,29,13; with Jcada 1,25,5; non-deliberative: 1,84,8 lcadá l nah2 susravad3 (perf. subj.) gíra4 índro5 angá6 "wannl wird el' unsere2 Lobe~­worte4 horen3

, einzig6 Indra5 1" (Geldner); 6,35,3 in a so-called rhetorical question: 10,27,11 lcásl tá~2 vidva~3 abhí4 manyate5 andhám6 "who, knowing3 (that, a girl is blind) would desi1'e4,5 her2 who is blind6 1": a merely ideal process without any reference to an individual girl. In retorting: 1;tV. 10, 10, il lcí~l bhr-áta2sat3 "whatl then is to become3 of a brother21" 01' rather, "what does a b1'othe1' mean 1"; he1'e the main difference with an indic. seems to consist in the 'consecutive-general' implication of this verb form, that is to say: the clause with subj., without an interval, follows another utterance, to which it refers; 01' it continues 01' repeats the verb of that utterance. We may compare, in Greek: Arist. Vesp. 760f. ep,o¿ m{)ov :: rí ao¿ ní{}wp,a¿; Nub. 87. General­izations, like exclamatory interrogations, are often more 01' less emotional: "what does a brother mean", and the subj. denoting ideal existence, helps to express the generalization.

This instance has led us to the clauses in the form of an echo to be found in Greek and Latin literature 4). To all appearance Aristophanes was the first Greek author to introduce into literature a construction

1) It is in itself an incontestable statement (Schwyzer-Debrunner, o.c., p. 318, sub 9y) that 1/; 74 the "direct construction" would have ¡tBVW (subj.). But who can warrant that the poet had this phrase in his mind before mouldillg the thought into the indirect constructioll? 01' that those who were the first to use the latter mechallically resorted to the subj. ollly because the direct construction had that mood?

2) See also A. Minard, La subordin,atioll dans la prose védique, I, París 1936, p.83f. Cf. e.g. also SatBr. 2, 4, 2, 1; 1,7,3,5 upa1 jánita2 yathá3smá4 áhuti1~5 kalpayáma6 "cherchez1 ,2 comment3 lui4 préparer6 une oblation5 " (Minard); 1, 6, 4, 7 ... yathe1da1~2 mayi3 srayátai4 tatho5pa6 jánita7 "avisez6 ,7 au:x: moyens de faire qu'lil2 co11e4 a moi3 !"-For the subj. without yathá TS. 2, 4,12,6; cf. Renou, o.c., p.26 (§ 37).

3) H. GraLlmann, W6rterbuch zuro Rigveda, 311. 4) See especially A. R. Anderson, Repudiative questions in Greek dramas, and

in Plautus and Terence, Trans. and proc. Amer. Phil. Ass., 44, p. 43ff. 6*

Page 46: Gonda.1956.The Character of the Indo-European Moods.pdf

------ --~---

84 The Indo-European moods

which may very reasonably be held to have existed before in popular speech: the repudiative question. This turn of speech isa question hardly more than in form and origin, and often the speaker uses it chiefly for appearances' sake. It differs from the ordinary subjunctive question in following a command. Phrynichus, in commenting upon Arist. Ranae 1133 naea¿vw aOb atOJnéiv . .. :: eyro atOJnw np~s; made the valuable remarks, first that the speaker in resorting to this construction wished to formulate an answer by means of a question, and in the second place that this answer involved indignation, or, anyhow, re­jection l ). Cf. e. g. also: Arist. Av. 1691; Lys.530; Eur. Ba. 1184 and especially the type Arist.Av.164 'tí aOb mf}ÓJp,sf}, ; :: 8 Tb ntf}r¡af}s; "wherein shall wefollow your advice? :: wherein you shall-follow my advice?"2). If however, a statement is rejected the repudiative question has the indic.: Arist. Ran. 26; PI. 371. In contradistinction to other scholars 3) who derive this use of the subj. from an original1y volitional function 1 would emphasize the ideal and <consecutive' (this term to be taken in the aboye sense) character of the construction: in immediate succession to the words of his partner the speaker, repeating the verb, repudiates, in the outward form of a question, the very idea of the demand 01'

suggestion implied in these words. In any case it is a subj. in its own right, not a subj. of indirect quotation.

Whereas the Greek authors showed great self·restraint in the use of exclamatory constructions of this kind, their Latin colleagues, especially Plautus and Terence, liked these echo-clauses very mucho Although the Latin <subjunctive' is the result of syncretism, the character of this <polemic' use of the mood is in the main identical: Plaut. AuI. 8lf. 1'edi nunciam intro atque intus serva :: quippini? ego (mark the < emotive' , pronoun) intus servem?; 829 i, redde aurum :: reddam ego aurum? :: redde, inquam, ut huic reddatur. In part oí the instances the speaker is

1) Phrynichus, in Bekker, Anecdota, I, 40, 32 {Jov),S7:at yde ).éysw uaf}' vn6ueto'w' eha Byeh Tovrep atwnf¡aop.at; olo'V a~t6'V eaTw ep.e Tépbe vnoaTsAAsaf}at ual vnetuew AoYOtr;;

2) As is well known, the repetition, in answering, of the construction of a question, is a wide-spread method of retorting or replying impolitely. Often in Shakespeare: R. and J. 3, 2 Will you speak well 01 him that killed your cousin? :: Shall I speak ill 01 him that is my husband?; A. Hu:x:ley, Point C. p. p. 173 (Alb.) Why will they insist on going? :: why will we insist on staying?; D. H. Lawrence, The whita peacock 3, 7 Why didn't you tell me? I asked :: why didn' t you ask me? she retorted. For the repetition of the verb alone in similar situations see, in Anc. Indian, Jaim. Br. 2, 100 ki1~ nu rájanyo dhyáyatiti :: na hi me dhyeyam?; with an opto SatBr. 1,2,5,25. Cf. e.g. also KiUidása, Urv. 2, 9+.-The aboye echo-question of the type Arist. Av. 164 (cf. also Soph. Ph. 974; Eur. H. F. 1417) is the only con­struction to include, in Greek, tha 2nd person ofthe <delib.' subj. (see: B. L. Gilder­sleeve, Syntax of classical Greek, I, NewYork 1900, p.151).

3) See Anderson, O. C., who insists upon the volitional character of this sub. junctive.

The Subjunctive 85

not merely echoing: Plaut. Capto 20sf. at fugam fingitis... :: nos fugiamus? quo fugiamus? l).

It is worth while to insist on a remarkable idiom in the J,tgveda which, if 1 am not mistaken, cannot be disconnected from the afore-mentioned usage. Not infrequently, the subj. refers to the further continuation of a process which is explicitly stated to have occurred up to the present time: J,tV. 1, 113, 10 "(the dawns) whichl have shone2 and4 which3

will, I'm sure5, shine6": yál vyú?úr2 yás3 ca4 núná1Jt5 vyuchán6 (cf. sto 11; 1, 4S, 3; 9,23,7 etc.). Here the poet so to say extends, in his mind, the line beginning in the past: the God Indra has always killed the demons; this activity is-in the poet's eyes-to be continued (J;tV. 9, 23, 7). A Greek parallel is A 262 oV yúe nOJ Totovs r~ov avéeus OV~B r~OJp,Ub. Cf. also J;tV. 1, 165,7; 7,26,3 with the emphatic núnam which significantly is translatable by "now; immediately; for the future; therefore; as sur­edly, certainly"; 8, 20, 15; see also 1, 124, 11; 4, 51, 1. An interesting passage is found in the AiBr. 3, 43, 6 yathii,1 hy2 eva3sya4 praya';Jam5

evam6 udayanam7 asatS "asl the beginning5 of it4 , S06 will beS (1 think, 1 foresee) the end7 ".-In the Homeric X 505 vvv ~' av nOAAa náiJnab a discontinuation is marked, 'Vv'V conveying the idea of opposition to what might have been under other circumstances.

In the Veda 2) a clause containing a subj. and introduced by the <inceptive' particle atha, which generally speaking announces a new element (including a change of the scene, of the subject 01' action etc.)­it occurs also in the apodosis oí a conditional sentence, after a temporal subordinate clause (yada ... , e.g. J;tV. 10, 16,2) and in the protasis ofthe second member of an alternative-expresses the result or consequence of the process of the preceding clause with which it is closely connected, or a process which is (foreseen, 01' expected, to be) made possible by it. The preceding clause has an impero (e.g.AV.14,1,21 Sá1Jtl sPTsasvá2tha3

... á4 vadasi5 "mingle yourselfl ,2 ••• , then3 shall you speak5 t04 ••• ",

cf. with a subordinate [accent!] first member: AV. 8, 1, 6 á l hí2 róha3 •••

átha4 ••• á5 vadasi6 "for2 do you ascendl ,3 ••• , then4 shall you speak6

t05 ••• "; cf. also J;tV. 8, 96, 7); or a subj. (27,22; 3,53,11 [where the

1) In Latin the mood of the <polemic' clause was also generally determined by the character of the antecedent sentence. The primary type was paratactic, but hypotactic variants did not fail to turn up. Cf. the < questions' introduced by ut: Plaut. Amph. 694 (see e. g. Ch. E. Bennett, Syntax of early Latin, I, Boston 1910, p. 186ff.; Hofmann, Lat. Umgangssprache, p. 51).

2) May it be suggested to e:x:plain the curious subj. in AthV. 6, 44, 1 ásthád1

dyaúr2 • •• tiE¡thád3 r6g04 ayá1p5 táva6 in a similar way?: not "The heaven2 has stood1

... may thisS disease4 ofthine~ stand3 " (Whitney-Lanman), but "The h. has come to a standstill ... your disease, I foresee, will come to a standstill". If this inter­pretation be. right it may perhaps shed light on the 'subj. instead of an imperf.', signalized by Reichelt, Avest. Elem., p. 304, § 631: Y. 9, 11 "he sprang forth from below the iron caldron, he was-I think-to overturn the water"; Yt. 5, 62 "he floated for three days' and nights, he was-I think-not to come downwards".

Page 47: Gonda.1956.The Character of the Indo-European Moods.pdf

86 The Indo-European moods

king's sacrifice is made possible by his previous victory]); cf. also J;tV. 10, 51, 7; TS. 2, 6, 12, 2 etc. 1). In Greek, we may compare B 235 ns({}sv' eyro Dé ~é Tm lMm xáew; JI 129 Dvaso TSVxsa {}(iaaov' eyro M us AaaV ays/em. Cf. also, in Avestan, Y. 50, 7.

Mention must be made here of the combination of a future 01' preso indic. and a subj. in Greek. Despite Walter's criticism Delbrück's view 2)

of such passages as ft 383 and 1121 probably was in the main right: ft 383 "if they do not pay me ... , I shall go dOWll (ovaoftal) to Hades and shine (rpas/vm) among the dead"; 01' "and then shine"; the second action is made possible by, follows in the wake of the former (not: "ist die natürliche Consequenz" Delbrück). A subj. in a subsequent 01' 'consecutive' -this term not to be taken in the traditional way-sense is also found in such passages as A 184 where the determined future in -d}v ftev eyro ... I néft1pm, is followed by the subj. in eyro Dé~' aym Bewr¡toa (notice the occurrence of the particle ~s); and, after an opt., in o 389ff. -r '.. ,,, , .lL~ 1 1 f3'.Q 1'" " ,~ , I I .Or y Sl nmr;; av uvValO ... /ls/la Sa'u'al, Or;; ~sv Tal Slnnaw ouov . .. . .. '51"" r54'" , ~at uS ~S Tal SlnnO'l ... ; ... ovu a'v Tal xealaftn ... In Vedic the subj. can however also be more 01' less 'conditioned'

by a preceding clause 01' sentence without the accompaniment of atha: J;tV. 10, 124, 1 "come to our sacrifice ... (if so; then) thou shalt be ... (asa(l,),'.

It is worth noticing that a subj. can also depend on an antecedent indic.: J;tV. 7, 63, 4 (the sun) úd eti "rises": nüná'lJ'l-1 jánal.~2 ... áyann3

árthani4 krfiÓ,vanns ápa'lJ'l-si6 "now1 people2 will (must) proceed3 to business4, (and) to be aboutS their duties6 " (01' " ... so that now ... "). Here the relation of the subsequent clause might even be called consecutive in the traditional sense of that termo Cf. also J;tV. 8, 43, 24 (see further on). We might compare in Greek e 417 f. Tro as xe~ o6ftsval . . . I ... eyro oé ué (/S ~As(m " ... so would I make your fame known"; and also in Latin: Plaut. Sto 177 hoc nomen repperi eo quia paupertas fecit ridiculus forem. In these cases the use ofthe subj., once again, marks the 'complementary' character of this particular 'parataxis', the subordinate nature of the subj. clause. It is a marker of ideal subsequence.

The force ofthe subj. can also be consecutive in the traditional (limited) sense of the termo E. g. KauSl. Br. 9, 4 yathii1 hotar2 abhayam3 asat4

tatllii5 kuru6 "do you, O hotar2, act6 SOIS) thatl there may be4 safety3"3). A1though the Greeks usually availed themselves of other constructions to denote the intended result of the action mentioned in the principal clause, the Thessalian dialect had the subj. after &a"CS ("in order that") 4).

1) Notice the different nuance conveyed by atha + fut., see p. 78, n. 3 and Delbrück, Altind. Synt., p.291.

2) Walter, O.C., p. 11f.; Delbrück, Conj. und Opt., p. 124, a) See also Renou, Déc., p. 25f. (37).-See also Reichelt, Awest. Elem., p. 378. 4) 1 refer to Liddell and Scott, Lexicon, S. V., p. 2041.

The Subjunctive 87

In trying to define the value of the 2nd pers. of the subj. in Vedic prose Delbrück 1) held that this form is used, "wenn del' Redende sich etwas ausmacht, oder eine nicht für die unmittelbare Gegenwart geltende Anweisung gibt". From SatBr. 1, 8, 1, lff., which he adduces as an example, it however appears that it is the 'consecutive' force of the mood-whether 01' not it be preceded by atha "then" -, which led to its being ,used here: yavad vai k?ullakii bhavamo bahvi vai nas tavan na?f1'a bhavati ... kumbhya'lJ'l- magre bibharasi. sa yada tam ativardhii atha kar?ü'lJ'l- khEitva tasya'lJ'l- mii bibharasi. The subjunctives used in this passage alternate with a < decided' future, which do es not any more than the subjunctives refer to "die unmittelbare Gegenwart": bibhrhi mii 1Jarayi?yami tva; with the periphr. future, which here and elsewhere adds the nuance of positive expectation: nirvoiJhii; tarhi va atina?tro bhavi­tasmiti; and with the imperative. Cf. also ll, 5, 1, 12 gandharva1 vai2

te3 pratar4 vara'lJ'l-s dataras6 ta'lJ'l-7 Vr'iJasa8 iti9 "tomorrow4 the gandharvas1

(certainly) will6 accord6 you3 the fulfilment of a wish5, then you must (will) choose8 that7 "9. SatBr. 12, 3, 4, 2 instruction is given by means of a verb in the s~bj.: "worship (imp.), I shall tell you (fut.) ... you should, then, (subj.) ... ". TS. 6, 2, 7,1 sa1b1'avid2 va1'a'lJ'l-3 Vr'iJcti4, sarvan5

maya6 kaman7 vy asnavatha8 "she1 said2 : let me choose4 a boon3 ; through me6 shall ye (then) o btain 8 alls your desires 7 ": "your o btaining", being the content of the boon, having not yet reached the state of actuality, is in a natural way, in the subj.2). Cf. AiBr. 1, 7, 4.

The 'subsequent' 01' < consecutive' position should be especially studied in connection with the marked tendency of this mood to appear after an imperative (01' a word of similar function) 3). In J;tV. 7, 81, 5 tád1

rasva2 bhunájamaheti3, which must be translated: "give2 (us) thatl, that we enjoy3 (it)"4), the clause containing the subj. expresses an anticipated consequence of the fulfilment of the request,-or, in general, of the performance of the process denoted by the preceding verb: we indeed also find J;tV. 8, 43, 24 agním1 i¡e2 sá3 u4 s1'avatS "je loue2 Agnil (en sorte qu')iP (m')écoute5"5), where i¡e is a preso indic.6) The construction sur­vives in post-Vedic Sanskrit: Mbh. 3, 64, 69 b1'ühi l ki'lJ'l-2 ka1'avamahai3

"speakl what2 can (shall) we d03 (for you)". Attention may also be

1) Delbrück, Altind. Synt., p.309f. 2) "Le continue de l'invite, la condition de l'accord", Renou, Déc., p. 21 (where

other instances may be found). a) Hirt (o. c., VI, p. 274), though mentioning the fact, did not explain it satis­

factorily.-In the older Veda imper. and subj. are not yet 'sup pletive' . 4) " ••• das gonne (uns): wir mochten (davon) zehren" rather thán: " ... (uns);

wir ... " (Geldner). Cf. Geldner's translation of 8,43, 24! 5) Renou, Llf décadence et la disparition du subjonctif, § 15, where more

instances are given (~V. 7, 61, 4 is ambiguous). 6) An impero can also be followed by a future; BatBr. 1, 8, 1, 2 bibhrhi má

páray'Í?yámi tvá it helps to express a promise and determination.

Page 48: Gonda.1956.The Character of the Indo-European Moods.pdf

88 The Indo-European moods

drawn to the type Soph. Phi!. 526 el ooue;; nUwwv; Kal. Sak. 3, yadyl anumanyase2 tad3 aham4 ... karavary,i 5 "ifl you permit2, 14 will make5

that3 ".

This < consecutive' implication also exists in sentences of greater length: AthV. 6, 32, 1 arádl rák~arp,si2 práti3 daha4 tvám5 agne6 ná7 nos grháry,am9 úpalO titapasill "from afar!, O Agni6, do thou5 burn4 against3 the demons2

, (so that) thou shalt not7 be hotll towardlO ourS houses9"1). See also ~V. 9, 97, 4 s6marp,l hinota2 mahaté3 dhánaya4 I svadúh5 pavate6 "set Somal in motion2 in order to acquire great3 wealth4 : that 'he, being savoury5, flows off clearly6" 2). The very context 01' situation may lead the poet to use the subj. in this way, no adhortation preceding: RV.4, 18, 2 bahúnil me2 álqta3 kártvani4: yúdhyai5 tvena6 Sárp,7 tvenaS prchai9, where Renou's translation ". ;. j'ai a combattre avec un tel, a m'entendre avec tel autre" is no doubt more convincing than Geldner's: "Vielesl n?ch nie Getane3 muB ich2 tun4 ; mit dem einen6 will ich kiimpfen5, mit emem anderens mich verstandigen7,9": "ce n'est pas tant sa volonté qu'il exprime, qu'une éventualité resultant des charges qu'il s'est im­posées" 3). It is clear that, so far as regards the <dependence' of the subj. from the preceding clause, Greek instan ces such as Z 340 are similar: d),),' aye vvv ¿níp,uvov deñi'a Tevxea ovw "tarry a while so that I can don my harness of war" is as good a translation as " ... that I ... " 01' "tarry ... : let me ... " . Cf. also X 450; P 71; Eur. Hipp. 567 ¿níax e7,'" avoi¡v 7:(»v gawihv ¿up,á{}w4); and the afore-mentioned E 235; M 129. That there is sorne room for individual preference to take the relation either final or consec­utive is no matter for astonishment: the logical relations between clauses were very often not or only vaguely indicated. The phrases Soph. El. 80 {}sAetr; I p,e¿vwp,ev; and Phil. 761 ¡JovAet Aá¡Jwp,at; though classified under the < conjunctivus deliberativus' are essentially similar. In all these cases the ~ubj. in dependence on, 01' in continuation of, actuality stated 01'

reqUlred expresses a process existing in the mind of the speaker, but reg~rd~d as realizable. In many .cases the preceding verb serves mainly to mdlCate what the vague subJ. alone leaves undefined. In Latin we find: Plaut. Asin. 605 sermoni iam jinem jace tuo, huius sermonem acci­piam; Caso 136 sine . .. deosculer; Curc. 313 da ... obsorbeam; cf. also Rud. 1010 tange! adjligam ad terram te 5).

1) And not "mayest thou not be hot ..• " (Whitney-Lanman). 2) To all appearance, M. Bloomfield, On instability m the use of moods in

earlÍest .Sanskrit, ~mer. Journ. of Phi1., 33, p. lff. (see p. 18), did not realize this syntactlC connectlOn expressed by the subj. (cf. Sama-Veda 1,535: indic.). Cf. e.g. also AthV. 18,4,40 and HirGS. 2, 10,6.

3) Renou, 1. C. 4) For other instances: Th. Barthold, on this passage (Eur. Hipp., Berlín 1880,

p. 58): "del' Konj. enthalt m diesen Fallen ... "wenigel' eme Aufforderung als del' Ausdruck emer Absicht". '

5) See Bennett, Syntax of early Latin, Boston 1910, I, p.256.

'The Subjunctive 89

In connection with SatBr. 3, 7,4,9 tatl prapnuhi2 yat3 te4 prary,0 5

vátam6 apipadyatai7 "erlange2 das!, daB3 dein4 Hauch5 in den Wind6

übergehe7" (Delbrück) Renou I) made the remark that "le subj. est celui de la < directive', secondairement intégré dans un cadre en tat ... yat". This idea may, in view of t.h~ infrequence of the. compound senten?e construction, be right, but even If It were not, the subJ. should-at least m the original idiom-be expected, because the process is not (yet) actual.

The phrases with aye, rpsee etc., to which :ve turno now, .ar~, though to a considerable degree stereotyped, to be Judged m a SImIlar way; they are in their entirety expressive of voliti?n, exhortation ~tc., n~t the verb form by itself. In its so-called voluntatIve and adhortatIve use ) the subj. was usually accompanied by one 01' more introductory 01'

adhortative words 3): L1 418 d),X aye oi¡ ... p,eoÓJp,e{}a "nay come, let us ... "; Arist. Ach. 4 rpérl lOW; Plato, Gorg. 455A rpéee oi¡ lOWp,SV; rpées oi¡ aus'IjJÓJp,s{}a' 1} 133 oevTs . .. ¿eÓJp,s{}a n)v ~8rVOV; JaimBr. 1, 12 eteI marp,2 jigi~ame3ti4 "kommt1, laBt uns versuchen illI~2 ,(den. T~d) zU,besiege~3:' (Caland); ~V. 10, ll9, 9 hántá,lhám24) prthwím3 ~marp,4 nt5 d~~han~6 "well then!, 12 will (shall, let me ... ) lay6 the4 earth3 down5

•••• The speaker conceives an idea or purpose to be execut~d by .himself 01'

others. Cf., in Latin, the type Plaut. Caso 234 vem d~cas vel~m. These sentences were probably often characterized by a special

intonation, perhaps not seldom also by gestures 5). In various languages, also in the ancient l.-E. idioms, the present indicative can also serve to express volition, stimulation etc. In Dutch a sentence like: je loopt niet op het gras, pronounced in a certain emphatic way, can mean: "you (people) must not walk on the lawn", je zegt het niet! "don't say that!" "Kom, we vragen het die vreemdeling" is a correct translation of {} 133 (see aboye). We should therefore conclude that the mood of the verb helps to formulate a desire, exhortation etc., not t~~t it expres~.es ~he~.

This is especially clear in the aboye cases of mnere AbhangIgkeIt eines Konjunktivsatzes von einem vorausgegangenen Satz" 6): X 139 d),),' ayeTe ... ¿ve¿uw " ... let me bring"; P 71; Eur. H. F. 1059 a;;y~, nvo~r; p,á{}w' rpées neor; ovr; ¡Já)'w; here the persons addressed must be sIl~nt m order to enable the speaker to listen; it is of course the latter's wIsh to

1) Delbrück, O,C., p. 322; Renou, Déc., p.27 .. 2) For particulars as to the occurl'ence of the dlfferent persons etc. the reader

may, for reasons of space, be l'eferred to the gramm~rs. 3) In Vedic texts such a word is, however, often omltted: J3.V. 10, 39,~; T~. 7,

1,5,1 praiáyárnahai; Jaim. Br. 1, 12 áiyáhuti1Jl' ,fuJ:aváva; 87 vírnau l!ar~hararna, etc. The occurrence of a particle 01' imperative lS m ltself no argument m favour of the supposition that the adhortative construction is not "original" as contended by Walter, o.c., p. 18. . ..

4) For the emotive aham see my papel' 1ll the ~cta Onentaha, ~9, p. 211ff. The particle hanta asks attention, expresses exhortatlOn to do anythl~g, etc.

5) See also Havers, Haildbuch, p. 25. 6) Brugmann, Grundn.l3, n, 3, p. 843.

Page 49: Gonda.1956.The Character of the Indo-European Moods.pdf

90 The Indo-EUropean moods

do so, but it is not exactly and merely the function of the subj. to express this. Compare ~V. 10, 124, 1 imá'f{/,1 n02 agna3 úpa4 yctj11ám5 ehi6 . " I as07 havyavat8 ... "Approach4,6, O AgnP, thisl sacrifice5 of ours2 ... in order to be 7 o blation -bearer8 ••• ". Of. also such instances as ~ V. 1,9,5 1 ).

If however a conjunction is used 2) Sanskrit has yatlúi, Gr. tJcp(}a, onwr; etc. and the relation turns out to be <final': final clauses, however, are not seldom regarded as containing a prospective subjunctive 3). Cf. e.g. a 89 'I-&á'Xr¡yfl Ba8A8vao¡.tat, tJcp(}a ... I ... -&sEw; (after an imper.) A 32 dAA' Wt, ¡.t~ ¡.t' B(}é-&d;s, aadYrs(}or; wr; 'X8 yér¡at; (after a fut.) e 37; AthV. 1, 29,5 údl (agád2) idá'f{/,3 mámaká'f{/,4 váca7y,5 yatM6ha'f{/,7 satruh68 'sáni9

"this3 spe1l5 of mine4 has "gone2 Upl", that6 17 shall be9 a slayer of ene­mies8 ". An interesting group are the instances of a relative clause with subj. preceded by an imper.: in the brahmal).as,e. g. PañcBr. 17, 1, 1 tael ehanda7y,2 pmyaeehata3 yena4srniin5 apnavan6 iti7 "give3 thatl metre2 by which4 they may reach6 us5 (in order to reach us by it)"7; Jaim.Br. 3, 105 tanl na7y,2 prayaeeha3 yaú·4 vaya'f{/,5 gho~i1Jir6 asama7' "übergib3 siel uns2, damit4,7 wir5 durch sie4 gerauschvo1l6 werden7" (Caland) the final sense is implied in the context in its entirety; in Avestan, e.g. Yt. 15,40 "(permit us to find a lord of a house) whol will take4 every care3 of us2" yjj1 njj2 hub'J),'Jtqm3 bara14; in Greek: <P 127 vnat~st lx-&vr;, or; 'X8 cpáYrJl1t ... or¡¡.t6y "shall dart up ... to eat"4). Attention may finally

1) The subj. of the verb as- "to be" is often fOlmd in this construction. Its general and vague sense makes it especially fit for paratactic addition. In Plautus the subj. of eSS6 is likewise very frequent in phrases such as fao sis frugi (Morris, o.c., p.284).

2) It is worth mentioning that under similar circumstances conjunctions may be facultative also in other families of languages, e.g. in Georgian: see H. Vogt, Esquisse d'une grammaire du géorgien moderne, Oslo 1936, p. 272. For the Latin constructions Plaut. Stich. 624 dixi equidem in carC6rem il'es and Capto 844 ita dico, magnus ut sit-it is immaterial whether ut is present or absent-see C. L. Durham, subj. substantive clauses in Plautus, Cornell Studies in Class. Phil., 13; Bennett, o. C., I, p. 209ff.

3) See aboye. 4) This does not mean that any case of a relative construction with the subj.

is easy to determine 01', rather, easy to bring lmder one of the traditional heads: see e.g. W. G. Hale, Trans. Amer. Phil. Ass., 24, p. 165 on r 459 (where néJ.r¡rac seems to mean: "will-as I foresee (express)-abide", although a final implication may be defended as well). Cf. e.g. also J;tV. 1, 107, 1 ("will", "so that ... ", "in order to ... " ?). In fJ 212; o 310, etc. a similar clause evidently conveys purpose. Cases such as ~ 158 are "ideal" rather than "prophetic" (Hale). In a negative clanse (~ 202) the e:x:istence of a process is, in general, negated. In constructions like O 756 there is on the part of the speaker no doubt as to the actuality of the future fact denoted in the principal clause (enéaaarat), whereas the subject of the subordinate clause only e:xists before his mental eye (8xnac), or-to e:xpress ourselves otherwise -the "mental image" to hold the halls will be realized in future (cf. also fP 103).­Notice alsoAthV.ll, 1, 13 tuSár¡¡,1 grh1Jitád2 yatamu3 yajñiyá4 ásat? "seize2 ofthém1

whichever3 is6 worthy of sacrifice4". The element of will, intention, hope, e:xpec­tation etc. can also appear, cf. e.g. J;tV. 5,4, 11.

The Subjunctive 91

be drawn to AthV. 7, 38, 4máméd ásas tvá'f{/, kévala7y,-where the accented verb indicates the dependent character of the clause-and the equi­valent 7,37, 1 yátMso máma lcévalalJ·

1 am convinced that we should not speak here of semantic transi­tions l), but only of the freedom of the ancient languages to use the subj. in munberless situations and in a considerable variety of con­structions.

In translating final clauses into Dntch, English 01' other languages no finite verb is needed: B 381 YVY ;Y l(}Xsa-&' sni as¿nyoy, tjJa ~VyáYW¡.t8Y L1(}r¡a, in Dutch: " ... om ons in de strijd te mengen". In the Veda this is one of the most freqnent uses of this mood : AV. 1, 29, 5 yátlúi1há'f{/,2 satr'uh63 'sani4 " ... in order to be1,4 a slayer3 of my(2) enemies3" 01' "that 1 shall be a ... ": although the process will (01' may) be realized in the future, the interest of the speaker (and the hearer) mainly attaches to the aim 01' result in view, not to futurity 2). Similarly, n 369 Tr¡Mpaxoy

Aox6wY'ísr;, lya cp<&taw¡.tsy éA6nsr; I aV'í6y ... Once again any discussion on the "volitional" 01' "prospective" character of the subj.3) may ling­nistically speaking be regarded as irrelevant, although, of course, a philologist mnst ask himself how far there are, in context and situation, indications of intentions on the part of the speaker 01' of the author. Of. also ~V. 2, 26,2; 4,57,6; 10,85,25; 174,3 etc., after an impero 10,37, 10' AV 1 16 4' 22 2,' RV. 2, 30,5 ával ksipa2 div63 ásrniinam4 ... yéna5 , ., , , , . . sátrum6 ... nijúrvaZ¡,7 "wirf2 herabl den Stein4 vom HimmeP ... , um7

dadurch5 den Feind6 zu7 zerschmettern7" (Delbrück). Other relative constructions are Av. Yt. 15,40 ctvall áyapt'Jm2 dazdi3 no4 ya15 nmáno. paitim6 vindama7 yo8 no9 hub'Jt''JtqmlO baraJ1l "give3 us4 thatl success2

that5 we find7 a master of a house6 wh08 will takell every care10 of us9".

Not infrequently, the nearest approximation in English to the original may be "is to (be)": ~V. 1, 107, 1 "Your favour may turn towards us which is to be the best procurer of free scope (ease)": ya varivovíttá?'ásat. OÍnitting the so-called relative we obtain the aboye <consecutive' con­st¡uction. Still more evident: 10,29,2 "the chariotl, which2 was to be3

the winning oné": 1'áthol y62 ásat3 sasaván4"; cf. also 2, 23, 14. In Latin, Planto Epid. 588 non patrem ego te nominem? "am 1 1l0t to call you fathed"; in Greek, Ti ná-&w "what is to become of me?" 01' "what am

1) See e.g. vValter, o.c., p. 17f. Needless to combat the author's view that in {} 31 the subj. is future in sense. Here also, the idea of time is of secondary impor­tance. For "'Villensau13erung, Aufforderung,. Befehl, Gebot" see also Slotty, o.c., p. 8; 19f. ("Wi1lenserklarung zugleich als Aufforderung"); Schwyzer-Debrunner, O.C., n, p. 313f.; for the opinions of other scholars see also Walter, o.c., p. 1ff.

2) The German um zu, like the Dutch 0111 te "bezeichnet die beabsichtigte Wirkung" (Bel:¡.aghel, D.S., n, p. 337). In French the airo can be marked by pom': Xen. Cyr. 1,2,15 lva 06 aarpéaul2.ov ol¡J.w{}ñ .•. "pour mieux mettre en lumiere ... " (Humbert, O. c., § 375); pO~~l' likewise expresses the intention to attain to an aim.

3) Schwyzer-Debrunnet, o.c., n, p. 313; Kühner-Gerth, o.c., n, p. 379.

Page 50: Gonda.1956.The Character of the Indo-European Moods.pdf

92 The Indo-European moods

1 to d01" This force ofthe subj. is especially evident when, in a dialogue, it resumes an impero of the partner: Plaut. Bacch. 731 scribe :: quid scribam?; Plato Gorg. 447C 8(!OV a1n:6v :: ·d l!(!Wfhal; "what am 1 to ask1" As the imper.-which generally speaking only occurs in positive principal clauses-is the expression of a concept without specialization, which assumes that the person addressed will understand it as a command without qualification, this use of the subj. puts its character beyond doubt: the speaker views the process as a concept, not as actuality.

Despite the rule to use, in Greek, the subj. after lvu etc. "in order that" if the verb of the main clause is primary, and the tendency to employ the opt., if the leading verb is secondary, the subj. is very common in the latter case. The author can adopt the mood (and tense) which the person himself would have used 1). Thus we find & 579 "this the gods wrought (Tevguv) ... that there might be (íva [í0'l) a song". Inversely, a final clause sometimes takes the opto in the case of a primary verb in the main clause: (! 250 tígw "1 shall take (him on a ship) ... lva fhOl PIOTOV 7COAVV tíAcpOl. In principIe, and leaving mechanisation out of account 2), all depends on the question whether the process is viewed as belonging to realizable, 01' to 'contingent', ideal (01' mental) existence. Whereas for instance in Xen. An. 1, 4, 18 7CAoía ~a7:é~avO'ev lva fh-Y} Kv(!ot; blaPñ all attention is focussed on the intention to prevent Cyrus from crossing over, the opto tíAcpOl in (! 250 shows that the person speaking reckons with the possibility that his expectation will not be fulfilled 3), and the opto 1Mmol in y 284f. that the speaker from a distance takes into account the contingent character of the process of burying which Mene­laus himself, (who was much more concerned by the accident) in con­ceiving the idea of paying funeral honours to his comrade had neglected: -"- , \ ".0. I 1" , " _o. I Cf l wt; o fhev sv'U'a ~a7:SO'xsTo ... , OCP(! STa(!OV 'U'a7CTOl. . e. g. a so ~ 26; 65. The fact that "une finale appartenant au passé soit sentie comme se prolongeant dans le présent" (Humbert) has in cases such as Xen. An. 1, 4, 18 not determined the use of the subj.; it is only conditioned by that state of mind of the author-who tends to identify his own point of view

1) " ••• mit lebendiger Zurüokversetzung des Spreohenden in Zeit und Seele des die Absioht Fassenden (Reprasentation)", Rehdantz-Riohter, on Xen. An. 1, 1, 5 and 5,4,21. "The subj. thus used for the opto makes the language more vivid, by introduoing more nearly the original form of thought of the person whose pur­pose is stated" (Goodwin, 0.0., p. 115, § 321).

2) See oh. VII. 3) Cf. also Lys. 20, 21. Humbert whose explanations are somewhat oumbrous­

he doe? not ado:pt the traditional distinotion between primary and seoondary tenses m the mam clause-attaohes too muoh value to temporal implioations. There is. no use in explaining A 94 fíAv{}er;, ¡¡!pea Ujn " ••. venu pour voir" by the assumptlOn that the speaker has .r:t {juetr;; in mind. Goodwin (0.0., § 322) was doubtle~s wrong i:r;t holding that there a:ppears no good reason for the anomaly opto after prImo tense m A 344 eto. A suffimenoy of examples is afforded by Goodwin, 0.0., p. 113ff. and Kühner-Gerth, 0.0., n, p. 378ff.

The Subjunotive 93

with that of the subject of his narrative-which prompted him to choose that categoryl).

It would be beyond the scope of this work to enter into a discussion of all combinations of moods and conjunctions 2). Suffice it to mention that the use of the subj. is natural after ocp(!a and lwt; in the sense of "until, till" 3) with reference to an event of mental, but realizable exis­tence (A 82; 887; MIO)4); after ocp(!a "so long as, while", in relation to processes which are not yet actual (''1' 47 ocp(!a CWO¿O'l fheTeíw "while yet 1 abide among the living"; L1 346; O' 133). Whereas the same 'temporal conjunction' when relating to present 01' past facts is followed by the preso indo or the impf. (cf. e. g. M 141; A 342), it never precedes a future verb form in the opposite case, not even if the principal clause takes a future: {J 204 oMé 7COT' lO'a Il!O'O'STUl ("nor shall-"), lJcp(!a ~ev fj ye btaT(!í­(JrJO'l'V J1XalOVt;5); Q 183; Soph. El. 225. The reason of this is clear: the processes referred to in these clauses appear to the mind of the speaker as a continuation, in the mental sphere, of existing events 01' situations, 01' as a sequel to these, not as decidedly future actuality: "l' 47 (see above) ; Soph. TI'. 148 lwt; nt; UVT¿ 7CU(!&éVOV yvv-Y} I ~Ar¡&ñ; Ai. 555.

Compare, in Vedic, yat + subj. in the same sense (~gveda): 7,88,4 "so long as heaven and earth endure"; 10, 68, 10 "so long as sun and moon (will) rise"; AV. 6, 75, 3 yávat1 súryo2 ásad3 diví4 "so long as1 the sun2 will be3 in the Sky4", and in Avestan: Y. 28,4 yavap isai2 (subj.) tava3éa4 , avat5 xsai6 "so long5 1 will teach6 as1 1 am (will be) able2 and4

(will) have power3 ". In idiomatic Dutch translations of the aboye pass­ages the 4present would be preferred to the emphatic future: Soph. El. 225 ov O'xf¡O'c¡; TaVTat; tíTut;, ocp(!a fhS (Jíot; l!XrJ Dutch "zolang ik leef", Germ. "so lange ich lebe"6); cf. y 353f. Whereas OTS "when, at the time when" is followed by the indico (impf., aor.) with reference to single events 01' actions in past time or (pres.) in descriptions of events which go on now or happen always, the future is only seldom used: O' 272 relating to a definite future: vvg b' l!aTa¿ OTS b-Y} aTvyS(!ot; yáfhot; UVU{JOAf¡O'Sl I OVAofhévr¡t; 8fhé&sv. EIsewhere epic and lyrical poetry has the subj.: 0207 this is 'general' and 'timeless': 80'&AOV ~a¿ TO Thv~Tal, or' tíyysAot; u'iO'lfha elbñ;

1) For A 444; Hdt. 7, 8, 2; Xen. An. 1, 6,6 and similar oases, i11 which purpose and intention are past and presentí see Kühner-Gerth, 0.0., n, p. 380f.

2) For oomparable phenomena in Celtio see Pedersen, Vergl. Gr. n, p.320ff.; Brugmann, Grundr:iJ32, n,3, p.850; for Armenian: Brugmann, 0.0., p.846f.; Meillet, Altarmenisohes Elementarbuoh, Heidelberg 1913, p. 137; for Latín e.g. Leumann-Hofmann, O. o., p. 762.

3) Hirt (0.0., VI, p. 274), too apodeiotioally holds that the subj. "in Folge- und Absiohtssatzen natürlioh futurisoh ist".

4) Whether the realization is supposed to ooour at an unoertain future time (Liddell-Soott, s:v.) is a matter of indifferenoe.

5) After an impero e.g. P 186. 6) (Sohneidewin-Nauok-)Bruhn10 unduly stresses "den futurisohen Sinn".

':

Page 51: Gonda.1956.The Character of the Indo-European Moods.pdf

......... -------~----------------------........ --------------------

94 The Indo-European moods

cf. L1 260; in anticipating happenings of uncel'tain futurity, which even may nevel' arise: T 337 "waiting evel' fol' woeful tidings of me" eh' ~nocp{hp,~'/Iow nv{}r¡r~l. After a series of emphatic future forms, den~ting m boastmg and mmatory speech pro ces ses viewed as (future) reality a ?rs clause referring to an action anticipated in the sphere of 'ideaÍ eXlstence', contains a subj.: rp 316ff.: "he shall (earal) have no need of a mound, when they make his funeral ({}ánrwaw)". Similarly the anti­cipation of Ilios' destruction is confidently stated to be reali~ed in the future: L1 164 eaasral i¡p,ae, 8r' 11.'/1 nor' O~WAn 'lAws fe1}. In connection with the 'iterative' 87:8 the subj. proves to be used irrespective oftime: cf.I 101 ~future); '/1 ~01 (pre~en~); E 522 (past); A 80 (timeless in a general say­mg): what lS essentlalls that the speaker, while conceiving the contents of the sentence in its entirety as actual, views the contents of the sub­ordinate ("temporal") clause as existing in the realm of the mind not ref~rrin~ to p~rt~cular actual ~ventsl). The subj. after net'/l may b~ ex­plamed m a SImilar way: e 9 she will not cease from wailing ... until sh~ se~s my ve-:y self": net'/l y' afn:6'/1 p,s tor¡ral: the process of seeing eXlst~ m the mmd of the speaker, not (yet) in reality. Of. also E 135. ~entlOn may be made here oíthe construction represented by Hdt. 7, 54 ,., Sil ér¡r"x ~" , , , "'" o:S" , • " • sv sro ... p,r¡USp,ta'/l Ol aV'/IrVXlr¡'/I rOlavrr¡'/I YS'/lsa{}at, fj p,w navasl ua7:aareS'IfJa~{}al rr¡'/I Evewnr¡'/I ne67:8eo'/l 1} Bni dep'aal ro tal BU8Í'Vr¡s yBYr¡ral where the mf. (cf. 7,2) would be the usual construction2).

We might here recall to memory Vedic instances oí yat + subj.: ~V. 5, 31,5 yátl te2 vf~a'iJo3 arkám4 árciin5 "whenl the bulls3 sang5 the song of praise4 for Thee2" 3); cí. 6; 1, 72, 3; 4, 55, 2 b. For the use of the subj. in. connection with .a temporal relative in Anc.lndian, see e.g. ~l~o ,,~almBr. 3, 122 ~adat.lva2 vayarfl,s yunajiimahii4 , athii5nviidhiivatiid6 ttt m dem Augenbhck, m deml,2 wir3 auíbrechen werden4 sollst du hinter (uns) h~r laufen6

" ~ (Oaland). In contradistinction to Delbrück4) who was convmced that m sentences of this structure the subj. corres­ponded to the Latin fut. :x~ctum, Minard 5), rejecting this supposition, held that such a nuance, lf lt existed, must be due to the mood not to ~he cOI~.junction. That is true, but the main point to be borne in mind ls-as ~t appears to me-the non-temporal value of the subj.: in the ~bove mstance the process of making ready íor departure is a mental lmage crea~ed ?y the. s~eaker's visualizing; it is an image called up in the hearer s mmd. SlmIlarly SatBr. 10,4,3,9 6 ) yadailva2 tvam3 etam4

1) ~e also find ~ subj. after an indefinite "when", the folIowing prooess being an obJeot of refleotlOn, no hio-et-nuno aotuality: see e.g. RV. 6, 35, 2; 3.

2) See Stein on this passage. . 3) F. Bollensen, ZDMG. 22, p. 574 regarded this subj. (and also those of 4 55 2'

6, 17, ll) as equivalent with the indio. ' , , 4) Delbrüok, Altind. Syntax, p. 325. 5) Minard, 0.0., p. 154f. 6) See Minard, 0.0., p. 154.

The Subjunotive 95

bhiigalp,5 hal'iisii6 atha7 ••• sarire'iJa8 mrto9 'sadlO yoll "quandl ,2 toi tu3 l' 4 auras pris pour6 ta part5, alors7 seulement immortel8,9 seralO

quill .•• "; both processes, that of taking and that of being, are, at the moment oí speaking, mental images; the relative chronology oí the processes is left unexpressed. In Greek parallel instances may be found: II00f. reO as xer¡ ... Uer¡fí'/lal 08 uai aAAq.¡, 87:' ay rwa {}vp,os aywyn which may be rendered by "may have promptedhim to ... ". That languages which possess a futurum exactum would use that íorm 01' construction in translating is another question. In case of indicative verb forms being used the internal chronological relation is not expressed either: SatBr. 3, 1, 2, 9 sal yadii2 kesasmaSl'u3 vapaty4 atha5 sniiti6 which must mean l ) "after having cut2,4 hair and beard3 , hel has a bath6 " (not: "when he ... , he ... "), the processes being successive, not contempo­raneous. It may be remembered that the relative chronological function oí the Latin íut. ex. has developed only in the course oí time in those subordinate clauses which implied anteriority: Plaut. Rud. 755 postea aspicito meum, quando ego tuum inspectavero 2). It remained foreign to Greek, where the aor. subj.-which is the source oí the Latin fut. II­must be in a modal, not in a temporal sense: A 128f.; e 549 3 ).

The subj. is, in Anc. Ind., incidentally also connected with yatra 4). As

to Bar Up. 3, 9,25 ahallikelti 2 ho3viica4 ••• yatrai5tad6 anyatrii7smanS

manyasai9, here the subj. which was not clear to Delbrück 5), was, 1 think, used because the last words of the interlocutor "on what is the heart based1" implied, in the opinion oí the speaker, the thought: "the heart is lfused on such and such a thing"; hence his answer: "you idiotl, that5 you think9 that it6 would be elsewhere7 than in ourselves8 ".

It may thereíore be concluded that this mood is especially apt to express a secondary process, that is to say: a process continuing, following aíter, succeeding to and resulting írom 01' being made possible by another process.

Hence also the very pronounced tendency to use the subj. in sub-ordinate clauses 6). There it is the counterpart of the imperative in

1) Cf. also Minard, 0.0., p.21. 2) For partioulars: Leumann-Hofmann, Latein. Gramm.6, p.563f. 3) Cf. also Kühner-Gerth, 0.0., n, 1, p. 181, Anm. 2. 4) Minard, 0.0., p. 132. 5) Delbrüok, 0.0., p.326. That yatra + subj. "auf den Zeitpunkt hinweist"

is no felioitous remark.-Cf. also JUpBr. 1, 13,2 tatrailva2 lcuru3 yatro4pajiváme5ti6

"make3 (it) there1 where4 we oan (may, will) live5 on (it)"6. VadhS. (Caland,A. Oro 4) p. 195 a subj. whioh does not depend on a oondition follows on oonditional opta­tives: yatrai1narn2 adhva1'yur3 apo4 'vaghrápayat5, tad6 apáyaccháthe7tiB "wenn1 ihm2

der Adhvaryu3 das Wasse1'4 zum Besohnuppern vorhalt6, S06 sollt ihr es weg­nehmen7"B. See also Renou, Déo., p.27.

6) In the anoient Vedio texts the 'indifferent' injunotive is signifioantly infre­quent in subordinate olauses: Delbrüok, Altind. Synt., p.358f.; Renou, Déo., p. 12.-For the deoline and disappearanoe of the subj. in Vedio, with whioh we oannot oooupy ourselves here, see Renou, 0.0., p.14ff.

Page 52: Gonda.1956.The Character of the Indo-European Moods.pdf

96 The Indo-European moods

principal clauses 1): Ath V. 1I, 1, 30 pánthiir¡¡,1 ... ádhi2 rohayai3nam4 I yena5 rohéit6 "make him4 ascend2, 3 the path1 ••• , by which5 he will ascend6 (to the firmament)". 01' it occurs, in similar formulas, beside an opto in main clauses: TS. 4, 1,2,3 yátal)} khánama2 "(tell us) whence1 we shall dig (him) Up2": 4 tátal}} khanema2 "thence1 let us dig Up2 (Agni)" 2). There are parallel instances in Greek. Applying to definite particular occurrences: r 285ft'. " ... let the Trojans give back Helen ... and pay recompense, even such as shall abide in the minds of men that are yet to be": fj Te •.• nSArríat3). That in Vedic the subj. was the mode oí subordination par excellence is also evident from the rare appearance of a fut. subj.4): l;tV. 4, 30, 23 yád ... kari§yáJ:¿.

The constructions of this type are often virtually conditional, 01'

temporal-conditiona1. l;tV. 10, 2, 2 sváhii1 vayár¡¡,2 kr'l}ávama3 havír¡¡,§i4 dev65 deván6 yajatu7 "when we2 prepare3 oblations4 with svahá1 the god5

must (shall) adore7 the gods 6 " but more exactly: "we-as a process oí mental existence-prepare ... (notice the accent!): the god shall ... "; 8,100,2; 10,101,3. The r61e of the subj. is obvious: "we prepare (in the mind) oblations ... " Le. "put the case: we p. O .... "5). A clause oí this type can be introduced and included by ya-: l;tV. 7, 88, 6; by yad: 1,54,5. In Vedic prose this construction is comparatively speaking frequent: the re1. clause usually precedes, the main clause containing another subj. serving to makeknoWll the speaker's wish orthe inevitabil­ity of the occurrence. These sentences express, inter alia, a threat (e.g. TS. 2, 6, 10,2; the main clause may be omitted: PBr. 14,6, 8 ya1 indrar¡¡,2 yajátii3 iti4 "ifl anyonel worships3 Indra2" 4), a promise, an agreement, the formulation of a wager (e.g. PBr. 21, 13,2 váco mithunena vijayamahai: yatare no vaco mithunar¡¡, na prativindár¡¡,s te para bhaván iti)6). That inevitability can be expressed by the subj. follows from its function to indicate also processes which in the speaker's eyes are impending, 01' are a logical necessity to follow from the realization of a condition. Whether 01' not the process denoted by the subj. is, in a given passage, inevitable, does not depend on the subj., but on the context in the largest sense of the termo When after 8e; "whoever", in the < collective hypothetical sense', the subj. is used, the clause likewise bears a general-conditional force: (l 165 "a son when his father has gone has many sorrows, when (if) there are no other men to be his

1) Cf. also the instances given by Bloomfield and Edgerton, Vedic Variants, I, § 93, who however did not notice this point; Renou, o.c., § 15f.

2) See Bloomfield-Edgerton, o.C., I, § 169. 3) For the original character of the so-called relative clauses see my relevant

paper in the Lingua, 4, p. 1ff. 4) Macdonell, Vedic Grammar, § 538a; Renou, Gramm. védique, p. 145. 6) For conditional clauses see chapter IX. 6) For particulars see Renou, Déc., p. 27.

The Subjunctive 97

helpers": 0 fl/Y} liAAOt ... gWGtyl). Cf. also E 81; Hes. Op. 327; cf. also T167.

The subj. is also, in a significant way, used in putting a case: TBr. 1, 5, 9, 4 yad va idam uccair yajñena carama (followed by a main clause in the indic.) "(posons le cas) OU nous procéderions au sacrifice (avec récitation) a voix haute" 2). The Latin subj. is likewise used in asserting one's conviction with regard to one's own aetion in a supposed case: Plaut. Capto 274 Thalem talento non emam; also (more generally) in eonnection with a supposed case: Plaut. Bacch. 139.

There is a strong tendency in several anCÍent l.-E. languages to use the subj. rather than the indic. in generalizations and predications of general application 3)4): l;tV. 8, 31,1 y6l yájiiti2 "(anyone) whol wor­shipS2"; AthV. 1I, 1, 13 "seize then of them whichever (yatamiiJ:¿) is (asan) ... "; l;tV. 5, 29, 14 yá1 cin2 nú3 vajrin4 kr'l}áv05 dadhr§ván6 ná7

teS vartá9 távi§yá10 astill tásyál}}2 "there isll nobody 7 who will cheek9

thineS strength10, whateverl ,2 Thou (wilt) boldly6 perform(est)5, O bearer of the vajra4 ". In Greek, a 351 men praise (ind.) that song the

t ,t{ , , " E 407 '.Il \" '.Q , mos, '/ Tte; .•. v8wTaTr¡ ap,ljJtn8/1.r¡Tat; ov ur¡vawe; oe; a'U'avaTouJt p,áxr¡Tat5); I1I7; N229; 8448; Hes. Op. 224; 327 6). In this language the subj. is tJ1>ical of the present general protasis as opposed to the indico in the present particular protasis: ~ 373; ef. A 192; A 218; and in 8e; Tte; clauses: v 335 yf¡p,aG~' 8e; Tte; ... nA8tG'ía n6enGty7). Thus, in ~ 32 the sentence taken as a whole is an assertion, but the 8'íte; clause exists only as an idea, no particular person being meant: oMe yae ovOs Tte; liAAOe;, 8Tte; ..• lur¡'íat, I h1}á08 ... or¡eov p,Bv8t. Subj. and indico can alter­nate: l;tV. 6, 52, 2 átil va2 y63 maruto4 mányate5 n06 bráhma7 vás yá7y,9 kriyámá'l}ar¡¡,lO nínitsatll "oder2 wer3 sieh mebr1 dünkt5 als wir6, ihr Marut4, oders wer9 die (von uns) getane10 feierliche Rede7 sehmahen willll " (Geldner) 8). A general-ideal sense must be attributed to this

1) For "0_ = Iláv 'n_" see Lingua, 4, p. 33f. In my opinion *jo- originally was, in the maín, an "emphasizing", "defining" and isolating element, the conditional relation being implied in the context and the construction of the sentence in its entirety.

2) Renou, Déc., p.29, who observes: " ... il est visible que le subj. dans les subordonnées des Br. n'a pas de valeur par lui-meme; la nuance éventuelle qu'il porte en général résulte du contexte plutót que de la forme meme, sinon l'on ne comprendrait pas pourquoi l'usage en est aussi strictement déterminé."

3) See e.g. Speyer, V.S.S., p. 85 (§ 272, 3). 4) Hirt was of the opinion that this subj. represented the 'injunctive' (indic.

aor., gnomic aor.), see: Indog. Gramm., VII, p. 151. Despite the similarity of the functions it would be difficult to substantiate this thesis (see chapter IV).

6) Hirt, Indog. Gramm., VII, p. 150, n.2 correctlyobserves: "Der Konjunktiv ist hier zeitlos."

6) For a collection ofpost-Homeric instances see Stahl, Krit.-hist. Syntax, p. 294ff. 7) For del or é"ácn:oTe in the main clause see Stahl, o.c., p. 310f. 8) See also Renou, o.c., § 23 "nuance d'indétermination, si l'on veut, résumant

une expérience générale~'. 7 Gonda, Indo-European moods

Page 53: Gonda.1956.The Character of the Indo-European Moods.pdf

98 The Indo-European moods

mood in such instances as ~V. 7,53, 3 asmé1 dhatta't'j1,2 yád3 ásad4 áskr­dhoyu5 "bringet2 uns1 das, was3 ungeschmiHert5 sein wird4". Cí. 5,53, 15. The speaker does not confine the contents to hic-et-nunc actualityl).

The aboye examples are all taken from subordinate clauses. The general character oí the clause is not always determined by the relative pronoun, in ~V. 1, 70, 7 the rel. is a substitute for Agni, but the subject of the sentence and the temporal location oí the process are indefinite: vá1"dhéin1 yám2 pÜ1"vtl}} k~ápa7y,4 "whom2 (Agni) they make increase1 in many3 nights4"; cí. 72,3; 6,17,112). The process do es not apply to particular actual processes, but to 'mental reality. In Latin we find the subj. in examples ofthe 'generalizing ideal second~erson': Plaut. M.G. 689 hoa numquam ve1"bum ex UX01"e ("from a wife") audias.

Now, is Miss Hahn3) right in contending that this use can be ac­counted íor "precisely on the assumption that these subjunctives were originally futures, and in no other way whatever"? Or Stahl4) in deíending the opposite view: "Dadurch daB nun diese generellen Satze schon bei Homer nicht mehr als reine Begehrungssatze empfunden wurden, konnte zeitliche Relation bei ihnen eintreten"? It may be argued that in passages such as 1116f. avd vv nOAArov I Aarov eC1UV av~e, ov Te ZBVr; ufíet cptA~C1rJ the verb does not refer to an actual process, to a "here and now", to an actual particular case, but to an ideal process predicated of Zeus with refence to man (in general, i. e. to any man whom the god loves). Here the act exists olliy as an idea ;-but the statement is applicable to any case which may arise, and so the idea can materialize anywhere and at any time. "Erwartete Verwirklichung" is too restricted a description5), because the judgement also applies to instances in the past (cf. e. g. ~ 373) and in the present (cf. 1116f.: 118 wr; vvv ... ); it may apply to ever recurrent events or to the periodic phenomena in nature (A 192), which are essentially timeless (not only expected in the future) and have nothing to do with volition.-The construction a 349 (ZBVr;) tJt~WC1tV I ... onwr; B{MArJC1tV, 8uáC17:q> can be judged in a similar way, where the act

1) Hence also, I suppose, the Latin phrase quod 8ciam (see W. G. Hale, The mode in the phrases quod 8ciam, etc., Trans. Amer. Phil.Ass., 22, p. 105ff.): Plaut. Men.1106 nil 1'eticebo quod 8ciam "nothing that I (not particularly now, but generally speaking)' know will 1 keep back"; hence also q.8. "so far as I know": Amph. 749 numquam tactum est, quod 8ciam "it never happened so far as I know".

2) Geldner's (Rigveda I, on 1,70,7) attempt to explain this use as a "Konj. bei Zeitangabe"-he translated the verbs by preterits, regarding them in a note on 6, 17, 11 as "eigentümlich"-has deservedly been combatted by Renou, O. c., § 23. I for one am inclined to judge the subordinate clause in J.1,V.6, 38,4 in the same way: the subject is "soma and sacrifice" in general; Geldner' s "starken moge" should be read: "starken".

8) Kühner-Gerth, O.C., Ir, p. 250f.; Hahn, o.c., p. 9. 4) Stahl, o.c., p.247. 5) Of. Mutzbauer, Grundbed. d. Konj. U. Opt., p.56 " ... ein Mann, dem

gerade, wie ich erwarte, von dem ich erwarte, da13 Zeus ihm seine Liebe zuwendet" (a translation not to be endorsed).

The Subjunctive 99

of willing does not bear upon particular actual ha~p~nings. As opposed to an unequivocally past tense the s8nse of the subJ. lS of course future: RV 1 113 10 yá1 vyüsúr2 yas3 aa4 nünám5 vyuahán6 "which1 have shone í~rti/ (hitherto), and4 which3 shall shi~e forth6 hereafter5"l).

Similar remarks apply to the so-called iterative subj., i. e. the appear­ance of the mood in a context implying iteration: Eur. Alc. 671 r}veyyvr; BA{}rJ {}áva7:or;, OMB¿r; f30VABrat (}Vf¡C1UstV. The apodosis expresses a repeated 01' customary action 01' a general truth. Here we should not say that the 'iterative' is a "futurischer Konjunktiv" 2), because the ideal process of coming can be realized in any particular case; since the aphorism giv~s evidence of the speaker's wordly wisdom, realization in the past lS included. Cf. also Dem. 2, 12 anar; A6yor;, av anñ rO: neáyp,am, p,árat6v u cpat11Brat ua¿ uBv6v "all speech, if deeds are wanting, _ appea.:s m.~re emptiness and vanity".-Cf. e.g. also SatBr. 1, 8, 1, 6 ya~a~ yavat as graduallyas" + subj ...... tavat !a.vat ;'so" + subj. The Ved1c lllle.(~V. 6~ 28 3) ná1 tá2 nasanti3 ná4 dabhat~5 taskam7y,6 must be translated. they do' not1 become lost3, n04 thief6 can har1ll5 (them)", because the context makes it clear that no special thief is llleant. If this, however, should. be the case the sentence would lose its general character and the translatlOn would be " ... the thief will not ... ". In AV. 4, 16, 4 y61 dyám2 ati­sárpat3 pa1"ástan4 ná5 sá6 muayatai7vá1"u1Jasya8 t'ájña7y,9 "he (of ideal existence) who1 steals3 far away4 beyond3 the Sky2, he6 ~oe~ not5• (ir: my mind's eye) escape7 king9 Varu1).a8 " the former clause m Itself lS mde­finite 01' general, the latter is determinate only i~ appearan~e. In trans­lating the subjunctives are, however, rendered d1fferently: though one steal away" 01' "whoso should creep ... ", and "he shall (can) not:' respectively3). Once again the specialized function traditionally attl'l­buted to the mood really belongs to other elements of the sentence or to the utterance in its entirety: hence the predilection of this subj. for clauses introduced by the so-called relative (or; ur;: v 335!; see above)4~.

The subj. in the indirect questions 5) can in my opinion also be conS1-dered to express 'ideal existence': f3 332 rtr; ~' ol~' sl ... I ... an6Ar¡mt "who knows if he ... may (will) perish ... "; X 245f. Curiouslyenough this subj. sometimes joins a future: C1 265 ovu ol~' BL U8V p,' &V8C1St (}sor; 1í usv aAww "1 do not know whether the god will bring me back, 01' whether 1 shall fall into the enemy's hand"; e 532. In using the future the speaker

1) See aboye. 2) Thus Hirt, Indog. Gramm., VI, p.275. . 8) J.1,V.9, 101, 9 vanámahai not exactly "siegen mogen" (Delbrück, Altrnd. Synt.,

p.318). Other instances: ibidem, p. 319f. 4) I fear that I cannot subscribe to the explication given by Humbert, o. c.,

§ 186. 5) For the term "indir~ct question" see Lingua, 4, J?' 27,0~. In a bro~der sense

the term is also used to rnclude the types 11 365 fFl2uCO>¡ts1} o:n;o>, ••• YS1I1),a¡ a~d similar constructions; O> 217 etc. Of. also Kühner-Gerth, O.C., I, p. 222; Goodwrn, o.c., p. 265f.; Stahl, o.c., p.310f. 7*

Page 54: Gonda.1956.The Character of the Indo-European Moods.pdf

100 The Indo-European moods

gives evidence of a more or less temporal view of the process. But one indico can suffice. It may be suggested to view this combination of fut. and subj. in the same light as the well-known 'omission' of the pre­position in the second member of phrases such as Thuc. 3, 67, 1 ual V:rCS(! vf1'ciw ual f¡¡u'iw, the addition of the preposition (e.g. Xen.An. 1, 1, 7 ual ua'ta yfj'V ual ua'ta 'l9áAana'V) being emphatic or a means of effecting clearness 1).-Passages such as E 33 "shall we not leave the T. and A. to fight", Ó:rc:rcO'tl3(!Ola¿ •.• ZSV~ ui5<5o~ o(!é~rJ admit of the same explication. There is for the rest much diversity among the l.-E. languages with regard to the use of the subj. in indirect speech (coniunctivus obliquus). For Greek, reference may be made to the aboye observations and the handbooks 2).

There is, in Vedic, an interesting 'particle of interrogation', used in direct and indirect questions, kuvíd: the verb following it---which is almost always in the subj.-does not lose its accent; that is to say: the clause containing this verb is characterized as subordinate. Cf. ~V. 3, 42,2 tám1 indra2 mádam3 á4 gahi5 ••• kuvín6 nv7 asya8 trp1}ávaJ:¡,9 "come4o,5, O Indra2, to this1 intoxicating drink3 ••• (where) indeed6 thou (shalt) become(st) satiated9 with it8"3), where the kuvíd clause depends on the preceding words. Cf. 6,23,9; 5,3,10. Often, however, the kuvíd clause does not follow another sentence with which it is closely connected 4 ) :

4,51,4 kuvít1 sá2 ... sanáyo3 návo4o va5 yámo6 babhüyád7 ••• "should7 this2

be7 an 0ld3 course6 or5 a new4 ••• F (with the verb in the opt.); 1, 143, 6 kuvín1 no2 agnír3 ucáthasya4 vír5 ásat6 "ob1 wohl1 AgnP an unserem2

Preislied4 seine Lust5 habenG wirdG?" (Geldner). Delbrück at the time surmised that these kuvíd clauses "von einem nicht ausgesprochenen Gedanken innerlich abhangig sind". It would be preferable to say that in these cases the mood and the accent of the following verb are condi­tioned by the mood, view or opinion of the speaker at the moment of pronouncing the more or less interjectional kuvid5). The subj., here

1) An emphatic eyro (see my paper in the Acta Or., 19, 1942, p. 211ff.) is not repeated either if a second verb is added.

2) See e.g. Schwyzer-Debrunner, o.c., n, p. 319. 3) " ••• ob du dich ... erfreuen mogest" (Delbrück); "gew:iL\ wirst du ... "

(Geldner). The 'original' sense of the particle which probably represents ku "howjwhere, anyhowjanywhere" (for the so-called interr.-indef. pronoun see Lingua, 4, p. 241ff.), and id "indeed", may have been "anyhow indeed ( ... 1)". Cf. Pal}.ini 8, 1, 30; Bhat;likas. 2, 7; H. Oldenberg, Zs. d. Deutschen Morgenl. Gesell. schaft, 60, p.738; Delbrück, Altind. Synt., p. 550f.; Renou, Gramm. véd., p.370 (who translates kuvid by "est-ce que par hasard 1").

4) In contradistinction to Delbrück, Le. I would defend the "subordinate" eharacter of sentences like J;tV.2, 16, 7 (cf. also Geldner's first interpretation: "Wettbewerb verschiedener Opferer um Indra's Gunst") (Geldner, Der Rigveda, P, p.297).

6) Geldner, o.c., nI, p.345 even goes so far as to render kuvit in 10,119,1 lcuvit sóma8yápam iti by "ieh merke, (daJ3 ieh Soma getrunken. habe)".

The Subjunctive 101

again, conveys the idea' mental existence of the process' , not infrequently in close succession to an antecedent clause: 2, 35, 2 "we would utter this hymn for him" kuvíd1 asya2 védat3? "perchance1 he will take note3 of it2 (?)"; the accent marks the subordination. We might compare hanta 1)

-an inceptive and exclamatory particle, expressive of exhortation­which falls under the same rule, formulated by PaI).ini 8,1,30: SatBr. 1, 2, 5, 2 hánte1má'l'(l-2 prthivf'l'fb3 vibhájamahai4 "come on tI let us divide4

the2 earth 3 among ourselves4 ". In the bráhmaI).as the hanta construction tends to join artother sentence pronounced by the same speaker 2) : JBr. 1,294 te abrütam: ittha'l'fb ced vai bhavi~yavo, na vai tarhi 8ak~yavaJ:¡, praja bhartum; hanta rüpa1Ji vyati~ajavaha iti. Compare also the-likewise afore-mentioned-Gr. <5SV(!o: 'l9 292 d8V(!O rptAr¡ AéU't(!O'Vd8 T(!a:rcstop,8'V; Eur. Ba. 341 <58v(!6 aov ad'l¡JOJ uá(!a; and <58V't8 in 8128 <5SVT lop,8'V. These indeclinable words, like the verbs aboye mentioned, help the hearer to understand what the mood alone leaves undefined.

In combination with néd "not" (na + id) the accented subj. expresses the idea of "lest", e.g. ~V. 5, 79, 9. Thus Ath V. 13, 1, 12 má1 má2 hasin3

nathit640 nét5 tvaG jáhani7 must mean: "let1 him3 not1 abandon3 me2,

lest5 1, a suppliant4, abandon7 thee6 "3). In the bráhmaI).as the use of ned becomes more frequent. "C'est paree que le subj. hérité tendait vers une nuance «subordonnée» qu'il s'est réfugié dans cet emploi et a orienté net dans le sens d'une particule fortement subordonnante."4o) The particle ned is not essential in this use: AiBr. 8, 6, 1I where na fulfills the same task. This construction was also taught by PáI).ini 3, 4, 8 ("apprehen­sion"): nej jihmáyanto naraka'l'(l- patama "de crainte que, nous detournant du droit chemin, nous ne tombions dans l'enfer" (Renou). See e.g. also SatBr. 4, 1,5, 7.

Mental 'reality' of the process must indeed be sought in the subj. after verbs expressing fear or apprehension: 1 244 al'Vw~ <58t<5olua ... p,~ ol a:rc8lAa~ I ¿u'tsAéaOJat 'l9sot "this is the great fear of my heart, lest the gods fulfil for him his boastings" (Murray) 5). This construction is usually, and rightly, explained from an original "1 fear (thinking) ... should not ... ". It must however be remembered that ¡d} rejects the realization of the thought expressed by the subjunctive verb form:

1) Curiously enough both kuvit and hanta can be followed by an enclitic word: since this does nót occupy the initial position, the interjections form part of the sentence.

2) Particulars: Renou, Déc., p. 19. 3) Cf. also Whitney-Lanman, Atharvaveda, p. 712; Renou, Déc., p. 24. For net

see also: Delbrück, Altind. Syntax, p. 316; 545; Vergl. Synt., nI, p. 288; 295. Other instances of the use of the construction: A. B. Keith, Rigveda Brahmanas, Harvard 1920, p.93f.

') Renou, Dée., p. 24. 6) See also Hirt, Indog. Gr., VI, p. 274. For the clauses with pI¡ see also ch. X.

Kühner-Gerth's (o. c., n, p: 394,4) opinion with regard to the following elr¡ is susceptible to improvement.

Page 55: Gonda.1956.The Character of the Indo-European Moods.pdf

102 The Indo-European moods

cf. B 195 "(did we not a11 hear what he said?)" pJ/ n ... esgr¡ uau6v "let him not ... ; beware lest he ... ". In a principal clause which constitutes the content of a command the same subj. is found: A 26 '>C(!aT8(!OV 0.' bd ¡.tv{}ov 8T8AA8' I ¡.tI¡ 118 UlX8tW ("Abmahnung"). For the particle ¡.tI¡ + subj. used in rejecting a thought compare also: X 123; lJf 407 '>ClXáv8T8, ¡.tr¡08 Atnr¡l1{}ov; q; 370 1); 8 356 W flOl eyw, fll¡ 'ite; flOl vq;atvr¡(J¿v 06AOV 2); E 684. Cf., in Anc.lnd., SatBr. 11,5,1,1 akama'YfL sma ma ni padyasai which comes to "the idea of your lying with me against my will is to be rejected" ; (and in Latin: Plaut. Poen. 251 ne ... vitia loquamur; Aul. 241 novi, ne doceas). The Romans said cave neges (Plaut. Most. 1025) just like fac abeas etc. 3).

In a sma11 number of cases the Vedic subj. is also accented after an <introductory' 01' leading verb 4): ~V. 5, 45, 6 éta1 dhíya'YfL2 kr1Jáváma3 "come on1, we will perform3 the act of devotion2"; an interesting in­stance is: 1, 81, 3 yuk~vái ... hári2 ká'YfL3 hána1J4 "yoke1 thy bay steeds2 ! whom3 wilt thou slay4?" The mental activity leading the speaker to formulate a process by means of a subj. verb form is sometimes indicated in the text: II 646 q;(!áC8'i0 -Dv¡.tfP, I nOAAa ... ¡.t8(!¡.tr¡(!tCwv, I fj ñor¡ ud uúvov ... I ... "EuTW(! I ... Or¡WI1r¡. Cf. also e 34 we have pity on the ... spearmen 01 U8V MI ... OAW')J7:al; II 435; n 74; '1jJ 140 5). In Anc. Ind. TS. 1, 5,9,4 SOl 'manyata2 : ima'YfL3 ... stavani4 "he1 reflected2: «let me praise4

him3 ... »" (MS. 4, 13, 4 etc.;) AiBr. 2, 7, 10 urüka'YfL1 manyamana2, ned3 vas4 tolce5 tanaye6 mvita7 mvat8 "deeming2 (them) an owll, lest3 in your4

progeny5,6 a howler7 may howI8"; KBr. 13, 3; AiBr. 3, 20, 1 avet ("he perceived") ... hanta + subj.; JBr. 1, 105 akamayante1má'YfL12 10káñ3 jayama4 "they desired1: may we win4 these2 worlds3". PBr. 6, 5, 12 after "to curse". Sometimes the expression of the thought which contains the subj. is closely connected with a yat clause, e.g. AiBr. 6, 18,9 yad1 enáni2 sa'YfLsanty3 ahinán4 svargá'YfL15 lokan6 ••• avapnavame7tí8 "the fact that1 they recite3 them2 (gives evidence of the intention formulated by the following words pronounced by them)8: «let us obtain7 the worlds6

of heaven5 without defect4»" 6). Attention may also be drawn to the subj. after the phrase vara'YfL vr1Jite "to choose a boon", e.g. TS. 2, 4, 1, lf. and after "choosing": e.g. JUpBr. 1,51,6. In Sanskrit the intention can, of course, also be indicated by iti: AiBr. 8, 24, 2 tasmad1 1'aja2 purodadhita3 deva4 me5 'nnam6 adann7 iti8 "therefore1 the king2 should

1) Of. e.g. Schwyzer-Debrunner, o.c., Ir, 316 in connection with A 470. 2) For particulars: Schwyzer-Debrunner, o.c., Ir, p. 317, 6; Kühner-Gerth,

o.c., I, p. 224f.; Humbert, o.c., p. 115; Slotty, o.c., p. 38ff.; O.Hentze, Bezz. Beitr. 28, p. 199ff.

3) See Durham, o.c., p. 54ff.; Morris, o.c., p. 379; Bennett, o.c., p. 232; 323; (Leumann-)Hofmann, o.c., p. 689, and Delbriick, Vergl. Synt., IIr, p.420.

4) See Oldenberg, ZDMG. 60, p. 731. 5) Notice also such constructions as f3 186 oweov nortoeyftsVOs, aZ us n6enaw. 6) For further particulars see Renou, Déc., p. 18. .

The Subjurl.ctive 103

appoint a purohita3 (expressing his intentions by the words:)8 «(then) the gods4 (will) eat7 !hy5 food6»8 (01': «in order that the gods ... »)"1).

We may, 1 think, with good reason explain the rise of the subj. in ~V. 10,27, 1 and similar passages, where it occurs in the very beginning of a literary unit, in a similar way: ásat1 sú2 me3 jaj'ita1J4 sá5bhivegá1J6 . .. "well2, that5 (1 see in my mind)l is1 my3 impetus6

, O invoker4";

10, 165, 1, where the label <voluntative' might do duty; cf. Ath V. 6, 98, 1; ~V. 1, 94, 4; 7, 96, 3: everywhere mental existence awaiting actualization. -Similarly, ~V. 10, 50, 5 (Indra) áso nú kam ajáro vardhas ca; 1, 176, 5. It is, in view of passages of this kind, not possible to maintain that the ancient Vedic subj. is always "un eventuel conditionné (dont) la réali­sation escomptée est soumise a d' autres événements qui ont place dans l'intervalle" 2).

In the ~gveda the subj. also occurs in passages cited from one of the persons mentioned in a text in order to make his thoughts 01' opinion known to the hearer. The verb of "saying" 01' "thinking" is always added: 8,93,5 yád1 ... ná2 mara3 íti4 mányase5 "if1 Thou (Indra) thinkest5 : «l shall not2 die»3" 4 ••• , which comes to "il est exclu que je meure3)" 01' rather "it is, in my opinion, out of the question that 1 die". In the mantras of the younger saJ:p.hitas the verb is usually omitted: Ath V. 10, 2, 5 k61 asya2 báhil,3 sámabharad4 virya'YfL5 kamvad6 íti7 "Wh01 brought together4 his2 two arms3, saying7 «he must perform6 (deeds of) heroism5})" 7? The intention of the subject of the antecedent clause is clearly expressed in passages like Ath V. 3, 12, 1 ihaí1va2 dhfuvá'YfL3 nt4

mínomi5 sálá'YfL6 k~éme7 ti~thati8 "just2 here1 1 erect4,5 my dwelling6

so that it stands8 firm3 in security7"; 5, 18, 2 sá1 ••• gám2 adyád3 adyá4

jivani5 má6 svá1J7 "it may be that3 he1 will eat3 the cow2 <thinking) «in order to live5 today4, not6 tomorrow7»".

Many authorities have endeavoured to derive the subjunctives oc­curring in subordinate clauses from the supposedly original types of usages (volitive, prospective etc.). Bennett, for instance, distinguishes between a subj. in "substantive clauses" developed from the volitive, a subj. in subst. clauses developed from the deliberative, a subj. of purpose etc. etc. 4). According to Ernout and Thomas "les propositions finales, indiquant le but visé ou le résultat cherché, sont au subjonctif,

1) Not: "wishing «may the gods»" (A. B. Keith, Rigveda Brahmanas, Harvard 1920, p. 339).

2) Renou, Déc. et disp. du subj., p. 9, § 14. 3) Renou, o.c., § 13. 4) Bennett, o.c., I, p. 208-347. Of. also (Leumann-)Hofmann, o.c., p. 708f.­

In Latin the subordinate clauses in the indefinite 2nd singular stand, as a rule, in the subj., the qrigin of which does not seem to be so obscure as was assumed by Bennett (o.c., I, p. 320): Ter. Phorm. 265 unum quom nOl'is omnis nOl'is.-Space forbids the discussion of the factors which have contributed to the establishment, in Latin, of the subj. in those constructions where it was not original.

Page 56: Gonda.1956.The Character of the Indo-European Moods.pdf

104 The Indo-European moods

employé conformément a son sens propre d'intention ou de volonté"I). Humbert and Schwyzer-Debrunner 2) are of the opinion that the pre­sence of the voluntative subj. characterizes a subordinate clause as final.

If w.e however remember the above tendency of the subj. to join a precedmg verb and to express a subsequent (01' 'consecutive') idea on the one hand, and the frequent cases of addition of a verb expressing mental activity on the other, we always find the same schema, whether the antecedent verb expresses a command-e. g. Plaut. Truc. 839 eloquere haec erae: puerum reddat-, a mere desire-Ter. And. 418 uxorem ducas volo-3

), or a request-Plaut. Merc. 992 modo pacem faciatis 01'0-, a warning, advice, admonition etc.-Lucil. 594 submittas alios censeo-, the idea of striving or seeing to-Cato, Agr. 73 bibant curato "let them drink! see to it!"; Plaut. Asin. 755 scribas, vide-, that of decision, resolution-Plaut. M.G. 729 statuit ... ut veneat; cf. also Caso 448 certumst hunc ... praemittam4 ), or other connotations. It is the context-and particularly the verb of the principal clause-which lends the idea of volition to the utterance, not the subjunctive 5). There is no fun~an:ental difference in construction between JBr. 1, 105 akámayanta .. . Jayama (see above) and uxorem ducas volo, between bdaxs-¡;' avCJ~'JI ... s'Xftá{}ro and Acc. 508 edicite ut omnes adhibeant, between Anc. Ind. hanta + subj. and opus eat or necease eat + the same mood, between other subjunctives and that in the Latin quid est quod metuas? "what is ~here for you to fead" > "why is it that you fead"6). In a similar way it lS the context, not the subj., which is "deliberative" in Plaut. Caso 1003 nulla causa est quin verberes; Poen.881 quid dubitas quin . .. faciat?; Rud. 1172 contineri quin complectar non queo. Nor is it advisable to call

1) Ernout-Thomas, 0.0.2, p.342. 2) Humbert, O. C., p. 229: " .•. seules méritent a proprement parler le nom de

finales, les propositions qui oomportent un subjonotif de volonté (ou l'optatif so~ substitute)"; Sohwyzer-Debrunner, 0.0., II, p.672 "Ausgangspunkt und Trager der finalen Bedeutung war überalI der voluntative Konjunktiv ... ", and p.311 "Prospektiv war von Haus aus überwiegend auoh der Konj. in Neben­satzen .•. ".

a) There is no reason to oonsider with Bennett (0.0., p.215 and 249), who agrees with Delbrüok in his reoognition of a wishinO' optative this oonstruotion to oontain an original optative: of. also in Greek, Soph. El. 80 do.et, p,6Ívwp,sv; Soph. Ph. 761 flOVAet Aáflwp,at; "would you have me take hold?"

4) There is, as far as I am able to see, no use in assuming, with Bennett, 0.0.,

p. 236, a speoial subj. of determined resolution, whioh then might have extended lts.el~ .fr~m the first person t? ~thers: in Plaut. Poen. 501 ... habeam, deeretumst m~h~ lt 18 the verb of the prmOlpal olause whioh expresses determination.

5) Needles ~o say that many oombinations of anteoedent verbs and subjunotives are due to the influenoe of analogy. Yet those who adopt a theory of jussive volitive or other 8peoial origin (see e.g. Durham, 0.0., p. 37) are more often ~der the ne~essity of resorting to this expedient than those who would rejeot it.

) Cf. Durham, 0.0., p. 77.

The Subjunotive 105

the subj. mood jussive~in the so-caUed stipulative ut-clauses 1): Plaut. Bacch. 875 ("do you ;ant to have so much money") ... ut tibi mala multa ingeram?; here as well as Curc. 660 the subj. by itself originally denoted a suggestion, conveyed an idea, 01' created a mental image; the character of a clause such as tu ut hodie cenam des may be considered jussive, but Bacch. 875 and Men. 217 can, 1 suppose, more convincingly be explained from a subj. of 'ideal existence' than through Bennett's hypothesis that all stipulative clauses originally contained a 'jussive'. Any discussion of the 'nature' of the phrase dignus quoi concrederet "worthy of his confidence" (Plaut. Asin. 80), of d. qui laudetur, in short, of the subj. phrases corresponding to the construction a~LO~ (}avftáaat 2)

resolves itself into an examination of the origin of the sentence in its entirety and the 'originar logical connections between the two clauses, not in that of the subj. alone.

An observation remains to be made on the phenomenon of the so­called modal attraction 01' assimilation. It would appear to me that real instances of assimilation in the proper sense of the term-i.e. mechanical assimilation, e. g. Plaut. M. G. 149 faciemus ut quod viderit (instead of vidit) ne viderit-are much more infrequent than many scholars have been inclined to believe 3). Many cases of this 'attraction' must on reflection rather be regarded as representing an opto or subj. in their own right. In the second verb of E 107 'JIV'JI CJ' s'lr¡ 8~ Ti]aCJé y' aftsí'JIo'JIa fti)n'JI s'JItanot is a potential: "if there were such a man, he could speak ... " . Although it is true that the whole utterance is so to say marked as contingent, it cannot be maintained that the second opto is mechanically conditioned by the first. So 1 fear that 1 cannot agree with those who would consider TS. 2, 6, 6, 2 an instance of this attraction 4): varar!1,l v!z¡,ai2 : yad3 eva4 ••• bahil;paridhi5 skandát6, tan7 mes bMáWz¡,á'f(L9 bhaga­dheyam10 asadll iti12 "let me choose2 a boon1 ; whatever3,4 ••• faUs6 outside the enclosing-sticks5, let that7 bell the share10 of mys brothers9 " 12.

Starting from the assumption that the subj. expresses 'ideal existence' both verb forms can be accounted for without supposing that they are dependent on each other. But, owing to their position in a subordinate 01' a principal clause, to the meaning of the verbs, intonation and other factors, yat ... skandát assumes a value translatable by "whatever falls" and tad asad that corresponding to our "let that be". In the following sentence, which imparts to the reader knowledge about the facts existing

1) Thus Bennett, Trans. Amer. Phil. Ass., 31, p. 223ff.; Syntax, I, p. 263ff.; (Leumann-)Hofmann, 0.0., p.761.

2) See e.g. Bennett, Syntax, I, p. 260f.; W. Kroll, in the Glotta, 7, p. 141. S) Cf.-e.g. also S.chwyzer-Debrunner, 0.0., n, p. 330 as oompared with Kühner­

Gerth, I, p. 255ff.; and Ernout-Thomas, Syntaxe latine2, p. 402ff. 4) Renou, Déo., p.29 uses the term "harmonie verbale" referring to Minard,

0.0., p.20, § 44 who speaks of "attraction".

Page 57: Gonda.1956.The Character of the Indo-European Moods.pdf

106 The Indo-European moods

in the present after this wish has been fulfilled, the indicative is the appropriate verb form 1).

From the comment made, in the preceding chapter, on the compound sentence with two optatives-type SB. 13, 8,4, ll, see further on-it may appear that 1 cannot adopt the view that one of the optatives has mechanically and by mere 'thoughtlessness' conditioned the appearance oí the other. This view seems to have arisen from the time-honoured assumption that, given the different character of the optatives under consideration and given the fact that as a rule only one oí them has one ofthose functions which are regarded as basic (cupitive or potential), the other must have been secondary. Thus a sentence such as SatBr. 13, 8, 4, II sal yatha2 kamayeta3 tatha4 kuryat5 "as2 he1 desires3 thus4 he should do5 " is, tacitly or explicitly, assumed to have arisen from sa yatha kamayate tatha kuryat 2 ). Both verb forms can, as we have seen, be more satisfactorily explained directly from the ideal and contingent character of both processes, or if one would express oneself in this way, from the visualization and contingency inherent in the whole utterance 3).

Cf. e.g. SatBr. ll, 4, 2, 7 "if one desires (opt.) it to be (opt.) ... one should (opt.) ... " as compared with 14,4,3,26 "when one who knows this departs (indic.) from this world, he enters (indic.) into his son": the whole utterance is a statement; 8, 1, 2, 1 "when the sun rises (indic.), the whole space comes into existence (indic.)", but ll, 1,4,4 "every time the moon does not rise (opt.), he should fast (opt.)". SatBr. 9, 1,2,12

1) It must be conceded that part of this disagreement may be due to a different interpretation of the term "attraction" which is less harmless than "harmony".­See also Minard, o.c., p.20, § 44, who after correctly observing that beside the construction with a future in both clauses "une subordonnée qui appelle ou tolere nuance d'éventuel passe au subjonctif" appears in connection with an impero in the main clause, concludes that "aux modes et temps du possible ... l'attraction joue aisément qu'exerce la principale"; 1 would repeat my view that the subj. does not express what is possible 01' "éventuel" and observe that in the sentences with two future forms such as BatBr. 13,6,2,12 (Minard, o.c., § 677) the future of the principal clause is clearly more emphatic, referring to "clear-cut" future events; here the future ofthe protasis introduced by yadi may, at least 'originally', be due to 'attraction' proper: yadi1 sar¡wthápayiwasi2 puru~a3 eva4 pur~am5 atsyati6ti7 "ifl you put (them) to death2, man3 wiIl eat6 man6"7, although 1 would prefer to ascribe sar¡wthápayi~yasi to the attitude of the speaker in pronouncing the whole sentence rather than to mechanical 'attraction'. As to BatBr. 6, 5, 2, 22 (see Minard, o.c., § 680) 1 would interpret yadi + the future bhetsyate as follows: "put the case that at a given moment in the future one (of them) goes to pieces", and bhar~yámaJ:¡, as "we shall ... ". We can of course say that the whole utterance is viewed as future actuality. As compared with a subj. construction this double future impresses us as more vivid and unambiguous, also, under other circum­stances, as somewhat more peremptory 01' straightforward.

2) Of. Minard, o.c., p. 18ff. 3) It may be important to recall that in Latin constructions of the type Plaut.

M. G. 736 qui deontm consilia culpet, stultus ... sit the hypothetical tone is likewise maintained through the whole sentence (see also Morris, Am. J. Phil., 18, p. 153).

The Subjunctive 107

is especially instructive: "if it is not broken (opt.), one should order it to be broken (opt.); for only when it is broken (ind.), the pain attacks (ind.) his enemy"; the former sentence is contingent, the opto bruyat having in this context the value of our "he should say", the latter sen­tence is viewed as actuality.

This is not to deny that the opto construction was apt to be mechani­cally reproduced and to turn up in those cases in which the subordinate clause was not contingento Thus SatBr. ll, 5, 1, 4 kcttha'f(i1 tu2 tad3

aviram4 ••• syad5 yatra6ha'f(i7 syam8? was correctly translated: "wie sollte es da keine Manner geben ... wo ich bin 1" 1); here the opto syam may be due to the phenomenon of 'perseveration' 2). So much for details; now to recapitulate: the harmony of the verb forms in both parts of a compound sentence 3) is, as it would appear to me, mainly due to homo­geneity in the speaker's view of the relation between (the process con­tained in) the utterance as a whole and reality, not to 'thoughtlessness', mechanical, and secondary developments. It goes, on the other hand without saying that these constructions were, in the idiom of the SatBr., traditionaI4).

An examination of those verbal categories in non-l.-E. languages which roughly speaking correspond to the l.-E. moods shows that one form class occurring at the same time in various functions fulfilled by the l.-E. subjunctive is no exeptional phenomenon. In the Southern

1) (R. Pischel-)K. Geldner, Vedische Studien, 1, Stuttgart 1889, p.245. 2) See R. J. A. Lagas, Syntactische perseveratie- en anticipatieverschijnselen

bij oudere Griekse dichters, Thesis Nijmegen 1941; H. Oertel, Über grammatische Perseverationserscheinungen, IF. 31, p. 49ff.; Havers, Handbuch, p. 72.-1 would prefer to explain the passage BatBr. 3, 9, 3, 32 discussed by Minard o.c., p. 190 (§ 690; cf. p. 19f., § 43) as follows: both yadi clauses contain an autonomous opto "put the (ideal and contingent) case ... ", in the first apodosis the author remains in the contingent sphere, in the second he turns to actuality. Similarly 3, 5,3,9 dealt with by Minard in § 698. Ibid. 4,5,2,1 (§ 688) the apodosis is a 'rhe­torical' question, in which the opto is the usual mood (Delbrück, O. c., p. 336); the conditions are however viewed as 'real'.

3) See also Minard, O. c., § 44. 4) Hale, Am. J. Phil., 8, p. 54, while generally rejecting the attraction theory,

overshoots the marle in assuming that the Latin mittat quem veUt contains two jussives, "because the modal feeling in the speaker's mind which expresses it­self in the main sentence is, in the nature of things, very likely to continue in the speaker's mind in the subordinated sentence ... " .-In Plaut. M. G. 1229 oro ... quod cupiam ne gravetur, the subj. cupiam would appear to be a normal case of the subj. generalization and ideal ex.istence rather than an instance of formal attraction (otherwise e.g. Bennett, O.C., 1, p. 307): " ... not grudge me my desire" (Nixon); cf. also Bacch. 653.-

In Latin, the subordinate clauses in the indefinite 2d sing. stand, as a rule, in the subj., the Ol;igin of which does not seem to be so obscure as was assumed by Bennett (o.c. 1, p. 320): Ter. Phorm. 265 unum quom noris omnis noris. - Space forbids to discuss here the factors which have contributed to the establishment, in Latin, of the subj. in those constructions where it was not original.

Page 58: Gonda.1956.The Character of the Indo-European Moods.pdf

108 The Indo-European moods

Bantu languages 1) the most common uses of the 'subjunctive' are, the

hortative 01' permissive, in clauses of purpose, in consecutive construc­tions, and after certain deficient verbs. In Nguni woza ubone means "com th t "." d" e a you may see 1. e. come an see. A sentence meaning "let-know the man that he-come" conveys the idea "tell the man to come". Sorne Indonesian languages have a formation which may fairly well be compared to the aboye subjunctive. A special formation exists for instance, in Old Javanese, and Bontok (Philippines), the forme~ using -a, the other idiom -(e)d. An analysis of texts shows inter alia the following uses of the Old-Jav. -a form 2): Bh Y. 1, 8 kino~1 lumakwa~ ijatenen (i.e. ijatena3 in4

) K.5 "they called upon him1 to g02 (a-form) towards~, 4 (prop~rl! speaking. an -a form of ijaten «approaching») K.5"; 9, 2 kad~1 mubura2n3 bala" san5 K.6 "as if1 he would destroy2 the3 army" ?~ the5 K.6"; 19,13 harep1 manugele (i.e. manugela2-i3) tengek4 san5 A.6

mtending1 to cut2,3 the throat4 of the5 A.6"; 9, 4 tambis1 meh2 sira3 matya4 "he3 carne near1, 2 being killed4 " (mati "die, dying"); Arj. Wiw. 6, 8 tulusakena1 tikun2 prih3 meh4 • • • ijatena5 "persist (-a form' in an «adhortative» function) in your2 endeavour3 till4 ••• she co~es5"; BY. 36, 14 sukii,1 matyapran (i. e. matia2 alJran3) .. , "1 would prefer1 (-~ form: suka-a) to be killed2 in battle3 (rather than ... )"; Arj. Wiw. 18,3 p~ra1 har~a2 sañ3 prabhu4 rumenwa5 denta6 mawuwus7 "how1 glad2 (will) the3 king4 (be) (if) he hears5 (heard5) you6 speaking7 ••• (your6 words7)".

The phrase wineh1 sira2 mastrya3 (i.e. ma-stri-a) means "they let him permitted1 him2 to take a wife3 "; ya1 ta2 pajarakena3 ri4 kami5 : "that1, ~ must be narrated3 t04 us5"; ucapa1 ta2 maMriija3 P.4 "now2 to speak Of1 king3 PY'; san1 pejaha2 "he-wh01 is to die2". an1 kapunguha2 can mean: "if1 ... be attained2"; yadyapi1 katona2 ~nswers to our "even though1 ... is seen2

". The -a form occurs after verbs of doubt and hesitation, after those expressing a command, a request, an intention, and also in sentences expressing .a reserved utterance (in the sense of "may" 01' "(to be) probably". Whereas this form is very frequent in Old-Javanese, a similar category is absent in those idioms which are linguistically and geographically, nearest to it 3). '

:) See C.~. Doke, The southern Bantu Languages, Oxford 1954, p. 73f. ) Many mstances were collected by R. Kern, Verspreide Geschriften, VIII,

The Rague 1918, p. 230ff. ~~rn describes this form as corresponding to a subj., an opt., ~ fut:! and a condltlOnal. For the 'future' see my papel': 'Tense in IN. ~an~ages '. BIJdragen Kon. Inst. no, The Rague 1954, p.240ff. It is true that n:dlvldual mstances of an -a form admit of a translation by an ancient l.-E. opta, tlve. C. C. Berg, Inleiding tot de studie van het Oud-Javaans, Surakarta 1928, p. 224ff. wrongly calls the -a form an "irrealis". . 8) In. part of its .functions the Tontemboan (N. Celebes) -um- runs parallel. An mte:;estmg feature lS, for instance, the occurrence of an -um- form after verbs of mot~~n: ma1ie~ kumu:z2 si8 [to'4 "gol and say 2 to uncle4 " (Dutch: "ga zeggen aan oom )~, ~mt ltS mam use ~nd.general character is different: see J. A. T. Schwarz, N. Adrlam, and M. L. AdrIam-Gunning, Roofdstukken uit de spraakkunst van

The Subjunctive 109

The concise description of the modern Georgian subjunctive given by Hans Vogt is worth quoting 1): "Les propositions subordonnées a vec rom ont le subjonctif si elles expriment le but, le désir, la peur, comme en fran<¡ais. Dans ces cas le subjonctif a lui seul suffit a exprimer la relation des propositions et la conjonction est souvent omise; surtout si le sujet est le meme dans les deux propositions" . " "Le subjonctif s'emploie dans les phrases relatives comme en fran9ais (e.g. «est-ce qu'on a jamais entendu parler d'un étudiant qui ne lise rien 1»)". The Arabic subjunctive, though much more limited in application than that of ancient Greek­it is practically a mode of subordinate clauses, mainly of those which are final in character-, has in some points a strong resemblance to the aboye forms 2). Mention may also be made of the syntax of a category such as the modern Hindi so-called "contingent future" 3): in simple sentences it expresses a permission, a possibility, a wish, a mild command, duty or propriety; in interrogations it helps to denote uncertainty or perplexity; in subordinate clauses it can express an intention, a pur­pose, a result (when not regarded as a reality), a concession; it is used in similes, denoting that with which the comparison is made, not as an objective reality, but merely as a supposed case; in clauses referring to an indefinite futurity; in the protasis of conditional clauses, when the condition is regarded as a mere possibility; in relative clauses implying a condition, when that condition is merely supposed to existo

Excursus on the form of the Vedic subjunctive

Before leaving the subject of this mood we must in continuation of our chapter on the injunctive come back to a few points of morphological interest. It has already been intimated that most scholars recognize a more 01' less original set of three modes, ind., subj., and opt., the injunc­tive being often regarded as a pro blem by itself. Yet the question has, now and then, not vainly been put as to whether there may have been inter­relations between the modes, whether the system of Vedic or ancient Greek modes was really original, whether the forms and the functions of ind., subj., and opto could, in a reasonably plausible way, be shown to have arisen from other forms 01' other functions. In addition to the

het Tontemboansch, The Rague 1908, p. 132ff.-Needless to say that such state­ments as: "the subj. can be expressed in Bimanese, but often the form cannot be distinguished from the indicative" (cf. aiso the subsequent expositions to be found in J. C. G. Jonker, Bimaneesche Spraakkunst, Batavia 1896, p.297) give a false impression of facts.

1) R. Vogt, Esquisse d'une grammaire du géorgien moderne, Oslo 1936, p. 272f. 2) See e.g. R. Reckendorf, Die syntaktischen VerhaItnisse des Arabischen,

Leiden 1898, p. 730ff. 8) See e.g. S. R. Kellogg, A Grammar of the Rindi Language8, London 1938,

p.455ff.

Page 59: Gonda.1956.The Character of the Indo-European Moods.pdf

110 The Indo-European moods

remarks already made on the foregoing passages it is now time to discuss sorne points in connection with the history and supposedly prehistoric development of the subjunctive.

Rejecting any theory which cannot be based on philological facts and distrusting any hypothesis which seems to be incompatible with our knowledge of the structure and development of semi-primitive langu­ages in general, attention here will mainly be focussed on the problems arising from the suffixes and endings by which the subjunctive forms are characterized. The 'orthodox' view is with sorne variation in detail in short the following: since the Orig. l.-E. period this mood is formed by the suffix -e- /-0- which may, 01' may not, be identical with the thematic vowelI), with which it combines to form -e-/-o- in the thematic, and by way of analogy also in many originally athematic, verbs. Although the secondary endings are not foreign to this mood, it is generally character­ized by the primary series. These fluctuations between both classes of endings are especially conspicuous in Indo-Aryan. With regard to this last point Renou has, in an important article 2), observed that, though it may be true that the Vedic subj. "reyoit tantót les désinences primaires, comme le présent; tantOt les secondaires, comme l'optatif", in the middle forms the only secondary ending is found in the 3rd plur., -anta indeed occurring more frequently than the primary -ante3). This detail and the spread of the ending -ai from the first sing., where it is normal, to other persons (2nd sing. -asai side by side with -ase) may, Renou argued, be regarded as due to the tendency of this mood to differentiate from the indico In the active the 1st duo and the 1st and 3rd pI. have, no doubt for the same reason, the secondary endings; the 2nd and 3rd dUo and the 2nd pI. have the primary endings only; the 2nd and 3rd sing. may take both forms, with a preference for the primary series. In this connection Renou also endeavoured to show that sorne distinctions between both classes of endings may be connected with a twofold character of those verbal forms which are commonly called subjunctives : those subjunctives which tend to be a simple device of subordination 01' a means of expressing general 01' undetermined thought have a preference for -si and -ti ("sub­jonctif éventuel", precursor of the "optative of mild statement"). The "modal" -i, e. final, consecutive, hypothetic, "prospective" and volitive -subjunctive, on the other hand, prefers secondary endings. Thus the very frequent asalp and asat (47 instances in the ~V.) always express

1) Compare e.g. the formulations chosen by Brugmann, K. V. G., p. 551 ("-e-/-o­ist unzweifelhaft der ... sogen. themat. Vokal"); Meillet, Introd.8

, p. 224; Meillet­Vendryes, Gramm. comp.2, § 298; Pisani, Glottol. Indeur.2, p.204; H. Krahe, Indogerm. Sprachw., Berlin 1943, p.119f. ("als Moduszeichen den mit dem Themavokal gleichlautenden ... Vokal e/o").-Cf. also Hirt, Indog. Forsch., 12, p. 213 and Indog. Gramm., IV, p. 294.

2) Renou, A propos du subjonctif védique, Bull. Soco Ling., 33 (1932), p. 5. 3) See Macdonell, Vedic Gr., p. 316. Curiously enough, Avestan has only -ante.

The Subjunctive III

one of these nuances: cf. e.g. 7,24,1; 8,100,2; 10,50,5; 1,89,5; 107, 11). Thus a subj. which follows an indic. tends to assume -ti 01' -si, a subj. which depends on an imper., -t 01' -s: l;{V. 10,85,25 prél t62

muñeámi3 ... yáthe4yám5 ••• subhágá6sati7 "1 releasel ,3 from here2 ... that4 she5 may be 7 happy6", but sto 26 grhánl gaeha2 grhápatni3 yáthá4salp5 "g02 t02 the house1, that4 you may be5 housewife3"; 1, 9, 52). Some value may, it seems to me, also be attached to ~V. 6, 23, 9 kuvítl tásmii2 ásati3

n04 bhá1'aya5 "gewiBl wird el' uns4 beistehen5 um diesen2 (Preis) davon­zutragen5" (after prrJata; followed by mrdhati); 2, 38, 3 asúbhisl eid2

yán3 ví4 mueati5 nünám6 "auch2 wer mit schnellenl (Rossen) fahrt3, soll jetzt6 ausspannen4,5" and similar instances which impress us as being more in the actual sphere than, e.g., 10, 95, 12 kadál sünúlp2 pitára'f!¿3 jatá4 iehat5 "whenl will the son2, born4 (of you), long for5 his father3 1"; 8, 24, 30 yátl tva2 preMt3 ... "whenl somebody will (might) ask3 you2 ••• ";

10, 108, 5 kás1 ta2 ena3 áva4 srjad5 áyudhvi6 "werl mochte die3 <dir2)

ohne Kampf6 herausgeben4, 51" (Geldner); 10,51, 7 yáthal yukt62 ... ná3 1'í?yalp4 "lestl, 3 you, being yoked2, be injured4 " etc., which may rather be said to reflect "mental reality".

In the aorist and also in the perfect, where no primary indic. exists, the secondary endings are in a large majority. Moreover the number of the -ti and -si forms increases in the course of time, the younger (01' more popular) parts ofthe l;{V. and the Ath V. containing a larger number of them than those parts of the former corpus which are usually regarded as ancient and may at least have originated and transmitted in more conservative milieus. In the later Vedic texts however, the primary endings tend to disappear and the so-called modal force of the mood is reinforced.

From those facts the conclusion may at first sight be drawn that it is the secondary endings which properly belong to the subj., the other class appearing under the influence of the indic. especially in those cases in which the syntactic value ofthe subj. approaches that of an indicative 3 ).

The use of the subj. in the Avestan Gáthás seems to agree with this view of the Vedic forms under discussion. Most of the subj. forms with secondary endings occur in clauses 01' sentences the verb of which more 01' less clearly refers to processes in the mental 01' ideal sphere 4) in 'modal' (final, e. g. Gá. 30, 4; 44, 1; consecutive: 30, 7 beabsichtigte Folge (Reichelt); conditional: 43, 4: mental reality) clauses; in clauses dependent on modal clauses and referring to processes which though not (yet) actual, are to be taken into account, may be expected etc.; 43,12

1) This is however not to contend that asasi, asati are always devoid of this nuance: in 2, 26, 2; 4, 57, 6; 10, 174, 3 they are completely subjunctive in funotion.

2) Cf. also Bloomfield-Edgerton, Vedic variants, I, p.80. 3) I cannot agree with the view expressed by Miss Hahn, o.c., p. 50, n.91. 4) See Kurylowicz, Roczn. Orjent., 3, esp. p. 176ff.

J

Page 60: Gonda.1956.The Character of the Indo-European Moods.pdf

112 The lndo-European moods

e. a. the clause depends on a command: uzir'didyail para2 hyat3 moi4. a.jimat 5 s'draoso6 "arisel before2, 3 my4 0bedience6 comes (shall come)5"; or on an interrogation: 29, 7 in Duchesne's translation: "Qui as-tu en tant que Bonne Pensée, qui prenne soín de nous deux ... 1"; in orders or commands, or when a desire, hope, expectation is couched in the form of an incitement: 29,4; 31,20; 47,1; in partial and subordinate inter­rogations (the person asking the question is interested so as to hope, to fear, to expect: 48, 2 katl asava2 • •• v'dnghat3 dr'dgvant'dm4 "will the righteous2, as it may be hoped, gain the victory over3 the wicked4 1l"); in general relative clauses of a conditional character (e.g. 31,20 "who (ever) comes to the righteous ... "). The forms with primary endings generally speaking occur in uncomplicated affirmative sentences; in interrogative sentences other than the above; in non-<modal' relative clauses; in temporal clauses; in gen. rel. clauses if the principal clause has a present indicative: cf. e. g. 44, 6 aS'dml syao-&anáis2 d'dbftzai ti3 .

armai tis4 "shallA.4 by her activities2 succour3 A.lf' Although the differ­ence in function is never very great, the distribution of the two cate­gories over the sentence types is remarkable. Kuryiowicz proposed to regard the second category as possessing the force of a simple future, the first as expressing "une action éventuelle soit un désir dont l'ac­complissement n'est possible que dans l'avenir". Would it not be pre­ferable to hold that the second, broadly speaking, refers to processes the realization of which is definitely taken into account, the first to processes which exist in the speaker's imagination, fear, hope, expecta­tion, which are the object of his curiosity1 May this distinction be re­garded as underlying the opposition 31, 5 yal noit2 va3 at:}hat4 at:}hai ti5 vá6

"whatl will4 not2 be4 or3, 6 will5 be51" Or is this question futile1 Shall we err greatly if we connect these functional variations and the

aboYe impression that the appel:j.rance of primary endings may be due to the influence exerted by the indicative ("Annaherung") with a fact revealed by a comparative examination of the ancient Greek data 1 Although scholars have differed in opinion with regard to Are. Cypr. Boeot. El. Dor. cpser¡, which is often regarded as etymologically identical with Ved. bhá?'át l ), there must be some value attached to the compara­tively conservative character of the dialects in which this type has been preserved, the more <modern' Ionic and Attic dialects possessing only cpser¡u;, cpser¡~. As, moreover, the explication upheld by Schwyzer-De­brunner: "cpser¡u; cpser¡~ aus *-er¡c; -er¡ durch Angleichung an cpSe8~c; cpsed '

1) See e.g. Brugmann, in H. Osthoffund K. Brugmann, Morpho1. Untersuchun­gen, l, Leipzig 1878, p. 182f.; Thumb, Handb. der griech. Dialekte, Heidelberg 1909, p. 277; Schwyzer-Debrunner, o.c., l, p. 661. Other scholars, denying that there is sufficient evidence for the survival of -1) from *-1)., suppose the 'dialect forros' in -1) to representthe usual-v (C. D.Buck, Comp. Gr. ofGr. andLat.3

, Chicago 1942, p. 299).

The Subjunctive 113

is no doubt correct, there is in principIe no objection to the supposition that also in Greek the subj. once had, whether or not beside the primary series, secondary endings, which were ousted by the former l).

Proceeding then to discuss the suffix by which this mood is charac­terized and identifying this without further comment with the thematic vowel, Renou 2) argues that "le subj. brávate est par rapport au présent brüté ce que !'indico stávate est par rapport a stute"; and, in addition to this that it is especially this "indétermination morphologique" which goes with the "indétermination sémantique" which has been signalized by many authors 3): that is to say, it is a form like lcarati which may be taken, and translated, as an indico as well as a subj., not krt}oti or krt}avat 4 ). So Renou is led to the conclusion that in the ancient Vedic literature a survival of an ancient subj. derived from an independent theme (such as kar- as compared with kr-, krt}o- etc.) must be recognized which was characterized by the thematic vowel and nothing else; "l'aspect et l'évolution meme de ce subj. indépendant serviront ensuite a suggérer la fayon dont la voyelle thématique s'est développée et fixée, tantOt s'orientant vers le sens indicatif, tantOt se précisant vers le sens modal". Now a weak point in Renou's argumentation seems to be that the indistinct and very frequent theme kara- of the present-aorist system of kr- "tú make, do, cause", which in a large majority of cases serves as a "subjonctif éventuel", uses in this very function the primary endings-which we expect--as well as the secondary. It seems also do~btful whether we should regard a form of this type as "hybrid" 5) : it may impress us as such on the strength of the traditional distinction between an athematic form karoti etc. and thematic verbs such as bharati, there is no conclusive reason to assume a mixed origin before the reverse has proved unacceptable. The question may further arise as to whether a form of the type karat cannot be considered an injunctive, and a form like karati cannot be classified under the so-called modal indicatives which have been discussed in another chapter ofthis volume6).

Commencing this argument with the latter point, serious doubt may be expressed as to whether the <interrogative-dubitative' character of karati in lines such as l;tV. 2, 35, 1 kuvítl sá2 supésasas3 karati4 "peut-

1) For subjunctives with seco endings in other languages see e.g. Brugmann, Grundri132, n, 3, p. 587; J. Schmidt, Kuhn's Zs., 19, p. 290f.; H. Jacobsohn, ibid. 45, p.342.

2) Renou, o.c., p. 13ff. 3) Cf. e.g. W. Neisser, in Bezzenb. Beitr., 7, p. 211ff.; Bloomfield, in the Amer.

Journ. of Phil., 33, p. lff.; Bloomfield and Edgerton, o.c., l, p. 53ff. 4) One might compare the relations between the fut. indico and the aor. subj.

in Greek. Cf. also V. Magnien, Le futur grec, Paris 1912, n, p.144ff. ó) Renou, o.c.,'p. 18,1. 4; cf.1.lf. "il s'agit a l'origine d'un theme étranger a

toute catégorie" : understand: " ... catégorie historique ou traditionelle" (see p. 28). 6) See aboye, p. 44.

8 Gonda, Indo-European moods-

Page 61: Gonda.1956.The Character of the Indo-European Moods.pdf

114 The Indo-European moods

etre1 (ApaIp. Napat) rendra4-t-il2 (mes chants) richement ornés31" must prove the verb to be a subjunctive. Dubitative interrogations are often expressed by sentences containing an indicative. After kuvit this mode occurs e. g. ~V. 5, 3, 10 kuvíd1 ... sumnám2 agnír·3 vanate4 • •• "wilP Agni3 (wird Agni wohl) win4 the favour2 ... "1); cf. also 10,131,2; SatBr. 4, 6, 6, 5. There may however have been a slight difference between one kuvit sentence and the other, both of them containing an indicative. We should not forget that the indicative does not record objective reality; it only indicates that the person speaking views the process as actual. In doing so he may within the bounds of what was common usage in his milieu, follow his own inclinations. It may there­fore be an hazardous undertaking to attempt, with Neisser 2), to establish which form with the -a- vowel is indicatively, which is subjunctively used. In 1;tV. 5, 86, 1 yám1 ávatha2 ••• mártyam3 I sá4 prá5 bhedati6

dyumná7 a subj. may, it is true, be expected in the second clause: "wel­chen1 Sterblichen3 ihr ... begünstiget2, der4 wird die ... Herrlichkeiten erbrechen5, 6" (Geldner). But why could the poet not add a slight tinge of subjective conviction and certainty to his statement so as to use a full-fledged indicative1 In such passages as 10, 96, 1 a subj. would, 1 think, be even less common: ... mádam1 ••• y62 ••• sécate3 "the intoxi­cating drink1 (soma) which2 is poured out3 ".

There is further also sorne difficulty in assigning such verb forms as jújot}at(i) , which derive from the perfect stem 3), to the same class as normal subjunctives, since they lack a marked and constant subjunctive character; their force is nearer to an indicative4). Curiously enough, there exist also several 'irregular' perfect imperatives and sorne pluperfects with an el- vowel (types mumócatam and acikitat)5). It must be conceded that regular 'thematized' forms derived from the weak perfect root, but the hypothesis of an uncommon 'thematization' 01', more cautiously, of a formation with -a- which had begun but was not canied out (Whitney) is in itself not more improbable than the supposition that the subj. suffix -a- had spread ayer imperatives and pluperfect forms.

So there may be sorne room for the consideration whether Renou's view of these forms could not be modified in a few details. Let us first turn to a discussion of the category to which they belong. It would appear that this "eventualis", "subjunctive", 01' "pré-subjonctif autonome" has as far as the forms with secondary endings are concerned, special connec-

1) Here as well as 2,35,1 Geldner translated: "gewill wird ... " (affirmative sentenoe).

2) Neisser, 0.0. Cf. also Renou, 0.0., p.24f. 3) Maodonell, Vedio Gr., p. 360. 4) See Renou, 0.0., p.26f.; Whitney, Sanskrit Grammar, § 815; P. Thieme,

Das Plusquamperfektum im Veda, G6ttingen 1929, p.27. 5) Whitney, 0.0., §§ 814; 820.

The Subjunotive 115

tions with the injunctive, firstly because of their vague and undetermined function, which has ah'eady been commented upon, and in the second place because of a considerable similarity in syntax and phraseology. Thus we find, for instance, beside ~V. 1, 186, 5 utá no ... máyas kaJ:¿ (inj.) and 1, 89, 3 sárasvatí naJ:¿ ... máyas !carat: both passages occur in a text addressed to "all the gods", in both of them a particular deity is implored to give delight to those on behalf of whom the poet is speaking (cf. also: 5,46,4; 8,18,7 etc.); 3,54,14 vít}'(iU'f(i st6masaJ:¿ ... gman; 9,40,2 gámad índra'f(i ... sutáJ:¿; 2,38,11 á gat: 5,42,7 agámat; etc. As already noticed by Renou 1) 7cámJ:¿, kárat, !cáran sometimes serve as regular injunctives: 1;tV. 5, 56, 7 and 8, 2, 20 we even find má ... !celrat: cf. 4, 18, 1 etc. ma ... kaJ:¿; it may be used instead of a nanative past, alluding to mythological events 2): 6, 18, 14 ká1'01 yátm2 váriv03 badhitáya4

"in which2 Thou (Indra) procured1 room (free scope)3 for the oppressed one4 ". Similarly yamat 5, 34, 2 must be an inj.: yád1 ím2 mrgáya3 hán­tave4 ••• vadhá'fJ~5 yámat6 "when1 he offered6 (him) the weapon5 in order to kill4 the aniníaP" 3).

It seems therefore reasonable to infer that the forms under discussion had something to do with the injunctive. Would it be a legitimate supposition that they originated in injunctives extended by an -a-1 If we are right in considering the injunctive a more 'original' category than the indicative, this supposition does not involve the existence of a parallel series of indicative forms beside kamt, yamat etc. The longer form of the type !camt could have been introduced after the model of the 1st pers. sing. karam and 3rd pers. plur. !caran (-am < *-'f(i, and -an <

j *-z~t respectively, hence kams instead of *kars etc.) and preferred because it was more distinctive than the ambiguous lcar, lcaJ:¿ < *kars and *kart. Whether such analogous processes, which cannot account for forms like loop,ut, 01' Lat. erit 4), were the main source of !caras, lcarat must be left undecided. There can however be no reasonable doubt that these forms, whatever their origin, were, 01' at least carne to be, used as normal injunctives if there existed indicative forms of the same stem: vahat: vahati, avahat, and as subjunctives 5), if there existed not only indicatives

1) Renou, 0.0., p. 17. 2) See A. Bergaigne, Religion védique, n, Paris 1881, p. 234; 276; Debrunner,

Festsohrift Thomas (New Ind. Ant.), Bombay 1939, p.7lf. 3) Aocording to Macdonell, this form is a preso inj. (Vedic Gr., p. 323) 01' a subj.

of a root aorist (Ved. Gr. f. stud., p. 409). 4) I refer to Brugmann, Grundrif;l2, n, 3, p. 524ff.; Schwyzer-Debrunner, O. C.,

I, p. 66f. and 790f. - The number of ancient 'short-vowel subjunctives' is rather limited; some forms, e.g. Ved. hanat, Av. fanal are only found in Indo-Iranian. See also Hirt, Indpg. Gramm., IV, p. 295 and especially W. Schulze, Kuhn's Zs., 33, p.134.

5) Cf. Renou, O. C., p. 28, n. 2 "Il est a noter ... que l'injonotif en fonction 'modale' a d'ordinaire la voyelle thématique".

s*

Page 62: Gonda.1956.The Character of the Indo-European Moods.pdf

116 The Indo-European moods / The Subjunctive

but also a short injunctive of the same stem: sparat: aspar [s], spar[ s JI). In these cases a plurality of co-ordinated forms increased the latent distinctive value of the -a- 2). Part of these forms could also adopt primary endings: this meant that they attained similarity, in form as well as in function, to the indicative. There is no reason to reject the supposition that particular present indicatives in -a- owe their origin to such -a-ti forms 3), but 1 do not see how we could substantiate a theory to explain for instance the numerous ancient thematic forms4 ) like l.-E. Mero: Gr. cpsew, Skt. bharami, OChSlav. be~'eti etc. from these sub­junctives or "pre-subjunctives". It would, in this connection, not be superfluous to state that most of the -at 01' -ati forms under consideration are typical of the ancient Vedic literature. That they did not survive shows that they did not participate in the vitality of the thematic classes. Lilce their relatives, injunctives and athematic subjunctives, they disappeared, no doubt because or their lack of formal and functional distinction and independence 5).

1) Cf. also the Avestan cara1 Y. 46, 4 (gath.), which functions as a subj. 2) It may however be remembered that it was the more distinctive -e-I-ó- forms

which, partly in the historical period, took over the task of conveying the sub. junctive idea: see e.g. Schwyzer-Debrunner, o.c., 1, p.791; Reichelt, Awest. Elem., § 244; Brugmann, o.c., p.532.

3) The opposite view is again defended by Hirt, Indog. Gr., VII, p. 148 (cf. also Delbrück, Vergl. Synt., II, p.349).

4) Part of the large number of identical forms (see Brugmann, o.c., p. 117ff.) may however be due to a parallel development.

6) For yamat: yachati etc. see Renou, o.c., p.20f.

1" I

B_

VII. Additional criticism of Miss Hahn's views 01 the optative and subjunctive moods

l It is now time more syste~atically to deal with Miss Hahn's views of

the optative and subjunctive. While considering the optative a future used "for more remote futurity (or potentiality)" 1) she must admit that "from potentiality to wish the step is less clear". The main difference between subjunctive and optative consists in the degree of vividness, the subjunctive tending to be used for more vivid futurity 2). Both theses seem to be open to controversy. The difference in vividness is, as far as 1 am aware, an accidental circumstance, a natural consequence of what 1 would regard as the main characteristic of these moods. In E 308 (n:~(]opat, 1í "8 cpserJC1t psya "eÚ'¡;O(;, 1í "8 cp8eotpr¡v "the unwelcome contin­gency, as the more likely, has the subjunctive, and the welcome one as the less likely, has the optative (cf. b 692)" and this peculiarity, Mis~ Hahn holds, is reducible to the aboye difference in vividness. In my opinion Hector, who is speaking here, express es by the subjunctive what he, in reality or pretendedly, expects, by the optative what he deems possible. Most of us would, under similar circumstances, resort to subtle­ties of the same kind. With this reserve it may be said that the "difference in mood reflects the relative degree of probability", that is to sayas visualized by the person speaking 3). Miss Hahn is of course perfectly right in contending that there "is really a bit of most subtle psychology" in the combination of optative and subjunctive in Q 653ff. 4); 'Te 383ff. etc.; as regards 'Te 389ff., there again the subjunctive expresses what the speaker expects (in casu: proposes), the optative what in his view might happen 5): pi¡ oí Xe~pa7:a .•• lJbwp8v I ... f¡ bé ,,' lJ'Te8tTa yf¡patTO .•. In

1) Hahn, o.c., p. 139, § 188. 2) Hahn, o.c., § 113. 3) Miss JIahn is in my view right in combating the opinion pronounced by

D. B. Monro; A Grammar of the Homeric Dialect2, Oxford 1891, p. 252f.: "where two alternatIves are not expressed by the same mood, the subj. gives the alter­native on which the stress is laid."

4) Cf. also C. Mutzbauer, Die Grundbed. d: Konj. u. Opto im Griech., Leipzig­Berlín 1908, p. 14.

ó) Th.e . n~gative statement made by F. Sommer, Indog. Forsch., 45, p. 60f. (the subJ. lS, m general, "mit einem im wortlichsten Sinn aktuellen Moment behaftet gegenüber dem .weniger personlichen-energetischen Optativ" in which appears "zum mindesten ein geringerer Grad des Begehrens"), who for the rest subscribes to Delbrück's opinion, caJ?not convince me: this aspect of the opto is, again, secondary.

.. 11

Page 63: Gonda.1956.The Character of the Indo-European Moods.pdf

I

118 The Indo-European moods

JI 646ff. it is a matter of secondary importance that "of the two alter­natives in his (Zeus') mind, the one uppermost is that this killer ... of his beloved son S. shall himself suffer death" 1): the main point seems to be that if the former alternative is chosen Hector is (with certainty) expected to kill P.; in case of the latter, P. might (was supposed, wanted to) continue to cause vexations to men (cf. X 166ff., where the alter­natives are on a par: two subjunctives 2)). Although 1 can, in the main, subscribe to the interpretation given of cp 193f., 1 again fail to see how the shift in moods is in line with a difference in vividness. A distinction between an action based on one's own initiative (subj.) and an action depending on some one else's (opt.) may indeed be noted in S 161ff.: 163 s'i nw~ l¡tst(!at'to expresses contingency (nw~ "in any way, at all, it might be supposed"), 165 XsV?1 refers to actuality visualized and, in casu, proposed. Cf. s 470ff.: si M "S'V ... "a'tao(!á1fw, s'i ¡ts ¡ts1fstr¡ ei:yo~ ... 013 ¡tOt {fn'Vo~ enéJ.1fot: going to rest (subj.), falling asleep (opt.) 3). Hence, 1 think, the impression of a "distinction between more or less vividness of futurity (which) can come to correspond to a distinction between will and wish"4). 1 decidedly disagree with Miss Hahn with regard to (J 30f.: 42f. 5). Aegyptius uses the opto s'inot, because in his view Telemachus might (possibly!) relate tidings of the army, whereas T. himself stating the fact that he had not heard any such tidings which he could relate, says: s'inw (subj.: if he had tidings he certainly would relate them) 6). It might also be doubted whether the possibility of co-ordinating both the subj. and the opto with an imperative is a proof of their similarity to each other 7).

It would also appear to me that in arguing that there often is no essential difference between the subjunctive and the optative, Miss Hahn not infrequently overlooks subtle differences of meaning. Are, for in-

1) Hahn, p. 86, § 114. 2) Cf. also Mutzbauer, 0.0., p. 49; 203f. 3) 1 also refer to the instanoes adduoed by Kühner-Gerth, O. O., n, 2, p. 479f. and

Sohwyzer-Debrunner, 0.0., n, p. 686 at the bottom of the page. 4) Hahn, p. 87. Ó) Hahn, p. 87, n. 199: "The reason is doubtless metrical convenience, but none

the less the shift could not be made if the two forms were not fairly equivalent ín sense."

6) With regard to r 71ft'. : 92ft'. it might be observed that the assumption of a contingent character of the opto in 74 (whioh is in the 2nd person; cf. also 255) sooms to oope with all difficulties; the subj. in 93 (1st pers.!) being 'propositive' 01'

'adhortative' is no mere substitute, illustrative of the interchangeability of the moods, but an indioation that the process referred to is, as far as the speaker is concerned, realizable, and that its realization is to be reckoned with. In repeating these words (255ft'.) the herald replaces an imperative oocurring in the original speeoh by the non-oommittal opto eno¿To-it may be interesting to notioe that he also prefers to make YVVT] "al "T'f¡p,aTa the subjeot-and retains the opto of the verb vato) 74 for similar reasons (va{o¿p,Bv 257).

7) Hahn, p. 88.

Critioism of Miss Hahn's views 119

sta~ce, the opto "CAatr¡'V and the forms xsvw, "'í8(!st~w, OWCfW in (J 218ff. eqmvalent, because they follow precisely the same sort of protasis?:

~ , , I !), I ';" -> ~\ , I St ¡ts'V "S'V na"C(!o~ ... VOCf"CO'V a"ovCfW, r¡ L a'V ... 8n "CJ.atr¡'V 8'Vtav"C6'V "if 1 shall hear (visualized as a fact) that my father ... is coming home ... 1 could (if other circumstances would not prevent me) endure for another year"; "(but if 1 shall hear that he is dead, then 1 will return and) heap up a mound for him (xsvw intention, actualityexpected), ... pay funeral rites (,,'tS(!st~w), and give (future actuality) my mother to a husband"l).

In dealing with the various ways of expressing 'unreality the American scholar inter alia comes to discuss the use of the optative to express, "in place of the past indicative", past unreal conditions 2). Rejecting the opinion maintained by Randiford 3) and other scholars that the optative was the original form in all unreal conditions, she defends the reverse thesis: the indicative with the 'modal' particle a'V, "S, "a is the earliest form, and-Homer's usage being confused-the optative once established in present unreal conditions was easily transferred to past unreal con­ditions. "It would be very hard to account for the use of the optative 4) there" 5). Refel'ring fol' the 'opto of the past' to our above argument and fo:- a detailed discussion of the 'unreal condition' to chaptel' IX, we would, wlth Schwyzer-Debrunnel' 6), maintain that ancient Gl'eek was also in this l'espect rich in possibilities expl'essing subtle shades of meaning; genel'ally speaking, we think it possible to account fol' the use of the modes in the sentences in question so as to be in keeping with theil' general functions.

It will be relevant to sul'vey fil'st those instances which are intel'­preted by Miss Rahn. In holding that the optative in cases such as P 70f. 7), E 311f. is "absolutely equivalent" to the indo (cf. E 679f.) she has not been happy: in P 70 b{}a "S ¡Jsi:a cpé(!ot "Av"((I 'tsvXea . .. I lh(!stor¡~, si ¡t~ ... "full easily A. would have borne off the glorious al'mour, if not ... ", in P 274 el ¡ts'V 'VV'V En¿ aAAq.> as{}J.svot¡ts'V l1xawt I i} r a'V EyW 'ta n(!ana Aa{Jw'V "J.tCftr¡'VOS CPS(!ot¡tr¡'V "if for sorne other's honour we Achaeans wel'e now holding contests, sul'ely it were 1 that should win the first prize, and beal' it to my hut" the pl'ocess indicated by the opto is not actual, but a mel'ely hypothetical event, and-this is the main point-viewed by the speaker as such. Cf. also JI 72; X 20 etc., whel'e the process is likewise not viewed as actual.

Specially interesting is {J 184, where the impf. indo and the preso opto are coupled: "(1 would that you had perished ... ,) then would you not

1) See Halm, p. 83f. 2) Hahn, p. 93ff. See also ohapter IX. 3) S. A. Handiford, The Latín Subjunotive, London 1947, p. 19. 4) 1'Vhioh, we lmow, was in her opinion a future. 5) Hahn, p. 96 and 31, n. 46. 6) Sohwyzer-DebrUlmer, 0.0., n, p.683ff. 7) See also Humbert, Syntaxe greoque2, p. 222.

-~

Page 64: Gonda.1956.The Character of the Indo-European Moods.pdf

120 The 1ndo-European moods

taIk so much (ovu áv 7:6aaa . . . ay6e8V8~: the person addressed would indeed have said in Dutch "dan had je niet zoveel gesproken") or be urging T. on in his wrath (ovdé U8 Tr¡U¡taxov U8XOAOJ¡tévov ¿M' aVt8ír¡~: this is a possibility, or even a probability foreseen by the speaker). Of. also d 546f. As to E 679f. uaí vv u' en nUova~ AvuíOJV W¡;áV8 Mo~ D., 8l ¡ti¡ ... "and O. would have slain yet more of the L., if not ... ": here O. has slain Lycians, and the poet in forming a mental picture of the course of events considers it a fact that the hero would have continued to do so. In translating we might insert a word like "doubtless", "verily" or even "certainly". In g 130 l!:vf}a U8 AOtyo~ 8r¡v ual a¡tl¡xava seya yévov7:o, I ... I 8l ¡ti¡ ... "then had ruin (certainly) come and deeds beyond remedy been wrought ... had not ... ", the ruin alluded to, though not being a fact, is viewed as such by the speaker with the eye of his mind, or, rather, in his imagination. Cf. g 90 uaí vv U8V evf}' ó yéeOJv ano f}vp,Ov 8Awa8v, I el ¡ti¡ ... "and now the old man would here (certainly) have lost his life, had not ... " (see also B 155; 8258; Z 73; H 104; g 217; A 311; s 426; 436 etc.). This construction even occurs in connection with an implied condition: r 56 "the Trojans are utter cowards: else would you (certainly) have been killed by stoning": i¡ 'té U8V ijdr¡ I Aá¡;vov saao xmova. An utterance like E 311, on the other hand, gives-or ifwe would prefer to express ourselves more cautiously: formerly gave-evidence of doubt on the part ofthe speaker: uaí vv usvl!:vf}' &n6AOt7:0 ... , I el ¡ti¡ ... "and now would (or: might) he have perished ... had not ... "1). In the passages t 24lf.; E 302f.; A 271f. the opto refers to processes, which, in the view of the speaker, have a remote chance of being realized. Miss Rahn is of the opinion that the sentences containing an opto in present unreal conditions-E 303 xse¡tádtov J,áfJs ... I ... 8 ov bvo y' fívdes cpseOtSV, I olot vvv fJeo7:oí slat-have facilitated the application of this mode to past unreal conditions, because they denote an action or state that is true of all periods. Since the hist. tenses occurring in the same passages refer to the past, and the optative also under other circumstances refers to past processes, this supposition does not seem probable or necessary. Leaving the results of a comparative examination of the relevant facts out of consideration, it might on the contrary be surmised that the imperf., being a more decided and definite mode of expression,::was destined to supplant the optative in some of its functions.

It is almost needless to add that an sl clause;containing an indicative refers to a process visualized as an actuality 2): in g 90 (see above) etc.: sl ¡ti¡ fíe' 0$-0 v6r¡as ... )1cpeobÍ7:r¡ the divinity had in fact been quick to see. Of. e.g. also Z 128f.; 142f. ("reine Annahme").

1) The reader may also consult Chantraine, Grammaire homérique, Ir, p. 226f.; 283f.

2) See e.g. Kühner-Gerth, o.C., Ir, 2, p. 466ff.; 477ff.; Chantraine, o.c., Ir, p. 283f.; 276ff.

Criticism of Miss Hahn's views 121

In passages such as B 80 1) 8l ¡tév 7:t~ 7:0V ovueov )1XauTw fíAJ.O~ evtan8 (past tense), 'ljJsiJM~ usv cpa'i¡tsv "were it any other of the A. who told us the dream we might deem it a falsehood" the speaker's imagination forms a picture of another man telling the dream; the verb form of actuality corresponds or at least formerly corresponded to the 'imaginary actuality' of the process of telling. Similarly Q 220.

In the subsequent part of her argument it becomes still more evident that the American authoress has attached too much value to the almost time-honoured distinction: subj. expressive of will: opto expressive of wish, that is to say: that she focussed her criticism too exclusively on those who think in terms of this oppositior( 2). In her detailed and ela­borate discussion of the ¡tI¡ and sl clauses she again attempts to show that snch utterances as X 123 ¡tI¡ ¡tw ¿yeh ¡tBv rUOJ¡tat lcóv "let it not·be that I approach him as a suppliant"3) and n 255 ¡ti¡ ... alva fJía~ an07:íaeat ¿Af}CÓV "let it not be that terrible will be thy coming to avenge violence" take the subj. not as an expression ofwill-"for the speaker's will does not enter into the situation" (1)-but as an expression of futurity. But this alternative is not essential and, in my opinion, even incorrectly posed: in both cases the speaker rejects or wishes to guard against the realization of a p1'ocess which is actual in his mind, of an impending occur1'ence which he sees before him, which oppresses and alarms him. Of. e. g. JI 128 ¡ti¡ bi¡ vfía~ SAOJat ... "(Up now, Pat1'oclus ... I see by the ships the rush of fi1'e.) Let it not be that they take the ships ... " 4). In these ¡tI¡ clauses the element of (negative) wish-Miss Rahn calls them "a type of wishing clause" -is inhe1'ent in ¡tI¡: cf. also n 381 ¡tI¡ n uauov éé$OJat "let them (may they) not do (us) some harm" or "beware, lest they ... ". With a view to proving her thesis, Miss Rahn, like her pre­decessor Walter, is not reluctant to adduce passages in which the same, 01' similar, contents are exp1'essed by a fut. indo We have seen that according to Walter the words uaí n07:é n~ s'bq¡at (Z 459) should be t1'anslated: "Und es wird einmal jemand sagen", mainly because in 461 it reads: w~ noré 7:t~ ¿eSst 5). I wonder hard if the verbs in s 415f. and

1) Hahn, O. C., p. 94. 2) Cf. e.g. Hahn, p. 99, par. 133 in fine; n. 259, p. 103 top; § 139; p. 113, § 152;

and especially p. 115, § 155. 3) See also Chantraine, O.C., Ir, p.208. A. Walter, Die Grundbedeutung des

Konjunktivs im Griechischen, Heidelberg 1923, who, like Miss Hahn, advocates the 'future theory' with respect to the subjunctive translates: "ich werde ihm doch nicht nahen" (p. 22f.), whereas Miss Hahn prefers: "may 1 not approach .•. (1 fear that 1 may)".

4) Walter, O.C., p. 23 "Sie werden doch nicht die Schiffe erobern", which might be misunderstood.

5) Walter, o.c.; p. 10. See also Hahn, o.c., p.78ff., who however does not suggest that the subj. and the fut. mean precisely the same thing. More correct is the view pronounced by Kühner-Gerth, o.c., Ir, 1, p.218.

Page 65: Gonda.1956.The Character of the Indo-European Moods.pdf

122 The Indo-European moods

tjj 563ff. may be put on a par: in both cases the future in the s~cOl:d clause has no connection with pi¡ in the first; if the process embodwd In the {ti¡ + subj. clause. will, .contrary to the ~peaker'~ wish, be realized 1), the contingency spemfied 11'1 the second will ?ertaI?ly happen. Or, to express ourselves otherwise: the process contamed m the. first sentence is viewed as realizable, it is expected, that referred to m the second sentence is imagined as future actuality. Cf. also A 587f.; n 87; a 76ff. 2) There is, in speaking, a shift in the view of the relation between the process denoted by these verb forms and reality on the part of the person to whom these utterances are ascribed. Of course both p:oces~es belong from a chronological point of view to the future, but thlS pomt is, in this connection, of secondary importance 3

).

1) Cf. also W. W. Goodwin, Syntax of the moods and tenses of the G~~~k.ve~b, London 1889, p. 132. Chantraine, 0.0., n, p. 354 needlessly translates: J al bIen peur que ... ".

2) Chantraine (0.0., n, § 410) devotes a paragraph to the subj. used if the author "envisage un avenir éventuel ou la répétition, ou ~ne notion général?". While 1 oannot agree with his additional remark that the partlOle T6 when ooourrmg in the relevant sentenoes emphasizes their general oharaoter (see my relevant papel' in Mnemosyne, vol. IV, p. 177ff. and 265ff.), 1 am on the other hand afraid that this peculiarity is mainly due to the oontext and often also to suoh words as :reov

h 1 " " t t' th " "anywhere, somewhere, per aps, suppose; :reoTe a some une 01' o el', ever (of. e.g. A 166; 205; p, 96 eto.). In A 81f. the poet has, it is true, the k~ng in gen~ral in view but his oherishing resentment (82) is visualizedas aotual, hIS swallowmg down his wrath (81) and the resultant prooess "till he brings all to pass" are assumed 01' viewed as to be taken into aooount. Cf. r¡ 204. "1 don't go to the oity" Eumaeussays; 373: "(aotual: leXop,at), unlesshaply (:reov) P. bidsr;t~ (subj. :suPl?ose~ 01' to be taken earnestly into aooount, to be expeoted), when tldmgs oome (Id.). Cf. also a 188. That the reality of the prooess is assumed 01' exists in the mind of the speaker, is espeoially olear in those frequent oases in whioh asimile is eXl?ressed; but as we have seen, the subj. does not refer here to the future, but to any tIme, the pro~ess being indioated independently of time: e. g. 1481 uaí p,' eCfíAr¡a' 0,; 6l ~6 :rea.i¡e 01' :reafóa q;tA~an "and oherished me as a (: any) fat~er oh~rlshes hIS son . Why should this truth based on experienoe have been proJeoted mto the future? Cf also A 116' JI 263' X 191 eto. The same remark applies to a general truth in other passages': :re 98 ,Zbrothers in whom aman trusts even if a great feud arises" (Oer¡Tat : let us suppose, there arises a feud). Cf. also A 81; r¡ 204; !t 96 eto. where the verb likewise does not exolusively 01' neoessarily refer to the future. In other oases it oan, of oourse, do so: in A 340 the person speaking sees, so to say, the event ooming: "let them be witnesses ... if so be hereafter there shall be need of me" : 6l :reOT6 ói¡ aiíT6 I xe6UD e/tefo yé1'r¡Tat; the word aiíT6 "again, furthermore" and ~he oontext (laTw1' in the prinoipal olause) leaving no doubt whatever. After avn,; and fut. in the pI'. 01.: E 258; eto. Cf. also M 245 (aAAol); 223ff.; P, 348ff. (oontext). The differenoe between assumed 01' impending reality with referenoe to the future and future aotuality implying oertainty and positiveness is especially olear in suoh oonstruotions as X 86f. el :re6e yáe a6 uaTauTáY17, oiJ a' 8.' eyÓJ y6 I uA~vaO~tat BY A6Xéwat. Cf. also 6221 aií "again, onoe more" + subj. 1 would prefer thIS VIew of these passages to Chantraine's opinion: "le sens de ~'hypothe~e .est indéfin~."

3) 1 can, with the same reserve, adhere to MISS Hahn s mterpretatlOn of X 123f.

Critioism of Miss Hahn's views 123

A critical remark 1 ) may be made in connection with Y 301; ro 544 {t17 nror; ... K(!OY[dr¡r; "exo}.wae7:G.t "let not Zeus be anywise (surely) wroth". Here there can be no question about Zeus' wroth. This construction does prove that {ti¡ clauses such as A 522 and E 845 likewise express "futurity and nothing else" 2). Although these passages, of course, imply futurity, "(go away)" {t1] n yoi¡a?7 "lest she notices something" in A 522 is first and foremost "prospective", i. e. modal, and "(she put on the cap)" fL'f¡ {tw LdOt "to the end that he should not see her" in E 845 before all also 'contingent'. After the primary tense and the present im perative the subjunctive is chosen because man, speaking about the consequences of things present, usually imagines these as ideal but impending actual­ities, as events of which he has to take full account and with regard to which has to decide on his attitude. Happenings belonging to the past especially when they have come to other people 3 ), are as a rule described in a more 'disinterested' way, so that references to their consequences, to the fears and desires aroused by them, can be suitably expressed by a form which essentially emphasizes potentiality 01' contingency, which in any case passes over the question as to how far they were actual, i. e. became realized. Another mode could not serve to express this implication. So 1 cannot subscribe to the conclusion that "we have so far as the future goes, a real system of sequence of tenses" 4). The verdict5) that "there is not a single type of dependent clause that cannot wherever futurity is involved take either a subj. 01' an opt., without the slightest distinction based on the question whether will 01' wish is involved" should, in my opinion be modified as follows: "there ... wherever non-actuality is involved take ... , the distinction being primarily based on the question as to whether the relevant process is visualized as impending, as mental actuality awaiting realization, as to be reckoned with etc. 01' as merely contingent" 6).

These critical remarks have led us to the so-called optativus obli-quus (in a wider sense of the term)7) in Greek. Attempts have been

1) Cf. also Hahn, § 161 in initio. 2) Cf. also Hahn, p. 113, § 152 in fine, eto.; § 154; § 155. 3) Cf. also the instanoes quoted by Miss Hahn in § 146. 4) Hahn, 0.0., p. 109. In oonneotion with N 648f., where the subj. is used after

a past ind., 1 would reoall the "objektivierende Darstellungsweise" of the historioal style (see Kühner-Gerth, 0.0., n, p. 392f., of. e.g. Hdt. 6, 9); the author so to say reproduoes the state of mind of the subjeot of the prinoipal olause, oarrying himself baok to the situation in whioh 1;he oontents of the subordinate olause were pronoun­oed. - See also my remarks made in Museum, 59, 189.

5) Hahn, 0.0., p. 115. 6) As to the optatives mentioned in Hahn, n.286 1 would prefer to regard

(jj 429 p,axo[aTo, as expressive of oontingenoy; Q 74 as a oourteous opt.; 75 el:rew as based on a 'propositive' subj.

7) Sohwyzer-Debrunner, 0.0., n, p. 333. See also Delbrüok, Vergl. Syntax, In, p.439ff. .

Page 66: Gonda.1956.The Character of the Indo-European Moods.pdf

• I 7 a

124 The Indo-European moods

made1) to derive this idiom, which is still developing in the early historie al period, from the cupitive (purely optative) or potential use respectively 2). The sentence E 2 oO'n,s f11j1o~ ... ly' B"or¡Ao~ ... yBYWíO has, for instance, been explained as having originally meant: "she gave him might, think­ing «may he be pre-eminent ... »" (and hence "she gave ... that he should be ... ") and e 368 steOY'ro Tt~ dr¡ "they asked who he was" ("wer er wohl sei") as having arisen from deOYTO' Tt~ aY s'tr¡~. It would, again, be better not to stress this alternative. Cases are indeed not wanting in which neither class of opto can be considered: Hom. h. Ven. 214 srreSY TS g"aaTa I w~ BOt a:{}áYaTO~ "aL . .. "he said everything, that he (his son) was immortal and ... (not: would be, might be, could be, etc.)". Here the more guarded and cautious character of this mood may have played a part: the author prefers to pass over the reality of the event at issue in silence; cf. also such instances as Xen. An. 2, 1,3 BAsyOY 8n Kveo~ P,eY TS-{}yr¡"SY, )1. oe nscpsvyw~ . .. s'tr¡ ("der Indik., die Aussageform der GewiBheit, li1Bt neben diesem furchtbaren Faktum das andere in den Hintergrund treten", Rehdantz-Richter) 3). The same factor-and under certain circumstances an inherent element of incomplete actuality­may be considered with regard to the opto in indirect interrogations (e 368 sle0YTO Tt~ s'tr¡ "they asked who he was") and to the <relative' sentences of the type t 402 s'teOYTO ... 8Tn S "17(JOt "and they asked him what ailed him" (beside e 120f. de STO ... 8TTSV xer¡tCOJY [,,6p,r¡y "he a,sked me ... in quest of what 1 had come": the coming is a fact)-to these two types this construction is confined in the epics-. It can hardly be admitted that the infrequent cases of direct interrogations with opt. 4

).

have exerted much influence. Cases such as A 838 are dubitative rather than interrogative; E 190 the opto is mild and courteous. 1 would rather suggest that apart from the tendency to be guarded in quoting thoughts or statements and in referring to facts which belong to the past and which the speaker often has not witnessed himself, and for the truth of which he will not make himself responsible-the opt.obl. is

1) Schwyzer-Debrunner, Le. 2) 1 do not eompletely understand Humbert, O. c., p. 122 who on the one hand

holds this opto to be the "optatif de possibilité" which assumes the eharacter of an opto of subordination, e:x:pressing a mere hypothesis, and on the other hand regards a sentence such as Q 249 TÓl' nOT' sydw ••• I a~w .•• , ll'a ¡Wt flíOTOV no,1.vl' a,1.rpOt which is without a shadow of doubt of a final character - ef. the same mood in final clauses in Sanskrit-as an argument in favour of the former assumption. The very absence of al' (even if the corresponding direct discourse would have subj. + al', e.g. Xen. An. 2,1,2 seems to show that for the Greek 'Sprachgefühl' this opto was no potential.

3) Cf. also Goodwin, o.c., p. 261, § 670. 4) See Kühner-Gerth, O. C., n, 1, p. 230f. It is perfectly understandable that

contingent processes are not often expressed in direct interrogations, whieh tend to be exactly formulated requests for affirmation, negation, 01' supplementary information bearing upon processes visualized as actual.

Criticism of Miss Hahn's views 125

mainly used if the subject is in the 3rd person and is not identical with the speaker or narrator 1 )-< analogical' processes are not foreign to the spread of this construction in Greek: an opto in the indirect construction often corresponds to an opto in direct speech: E 300 TOY "Táp,syat p,sp,awc; 8~ Tt~ TOV y' aYTtO~ BA-{}ot "eager to kill the man whosoever should approach him"2). Another starting-point may be seen in the opto corresponding to a subj. in direct speech: see aboye, and in general those clauses which express fear, doubt, intention, purpose, expectation, etc. 3) Passages such as o 458 tiyysAov i}~aY, OC; ayystAStB yvyat"t containing an opto which, along the traditional lines, could be variously classified, are typically contingent: there was a chance that the messenger would not succeed in bearing the tidings to the woman, although he was instructed to do so. Of. also C 50.

With regard to the aboye tendency not to emphasize the reality of past processes which are referred to in the narrative and to adopt a a disinterested attitude towards them or to relate only the intentions of the characters in thest ory-cf. e. g. E 845 OVYS ... "Vysr¡y, p,~ p,w 'toot ... Y:l.er¡~ "(Athene) put on the cap ... to the end that Ares should not see her"-it may be observed that this mood-which in itself does not refer to any time-was suited to be used in connection with an historie al tense in the principal clause 4). This use spread in early post-Homeric times. Enunciative sentences (Aussagesi1tze) were not from the beginning formed in the same way5): the actual character of the verb contained in them could no doubt for sorne time prevent the opto forcing its way into them 6).

The optativus obliquus in the strict sense of the term-the oldest instance is h. Ven. 214-is that constructioll by which the speaker or

1) Compare also the use of the Slavonic conditional: Vondrák, o.c., n, p. 398f. 2) This point was long ago emphasized by F. Urtel, Homerischer Gebrauch des

Opto del' abh. Rede, Programm Weimar 1884; cf. also Behaghel, Der Gebrauch del' ZeiVormen im Konjunkt. Nebensatz des Deutschen, Paderborn 1899, p. 161ff.

3) Schwyzer-Debrunner, o.c., p. 332f.-It may be observed that neither this digression nor other similar pericopes in this discourse pretend to be exhaustive. Cf. also Brugmann, GrundriJ3, n, 3, p. 874ff. Behaghel, Deutsche Syntax, nI, p.699 observes that in indireot speech "die Redeeinleitung ... der Rede ein subjektives Geprage gibt, die der Rede selbst für sich allein nicht zukommt".

. 4) c~. also 8240 and simila~ instances: Cal!pso planning Odysseus' departure, glves h1m an axe and leads h1m to a spot wlth tall trees, '1:á oí n,1.cóot13l' s,1.arpQwt; "which would float for him lightly". The process.of floating is neither actual nor present as 'virtual actuality' in the mind of the narrator: it may have been such in Calypso's mind, but the poet who only interprets and relates her intentions is from his standpoint thrown on that form whioh can express contingency with reference to the pasto

5) Schwyzer-DebrUnner, O. O., n, p. 333. 6) For Slavonic parallels see Vondrák, o.c., n, p. 398f. The German develop­

ment may for the sake of brevity be read up in Behaghel, O. C., nI, p. 670ff.; 699ff.

Page 67: Gonda.1956.The Character of the Indo-European Moods.pdf

126 The Indo-European moods

narrator does no longer in a vague 01' cautious way refer to, relate, '01' interpret, the thoughts of another person, but quotes his words without such restrictions, referring to historical facts: Lysias 12, 74 81Je8v 8-n ovMv aV"ít¡> flsAot . .. , 8Je8U5i¡ . .. 8Meír¡; cf. also Soph. O. R. 1161 8lJeov w~ ootr¡v (1157: eowxa). In this stage the obl. opto may be said to have become a means oí indicating the so-caIled intern dependence. In order to avoid obscurity with regard to the actual character of the relevant process many authors, however, often preíer the imperfect to the oblique opto in such cases as Xen. An. 1, 2, 21 ~X8V t1yy8Ao~ ASywv, 8-n A8Aot7CCV~ e'ir¡ I:. "tC1, t1xea, 8Jeel fíaih"ío, 8"ít ... (the speaker of course said: UAotJe8I: . ... , 8Jeel fía{}8"íO ... ). Thus, aIl dependent secondary tenses of the indicative remain, in indirect quotation of various complex sentences, unchallged after aIl tenses: Thuc. 7, 80 'iíAmCov "íOV~ I:tX8AOV~ "íav"írJ, o1J~ fl8"íSJe8fl'!fJ8V, d"WV"íf¡a8a{}at "they hoped that the S. whom they had sent for (historical fact) would meet them there".

A word must also be said on Miss Hahn's view ofthe W~, 8Jew~, (frpea, and ¿va clauses l ). Her conclusion that they are made after the pattern of the corresponding dependent flf¡ clauses, because they cannot be traced back to independent utterances, does not convince me. I would con­sider the use of the modes in these clauses in connection with that in the clauses introduced by the so-called relative pronoun 8~, to which dk, 8Jew~, and ffrpea etymologically belong, and recall the frequent occurrence of subj. and opto in the equivalent constructions in Sanskrit alld A vestan. Both subj. and opto frequently occur in clauses introduced by the so­called relative pronoull *jo-, ~V. 9, 101, 9 rayí'Y(bl yéna2 vánamahai3

"wealthl by which2 we see ourselves being masters3" (in contradistinction to Delbrück's " ... wir siegen mogen" alld Geldner's "wir Sieger sein werden") 2); 10, 53, 4 yéndlsuran2 abhí3 devá4 ásama5 "by whichl we, gods4, shall conquer3,5 the asuras2", 01' more precisely, "in order that we, gods4, conquer3 , 5 the asuras2 by itl " (Geldner: "überwinden konnen"); 2, 30, 5 yénal sátru'Y(b2 ... nijiírval}3 "in order to destroy3 the enemy2 by it it l ", 01': "by whichl I see you destroying3 the enemy2"; SB. 6, 3, 1,20 ... yá1 etát2 sárva'Y(b3 vindát4" ... welcherl dieses2 alles3 erwerbe4

" (Delbrück); there is in my opinion a similar difference between these subjunctives and futures to that existing in German: the clause containing the subj. involves a purpose 01' end in view (Absichtssatze). This is not always the case. 1, 113, 11 ól té2 yanti3 yé4 apart~u5 pásyan6 "they2 are to comel ,3, wh04 are expected to see6 (them) in the future5". In the frequent construc­tion 1,132,6 yól nal}2 prtanyád3 ápa4 tár(lAam5 íd6 dhatam7 ("wer uns be­kampft, wenn uns einer bekampfen wird" Delbrück, "del' uns allfeinden

1) For the original character and function of l.-E. *jo- see Lingua, IY, p. lff.

2) K. F. Geldner, Rig-veda, nI, p. 105.

Criticism of Miss Hahn's views 127

solIte" Geldner) reference is not made to actual hostilities, but to ima­gined attacks, to acts, anticipated by the person speaking: "whoeverl ,5

shall attack3 us2, beaV him5 off4" seems to be the nearest approach in translating into a language which does no longer possess a subjunctive. Similarly ~V. 2, 23, 4; 1, 156,2; 4, 30,23; in prose texts the construction represented by AiB. 1, 24, 6 tabhirl eva2 nal}3 sa4 na5 sa'Y(bgachatai6 yo7 na8 etad9 atik?'amatlO "he4 of us3 wh07 transgresseslO this9 of us8 is (in the reality seen by the speaker's mental eye) not5 united6 with theml ".

The subj. in the rel. clause and that in the principal are completely on a par, cf. e. g. also TS. 2, 4, 1, 2 vara'Y(bl vr'IJamahai2 (subj.); yad3 asurañ4 jayama5 (subj.) tan6 nal}7 sahéisad8 (subj.) iti "let us (we shall) choose2

a boonl ; what3 we win (shall win)5 from the asuras4, let8 that6 be shared8

between US7"1). Avestan instances are: Y. 16, 10; 50,42); Greek: E 407 ov or¡vau3~ 8~ d{}avá"íotat fláxr¡rat3) "he endures not for long who fights (wiIl fight) with the immortals"; (jj 103 4), and especiaIly such cases as 1 165 XA1]"íOV~ O"íeVVOfl8V, ol X8 ráxw"ía I eA{}wat "(to) send chosell men to go forthwith ... " Compare also the Latin instances: Plaut. Most. 344 da illi quod bibat "give him a drink", where the subj. must be put on a par with that in date bibat5); Capto 103 nihil est quo me recipiam "there is nothing with which I shaIl (be able to) go home", etc. Similar constructions are also proper to the Slav. languages: Polo Flor. Ps. 49, 23 nie badzie, jenze by odjal "ne sit qui eripiat" 6).

We now mention sorne cases in which the Sanskrit 01' Greek optative appears under the aboye circumstances: ~V. 6, 19, 7 yásl te2 mádal}3 . .. índra4 tá'Y(b5 na6 á7 bhara 8 ••• yéna 9 tokásya lO tánayasyall sataú I2 ma'Y(b­simáhi I3 jigivá'Y(bsal}14 "deinen2 Rausch3 ... bring7,8 uns6, Indra4, mit7, •••

durch den9 wir glauben dürfen13 bei dem GewinnenI2 von leiblichem SamenIO,11 Sieger zu seinl4" ; AV. 8, 10, 9 tádl veda2 yád3 ubháya4 upajt. vema5 "she knows2 thatl upon which3 we of both classes4 may (could) subsist5"; SB. 11, 5, 1, 13 agnerl ... tanür2 asti3 yaye4~tva5smakam6 ékal}7 syat8 "there is3 the manifestation2 of AgnjI by which4 if one worships5

. (with it) one could become8 one7 of us6" 7). In Greek B 687 ov yae er¡v, 8~ rt~ arptv ... f¡yf¡aatro "since there was nobody to lead them".

1) Other instances may be found in Delbrück, Altind. Syntax, p. 317ff. 2) See also Reichelt, Awest. Elementarbuch, p.368. a) Walter, Die Grundbed. d. Konj. im Griech., p. 34ff. is right in combating

the "fallsetzenden Konjunktiv", wrong in believing this form to be essentially "futurisch" .

4) For the subj. in the Homeric símiles (O 624 8j1 (j' enea', w, (he ')(,vf1u Doñ 8V jl71t néanatv see especially Kühner-Gerth, o.c., 1, p. 251. The subj. is used when the situations and phenomena described are not "zu jeder Zeit gegenwartig VOl' Augen stehend".

5) See W. Kroll i~ the Glotta, 5, p. 338; 7, p. 140ff. 6) Fol' pal'ticulal's see W. Vondrák, O. C. n, p. 398. 7) See also Delbl'ück, Altind. Syntax, p. 192ff.

J

Page 68: Gonda.1956.The Character of the Indo-European Moods.pdf

128 The Indo-European moods

So there seems to be no room for doubt that these modi when occurring in similar sentences introduced by relative conjunctions, which are often derived from the stem *io- are, also in Greek, original, i. e. prehistoric, and not introduced, as is held by Miss Hahn, after the pattern of the fl1}

clauses. Cf. e. g. the correspondence between ~V. 10, 85, 26 grhán1 gacha2 grhápatni3 yáthá4sa7y,5 "g02 to the house1 that4 you will be5 (in order to be5) housewife3" and Ll 205 uaAéet . .. A. I orpr¿a 'to'{}r; M. "A. calls you to see M."; ~V. 10, 131, 1 ápal ... amítrán2 ... nudasva3 I ... uraú4 yátha5 táva6 sárman 7 mádema8 "vertreibe1 , 3 die Feinde2 ••• auf daB5 wir uns unter deinem6 breiten4 Schirm7 erfreuen konnen8 (Geldner) and the final wr;, (Jnwr;, orpr¿a, Zva + opto in Greek.

We append here a brief discussion of a point which was already touched upon in the preceding chapter, to wit the frequent interchanges between the moods in Vedic texts. Does the fact truly observed by Bloomfield and Edgerton 1) that, as far as we can see, any one of the moods may interchange with any other, justify the conclusion drawn by Miss Hahn2), that the entire modal system is in a highly fluid state ~ It is clear that this verdict can only be substantiated if we succeed in demonstrating that the relevant verb forms are completely interchangeable in syntactic wholes (clauses, sentences) of exactly identical meaning and purport and uttered under completely identical circumstances. If we cannot succeed in demonstrating this complete identity we are taking the risk that our verdict is not altogether dissimilar to the judgement: in modern Dutch the ·tense' system of the verb is in a fluid state, because "ik heb mijn broer gezien" can interchange with "ik zag mijn broer". The former construction, is, however, unlike the latter especially used with reference to what for the sake of brevity may be called the actual past; besides, in popular usage the simple past is, especially when the verb denotes an activity, rather rare 3). The objection that modern colloquial construc­tions may not be compared to the stereotyped, repeated, 01' at least recurrent phrases of the Vedic texts, is not valid in all respects. The patterns of a modern spoken language are also repeated unalterably under similar circumstances, and even a slight motiveless deviation does not fail to strike the hearer.We must not lose sight of the fact that the Vedic mantras, although they are repeated, with 01' without variations, in two 01' more texts, were later on not always used under the same circumstances under which they were composed. This fact could easily lead to their being adapted and changed to serve a new purpose. Further, many alterations may be supposed to have been introduced unconsciously, not seldom under the influence of similar text units.

1) M. Bloornfield and F. Edgerton, Vedic Variants, I, Philadelphia 1930, p. 94. 2) Hahn, o.c., p.40 with n. 60. 3) See G. S. Overdiep, Stilistische grammatica van het Nederlandsch, Zwolle

1937, p. 320 and 315.

Oriticism of Miss Hahn's views 129

Directing our attention first to the interchange of the optative and other moods, it may be remembered 1) that it alternates with the indi­cative only in a moderate number of cases. The authors of the Vedic Variants already observed that the latter mood makes here as usual the impression of greater certainty 01' insistence. This may' be one of the reasons why the AV. sometimes prefers avase1 havámahe2 "we ca1l2 to aid1

" (indic.) to avase huvema (opt.)2) (~V. 7, 96, Id has the intensive johavimi). The opto "we would call" is in itself perfectly understandable and a diaskeuast may have preferred it, inter alia, for its more courteous nuance. We should, moreover, guard against misunderstanding in calling these alternations 'variants'. It is true, that, for instance, AV. 7, 63, 1 and TA. 10,2, 1 are parallel stanzas but the former text has, in the second quarter: (agnim) ukthaír havamahe paramát sadhásthat, the latter: (u?ram) ~gni'f(b huvema p. S.; the alternating moods are not the only pomt of difference. If two otherwise identical mantras differ not only in t~e moo,d, but also in the person of the verb 3), e. g. ~V. 4, 58, 1 úpá1'f(b­suna2 sam3 amrtatvám4 anat5 "by the soma plant2 he (it) has reached3,5 immortality4": ApMB. 2, 3, 2 U. s. a. aByam "by ... may 1 reach i."; (cf. also TS. 4, 3, ll, 4: KS. 39, 104)), the tenor and meaning of the text­unit diverge to such a degree as to preclude us from speaking of "fluidity of the moods." It is evident, to quote a striking instance5), that AV. 17, 1, 25 dditya1 návam2 áruk~a7y,3 satáritrá'f(b4 svastáye5 "O Aditya1 (the Sun), thou hast ascended3 a boat2 of a hundred oars4 in order to well­being5

" has Httle in common with VS. 21, 7 sunavam1 aruheyam2 .. , 8. 8,

"m.ay 1 mount2 (1 would like to m. 2) this good Ship1 ••• " (to wit: sacrifice, WhICh b.ears man to heaven, the oars being the hymns of praise) which occurs m a text addressed to Varul).a. The line indrasya1 sakhyam2 amrtatvam3 asyam46

) "may 1 obtain4 (1 would obtain4) the friendship 2 of 1.1, immortality3", occurring TB. 3, 7, 6, 14 and ApS. 4, 8,4 corre­sponds to ~V. 10, 62, 1 i. s. a. ánasa "Ye have obtained ... ", but the ?ontexts .are differe~t: the former unit forms part of the texts accompany­mg the rItual touching of the oblations and is pronounced by the institu­tor o~ the sacrifice, the latter is addressed to the Aúgiras; besides, the ~V.lme forms part of a relative clause. If not only the mood, but also the person and the very root of the verb vary, it would be unwise to use a recurrent line as an argument in favour of the view that the moods can freely interchange: cf. e. g. ~V. 3, 59, 1 juhota as against KS. 35, 19 vidhema in an otherwise identical unit.

1) See Bloornfield-Edgerton, o.c., p.77. 2~ AV. 7, 6, 2; 7,40,1 (where the Paipp. rec. has juhuvema); 3,16,1 and 2

(Palpp. huvema), etc. 3) Of. Bloornfield.Edgerton, o. c., p. 86. 4) And the relevant instances quoted by Bloornfield-Edgerton, o. c., p. 101 f., § 160. 5) Mentioned by Bloornfield-Edgerton, o. c., p. 86. 6) See Bloornfield-Edgerton, o.c., p.90.

9 Gonda, Indo-European moode

Page 69: Gonda.1956.The Character of the Indo-European Moods.pdf

... Wiii#\i

130 The Indo-European moods

N or can cases like the folIowing, in which the unit forms part of differ­ent constructions, be regarded as variants either: AV. 19,24,2-3 yáthai1nam2 jaráse3 náyat4 "thatl it will conduct4 him2 unto old age3

"

(subordin~te clause with subj.): HG. 1, 4, 8 athai1narp,2 jarimii3 nayet4. "then1 may (could) old age3.lead4 him2 away4". Bloomfield and Edgerton already pointed out that KS. 18, 13 changing a vocative to a genitive makes the line so 'harp, vajarp, saneyam agne no longer a direct address to the god, and therefore (perhaps) "less in tune" with the optative mood.

There is of course also the metrical aspect of the question. As long as the metre of the AV. and other texts has not been thoroughly investi­gated, we should be cautious in calling such alternations either corrections or corruptions. Although metrical convenience is often associated with a shift this does not mean that both variants are completely equivalent in meaning: ~V. 7, 66, 16jtvema saráda7y, satám "may we live a hundred years" is a wish, but 10, 85, 39 jivati s. s. expresses the speaker's con­viction that the husband shalllive a hundred years now that Agni has given back the bride together with a long liíe.

Even ií there is no océasion for the above considerations cases oí interchange between the mood are often explainable. The predilection for the imperative rather than the optative in the second person 1) may be connected with the very frequent use of the direct and coercive im­perative in these texts which almost always verge on magic; the optative impresses me as more considerate: cf. e. g. TS. 4, 1,2,5 arah}asa1

manasa2 tap j'U~asva4 "(O Agni) enjoy4 it3 with harmless1 mind2": KS.

16,2 a. m. t. j'U~etha7y,. The interchange between subj. and opt., which is most írequent in the first person, where the subj. expresses the speaker's intention and self-incitement 2), may be regarded as being produced by similar causes.Whereas the consistent use oí the opto in a prayer for long life and prosperity (AV. 19, 67), which is traditionalIy held to belong to the worship oí the sun, is quite understandable: páByema1 sarádaJ:¿2 satám3 "may we see1 a hundred3 autumns2", the shiít to the subj. in the version oí the same text in VS. 36, 24-where it is used in the pre­paratory rite of the Soma sacrifice-may have arisen írom other reasons than a complete interchangeability oí the verb forms: after "may we live a hundred years" the words "we will (intend to) praise for a hundred years" (pra bravama s. s., subj.) may be due to a conscious alteration on the part oí the diaskeuast 3). In the version presented by HirGS. 1, 7, 10-where the text forms part of a series of rites connected with studentship-the first two verbs are in the opto "may we see a h. y.,

1) Bloomfield-Edgerton, o.c., p. 101. 2) Cf. Bloomfield-Edgerton, o.c., p. 107. 3) See also W. D. Whitney-Ch. R. Lanman, Atharva-veda Sarp.hita, Cambridge

Mass. 1905, p. 1006.

Criticism of Miss Hahn's views 131

may we live a h. y.", the next four in the subj.: "let us enjoya h. y. etc."; here too I would prefer to take into account the possibility that the < author' in formulating this text shiíted his standpoint: commencing with the prayer he soon came to anticipate the future (on which he had set his heart), to experience it in advance. Of. also ~V. 10, 85, 37d: AV. 14,2, 38d.

Part oí the instances of an injunctive alternating with an opto may be due to the tendency to replace a verb oí modally indistinct function by an unequivocal íorm: MS. 4, 13,9 asyam rdhad dhotrayam: SB. 1, 9, 1,22 a.1 rdhed2 dh. 3 "may he succeed2 in this1 sacrifice3".

It may finalIy be observed that the authors oí the Vedic Variants 1) mention only one instance of an opto in interchange with a future.

1) Bloomfield-Edgerton, o.c., p. 111.

9*

Page 70: Gonda.1956.The Character of the Indo-European Moods.pdf

VIII. The lunction 01 the so-called modal particles

Miss Hahn following Sturtevants's extreme hypothesis is convinced that "lndo-Hittite", the supposedly earlier form of lndo-European, did not in contradistinction to a later stage of this parent-tongue possess subjunctive and optative. Modal distinctions, she maintains, were in the Hittite manner marked by means of particles rather than verb forms. Although she may, broadly speaking, perhaps be sofar right, that once so-called modal notions could to a certain extent be expressed by means of particles at a period in which subj. and opto had scarcely developed, any assumption to that effect would be mere speculation. Miss Hahn relies too much on the lndo-Hittite hypothesis 1); besides, the question may arise whether she is right in taking for granted that the Hittite and the supposedly original particles had exactly the same functions as the moods in early historical times.

l cannot attach decisive value to Miss Hahn's argument derived from the double expression of negation (Gr. ov and fli¡) which, she holds, proves an obvious similarity of the l.-E. languages in the matter of preferencefor particles. The use of two negative particles is, as we have seen, a phenomenon of wide occurrence: cf. also the two negative verbs in Finnish, the relevant constructions in Bantu idioms etc. It seems therefore pointless to conclude from ov : fli¡ etc.-which moreover do not agree with Hitt. natta: le etc. in their outward appearance-that l.-E. generally still used particles as lndo-Hittite had done. Negation is on the other hand in the l.-E.languages generally expressed by particles, not by special forms of the verb: there is no sound reason for supposing the parent-tongue to have developed a negative or prohibitive verb form, for instance in the Finnish or Japanese manner. Moreover, negation and the 'modes' indicated by the subj. and opto cannot be put on a par. l.-E. negatives are as a rule associated with the 'notional moods', not a sort of mood by themselves.

The supposed congruency between Greek and Hittite with regard to the use of a 'modal particle' 2), cannot, in my opinion, prove the cor-

1) For this hypothesis see now also T. Burrow, Sanskrit, London 1955, p. 17f. 2) Unless used in the sense of "partiele joining a modal verb form" this term

is objeetionable, the more so as it is variously interpreted (see also ehapter 1). Sometimes any modifying partiele whether it be expressive of emphasis, negation, exelamation or interrogation is styled modal (see e.g. Speyer, S. Synt., p. 310); authors of Greek grammars apply the term to those little words whieh are eonsidered "to emphasize the sense of the moods", espeeially to di)) and ue(v), ua (see e.g. Sehwyzer-Debnmner, o.e., n, p. 305).

The modal partieles 133

rectness of the hypothesis that particles of this category originally ful­filled the function ofthe moods of the historical period 1): firstly Greek and Hittite are alone in possessing these elements, secondly /J.y and man are etymologically different words, and lastly, they are not used in the same manner. Besides in Latin and Gothic the no doubt identical an has other functions. For the fact that the Greeks combined the particle /J.y with tenses (which denoting aspect not time, point according to Miss Hahn to the original character of this construction) l propose, in the following pericopes, another explanation.

What was, to begin with, the function of particles in general 2) 1 Broadly speaking, they serve sometimes notionally, but often emotionally, to qualify the content of a sentence or clause or the sense of a word or phrase. They are indications of the speaker's attitude or disposition with regard to the utterance, intimating the general sense in which it should be taken. They show in what particular manner the speaker is interested in the content of the utterance; for instance, he may wish to emphasize it or part of it, or to express doubt or strong suspicion as to its truth, or to admit the truth with some reserve, to express his surprise or other feelings etc. They can also help him to bring out the mutual relations of the components of his utterances, to establish a relationship between ideas 3).

1) Miss Hahn is not alone in defending this thesis with regard to the partieles. Delbrüek, Conj. U. Opto (= Synt. Forseh., 1), p.90 for instanee gives evidenee of the opinion that (iv and U8V "nur ein beredterer Ausdruek dessen (sind), was aueh dureh den bloI3en Conj. und Opto ausgedrüekt wird", a verdiet whieh was readily endorsed by A. Walter, Die Grundbedeutung des Konj. im Grieeh., Heidelberg 1923, p.65. But even if onee particles fulfilled a funetion similar to that of the modal verb form, it would not follow that the latter were not modal in eharaeter. Besides the very faet that the partiele from the beginning joins more than one mood shows that it does not eonvey the idea implied in partieular moods. The view expressed by Meillet-Vendryes, Gramm. eomp.2, p. 256 ("la valeur expressive des modes opto et subj. s'est affaiblie, si bien qu'on les a renforeés par l'emploi de partieules modales") must, for these and other reasons, be modified: in the eourse of time the originally faeultative partiele beeoming obligatory in definite positions, developed into a means of differentiating one prineipal funetion of the opto and subj. from the other. Waekernagel, Vorl. über Synt., 12, p. 223f., maintaining that (iv and ue "ursprünglieh der modalen Bestimmung des Verbum finitmn dienen", says that Eur. Alk. 182 (iv "das awrpewv in die Sphare der Hypothese rüekt"; 1 would prefer: it indieaws that the woman's being a. is eonditioned by eireum­stanees: "then" > G. "wohl". For Hitt. man see J. Friedrieh, Heth. Elementar­bueh, 1, Heidelberg 1940, p. 79; W6rwrbueh, p. 134.

2) Bibliography: Brugmann, Grundrill 2,n, 3, p. 969f. (up to 1916); Sehwyzer­Debrunner, O. e., n, p. 553; Renou, Gramm. sanser., p. 520f.

3) The definition of partiele given by J. D. Denniston, The Greek partieles2, Oxford 1954, p. xxxvnruns asfollows: " ..• awordexpressingamode ofthought eonsidered either in isolation or in relation to another thought, or a mood of emotion." Cf. also the definitions and eomments in Speyer's Ved. S. Synt. § 225; L. Bloomfield, Language (1935), p. 199ff.

Page 71: Gonda.1956.The Character of the Indo-European Moods.pdf

134 The Indo-European mooels

So many particles qualify a whole predication, or at least a phrase; they are a means of expressing an attitude of the speaker: doubt, sur­prise, questioning, a social relation in addressing an inferior or superior etc. Thus the Vietnamese a! denotes, at the end of a declarative predi­cation, the speaker's great surprise at the content of that utterance 1

).

Other particles, in various languages, serve to express syntactic relations. Thus the element t in the Tagalog buto t balat "bone and skin" indicates serial relation; y in sya y sumusúlat expresses, in the same IN. language, the predicative relation: "he (she) is (was) writing". A widespread class of particles act as 'attributes' of full words. Thus the Tagalog sak~t means "sickness", but rnáy sak~t "having sickness, (being) sick". They modify the syntactic value of a full word. In Tahitian 2) a, preceding the predicate, indicates that it is an imperative: a tuama te mori "light the lamp!" The particle ua, in the same position, may be considered to bear the force of a sort of aspectual confective: ua po "it is night" (po "night"). "Particules perfectives ou imperfectives, particules d'éloigne­ment ou d'approche, particules de renforcement, d'intensification ou d'atténuation et de réduction, tout cet appareil de mots auxiliaires permet au sujet parlant de définir avec une grande précision les différents aspects de sonénonciation, sans que celle-ci rentre, toutefois, dans le cadre des catégories verbales que nous sommes accoutumés a utiliser dans nos langues occidentales, ... "3).

Much-used, and much-misused terms are: "modal particle" or even "modal auxiliary". They suggest the idea conveyed by the ancient l.-E. moods and they are, indeed, often applied to words of the class under consideration by those authors who confound the Greek or Sanskrit categories with any word or phrase in a foreign language which might be translated by English verb forms accompanied by the so-called modal auxiliaries 4). They are often, indeed, incorrectly identified with our can, may, must, ought, need, should etc. But the Malay hendak in hamba1 hendak2 djadikan3 tuanhamba4 saksi 5 "11 wil12 (shall) ca1l3 you4

to witness3,5" is neither an exact counterpart of our verb forms will etc., nor that oí a modal suffix in Greek 5). The 'sense' ofthe Mori (Celebes) koa 6 ) which in translating can be rendered by a great diversity of ex­pressions, does not properly speaking differ essentially from the effect of a special intonation, of certain gestures, etc. After an imperative it is slightly emphatical, after other verb forms it can, roughly, be "indeed",

1) Emeneau, Stuelies in Vietnamese gl'ammal', p. 211. 2) See A. Sauvageot, Le Tahitien, Confél'ences de l'Inst. ele Ling. de l'Univ.

de Paris, 10, Paris 1951, p. 92f. 3) Sauvageot, o.c., p.93. 4) 1 refer to chapter 1. 5) See my papel': "De zg. hulppl'aedicaatswoorden in Maleis en Javaans",

Bijdr. Taal-, Land- en Vollr., 102, p. 95ff. 6) S. J. Esser, Klank- en vol'mleel' van het Mol'isch, n, Bandung 1933, p. 216.

The modal particles 135

or "merely, only"; the latter sense can also be attributed to the particle when it accompanies a substantive.

It must, indeed, be emphasized that often the identical particle accompanies words of various classes 01' occurs in different types of sen­tences. Thus the Skt. jatu which in a sentence containing an opto helps the speaker to express censure, disbelief 01' intolerance (see PaJ}ini 3,3, 147), can in a certain type of interrogation convey the sense of "at all" (e. g. kirrdenajatujatena "what is the use at all ofhim being born 1"); elsewhere it means "possibly, perhaps" (Mbh. 12, 180,50 + opt.). The Skt. api which at the beginning of a sentence containing an indo helps to express disapproval (PaJ). 3,3,142), and which withanimper. strength­ens an exhortation, also occurs in interrogations, etc. etc.; while it can, in connection with an opt., convey the sense of "doubtless" (PaJ). 3, 3, 152) it is also often added to the cupitive 01' pure optative: Ram. 2, 43, 9 apídanírp, sa 7cala7y, syat "if that time were already present". These con­siderations and the fact that e. g. api + cupitive is not obligatory lead us to the conclusion that notwithstanding a certain predilection for a specified particle in sentences of a definite type containing a definite mood or tense, the function of the moods on the one hand and that of the particles tending to accompany them are differentl).

Coming now to the Greek particles discussed by Miss Rahn, to wit aY and ?lS2), 1 would not regard them as indices of modality (in sensu stricto). Row could this function belong to an element that is compatible with all three modes 1 It would appear to me that they essentially served to express doubt with regard to the absolute validity of the utterance 3).

They expressed a restriction, denoting that the assertion made by the verb is limited by circumstances 01' dependent on a condition, which may be expressed or implied. Whereas the modes indicate the speaker's view of the relation between process and reality, these particles served to modify the assertion with regard to the process itself 01' to limit the effect of the statement. Rence its absence from sentences referring to such unconditioned processes as are denoted by imperatives, from the

1) Passing mention may also be made of the Vedic nú, nu for the use of which see Gra.Bmann, W6rterbuch zum Rigveda, p. 744ff.; Renou, Gramm. véd., p. 376.

2) For the difference between av and %8 see Goodwin, Syntax of the moods and tenses, p. 65; Monro, Gramm. Hom. dial., Oxford 1882, p. 265ff. The conclusion arrived at by H. Seiler, Word, 11, p. 133 that the use of av must be of comparatively recent date, not even Proto-Greek, since in the Aeolic and Doric dialects other particles are used, is not convincing. The dialects seem to have made a different choice.

3) 1 cannot adopt the formulation: av dient "dazu die Bedeutung del' Modi und Tempora zu verstarken" (Pape, W6rterbnch). N 01' would we seem happy in adopt­ing for thE¡ original stage of development, Schwyzer-Debrunner's view (o. c., n, p. 306) that "die Partikeln lediglich die schon bestehenden Modusftlliktionen ver­starken" (cf. also Meillet-Vendryes, Traité de grammaire comparée2, Paris 1948, p.256).

Page 72: Gonda.1956.The Character of the Indo-European Moods.pdf

136 The Indo-European moods

present indicatives 1), or pure and unconditioned volitives and cupi­tives 2). Examples: r 138 A. and M. will fight, ?:<p bé ue v¿u~O'avn cpíJ..r¡ ueuA~O'rJ aUOt'tu;; b 546f. ("either you willl find Ae. alive") fi uev 'O. I wce¿vev ... ,3) O'V bé uev 'tácpov avn(JoA~O'atr; "or by chance O. has slain him (impf.), and in that case you may chance upon (opt.) his funeral feast"; E 235 neHlsv' eyc1 bé ué 'to¿ lbéOJ Xá(!W "obey, and if so 1 shall be (subj.) grateful". Cf. e. g. also A 100; 225f.; 2944); Soph. AL 88; O. R. 95•

That the character of av must have been as stated aboye, also appears from its occurrence, in early poets and sporadically in Attic in clauses with a future indic.: 11 176 uat ué 'nr; (M' S(!éet "and in that case many a one shall say" (fut. ind.); E 267 aAA' t&' eyc1 bé ué 'tOt ... I bwO'OJ "(if you fulfil my wish) then 1 shall give you"; A 139 (under the circumstances alluded to); X 66 "myself then last of all"; 1 167 (dependent on the context of 165); A 174; e 539 el . .. llM}o¿ I al1jJá ue ... fut.; cr. fk 345; from the so-called av iterativum: Xen. Anab. 1, 9, 19 (after el + opto in itero sense) "dann allemal" el bé nva O(!CPr¡ ... n(!oO'óbovr; notovv'ta, ovbéva tiv nó:mo'te acpeO.e'to; 2,3, n; Thuc. 7, 71; Xen. Cyr. 7, 1, 10 onó'te ... 'to'te fkev eInev av; Arist. Vesp. 269 ov fk~V n(!o 'toiJ y' ecpoAUor; i}v, aAAa n(!éi'nor; 1¡fkwv l1¡ye¿7:' tiv ... 5); Nub. 977, where "under these circum­stances, in those cases, (then)" (words like Dutch dan; Germ. da) is-in languages which have such words in this particular use-always an appropriate translation. cr. also Hdt. 2, 174 oUOJr; bé fkw smAe{no¿ ntvov'ta ... , uAén't8O'ue av ne(!uwv; 3, 119 cpot'téovO'a ... uAat8O'Ue av; and also 4, 42; 78; 130; here too the particle may be explained in the aboye way, cf. 1, 100; 186; 196; 4,43, where av is absent and Dutch "dan" would be unusua1 6). Cf. also Soph. PhiI. 290ff.; Eur. Ph. 401; Arist. PI. 982 (980

1) For particulars: Goodwin, p. 65ff.; Kühner-Gerth, o.c., 1, p. 210f.; W. Pape-M. Sengebusch, Griech.-deutsches Hdwtb.3, 1, p. 147.

2) In a subjective wish an objective limitation can often be dispensed with. The instances of a cupitive opto with the particle (see Kühner-Gerth, o.c., 1, p. 230) are dealt with further on. - It is also significant that (J.v is often omitted in apodosi, on the one hand with verbs e:x:pressing obligation 01' propriety, on the other hand with those expressing possibility. Nor does it accompany the purely adhortative subj. For its absence from "deliberative" clauses containing a subj., see Kiihner­Gerth, o.c., 1, p. 223; W. G. Hale, in the Class. Review, 8, p. 27f. In the Homeric comparison the particle is rare (see Schwyzer-Debrunner, o.c., n, p. 313), probably because of the general, unconditioned character of the clause containing the ideal process.

3) For the combination of the particle and an hist. tense see also Kühner­Gerth, o.c., 1, p. 211ff.

4) Being concerned with the 'original' function of the particle we cannot dweU upon those innumerable passages where the (J.v phrase is stereotyped.

5) See J. van Leeuwen, Aristophanes Vespae2, Leiden 1909, p. 51, who, however, is not correct in saying that it is (J.v that indicates iteration. N 01' is it exact to hold, with Stein (adHdt. 2, 174), that "bei den iterativenFormen (in -aue) (J.v"eigentlich überflüssig ist".

6) The particle C(1n of course be left out; Hdt. 7, 119.

The modal particles 137

luáO''to'te); 11 79ff. etc. 1) ; from the use of the particle in sentences ex­pressing what might, or might have, occurred if certain conditions (which are specified or alluded to) were fulfilled: r 56 "nay, verily, the T. are cowards: otherwise you would before now have donned a coat of stone": ñ 'té uev fibr¡ I ... §O'O'o; Soph. O. R. 82; from such constructions as 1jJ 47 lboiJO'á ue 1}vfkOv láv1}r¡r; where a condition is implied in the parti­c~pl~ (cf. P 398); from ¿ 79 uai vv uev ... luófkr¡v ... aAAá fke ... and sImilar passages (A 630 etc.); from the • emphatic' double av in cases such as Aesch. Ag. 340 and Xen. Cyr. 1, 6, 18 2): in the latter passage av shows that the verb is to form a 'virtuaP apodosis. This particle is often found in sentences beginning with 'tw ('té[¡) "therefore, in this wise, then"; ev1}a "then"; enet'ta "thereupon" and similar deictic-anaphorical elements: v 273 "Zeus did not suffer it, else would we have ... ": T 61; ifJ 280; X 427; 11 539 3); (e 426;); P 319.

Let us for a moment dwell on av with the indicative of historical tenses. It is a well-known fact that the particle often occurs in the apo­dosis of conditional sentences which express what would be 01' would have been the case if a condition were or had been fulfilled: Thuc. 1,9 ovu tiv v~O'o)'V eu(!á7:8¿, el fk~ 'tt uai vavnuov elXev; b 178 4).

Special mention may be made of av accompanying, in contradistinction to a well-known rule to wit its omission in apodosi with verbs expressing obligation, propriety, probability, possibility, etc., i3be¿, x(!fív, uaAWr; elxev, elmr; i}v: it is addedifnot onlythefulfilmentoftheobligationetc., butthe obligation, propriety, necessity, possibility etc. is denied or presented as contrary to fact 5

): Thuc. 1, 74 el n(!oO'ex())(!~O'afkev n(!6'te(!ov 'té[¡ M1jbep ... ovbev tiv en ebet vfkéir; vavfkaxe¿v "so hiittet ihr nicht mehr zu schlagen gebraucht (weil ihr es namlich nicht gekonnt hattet)"; Plato, Rep. 328 C. On.ly in these cases-in which attention is focussed on the necessity, propnety etc. of the act, rather than on the act itself which is expressed by the infinitive-the verbs of obligation 01' propriety-which in Greek denote the process in an absolute, unconditioned manner-6), admit of

1) The interpretation given by Kühner-Gerth, O.C., 1, p. 211 "je nach den Umstanden, eventuell, wohl" etc. is not completely satisfactory.

2) See Goodwin, O.C., p. 73f. 3) N ot: "da tadelt etwa" = "hiitte getadelt": lvf}a uev .•. óv6aat1:o (Schwyzer­

Debrunner, o.c., n, p. 328). Similarly, P 398; P 70 where 1 doubt the correctness of their interpretation: "das ist wohl leicht am Wegtragen" = "hi1tte wegge­tragen": lvf}a ue éeía cpé(}ot.

4) 1 also refer to Liddell-Scott, S.V. (J.v AL The protasis is often understood: e.g. ¿j 421: see e.g. Kühner-Gerth, o.c., 1, p. 212. The particle may be omitted: "Del' Redende nimmt dann auf das tatsachliche Verhaltnis keine Rücksicht ... " (Kühner-Gerth, O.C., 1, p. 215 where particulars may be found).

5) For det~ils: Kühner-Gerth, o.c., 1, p. 206, A. 2; Schwyzer-Debrunner, o.c., n, p.309; Goodwin, O.C., § 421ff.

6) For these verbs in, Greek and Latin see also Wackernagel, Vorl. Synt., 1, p.227. ,

Page 73: Gonda.1956.The Character of the Indo-European Moods.pdf

140 The Indo-European moods

Oó¡;, q;íAOC; . .. 8yW fJé ué as UAsíw "give me, friend ... so would 1 make your fame known"; cf. also r 417. The combination subj. + us also occurs after a protasis having sl + opt.: A 387; ~ 388f. (cf. 391); cf. also Q 655 uat usv "and so". Similar remarks may apply to E 308 aVTr¡v \ aT~aolJ,a¿, 1) us q;éerJa¿ p,éya ueáTO¡;, 1) us q;seoíp,r¡v. The conditions on which the clause containing the particle depends may be implied: e 354; a 396; 15 692; or they are not expressed:A 433. It would appear to me that the same value can be attributed to the particle in relative clauses. See e. g. o 311; u 539; a 86 nép/lpw as ... \sl¡;'E ... \ 8¡; u' ano Ifiva Táp,rJa¿l); (J 213; n 349 etc. after an imperative 2); cf. also ¿ 356; 1424; e 385; after an infin. Q 119 etc. "to bring him gifts which will (then, i. e. ifthey have reached him) make him glad"3). In Homer the particle is commonIy used when, in more or less vivid style, the relative clause does not express a general thought: 1397 Táwv 1}v u' UHAWp,¿ q;íAr¡V nO¿17aoP,' auomv. In utterances of a general import such as the following the poet prefers the simple relative: v 214 "Zeus watches over all men, and punishes him who sins" (8¡; n¡; tÍp,áeTrJ); '188; E 407; cf. also A 163. If however the verb of the antecedent clause denotes a repeated 01' customary action or a general truth, while the relative clause refers in a general way to any act or acts of a given class 4), the particle is, as might be expected, added: e. g. 1312 5).

The conditions are implicit or left unexpressed: B 231; K 235 etc. etc. Cf. also ¿ 16 "1 shall tell you my name, that you may know it and that

1) In this construction too other languages may often be preferable, or no translation at aH, the weak and vague "then" bJending into something like the German "wohl", a weak "opinor" or into no translatable sense.

2) I cannot subscribe to Delbrück's (Conj. U. Opt., p.42) interpretation of H 171 "A1Íeq> vvv nsnciAstJf}s (j¡afmseé~, o~ "S AcixntJw " ••• da wird es ja wohl (Er­wartung) erhalten".

3) {J 43 not "die will ich euch wohl sagen" (Delbrück, O. c., p. 41), but dependent on the following O-,;S + opto

4) Cf. Goodwin, o.C., § 532; Chantraine, o.c., H, § 362ff. 6) Mouro, Gramm. Hom. dial., p. 202ff. enumerated a large number of Homeric

passages in which the particle accompanies the subj. in a general statement (see also Chantraine, o.c., H, p. 247); however, in Z 228 the rel. clause refers to un­specified individuals which out of the definite number mentioned in the preceding line, are under certain circumstances offered to be killed by the godo In r 279 not all men in the world below are punished, but only those which-in cases similar to those alluded to in the situation given in the context-have sworn a false oath: so the character of the sentence in its entirety may be general, the relative clause admits of a limitative particle. Here the sense of "S seems to verge on that of a weak "at least" (to be inserted, in translating, before the relative pronoun). Cf. also II 621; P 99. Or the force may be a very weak "if such be the case, when occasion arises, on aH occasions, should the case arise, in each case, anyhow": cP 24 f-lciAa ycie Te "a-';8IJf}{st 01' "S Aci{JntJtv; 484; S 416;!J 335; (j 196; r¡ 33; f} 32;" 22; 74; 328 etc.; cf. also T 228. Other passages may give us the impression that "S can be a means of mitigating the absolute character of a statement, cf. e.g. T 167.

The modal particles 141

1 (this being the case) hereaftermay be your host": tJq;ea ual v¡te'i¡; I eYbs7:' 8yW 15' tlv . .. I v¡ttV ~StVO¡; sw. ' . The particle accompanying BW¡; is already found in Homer, though lt may be absent (e. g. N 141, in a comparison)1). That it is longer in joining neív, nelv 1), n(!67:8eOV 1)2) may perhaps be due to the original character of the clauses introduced by these words with the limitation conveyed by the particle. Are there similar reasons (incompatibility) for the remarkable fact that the final ¿va is in contradistinction to the final w¡; and 8nw¡; also in post-homeric usage, never accompanied by av1: in connection with av and the subj. the local adverbial rela­tives 3) mean "wherever" 4), an indefinite nuance unfit for these final clauses 5).

As is well known Homer, in the cases where Attic usage requires Báv + subj., has also sl without us or av 6 ). Both constructions differ in that, broadly speaking, the latter is decidedly preferred in general suppositions ; here, Pindar has it always. Cf. K 226 ¡t0VVO¡; ~' sr née TS vo~ar/ "if one l " A ' a one . ..; 81; 262; a 167 (n¡;); r¡ 204; ¡t 96 (sr noih) etc.; Pind. O. 6, 11

nOAAo¿ ~8 ¡té¡tvavTa¿, uaAOv sr n nova{}fí and hence also in comparisons and similes: II263 sr nse ... Tí¡; ... av1}ewno¡;; A 116 ("a hind"), f/J 576 etc. The few instances of the simple sl in the Attic poets likewise concern general conditions 7). The particle is on the other hand very often added already in Homer, when the content of the clause bears upon a parti­cular caseS): A 137 sl ~é us ¡ti¡ bÓJwaw i. e. "if the Achaeans do not give me an equal recompense"; B 364 el ~é usv &¡; se~rJ¡; ... yVÓJar/ "if you do thus ... you will know"; ~ 391 "he will tell you", ar u' 8{}éArJa{}a; E 306; r 281 "if A. slay M. (sl . .. usv ... uaTanéq;vr/) then let him keep H."; a 279 "1 will give you counsel," ar us ní{}r¡a¿; cf. also A 580; Ll 98; H 118;

1) See e.g. Kühner-Gerth, o.c., H, p.449, A.4; Schwyzer-Debrunner, o.c., H, p. 650; Chantraine, o.c., Ir, p. 261; Humbert, 0.c.2, p. 212 (§ 347).

2) See Kühner-Gerth, O. c., Ir, p. 455, A. 1; Schwyzer-Debrunner, o. c., H, p. 313; Stahl, o. c., p. 294ff.

3) For the original function of ¿ya ("where" > "in order that") see M. Leumann, in the Mélanges de Phil., de litt. et d'hist. anc. offerts a A. Ernout, Paris 1940, p.235. " 4) ,?f. Eu;-. !on 315 lJ.nav f}sov f-lOt OWf-l', tI" (ly Aci{Jn f-l' vmo,; Thuc. 2, 11 [nstJ{}s onn ay Tt, r¡yr¡-,;at.

6) The statements on the use of the opto + ay in subordinate clauses made by Wackernagel, VorI. Synt., I, p. 238 should be completed by consulting the data to be found, for instance, in Kühner-Gerth, O. c., vol. Ir (see p. 628).

6) See e.g. Kühner-Gerth, o.c., H, 2, p. 474; Goodwin, O.C., § 444; 453; 468 etc. 7) I refer to Goodwin, o.c., § 471. B) Although for practical reasons "S, al' joining sl, ensl, lfrpea etc. be distinguished

from other use of the same elements (Goodwin, o.c., § 192), their functions are fundamentaHy one.-In si "S with the opto which is sometimes found in Homer the particle lili:~wise has a similar function: cf. Goodwin, O.C., § 460; Chantraine, o.c., Ir, § 408 (lt would perhaps be preferable to regard SI "81' in cases such as Z 50 as instances of "persevel1ation").

Page 74: Gonda.1956.The Character of the Indo-European Moods.pdf

142 The Indo-European moods

1358; 412; 604; K55; 105; a287; fJ 188; o 391; o 312; e 549 etc. etc.!). In a small minority of cases Homer does not add the particle although the passages are (01' may be) similar: X 86; P, 348 etc. This idiom survives in a few passages of the Attic drama: Soph. O. C. 1442 etc. 2). They show that the addition of the particle originally was not obligatory. Its earIy occurrence in <general suppositions' on the other hand-A 166; M 302; A 159-shows that it soon spread beyond its original domain 3). This development may to a certain extent be ascribed to the more vivid and distinct character of the sl clauses with the particle,-vivid construc­tions tending to be automatically reproduced-, to a tendency to differ­entiate between the contingent sl + opto and the construction under consideration4 ) and to a preference for an uniform al aY, eáy + subj. construction.

Sometimes us is repeated in more 01' less emotional passages, in which one reference to the conditions and circumstances does not suffice ("Perseveration") 5): cf. 1359; also when a double reference, in protasis and apodosis is required: A 137; (opt.) E 273; 1141.

Returning to the opto it may be remembered that the potential with the particle occurs, in early Greek, side by side with the unqualified potential: L1 17 ff. sl o' aiS nw(; 't60s niicn cptAov ual f¡ov yéyot7:o, I i} 'tOt p,eY oluéott'o n6At(; "if this might in any way be welcome ... to all, then might the city still be inhabited ... ": E 3Il uat vv USY 8Y{Y an6AOtt'0 ... , Isl p,i¡ . .. "and now would he have perished, had not ... ". In the latter sentence the condition expressed in the subordinate clause is alluded to in the

1) See also Chantraine, o.c., JI, § 411ff., whom I cannot follow in some parti­culars: L1 415 e.g. is not, to my mind, "si un jour les A. détruisent les T.", but "if (then, when Ag. will be in command) ... "; fJ 218 "if (then)"; in "meme si tu t'en vas ... " rendering e 477 there is nothing to suggest the value of Ue which may refer to the eventuality mentioned 466; as to r 281ff., I am not convinced that "e ••• Ue "souligne l'opposition entre deux hypotheses" (el ftSY ••• el os ... ); the particle, here also, refers to the circumstances implied in the situation, which continue to be present to the speaker's mind; hence the repetition (cf. also 288 el o' ay .. . ); cf. also X 99ff.; A 110ff. etc. "if (then, there) ... ". Why should we expect the simple el in A 391, which is no "general supposition"? ~ 273 the idea of "au contraire" does not seem to be expressed by (11', nor does it bear an emphatical force in E 224 (for el: nel2 see J. D. Denuiston, The Greek particles2, Oxford 1954, p. 487 f.; cf. also E 232); r 25 el: nel2 "even if" (Denniston, O. c., p. 488); cf. el: nel2 (11' + opto B 597).

2) For particulars: Goodwin, O.c., § 454. The interpretation of some passages quoted by this author may be a matter of opinion; for instance A 340 el: n07:e may be taken as a general 'supposition', without reference to the particular circum­stances mentioned in the preceding part of the contexto

3) A similar remark may apply to the phrase rel. prono + parto above mentioned. In some cases the particle is absent where it might have occurred (e.g. 1 554), elsewhere Homer has it in the cases specified by Goodwin, o.c., § 542.

4) For the character of this construction see chapter IX and Humbert, 0.c. 2, § 357.

6) See also Liddell and Scott, S. V. (11' D, JI, r.

The modal particles 143

pr~nc~pal clause; in both ir:stances the speaker visualizes the process of the prmCIpal c1ause as a contmgency. 1) Although the condition 01' limitation does not necessarily require the partic1e, the difference between both constructions is not seldom perfectly evident and consistent with the aboye distinction: A 838 mv(; t" al!' 80t 'táos 8eya; "how may these things be" (no condition); ; 122; 407; T 321 "nothing more grievous than this could 1 suffer" (n&{}otp,t); 331 (1 hoped that you would return), "that so (ay) you might take (e;ayáyot(;) my child"; cf. also K 2uf. "if somebody went among the Trojans, he might learn (us 01' 'ts ... nv#ot'to); and he might come back (8A#Ot) unscathed; in that case (usy) his fame would be (Sl'/}) great".

In combating the opinion fostered by Stah1 2) and other scholars that t~e potential without 0.1' is no correct Attic Slotty was generally speaking nght: the construction survives in all literaty dialects and cannot be regarded as a reminiscence of the Homeric style. It may be added that there are also many post-Homeric instances in which the absence of the particle is (on the strength of the original function attributed to it earlier in this chapter) only to be expected:such exclamations, emo­tive 01' exclamative questions as Aesch. Ch. 595 aybeo(; cpe6y'/}p,a 'tt(; UYOt; (cf. Soph. Ant. 605; Plato Gorg. 492B and T 321; f[> 274); such general observations as Eur. 1. A. 1210 ovbsl(; neO(; t'áb' ayt'stnot 01' general questions as Plato, Euthyd. 286B o be p,i¡ UywY np Aéyoyt't nw(; aynUyot; 01' such g~ner~l subordinate c1auses as Soph. O. R. 979 sluñ Ueát'Ul'tOY Cfjv 8nw(; OVyatW t't(; (cf. also Aesch. PI'. 292 'X,wet(; t's yéyov(; ovu 8<1ny 8t'q> ftStCovu p,o'ieuv ystp,utp,' ~ <1oC) do not convey any limitation 01' qualification. Unconditioned potentiality is also expressed in sentences like Plato Lys. 214 D (] be uv'to uvnp ayóp,owy Sl'/) UUt btácpoeoy, <1'X,oAñ ys 'tq> aAAq> op,o'iOY I} cptAOY yéyott'o. In other cases an explicit qualification need not be emphasi­zed: Soph. AL 921 (exclamation); Plato, Euthyd. 291 E; Xen. An. 2, 5, 14

, s' I '] I t: Ji): , , , st uS n(; <1S AVnOt'/}, W(; uS<1not''/}(; UVU<1t'e scpow. The association of the cupitive opto with the particles 0.1'3) 01' us has

often been denied, suppressed, 01' hesitatingly admitted 4). However, it cannot be a matter for doubt that in cases like 1 .G. V, 2,343, 46 (Arcadia) ovb' (11' avt<1t'uCp,UY anv t'o'i(; E. "nor shall (will) 1 detach myself from the E." (pro~ise on oath)5) the opto is no common potentialis. Slotty who

1) For particulars See Schwyzer-Debrunuer, o.c., JI, p. 324f.; F. Slotty, Del' Gebrauch des Konj. und Opto in den griech. Dialekten, Gottingen 1915, p. 82ff.; p.140ff. .

2) J. M. Stahl, Kritisch-historische Syntax des griechischen Verbums del' klassischen Zeit, Heidelberg 1907, p.298ff.

3) Miss Hahn is right in observing (o. c., p. 101) that there is no such clear-cut an~ invariable ~istinction between the cup. opto and the poto as in Attic Greek.

) See e.g. kuhner-Gerth, O.C., I, p. 230; Chantraine, Gramm. hom., JI, p.218; J. ~~bert, Syntaxe grecque2

, Paris 1954, p.119. Cf., in general, Slotty, o.c., p. 93ff. ) Dlscussed by Slotty, O. c., p. 94.

Page 75: Gonda.1956.The Character of the Indo-European Moods.pdf

144 The lndo-European moods

at the time amply discussed these constructions-no doubt exaggerating their number and neglecting certain shades oí meaning of the pot. + av-, though considering them as voluntative, did not go into the problem oí the sense of the particle. Could there be room íor the suggestion that this in accordance with its general function bears the force of a reference to a situation implying an: "under, 01' on, certain conditions " , i. e. "possibly, if, and as fal' as possible, within the limits oí possi­bility" 1 The addition of the particle, though apparently limiting the action, might perhaps have conveyed the implication oí "by any possible means, to the best of my ability"; although a simple "then" (01', rather, the Dutch "dan") would probably be a better translation. May such wishes as Z 28lf. W~ ?-e8 oí a'Úih I yaía Xávot "would that the earth might straightway gape for him!" also be regarded as implying a condition: "if such were possible" 1 W ould it, further, be warranted to suggest inter­preting passages such as cp 113 in the same way1: (don't turn away from the drawing of the bow ... ) ?-eal b8 ?-e8V aV'id~ syw 'iOV 'i6~ov 7I:8t(!r¡uatp,r¡v "yea, and 1 would myself (?-e8V suggesting: «if the occasion arises» 01'

«possibly» 01' «then; indeed») make trial oí that bow". Mention may in this connection also be made of Hdt. 4, 97; Bacch. 5, 165 "is there an unwedded girl ~ If so, 1 would willingly wed her": 'iáv ?-e8V . .. S~éA(¡)V 1}efp,av a?-eot'iw. This function is essentially identical with that of av, ?-e8

when accompanying a so-called potentialis: cí. e. g. also y 365 l!v1}a ?-eS A8~atp,r¡v ?-eotAn 7I:aea vr¡l "da will ich mich dann legen" (Faesi-Hinrichs), i. e. under the circumstances and conditions specified and implied in the preceding lines (cf. the future in 359 and elp,t in 361 and 367)1). Odysseus, incognito, unfolds his plans (o 307ff.): (1 am minded to go forth to the city (AtAaCOp,at) .. . ; give me good counsel (imper.) ... 1 will wander (fut.) through the city) ?-eal?-e' sA1}WV 7I:ed~ brhp,a'í' 'Ot5v(j(jijo~ ... 1 ayy8Atr¡V 8L7I:OtlJ,t ... JIr¡V8A07l:8tn "and 1 would (if the opportunity presents itselí) go to the house of O. and bear tidings to P." (similarly, 315; 317). Cf. also Q 370 "1 will nowise harm (fut.) you, 1 will even (if such a case should arise: ?-C8) defend (opt.) you"; I 417; h. Ap. 51; b 348; 637; 0449 (cf. 448 OLU(¡); after 'irp "then, in that case": o 181.

In v 326 the construction is followed by an 8l clause, containing an explicit < condition'. The particle appears in connection with the "delibe­rative" optative: 1} 570 'ia <,)é ?-eBV 1}8d~ .¡¡ 'iSAéuBt8V I .¡¡ ... "and these things the god will haply bring to pass" (Murray). Cases are not wanting in which the function of the opto is doubtful (cf. e. g. Y 427; v 390; X 325). Here too it would be best not to emphasize differences which, though becoming pronounced in the course of time, can hardly have been fully developed at an early periodo

1) Of. 't" 598 (see e. g. Faesi-Hinrichs: reminiseence of y 365): is a faint suggestion of "as well as l, under the circumstances can", accompanied by a shrug or similar gesture?

The modal particles 145

If these observations are in the main apposite, and the examples which could be multiplied, convincing, they may fortify us in the opinio~ that the functions of moods and that of the particle originally were different.

The particle av has early and often been identified with the Lat. an1) and the Goth. ano About the question whether the Greek use or that prevailing in the other languages was more original scholars were, however, not unanimous. Leo Meyer 2), considering the Latin function in the <Doppelfrage> to have been the most ancient, fostered the opinion that the word originally meant "or" (Germ. "oder, sonst, andernfalls"); from this the meaning "moglicherweise, unter gewissen Umstanden, vielleicht, etwa" would have arisen in such passages as JI 687 Bl . .• rpvAa~Bv, 1] 'i' avv7I:é?-erpvyB, which he explained as follows: "Er bewahrte ... nicht, oder (= sonst) er entging (= «ware el' entgangen»)". This inter­pretation as well as the comment added to it were deservedly rejected by Brugmann 3), who rightly observed that Meyer attributed to &/1 a force which was properly inherent in other components of the sentence. He preferred to regard the Latin sense of "or" as younger and to assume an original meaning "moglicherweise, eventuell, vielleicht, etwa". Remainders oí this use in Latin are, perhaps: ¡orsan, ¡o1'sitan < 101's sit an; eho an vieimus? "sag, haben wir vielleicht (etwa) gesiegt1"; haud seis, an ita sit "ich weiB nicht, vielleicht (eventuell) mag es so sein". "In das an der Doppelfrage (dieesne mihi, an taeebis?) zog dann der Sinn «oder» erst dadurch ein, da13 der Satz oder das Satzglied mit an mit Rücksicht auf eine vorausgehende Frage gegensatzlichen Inhalts hin­gestellt ... mit ihr zu einem Satzganzen ... verschmolz 4)." In the opinion of other scholars "wohl, ja, denn" (Hofmann and others 5», 01'

"*an servant a affirmer" (Meillet-Ernout) must have been the original force. Wackernagel 6

) was no doubt right in emphasizing that the Greek and Gothic 7) uses of av and an are no impediment to identifying them. In Luke 10, 29 an hwas ist mis nehwundja? (?-eal 'it~ SU'itv p,OV 7I:Ar¡utoV;) an no doubt serves to introduce an emphatical question replying to,

1) In addition to the publications to be mentioned in the following lines, refer­ence may be made to Waekernagel, Vorl. Synt., l, p. 222f.; (Leumann-)Hofmann, 0.e.5

, p. 65lf.; Hofma1l11, Lat. etym. Wtb.3, l, p. 44; Ernout-Meillet, Diet. etym.3, l, p. 55 (the author last mentioned, like Hj. Frisk, Griech. etym. Wtb., Heidelberg 1954, p. 100, rightly agrees in combating Wackernagel's adherenee to the opinions expressed by Ebel and Skutsch: an < anne < *at-ne).

2) Leo Meyer, An im Griech., Lat. und Got., Berlin 1880, esp. p. 11 ff. 3) K. Brugma1l11, in the Literarisches Centralblatt, Leipzig 1880, 1668ff. 4) The paper by G. H. Mueller, Hermes 1890, p.463f. may be passed over in

silence. 5) Hofma1l11, in Lat. Gramm., l. C. 6) Wackernagel, l.c. 7) See Meyer, o.c." p. 9ff.; Streitberg, Got. Elem.5, p.219.

10 Gonda, Indo·Europ~an moods

Page 76: Gonda.1956.The Character of the Indo-European Moods.pdf

IU_lbl ••••• 7.11 ____ •• _-_i1 .. _IIII __ tA_-------,...".,--~~----~---~~

146 The Indo -European moods

01' resuming the words of the partner. Similarly 18,26 ("dan" ~n. th~ Dutch translation)' John 9,36; cf. also Luke 3,10 an hwa tauJa~ma. (-tÍ o15v 'J'tot~f1o)¡.tsv) 'in Dutch: "wat zullen wij dan doen 1"; without an interrogative adv~rb: John 18,37 an nuh piudans is pu (ovuovv (3aatAS~~ 131 f115;), in Dutch: "Zyt Gij dan een koning1" The analogy between th~s construction and certain cases in Greek, when taken_ by themselves, I~ clearly discernible: Xen. R. 1, 9 si yae OV'io) 'iav'í' sixs, 'J'tw~ (lv 'J'tOAAOt ¡.tF:v s'J'ts-&15¡.tovv 'iVeavveiv ... ;; Plato, Gorg. 514 E; Dem. 50, 67.

The closest parallel in Latin seems to be the rather frequent construc­tions represented by Plaut. Most.333 quo ego eam?:: an nescis?1);

, . 2 " d 't Amph. 745 an etiam id tu scis?; Asin. 837 an tu ... putas. y?U o~ 2

think ... "; Poen. 991; 1066; Ter. Eun. 382ff., by Asin. 524 qu~d ded~t. quid ad nos iussit deportari? an tu tibi \ ve1:ba blanda esse a~rum rere, where the question with an refers to somethmg the speaker hlmself ha~ said and a Dutch translation might require "dan"; by Amph. 456 ub~ ego formam perdidi? \ an egomet me illic reliqui ... ?; 661 (and hence also the nuance represented by Yergo Buc. 9,1); by Pla~~. Me~ .. 962 "1 am perfectly sound. Can it be they are insane ~hemselves? .: an ~ll~ ... ipsi insaniunt?; Rol'. Sato 2, 7,109. The. ?ipart:t~ const~uctlOn ~f the 'Wahlfrage' can be explained from an ol'lgmal VIVId self-mterruptlO~ of the speaker: the starting-point of the type Plaut. Mere. 90~, sed ~U~ ego istuc c1'edam? vidistin an de audito nuntias may have been: ... dld you see her ... is your news then mere hearsay 1" 2); cf. also Cure. 589 maneam an abeam?3). Rence also the type Rud. 104 utrum tu mas~e an femina es ?4) A similar original function ~ay be assum~d ~?r an I~ indirect questions: Plaut. Epid.· 543 haud scw an congredtar 1 don t know whether to advance upon him" < "1 don't know: am 1 (then) to a ... 1" 01' " ••• shall 1 (then) a •.. 1"5); cases without.t~e introduc-tory particle are not wanting: Truc. 779 volo sGÍre fa!eamtm. .

So 1 would conclude that, despite the divergent roles o~ the partlCle in the three languages, its <original force' can be traced Wlth some cer­tainty. The particle *an seems to have vagu~ly ~enot~d:, not so much "Unbestimmtheit" (Rofmann), "UnzuverslChthchkelt (Schwyzer-

1) "an niromt stets auf etwas Vorausgegangenes Rücksicht" (R. Klotz, Hdwtb. d. lato Spr., I (1866), p. 380); "altera persona alteram interrumpit loquentem vel respondet indicans se non capere vel non credere vel improbare quae alt,:ra modo dixit: dU' f¡" (Thes. Ling. Lat. S. V. an I a, where an abundance of mstances may be found).

2) See also (Leumann-)Hofmann, o.c., p.650ff. ".. 3) Cic. Verr. 5, 109 cum homine ... res est an cum jera? ( vlVaClté de la langue

familiere", Ernout-Thomas, O.C., p.160). 4) The double questions assume a great variety of forms: see Bennett, O. C'i I,

p. 332f. ? . h 5) (Leumam'l-)Hofmann, O.C., p. 696: "son ich wohl an ihn herantreten 10

weiJ3 nicht recht".

The modal particles 147

Debrunner1), 01' "affirmation" (Ernout-Meillet, Musi( 2)), but reference to a preceding part of the discourse, whether expressed 01' implied, or to the situation. This function could also involve more or less emphatic resumption (Goth., Lat.), even to give lively expression to a feeling of surprise, incredulity, or remonstrance-, 01' more 01' less distinct restric­tion or limitation (Gr.). Anyhow, the particle could serve to detract from the absolute character of an utterance, to diminish its objective certainty, or to throw doubt upon it (Lat., Gr.).

Excursus 1

In contradistinction to Cobet and other scholars who denied the possibility of av + (ind.) fut. in post-Romeric Greek, this construction has, some years ago, again be defended by A. C. Moorhouse (Class. Quart. 40, 1946, p. lff.), who rightly starts from the consideration that "if it is possible to show that a special meaning can be attached to the construction, we shall have greatly increased the chances of its admission in our texts". While explainingA 139; Y 523; .1176; 1167 by the suppos­ition that, here too, av limits the statement in the aboye way (A 139 ("given the condition thatA. chooses for himselfthe prize rightly belong­ing to one of the others"), he-rightly too, as it would appear-holds that elsewhere the condition on the fulfilment of which the validity of the statement depends is left unmentioned. Cf. P 514f. uSV • •• ¡.¿eAr)f1SL "it will (haply) be the care of Zeus".

Excursus 11

As to the subj. + us in conditional clauses it is difficult to be convinced of the truth of the thesis recently upheld by Chantraine (o. c., n, § 411) that the particle emphasizes "a particular case", unless this phrase is explained by "alors, dans ces conditions" (p. 211). The force of us may be regarded as similar to that of us and av when accompanying an optative. In {3 218, which may be adduced in illustration, the addition, in translating; of "le jour (ou j'apprendrai)" seems to be quite unneces­sary: sl ¡.tBV usv 'J'ta'ied~ (3tO'iov . •. auo15f1O): Telemachus after declaring that he will go to Pylos to gather information about his father continues: "if, then, (if, in carrying out this plan) 1 shall hear that he is alive ... ". Similarly 1359 aZ u' S{}BAllf1{}a "(supposing in that case you would)" i. e. "if so be you will". Often, however, a weak "then" or "now" seems to be the strongest translation possible, cf. A 404; g 477. In instances such as r 281ff. si ¡.tBV usv ••• si éjé us the opposition is, as far as 1 am able

1) Schwyzer-DebrUilller, o.c., n, p.558. 2) A. Musió, in the Rad of the ,Zagreb Academy, 237 (1929), p. 194f.

10·

Page 77: Gonda.1956.The Character of the Indo-European Moods.pdf

..... r

148 The Indo-European moods / The modal particles

to see, expressed by /ú/v ... M, not by the repeated us. Passages with av like@ 556-also referring to special occurrences-may be judged in the same way. Moorhouse, in discussing (in the Olassical Quarterly 40, 1946, p. 1ff.) the phrases consisting of av + fut. was no doubt right in observing that the condition, on which the statement depends, is not always a definite line of action, clearly stated in, or implied by, the contexto It might, indeed, have been resorted to for reasons of politeness, implying "if you don't object" or a similar "condition".

In another part of his book the above-mentioned French author upholds the view that us and av "soulignent un cas particulier (et), marquent une emphase". In rendering the idea conveyed by these words by "alors" (in passages such as A 137 etc.) he is no doubt right, but apart from those cases in which they are used in both the principal and the subordinate clause I fail to see any 'emphasis' in them.

I

IX. The conjunction el and the conditional sentences

In perusing the recently published book on the Homeric sl clauses by Tabachovitz 1 ) we are unable toescape the conviction that his criticism on certain almost traditional views held by many of his predecessors is a greater success than the theory on the original character of this con­struction for the defence of which he has written this dissertation. The verdicts given on this point by the authors of the modern handbooks are indeed not very convincing. According to Meillet-Vendryes 2), for instance, the main source of the conditional sl construction must have been a combination of utterances such as is represented by H 28 aAA' s;; ¡.tot Tl 'ltt{}OLO, .. 6 usv 'ltOAV usel5LOv dr¡ which originally meant: "puisses-tu me croire! cela voudrait mieux" but in the course oí time coalesced into: "si tu pouvais me croire, cela vaudrait mieux." Schwyzer-Debrunner are of the same opinion 3): "Sprachgeschichtlich gehort zum Kupitiv auch ein Teil der Optative der Konditionalsatze", and Oh. Bally4), quoting the same line H 28, even arrives at the conclusion that "la plupart des tours prépositionnels ou conjonctionnels (cela revient au meme) sont nés par condensation de deux coordonnées, et plusieurs ont conservé la forme segmentée". The 'paratactic' origin of the conditional construction is likewise upheld by the authors of the two very recent French books on Greek syntax 5). This theory and even the much quotedexample H 28 go back to Lange's ample discussion of the use of el in Homer, published in 18746). In arguing that Lange's views must be judged in connection with the opinions prevailing at that period, Tabachovitz 7) is in the main right. As far as Homer is concerned, his conclusion that an 'originar combination of two main clauses does not appear in the texts, is largely justified. His criticism with regard to the aboye "Schul­beispiel", which is usually wrenched out of its context is, in general, convincing: it can refer to what follows. It might be added that at least part ofthe considerations leading modern authors to subscribe to Lange's

1) D. Tabachovitz, Homerische el-Satze, Lund 1951. 2) A. Meillet-.J. Vendryes, Traité de grammaire comp.2, § 946. 3) E. Schwyzer-A. Debrunner, Griechische Grammatik, n, p. 323. 4) Ch. Bally, Linguistique générale et linguistique fran<;aise2, Berne 1944, p. 67. 6) P. Chantraine, Grammaire homérique, n, Paris 1953, p.275; .J. Humbert,

Syntaxe grecque2, l;'aris 1954, p.219. 6) L. Lange, Der homerische Gebrauch der Pal'tikel el, Abh. phil. hist. Kl.,

Sachs. Ges. d. Wiss., Leipzig, vol. 16. 7) Tabachovitz, O. C., p. 9. '

Page 78: Gonda.1956.The Character of the Indo-European Moods.pdf

.. p 1 ...

The Indo-European moods

~ly stand criticism; such well-known a priori reasonings ; "Griechische Grammatik" 1), can only be substantiated

éthno-psychological' examination of the relevant facts: /J.~ dem Temporal-, Final-, Befürchtungssatz steht del' Bedin­

.. Gz im engern Sinn auf einer andern Ebene; jene wurzeln schon cln'lÍtivsten Lebensverhaltnissen, die erzahlend wiedergegeben werden

/K6~nen, dieser gehort zunachst zum Handel und W andel und .z~m Recht" ; ./ curiously enough, the observation is added that the condltlOnal clause

occurs in many languages all ayer the world. But in reversing Lange's line of reasoning and in defending the oppo­

site view Tabachovitz seems to have defeated his own object. In his opinion the complete bipartite construction must be considered the histori­cal (01' prehistoric) starting-point and the 'interjectionaf use of el with an imperative to express a wish originated in an ellipsis. In defending this thesis he continually manifests his dissatisfaction with, nay his dislike of, comparative methods, preferring to explain the phenomena at issue psychologically and stylistically. In so doing he overlooks ~he famous 'distinction' between speech and language. The person speakmg 01' the author writing makes his selection from the possibilities placed at his disposal by the language as it is spoken (01' used in poetic tradition) in the milieus with which he comes into contacto We may attempt to account for his choice by stylistic methods, inter alía by detecting psychological motives. But in the language which that person uses, there are norms, there exists a usage which cannot be transgressed without running the risk oí being unintelligible 01' ridiculous. These norms are no matter for psychological interpretation oí particular

texts 2).

In endeavouring to establish his theory Tabachovitz 3) too often detects a "Nebenbedeutung des Wunsches" in conditional sentences, maintaining that this, like the nuance of exhortation etc., is a later acquisition. 1 for one am unable to follow him for instance in discerning a "kupitiver Nebensinn" in such passages as II 618 where si + indic. is said to intimate that "die nackte Wirklichkeit unseren Wünschen und Hoff­nungen so oft zuwiderlauít". Of course the last words oí the sentence ... 'táxa UBV as ud oexr¡(J7:f¡v 'TeSe B6v'ta I l!yxo~ eptJv· ua'tB'TeaVaS ... , sr a' l!PaA6v 'TeSe admit of the translation: "hatt' ich getroffen", but 1 would preíer considering l!paAov to express that the process, being contrary to fact, exists only in the speaker's imagination 4): a bald fact, and si to ~~t a case. Aeneas, speaking emotionally says that-if oí course condItlOns and circumstances had permitted him to do so (usv)-he probably ('táXa) would have made Meriones cease dancing, if he had struck him. Whether

1) Sohwyzer-Debrmmer, 0.0., p.682. 2) 1 also refer to Chantraine, Revue de philologie 1953, p. 215. 3) Tabaohovitz, 0.0., p.47. 4) See further on, p.181ff.

The oonditional sentenoes 151

he denied killing Meriones is left unstated. Similar more 01' less emotional passages are ¿¡ 732; 7jJ 21; ro 284. e 366ff., where Athene likewise visualizes a;n occurrence which might have occurred in the past as actual; the situa­tlOn can, however, be regarded as implying a wish 1).

We might, íor the present, connect the aboye and similar cases of a past indicative with the tendency to express oneself, in an emotional manner, by means oí actual facts instead of possibilities 01' expectations 2).

Not infrequently the indicative clause contains a fact which has actually occurred: E 454 "they would have sacked the city, ií Apollo had not (si ftf¡) slain Patroclus"; does Thetis who is speaking here really express a wish1

The same cupitive nuance is admitted by Tabachovitz in el clauses containing the subjunctive. Whilst denying the existence oí the prospec­tive subj., he does not tell us what is the function of this mood in these clauses. Its use, he says, is "formelhaft": H 287 ar UBV ftOl own Zev~ ..• + inf. is substantially identical with A 128; II 87; E 260. But the question remains why the verb oí this stereotype phrase is in the subj. Besides, verbs for giving occurring in el clauses are not infrequently in the opto : 1379; l 229; 317; ~ 132; X 61. The difference seems to be similar to that existing between these moods in other clauses. In these el + opto clauses the processis contingent; it may even be imaginary; the apodosis contains an opt., 01' the el clause expresses a subordinate thought (l 317). InA 128 etc. the main clause often contains a future 01' expresses a com­mand; in 1363 a cautious opto follows el 08 uev . .. own. These remarks are of course not to deny that here too an el + subj. clause can sometimes imply a wish: Z 309 ar u' eAe17an~.

The use of el in dependent interrogative clauses ("whether") has been explained as originating in the cupitive 01' conditional el 3 ). Tabacho­vitz, without applying the comparative method to which he is averse, pronounces the opinion that the conditional and the interrogative el clause are, both oí them, ancient, because they occur in Homer, and because they cannot be shown to have originated in a main clause 4).

In attempting to demonstrate that the interrogative el clause has arisen from the conditional, he draws attention to a frequent interrogative nuance oí the latter, which being a protasis, often has the form of an interrogation (willst du es tun, 80 beeile dich). This order of words in the

1) See also Kühner-Gerth, Ausf. Gramm. d. griech. Sprache, n, 1, p. 207, Anm. 3. .2) Cf .. e.g. also L. Spitzer, Jahrbuoh f(ir Philologie, n, Münohen 1927, p. 270ff.;

Stllstudwn, I, Münohen 1928, p.258ff. See also W. Havers, Handbuch der er­klarenden Synta:x, Heidelberg 1931, p. 41ff. and 219f.

~) See e.g. D.elbrü.~k, Vergl. Syntax;, nI, p.414; K. Brugmann-A. Thumb, Grlech. Grammatll~, Munohen 1913, p. 617 (two sources: after verbs of attempting, and after verbs of saying); Schwyzer-Debrunner, O. C., n, p. 687 « the oonditional el and < the cupitive sl). Other authors leave the point undiscussed.

4) Tabachovitz, o.c., p. lll.

Page 79: Gonda.1956.The Character of the Indo-European Moods.pdf

152 The Indo-European moods

conditional protasis, which is indeed in keeping with a frequent con­struction of the interrogation, does however not prove that the expression of a condition has generally developed from an interrogation 1 ). As 1 have shown elsewhere it is not possible to demonstrate that the condi­tional construction ofthetype 'willstdu es tun, so beeile dich' is of a secon­dary character with regard to the construction with a conjunction 2). Besides, Tabachovitz does not consider the preponderatingly antecedent (prior) position of the cond. clause and the posterior position of the indirect question.

So it does not appear to me to be a sound method to found the dis­cussion of the Homeric el clauses on the small basis of a psychological and stylistical interpretation of the epics alone. As long as there is any hope of making progress by applying the comparative method in order to get an insight inta the pre-homeric situation we must adhere to it, being however aware of the possibility, or even the necessity, of taking the results of a psychological approach to language into account. If the objection is raised that the application of the comparative method forces us to face such issues as the origin of the word el, the etymology of which cannot be definitely demonstrated, the answer must be that any etymology bearing on prehistoric periods is hypothetical in character, that comparative philology as a whole is a complex of hypotheses which supporting each other have proved to be an indispensable means of deepen­ing our insight into the historicaJ. processes at work in the languages in which we areinterested. If we could succeed in harmonizing philo­logical facts-comprising those established by a "psychological inter­pretation" of Homer-and linguistic data, deriving from a methodic examination of all elements of the constructions at issue and of all relevant factors, we should to my mind feel easier than those who follow Tabachovitz inter alía in taking for granted the original character of a complete bipartite el periodo

My view of the problem under investigation, to wit the origin and the interrelations of the Greek el constructions, would, for the time being, be as follows. The particle may be assumed to have óriginally expressed the notion of "under this (that) circumstance (cf. E 104), in this case, this being the case, in case (cf. A 116), thuS"3). The function may be supposed to have been the starting-point for all the historical construc­tions, el developing in the course of time into a conjunction or particle

1) For this point the reader may be referred to my 'Remarques sur la place du verbe dans la phrase active et moyenne en langue sanscrite', Utl'echt 1952, p.67ff.

2) See Remarques, p. 74f.- Part of the subsequent l'emarks made by Tabacho­vitz are not very happy: the double function of the Fr. si for instance is a legacy from Latin (cf. e.g. Tert. Apol. 21 quaerite ergo si 11era est ista divinitas Ohristi).

3) Attention may perhaps be drawn to the construction ar yae ••• Tro ue "in case ••. then (would, could etc.)", see y 223f.; 456ff.; cf. B 371ff.; LI 288ff.

The conditionaI sentences 153

of special function and application. Whether Brugmann's etymology of el!) be correct-which 1 am inclined to assume-or not, is in itself irre­~evant. ~he cupitive, c?nditional, 01' i~terrogative force was originally mherent m the clauses mtroduced by el, tbe modulation of these clauses being an important factor in the context and the situation; the particle originally was, in my opinion, a means of referring to, or connecting the following utterance with, these 2) 3).

It would appear that this hypothesis can be substantiated by the following facts and considerations. All three expressions (wish, condition including concession, indirect question) can in many languages be formed without any introductory particle or conjunction 4). The conjunctions used in these utterances can accordingly often be shown to have deve­loped from words belonging to another class. Besides, the same element often serves like el in two oí the aboye functions, or sometimes even in a11 of them.

It must never be forgotten that these relations which we at least in written language, as a rule, express by means of conjunctions, are often implicitly or explicitly expressed in various other ways 5). In the speech

1) Brugmann, Grundri131 (1892), Il, p. 768; 2Il, 3, p. 990; KVG. p. 616; 669f.; Die Demonstrativpronomina del' idg. Spr., Abh. Sachs. Ges. d. Wiss., 22, 6, Leipzig 1904, p. 118; Brugmann-Thumb, o.c., p. 582.

2) For the role of the particIes see also A. Sauvageot, Le Tahitien, Conf. de l'Inst. de Ling. de rUniv. de Paris, 10 (1951), p. 93.

3) We are no doubt justified in assuming that the Greek conjunctions enet, snetíj~, snehe and Hom. Att. enetTa, Ion. Dor. enetTe, the adverb, originally were identical. They are correctIy considered to represent sní "upon, on, etc." and the ~bove el. In specified cases snd camlOt be considered a subordinating conjunction 1ll the proper sense of the termo Pind. O. 9, 28ff. dya{}ol óe ual IJorpol uaTd óaífl,ov' avóeer;; syéVOVT" snel dVTtOV nror;; UV TetÓÓOVTOr;; 'HeauAér¡r; IJUVTaAov Ttva;e Xe(}IJív; the meaning of the word is "for otherwise" (cf. e.g. also Arist. Nub. 688f.; Soph. O. T. 433; Herondas 2, 72). It seems possible to disagree with Schwyzer-Debrlmner (o. c., Il, p. 660) who regard this construction as a case of "Lockerung del' Ab­hangigkeit" and to expIain it from an original "in that case, then".

4) Hirt's thesis, Urgermanische Grammatik, IIl, Heidelberg 1934, p. 205: the Greek way of expressing a cOlldition by means of a particle "ist jedenfalls das alteste" does thel'efore not seem to be tenable. The same author is however right in questioning Behaghel's opinion (Deutsche Syntax, IIl, p. 637) that aU conditional sentences without a particle originate in interrogation.

5) Korean, for instance, feels no need for conjunctions (G. J. Ramstedt, A Korean Grammar, Helsinki 1939, p. 187). In Tonkawa (Texas) there are three subordinating enclítics, answering to our "because", "as", "while" and "but" respectively (see H. Hoyer, in Linguistic structures of America, NewYork 1946, p.31O).-For the expression of a condition by means of special paradigms of verbs see M. R. Haas, A grammatical sketch of Tunica, ibidem, p. 35lf. In the African (Bantu) Kamba language conjunctives are not often used. (See E. Brutzer, Handbuch del' ;Kambasprache, Mitt. Sem. f. Oriento Spr. Berlín, IX (1906), Afr. Stud., p. 55). The vaIue conveyed by our "that, in order to" is expressed by a special verbal form; that conveyed by "while" by another verbal category. For facultative 01' superfluous' conjunctions see e. g. K. Gronbech, Del' türkische

Page 80: Gonda.1956.The Character of the Indo-European Moods.pdf

154 The Indo-European moods

of many persons, in many languages, conjnnctions are often little used 1). The Brahni ni kiisa, i néto baréva may mean "you are goillg, yon say-I will come with yon"; "are yon going1 1 will come with yon" , or "if yon are going, I will come with yon" 2). In the Bantn idioms, in which conjunctions are of infreqnent occnrrence, and which, like the l.-E. langnages, often have etymologically different words to fnlfill this fnnction, an original term for "time" serves to express "when", a word for "way, manner" expresses the idea of "thns", a pronominal "jnst these (things)" that of "likewise" 3). In the Indonesian idioms Gayo and Achehnese the Skt. artha- "thing, matter, object" developed into a 'cansal conjnnction': "becanse, as"4).With regard to Snmerian it may be stated that apart from a special conditional particle nsed in laws, contracts, etc., the relation expressed by the word is very often indicated either by the context or by a clanse containing an imperative 5

).

A few words may be said here on the way in which conditions are formnlated in Chinese6). The conditional clanse, which always precedes the principal clanse, is not always characterized by special elements. As a rule, however, one of the following characteristics may be distinguished. The most 'primitive' form oí the conditional sentence has the snffix tsjé after the subject and is closed by ye 7). This is an nndifferentiated 'snbordinate' clanse which, nnder certain circnmstances, can also be translated throngh an English clanse beginning with "as, althongh, etc.". The word tsée "then" in the apodosis serves to exhibit the condi-tional fnnction of tsjé ... ye in the protasis; at a later period tsée can also occnr withont tsjé . .. ye. A second means of expressing a condition

Sprachbau, I, Copenhagen 1936, p.55, whose statement: "Eine nahere Unter­suchung der Konjunktionen (ware) eine sehr lohnende Aufgabe und würde auf die allmahlíche Verschiebung des Sprachtypus manches Licht werfen konnen" is also applícable to the other languages.

1) Many instances of conditional periods without any conjunction taken from Albanese may be found in M. Lambertz' article in Indog. Forsch., 34, p. 157ft'.

2) Denys de S. Bray, The Brahui language, Calcutta 1909, p. 221, who does not inform us about the modulation of the sentence. In view of the possible ambiguity in this style of speaking the Brahui makes, often under foreign influence, an increasing use of conditional conjunctions, the forms ofthe verbs being left unaltered.

3) The reader may be referred to D. Westermann, Grammatik der Ewe-Sprache, Berlín 1907, p.l07; C. Meinhof, Grundzüge einer vergl. Grarnm. der Bantu­sprachen2, Hamburg 1948, p. 125f. It may be remembered that there e:x;ist instances of a comparable shift in Englísh: in such sentences as he turned pale the moment he saw her and it is a disgmce the way he drinks the elements the moment and the way carne to play the part of conjunctions. See also R. W. Zandvoort, A handbook of English Grammar3, Groningen 1948, p. 231.

4) See my 'Sanskrit in Indonesia', Nagpur 1952, p. 393. 5) See A. Deimel, §umerische Grammatik2, Roma 1939, p.232. 6) See J. L. M. Mullíe, Grondbeginselen, lII, p. 395ft'. 7) For the combination tsjé ... ye as an inde:x; of subordination see Mullíe, o.c.,

lII, p.475 and esp. 357ft'.

The conditional sentences 155

consists of the use of the verb-conjnnctions zjóe, "to be similar to, to be such as, to be so (as), thns"; zjóe "to be similar to", both ofwhich assnme the sense of "if", etc., at first in addition to tsjé ... ye, bnt later withont these. A third method is the nse of yén "so" at the end of the conditional clanse. In the spoken language the conditional conjnnction can be omitted if the sense is obvions 1). Where it is necessary the conjnnction yao4-shih, or simply yao4 "if" is used. The connection between the main clanse and the conditional clanse mnst, however, necessarily be indicated by the addition of cMu4 or other elements in the former. The condition can be expressed irrespective of time.

In the Indonesian Bnginese, besides a word for "to try, attempt" nsed in the sense of" snppose, pnt the case that, if" we also find: iya pa lit. "that + an emphasizing particle" which means "(if) that happens"; nakko "if" probably contains a demonstrative na and ko "thns". In the same langnage and in closely related Macassar the aboye postpositive emph. pa itself is nsed in cases where we ha ve "if" : Mac. mántañ,l-pa2-kó3 "if2

yon3 stay1". In Mac. ka which invokes attention and kade (= ka + de "went, come!, say!") stand for utinam (introdncing a wish) and for "snppose, if"2). In Roti a word for "see" has come to fnlfill the fnnction of "if"3). The Leti mana "if" primarily is an emphatic particle. The Bare'e (an IN. idiom of Celebes) ane, which is related to Mac. anne "this" and "as to", not only serves to call a person's attention, bnt also to render onr "if, in case" 4); besides, it helps to express the idea of "if there is qnestion of, if the qnestion is ... ; as to". An interesting little vocable occnrs in Madnrese 5): jhaq, which in sorne constructions emphasizes the following word: jhaq1 taretan2 elaban3 "as to1 a brother2 (as far as a b. is concerned, now that it is a b.), he offers resistance3" or: "to a bróther he o. r."; elsewhere it can be translated by the cond. "if", by "when", or by the indo interr. "if" 01' "that": "he says1 that2 mother asks": pabalana1 jhaq2 ... A sentence like señkoq1 taq2 taa3 jhaq4 baqna5 sala6

can mean either: "11 did3 not2 know3 that4 yon5 are gnilty6 «01'» 11 d03n't2

know3 whether4 yon5 are gnilty6". The Minangkaban (Snmatra) ka' seems to have still more 'functiolls'.

In varions strnctnres ít bears the force of "perhaps, perchance, it may be that"6): ka1 pai2 badan3 ka'4 mati5 "(if I) shalF g02 13 shall perhaps4 die5"; bantie1i1 den2 ka'3lah4 lapeh5 "my2 cow1 may3 have4 got4, 5 loose5", or "what if my cow has got loose". If the sentence introdnced by leO' is

1) Mullie, Structural principIes, I, p. 64; lI, p. 9; 602. 2) See B. F. Matthes, Boegineesche Spraakkunst, 's-Gravenhage 1875, p. 251f.;

the same, Makassaarsch-Hollandsch Woordenboek, 's-Gravenhage 1885, p. 32. 3) J. C. G. Jonker, Rottineesche Spraakkunst, Leiden 1915, p.712f. 4) See N. Adrianj, Spraaklmnst der Bare'e-taal, Bandung 1931, p. 406. 5) See H. N. Kiliaan, Madoereesche Spraakkunst, Semarang 1911, lI, p. 101. 6) See also J. L. van der Toorn, Minangbakausche Spraakkunst, The Hague

1899, p. 198ft'.

Page 81: Gonda.1956.The Character of the Indo-European Moods.pdf

156 The Indo-European moods

concluded by ba-á "how?", it expresses a wish: kel kayó l awa'2 lah ba-á "we wish we were2 rich 1, if only we were rich". The word ka' can also answer to our "if": malaml basalimu~'2 ambun3, ka'4 sian5 bapayuan 6 awan7 "at nightl the dew3 is his cover2 ; if4 it is day5, the clouds7 are his shelter6". (Notice the absence of the particle in the former sentence which, however, is completely on a par with the latter). The 'transition' "perhaps", "if" seems to be manifest in such cases as ka'l angan2 dilim­bUrJ,3 pasan4, usah5 barumah6 di7 tapi8 pulau9 "perhapsl (you are) un­willing2 to be flooded3 by the tidal flow4; (you) should not5 take Up6 (your) residence6 on7 the beach8 of an island9", i. e. "if you are ... ". Cf. also ka'l la~2 untuañ3, kumbali4 pulan5; ka'6 tida'7 hilan8 di 9 1'antaulO "perhapsl (1) am2 lucky3, (1 shall) return4, 5; perhaps6 (1 am) not7, (1 shall) be lost8 on9 foreign shoreslO" i. e. "if 1 am ... , 1 shall; if not ... , 1 shall". In this connection mention may be made of a 'disjunctive' use of the particle, followed by nó lah, by which the sentence is concluded: ka'l hidu~'2 mati3 nó4 lah5 "perhapsl live2 dead3 he4 (lah5 emphasizes)" i. e. "either alive 01' dead". A single 01' repeated ka' can also occur in a subordinate construction: ka'l harimau2, ka'3 á4 lah5 tu6, tahitam7 samien8

tampa'9 dilO denll "perhaJpsl a tiger2, perhaps3 whatever4,5 it6 (was) , something black7 (only8) presented 9 itself9 tolO myll eyes9". In cases like Guból lah2 ka'3 lé4 lama'5 di6 an7 "won't you tryl, 2 whether3 you7 like5 it" and after verbs of being sure, knowing etc., when the context implies a doubt, ka' corresponds to the English subordinating "if, whether". It also combines with manta n "although, supposing that" to introduce a concessive clause: ka'l mantan2 samikin3 ikó4 hidu~'5 hambó6, ••• "al­thoughl,2 15,6 am S04 poor 3,4 (1 canafford ... )". In combination with bana( i) "true" it also forms concessive clauses: ka' 1 gadan2 bana( i P aie4 . .. "it may bel true3 the water4 is high2 (1 shall cross the river)", i. e. "although the water is high ... ". The particle also forms phrases expressing suppositions: the combination ka' ... santannó íor instance can be rendered by "if by any chance". Finally, ka' can correspond to our "as to": ka'l kayó2 tuan3 U. M., lah4 tahu5 den6 di7 kayónyó8 "as tol the wealth2 oí Mr. 3 U. M., l6 know5 his8 wealth8".

In Bolaang-Mongondow (Celebes)l) mOlio, which conveys the notion oí uncertainty, vagueness, 01' indecision, and is sometimes translatable by "perhaps" (cí. lima1 m01i02 onom3 "fivel it may be2 six3" i. e. "five 01' six"), can also introdtice subordinate parts oí the utterance: libo'l­pamonag2 mono3 ulan4 i 5 ine6 ... "go down2 and askl t05 whom6 (belongs) the snare4 ... "; libo'on-pal muna2 k03 inia4 mon05 inin6 ... "askl him3,4 first 2 whether5 he does not mind6 ... ". Sometimes it corresponds to the cond. "ií". The Banggai kalu stands for "supposing that" and "whether".

1) The reader may oonsult: W. Dunnebier, Spraakkunst van het B. M., in the Bijdragen Taal-, Land- en Volkenkunde, 85 (1929), p. 426; 86; p. 156; the same, B. M.-Nederlandsoh Woordenboek, The Hague 1951, p. 250.

The oonditional sentenoes 157

Esser l) at the time niade some interesting observations on the Mori (Celebes) particles ke and bao 'rhe latter may be considered an abbrevia­tion oí bara---which, meaning "perhaps" in the closely related Bare'e, is etymologically identical with the Malay etc. baran ("thing, goods, something") 2)_; it answers to our "if, when" 01' introduces indirect questions: "whether, how" (also in sentences like "1 don't know what he is called"). The primary sense oí ke may be that oí "perhaps". It can however often be indiscriminately translated by "ií", "whether" 01' "perhaps": "perhapsl P shall die2, but you will not be responsible íor it", 01' "ií 1 shall ... , yOU ... " 01' "whether 1 shall ... will not be your concern": kel mate2 aku3 ... In a related idiom, Tinompo, ke is an interr. particle; in dialects and in Bungku it means "if".

The 'double sense' of "as to" and "ií" is indeed frequently found in Indonesian words oí this class: a Bimanese clause "ií it is a buffalo ... ", introduced by1Ígara, canalso mean "asto a b .... , with regard to a b .... "3).

The Toba-Batok ia is according to Van del' Tuuk "properly speaking nothing else than a preliminary interjection (aanloopje) by means oí which the speaker brings about a transition to another subject 01' another utterance"4): it often occurs as an isolating, emphasizing, 01' annuncia­tory particle beíore a word (subject) which is placed at the beginning of an utterance; under certain circumstances it can be translated by "as to", 01' by "then, now". It can also mean "if", but evidently because the conditional force is implied in the utterance: ial pinisat2, bottar3 gotana4 "thenl one2 pinches2 it2 : its gum4 is white3, i. e. "if one p. it, ... ". The word often introduces, in this way, two antithetical sentences in close succession. In the same language the word ango preceding a term which is in an emphatical way placed at the head of a sentence-trans­latable by "as to" 01', rather, by a special intonation 01' also by "to be sure, namely, to wit" -has again a conditional sense: indadonl Ol02 mulak3 au4, a1Íg05 indadon6 dapot7 au8 "14 am notl willing2 to come back3,

to wit5, J8 don'ta find7 it7" > "1 ... back, if (implied in the context)

1 ... it". The Toba tun, finally, expressing possibility oíten occurs in a conditional clause, "fulfilling the function of a conjunction": tun1 di­suru2 h03 manaek4 au5, unan6 h07 d08 "ifl he mightl order2 you3 to ascend4 me5, don't6, 7 be willing8", as against molol tun2 dida3 halak4 "ifl another (person)4 might2 see3 it".

1) S. J. Esser, Klank- en vormleer van het Morisoh, n, Bandung 1933, p.201, n.1. .

2) There may be room for the observati~n that Mal. barañ oan also help to express a wish: barañ1 dipeliharakan2 Allaha • •• "may (it happen in some way that)1 Goda protects2 (him) " .

a) For partiCl;tlars see J. C. G. Jonker, Bimaneesche Spraakkunst, Batavia 1897, p.435.

') H. N. van der Tuuk, Tobasche Spraakkunst, Amsterdam 1867, p. 339f.; 294ff.

Page 82: Gonda.1956.The Character of the Indo-European Moods.pdf

----------------------------------

160 The Indo-European moods

It is difficult not to remember the intricate and disputed history of the Rumanian del). In Ancient Egyptian 2) certain emphatically antecedent elauses serve to formulate a condition: a "as to thou art ... " can also be rendered by "if thou art". The use of the element ir, occurring in emphatically antecedent elauses is optional3).

Words for "supposing, put the case etc." not infrequently develop a conjunctional use. In the IN. Gayo (Sumatra) señkíro "at a rough esti­mate, in (my etc.) opinion" (: kiro "estimating, calculating") is also used for "put the case, if for instance, if, to say so, if then, if once": it is often preceded by ka, i. e. ikO "as to, if": ka S.l ara2 beta3, berbagi4

kit65 besito6 "if1 it is2 thus3, let us5 now 6 make an allotting4"; it can also combine with uta "under these circumstances, then, further": uta1

S.2

k'3 umah4 leam5 ••• "if2, then1 , you5 have reached3 the house4 ... ". In the Melanesian Jabem (N. Guinea) condit. clauses are introduced by embe which is no conjunction proper, but a verbal form: -be, -mbe means "to think, be of the opinion that, to wish etc."4). In Assyrian a condition is very often asyndetically expressed, but 8umma "supposing" -intro­ducing a <main clause'-is far from rare 5). Compare, in French: a supposer qu'il parte maintenant, peut-étre arriverait-il encore a temps, and the popu­lar construction: une supposition que nous donnions tout notre avoir liquide pour . .. 6), and, in English: supposing you miss your tiger, he is not lileely to mis8 you.

We might even draw attention to such functional shifts as that of the Karo Batak (IN., Sumatra) bicara "usage, use, practice; manners; nature, character" and "in case, in the event of, as to, supposing that, if" 7). In Busang (Borneo) barañ, the primary sense of which is "thing, goods, something", also means "as to, as far as ... is concerned" and, in elauses of a negative or antithetical purport, "if, when, perhaps": barañ1 aleui2

A.3 "as to1 (me, am) 12 A.3r', "am 1 A.1" i. e. "(well) 1 am not A, to be sure"; b.1 akui2 8ayu'3 ni4,. ;" "if! 12 were (not ill, but) healthy3 ... "; a meaning like "the point, the fact, the question is" seems to have been the starting-point for this use.

1) For which see the brief comment by E. Bourciez, Elements de linguistique romane, Paris 1923, p. 564ff.

2) See also A. H. Gardiner, Egyptian Grammar, Oxford 1927, p. 163f.; A. de Buck, Egyptische grammatica, Leiden 1944, §§ 117; 252.

3) For optional conjunctions see e.g. also P. Arndt, Grammatik del' Ngad'a Sprache (IN., Flores), Bandung 1933, p. 31; C. M. Churchward, Tongan Grammar, Oxford 1953, p. 51; De Bray, Brahui language, p.220ff.

4) O. Dempwolff, Grammatik del' Jabem-Sprache, Hamburg 1939, p.90ff. 5) For particulars see W. von Soden, Grundri13 del' Akkadischen Grammatik,

Roma 1952, p. 211ff. 6) See F. Brunot, La pensée et la langue, Paris 1936, p. 876. 7) For this word, which originates in Skt. vicára- "procedure, consideration"

see my <Sanskrit in Indonesia', p. 393.

The conditional sentences 161

In this connection we may also remember such phrases as the Skt. evar!l,1 8thite2 ••• (e. g. in the Pañcatantra) "it being2 SOl ••• ", and as the Engl. in ca8e of my not seeíng you; you'd better take an umbreUa with you in ca8e ít rains. French former1y possessed the phrase en ca8 que expressing a hypothesis or condition: je pourrai8 . .. compter sur la connivence du ... président, en ca8 que la chose luí fút bien recommandée (Voltaire); cf. also: au ca8 ou il se pré8enterait VOU8 le recevriez, n' est-ce pa8? In Dutch in geval (litt. "in case") can be used instead of indien, al8 "if". In Swedish, ifall "if" < i "in" + faU "case" (cf. isa faU "in that case") is a common conjunction.

Pronouns not seldom develop into conjunctions, as Engl. that and Germ. da(J, the history ofwhich will, in the main, be known to the reader 1). In Indonesian languages instances of a similar development are not rare: Leti leaa, for instance, is a pronominal element which has assumed the function of our "if"2). Ledo and Kulawi (Celebes) ane "if" must doubtless be connected with Macassar anne "this". Gayo (Sumatra) iko "as to; if; though, even if" is no doubt identical with the Minangkabau iko "this", Jav. ika "that" etc.: iko1 aleu2, {Joro3 mera4 "as far as1 12 am concerned1,

1 don't3 like4 t04 ••• ; ileo1 uron2 lañ3, aku4 {Joro5 bloh6 "if! it is raining2 tomorrow3, 14 shall not5 goG"; iko! tuo2 pé3 ko4, te{Jor5 ilon6 "though1 you4 are old2 (pé emphasizes), you are still6 strong5". The word can moreover also occur in questions expecting agreement or approval: "isn't it". One might compare Jav. ileu "that" which can also be used as a 'demon­strative conjunction': "so, then, now, therefore": iñkañ1 punilea2 (the krama substitute) paman3, 8inten4 • •• "as to this1,2, unele3, wh04 ••• "

(iñkañ, leañ is an isolating and annunciatory< ineluder' 3)), i. e. "now, therefore, unele, who ... ". This iñkañ punika, or kañ iku, apparently refers to the preceding words, adding something in connection with them, or with reference to them, or as a consequence of the thought expressed by them 4).

In another paper 5) 1 discussed the widespread expression of the con­ditional relation by a paratactic pair of clauses which is no doubt as a rule characterized by a special intonation 6): the Latin (Petr. 44, 3) serva me, 8ervabo te; the Russ. za choéu, poljublfu "(if) 1 wish, 1 shalllove" 7);

1) For particulars see e.g. O. Behaghel, Deutsohe Syntax, IIl, 1928, p. 128ff.; L. H. Gray, Foundationa of Language, NewYork 1939, p.l71f.

2) Cf. J. C. G. Jonker, Lettineesche Taalstudien, Bandung 1932, p. 118. 3) I refer to my papers 'Indonesische relativa', Bijdragen Taal-, Land- en Volken-

kunde (The Hague), 102, p. 501ff., and 'The r.-E. rel. pronoun jo-', Lingua, 4, p.lff. 4) C. Poensen, Grammatica del' Javaansohe Taal, Leiden 1897, p. 271. 6) Lingua, 4, p. 35ff. 6) This point is overlooked by R. Van Pottelbergh, Over de geschiedenis en de

betekenis van den sl-zin in het Grieksch, Gent 1939, p. 9ff. 7) For Slavonic see W. Vondrák, Vergl. Slav. Gramm., II, G6ttingen 1928,

p.545; for Celtic, H. Pedersen, Vergl. Gramm. d. kelt. Spr., II, G6ttingen 1913, p. 256; 318. See also Hirt; o.c., VII, p. 174. 11 Gonda, Indo·Euxopean mooda

Page 83: Gonda.1956.The Character of the Indo-European Moods.pdf

--,.,~~-~~_.~

! !

162 The Indo-European moods

the medieval Dutch ghelovet an hem; du. wordes wel ghesont; in French un an de vie de plus, on l'aurait tmuvé; the popular construction: une supposition, il vient, je le chasse 1). The same intonation could be main­tained if an anaphorical or other pronominal element was added (see further on). In addition to the instances quoted in the afore-mentioned publication 1 first add some other Sanskrit sen~ences: ~h Up .. 6~ ~ 1, 2 asya1 yad2 ekii'fJ13 sakM'fJ14 jivo5 jahati6, atha7 saB SUí}yat~9; dvtt~ya1(l'10 jahatill, atha12 sa13 su~yati14 . .. "if2 the life5 leaves6 one3 branch4 of ltI, then7 itB dries Up9; (if) it leavesll a secondlO, then12 that13 dries Up14 ... "; Mbh. 3, 72, 17 athava1 tvarate2 bhavan3, e~a4 yati5 sivaJ:¿6 pantha7, yahiB

var~'(/,eyasarathi?~9 "or1 if you3 are in a hurry2-the4 :vay7 g,~es5 smo~th and straight6-goB then with VaWI).eya as your charlOteer9 ; amI wlth the same initial position of the verb in Greek: ~ 193 s'tr¡ p,Sv vvv vwi"v ... ¿amor¡ I ... I rJr¡i"Oím~ xsv lhrsl'ta ... I ov 7:l alan(!f¡~a,lf,l}.~ymv2). • .

In the Indonesian language Toba-Batak molo lf lS also, accordmg to the formulation given by van der ~r-q.uk, often left out; the original sense of this word must have been "to be thuS"3).

Now it is a well-known fact that in Vedic texts-where the verb of a subo;dinate clause introduced by a form of ya- etc. is accented (see Pal,lÍni 8, 1, 66)-the first of two clauses which, th?ugh otherwise .in form a principal one, is virtually subordinate, occaslOnally accents Its verbo Such a sentence can be translated as if it were introduced by a

'1 't'2 3 ' 4 > -'k 5 'ndra6 conjunction: J,tV. 6, 47, 31 sa'fJ1 ... camn ~ no naro, sma am ~

t'athíno7 jc¿yantuB "when our3 men4 ... come2 t?getherI, 0~r5 car-fight~rs7 mustB, O Indra6 , win B the victoryB". But, in pomt of fact, lt was the hlgh tone which indicated that the first part of the utterance was no separate unity, that it was to be continued 01' complet~d. ~articles. occ~rring in these constructions are not necessarily subordmatmg conJunctlOns: cf. Gothic rfJ4). The former of a pair of sentences which contain, both of them, the particle ca-this element is, as 1 have sho:wn elsew~ere5) a marker of complementary connection-also accents Its verb m cases such as gardabM'fJ1s1 ca2 kálayati 3 vi'(/,a'fJ14 ca5 vadayati6 (Pal).. 8, 1, 59) "he drives3 asses1 and2,5 plays6 the lute4", and, moreover, in virtually

. 'd '1 2 l' -t'3 4 '5 h6 conditional clauses like J;tV. 2, 41, 11 ~n ras ca mr aya ~ no, na na. pascád7 aghá'fJ1B nasat9 "if2 1.1 be gracious3 to us4, n05 hurtB will9 there-

1) See also A. Meillet, Ling. hist. et ling. gén., I, Paris 1921, p. 173f. 2) See also Schwyzer-Debrunner, o.c., n, 682f.; Spey~r, Ved. Skt. Synt.,

§ 284; F. A. Stoett, Middelnederlandsche Spraakkunst, Syntaxls3, The Hague 1923,

p.217. 3) H. N. van der Tuuk, Tobasche Spraakkunst, p. 123; 334. 4) See Delbrück, in Paul U. Braune's Beitrage, 29, p. 271; Schwyzer, LF. 23,

p. 163f.; Streitberg, o.c., p. 246; 250. . . . 5) See my paper on the r.-E. particle kUe, especially m Greek a?-d La.tm , m

Mnemosyne (Leyden), IV, vol. 7 (1954), p. 177ff. and 265ff. A spemal artwle on the r. Ir. ca is to follow in the Indian periodical Vak, Poona 1956.

The conditional sentences 163

after7 reach9 us6". In these clauses ca had, in the course of time, come to convey the sense of "if" which originally was inherent in the contexto There can scarcely be any doubtthat it originally was the mere juxta­position of the clauses (and the intonation) that implied the conditional force 1). So these Vedic constructions enable us to get an insight into the processes leading to the development of conditional conjunctions.

Combining with id this ca develops into the conjunction ced2): J;tV. 7,72,4 ví1 céd2 uchánty3 ••• u~ásnJ:¿4, prá5 vam6 bráhma'(/,i7 kiirávoB bhnmn­te9 "when2 the dawns4 grow brightl, 3, the poetsB present5, 9 you6 theirholy texts7". The verb of a ced clause which sometimes verges on a temporal sense is accented (Pal).. 8,1, 30). According to Indian sources 3) the con­junction mainly serves to mark an individual or particular case (pak~an­tnre). In this function it is extremely frequent. But sometimes and, as far as 1 am able to see, especially in comparatively ancient texts, a non­conditional force is still apparent: cf. e. g. AthV. 2, 30, 2. The second component of ced, to wit id, is a very frequent Vedic particle which mainly serves to express emphasis 01' affirmation, conveying in a more 01' less antithetical clause the sense of "indeed, assuredly" -cf. J;t V. 1, 147, 3 dí1psanta1 íd2 rilJávo3 náha4 debhuJ:¿5 "the enemies3 wishing1 indeed 2

to hurtl were in nowise4 able to hurt5" which tends towards a concessive force: "although" .. "-; it may be considered to have arisen from a stereotype neuter acc. of the dem. prono i-.

The other Skt. word for "if", yadi, is a combination of the neuter prono yad and the deictic element -i 4). In Vedic yad could by itself inte1' alia fulfill the same function. Elsewhere5) 1 ventured the supposition that yad (from ya-) here, too, originally was an <emphasizing', isolating, < defining' element which in constructions implying a conditional relation or hypothesis-cf. e. g. J;tV. 2, 28, 10 sten61 va2 y63 dípsati4 ... tásmiid5 •••

pahy6 asmán7 " ... or2 if a thief3,1 intends to injure4 (us) ... , protect6

us7 from that5" -came to as sume the character of an explicit means of empressing that relation. But yad and yadi are, in harmony with their origin, plurivalent 6): the latter also serves the subordinate clause which depends on verbs of knowing, asking, saying, thinking, observing, expecting, doubting etc., and the former is e. g. found afterviduJ:¿ "they know": J,tV. 1, 131, 4 vidú~l te2 asyá3 viriJnsya4 ... yád5 indra6 ... aváti­raJ:¿7 "they took cognizance Of1 this3 heroic deed4 of thine2 that5 thou, O Indra6, overpowered 7 ... ". Here the word yad has the same distinguish-

1) See my observations in Mnemosyne 1954, p. 20lf. 2) For particulars, see Grassmann, O. C., 205f.; Delbrück, Altind. Syntax, p. 495. 3) See the Petr. Dict., n, 1054. 4) Of. also Renou, Gramm. véd., p. 329; 391. 5) In the Lingua, 4, p. 36. 6) One might to a certain extent compare the history of Slav. jako, the neuter

of á rel. pron., which can serve as a modal, declarative, causal, consecutive, tem­poral, final, and hypothetical-conditional conjunction (see Vondrák, o.c., n, p.505ff.). .

n*

'j

Page 84: Gonda.1956.The Character of the Indo-European Moods.pdf

164 The Indo-European moods

ing, annunciatory, and isolating character: because it includes the following words in a larger whole it as sumes the fundtion of a conjunction. There can be no doubt that constructions such as: kathaya1 me2 yadi3

•••

"tell1 me2 whether3 ••• "; vicaryatli'l'fl- yadi... "refl.ect if ... "; M. S. 1, 4, II ná1 vái2 tád3 vidma4 yádi5 brahmw(uj,6 vfi,7 sm6s 'brahma'(l-a9 va10

"we don'tl know4 whether5 we ares brahmans6 or7,10 non-brahmans9"

have the same origino After verbs of doubting or not believing yadi must be translated by "that": Ram. 2, 51, 14 na1 sa'l'fl-se2 yadi3 jivanti4 .

"1 do not1 expect2 that3 they are alive4". The phrase du~7cara'l'fl- yadi, lit. "it is difficult if perhaps" came to mean "hardly, scarcely". In such sentences as Mbh. 5, 177, 13 mama1 • •• hrdi2 kiim03 'bhivartate4 I ghii­tayeyar(l-5 yadi6 ra'(l-e7 bhi~mams iti9 the conjunction has the force of "perhaps, if perchance": "the desire3 obtains4 in my1 heart2 if perhaps6 1 could Slay5 Bhif?mas in battle7". Many cases of a yadi clause without an apodosis may be ellipses, but sometimes yadi + opto may be supposed to express an independent wishing "perhaps": Ram. 3, 54, 3 Sita when being carried away casts offher jewels among a band of apes yadi1 ramaya2

sa'l'fl-seyur3 iti4 "thinking<l: perhaps1 they will mention3 (them) to Rama2".

The conclusion appears to be justified that yad + i, i. e. yadi was preferred, if the subordinate clause expresses doubt or incertitude1).

The history of OLat. sei > Lat. si also deserves passing notice. Its origin in a sing. loe. of the demonstrative so-, though disputed in irre­levant details, has been generally accepted by the main authorities 2

).

The ancient Latin si dis placet must have meant "thus it pleases the gods": cf. Plaut. Capto 454 expedivi ex servitute tilium, si dis placet. Actually the phrase meant "God willing" or "if it please the gods"; Most. 1193 quiesce, si sapis "quiet down, (in that case» if you have any sense."

The MRGerm. so and Dutch zo (,,-, Got. swa, ORG. etc. so, Engl. so etc.) 3) may also be compared: so aber der Mann stirbt, soist sie los (Luther); Dutch zo de Heer wil "God willing"; zo ja "if SO"4). The common Dutch conditional conjunction als derives from al + so (just like the Engl. also < all + so, but this is not subordinative).

The interesting developments of Slavonic da (which deriving from a prono stem *do- and originally meaning "thus" came into use for "if", "in order to" etc., but in OChSlav. it is still in use as an emphatical particle; when joining an impero etc. it serves as an adhortative particle;

1) Cf. also the opinions brought forward by Delbrück, Altind. Synt., p. 348; Speyer, Ved. S. Synt., p.90, n. 1; Renou, Gr. véd., p.391.

2) I refer to Walde-Hofmann, Lat. etym. Wtb., n, p.530; Ernout-Meillet, Dict. etym.3, p. 1097ft'.; 1112f.; Walde-Pokorny, Vgl. Wtb., n, p. 458; Hofmann, in Leumann-H., Lat. Gramm., p. 771; Meillet-Vendryes, Gramm. comparée2

, p. 654. 3) I refer to Feist, Vgl. Wtb. d. Gotischen, p.462; F. Kluge-A. Gotze, Etym.

Wtb. d. deutschen Spr.15, Berlin 1951, p. 729. ') See also H. Paul, Deutsches Worterbuch, S. v. 80.

The conditional sentences 165

(under other circumstances it helps to express a desire) may also be re­called to memory 1).

In considering Armenian et'e, t'e which apart from introducing a conditional clause and a double interrogation of the complementary type serves also to announce a sentence and to introduce direct speech 2), one arrives at the conclusion that "thus" may have been the 'original' function of this elemento

Returning now to the problems connected with the Greek el we must first consider the etymology of this particle. Most scholars agree in main­taining that it is identical with Gothic ei 3 ), both forms constituting a sing.loc. ofthe pronoun *e-: 0- "this". There is indeed no serious reason to differ in opinion 4).Whereas Ion. Att. Are. el can be explained as a fixed form of a neuter or masculine locative, Lesb., El., N.W. Gr., Dor. al 5)

may be regarded as the corresponding feminine (stem *a-), Cypr. ñ representing a fem. instrumental. We might also compare the parallel *tej, represented by Got.pei S'u, Z'Va (rel. and conj.) and Dor. -ceibe "here"6).

Now forms belonging to the pronominal stem *e-/o- often referred to following or preceding elements of the enunciation 7). These forms are not relative, but cases like the following show that they could be freely used: J;tV.l, 73, 4 tá'l'fl-l tva2 nár03 ••• ágne4 sácanta5 ••• I ádhi6 dyumná'l'fl-7 ní8 dadhur9 bhúrylO asmin,u "O Agni4, the man3 follow5 Thee2 ••• They have bestoweds,9 much10 honour7 upon6 himll (= Agni)". They could also refer to a word in another clause of the same sentence: J;tV. 3, 13, 1

1) For particulars see Vondrák, o.c., n, p. 515ft'.; see alao Meillet, Ling. hist. et ling. gén., I, p. 17f. .

2) Meillet, Altarmenisches Elementarbuch, Heidelberg 1913, p. 139ft'. Etymolo­gical arguments point in the same direction: Walde-Pokorny, Vergl. Wtb., I, p. 98.

3) See e.g. Brugmaml, GrundriLl, n, p. 768; 786; idem2, n, 3, p. 990; the same, Die Demonstrativa der idg. Sprachen, in Abh. sachs. Ges. d. Wiss., phil.-hist. Kl. 22, 1904, p.32ft'.; 118; and in Berichte sachs. Ges. d. Wiss., phil.-hist. Kl. 63 (1911), p. 166f.; Griech. Gramm. 3, p. 243; 507; 533; Brugmann-Thumb, ibidem4, p.616; Schwyzer-Debrunner, Griech. Gramm., I, p.613; Boisacq, Dict. étym. (1923), p. 219 (who does not explicitly identify sl and Goth. ei); Hofmann, Griech. etym. Wtb., p.70; W. Streitberg, Gotisches Elementarbuch5 (1920), p.226, § 344; Feist, Vgl. Wtb. Got. Spr., p. 130. Lack ofspace prevents me from discussing also the various connotations of Gothic pat-ei, for which see the aboye books by Streitberg and Feist (cf. also W. Krause, Handbuch des Gotischen, p.299).

4) Schwyzer-Debrunner, o.c., n, p. 557 prefer to regard sl as an ancient inter­jection (see also n. 1).

6) For particulars see E. Hermann, Die Nebensatze in den Griech. Dialekt­inschriften, Leipzig-Berlin 1912, p.270ft'. (see. also C. Hentze, in Bezzenberger's Beitrage 29, p.280ft'.; in KZ.41, p. 356ft'.; 42, 131ft'. and in TlÍllw;, Festschrift A. Fick 1903, p. 77ft'.).

6) See also A. Musió, in Beitrage zur Gesch. der Deutschen Sprache und Lit., 53, p. 228ft'.; 1. Dal,in Norsk Tidsskrift for Sprogvidenskap, 9, p. 193.

7) I refer to Brugmann, Grundri132, n, 2, p. 333; Speyer, S. S., § 274; the same, V. S. S., § 133. Cf. Geldner, Der Rig-Veda, 12, p.350; Debrunner-Wackernagel, Altind. Gramm., nI, p. 518,

I ~

Page 85: Gonda.1956.The Character of the Indo-European Moods.pdf

166 The Indo-European moods

p1'á1 V0 2 deváya3gnáye4 bá1'hi~tham5 a1'cá6smai7 "start (a eulogy)l in honour of 3, 4 your2 god3 Agni4, st¡¡,rt6 the mightiest5 (eulogy6) in his honour7".

That demonstrative pronouns could under favourable cITcumstances develop a relative value is a matter of common knowledge 1 ).

If the aboye etymology of el is correct, its original sense must have been "in this, in this case, with reference to this, in view of, with regard to this, there, in that way, etc." 2). This sense is, if appearances are not deceptive, reflected by the use of Gothic ei 3 ), which serves-as a rule after an expression of time-as a 'relative': Luke 1, 20 (Cixel i]~ f¡fhSea~ ySY'Y)Tat TavTa) und pana dag ei waí1'pai patcl; Neh. 5, 14 (ano T* f¡fhsea~ i]~) jah f1'am pamma daga ei anabaup mis; Col. 1, 9; once the Greek OY Te6noy ¡jé

is translated by pamma haidau ei: 2 Tim. 3, 8. In a few passages ei is < demonstrative' 01' < copulative' 4): J ohn 16, 17 (fhlUeOY Mi ov {}sweú-d fh8) leitil ei ni sailvip mik "ein Kleines, da seht ihr mich nicht" (Streitberg). In other cases an original "in the matter of, on the subject of, on account of, with reference to the fact, etc." may have been the starting-point: Luke 2, 3 (Ual enOes1JOYTO náYTS~ anoyeácpsa{}at) jah iddjedun clllai, ei melidai we8eina; John 6,5 (n6{}sy ayoeáawfhSY CieTOV~, Zya cpáywaw OVr:Ol;) lvap1'O bugjam hlaibans, ei matjaina pai?; 1 COI'. 15, 9 (ovu slfhi ZuaYo~ uaAúa{}al an6aToAo~) ikei ni im Wai1'p8 ei haitaidau a.; Luke 9, 54 ({}SASl~ srnwfhsy;) wileizu ei qipaima?; Matth. 5,17 etc.; compare also sl: on before a quotation: John 18,9 (slnsy on OV~ osowuá~ fhOl ... ) qClp ei panzei atgaft mi8. The combinationpa1'ei is local in sense: onov Matt. 6, 19. The relative pronoun is saei 5

), i. e. the demonstrative sa + ei, but in a minority of cases the original demonstrative force of the word group is still evident: Tit. 1,5 (WVTOV xáew ansAm6y as ey Kef¡T17) in pizozei waihtai8 bilaip pU8 in K1'etai. This demonstrative is emphatical.

It would appear to me that the various fimctions of Greek sl may be explained from one source which is identical with that of the Gothic ei. To discuss the interjectional use first, the ancient 6) explanation el o' Ciys (cf. A 302) < el os <(JOVASl) Ciys did not account for sl + opt.; besides, the idea expressed by (JOVASl did not often suit the con-

1) See e.g. Brugmann, GrundriJ3, n,2, p. 348. 2) E. Hermann, o. C., p. 276 derives the femction of the particles al, si in main

clauses from an original sense: in Germ. "da, so; gehe".-The semantic starting­point mentioned in the text ("with reference to this case etc.") can also account for the conditional-temporal or temporal sense of the conjunction found in the dialects of Locris, Delphi, Laconia, and for the more or less causal implications which are sometimes proper to the conjunction (this force was particularly empha­sized by Stahl, o. c., p. 513ff.).

3) See Streitberg, o.c., § 344; 353ff. 4) See Brugmann, GrundriJ32, n, 3, p. 990; Streitberg, o.c., p. 227. 5) See Streitberg, O.C., § 347. 6) I refer to Eustathius 107, 1519, quoted by H. Ebeling, Lexicon homericum,

Leipzig 1885, I, p. 347. See also W. Pape, Griech.-Deutsches Wi.irterbuch2 (1888), p.723.

The conditional sentences 167

textl). The supposition that .sl was an ancient interjection 2)encounters difficulties: in cases like K 111; O 571 (see further on) it is immediately followed by an enclitic 3); besides, it is often followed by yáe which as a rule stands after the first word in a clause and not at the beginning, the derivation of yáe fromthe enclitic y8 and Cie( a) being pretty generally accepted 4).

The particle may, in my opinion, have referred to the situation 01' to the preceding part of the utterance, conveying the sense of "in this case, in this, with reference to this, with regard to this, nnder these cITcumstances, so etc.". Thus after a preceding part of the utterance in which a case is put, a supposition is formulated, reference to a situa­tion is made: o 831f. "if thou art indeed a god ... in that case (so) well then, tell me": el fhSY MI {}s6~ eaal ... [ ,sl o' Ciys fhOl ... uaTáA8~OY; X 381 (after enst); 1262. In 8 37 Bl may have had the value of: "this being so, in consideration of (in view of) these facts". Cf. also (with Ciy8) A 302 (at the end of an address); e 18; a 271; A 524; T 108; Z 376 (at the beginning of an address: H. did not find his wife; he asked the women el o' Ciy8 fhOl ... fhv{}f¡aaa{}8, "where did she go 1"; cí. 11 6(7). Needless to say that the element of adhortation was, at least originally, contained in the verb and in Ciys. Without Ciy8 1262 "A. offers yon worthy gifts, so you will ceas e from yOlir anger. So listen to me, and I will tell yon ... ": sl os av fhSY fhsV Ciuovaoy, eyO; os us TOl uaTaAS~w.

As the so-called cupitive sl, in sentences like O 571 it refers, in a similar way, to the preceding part of the context, the optative expressing a suggestion 01' polite order: "A., nobody is younger 01' swifter than you: so I would you might leap forth and smite sorne man of the T.": sr Twá nov Tewwy e~áAfhSYO~ Ciyoea {JáAow{}a (Dntch: "jij moest dan eens; als jij dan eens ... "); K 111 aAA' sr n~ Ual ToVa08 ... uaASa8lSY lit. "another thing, then one could (should) summon these also". In this use al, el are frequently followed by yáe, which may be regarded as etymologically representing y8 + CieaS). Now ys tends to attach itself to pronouns; in

1) For other explications see Schwyzer-Debrunner, O.C., n, p. 557. In associatillg this si with the interjection slev Schwyzer-Debrullner, 1. c. seem to overlook its different meaning and syntactical function. This word is as a rule used in the sense of "very good, quite so, well" (in passing on to the next point). The interjection sla is mainly found in such phrases as dU' ola, Jj da, ola dí, ola o1}, but also: Eur. Med. 820 ela xwes~ etc.; cf. Aesch. Aeg. 1650!

2) Thus G. Gerland, Der alt-griech. Dativ, Marburg 1859, p. 15; Schwyzer­Debrunner, o.c., p. 557.

3) Of. also f3 33; Soph. O. T. 863; Eur. Hec. 836; Xen. Mem. 1, 2, 46 etc. 4) Schwyzer-Debrunner, o.c., p.560 even suggest writing a¿ (sl') y' ae "wenn

folgerichtig, der Sachlage gem¡W", adding that "das hervorhebende ys von yáe ... anfanglich zu:¡n vorhergehenden ersten W ort des Satzes bzw. zum Satzganzen (gehi.irte)". We shall revert to yáe further on.

5) Ireferto.r. D. Denniston, The Greekparticles2, Oxford 1954, p. 56; Schwyzer­Debrunner, o. c., n, p. 560.

Page 86: Gonda.1956.The Character of the Indo-European Moods.pdf

168 The Indo-European moods

Homer it even occurs after words of this class much oftener than in other combinations 1). This tendency is to all appearances ancient, since the Vedic gha (ha) to which it corresponds shows the same predilection 2) : cf. e. g. ~V. 8, 1,30 stuhi1 stuhíd2 eté3 gha4 te5 máñhit}thaso6 magh6nam7 "lobsingel, lobsinge doch2 ! Diese3,5 sind ja4 unter allen Freigebigen7 an Gabe die Freigebigsten6" (Geldner); 7, 86, 3 ayá'rfl-l ha2 túbhya'rfl-3 váru~o4 hr~ite5 "this1 VaruJ)a4 is wroth5 with thee3". The Ancient Indian ha very often follows a fixed case form of a pronominal stem, 01' an adverb deriving from a pronoun: atha ha, tatra ha, iti ha etc. 3). In argumentations, reasonings, expositions this particle is frequently found after the first word of a sentence which joins a preceding part of the discourse. "Es betont den AnschluB"4): Ait. Br. 4, 25, 5 Prajapati and the seasons find support in one another. eva'rfl-l ha2 vava3 sa4 rtviji5 pratiti§thati6 yo7 ... "SOl

then2 verily 3 he4 wh07, ••• finds support6 in the priests". Like ye, gha and ha can be followed by other particles.

Without entel'ing here into a discussion of the various values conveyed by ye, ae, and yáe, it may be observed 5) that not infrequently the preced­ing part of the utterance contains a ground to which the phrase el (aX) yáe refers: B 370 "you are pre-eminent in speech ... I would that I had ten such counsellors" (al yae ... elsv); L1 288; II 97; r¡ 3Il; that the wish is, in other passages, for something supplementary to and usually similar to, a fact just stated as true: K 536; Y 205; cp 372; that, in dialogue this phrase helps to formulate the wish that something stated or wished by the previous speaker may come (might have come) true: e 496; T 309 etc. In all those cases al (sl) yae . .. cannot be disconnected from what precedes.

I cannot subscribe to Musiií's theory6) according to which the condi­tional sentences have arisen from concessive clauses: the primary stage being represented by cases such as A 280f. the secondary by cases like y 371f. In defending this view the author too onesidedly emphasizes an • assertive' force of sl. That this element can convey a concessive sense is beyond dispute 7), but this use does not appear to be the onlyoriginal one. The concessive force may rather be supposed to have been inherent

1) See Ebeling, o.c., p.247ff. (very often after demonstrative pronouns); Denniston, o.c., p. 121ff.

2) GraJ3mann, Wtb. zum Rig-Veda, s. v. gha (419f.): gha alone "nur zweimal nach einem Nomen, einmal nach einem Verb, sonst irnmer entweder nach ná (cf. the use of ')18 in negative answers: oot5év ')18 "not at all" etc.) oder nach Pronomen oder nach Richtungsw5rtern"; for ha, see ibidem 1640. 1 also refer to Delbriick, Grundrill, IV, 2, p. 499.

3) I refer to the Petrograd Dict., VII, 1485. 4) Delbriick, Altind.Syntax, p.498. 5) Compare Denniston, o.c., p. 91ff. 6) A. Musió, Beitrage zur griechischen Satzlehre, Zagreb 1927, p. 1ff. 7) Kiihner-Gerth, o.c., Il, p. 489, A. 2. The original sense of 8l must have been

similQ,r,

The conditional sentences 169

in the context, especially in the combination oí th;e particular protasis and apodosis. The ·conjunction' and in cases other elements-e. g. 8p,w~ in Xen. R. eq. 1, 17 el os TWS~ av~av6p,evo~ p,sTafJáAAovaw, lJp,w~ o'Í57:w -&aeeOfiV'íS~ oox~p,áCo~sv (av)-served to emphasize this force, or to make it explicito 1 would therefore explain A 28Of. el 08 av xaeTee6~ eaC1t ... I aAA' 80s cpse7:se6~ eanv ... as being based on an original: "with regard to this, to the case (as to this; in this [case]; it being the case): you are the valiant one, yet he is the mighty one"; the actual character of the process contained in the el clause is explicitly conceded. So a contrast which rightly impresses us as a concessive relation· comes to be explicitly indicated if the clause introduced by aAAá is added. Com­pare alsoA 290f. el os p,w alxp,r¡Ti¡v lf{}scrav -&eol ... l7:ovvexá ol ... ; "as to the case: they made him a warrior ,do they therefore ... ?" In Y 371 xal sl 7tVel xe¡;ea~ Eo~xev means "even if (01': though) his hands be as fire"; this force though indicated by Ual sl is likewise implicitly borne by the sentence in its entirety.

The analysis proposed by the same scholar 1) of sentences of the type A 137f 'j¡ , \ j¡ I ,\ j¡ I , \ "1 1" \ "A" . e~ ue us p,r¡ uwwaw, eyw ue usv aVTO~ el'.wp,a~ r¡ 7:SOV r¡ taVTO~ ... which, in his opinion, stands for eyro ... /!Awp,a~ (ysea~ 7:tv6~), el7:e (usv í!Awp,a~) 7:s6v, et-rs AlaV7:0~ ysea~ ... is far from convincing; the German translation of "sei es" for elTe cannot be an argument in favour oí the essentially concessive character of the Greek s'hs. In Arist. Nub. 269ff. eA-&S7:8 oif7:', d) NScpSAat, el7:e ... uá-&r¡a-&e, el7:e ... Ya7:a7:e, e'he ... , 1j EXS7:S, el . .. sl may be supposed to have had its starting point in: "in case ... in case", i. e. "be it that ... be it that", TS expressing complementary connection. The disjunctive character of el7:S ... s'hs (expressing that two 01' more cases are put as equally possible 01' equivalent) is implied in the juxtaposition of the parallel clauses containing this particular opposition of ideas. Cí., for instance, also Eur. TI'. 942 rjJ.-&e ... aAáa7:we, sh' )1A8~avoeOv -&SASl~ I OV6p,a7:l 7teOC1cpwveiv vw s'he ual IIá(!w (" ... supposing on the one hand you would call him A., supposing on the other hand ... P."), which developed into ~he disjunctive sense oí "either ... 01'''. After these considerations a detailed criticism of MusiC's argument does not seem to be necessary. Suffice it to express the opinion that there is no reason to regard X 49ff. 2) as having originated in concessive-conditional constructions and as having been "Glieder einer konzessiven Disjunktion, als deren Hauptsatz ein Verbum zu denken ist, das die del' Situation ent­sprechende Folge del' Handlungen beider Glieder del' Disjunktion be­zeichnet und in denen andere Yerba enthalten sein konnen, die spater in den Apodosen del' zu hypothetischen gewordenen konzessiven Protasen erscheinen sollen, in Verbindung mit del' entsprechenden Form des Pronomen indefinitum; also: ei7:e CCÓOVC1t Sl7:e 7:e{Jv{ial, V7top,svofip,sv n

1) Musió, o.C., p.11ff. 2) Musió, o.C., p. 12.

Page 87: Gonda.1956.The Character of the Indo-European Moods.pdf

170 The Indo-European moods

«sei es daB sie leben, sei es daB sie tot sind, wir werden etwas ertragen». Das Pronomen indefinitum verlangt nach konkretem Inhalt, den es dadurch erhalten kann, daB zu jeder Protasis eine eigene Apodosis hinzutritt". Apart from the aboye objections the complicated character of this theory is as great a disadvantage as the a priori and quasi histori­cal bias to which it is subject: "Da sich die Entwickelung der Konzessiv­satze zu Bedingungssatzen in Disjunktionen bewegt, sollten diese eigentlich immer zu zweien oder mehreren auftreten. Diese Erwartung ist auch begrifflich vollkommen gerechtfertigt. Bedingungssatze enthalten namlich Moglichkeiten. Dieser müssen aber immer mehrere dasein; eine einzige Moglichkeit ist keine Moglichkeit mehr, sondern eine Notwendig­keit. Daher erscheinen Bedingungssatze oft zu mehl'eren ... " 1) Musiií's view that the indicative in both parts of an hypothetic sentence must be explained from its original concessive nature-"in einer konzessiven Pro­tasis wird ursprünglich eine Tatsache zugestanden"-cannot, I am sure, be substantiated. N 01' should we adhere to his opinion that the basic form ("Grundform") of all hypothetic sentences-apart from that with the in­dicative in both clauses-was a construction with either the • futuric' (pro­spective) subjunctive 01' with a real future in both clauses (type: sr us C1j7:f¡ t1vr;, s{Jer¡r; uS'V = si C1j7:17t1str;, SVe1]t1str;2). How can we know for certain that thenuances expressed byother combinations ofmoods and tenses-for which I refer to a later part of this chapter-came into existence after, or even arose from the aboye type of sentence? The argument that "es kann ... vorkommen, daB die Ausführung del' Handlung del' Protasis nicht erwartet, sondern bezweifelt wird"3), and that "dann die Zeit, in der sie als erwartet ausgesagt werden konnte, vergangen (ist) und del' Konjunktiv seinem Pratel'itum, d.h. dem Optativ Platz machen (muB)" cannot be regarded as adequate to cope with this difficulty. That I cannot agree with Musiií in considering the opto to have been a preterit-­"zu dem ihn seine Form (sekundare Personalendungen) und sein Ge­brauch unwiderleglich stempeln"-, and, especially, a syntactic preterit ofthe subj. will be obvious to those who have read the previous chapters of this book.

The so-called conditional sentences occurring in the ancient languages have very often been described in too schematic a manner. EspecialJy the traditional school grammar has, as far as I am able to see, laid too much emphasis on the constructions prevailing in the classical periods, considering the unsettled usage of the earlier periods as exceptions and· anomalies. Thus Goodwin in his detailed and still indispensable discussion of this subject4) based his rules on the classical idiom, adding to each chapter a paragraph on Homeric, 01' Homeric and other poetic, "pecu-

1) Music, o.c., p.14f. 3) Music, O. C., p. 24f.

2) Music, O.c., p. 18ff. 4) Goodwin, O. C., p. 137ff.

The conditional sentences 171

liarities". Classical scholars are apt to follow this practice acquired in their school-days and almost necessary in courses for beginners. The relevant facts are, on the other hand, very often presented, in accordance with the time-honoured schemata arranged by logicians; in teaching them in this way grammarians have also confused logical and grammatic­al terminology. Goodwin, in his introductory remarks on the classi­fication of conditional sentences, distinguishes, inter alia, those present and past conditions which imply nothing as to the fulfilment of the condition from those which imply that the condition is not or was not fulfilled 1 ). In the very recent book of Humbert 2) a distinction is made between "réalité", "éventualité" and "possibilité", the latter category including also such cases as si xs'iear; lixo'V fJ6sr; "if cows had hands". After having described those constructions which answer to the logical schemata and focussed their pupils' attention mainly on specified com­binations of modes and conjunctions, and on definite pairs of protases and apodoses, grammarians often present the remaining cases-which are very numerous almost everywhere-under the heading "mixed constructions", stigmatizing even some of them as "not strictly logical combinations"3). Nor does it seem to serve a reasonable purpose to start from suppositions based on too small a foundation of facts: why should we, for instance, as sume that at some early stage Greek had two perfectly analogous forms in future conditions, one with two subjunctives, and one with two optatives, the particle us being allowed in both of the conditions and conclusions 4)? Thus the description of these constructions is not rarely to a high degree unhistorical and, from the linguistic point of view, unsatisfactory.

It may therefore be of some use briefIy to survey the main relevant sentence types in the principal ancient l.-E. idioms according to compara­tive and historical lines, omitting many details without comment ancl clwelling mainly upon the controversial matter 5). In so doing we shall endeavour to establish the function of each element of the constructions (conjunctives, particles, modes, aspects etc.) so far as possible indepen­dently of the others, in consideration of the fact that in those early documents which are not too fragmentary, any combination appears to have been possible. As special combinations are, from the very nature of the double thought expressed by these sentences, as a rule more numer-

1) Ibidem, p. 139; cf. also p. 188 (~498). Of. e.g. also Kaegi's school grarnrnar (P. V. Sormani, Kaegi's Grieksche Grammatica3,"Groningen 1915, § 179ff.); Speyer, Sanskrit Syntax, p. 377; Reichelt, Awest. Elem., § 79I.-Hirt, Idg. Gramm., VII, p. 177 cannot convince me of the correctness of a classification based on temporal principIes.

2) Humbert, Sy;nt. gr.2 (1954), p. 219ff., cf. however § 359. 3) See e.g. Goodwin, O.c., p. 190 (§ 504). 4) Thus Goodwin, o.c., p. 144. 5) The reader may be referred to the surveys in the well.known handbooks,

almost aH of which are quoted in this treatise.

! 11

Page 88: Gonda.1956.The Character of the Indo-European Moods.pdf

172 The Indo-European moods

ous than others, the 'regular' types as they mainly appear in classical writings of a considerably intellectual standard stand out, however, already from the beginning, by their comparatively high number.

As to the so-called realis expressing a simple statement of a present 01' past particular supposition, implying nothing as to the fulfilment of the condition, the indicative in the protasis is wide-spread: Eur. Bell. Fr. 294 el {}eot 7:t ~eéíJaw alaxeoy, ov'X elaiy -&eot1). This el protasis was, among others by Van Pottelbergh 2 ), assumed to have arisen from an independent wish: whereas in the opinion of this author A 59 (opt.) clearly expresses a wish, and in Z 50 (opt.) though expressing a 'hypothe­tical relation' the underlying wish is still perceptible, 1JI273 though based on a wish clearly expresses a potentialis; E 897 (ind.) is an unreal hypothesis, the wish being unrealizable, and A 564 (ind.) no wish is implied and a positive fact stated. This theory, though ingenious, is needlessly complicated, and in my opinion superfluous. The aboye passages represent a collection of the various nuances which can be con­veyed by this construction, not, I am sure, a series of stages of develop­ment.-In many cases el has the force of "just as (equivalence in manner etc.)" LI 321; that of "as surely as", of "as, seeing that": A 39; the well­known correctness of the protasis warrants the correctness of the main assertion; cf. in Sanskrit, Mbh. 3,297,98 yadi1 me2 'sti3 tapas4 taptam5

... pU1Jyii,sstu7 sarvari8 "ifl 12 have3 done5 penance4 ••• the night8 must be7propitious6" •

The apodosis can contain an imperative: Soph. Ant. 98 aAA' el ~o'Xe¿ aOt, a'íe¿xe; Z 142; 1;tV. 7, 56,15; Av. Vend. 21, 5; Goth. Rom. 13,4; a subj. Soph. Ph.526 (exhortation); 1;tV. 1, 161,8; Av. Yt. 10, 136 3);

cf. also 1;tV. 8, 8, 6; a future (with ay): x 50 4); cf. also Mbh. 5,186, 35f. In Plautus this construction is with nisi usually resorted to in (minatory) commands referring to the intermediate future 5): Amph. 357. Similarly with future forms in both clauses: f-t 382f. A future in the protasis ("if in sooth") can correspond to a past tense in the apodosis: O 185f. The occurrence of the future in a protasis in cases such as E 350 does not prove, as Walter believes 6), the absence of any dift'erence between this form and a subjunctive; this is not to contend that the speaker alway views the process as "sicher eintretend": in the passage quoted he simply

1) The indo is sometimes found in a general condition: Goodwin, o.c., § 467. For particulars see ibidem § 402ff.; Kühner-Gerth, O. C., II, p. 466ff.

2) Van Pottelbergh, O.C., p. 12ff. Nor can I completely follow him on p. 11; the standpoint of Meillet and Vendryes, Traité de gr. comp.2, p. 655, § 947 is not convincing either.

3) For Oscan and Umbrian cf. Buck, O. C., p. 220, § 320. 4) See also Kühner-Gerth, O.C., I, p.209.

. 5) A survey of the combinations found in Latin: (Leumann-)Hofmann, o.c., p.773.

6) Walter, o.c., p.58f.

The conditional sentences 173

makes a statement about a fact which he considers as belonging to the future: "if you (in the future) shall ... ". If in Homer words spoken by the partner are repeated, they are, in an el clause, regularly in the fut. ind.-el being sometimes followed by ~~-, irrespective of the verb forro by which the future was denoted in the original utterance, "oft'enbar deshalb weil er 'Xe im Konj. zur objektiven Wiedergabe del' AuBerungen eines Andern nicht zugeeignet war" 1); cf. e. g. O 185f. (: 179f.); rp 168ft'.; g 153 (: 148f.). Thus a great variety of combinations were in more 01'

less frequent use. The subjunctive 2) in the protasis has its ordinary fundamental char­

acter. The apodosis often contains· an imperative, cf. cases such as 1;tV. 1, 48, 15; 8, 13, 21; 32, 6 yádi1 me2 rará1Ja7!} suté4 ••• I arád5 úpa6 svadhá7 gahi8 "wenn1 du an meinem2 Safte4 Freude3 hast3 ••• , so komm8

docb 8 nach Belieben7 aus del' Ferne5 her6 !" (Geldner); Av. Y. 31,4; r 281; f-t 53 el ~é '>ee Ataar¡at É'íáeov~ I oE ~é ae ... &~éY7:my running parallel; with "inf. pro imp." E 260. N ot seldom the subj. occurs inboth clauses: J;tV. 1, 123, 3 yád1 adyá2 bhiigd'f(l-3 vibhájasi4 ••• I devó5 nal¡,6 ... ánagaso7 vocati8 "if1 (when) today2 thou assignest4 the portion3 ••• then the god5

declares8 us6 (will d. us) blameless7": both processes are of course in the future, if one would focus the hearer's attention on concepts of time, but-in my opinion-the poet in creating a mental image by calling up the process mentioned does not emphasize this 3). The subj. in the apodosis may be volitional: J;tV. 10,27,2; 1,27,13 yájáma1 deván2

yádi3 saknáváma4 "wir wollen die Géitter2 verehrenl, wenn3 wir es ver­méigen4" (Geldner). I would be guarded in assuming too many cases of mechanical 'modal attraction' 4): cf. 10, 2, 3 aganma1 yác2 chaknáváma3 tád4 ánu5 právo]hum6 "we1 have come in order to carry forward5,6 what (ever)2,4 we are (shall be) able3". See also Av. V. 7, 36; in Greek, with

1) Hentze, K.Z. 42, p. 132. Cf. also p. 136 "Den prapositiven Fallsetzungen mit el und Ind. Fut. entsprechen (in Romer) im Nachsatze futurische Aussagen oder Willenserklarungen im Ind. Fut ... , selten in anderen Ausdrucksformen. Es hat sich mithi11 auch hiel' ein fester Typus von Satzgefügen gebildet, in denen Vorder­und Nachsatz sich genau entsprechen. Diese Satzgefüge aber haben das besondere, daJ3 sie sich vorzugsweise einer Aufforderung oder Abmahnung gegensatzlich an­schlieJ3en, so daJ3 der Sprechende den seinem eben ausgesprochenen Willen ent­gegengesetzten Fall setzt, wo mr eine irreale Fallsetzung eintreten zu lassen pfl.egen. Die hom. Sprache zieht es vor, diesen Fall als wirklich in Zukunft eintretend bestimmt in das Auge zu fassen, um im N achsatz ebenso bestimmt die voraus­sichtlich eintretende Folge hinzustellen." Cf. e.g. B 379; M 248; N97.

2) Classifying the data according to the tenses occurring in both clauses (Rirt) would in my opinion be a far less satisfactory procedure.

3) English usage allows us to say "declare" as we11 as "will d." I have my doubts about the, correctness of "mage" (Delbrück, Ved. Synt., p. 323) or "sol1" (Geldner, l;tg-veda, I, p. 170).

4) Speyer, Ved. S. Synt., § 281 and in connection with l;tV. 1, 27, 13, also § 283. See also ch. VI , p. 105 ff.

I

i

~

Page 89: Gonda.1956.The Character of the Indo-European Moods.pdf

174 The Indo-European moods

&)1 or us cf. e. g. {J, 3001); sometimes with the particle in bot~ clauses: A 137 sl 08 us {J,'Y¡ owO)aw, eyro Oé usv av'to¡; IJ),O){J,m (pers~veratlOn of the force borne bythe particle) 2). Wiyh the opto in ~he ap~dosIS: ~y ~ 4, 4~'.;1 yád1 didyávaJ:¿2 pftanasuS prakrilan4 tásya5 varp,6 syama7 santtara8 aJeZ~9 "when1 (if) the missiles2 in the battlesS begin their play4 ('mental pro­cess'), we are7 (may be: contingent process) your6 winners8 of that5

contest9"; not " ... wollen wir ... die Gewinner sein" (Geldner), al­though, of course, the wish to come off victorio~s is at the ?ac~ of the poet's mind. Of. also Av. V. 4, 44; in Greek wlth the partlcle m both clauses: 1 362f. "if P. grants me (Mm) fair voyaging, 1 may (shall) reach (ZuoíW¡V) ... ": the processes are mentally anticipated, the arriva} ~s anyhow contingent; a 287f.; A 104 etc. S). It seems doubtful w~lether lt lS

correct to believe that this opto is merely a softened expresslOn for the future indicative 4 ); 1 for one would prefer to say that, as far as the verb is concerned there is no conscious feeling of any definite pheno-, , ~, menal futurity on the part of the speaker: A 111; Hes. Op. 48~ Sl uB UBV lh¡l ae6aU¡;, 't60s U8V 'tOl cpáe{J,UUOV etr¡ "but if you plough late, thIS may be your remedy"; Soph. El. 554; Phil. 1260. ~hen howev~r ~he same protasis has a future in the apodosis, the latter IS more defimte m charac­ter: cf. e. g. f3 218 sl {J,8V usv na'teO¡; ... v6a'tov auaúaO), I i1 't' tiv 'tevx6{J,sv6c; , I ' ~, _.Q ~ " I ~ , ,< , nse lJ'tl 'tAaír¡v eVlav'tov' Sl uS us 'tó"U'vr¡O)'tO¡; UuovaO) . .. ar¡{J,u 'te Ol XSVO) ud ent U'tBe sa u'tseet~O) (firm intention); Hdt. 1, 71 (see further on); Xen. Anab. 2, 4, 19f. (notice the psychology): and cases such as A 128; B 364; r 288; LI 415; 1 359 etc. where the apodosis often is more 01' less em-phatical.

After a consideration of the opposite views of Ebeling 5), who regarded

f3ovAeVa0{J,sv B 379 as a fut. indo and ofMutzbauer 6) who took it as a subj.,

Walter7) is not happy in concluding that "es sich um eine Voraussetzung handelt, die in del' Zukunft als verwirklicht gedacht wird", so that "der Streit ganz zwecklos ist", because a subj. as .well as a ~ut. are pos~i~le: like Mutzbauer, he disregards any difference m emphasls 8

). The opllllon

1) I refer to Delbrück, Conj. U. Opt., p.180. , 2) For the frequent slns(! (without a'V 01' "s) see Denmston, Gr. Part., p. 487ff. 3) Eur. 01'. 1147 Idl ya(! ovv cqJrI'V ¡irt, I ijv p,Y¡ 'n' Bust'VD <páuya'Vo'V unaawp,sf}a

contains, it is true, an opto of wish in the apodosis, but un. :nay also be a future (Goodwin, o.c., p. 166; Kühner-Gerth, O.C., II, p. 475 a sub].); cf. Eur. Aeg. fr. 5 el p,Y¡ uaf}é~st, yAwuuav, gUTat UOt "a"á; see however 1149f. .

4) Goodwin, o.c.,p. 191 (§ 505). 5) Ebeling, Lex. hom., I, p. 347. 6) Mutzbauer, Die Grundbed. d. Konj. U. Opt., p.120. 7) Walter, Die Grundbed. d. Konj., p. 58. 8) VanPottelbergh (o.c., p. 32), though upholding a somewhat different view

of the subj., is in my opinion right with regard to the m~re temperan:;tental character of Dem. I, 12 n(!or¡u6p,sf}a as compared with 3, 9 n(!owp,sóa.-Enghsh phenomen~, described by Jespersen, A modo Engl. Gr., IV, ch. 15, 9 and 18,4 are to a certam extent comparable.

The conditional sentences 175

expressed by Kühner-Gerth 1) that the speaker in using sl + fut. indo "von del' Wirklichkeit geflissentlich absieht, den Satz als eine rein logische SchluBfolgerung ausspricht, wahrend el' mit Báv C. conÍ. auf die eventuelle Verwirklichung hindeutet" cannot be maintained. Hdt. 1, 71 sl vlu~aslc;, 'tí acpsuc; .anule~asal : .. ,' fJv vlur¡{}ñ¡;, flá{}s Dau ... a.nof3uABslc;, can hardly be quoted m substantIatlOn because, here too, the main difference consists in the different view of the relation between the processes and reality to be suggested to the hearer: "this use of the modes characterizes the c?urtier" 2). Thus this variation in mode can be utilized to express different degrees of probability: cf. 3,36 si fleV flS'tuflsA~asl 'teO K., oE oe ... Aáfl'1fJov'tUl, fJv oe fl1] flS'tufl8Ar¡'tm . .. , 't6'te ua'taxe{ia{}m-. 'Ohantraine 3)

does not seem to be correct in saying that the hypothesis is indefinite in cases like E 258; X 86; s 221; the occurrence oí n¡;, no'ts and similar words is a thing by itself. Generally speaking the pI'. subj. is used if the processes expressed in both clauses are taken as synchronous (Plato, Lach. 201 O fj~O) nuea ae aVelOV, eav {}soc; e{}ékn), the aorist ifthat contained in the protasis is regarded as preceding (LI 169 ulvov (1xoc; a8{)Bv 1!aas'tal . : . I al UB {}áv'(j¡;): He~e the subordinate clause essentially poses a condi~ tlOn, of no duratlOn, lt does not mainly express futurity 4). Oompare in Ve~lC, ~V. 3, ,31, 6 vidád1 yádi2 . .. sadhryak3 lcaJ:¿4 "if2 she finds1 (in Latm: tnvenent) ... she shall make4 rightS"5).

Sometimes a pI'. indic. occurs in the apodosis to suggest instantaneous r~aliza~io~ ,01: the, conti~uat~~n of the yroeess: Eur. Andr. 381 fJv {}áv'(jc; av, nUl¡; 00 sUcpsvysl floeov ... the hfe of this child remains spared'" ef. ~V. 8, 61, 10 ... yádi1 me2 8!1:/'ávad3 dhávam4 I ... st6mair5 índ1'am~ havfhnahe7 "if! he should hear3 my2 ea1l4 ••• , we invoke7 lndra6 with hymns oí praise5

". See also Xen. Oyr. 5,4, 35. (Of. Latin instances sueh as Plut. Baeeh. 364.)

Very often however a present in the a podosis helps to express a eustom­ary 01' repeated aetion 01' a general more 01' less <timeless' truth: cf. ~V. 4, 41, 3; 5, 3, 10; 6, 25, 6; Av. V. 13, 40; ~ 372 "1 dwell aloof with th~ s~e, nor do 1 go (1!eXoflm) to the town, unless P. bids ((J'tIlVVUaw) tlllther ; cf. alsoA 391; Eur. Ale. 671; Dem. 2, 12 "all speeeh, if deeds are wanting (anñ), appears (cpatvs'tul) mere emptiness and vanity"6). Henee the use of this eonstruction in similes, the subj. in the conditional 01'

coneessive-eonditional protasis ealling up ideal existenee, the present in the apodosis intimating that given the circumstances mentioned in the

1) Kühner-Gerth, o.c., II, p.474. 2) B. A. van Groningen, Herodotus' Historien, III, Leiden 1946, p. 42; cf. also

p.228. 3) Cha,ntrainll' Grarnm. hom., II, p. 279 (§ 410). 4) As lS held by Hirt, O.c., VII, p.179. 5) For Latin see e.g. (Leumann-)Hofmann, O.C., p. 773. 6) I also refer to Humbert, o.c., p.220 (§ 357).

Page 90: Gonda.1956.The Character of the Indo-European Moods.pdf

176 The Indo-European moods

protasis such and such a process usually occurs: JI 263; 1481; (fJ 576;

X 191 etc. 1). The question remains to be posed why aY came to be practically oblig-

atory in sl clauses containing a subjunctive, Homeric and poetic usage often omitting the <modal' particle2). The answer seems to be that a process which exists only in the mind of the speaker is dependent on certain conditions and circumstances in a larger number of cases (remem­ber the 'subsequent' subj.!) than actual or merely contingent processes.

An opto in the protasis with a present indico in the apodosis expresses a general statement, which is not seldom definite in character, depending on a contingent (which often means: more or less remote, uncertain, problematic, undecided) condition: a 414 ov-c' oVv &:yysAtn sn nst{}o¡tm, si no{}sy s}.{}Ot "no longer do 1 put trust in tidings, whensoever they may come" (Murray"); r¡ 52; Isocr. 14, 39; sometimes the enunciation is even emphatic: cf. also {} 138 ov yae syOJ "lB n q;r¡¡tt ... al . .. sir¡3). Does inA 389

4)

ovu dAByOJ, roe; si ¡ts yVY~ f3á),ot the opto help to give utterance to the speak­er's utter indifference: " ... any more than if a woman has struck me (or not even struck me)"5)~ Ancient Indian instances: AiBr. 3, 46, 6 tWf!¡,l yady2 ete~árp,3 trayár,¿ám4 ekar(bcid5 akámam6 abhyábhavet7, tasyá

Bsti

9

vámadevyasya10 stotrell práyascitti7}12 "if2 against his desire6 hel

should have7 one5 of these3 three4, there is9 in the stotra1l of the Vamadevya

10 an

expiation12 for it B"; ~V. 5, 74, 106); Varah. Yog. 1,4 syácl

chidram2

ekam3 api4 cet5 tata6 eva7 sarvarp,B násarp,9 prayátilO "even4 if5

there should bel one3 weak point2, then6 allB is destroyed9, 10"; Kal. Kum. 6, 61 "1 know nothing 1 could do for you; if there should be (syác cet kim), all is at your disposal" (kirp, nopapadyate) 7). In the corresponding Gotbio sentences the process mentioned in the apodosis is likewise represented as a fact: Gal. 5, 11 (sl nset-Co¡t~Y ur¡(!ÚaaOJ, -ct sn (juhuo¡tm) jabai bimait merjau, dulve panamais wrikada (conc.-cond.); Luke 6,33. It may be doubted whether BehaghelB) is right in assuming that "das Got. mit seinen Fallen des Konj. Pras. (i. e. an opt.) nichts Ursprüngliohes be­wahrt hat": the Greek, Indian, and Gothio sentences run parallel, and

1) For (hav + subj. see Schwyzer-Debrurmer, o.c., n, p. 684, n. 1, 1. Incorrect

Hirt, o.c., VI, p.275. 2) For particulars see e.g. Kühner-Gerth, o.c., n, p. 474, See also Schwyzer~

Debrurmer, n, p. 684f. 3) Humbert, p. 223; Ohantraine, O. C., n, p. 277; and Schwyzer-Debrunner,

o.c., n, p.685 are silent on this point. 4) See Ohantraine, o.c., n, p.277 (§ 407). 6) For the difference between an opto and a fut. in the protasis cf. A 59ff.

("should we even escape death" rather than "pour voir si ... ", Ohantraine, p. 277). 6) See also Delbrück, Ved. Syntax, p. 348. 7) Not: " ... was würde wohl zu tun unmoglich seín?" (O. Walter, Der Kumara-

sarp.bhava, München-Leipzig 1913, p. 58). 8) Behaghel, Deutsche Syntax, nI, p. 640.

The conditional sentences 177

the .concessive nuance is not foreign to the other languages l) N d Latm stand 1 f· L' 10 19' . . or oes . . ~ oo.. IV.., ,17 (m an archalzing prayer) Bellona si hod~e nob~s vwtor~am du~s, ast ego templum tibi voveo' Plaut Mer 4' 30 at ego s· l' . d ' . C. ~ .ve ~m, .wm antur; Rud. 566 amare utramvis possum, si probe adpotus swm; MIl. 685 2); as already observed by Hofmann 3) the emph t' charact.e~ of the ap,odosis is often unmistakable: Plaut. Truc. 232 n: ~: sunt, n~s~ ... dare ~am lubeat denuo. g . vy e may now attempt an explanation of the occurrenoe of the opt .

sImIles and comparisons, or, in general, in "as if" clauses. In Satr:n

ll, ~,6, 6_ sal yathá2 nadyai3 párarp,4 parápasyed5 evam6 svasyá7yusahB p~~ rarp, paracakhyaulO "he sawlO the limit9 of his7 lifeB a~ if2 one sees5'the fa ther bank4 of ~ rivera" 4) '. the process of the yathii-clause is contingento I: other cases thlS process lS at complete variance with established fact see ~. g. ll, 5, 7, 10 "the waters, the sun, the moon the asterisms mo:~ (p:: md.) yathá1 ha2 vaa etá4 devatá5 ne6yur7 naB kuryur~ evamlO haill a12

brahma~wla bhavatil4 "as lil these4 divinities5 did7 ~ot6 'move7 ~r a~t'9' thuslO isl4 a br.ah~anla (when he does not study)". Having no othe; ~erb .forms at hlS dlsposal, the author has, in referring to a process which

e wlShes to represent as neither actual nor merely mentally conceived recourse to the opto which enables the hearer to understand that the pro~ cess, at least, may be unreal. Compare in Greek: A 467 (contingent); e 366 náY-co~s xs~f 0eBYOJY, roe; sl5) n-cOJxoe; ná),at sir¡ (originally si had the'above m.eanmg) as though he had long been a beggar" (which is at variance ~lth real faots): u .416 .. Thus this turn of speech is also suited to express­mg referen~es to lmagmary processes in illustration of the contents of the ~POdO~lS: B 780; X 411; ¿ 3146), and in Sanskrit, Ram. 2 63 12 Cf., m Latm, velut si eto. ' , . I~ Avestan. and earlier Greek the apodosis incidentally contains a

SU?J~; A 386 (If you would make trial of me (opt.)), ovu aY -Cot xeata¡tna¿ f3tOe; then yo~r b~w wou~~ not (in my opinion,) help you"; (j 388; a ';01. or adhort. subJ. (wIthout ay) P 893. Afuture is not rarely emphatic: 1388

1) For a concessive force in Latin si-clauses see Kriege De t' t' cessivis apud Plautum et Terentium, Halle 1884' H O Nu'ttrn' g eSntund~a l~ ctohn­sil U' , .. ,UleSln e el-~na~:~ek:n~ü~nf O~GI., tOhl. Phil., 111,2, p. 35ff.; Bermett, 1, p. 78f. For concessive

er er ,o. c., , p.489f. M 2) See H. Bl~se, Der ~onj. im bedingenden Satze, Glotta, 11 (1921), p. 152ff.·

f. Schuster'lWlen~r,Stu~ren, 44, p. 122 on the occurrence of a fut. in the apodosi~

o an unrea condltlOn rn Latin 3 • 4) Hofmarm, in Leumarm-Hofmann, Lat. Gramm.5, p.774f. 6) Of. als? ~elbrück" ~onj. u,,?pt., p.23lf.

dos~S,~h(~re dlS l?- my oprnlOn no unconscious suppression of the verb of the apo-1 00 wrn, o.c., p. 179f.). For w,; see Lingua 4 p 9ff It do s t ne t d ' , ..- e no seem

S cesslary o evote many pages to this construction (cf. Lange O.C. p 433-441)

ee a so Mutzbauer, o.c., p. 122. ' ,. . .6) We can pass over the cases of a subj. (1481) or indo (N 492) in the protasl's

wlthout comment. 12 Gonda, Indo-European mooda

Page 91: Gonda.1956.The Character of the Indo-European Moods.pdf

--~------'-------~---,--------- ,-,---,----

178 The Indo-European moods

, s', , I ovo' si "the daughter ... I will not wed, not uover¡v u ov yafhsOJ • • . . . . h t' t d th h'" K 222 1). in Latin 2) the more 01' less emp a lC presen an f ~ug c~me to gain the majority in Tacitus; in Skt. cf. Mbh. 3, 72, 38. uIureG k the contents of the apodosis may be qualified by av: fh 345ff. si

n ree, - 1'" ill t . ht oé usv sir; 1{}áur¡v acptuotfhef}a .•. I al1jJa usv . .• -r8V~ofhsV we w s rarg , -( k the adverb al1lJa) build (a rich temple)"; the speaker however way mar T ill t) th .

continues: BV oé us {}sífh8'JI ayáAfha-ra "and put (not:. we. w pu erem 1 . g gl'fts'" the wave of animation and determmatlOn ebbs away. '

p easm . d'd h th < S t 'mes the form alone3) does not a110w us to eCl e w e el' a

ame 1 "h will t . ht " subj. 01' a fut. sense' is meant: e 539f., where I prefer e s ralg way to "s. would he" (Murray). .

An imperative in the apodosis seems to occur only sporadlCa11y:

RV. 8, 73, 54). ~ 1 t . The opto in both parts of the sentence often helps ~o orm~ ~ e _a

. . d' t' 5). TS 2 1 4 3 yádi1 bibh~yád2 duscarma3 general mstructIOn 01' !rec IOn. .", bh . ~ -4t'S syamám6 á7 labheta8 "if1 he (i. e. any person whom av~syam~ ~ ... " 11 b 4-

this i~formation might concern at any moment) fe,~rs2. I s~~. ec?me diseased3 in the skin3,5, he should offer 7,8 a dark ( ammal) .. the mde­finite and general use ofthe mood. In an interesting passage, AIBr. 2, 7, 2ff. it reads' sorne authorities say that he (anyone concerned) should not make ~ention of rak~ases, others however maintain that he . sh~uld make mention; then the author observes: "if2 ~e1 ,~ak_e me~tl~n he should do S05 inaudibly4": sal yadi2 kirtayed3 uparq¿8~ hrtayet: m the protasis a case is put which will no doubt often arlSe, ~ut may, also under the influence of the authorities last mentioned;, fail to, turn ~p. See also Manu 2, 243, etc. In Gothic: e. g. Matth. 5, 41 (00'-r~r; .O's ayyaes~O'St fh tAtoV /lv, {5nays fhs-r' av-rov ovo) jabai /vas puk ananauPJa~ rasta a'bna,

• ~h 'mma twos I am afraid that I cannot fo11ow those who held gagga~s m~J! ~. . h . that the opto in the protasis is due to attractIOn 6); compare suc mstruc-tive passages as 1 COI'. 7, 8f. 7

). , 2

With regard to the frequent turn of speech l;tV. 8,44,23 yad1 agne h ~ -'9 h '10 '11 te12 satyá13 ihá14_ syám3 ahárq¿4 tvárq¿5 tvárq¿6 vá7 g a8 sya a am SY~~I10' Th 12 15

,. h15 "'fl O Agnl'2 14 were3 thou5 01'7 thou6 wert , y prayers s~sa 1, , 2 8 w~uid now (in this case)14 bell granted13" (cí. also 1, 38, 4; 7,3, 1 ;

1) In Dutoh "zal" not "zou" in the apodosis (of., however Van Pottelbergh,

0.0., p. 5). . th th th" ormal" 2) The term "Inkonzinnitat" applied to oonstruo~lOns o er . an e n t "t 8;t (Hofmann 1. o.) may lead to mISunderstandmg. ype8~~.... '. ", tW

3) That is to say: without takmg stress or mtonatlO~ mto aoooun . e mar com are the emphasized verbal forms in Dutch, Enghsh, German, e. g. 1 don t kno! when 1 8hall die, but 1 do know that 1 8hall die 80me day (Jespersen, Mod.

Eng1. Gr., IV, p. 283). , B t 1904 221 4) See also O. D. Buok, A Grammar of Osoan and UmbrIan, os on , p. . 5) See e.g. Delbrüok, Oonj. U. Opt., p. 223; Ved. Syr:t" p. 346f. 6) Of. Behaghel, 0.0., nI, p.640. 7) See Str61tberg, 0.0., p.247f.

The oonditional sentenoes 179

8, 14, 1; 19,25), the question may arise whether scholars were right in considering it as an instance of an irrealis in the strict sense of the term 1). A11 depends on the state of mind of the speaker on the one hand and on the definition of the term <irrealis' on the other. According to sorne scholars any statement of a hypothetical situation which is contrary to actual facts is an "irrealis" 2), according to others any supposition which is incompatible with, 01' logica11y opposed to, reality3); whereas sorne apply the name to any construction implying that a condition mentioned in the protasis is unfulfi11ed, others hold that only an unrealizable hypo­thesis must be considered ·unrear 4). The confusion ofideas has increased because in hypothetical statements ·the three divisions of time' are not always easily kept aparto As, for instance, it is genera11y speaking humanly impossible to deny anything with regard to the future, a condition 01' hypothesis bearing upon things to come usually implies nothing else but uncertainty 01' at most probability. Dnder many circumstances, e. g. those conditioned by ·primitive' thought and ·primitive' social life, the same remark may hold good with regard to many processes in the past and in the presento As long as no sharp line is drawn between gods and men, between mortal and immortal, between the seen and the unseen, as long as anything is deemed possible provided the right methods 01' technique are known and carried out, it is not surprising that special idioms to express the 10gica11y impossible and practica11y unrealizable are wanting. It does not, therefore, seem correct to confuse possibility 01' imagination and the negative aspect of realityS).

A sentence of the type represented by SatBr. 14, 1, 1, 19 etárq¿l ced2 anyásma3 anubl'üyás4

, táta5 eva6 te7 síras8 chindyam9 "wenn2 du diesen1 einem anderen 3 mitteilen solltest4, S05,6 würde ich dir7 den Kopf8 ab­hauen9

" is, therefore, incorrectly quoted as a case of irrealis 6): it refers to a contingency depending on another contingency. For us the sentence l;tV. 8, 44, 23 (see above) is on a different plane, but the processes to which it refers likewise belong to the sphere of contingency. A supposition expressed in this way may from the objective point of view be utterly improbable. In pronouncing the l;tgvedic passage the poet did not, I am sure, consciously consider the question whether his supposition-which was, of course, contrary to fact-would ever be realizable. Dnder these

1) Of. e.g. Hirt, Indog. Gramm., VII, p. 181; Renou, Gramm. véd., p. 391. 2) See e.g. Goodwin, 0.0., p. 147; M. A. Pei and F. Gaynor, A diot. oflinguistios,

NewYork 1954, p. 224. 3) See e.g. Sohwyzer-Debrunner, 0.0., n, p. 686 ("als nioht durohführbar hin­

gesteHt"); Humbert, O. o., p. 223ff. 4) See e.g. J. Marouzeau, Lexique de la terminologie linguistique, Paris 1933,

p. 106. Often ho~ever no definition is given at aH. 5) Oompare also Oh. Bally, Linguistique historique et linguistique fran<;aise,

Berne 1944, p.49, n.1. 6) Hirt, 1. O.

12'

Page 92: Gonda.1956.The Character of the Indo-European Moods.pdf

·r---------------------.. -·-------""'I~~,,~~~.J.f--~~~-~-~-~~~~­l.

t

180 The Indo-European moods

circumstances I would avoid using the term <irrealis'. Compare also SV. 1, 82; SatBr. ll, 6,2,6 1). The same ideal conditions can be ex­pressed in Avestan, see e. g. Yt. 8, ll; 24; 56 "if the Aryan countries would think (opt.) of ... then a hostile army will not come (opt.) ... ". Greek counterparts are: A 255 ij usy yr¡{}f¡aat JI . ... sl 'tábs náv"ía nv{}ota'to "P. (and the others) verily would rejoice ... were they to hear all this". The contents of the protasis are at the moment contrary to reality­though there is nothing in this passage to indicate this explicitly-, but by no means unrealizable, and the person speaking-and that is the real point-bases for a moment a conclusion upon the fulfilment of the condition implied in them. Cf. also y 223; Aesch. Prom. 979; Xen. Anab. 7, 7, 11. In these- cases no opinion of the speaker as to the untruth 01' impossibility of the supposition is implied; he only expresses himself cautiously (01' unobtrusively, politely etc.: cf. e.g. s 178f.)2). In Attic Greek this form of a purely hypothetical sentence is consistently distinguished from the irrealis, to which we shall have to revert further on. In Homer and other early l.-E. idioms the same construction could also be used if the speaker avowed his knowledge of the <unreal' character ofthe protasis, i.e. ofits containing a process which is factually inexistent, logically impossible, practically unrealizable, that is to say, this construction originally was also used to express a real (logical) <irrealis' and in part of the l.-E. territory it continued to do so. Cf., in Sanskrit, M"fcch. 3, 18 diparp,l cá2pi3 na4 mar§ayed5 abhimukharp,6, syál7 lak§yasuptarp,8 yadi9 "and2 if9 this (pair) be7 feigning sleep8, it could not4 be bearing5 the light1 straight on (its) faces 6 " (the person speaking has just made a trial of this); in early Greek: 1 515 el p,8V yae p,i¡ Miea cpSeOL ... I ovu av syw ys ... uSAotp,r¡V ... I vvv b8 ... nOAAa bLbo¿3); P 274 sl p,BV vVv snl aAAq.> G.S{}AeVOLp,SV, I ij 't' av syw 'ta neoYr:a cpseotp,r¡v "if we were now contending ... " (so-called unfulfilled present condition) 4). For Albanese the reader may be referred to the comprehensive study by Lambertz 5). The optative is also found in both clauses of the Lithuanian

1) Quoted in full by Delbrück, Conj. u. Opt., p. 233. See also Speyer, Sanskrit Syntax, p. 264, sub d.

2) "Doch kann statt des Indicativs sowohl im Vordersatz als im Nachsatz del' Optativ gesetzt werden, wenn die Rede als eine gemilderte oder gar als ein halber Wunsch ausgedrückt werden sollte" (F. Kurschat, Grammatik del' Litauischen Sprache, Halle 1876, p. 429, § 1598).-In particular cases-by no means always! -the speaker may indeed have, consciously 01' unconsciously, wished what he says in the protasis (cf. e.g. Z 284; cf. 281), but it does not follow (cf. Chantraine, o.c., n, p. 276) that these conditional constructions have originated in wishes (cf. e.g. A 255).

3) Cf. Humbert, o. c., p. 222 (§ 360; Phoinix is speaking); as to Soph. El. 548, cf. also Aesch. Ag. 37 and Lucian., De Luctu 13.

4) See also Delbrück, Vergl. Synt., n, p. 401. For Latin (Leumann-)Hofmann, o.c., p. 773f.; Meillet-Vendryes, o.c., § 950.

6) M. Lambertz, Die hypoth. Periode imAlbanischen, Indog. Forsch., 34, p.44ff.

1 The conditional sentences 181

<irrea~is': kad1 asz2 zinócziau3, ta14 sakycziau5 "ifl 12 knew3 it, then4 I would say5 it" I).

Contrary to Attic usage, in which the original scope of the opto has been limited, Herodotus still used the <optative of subjective suppo­sition' with reference to the past: 7, 214 slbstr¡ p,8V yae av ... 'tav'tr¡v 'ti¡v G.'teunov 'Qvf¡'tr¡q, si 'tñ xwen nOAAa Wp,LAr¡uwq dr¡ "el' mochte (dürfte) gekannt haben ... ".

We now turn to the protasis with an opto followed by an apodosis with an historie al tense. In general conditions the < gnomic' aorist, and past forms expressing a customary, 01' repeated, action may be used in Greek 2): Thuc. 7, 71 sr 'tLvsq lbotBV nn 'tovq acps'tSeovq smuea'tovvruq, G.vs{}áear¡auv av "whenever any saw their friends in any way victorious, they would be encouraged (i. e. they were encouraged in all such cases)" 3) ; Xen. Anab. 2, 3, ll; Hdt. 1, 100 Sl 'tlYU nvv{}ávo¿'to 15{3etCov"ía, 'tovrov s&uutsv; Eur.Alc. 755; Thuc.1, 70; 7,79 etc. (without ay). There is a single case in Homer: Q 768. This rarity in Homer and the seeming absence of this construction in the other languages is the more remark­able as its significance results immediately from the force borne by its components.

The indico of an historie al tense with av can, however, in a small number of cases alsoserve to deny the reality of the result: Eur.Or.1l33

• sl p,8V yae slq yvva¿ua . .. I ;tcpoq p,s{}s¿p,sv, bvauAsi¡q av ijv cpóvoq "if we would plunge the sword into a woman, the murder would be shameful" the speaker continues: vvv b8 ... bwas¿ Mur¡v 4 ). As no satisfactory explanation of this idiom in the ancient l.-E. languages has come to our notice, we shall enlarge here on a remarkable function of the durative past tenses in general. The phenomenon in itself, the differentiation, in Greek and other languages, between the expression of the <irrealis' and that of contingency, may broadly speaking be attributed to the well­known tendency to differentiate, in case of <functional overburdening', under the influence of a cultural development 5).

That an infective durative past tense, which is often used when the person speaking dwells upon a process 01' takes a particular interest in it, can serve to denote, without temporal import, imaginative thought and vivid visualization, is a thesis first put forward by the Romanist

1) "Gegenteiligkeitssatze": Kurschat, o.c., § 1600. Cf. also A. Schleicher, Lit. Gramm., Prag 1856, p. 335ff. ("Opt.: reality in tha- past 01' mere supposition in the present").

2) See also Goodwin, o.c., p.462ff. 3) For the hist. tense with Ü'v (the particle referring to "certain conditions 01'

circumstances") see ;Kühner-Gerth, o.c., 1, p. 211; cf. n, p. 476. 4) N 825 quoted by Hirt, O.C., p.182 is foreign to the matter under discussion. 5) An opto and a past tense can, in the apodosis of an unreal condition, occur

alternatively: § 184f.

I

Page 93: Gonda.1956.The Character of the Indo-European Moods.pdf

182 'The Indo-European moods

Lorck 1), and emended, amplified, and vigorously ma~ntained by ~is collea­gue Lerch 2). According to the latter scholar the pnmary functlOn of the French imparfait was "einer lebhaften Vorstellung des Vorgangs Aus­druck zu geben" 3): Lafontaine, Contes IV, 401 G. courut a la chambl'e des femmes, n savoit que l'infante étoit dans ce vaisseau; et ... ~ll'emp?rtoit comme un moineau (in narrative style the impf. often occurs m emotlOnal passages4) and especially in connection with those situations and pro­cesses which are apt to stimulate the speaker's imagination5

)). The speaker, seeing the occurrences from acertain distanc~ and a certain perspective, dwells upon them in his mind 6). In relatmg dreams and

1) E. Lorck, Germanisch-Romanische. Monatsschr~ft, 6, p. 186ff.: "Das Iml?erf. (the author is writing on the French lmpÍ;) hat slOh durchweg m all~n semen verschiedenartigen Verwendungen als eine AuJ3e~gsfo:m des Phan~asledenkens (1 would prefer: imagination) nachweisen lassen. D.lesem l~t das Verweüen. auf dem Gedanken, die Auffassung des Geschehens oder Sems als 1m Verlauf ?egriffen und als unabgeschlossen vorliegend eigen. In diesem Phantasiedenken 1st auch del' Dauerwert des Impf. begründet, nicht in del' objektiven Dauer des Ausgesagt?n. Wenn es ZUlll Ausdruck gedanklicher Beziehungen dient, wenn es anschauhch erzahlt wenn es in die innere Welt del' V orstellungen und Gefühle versetzt, so sind dies alies Funktionen, zu denen es kraft seiner Eigenschaft als Phantasiedenkart berufen ist."-'The qualification "Mischformen" applied to E 897f.; Plato, Apol. 31D (Schwyzer-Debrunner, o.c., n, p. 686) may lead to misunderstandings.

2) E. Lerch Das Imperfektum als Ausdruck del' lebhaften Vorstellung, Zs. f. rom. Philologi~, 42 (1922), p. 3Uff. and 385ff.; the same, Historische franz. S~~tax, n, Leipzig 1929, p.222. See also Meillet, BSL. 3? ,(comp. rend.), p. 136: avec l'imparfait on présente les choses comme non realisees, en les transportant dans un passé imaginaire OU en effet elles ne pouvaient l'etre." 'The same ~uthor quo~es from Merovingian Latin: si p¡'evalebant Ohuni haee ... defende¡'e, med~etatem Galltae ab Ageeio perciperint. . . . . .

3) Lerch, Zs. f. romo Phil., 42, p. 318. In descrlbmg dreams 01' vlsIOns thlS form is often preferred. Compare also the Greek idiom Eur. Ion 1501 (see Schwyzer­Debrunner, o.c., n, p. 280, 8).

4) Some Rumanian and Portuguese instances (without a satisfactory e:s:pl~nation) may be found in W. Meyer-Lübke, Gramm. d. romano Sprachen, 111, .~elpzlg 1899, p. 123 (§ 104); for Spanish see Lerch, o.c., p. 331 (cf .. e.g. Don QUIJote, 1,6 que me place, señor mio, respondia ella, y con mucha alegria eJecutaba lo que era mandado); for Italian Lerch, o.c., p.385f.

5) 'This impf. is e.g. very often used in descr~bin~ milital'y eX1?loits. ando oc­currences: Tout cil qui estoient ataint .•. , se restra~ndMent. La pe1'do~en:t lt plu~seur force et alainne, et tresbuchoient l'un SUB l'autre, et s'estindoient et morMent sans cop férir (Froissart, quoted by Lerch, o.c., p. 386f.). AIso ifthe verb expresses a move­ment, speaking, crying, asking, calling etc. (Lerch, 'P' 392ff., who ~lso y draws attention to Latin instances: cf. e.g. Vulgo Matt. 27, 23 m accordance wlth BUea~ov etc. in the Greek texto In Greek ASYBtv, UeABVBW, eeon:av, n8l1}etv, nS¡¿n8tv often are in the imperfect (see Schwyzer-Debrunner, o.c., 11, p. 277f.). .

6) A propensity not e:x:actly to imagination but to mental reproductIOn, to dwelling upon the occurrences spoken of, viewing them not a~ historical events but as scenes existing in the mind, is no doubt an element m the remarkable predilection for eulÍA6v8v, l!n8t{}8v etc. in Greek (see A. Svensson, Zum Gebrauch del' erzahlenden 'Tempora im Griech., Llmd 1930; cf. e.g. 0/539; Ll104),

'The conditional sentences 183

visions the Romance languages have a distinct preference for the same form of the verb 1).

Uses of the imperfect which are related to the imaginative preterit were not foreign to the ancient l.-E. languages 2). Leaving those many instances undiscussed in which the imperfect appeals more to the eye than to the intellect 3 ), we first call attention to such cases as are fre­quently.found in the Vedic hymns. In saying ~V. 2, 12,2 yá7y,l prthivt'f]'/,2 vyátharniinam3 ádr'f]'/,hat4 "WhOl made firm4 the quaking3 earth2 " the poet so to say invited the he arel' to dwell upon the process executed by the god,not to see it as a deed completed in the pasto We may to a certain extent also point to the Greek idiom illustrated by Eur. Ion 1501 8; sftov ,.' oVX S(n' lif}v'llauer; "und dich sah ich im Geiste ohne Recht durch meine Hand dem Tod entgegengehen"4).

What interests us much is that the usual occurrence of the impf. in conditional clauses can be explained in a similar way: it may be considered a preterit of imagination, preferred because it enables the speaker to express 'mere thought' 5): certes, si nous avions passé huit jours ensemble, je l'épousais (= "1 would have married her"; Maupassant) is, in a way, equivalent to ". . . imagine: 1 married her", the situation being so to say in a vivid and emotive way viewed by a witness in the past who did not know that the condition was not to be fulfilled. Anglists were not long in establishing a similar theory with regard to English 6).

This preterit can, however, occur also to indicate 'unreality' in the present time. Cf. e.g. Shakespeare, COI'. 3, 1, 19 1 wish 1 had a cause to seeke him there; Marlowe, H. a. L. 2, 53 T.B. supposin(J nothi'(lc(J else was to be done; Fielding, Works2 8, 641 a murder behind the scenes will affect the audience with greater terror than if it was acted before their eyes; cf. also the construction represented by Stevenson, M.M. (1896), 281 it is

1) "Del' Konj. würde ebenlediglich (in logischer Weise) die Unwirklichkeit del' getraUlllten Geschehnisse ausdrücken-das Impf. dagegen bringt (in anschaulicher Weise) ZUlll Ausdruck, daJ3 dem 'Traumenden diese Geschehnisse lebhaft VOl' Augen stehen, und es ist bezeichnend, daJ3 die Volkssprache sich für diese zweite Mog­lichkeit entschieden hat" (Lerch, O. C., p. 405).

2) See also Meillet, Bull. Soco Ling., 30, (C. rend.), p. 136. 3) B. L. Gildersleeve, in Amer. Journ. of Phil.,23, p.250 "the aorist appeals

more to the intellect, the impf. to the eye". For the graphic and narrative impf. the reader might be referred to Schwyzer-Debrunner, o.c., n, p. 276f.; Humbert, O.C., §235ff. (esp. §238); Debrunner, in "Volume ofeastern and Indian Studies" in honour of F. W. 'Thomas, Bombay 1939,. p. 7lff.; Jespersen, Phil. of Gr., p.275ff.

4) Schwyzer-Debrunner, p. 280. 5) See also A. 'Tobler, Vermischte Beitrage zur franz6sischen Grammatik, n,

Leipzig 1906, p.158 "(etwas) bloJ3 Gedachtes". 6) See e.g. M. Deutschbein, Sprachpsychologische Studien, Cothen 1918, passim;

cf. also Jespersen, A modern English Grammar on historical principIes, IV, London 1932, esp. ch. 9. '

Page 94: Gonda.1956.The Character of the Indo-European Moods.pdf

184 The Indo-European moods

high time that the omission we1'e supplied; cf. also: and what were the du­rability of love without the powerful alliance of habit?I)

This use of the <imperfect' is no doubt widespread. When Dutch children in the heat of discussion, for instance in debating the question which game will be played next, suggest a plan, they often resort to the imagina­tive imperfect: nee, we speelden schooltje, en ik was de juffrouw en juUie waren de kinderen (litt. "no, we played at schools, and 1 was the teacher, and you were the children"). Here the same explanation may hold good, although there may perhaps also be room for the following considerations : As is often the case when in emotional speech the speaker enters the world of imagination his thought guided by his fantasy anticipates the events 2), the boundary-line between actual reality and wished 01' pro­posed existence is wiped out, and the child sees, from a certain distance, himself already playing: his thought having already hurried forward, he describes the immediate future filled with the performance of the much desired game, by means ofpast verb forms. This is not to say that this construction is a parallel of the construction-a favourite of Dutch shop-assistants-wat wou u gehad hebben? litt. "what would you have had1" instead of: wat wilt (01': wilde) u hebben?3). When a Dutch child indulges in fancies it very often chooses the narrative (rather: the imaginative) imperfect,

A peculiar, but sporadic use of the imperfect in Sanskrit may, as far as can be seen, likewise be explained as an <imaginative': Ram. 3, 81 (79), 10 saumitre1 mám2 •• , santápayati3 manmathaJ:¿4 I . .. kokil05 mam6

iva7hvayat8 "O S.l, the god of 10ve4 tortures3 (pres. ind.) me2 ••• , the cucko05 is as if7 inviting8 (impf.) me6 ".

1) See Jespersen, a.c., nI, p.647 and IV p. 123f. 2) The reader may for the sake of brevity be referred to L. Spitzer, Jahrb. f.

Phil., n, München 1927, p. 270ff.; the same, Stilstudien, n, Müuchen 1928, p. 258ff.; Ravers, Rdb. der erklarenden Syntax, p.41ff. and 219. In connection with the ,German idiom Schiller 14,365,2060 mit diesem zweiten PfeiZ durchschof3 ich euch, wenn ich mein liebes Kind getroffen Mtte ("Ind. Pras. zur Bezeichnung irrealer Vorgange, die unter bestimmten Bedingungen hatten eintreten kannen, aber nicht eingetreten sind"); Behaghel, De. Synt., U, p. 290 already called attention to "eine Art van voreiliger Erzahlung".

8) For the perf. inf. in this connection see Ravers, O. c., p. 42. The Dutch idiom U kreeg een kop koffie (assertive or interrogative) (lit. "you got a cup of coffee [?]"), frequently used by waiters in bringing the drink ordered, does not appear to belong here: it may rather represent a "you are the one, who (a moment ago) ordered a c.?" (when you ordered, I said to myself, or I noted down: this person gets, i. e. will get a c.: hy krygt een k. k.). We might compare Soph. O. C. 117 'r:t~ ár/ i)v; nov vatet; (cf. 78f.). In discussing the comparable German was bekamen Sie? instead of was bestellten Sie? or was bekommen Sie? W. Rorn, Arch. Stud. Neuer. Spr. 180 (1942), p. 113, and A. Debrunner, Mus. Relvet., 1, p. 46 gave different explanations (result instead of cause and a courteous reference to "die kOlmnende Erfüllung des zu erw&rtenden Wunsches als schon geschehend").

The conditional sentences 185

Another idiom exemplifying the imaginative preteritl) is the so-called 'impf. futuri: Voltaire, Zai:re 2, 1 enfin, lorsqu' elle touche au moment souhaité qui la tirait du sein de la captivité (i. e. which, inher opinion, will sether free): the author enters into the feelings of his subject for whom the liberation does not belong to the future, but to the presento Cf. in Latin: Ter. Hec. 171 interea moritur cognatus senex . .. ; ea ad hos redibat lege hereditas 2 ).

In certain phrases, especially exclamations, Swedish has the imperfect where other languages would have the present: det var snallt av dig "that is kind of you" (German das ist nett von dir, Dutch dat is aardig van je). Compare, in French: qu' elle était jolie! Similarly in the idiom of ancient Greek comedy: Arist. Ach. 767 -rov-rt -ri ijv -ro 'TC(!i1y¡tu; Ran. 39 Bbré ¡tOb,

-rov-rt -rt ijV;3) The former passage may be translated into French by "qu'est-ce que c'était encore que cette histoire-la,".

In various languages the <impf. modestiae' is, in certain situations, a means of referring in a modest, unassuming, guarded 01' submissive way to the present4): Moliere, M. 1,3,20 hélas, faut-il queje perde mon pere, la seule chose qui me restait au monde; Becque, Polich. 1, 3 je venais m'informer de son retour (i. e. "1 come ... "); Musset, Fant. 1, 2 je voulais demander ... si la princesse est la cause . .. ; It. Manzoni, Prom. Sp. 52 volevo di1'e, non intendo dire, cioe volevo dire; Dutch ik kwam eens horen of u nog wat te bestellen hebt; ik kom niet om cente, 'k had een plán. The verb form expresses the contents of thought rather than past time. In the Latin dramatists the speaker sometimes resorts to the imperfect in cases which would appear to be similar: Plaut. As. 452 si domi est, Demaenetum volebam "1 wished to see D. if he is at home"; Trin. 195 istuc volebam scire5 ) we also find the perfect: Cas.440 volui Ohalinum, si domi esset, mittere "1 intended to send him, if, .. ". Compare also the Greek idiom sfJovM¡tr¡v "1 would" (in the present); &S<pBAOV "would that" will be discussed further on.

This preterit can also serve in persuasive 01' considerate address: Dutch we zullen het maar doen, dacht je ook niet?; had je z6'n honger, ventje?; ... we waren toch vrienden? In many situations the present would be too blunt and inconsiderate a verb formo The imperfect, removing the idea of the verb from the actual present, keeps the process denoted at SOme distance. The speaker refers the hearer to the realm of imagination

1) "Das Impf. bezeichnet den Inhalt eines Denkens", Lerch, a.c., p. 422. 2) B. J. Wheeler, in Amer. Journ. ofPhil., 24, p .. 174. 8) See J. V. Leeuwen, on the Ranae (Leiden 1896), p. 12 (st. 39): "de re insolita,

absurda, ridicula, quae aculas ferit, ita loqui solent." 4) See e.g. Meyer-Lübke, Gramm. d.langues romanes, UI, Paris 1900, p.128

(Germ. ed. p. 123); F, Brunot, La pensée et la langue, Paris 1936, p. 506; 570; G. S. Overdiep, Stilistische Grammatica v. h. modo Nederlandsch, Zwolle 1937, p.316.

5) Other inst&nces Bennett, o. c., I, p. 36.

Page 95: Gonda.1956.The Character of the Indo-European Moods.pdf

186 The Indo-European moods

and is in so doing, able to attenuate, for some reason or other, the utterance.--The preterit is therefore, for instance in English, also used if an author wishes to dissociate himself from stating a factl).

Very often, if context and situation leave no doubt as to the con­trariety of the imagination expressed by the preterit and actuality, this verb form is a favourite means of indicating unreality. The words il only 1 knew! pronounced by aman who obviously does not know imply, nay indicate, his lack of knowledge, and are therefore often used to express the wish to know. The sentence 1 wish 1 did not have to go means, from the intellectual point of view: "I'm sorry 1 have to go." In Dutch the exclamation had ik maar wat! pronounced in reply to wat heb je?, unequivocally indicates that the speaker does not have and wishes to bring this deficiency to an end.-If however a fact is meant to be ex­pressed the present tense is also used: it is time that 1 give you some idea 01 . .. (cf. : it is time we gave a thought to ... ).

In English such sentences as 1 wish 1 could help you originally had the preterit subjunctive only, and the unreality was denoted by the mood rather than by the tense 2). In the course of time the distinction between the forms of both moods was, with the exception of was: were in the sing.-which was formerly frequent in main sentences of imagination­blotted out, and the preterit became the normal form to indicate un­reality3).

In German "besagen Ind. und Konj. Pras., daB beim Sprechenden kein Zweifel darüber besteht, es liege die Erfüllung der Bedingung im Bereich der M6glichkeit; der Konj. Prat. kann besagen, a) daB für den Sprechenden die Erfüllung nicht im Bereich der Moglichkeit liegt; b) daB (für ihn) die Erfüllung zwar im Bereich der Moglichkeit liegt, daB er sie aber doch für unsicher ansieht; oder aber del' Sprechende gibt aus Hoflichkeit seiner Aussage die Form der Unsicherheit"4). Cf. Nib. 606, 2 des waere lützel not, ob iu daz ieman sagte. Incidentally the reali­zation of the process is, in the eyes of the speaker, out of the question.

Returning now to the construction represented by Eur. Oro 1133 it may in my opinion be asserted that, here too, the imperfect in this context and situation enables the speaker to refer the idea contained in the apodosis to the realm of imagination; in so doing he prevents the he arel' from taking it as actual. We might have expected etr¡ instead of

1) Of. E. Kruisinga, A handbook of present-day English4, II, 1, Utrecht 1925, p.40.

2) Of. Jespersen, a.c., IV, p. 113f. 3) We cannot attempt here to go into particulars and to sketch the historical

development of the l.-E. 'irrealis' constructions in general 01' in part of the l.-E. area. A comparative historical study of the relative German, Romance, Indian, Slavonic etc. constructions would however be an undertaking of gl'eat interest.

4) Behaghel, Deutsche Synta:x:, III, p. 642f.; 777f.

The conditional sentences 187

1]'1', the opto in protasis and apodosis being the usual form of the condition­al sentence. The latter construction would however denote what would be the result if the condition of the protásis should be fulfilled; now the person speaking declares that the result being imaginary does not answer to any realityl).

So the historical verb form serves, in this construction, to express the content of the speaker's vivid thought 01' imagination. But ancient Greek verb forms had an aspectual value. The speaker had to take his choice between a confective and an infective form, and this choice conditioned an alternation of present and aorist forms which eventually resulted in a variety of constructions, so me of which became for obvious reasons

. more common than others. In a case such as E 897f. el CJé TSV 8~ ¿LUov ye f}ewy yéyov . .. I ua{ uey MI náAat 1]af}u. 8yéeTeeOt; "but were you born oí any other god ... , then you had long been lower ... ", the process of being born is viewed as confective; as it is clear from the context that it is antecedent to the process expressed in the apodosis, it is, from the temporal point of view, situated in the past: cf. such constructions as B 513 etc. In Plato, Conv. 180C el fl,By ya.e ert; 1]Y o "Eeillt;, uaAWt; a.y elxe both processes are infective 2). This is the reason why this construction­which in Attic etc. is the normal expression of <present irreality' -was also used when the verbs refer to durative processes in the past3): Thuc. 1, 9, 4 ovu a.y ow Y?}aillY ... 8ueáT8t, el fJ,?} n ua¿ yavnuoy elX8y4). Similarly, Q 713ft'. Instead of the term <Bedeutungsverschiebung' used 5) in connection with the post-Homeric cases of the type Plato Conv. 180C 1 would preíer to say that in literary Greek the function of the imperfect was in the constructions at issue extended to those containing processes which takeplace in the presento The past verb íorms could refer to the real past time and to <imaginary time'.

1) Fol' othel' instances see Goodwin, O.C., 508 to whose judgement (anacoluthon) 1 cannot subscribe. Xen. Oyn. 12,22 el ovv el8eiev TOVTO, 8n 1}eáTat (subj. f¡ a/2ení) aVTov" ZevTo av snl TOV, n6vov, after the author has observed: vno 86 nj, a/2eTf¡, ovu ol6pevot smauonúa1}at noJeJea uaua .•. svavTtov noWVatV.

2) It may be questioned whether Brugmann, Grundri13, II, 3, p. 884, was right in contending that "del' Übergang zu dieser Ausdrucksweise (to wit the use of the past instead of the opt.) sich um so besser vollzog, als die Partikel ... dem Satz seinen optativischen Oharakter wahrte". For the function of the particle see ch. VIII.'

3) In modern Greek the same constl'uction is (impf. prot., condit. apod.) used irrespective oftime; see e.g. A. Thumb, Handb. d. neugriech. Volksspr. 2, Stra13burg 1910, p. 185. In Albanese "konnen die gegenwal'tige lilld die vergangene Periode del' Irrealitat dieselbe Konstruktion haben. Sowohl del' Ind. d. Impf. im Vorder­und Nachsatze, wie del' Ind.lmpf. im Vorders. mit ... Impf. Fut. im Verbals. konnen die il'reale Periode, als del' Vergangenheit und als del' Gegenwart angehorig bezeichnen", Lambertz, Indog. Forsch., 34, p. 75.

4) Of. also J. H. Moulton, EinI. in die Spr. des Neuen Testaments, p. 316. 6) E.g. by Wackernagel, VorI., I,p. 226.

Page 96: Gonda.1956.The Character of the Indo-European Moods.pdf

188 The Indo-European moods

That the aorist does not necessarily refer to the past is evident from cases such as Soph. Ant. 755 8l pi¡ na-r:i¡e rja-&', slnoy aya' OVU 80 <peOY8¿y1). In this passage the aorist points to a mere fact dependent on a condition rather than to a "sinnendes Verweilen" on an imaginary occurrence 2).

Hence 1 would suppose, the occurrence of the two aorists in those passages which in a narrative style-as a rule characterized by a pre~ ference for confective verb forms-deal with 'unreal processes' situated in the past-@ 90f. uat YV USY ly-&' ó yéewy ano 1JvpOY 8A8I1a8Y I 8l pi¡ ae' o;v Y6r¡a8 ... ¿J. "now would the old man here have lost his life, had not D ... been quick to see" 3).

The poet, it is true, only expresses a subjective thought, but does not dwell upon the event conjured up by his imagination. Here the verb form referring to a mere fact, does not allow the speaker to ponder over the process. The aorist could therefore serve to point to reality in the pasto The particles U8 and el fl/1 could suffice to indicate that he díd not 'relate real historical events. These words and the verb form which 'in the beginning' did not refer to a particular time-the Homeric instance quoted contains two unaugmented forms or injunctives-helped him to express a mere thought, an imaginary case, an hypothesis which was. contrary to reality. In 'mixed constructions' such as e.g. P 70 lY-&a U8 ésta <péeOl ... 8l pi¡ ol ayáaaa-r:o <P. the consideration of the force proper to the verb forms and the particles will suffice to understand the poet's intentions 4).

It may be observed that the aboye interpretation of the use of the imperfect indicative in the unreal condition and the explanation sug­gested by Miss Hahn 5), which is based on the occurrence of the same form to indicate an iterative action, are in themselves not mutually exclusive. Yet 1 faíl to see how the iterative nuance contained in the ex­ample ~ 178 uat U8 -&áp' h-&á~' 86Y7:8; eptay6ps-&' ... -where it for the rest is mainly conveyed by -&ápa "often"-could predispose this verb

1) In connection with the Greek aorist the reader might also consult Lerch, Zs. f. romo Phil., 42, p. 413ff.

2) One might compare the Italian construction egli amerebbe < Lat. amare habuit instead of egli ameria < amare habebat in the apodosis of a conditional periodo Significantly enough the latter idiom has been preserved in popular and poetic usage.

3) A similar harmony between the verb forms of both clauses-an indication of the uniformity of thought prevailing in the utterance-is also found in other languages.

4) Cf. also B 81; E 311 etc. In later usage this construction is rare 01' dubious (see Goodwin, O.c., p. 160ff.; Mouro, Gr. Hom. Dial., p. 218; Kühner-Gerth, o.c., H, p. 472; Schwyzer-Debrunner, o.C., H, p. 348). Cf. in German (S chiller) war ich, wofür ich gelte, der Verrather, ich Mtte mir den guten Schein gespart. Cf. also Behaghel, o.c., HI, p.645f.

6) Hahn, O.c., p.90f.

)

The conditional sentences 189

form to the function which it fulfills in the irrealis. It may be true that the particular action denotedin this example was likely to be continued in the present, the processes denoted by these imperfects belong in Homer, exclusively to the past and here too the correct translatio~ is "then, living here, should we ofteri have met" (infective: cf. 174 "in Argos 1 would have buílt a house": l-r:8v;a, confective).

Needless to say that the acceptance of the aboye view of the use of historical forms in unreal conditions can dispose of the distinction made by Schwyzer-Debrunner 1) between a "kupitive Vorstufe" of an irrealis (e.g. ~ 732) and a "potentiale" (e.g. E 679).

A Latín example is Lucil. 150 nam si, quod satis est homini, id satis esse potisset, hoo sat erat. Here sat erat is equivalent to sat esset. Poteram in the sense of "1 could": Lucr. 2, 1035 quid magis his rebus poterat mimbile dioi 2).

Mention must also be made here oí the verbs of obligation, propriety or possibility like l~8l, 8luo; rjY, neoaijUBY etc. which can refer to the past 01' present time, to the aorist W<p8AOY "ought" (and the impf. W<p8AAOY) used to express a present or past unattained wish, and similar cases. The current opinion seems to be that this idiom has undergone a shift from an exclusively past tense to a more or less tenseless device to express the aboye ideas even whert they are situated in the present3).

"Der Gebrauch der ursprünglich nur für die Vergangenheit geltenden Ausdrucksweise auch für die Gegenwart vollzog sich, indem z. B. W<pSA8 ~OV')lal "el' hatte damals die Pflicht zu geben (gab aber nicht)" auch die Bedeu~ung ,:'er hatte langs~ die Pflicht zu geben und hat sie noch (gibt aber lllcht) annahm: der Ubergang fand also statt, wenn die Forderung von der Vergangenheit in die Gegenwart hineinreichte" (Schwyzer­Debrunner). This consideration, wruch may hold good in particular cases, can hardly explain the spread of the phenomenon, which, as we have seen, has its counterparts in other languages. In using a past tense the speaker looks for reasons of reserve, modesty or prudence, away from the present resorting to the realm of thought and imagination.

1 am not convinced that Wackernage14 ) is correct in considering Latin constructions such as Tac. Ann. 1) 63, 6 trudebantur in paludem ... ni Oaesar produotas legiones inst1'uxisset as elliptical, properly meaning: "sie waren daran, in den Sumpí getrieben zu werden (und es ware dazu ge­kommen), wenn es nicht ... ".

1) Schwyzer-Debrunner, o.c., H, p. 349. 2) For a discussion of these Latin constructions in general I also refer to

(Leumann-)Hofmann, O. c., p. 566f.; Blase, Hist. Gramm. del' lato Spr., HI, p. 148ff. I ~m not .entirely convinced by Wackernagel's argument (Vorl., I, p.227).

) See e. g. Schwyzer-Debrunner, O. c., H, p. 308; Chantraine, O. c., H, p. 227ff.; Humbert, o.c., p. 118f.; Wackernagel, Vod. I, p.228.

4) Wackernagel, Vorlesungen, I, p.291.

Page 97: Gonda.1956.The Character of the Indo-European Moods.pdf

190 The Indo-European moods

In Vedic an unreal condition referring to the past is formulated by means of perfect opto in the apodosis: ~V. 10,28, 1 "my brother-in-law has not come (perf.): (if he had come) he would have eatenl fried grain2 andS drunk5 soma4 ••• ": jak§iyádl dhiiná,2 utáS s6mam4 papiyat5• The 'original function' of this not very frequent verb form can scarcely have been other than that of a 'contingent resultative'. Cf. Xen. An. 5,7,26 lffleu1uv p,i¡ A:VO'O'U f¡p,'iv ep,nen'íÓJuo~ "they feared lest madness might prove to have fallen upon US"l).

In Avestan an assumption which is inconsistent with evident facts can be expressed by means of the opto of the perfect stem: Yt. 8, 52-54 ye(Jil zi2 aZ'Jms noit4 dai flyqm5 ... tistrim6 ••• pairika7 ••• ava.hisiflyap "verily2, ifl ra had not4 created 5 T. 6, P. 7 would destroy8" (ava. hisiflyat is perf.; daiflyqm is ambiguous: pI'. 01' perf.)2). In this passage the idea of destroying is presented as an accomplished fact. This use is not foreign to other forms of the perfect: Y. 32, 15 anais a vi. n1Jnasa ya karapo. tas éa k'Jvitds éii (in a series of imprecations "deshalb ist die Karapan- und Kavayschaft dem Untergang verfallen" S). Cf. in Greek e.g. 0128; Soph. Phil. 75.

Some words may be added here parenthetically 011. the syntax of the perfect optative in the Veda, 011. which most authors are silent. It is however worth discussing. There is a number of passages in which this mood seems to have been used to give vent to a certain emotion which prompts the person speaking to resort to anticipation 01' to a 'fait ac­compli'. ~V. 10, 10,7 Yami attempting to seduce her brother says that she wishes to offer her body to him: jayé1va2 pátyeS tanvarp,4 riricyam5 "like2 a wifel to her husbands 1 may have left5 my body (myself)4 to you", i. e., at least before the construction was stereotyped, "1 desire to be in the situation resultant 011. the act proper". ~V. 1,24, 13 an ardent wish is formulated: ávailnarp,2 rájaS váru1JaJ:¡,4 sasrjyat5 "KonigS Varu\la4 moge 5 ihn2 freigebenl , 5" (Geldner), but more 'to the letter': "freigegeben haben", i.e. "may his liberation be an accomplished fact"; 1, 143,6 in the last of a series of three emotive questions: codá~l kuvít2 tutujyátS satáye4 dhíyaJ:¡,5 "ob2 el' wohl2 als Anspornerl uns (fromme) Gedanken5

erregen wirds zum Gewinne4 1": the process of animating is if 1 am not mistaken viewed as having reached its end and having resulted in a certain situation; 5,7, 10 in an emphatic wish at the end of a sükta: ádl agne2 ápr1Jat63 'tri~4 sasahyat5 "may5 Atri4 nowl have conquered 5 the miserss, O Agni2" 01' rather: "may he prevail ayer them, be their master";

1) Other examples: Goodwin, o.c., p. 32. 2) See also Reichelt, o.c., p.386 (§ 792fJ). 3) Reichelt, Awest. Elem., p. 311. Duchesne-Guillemin, Zoroastre, Paris 1948,

is right in explaining "les feront périr" (p. 253) though translating "sont réduites a périr" (p. 256). Cf. also H. Lommel, in W5rter und Sachen, N. F. 1, 1938, p.257.

The conditional sentences 191

this verb is signifieantly often found in this opt.: 1, 132, 1 tváyal vayám2 . .. sasahyama3 prtanyeltáJ:¡,4 "with thoul may we2 prevail over3 the assailants4 "; 8, 40, 7; 9, 61, 29; 6, 45, 18; ef. also 4, 24, 5; 6,50,8. •

There seems to be no reason not to compare the well-known instances: Fr. soyez armés! instead of armez-vous; Germ. bist du noch nicht weg; Eng. begonel ). Greek examples of this idiolll"are not lacking: 0128 p,uw6p,eve, f{Jeévuf: f¡Aé, &écp#oeuf:; Soph. El. 674 ot 'ych 'íáAUW', OAWAU 'íñfl' ev f¡p,éeq.2). .

Referenee could also be made to modern Persian: in this language the present unreal eondition is not distinguished from the past, the imperfect 01' preterit being the verb form of both clauses. The classical language usuallY had the conditional. However, a past unreal eondition is some­times expressed by a pluperfect.

The ancient German construction is remarkably similar to the Avestan. In Gothic, the 'unreal condition' (irreale Bedingung) is expressed with jabai followed by the pret. opt.; the apodosis, which as a rule is likewise 'unreal' -that is to say: the contents of which are viewed by the speaker as contrary to faet-, has the same verb formo This eonstruction is used irrespeetive of time: in aceordance with the nature of the opto no distinc­tion is made between processes in the past and those in the present: John 8, 42 (el ó#eof: na'íi¡e vp,wv ~v, f¡ywciJ.'íe Clv ep,é}J'abai gup atta izwar wesi, friodedeip pau mik; Matth. 11,23 ( ... el ev J:. eyévov'ío af flvváp,e~f: a[ yBV6p,eva~ 8v O'ot, lfp,ewuv Clv p,éxeL 'ífíf: O'1}p,eeov) jabai in S. waurpeina mahteis pos waurpanons in izwis, aippau eis weseina und hina dag; cf. also John 15,19. Like the aboye Avestan example these passages may be considered to have originated in the widespread preference for the 'fait accompli'. If in Old Norse the 'condition' expressed in the protasis is presented as a mere thought, supposition 01' assumption ofthe speaker, the preterit (imperf.) subj. is used in both clauses 3): H. H. 2, 33 pá vmri Pér hejnt Helga dauaa, ej jú vmrir vargr á viaum úti; Hom. 63, 11 ek mmtta, ej ek vilda, mea einu orai senda yar til helvítis. The same rule holds good if one wishes to formulate a condition which is at complete variance with the real facts ("med et bestemt faktisk forhold for 0ie vil have betingelsen udsagt som antaget mod virkeligheden"); however, the impf. subj. is used if the utteranee has reference to the present, the plqpf. (01' often also the impf.), ifit has a bearing 011. the past: Hom. 104, 27 eigi vmri eilíjar kvalar, ef endir yrai á; Nj. 112, 11 karlmannligt verk vmri petta, ej einn hejai at verit; Sig.3, 3 hann um mtti, ej hann eiga

1) See Havers, Handbuch, p.42. 2) See Kühner-Gerth, O. C., 1, p. 150; Schwyzer-Debrillmer, O. c., II, p. 287. 3) Cf. M. Nygaard, Norr0n Syntax, Oslo 1905, p. 213f. (§ 193f.): "Naar i det

betingede udsagn betingelsen udsiges som noget, den talene alene trenker sig eller antager, srettes imperf. konj. saavel i betingelsessretningen som i hovedsretningen."

Page 98: Gonda.1956.The Character of the Indo-European Moods.pdf

The Indo-European moods 192

. first that here again imaginative thought knretti. What is relevant /~ ret~rit be it a pret. opt.; in the second is expressed by means o t Pt' .' used if imagination is contrary to

1 th t th same cons ruc Ion lS d·.rr pace, a eh" the latter case a tendency to lUeren-f: ctl); and finally tha t t ere lS m ., a , t and past 'unreahty .

tiate betwee~ presenld

lrish unreal conditional sentence consisted ofthe As to Celtlc, the O . t" t've 2» in the protasis and the same

pret. subj. (which is an ~:~~e~)l:;u:plf. in the apodosis, irrespecti,:"e of form or the pret. fut. (d that the Celtic subj. denotes, mter t · (It may be recalle to memory h th e lme. . 'bl' several cases other t an e on alía, the unreal and th~ lmposs~ le;b 11 a 22 docoith dígal for1'u matis under discussion.) For mstance. r. .

tuicsi ni rigad. . d t he e that the Sanskrit perfect is sometimes It may als~ "?e pomte ou 3 ~ Ram 5 67, 38 yat1 te2 pr§fha'rfo3 si§evé

used in condIt1Onal sentences): k '. '- '10 "'f1 J5 of my own accord6 'ham5 svavasá6 ••• tatrá7ha'rfo8 k~'rfo9 an§yam~ 1 t ll7-10"

• 2 b k 3 it will be no use a a . would climb4 on your ac ... ,. . o t the so-called

The Old-Slavonic representativde of th~1'preytevr;:b4)~bearS a similar .. 1 t ret part an an aUXI lar condlt1Ona -an ac . p.. t- OChSlav. John n, 21

. h 'th reference to the pas e.g. 'v

force, eIt e~ ';,1 , I , 'o A 6 ov) aste bi s'íde bylU, ne bi bratu mOJ~ (el r¡~ (Me ovX aY ane{}ayeY o a e cp ~ ¡t t L k 17 6 (el l!xe'te

, M th 26 24-- or to the presen - u e, , umrulU5); at., 't b doubted whether Brugmann s nta'tlY ... , sABye'te aY .. ')'hl m(.a

y l~eady in OChSlav.) der Zustand opinion 6) that "damals sc ont l.t~· a Iwas used in interrogations with

. ht " that this cons ruc Ion 1 h erreIC war. th ht "schlechthin nur als so c en, kuto ("who") in order to express a ou~." -e . Luke 9,46 (OlaAo-als der Sphare der Vorst~;lung ,a~ge)horlgku~to J'i'cXu Vl2stej'í bi bylu-was

, " "y eZr¡ ¡te¿<,OJY av'tOJY ... lt y¿a¡to~ ... , 'to 'tt~ a . to regard this use as the resu correctly formulated. There lS no rea~on Passing reference may also be of a comparatively late devel.op~~n :-pf is in both clauses, also usual made to the fact that in Bulgarl~n ~.~~n '7).' cf for instance the Ruma­in past as well as present unrea con 1 10ns , .

nian usage. d t' d that in Ancient lndian the so-Other authors have aIrea y no ICe TI quently and necessarily, can

called conditional, though by no means re

1 A H ler Altislandisches Elementarbuch, Heidelberg 1921, p. 134f.

1) See a so . eus ,

(§ 419). 354ff 2) Thurneysen , o.C., I~, p. urnai of the Oriental Institute, Baroda, I (1952), 8) I refer to N. Sen, m Jo substitute of the future.

p. 305 who considers the perf~ct to be I~ 136f' A VaiUant, Manuel du vieux 4) For the forms see Vondrak,;.~, k.' p. Altb~lg. ·Gramm., Heidelberg 1919,

slave, l, Paris 1948, p.241f.; . es len,

p. 213f. 'L\2 II 3 871ff 6) See Brugmann, Grundrl , , ,p. . 6) Brugmann, o.C., p. 873f. . 142 7) See G. Weigand, Bulgarische Grammatlk, p. .

The conditional sentences 193

function instead of an optative in sentences expressing 'unreality'l). The conditional ha~ indeed in a modest way played a part in Sanskrit comparable to that of the other augmented forms in Greek. As this verbal category, which occurs as a rule in both protasis and apodosis, signifies what would happen or have happened, if something else should have occurred, which really has not taken place 2), it serves as an ex­pression for the past as well as the present 'irrealis': Kal. Kum. 6, 68 gám1 adhásyat2 katha'rfo3 nágal.é ... , tvam5 aválambi§yathá6 na7 cet8 "how3

would the serpent4 bear2 the earth!, ... if8 you 5 had not7 lifted (it) Up6". The protasis however almost always refers to a process in the pasto The origin of this idiom may-given a context implying an 'unreal condition' -be sought, on the one hand, in the abovementionedmodestor 'distanc­ing' imaginative functioll of past verb forms 3), on the other hand in the well-known tendency of Sanskrit to use the future4) to express possibility, doubt, wish, intention, probability, and so on 5): Mbh. 1, 76, 32 vyakta'rfol hato2 mrto3 vá4pi5 kacaé táta7 bhavi§yati8 "1 am sure!, father7,

Kaca6 will haves been injured2 or4,5 has died3 "; 3,55,4 yadi1 ••• mál?~2 pmtyákhyásyasi3 ••• , vi§am4 ásthásye5 "ifI you shall reject3 me2, 1 will resort5 to poison4". A construction such as Gop. Br. 1, 3, 14 eva'rfol cen2 ná3vak§yo4 mürdhá5 te6 vyapati§yat7 "jf2 you should not3 speak4 thus!,

1) See Speyer, Sanskrit Syntax, p. 269f.; Ved. S. Synt., p. 60; 89; Delbrück, Altind. Synt., p. 365f.; Renou, Gr. sanscr., p. 462f.; Whitney, Sanscrit Grammar, p.339 (§ 950); A. Holtzmann, Gramm. aus dem Mbh., Leipzig 1884, p. 33; 35ff. See also Pa:r:lini 3, 3, 139ff. and the notes by Renou, La Grammaire de P., Paris 1948, p. 160f.

2) Or, to mention the lndian term, in case of kriyatipatti "non-realization of a process (as a result of impediments etc.)".

3) Delbrück is no doubt right in describing the 'Bedingungsperioden' in which most cases of the conditional occur as being expressive of 'reflexion' ("in welchen über etwas reflektiert wird, was in der Vergangenheit nicht eingetreten ist, aber hatte eintreten konnen"). This 'reflective' force is however not limited to the pasto

4) Some objections may be raised to the derivation of the Sanscrit conditio­nal proposed by Hirt, Ind. Gramm., VI, p. 286: the explication of the wrongly so-called Aeolian optative defended by Wackernagel, Vermischte Beitrage zur griech. Sprachkunde, Basel 1897, p. 42ff. is in spite of its ingenuity not entirely beyond doubt, the hypothesis of an analogous influence of the subj. of the type TetasTS (TstrJ1)TS) being in itself sufficient to account for the existence of an opto stem ulasl- (see also Meillet-Vendryes, Traité de gramm. comp.2, p.260); for Hirt's hypothesis Lat. amarem < *amasejem (Indog. Forsch., 35, p. 140) see Leu­maml, Lat. Gramm., p.326, sub 4; Meillet-Vendryes, o.c., p. 297; whatever the origin and primary function of the augment may have been, it cannot be said to be a mattel' of indifference: although sporadic instances of the augment joining an opto can be found in later texts (Renou, Gramm. sanscr., p. 403: apibeyamRam.), there is no reason to suppose this combination to be ancient 01' comparatively frequent. From the syntactic point of view some connections between the fut. and cond. cannot be disregarded.

6) Speyer, S. S., p.266ff.; V.S.S., p.59f.; Renou, Gr. S., p.46lf. In later Sanskrit even in connection with the pasto 13 Gonda, Indo-European mOQ(Ia

Page 99: Gonda.1956.The Character of the Indo-European Moods.pdf

194 The Indo-European moods

your6 head 5 would fly off7 ", situating both processes in the realm oí imagínation-the text continues: hantatu te tad vak~yami yatha te na vipati~yatiti-, does no longer express a real hypothesis but an unreal condition 1).

The case represented by ChUpo 6, 1,7 yad1 dhy2 etad3 avedi~yan4, katha'f(b5 me6 na7vak~yan8 "for2 if1 they had known4 it3, why5 would they not7 have told8 it to me6 ?"; Asv. Bc. 8, 41 etc. seems to show that this construction, like that with the optative, could be used irrespective of phenomenal time: the augmented form may be taken to convey 'imag­ination' projected into the past2), and in combination with the con­junction and the context, unreality, the future stem to emphasize one of the aboye modifications (possibility). In Pañc. Br. 17, 1, 1 SOl 'man­yata2 kva3 ho~yami4ti5, sa6 tad7 eva8 na9vindad10 yatrallho~yat12 "he1 re­flected2: where3 shall 1 sacrifice4 ? ; he6 did not9 find10 that7 wherell he could have sacrificed12 ", the augmented form, though referring to a real past, is always expressive of mentalreflection on the part of the author, the conditional being, in a narrative passage, a substitute for the 'dubitative future' in direct speech. Similarly, SatBr. 14,4,2, 3 kasmad1 dhy2 abhe­fJyat3 "denn2 woher1 hii.tte el' sich íürchten sollen3 1" (principal sentence). The only case in the Vedic mantras, ~V. 2, 30, 2, does not seem to be essentially different: y61 vrtráya2 sínam3 átra4bharifJyat5 prá 6 tá'f(b 7 •••

uvaca8 "him7, who1 here4 could5 pay5 Yrtra2 his deserts3, she proclaim­ed6, 8 ... " 3).

In hypothetical sentences we find this form in one oí the clauses, ~ostly in the apodosis: Mbh. 8, 70, 27 bhratara'f(b1 ... yadi2 hanyalJ3 I hm4 uttaram5 akarifJyas6 tvam7 "gesetzt2 du würdest3 ••• deinen Bruder1

toten3 (opt.), was würdest6 du7 denn weiter5 tun6 ?" (Holtzmann): here the qualification 'imaginative 01' modest past of a dubitative future' does not seem to be incorrecto Hence, 1 suppose, such passages as Mbh. 3, 268, 19 where the commentator Nilakal,ltha explains the cond. by a fut., to which the form, indeed, comes very near: yada1 saran2 arpayita3 ... , tada4 manas5 te6 kim7 iva8bhavifJyat9 "when1 he will cast3 his arrows2

(into) ... , what7 then4 will you6 think5,9; ef. also 5,48,55 and other instances. Sporadically, the same funetion is found in a principal sentence:

1) In Kal. MiHav. 2, 10 + anyathá} katharp,2 tvám3 ••• ná4rcayif1yáma~t5 "otherwise1

how2 could we have omitted4 to honour5 you3 ••• ? ("comment ne t'aurions-nous pas rendu hommage?") and similar instances the future itself can function in the same way.

2) Compare Renou's observation (o. c., p. 463): "ehez les grammairiens ... le conditionnel transporte au passé certains emplois semimodaux du futur, a savoir apres particules interrogatives impliquant un étonnement." In the instance quoted the element of refiection is, again, far from absent.

3) According to Whitney the condition is here "most originally and properly used to signify that something was going to be done". Delbrück dismissed the passage as "nicht ganz deutlich".

The conditional sentences 195

Mbh. 12, 167, 34 nasti násin nábhavifJyat "gibt es nicht, gab es nicht und wird es wohl auch nicht geben". See also Bhag. Pur. 10,47,21. A cu'rious parallel of the Gr. eUJe 01' el yál2 preceding OXpeAOY to express a wish is in a way, Bhag. PUl'. 10, 28, 11 api1 nalJ2 svagati'f(b3 ... upádhCisyad~ isvaralJ& "may1 the Lord5 teach4 us2 his own state3 " 1). ,

A brief reference to the 'conditional' in another language may suffice to show that the aboye view of this category in Sanskrit, if it be in the main correct, is not entirely without parallels: the well-known French turn s'il venait, je partirais originates in such Latín constructions as e.g. Ps. Aug. Serm. 253, 4 sana1'e te habebat deus ... ,si tatereris 2). This originally popular idiom no doubt involved an element of emotive anticipation.

Returning now to the idioms connected with the Greek el we must for a moment, dwell on its well-known occurrence after verbs denotin~ wonder, delight, indignation, disappointment, contentment, and similar emotions. One example may suffice: Dem. 4, 43 #avp,áCw el p,r¡oeir; VP,WY p,frr:' §Y#Vp,eL-ía¿ p,rJ7:e ... "1 wonder that no one of you is either coneerned 01' ••• ". There are parallels in other languages: in Latin, miror, mirum est, gaudeo, and indignar are followed by si: Ter. And. 755 mirum si tacit3). In Sanskrit yadi is possible with phrases expressing wonder and unbelief, doubt etc. and also in áscarya'f(b yadi (cf. Pal,lini 3, 3, 151) whieh, beside ascarya'f(b yat expresses astonishment (Fr. "il est étonnant que"). Fr. si instead of que after verbs of emotion (esp. wonder 01'

admiration): tu avais voulu le tuer, et c'était miracle s'il n'était pas mart sur le coup. The construetion is, beside that with que, already found in Ane. French4). Similarly, in Albanese 5), and other languages.

In elucidation of this idiom attention may be drawn to the widespread tendeney to eonceal such emotions as wonder, disappointment, and so on, 01' at least, to express them with reserve. The phrase {}avp,áCw el, which was especially favourite in 'urban' and polite Attie 6), eautiously withholds complete astonishment, originally stating only that "if sueh­and sueh be the case, 1 certainly experience wonder". The eorreet translation of Skt. adbhutam1 andho2 náma3 girim4 árohet5 is "it would be marvellous1 if a blind man2 would elimb5 a mountain4 " (Fr. "ilserait surprenant qu'un aveugle grav~t . .. "); the same idea may be expressed

1) For interesting blends of future and cond. see Edgerton, Buddh. Hybr. Skt. Gr. and Dict., I, p. 152.

2) It may be recalled to memory that alsó in Latin so-called mixed conditionals (constructions with different forms in either clause) were far from rare, with the ::esult that the impf. and the pluperfect indicative tended to supplant the subj. m unreal apodoses.

3) Cf. Hofmann, Lat. Gramm.5, p. 776f. 4) Cf. Lerch, Hist. franz. Syntax, Leipzig 1925, I, p. 288. 5) Lambertz, Indog. Forsch., 34, p.57. 6) I refer to Kühner-Gerth, o.c., n, p. 369.

13·

Page 100: Gonda.1956.The Character of the Indo-European Moods.pdf

196 The Indo-European moods / The conditional sentences

otherwise, but in case yadi is used the opto is, according to Pal).ini 3, 3, 151, obligatory. In Latin the contents of the subordinated clause are not considered to be actual facts. In Sanskrit the construction is also found after verbs of expecting, after phrases expressing hope 01' unbelief 01' meaning "1 cannot bear" etc. 1), which from the psychological point of view, also belong to the same category: Oomm. on pal).. 3, 3, 147 yadi1 bhavadvidhaJ;,2 k~atriyar¡¡,3 yájayen4 nas mar~ayámi6 "1 cannotS endure6

that1 any one like you2 sacrifices4 for a k~atriya3". In Ram. 2, 51,14 ná1sar¡¡,se2 yadi3 jivanti4 "1 d0 2 notl think2, they are alive4 " the speaker literally "does not make a statement with regard to the case: they are alive"; Ram. 2, 59, 3 áiayá1 yadi2 már¡¡,3 ... sabdápayed4 itiS may origin­ally have meant: "hoping1:5 with regard to the case 2 (as to > perhaps, notice the opt.) he will address4 me3". In the sense of "if perhaps"; "ob nicht vielleicht" the word is rather frequent. After "to know" yadi expresses a similar nuance, if the content of the sentence involves uncertainty 2): Maitr. S. 1, 4, 11 na1 vai2 tad3 vidma4 yadiS bráhma1JáJ;,6 "we4 do notl know4 [that3] whethers we are brahmans6 "; Kathas. 24, 51 jánihi1 yadi 2 ••• "do you know1 whether2 ••• 1"; under similar circum­stances also after verbs of saying" etc.: Ka!. Kum. 5, 44 vada1 ... vibhavari2 yady3 aru1Jáya4 kalpate5 "say1 if3 (here Dutch has: zeg 01) the night2 goes with5 the sun4"; Mbh. 1, 154, 4 yadi1 ••• tvar¡¡,2 devatá3 ... ácalc~va4 mamas "as to the case1 you2 (are) a divinity3 ... te1l4 me5

> perhaps1 ... > whether1 ... ". Thus du~karar¡¡, yadi means "hardly, scarcely" < "difficuIt with regard to, difficuIt if".

The Greek construction of si after verbs ofknowing, saying, considering, attempting and other words implying uncertainty needs no comment3): A 371 sine . .. , sr "tt'Jia¡; ... i'bs¡; "say it in case you saw" > " ... whether"; A 83 cpeácrat, si ... "consider if"; A 420 slftt ... ar us nHJrrr:at "1 will go myself ... in case (> in hope that) he may be won over"4).

1) And, of course after verbs of cursing or swearing: Speyer, S. S., p. 371. 2) This character of yadi is also apparent from va yadi va and similar phrases:

Speyer, Sanskrit Synt., p. 334; 340. 3) Schwyzer-Debrunner, o.c., II, p.687f. 4) For German ob (Dutch 01; Engl. if) see Behaghel, o.c., III, p. 234ff.; H. Paul,

Deutsches Warterbuch, S. V.

X. The particle *me

How are we to account for the alleged fact that the negative particle accompanying the optative originally was not *me, but *ne 1 Miss Hahn 1) is of course right in observing that in this other way of expressing prohibition the <modal force' -if we understand by this term the attitude of mind of the speaker-is in thé form of the verb, not in the particle. In explanation of the curious difference it might perhaps be noticed that the opto with na, though appearing in the Vedic mantras, is increasingly common in later texts. Being the counterpart of the so-called prescriptive opto which essentially presents a request, injunction, exhortation, 01' prescription in the form of a <possibility' 01', rather, contingency, it was, at least in the beginning, no doubt a polite 01' deferential construction. It merely denied the (subjective) statement with regard to the contingen­cy of a process; it dissuaded 01' advised against performing specified actions: Manu 2, 144 na1 druhyet2 kadácana3 "he2 must2 never1,3 offend2 (him)", a type of prohibition which at first may have meant something like: "1 (the person speaking) deny that he (shall) possibly offend(s) him", 01' "1 exclude the possibility ofhis offending", and similarly ~V. 1, 91, 8 ná1 ri~yet2 tvávataJ;,3 sákha4 ("1 don't [supposeJl it will be possible2 that a friend4 of one who is as great as thou art3 will be harmed2 ") "del' Freund von einem, wie du bist, solIte (Eng. may, might) nicht zu Schaden kommen" (Geldner)2) ; ~V. 1,41, 9 ná1 duruktáya2 sprhayet3 "one should3 notl be eager3 to bad words2 "3). One might remember the na phrases with arhati: Mbh. 3, 55, 7 mar¡¡,1 na2 pre~ayitum3 arhatha4 "please do not send me" < "you are not the sort of people who ... , you are not2

qualified4, required4 to send3 me1". Hence also the general character of the prohibitions expressed by the

opto with na in Vedic prose, where it already is a familiar construction, and in later texts: Taitt. SaIJ1h. 2, 5, 5, 6 tasyai1tad2 vratar¡¡,3 nii4nrtar¡¡,5 vadet6 "his1 (i.e. of any person concerned 01' falling under this head) vow3

is as follows2: he (one) sha1l6 not4 speak6 untruth5 ". This use, too, may, it would appear to me, be explained from the hypothesis that na + opto fundamentally served to deÍlya contingency, not to ward off a particular

1) Hahn, o.c., p.41. , 2) "na + Opto im Sinne des griech. Opto mit OV7<: (J.y", GraJ3mann, Warterbuch

zuro Rig-Veda, 700.-0therwise: Delbrnck, Altind. Synta:x:, p. 337. 3) Cf. also J;tV. 8, 1,5; 23, 15; 96,2; 10,10,12; Maitr. Sa:rp.h. 2, 1, 3; Sato Br. 3,

8,3,11; 14,6,5,1.

Page 101: Gonda.1956.The Character of the Indo-European Moods.pdf

198 The Indo-European moods

evento The latter task was allotted to *me + inj. The particle *me keeps off or prohibits, or rather it rejects the realization of a process, and the injunctive refers to a process l) without being explicit in point of mood and tense: so mií + inj. wards off the realization of a particular process 2). When this construction is chosen the subject, object, or attending circumstances are as a rule known, definite, or specified, a fact not duly emphasized by my predecessors: cf. e.g. J;tV. 1, ll, 2; 23,9; 38,5; 6; 105,3; 162, ll; 15; 20; 2, ll, 21; 28,5; 32,2; 42,2; 3,15,6; 53,20; 4,18,1; AV. 1, 19, 1 etc. etc., and to quote some instances in full, AV. ll, 2,29 má1 no2 hi~siJ.¡,3 pitára~4 matára~5 ca6 "do notl harm3 our2

father4 and6 mother5"; 5, 7, 1 mál no2 rak~ir3 dák~i1Ja~4 niyárniinam5

"dol notl hurt3 our2 sacrificial gift4 as it is led away5" 12, 1, 31 má1 ní2

papta~a bhúvane4 sisriya1JáJ.¡,5 "letl mea notl falla down2 when supported5

on the earth4". However, na + inj. occurs also. "Wenn der Injunctiv andre als prohibitive Bedeutung hat, findet sich na", says Speyer3), Wie J;tV. 7,20, 6 núl cit2 sáa bhre~ate4 jáno5 ná6 re~at7 "nicht1 wankt4

das3 Volk5, noch6 wird es zu Schaden kommen7 ". 1 am under the im­pression that this construction simply negates the process expressed by the verb without preventing speaker and hearer from implying modal or temporal distinctions or giving utterance to them by other means.

There is, as far as 1 am aware, no reason why ma + opto should not have occurred also at an early period 4). This construction, which accord­ing to Miss Hahn 5) must be the result of contamination of the two ~bove types of expression and which because of its rarity is generally consldered an anomaly6), may however be explained by the need of an expedient for rejecting or keeping off, in emotion, a contingency: 4, 3,13 má1 sálchyur2

dák~a~3 rip6r4 bhujema5 "let it not be that we experience ... ", "wir mochten5 nichtl den Anschlaga eines schelmischen4 Freundes2 aus­kosten5 1" (Geldner)7); Hariv. 7944 pasyeyarp,l caiva2 ma3 yamam4 "and2

mayl 11 not3 seel the god of death4 1"S) This use is at the root of ma +

1) Process = any state, event, occurrence, act, action etc. expressed by a verbo

2) For the strong force of má alone see e.g.Speyer, Sanskrit Synta:x:, p. 318. 3) Speyer, Ved. U. S. Syntax, § 237. Of. also Delbrück, Altind. Synta:x:, p. 358.

The opinion that "*mé + 'Primitiv' (Injunctiv) zum Ausdruck des Verbots" "als indogerm. Geltungsbereich von wl¡ betrachtet werden darf" is e. g. also expressed by Schwyzer-Debrunner, O.C., II, p. 591.

4) Of. also Hahn, O. c., § 143. 5) Hahn, o.c., p.41, n.65. 6) See e.g. F. Edgerton, Buddhist hybrid Sanskrit, New Haven 1953,1, p. 200. 7) See GraJ3mann, O. c., 1027, 3. 8) Other instances: Petrograd Dict. V, 680,3 and Kulkarni, Annals Bhandarkar

Oriental Research Institute 24, o.C., p. 90f. Whitney's qualification (o. c., § 579b) "very rare" seeUlS to be somewhat e:x:aggerated.

The particle *mé 199

opto in so-called negative purpose clauses, where the particle has assumed the sense of "in order that not, lest": Mbh. 1,30, 15 miil sahasa~2 karsih3 ... mii4 tva~5 daheyu7J6 "do notl commit3 a rash act2 lest4 they might6 burn6 you5 to death" ("let the possibility of burning be preclu­ded"); cí. also 1, 153, 52 bhimol baladeva~2 vicakar~a3 ... 'mEi4 sabdaJ.¡,5 ... bhaved6> iti7 "Bh1mal dragged3 Baladeva2 away3 S04 that4 (his) yell5

might6 not4 ... ", lit. " ... away, 'thinking" "let there not be his yell". It may be doubted whether this use is a piece of evidence of a "Process der Zerrüttung"l). In other constructions mii can bear the same value ("lest"): 3, 269, 22 etanil vartmany2 anuyata3 sighra~4 ma5 va7J6 kalaJ.¡,7 k~iprams ihCi9tyagad10 vai "g03 speedily4 by3 thisl way2 lest 5 time7 should1o

ra pidlyS golO beforelO you 6" ("let loss of time be precluded") 2). In contradistinction to what seems to be the common opinion 3) it

would appear to me that the starting-point for the idiom ma + opto must date back to the prehistoric period: A vestan instances are numer­oUs 4 ), although they only belong to the so-called younger, but certainly not in all respects more recent, remainders of the ancient Iranian litera­ture; sometimes this phrase occurs after a parallel affirmative sentence in the opt.: e.g. Y. 9,21 paUrval tayüm2 ... büioyoima ioe3 ma4 éis5 paUrv66

büi oyaeta7 nos "may we become aware3 of the thief2 firstl; let4 (it) not4

(be that) anybody5 be aware7 of uSS first6 ". Old-Persian 5) and Homeric parallels can likewise be adduced: a 386 ¡,ti¡ aé y' Bv ..• l{}áurJ (JuatAf¡a Keov/wy I :reOtríaBtBV "m ay the son of Cronos never make you king in Ithaca"; O 176; cf. :re 372 ¡,tr¡c)' i]¡,tw; V:reBUrpVyOt. In those passages which have come to my knowledge the possibility 01' contingency of a process is, generally speaking 6) (emotionally) rejected or warded off: cf. also y 229; O 476. In Homer, ¡,trí alone can also introduce a dependent clause: with subj. A 522 dn6auXB ¡,ti¡ .¡;'t voi¡a'{} "Her¡ "go away, lest Hera becomes aware of it"; with opto E 298. In Greek and Avestan the optative also combines with the other negative particle: E 303 o OV ovo y' (ivoeB cpéeOtBY

1) Speyer, Ved. U. S. Synta:x:, § 237. . 2) The augmented form, which is also found elsewhere, may be ascnbed to

"Zerrüttung" if it is desired to use this term. Some instances of má + augmented aorist were collected by O. Bohtlingk, in the Zs. d. Deutschen Morgenl. Ges., 43 (1889), p. 56f.: Ram. 1, 27, 3; 74, 25 etc.; miZ + preso in the sense oí "lest": 1,65, 21; ma + fut. in the same sense: 2,65,39; pura • •. má + preso "sonst, andern­falls" .

3) See Brugmann GrundriJ32, II, 3, p. 807; Delbrück, Vergl. Synt., II, p. 355f.; Schwyzer-Debrunne~, o.c., II, p. 591. Of. also Meillet, Introduction a l'étude comparative des langues indo-européennes8, Paris 1937, p. 371.

4) 1 refer to Reichelt, O. c., § 655. . 5) 1 refer to A. Meillet-E. Benveniste, Grammaire du Vieu:x:-Perse2, Parls 1931,

p.244. 6) In sorne of the relevant A vestan passages, especially in the late Vendidad,

one might on the strength of the aboye ludian e:x:amples, e:x:pect nóil (,...., AncInd. na, Gr.ov) instead of '1lw" the context being prescriptive.

Page 102: Gonda.1956.The Character of the Indo-European Moods.pdf

200 The Indo-European moods

"whích two men could not bear" ; Yt. 11,5 ni5i11 dim2. •.. drvd 3 ava. spasti­cina4 aoi5 ava.spasni5i16 "(then) not (never)! a wicked3 one can6 see5

,6

him 2. by any spying4". The function of this idiom has already been com­mented upon: it negates a contingency.

The subjunctive, finally, could likewise be accompanied by both partícles: cf. ~V. 10, 34, 5 yád1 ádtdhye2. ná3 davi~á1Ji4 ebhily,5 ... "when1

P think2. to2. myself2 : 14 will4 not3 go4 with them5 ••• "; 8,81,4 ná1

rádhasá2 ma1'dhi~an3 naly,4 "he3 will3 (we expect3) notl neglect3 us4 in (granting his) bounty2"; Av. Y. 12, 3 nOip ahmá12 á3 ryánim4 ni5i15

vivéip'Jm6 xstéi7 "henceforth 2,3 P shal17 commit7 neither1 damage4 nor5

destruction6 "; H 197 ov yáe -dr; pe pín ... &é)-cov'ia Oír¡'iW "by force no man shall drive me in fiight against my will" 1), where the process is simply negated-although very 'subjective' motives may induce this more 'objective' negation-, as against: SatBr. 11, 5, 1, 12) where a warning is expressed akámá'l'f/,l sma2 má3 ní4 padyéisai5 "against my will1 you5 must3,5 nots He5 down4 (with me)"; Av. 7, 38 méi1ca2 ... irisyá~3 "and2 notl ... shall (must)2,3 he hurt3"; Ll 37 pi¡ WV'iÓ ye ve¿)-Cor; oníaaw I aOl )-Cal epol psy' lieLapa y8Vr¡'iW "let not this quarrel be a grievous cause of strife"; a 334 pI¡ -rír; -rOL 'iáxa 'leov apeívwv liAAOr; avaa'i'ñ "beware, lest presently another better than Irus shall rise up against you"; M 216 pi¡ Yopev "let us not go": in these passages a process which is taken into account, which is impending or is anticipated is rejected or warded off. I cannot imagine why this combination should not occasionally have been resorted to in prehistoric times.

It has often, and correctly, been observed that in Greek the subj. (aor.) + pI¡ must have replaced the ancient inj. (aor.) + pI¡, traces of which have already been mentioned on p. 45. The most satisfactory idea we can form of this substitutional process seems to be founded on the assumption that a pre-existing *me + subj. ousted the other con­struction. I am under the impression that this well-known Greek phrase: pI¡ + subj. (usually of the aor.) 2nd or 3rd pers. originally was more significant and expressive than *me + inj. because of the special connotation of the mood: in Ll 37 f. (see above) for instance an impending (danger of a) controversy is warded off, not only the act of quarrelling in general or at least without regard to time: "let not this quarrel in time to come be to you and me a grievous cause of confiict between us" ; cf. also E 487f. May the combination imp. (inj.) + negated subj. be considered to refiect a case in which the 'stronger' subj. was especially favourite: 1]1 407 'brnovr; ... )-CLxúveT8, pr¡CJs Atnr¡af}ov "overtake the horses with speed, and be not outstripped of them « let the impending reali­zation of the idea of being outstripped be warded off)" ~ It is in this

1) See Delbrüok, Altind. Syntax, p. 315f.; Reiohelt, 0.0., § 649. 2) The only example quoted: see Whitney, 0.0., § 579d.

The partide *me 201

connection worth noticing that this construction served to express an individual, not a general 'prohibition'. In Homer it is used to express a warning, admonition, or threat (e.g. a 334; A 26; hence pI¡ in dependent "lest" phrases), or in statements of fear (e 415 pI¡ nwr; p' e)-C(Jaívov'ia páAn Aíf}a)-CL nO'il n8'ten )-Cvpa "haply were I to seek to land, a great wave may dash me against the jagged rock" (Murray) or in Dutch "als nu maar niet een grote golf mij ... zou slaan" 1); hence the use of pI¡ in dependent clauses expressing fear or apprehension 2)-in these cases the character of the subjunctive is manifest and the preference for the abOYe postulated construction understandable-; besides, it occurs in negative incitements or hortations: M 216 pi¡ Yopev Llavao¿aL paxr¡a6pevoL "let us not go forward to fight with the Danaans": here the corresponding Vedic construction is má + inj. (~V. 2, 28, 7 má .. . ganma; 10,57, 1 má prá gáma). As already intimated, this was, probably under the infiuence of the com­binations of *me and subj. and 'for reasons oí expressiveness' given up in Greek.

In Latin a further development led to the use of ne instead of the ancient *me: ne feceris: pi¡ n0t1]anr;3).

A word should be said here about a remarkable parallelism in Greek and Sanskrit with regard to the history of *me. "Viel haufiger als ov statt pI¡ steht hellenistisch pI¡ für OVo Das Vordringen von pI¡ gegenüber ov ... hat sich bis zum Ausgang des Altertums fortgesetzt, aber nicht etwa mit einem Siege von pf¡ geendet. Die Gründe waren verschieden: 1) die Analogie ahnlicher Fügungen und 2) die N eigung zu mechanischer Regelung; 3) Attraktion an prohibítivem Satzgehalt; 4) der affektische Gehalt von pI¡; 5) Vermeidung van Gleichklang und Hiat; 6) Abwechs­lungsbedürfnis; 7) spater auch mangelndes Gefühl für Feinheiten lite­rarischer Sprache ... und dadurch veranlaBter übertriebener Gebrauch des als fein geltenden pl¡"4). As to 1) and 3); in contradistinction to the ancient Vedic 5) usage má has in classical Sanskrit become the proper negation to be used with the imperative (the Vedic instances 6 ) quoted in the Petr. Lex. S. V. méi concern the 2nd p. plur. which originally was an ínj.: ~V. 4, 5, 2 má nindata "do not blame"; 8, 1, 1; also AitBr. 2, 16,4 má bibhita "do not fear") epie, narrative, popular, 'Buddhist' Sanskrit use méi with the imperative: Mbh. 1, 154,23 tvarasva ... má

1) See also Waokernagel's observations in oonneotion with K 65: P 71 (Vorles. Synt. Il, p. 275).

2) For partioulars see the grammars, and D. Tabaohovitz, in the Eranos, 49, p.93ft'. .

3) See espeoially W¡;okernagel, O. O., Il, p. 259. ') Sohwyzer-Debrunner, 0.0., Il, p. 594f. 5) Má + imper. also ooours in Y. Avestio. 6) For the Vedio (of. e.g. ];tV. 8, 103, 12) ooourrenoes sea also H.Oldanbarg,

Textkrit. und axegat. Noten, Il, Berlin 1912, p. 379; A. Fürst, Spraohgabr. del' alteren Upan., KZ.47, p. 21.

d

Page 103: Gonda.1956.The Character of the Indo-European Moods.pdf

202 The Indo-European moods

kl'irJa "be quick, do not play!"l) Ram. 2, 1I6, 5 it reads ma gantum al'hasi instead of na g. a. In Greek this construction is usual already in Homer: A 32 fhi¡ fh' e(2é-&lCe "don't rouse me to anger" 2).

In Greek and Sanskrit there are stray occurrences of *me + fut.-e. g. Arist. PI. 488; Mbh. 3,147,5 Sukth. 3)-, conditioned by the speaker's wish to ward off, or to warn against, an occurrence visualized as going to happen after the moment of speaking. In Buddhist Sansk;rit texts this construction is very common. There ma is also much used to express a strong negative wish or hope (often translatable by verbs of fearing), a strong determination, etc. In such sentences as Lalitavist. 186, 14f; ma + fut. does, ít is true, not express a prohibition proper or negative command, but certainly expresses the speaker's resistance to a future occurrence: (the king established guards around his son's dwelling) "thínking: the prince is not going to (1 am determined that he shall not) depart (from wordly life)": ma khalu (kh.: a strengthening particle) kumal'o ... abhini~kl'ami~yat'íti. This usage reminds us of the Greek ídiom: wl¡ in oaths and asseverations: K 330 '(cn:w vvv Zev~ uv'td~ . .. I fhi¡ fhBV 1'0 r~ 'lnnOlCJW avi¡(2 enox/we'tat aAAo~ I T(2ww'V "N ow be my witness Zeus himself ... that on those horses no other man of the Trojans shall mount" (cf. 041; T261; fhi¡ fhév: cf. AInd.ma sma). In Buddhist Skt. disquietude or fear of the future is also very often the dominant emotion: ma "1 am afraid that, 1 hope that not". The particle can also, in the sense of "lest, in order that . . . not" introduce a dependent clause (For Greek see above).-As to 4) cf. e.g. Mbh. 3,281,24 Su. mal vai2

dvitiyarp,3 ma4 trtiyarp,5 ca6 vañche7 "1 do notl desire7 a second one3, nor4, 6 a third 5" .-Although the special force of the particle usually remains perceptible, there are cases in which it is merely a substitute for na: in the Buddhist Dvavirpsatyav. it reads: mal kascidl'ujaya2 vapuJ:¿3 "his body3 is notl afflicted by any disease2" 4); in Pali: Vino 3, 23 masakkhimha "we could not".

Edgerton 5) is in my opinion right in saying that sorne má phrases were questions: "the concern, perturbation, or deprecation of a feared con­dition . . . is certainly at times cast in interrogative form". See e. g. Divyav. 9, 16f. mal si2 tr~ito3 ... "are2n'tl you2 thirsty~" Sometimes the particle (and a special intonation, of course) marked a question to which the speaker hopes for, or expects, a negative answer: Mülasarv. 2, 67, 21

1) For particulars see Speyer, Skt. Synt., p. 273f.; Ved. Skt. Synt., p. 58; 73; Renou, Gramm. sanscr., p. 413; Edgerton, Buddh. Hybr. Sanskr. Gramm., p. 200, see also my aboye remarks (p. 44ff.).

2) See also Schwyzer-Debrunner, o.c., II, p. 315; 343. a) 1 refer to Liddell and Scott, Gr.-Engl. Lex., II, 1124 A 3, and Kullrarni,

Ann. Bhand. Oro Res. Inst., 24, p.92f. 4) See R. L. Turner, in the .Tournal of the R. Asiatic Society 1913, p. 297. 6) Edgerton, o.c., p. 202f.

The particle *me 203

mal kirp,cit2 pal'ihiyate8 "1 hopel nolthing2 is lacking3 ". In Greek fh?} occurs in a similar way in direct questions mostly anticipating a negative answer: Aesch. Pers. 344 fhf¡ O'Ol oo~ovfhe'V 'tñr'Je AelcpfMj'Vul p,áxn; "does it esem to you that we were inferior in that battle ~ " 1).

1) Some words remain to be said on the much debated problem of the alter­native el flf¡ and el ov in Greek. In considering the paraIlel constructions used in the ancient l.-E. languages-to wit the Anc. Ind. yadi na, the Av. yezi nait, the Latin negative conditional clauses formed without ne-we seem to be driven to the conclusion that the usual Greek el flf¡ 01' at least its frequency, may be an innovation. Tabachovitz (Homerische el-Satze, Lund 1951, p. 22ff.), who recently devoted a chapter to a discussion of some authoritative opinions of the problem as to how far both Greek constructions are different in meaning is no doubt right in most of his critical remarks. Lange's interpretation '(Del' homer. Gebrauch der Partikel el, Abh. Sachs. Akad., 16, p. 460): el fI~ 6w(la rpS(lOl properly meant: "gesetzt das Zugestandnis: fern sei der gesetzte Fall, er brachte Geschenke" is indeed untenable, since the originally concessive character of the opto is, as we have seen, completely improbable. The explication proposed by Goodwin (Syntax of moods and tenses, p. 138) and recently upheld by Chantraine (o. c., rr, p. 333): when ov is found in a protasis, it is generally closely connected with a particular word, esp. the verb (cf. Plato, Apol. 25B eáv 7:8 ••• ov rpfíTe eáv 7:8 rpfíTe), though-in some cases-acceptable (cf. Humbert, Synt. Gr. 2, § 623) and elsewhere plausible at first sight, is-certainly in passages snch as r288; Y139 (Wackernagel, Vorl. über Syntax, II, p. 261; Tabachovitz, o.c., p.32)-open to controversy. The opinion maintained by Hale (G. Hale, The origin of subj. and opto conditions in Greek and Latin, Harv. Stud. 1901, p. 109ff.) and other scholars (cf. e.g. Monro, Gramm. Hom. dial., p. 288) that the choice of the negative particle mainly de­pended on the mood contained in the conditional clause forces these authors not onIy to adopt intricate analogical processes, but also to disregard those facts which are discussed in the first part of this chapter. The solution suggested by Tabachovitz (o.c., p.33): "vielleicht spielten dabei rhythmische und metrische Bedingungen die gri:i13te Rolle" cannot be regarded as satisfactory either: the difference in function and connotation between ov and flf¡ was, generally speaking, too great to permit a poet to interchange them for reasons of versification. Another explication of the same scholar is however worthy of serious consideration; the use of flf¡ must have originated in those sentences in which flf¡ is the natural negation.

It would appear to me that we are confronted with two problems: first, why has Greek so obviously preferred el flf¡, both particles being possible in clauses of this character?; second, what is the difference between el flf¡ and el 0V? N ow we may scarcely expect to find out which processes led to the enormous spread of the el flf¡ phrase which is already in Homer much more usual than its rival. It seems however warranted to suppose that some of the factors which determined the preference for flf¡ inlater Greek and for rnii in later Sanskrit have exerted in:fl.uence. If we are right in the opinion that flf¡ essentially served to reject the realization of a process, it involved an element of emotion and subjectiveness (cf. also Stahl, O.C., p. 233) which, being foreign to the more intellectual OV, has no doubt contrib­uted to its spread. By way of hypothesis an original meaning: "supposing, then, let the mentaIly existing process of his coming not be realized (-fling me down from a cliff)" may'be considered to have been at the root of the type of utterance found ~ 398 el os ¡,:e fl1] 8J..1}naw ..• ; "let him, in that case not give" or "let us, on the other hand, suppose, he willnot give" at the root of A 324 el 6s ¡,:e fI~ &ónaw. Cf. also 1515. Did H 98 el fI,f¡ nr; ••• elaw 'originally mean' something like: "(this

Page 104: Gonda.1956.The Character of the Indo-European Moods.pdf

204 The Indo-European moods / The particle *me

shall be a disgrace) if be rejected the fact of any man's going ... "? On the other hand, the construction oV + indo (pr., fut., aor., perf.-pr.) in an antecedent protasis simply negates a supposition. The objective character of this construction may involve greater force 01' emphasis (Humbert, o.c., § 625: "fait mena<;lant"?). It is worth noticing that the negation of the apodosis, if any, is also ov (Ll 55; Q 296; (J 274; cf. also 1435). The character of the utterance in its entirety has no doubt been an important factor also: contrast for instance r 288ff.; 55f.; O 162; Q 296 with J. 159; II 32; A 137; B 261. Parallel constructions in other languages may also help to the solution of the riddle: in Malay lamun "provided it be" is followed by lañan "fJ,~" : "1 offer you ... provided1 you4 don't2 slay 5 me3": lamun1 fañan2 aku3 eñkau4-bunuh5

• In Tongan 'oua is the negative particle used in prohibitions and after t~e ~o-c~lled prospective ke (Churchward, T. gr., p.58). In Hebrew 'al, in contradlstmctIOn to lo, expresses a subjective and conditioned negation.

Select Bibliography (including those titles of works which are not always cited in full)

A.R. Anderson:

J. Avery:

Repudiative questions in Greek dramas, and in Plautus and Terence, Trans. and Proc. Amer. Phil. Ass. 44. The unaugmented verb forms ofthe Rig- and Atharvavedas, Journal Am. Oro Soco II (1885).

J. Th. Beckmann: Das Gebet bei Homer, Thesis Würzburg 1932. O. Behaghel: Deutsche Syntax, Heidelberg 1924. Ch. E. Bennett: Syntax of early Latin, Boston 1910. E. Benveniste: Prétérit etoptatifenindo-européen,Bull. Soc.Ling.,47 (1951). IV!. Bloomfield and F. Edgerton: Vedic Variants, Philadelphia 1930. K. Brugmann: Griechische Grammatik, Stra13burg 1899. K. Brugmann: Grundri13 del' vergleichenden Grammatik der indoger-

manischen Sprachen, Stra13burg 1916. K. Brugmann: Kurze vergleichende Grammatik, StraJ3burg 1903. K. Brugmann-A. Thumb: Griechische Grammatik4, München 1913. C. D. Bucle A dictionary of selected synonyms in the principal l. E.

O.D. Buck: C.D. Buck: E. Cassirer:

P. Chantraine: B. Delbrück:

B. Delbrück:

B. Delbrück:

B. Delbrück:

J. D. Denniston: C. L. Durham:

H. Ebeling: F. Edgerton:

languages, Chicago 1949. A grammar of Oscan and Umbrian, Boston 1904. Comparative grammar of Greek and Latin, Chicago 1933. Philosophie der symbolischen Formen, 1, Die Sprache, Berlin 1923. Grammaire homérique, vol. Il, París 1953. Altindische Syntax (= Syntaktische Forschungen V), Halle 1888. Del' Gebrauch des Konjunktivs und Optativs im Sanskrit und Griechisch (= Syntaktische Forschungen I), Halle 1871. Die Grundlagen dergriechischen Syntax (= Syntaktische Forschungen IV), Halle 1879. Vergleichende Syntax (= GrundriJ3 dervergleichenden Gram­matik der indogermanischen Sprachen1, vol. IIl, IV und V). Greek Particles2, Oxford 1954. Subject substantive clauses in Plautus, Cornell Studies in Classical Philology 13. Lexicon homericum, Leipzig1885. Buddhist hybrid Sanskrit, Grammar and Dictionary, New Haven 1953.

A. Ernout-F. Thomas: Syntaxe latine2, Paris 1953. W. ViT. Goodwin: SyntaxofthemoodsandtensesoftheGreekverb,London1889. H. Gra13mann: vVorterbuch zumRig-Veda, Leipzig 1936. E. Adelaide Hahn: Subjunctive and Optative. Their origin as futures, New

York 1953. H. Hammerschmidt: TIber die Grundbedeutung von Konjunktiv und Optativ,

W. Havers: E. Hermann: .

H. Hirt: J. B. Hofmann: J. Humbert:

Erlangen 1892. Handbuch der erklarenden Syntax, Heidelberg 1931. Die Nebensatze in den griechischen Dialektinschriften, Leipzig-Berlin 1912. Indegermanische Grammatik, Heidelberg 1934. La,teinische Umgangssprache3, Heidelberg 1951. Syntaxe grecque, Paris 1954.

Page 105: Gonda.1956.The Character of the Indo-European Moods.pdf

206

O. J espersen:

Select Bibliography

A modern English Grammar on historie al principIes, London 1932.

O. Jespersen: The philosophy of grammar, London 1924. R. Kühner-B. Gerth: Ausführliche Grammatik del' griechischen Sprache, Han­

novel' 1898. E. D. Kulkarni: F. Kurschat: G. van del' Leeuw: G. van del' Leeuw: E. Lerch:

Annals of the Bhandarkar 01'. Res. Inst., Poona. Grammatik del' Litauischen Sprache, Halle 1876. L'homme primitif et la religion, Paris 1940. Religion in essence and manifestation, London 1938. Das Imperfektum als Ausdruck del' lebhaften V orstellung, Zs. f. rom. Philologie, 42 (1922).

M. Leumann-J. B. Hofmann: Lateinische Grammatik, München 1928. A. Meillet: Introduction a l'étude comparative des langues indo-euro­

péennes8, Paris 1937. A. Meillet-J. Vendryes: Traité de grammaire comparée des langues classiques,

Paris 1948. A. Minard: D. B. Monro: E. P. Morris: J. Murphy: A. Musio: C. Mutzbauer:

La subordination dans la prose védique, Paris 1936. A grammar of the Homeric dialect 2, Oxford 1891. The subjunctive in Plautus, Amer. Journ. ofPhil., 18 (1897). Primitive Man, Oxford 1927. Beitrage zUr griechischen Satzlehre, Zagreb 1927. Die Grundbedeutung des Konjunktivs und Optativs und ihre Entwicklung im Griechischen, Leipzig-Berlin 1908.

W. Neisser: Zeitschrift für Indologie und Iranistik, 5 (1927). H. Osthoff-K. Brugmann: Morphologische Untersuchungen, Leipzig 1878ff. R. van Pottelbergh: Over de geschiedenis en de betekenis van den Bl-zin in het

H. Reckendorf: H. Reichelt: L. Renou: L. Renou:

L. Renou:

Grieksch, Gent 1939. Die syntaktischen Verhaltnisse des Arabischen, Leiden 1898. Awestisches Elementarbuch, Heidelberg 1909. A propos du subjonctif védique, Bull. Soco Ling., 33 (1932). La décadence et la disparition du subjonctif (Monogr. Sanskr., I), Paris 1937. Les formes dites d'injonctif dans le ~gveda, in the Etrennes de linguistiques offertes a E. Benveniste, Paris 1928.

L. Renou: La grammaire de Pa:r;tini, Paris 1948-1954. L. Renou: Grammaire sanscrite, Paris 1930. L. Renou: Grammaire de la langue védique, Paris 1952. E. Schwyzer-A. Debrunner: Griechische Grammatik, München 1940. F. Slotty: Del' Gebrauch des Konjunktivs lmd Optativs in den griechi­

F. Sommer: J. S. Speyer: J. S. Speyer: J. M. Stahl:

T. G. H. Strehlow: W. Streitberg: D. Tabachovitz: W. Vondrák: J. Wackernagel: A. Walter:

W. D. Whitney:

schen Dialekten, G6ttingen 1915. Vergleichende Syntax del' SchuIsprachen, Leipzig 1921. Sanskrit Syntax, Leyden 1886. Vedische und Sanskrit Syntax, Stra13burg 1896. Kritisch-historische Syntax des griechischen Verbums del' klassischen Zeit, Heidelberg 1907. Aranda grammar, Oceania, 13 (1943). Gotisches Elementarbuch 6 (1920). Homerische el-Satze, Lund 1951. Vergleichende Slavische Grammatik, G6ttingen 1928. Vorlesungen über Syntax, Basel1926. Die Grundbedeutung des Konjunktivs im Griechischen, Heidelberg 1923. A Sanskrit Grammar 5, Leipzig 1924.

Index of subjects

(only leading references are given)

adverbal prepositions 43; 44 aorist 188 aorist, gnomic 181 apodosis 138f. apodosis with imperative 172f. apodosis with preso indico 175ff. apodosiswith subj. 177f. aspects 10; 187 aspects in Chipewyan 12 aspects in Tibetan 11 aspectual view of processes 71 attraction 72; 105ff. attraction, modal 105ff. augmented tenses 47 auxiliaries in MH Germ. 5 auxiliaries absolutely used 7 auxiliaries, modal 5; 6 auxiliaries, modal Greek 5; 6 auxiliaries, modal Latin 7 auxiliaries, modal Malay 6; 18 auxiliaries periphrastical 8 auxiliaries with inf. 7 auxiliaries with real mood 7 concessive clauses 168; 177 concessive opto 57 condition, expressed without conjunc-

tion 153ff.; 162f. conditional 192ff. conditional, Old-Slavonic 192 conditional clauses 153ff. conditional conjunctions, origin of, in

various languages 155ff. conditional constructions 96 conditionaI opto 58 conditional sentences IX conjunctions and subj. 90 consecutive subj. 83ff.; 9lf.; 104 construction of a question repeated 84n. contingency 51ff.; 70; 92; 106; 179; 197 contingent future, in Hindi 109 cupitive optative 49; 52f.; 55; 58 deliberative opto 55f. demonstrative pronouns 166 durative past tenses 181ff. echo-clauses 83f. emphasis 76; 78f.; 82; 148 emphatic future 94 endings, primary 47f. endings, secondary 47 exclamatory interrogation 56; 81; 83

exhortation 89 extended deliberative 82 final clauses 90f.; 126f. future 69; 7lf.; 74ff.; 98; 99f.; 117 future, dubitative 194 future, emphatic 177 f. future, gnomic 78n. future, in Chiricahua Apache 12 futurum exactum 94f. general optatives 52; 56f. Gesetz del' wachsenden Glieder 76 'hope' 01' 'expectation' 50; 51 hypothesis 179f. hypothetical sentences 194f. indicative 115f. injunctive IV; 115f. imaginary time 187 imagination 188 imaginative preterit 182ff. imperative 49; 53; 54n.; 95f.; 130 imperative use of the optative 49 imperfect, persuasive 185f. imperfect in conditional clauses 183 imperfect, in French 182f. imperfectum modestiae 185 indicative 3; 6; 47f.; 95 indicative of hist. tense with &v 181 injunctive 47; 131; 198 interchange between present and past

in Vedic 36; 39; 40 interchanges between the moods in

Vedic texts 128ff. interrogations with subj. 83 irrealis 179f. liñ 52 mantras 39 mental reality 72; 73f.; 80; 82f.; 92;

101 'modal auxiliaries', Indonesian 134 modal distinction in Hittite 132f. modal distinction in Modern Indo-

..Aryan 8 modal 01' temporal force 42; 43 modal particles VIII modal particles in Chinese 13 modal view of processes 71 modality 9n. modes, Aranda 10 modes, Chiricahua Apache 11; 12 mood 2; 3; 5; 6; 9n.

Page 106: Gonda.1956.The Character of the Indo-European Moods.pdf

208 The Indo -European moods

morphology 6 myths, language of 40 negation 132 negative particles 132 negative purpose clauses 199 objectivity 3 optative V; VII; 3; 6; 78n.; 106f. optative, adhortative 53f.; 55 optative, (cupitive) with av or ue 143f. optative, deliberative 144 optative in final clauses 92 optative in indirect interrogations 124 optative in interrogations 55 optative in Middle-Indian 8 optative in protasis and apodosis 178ff. optative in similes 63; 177 optative of imaginary process 59 optative of logical possibility 57f. optative of past processes 64ff. optative of probability 59f. optative of subjective supposition 181 optative of supposition 58f.; 60 optative of the 'gemilderte Behaup-

tung' 6lf. optative, paraphrased in Sangirese 6;

17 optative, perfect 190f. optative (potential) with av 142f. optative (potential) without av 143 optative prescriptive 197 optative with na 197f. optativus iterativus 62ff. optativus obliquus 123ff. optativus urbanitatis 61 parataxis expressing conditional re-

lation 16lf. particles 133f. . past unreal conditions 119 perfect 192 perfect optative 190f. periphrasis 6 perseveration 142 politeness 53f. possible future 74n. potential optative 49; 52; 55; 58; 60f. 'prepositions', adverbal 43f. preterit denoting unreality 183f. primitive mentality nI primitive mind 49f.; 179 prohibition 44; 197 prohibitive 44 pronouns, developing into conjunctions

161 protasis with optative 176 protasis with opto and apodosis with

an hist. tense 181 questions of doubt 82 realis 172 rejection of the realization of a process

198

relative clauses 140 relative pronoun 62f.; 96 repetition, emotional 76 repudiative question 84f. stipulative ut-clauses 105 subjunctive VI; VII; 3; 200 subjunctive, accentedin Vedic 100; 102 subjunctive, adhortative 79;" 89 subjunctive after imperative 87f.; 90;

95f. subjunctive after verbs expressing fear

10lf. subjunctive and future 86; 99f. subjunctive and future, alternating 75f. subjunctive, deliberative 68; 8lf. subjunctive endings 11 Off. subjunctive in the Avesta 11lf. subjunctive in Bantu 107 f. subjunctive in Georgian 109 subjunctive in Indonesian languages

108 subjunctive in continuations 85 subjunctive in final clauses 90ff. subjunctive in generalizations 97 subjunctive in indirect questions 99f. subjunctive in interrogations 83 subjunctive in similes 74 subjunctive in subordinate clauses

95ff. subjunctive, iterative 99 subjunctive of the past 72f. subjunctive, origin 115f. subjunctive, periphrastic in Medieval

Dutch 6 subjunctive, periphrastic in OHGerm. 7 subjunctive, prospective 68; 74f.; 91 subjunctive, suffix 110; 113ff. subjunctive, voluntative 68; 71; 89; 91 subjunctive with ue in condit. clauses

147f. subjunctive with pI¡ 200f. subjunctivus obliquus 100 suffixes, .Japanese 12 syntax 6 'temporal adverbs', Chinese 12 temporal relative with subj. 94f. tense In tense, in Aranda 10 tense, in Maya 11 timeless subjunctive 73 unreal condition 179; 188ff. unreality 119; 193 verb, accented in Ve die 162 verbs of obligation etc. 189 verbs of wishing 52 visualization 51; 53; 69ff.; 73; 81; 94;

106 vividness 117 f. will 48; 51; 69; 121 wish 48; 51; 121

Indo-European: *an 145ff. *e- / o- 165f. *io- 126' 128 *me X ' *me with future 207 *me with injunctive 198 *ne 197

Ancient lndian: atha 85 api 135 aham 78 iti 102f. U, usu 42 katham 55 katha 56; 83 kuvíd 100f. gha 168 ca 162f. ced 163 jatu 55; 135 na+ subj. 80 nu 42 nünam 85 ned 101 pura 64 ma 44ff. ma instead of na 20lf. ma with inj. 36 ma with opto 198ff. ma sma 202 ya- 96; 162f. yat, yad 94; 102; 163f. yatra 95 yadi 163f.; 195f. yat 93 sma 42; 64 ha 168 hanta 101; 102

Greek: aye 89 al 165 av 133; 135ff. av and fut. 147f. av in an apodosis 138f. av in el clauses with subj. 176 av in relative c1auses 140 av with indic. 137f. ~v with verbs of obligation etc. av or ue with subj. 139f. yá(! 167 ye 167f. Sdv 141 eyro 78; 82 6{}éAet'V 7 el IX

Select glossary 209

Select gIossary

137f.

e! clauses 120f.; 144; 165ff. e~ clauses expressing wonder etc. 195f. et clauses with subj. 151 el, cupitive 167 el, interjectional 150 el, original function of 152ff.; 166 el "whether" 151 f. el with fut. indo 174f. el yá(! 168 el pI¡ 188 el pI¡ and el oV alternating 203f. n. eha 82 e't-r:e 169 loor; av 141 {}avpá!;oo el 195 ¿va 92; 141 Eva clauses 126f. uat(!6r; 26 ue 135ff.; 188 ue repeated 142 pI¡ 132 pI¡ clauses 12lf.; 126 pI¡ in questions 202f. pI¡ instead of 201 pI¡ rejects realization lOlf. pI¡ with fut. indo 80 p1 w~th subj. 80; 102; 200f. ¡trI p,ev 202 8r; 96f.; 126 87:e 93f. ov 132 ov + future 80 ovu av + opto 139 orp(!a 93 orp(!a clauses 126f. ¡¡;(!{p 94 ¡feia 59 7:áXa 59f. rpé(!e 89 éJJ(!a 26 ror; clauses 126f. wrpeAov 185; 189

Other languages Lat. an 133; 145ff. Lat. habeo 8 Lat. si 164 Fr. pouvoir 5 Goth. an 133; 145ff. Goth. ei 165f. Goth. ip 102 Goth. Iveila 27 ORG. mal 27 MHGerm. so 164 Dutch kunnen 8 Eng. may 8 Mod. Germ. mogen 5; 7 Slav. da 164f.

Page 107: Gonda.1956.The Character of the Indo-European Moods.pdf

210 The Indo-European moods

Index of names

Bailey, H. W. 65 Bally, Ch. 149 Basset, A. 11 Behaghel, O. 76; 77; 176 Bennett, Ch. E. 103; 105 Bloomfield, M. 128 Brockelmann, C. 159 Brugmann, K. 5; 33; 47; 48; 68; 145;

153; 192 Buck, Ch. D. 51 Burgt, J.M.M. v. d. 14 Chantraine, P. 57; 58; 59; 122n.; 147;

175; 203n. Cobet, C. G. 147 Debrunner, A. 104; 112; 119; 147;

149; 184n.; 189 Delafosse, M. 10 Delbrück, B. 1; 33; 34; 38; 47; 48; 51;

69; 86; 87; 91; 94; .95; 100; 126; 133n. -

Ebeling, H. 174 Edgerton, F. 128; 202 Ernout, A. 103; 145; 147 Esser, S.J. 157 Geldner, K.F. 40; 41; 88; 126 Gesenius, W. 158 Goodwin, W. W. 1; 2; 72; 170f.; 203n. Hahn, Miss E. A. 2; 4; 22; 35; 36; 47;

66; 98; 117ff.; 126; 128; 132f.; 135; 188; 197; 198

Hale, G. 203n. Ha=erschmidt, H. 4 Handiford, S. A. 119 Havers, W. 53 Herbart, J. F. 48 Hirt, H. 1; 22; 34; 35; 49; 76; 193n. Hofmann, J. B. 145; 146; 177 Horn, W. 184n. Humbert,J. 4; 5; 92; 104 Jespersen, O. 1; 62 Kühner, R. 81; 175 Kurylowicz, J. 36; 112

Lambertz, M. 180 Lange, L. 149; 203n. Leenhardt, M. 17; 18; 29 Lerch, E. 182 Li Fang-Kuei 12 Lorck, E. 182 Macdonell, A. A. 35 Meillet, A. 133n.; 145; 147; 149 Meyer, L. 145 Miller, C. W. E. 34 Minard, A. 94 Moorhouse, A. C. 147f. Music, A. 147; 168ff. Mutzbauer, O. 4; 68; 174 Neisser, W. 114 NllakaJ;ltha 194 PaJ;lini .52; 101 Pedersen, H. 1 Pottelbergh, P. van 172 Reckendorff, H. 159 Reichelt, H. 111 Renou, L. 37; 38; 55; 59; 88; 89; 101;

110; 113ff. Schwyzer, E. 5; 6; 48; 104; 112; 119;

146; 149; 189 Seler, E. 11 Slotty, F. 143 Speyer, J. S. 198 Stahl, J. M. 80; 98; 143 Strehlow, T. G. H. 10 Streitberg, W. 34 Sturtevant, E. H. 132 Tabachovitz, D. 149ff.; 203n. Thomas, F. 103 Tozzer, A. 11 Tuuk, H. N. van del' 157; 162 Vendryes, J. 133n.; 149 Vogt, H. 109 Walter, A. 75; 76; 77; 86; 121; 133n.;

172; 174 Wackernagel, J. 44; 133n.; 145; 189 Whitney, W. D. 1; 35; 47; 114

Selected list of passages quoted

AthV. 4, 8, 2: 41 Bhiig. PUl'. 10, 28, 11: 195 Mbh. 3,268,19: 194 Mbh. 8, 70, 27: 194 ~V. 1, 24, 13: 190 ~V. 1, 113, 10: 85 ~V. 1, 123, 3: 173 ~V. 1, 132, 6: 126f. ~V. 1, 143, 6: 190 ~V. 2, 12, 2: 183 ~V. 4, 41,11: 174

~V. 7, 81, 5: 87 ~V. 8, 44, 23: 178f. ~V. 9, 101, 9: 126 ~V. 10,2,2: 96 ~V. 10,27, 1: 103 ~V. 10, 28, 1: 190 RV. 10, 53, 4: 126 SatBr. 11,5, 1,4: 107 Sat1?r. 13, 8, 4, 11: 106 SatBr. 14, 1, 1, 19: 179 TS. 2, 6, 6, 1: 105

Selected list of passages quoted

Dar. Bis. Al, 50 Homer, A 64: 139

A 137f.: 169 A 184: 77 A 262: 75; 85 A 280f.: 169 A 290f.: 169 A 324: 139 A 555: 72 B 340: 57 B 379: 174 r 281ff.: 147f. E 2: 124 E 679f.: 120 E 897f.: 187 Z 28lf.: 144 Z 459ff.: 75; 121 H 28: 149 H 29ff.: 77 f. H 87ff.: 75 H 171: 140n. 1 359: 147 1 362f.: 174 K211: 57f. A 389: 176 M618: 150f. M687: 145 P 70: 188 L: 308; 117 y 301: 123

y 371: 169 X 49ff.: 169f. X 123: 121 X 505: 139 Q 119: 140 f3 30f.: 118 f3 43: 140n. f3 184: 119f. f3 218: 147 e 415: 201 I¿ 383: 86 :re 255: 121 :re 437: 75 e 368: 124 rp 113: 144 Ij! 73: 77 Q) 544: 123

Arist. Nub. 269ff.: 169 Bacch. 5, 165: 144 Eur. Ion 1501: 183 Eur. 01'.1133: 181; 186f. Eur. TI'. 942: 169 Hom. h. Ven. 214: 124ff. Plaut. Amph. 456: 146 Plaut. Asin. 524: 146 Plaut. Epid. 543: 146 Plaut. Men. 962: 146 Plaut. Mere. 903: 146 Plaut. Most. 333: 146 Tac. Ann. 1, 63, 6: 189

211