·ORIENTALIA
RHENO-TRAIECTINA
EDIDERUNT
J. GONDA et H. W. OBBINK
VOLUMEN QUARTUM
J. GONDA
REMARKS ON THE SANSKRIT PASSIVE
ans
TUTA
RALLAS
LEIDEN
E. J. BRILL
1951

·w
REMARKS ON
THE SANSKRIT PASSIVE
BY
J. GONDA
UTRECHT
SUB
bini
PALLAS
EJB
LEIDEN
E. J. BRILL
1951

·Copyright 1951 by E. J. Brill, Leiden, Holland
All rights reserved, including the right to translate or to reproduce
this book or parts thereof in any forni
PRINTED IN THE NETHERLANDS

·CONTENTS I 9 9 9 21 22 I-INTRODUCTION II_TEXTS Satapathabrāhmana and other brāhmanas 1—The verbal adjectives II–The gerundives III—The present-system, formed with Já- IV-Other forms and Jaiminīyabrāhmaņa Aitareya-Aranyaka. Bșhadāranyakopanişad The story of Pausya (Mbh. I, 3), the stanzas excepted Sakuntalā by Kālidāsa, acts I and VI, prose Dasakumāracarita by Dandin, ch. 8. Pañcatantra, Hitopadeśa, prose Vetāla-tales, etc. Vetālapañcavimsatikā by Sivadāsa Rgveda Atharvaveda Kathopanişad Mahābhārata The story of Nala 1-19 (= Mbh. 3, a. 52-70) Rāmāyana: the texts in Boehtlingk, Chr.3 and 3, 11-15 . Saundarānanda of Aśvaghoşa, C. 1-6, 10, (14, 15) Meghadūta of Kālidāsa Raghuvamśa of Kālidāsa, C. I, 3, 5, 15, 16, 19 Kirātārjunīya of Bhāravi, c. 7-9 Kathāsaritsāgara 75, 21 sqq.; 76; 77; 83; 87; 88; 98; 9, 1-41 (= Vetālapañcaviņsatikā, Introd., Ist, 2d, 3d, gth, 12th, 13th, 14th, 24th, 25th story) III–CONCLUSIONS AND ADDITIONAL REMARKS. 27 29 31 34 37 39 41 46 47 53 . 5-7 55 56 60 62 65 66 68 + 70 73 7

·No text on this page.

·7 I-INTRODUCTION Although the use of the passive voice in Sanskrit is, in the main, known to the students of this language, it would not appear superfluous to examine the history and development of this verbal category and to make research into some questions connected with it. To begin with, two important questions arise: what is the function of the so-called Sanskrit passive, and how often is it found? After Whitney in his Sanskrit Grammar 1 had formulated the rule that the forms of the present-stem, which is characterized by an accented -y'á- added to the root, are used "only in a passive sense'', various scholars have repeated this statement in a less absolute and apodictic manner. The sense of the Vedic - vá-forms is, according to Delbrück 2, "überwiegend passivisch": "Neutral oder medial sind dhar- und mar-, und etwa noch dars- dhu- parc-11p-tac- vaſic-, Zweifelhaft bleiben tuj- und pan-". Speyer 3, who was guarded in his statements, taught with good reason that the passive voice, which is much used in Sanskrit both personally and impersonally, serves also to indicate such intransitive actions as the rice boils, the wood splits; "it is, however, not allowed to use that 'intransitive passive' of all verbs”. In another paragraph 4 the same scholar remarked that Sanskrit has a decided predilection for the passive voice; he added 5 that the use of impersonal passives is highly common. In his later book 6 he goes into some particulars: "Im Vedischen herrscht der activische Ausdruck bei weitem vor; das Skt. dagegen hat eine mit der Zeit zunehmende Vorliebe für das Passivum, auch von intransitiven Verben". It would, however, appear to me that this could be expressed more exactly. Renou 7 has already remarked: "le fait (viz. la faveur dont jouit l'expression passive), en tant que générique... va de pair avec le progrès du style nominal qui multiplie les adjectifs passifs en fonction de verbes”. On one hand one has more systematically to distinguish 1 W. D. Whitney, A sanskrit grammar, $ 768. 2 B. Delbrück, Altindische Syntax, p. 268. 3 J. S. Speyer, Sanskrit syntax, § 319. 4 Speyer, O.C., § 7. 5 Speyer, O.C., $ 8. 6 Speyer, Vedische und Sanskrit-Syntax, § 245. 7 L. Renou, Grammaire Sanscrite II, § 367.

·2 REVARKS ON THE SANSKRIT PASSIVE the various functions of the so-called passive verbal forms, on the other hand one has to inquire into the frequency of each category. If we neglect to make these distinctions it might lead to misunderstandings. Reichelts, for instance, is not right in asserting without any qualification that "im Indischen... die passivische Konstruktion die aktivische immer mehr verdrängt". I need not dwell upon the misconception that a formal category, e.g. the Skt. -d-class must of necessity have only one function. On the contrary, plurality of function is a common occurrence; compare such categories as the Latin rerbs in - and the French turns with se. Nor is it necessary to dispute the erroneous conception that the so-called passive is a 'passive' in every sense that expression may have, that the distinction between active and passive in the linguistic sense is congruent with the distinction between phenomenal, bodily, mental activity and passivity and the foolish, but almost ineradicable opinion that any active turn may be substituted by or converted into a passive construction. These mistaken opinions have now been sufficiently discerned by other authors, so that my remarks may be confined to these 10. In modern western languages the active turn does not only occur in such sentences as the man kills the horse, where horse is the ‘passive object', or the artist paints a portrait, where portrait is the ‘resultative object', or he talks politics (ʻInhaltsobjekt'), but also in I see the dog, I hear the lark, the father loves his son, he feels the beauty of the landscape, where the verb does not express an act resulting from a decision of the subject and the relation between object and subject is not the same as in the other examples 11. The active turn may occur without an object: it rains, the dog sleeps, the horse lies. We see that events, states etc. are often expressed by other forms than S H. Reichelt, Awestisches Elementarbuch, $ 615. 9 This point has been insufficiently discussed by Alice Berger, Der Ausdruck der passivischen Idee im Altfranzösischen (Berliner Beiträge zur romanischen Philologie 4:1; Jena-Leipzig 1934), p. 14 and by. E. Wistrand, Über das Passivum (Göteborgs Kungl. Vetenskaps- och Vitterhets-Samhälles Handlingar, 6 AI; Göteborg 1941), p. 12. 10 See e.g. H. Kallin, Etude syntactique sur l'expression du rapport de l'agent dans les langues romanes (Diss. Uppsala; Paris 1923). 11 I refer to Jinendrabuddhi's Nyasa or Kāśikāvivaranapañjikā I, 4, 49, who distinguishes between three kinds of karman "transitive object”: nirvartya „celui qui résulte d'une transformation", ex. katam karoti „, faire une natte"; vikārya ,,celui qui résulte d'une transformation", ex, kāşthāni bhasmikaroti „réduire du bois en cendres"; prāfya ,ce qu'on atteint (sans autre spécification de l'acte)", ex. ādityain paśyati „voir le soleil" (quoted from L. Renou, Terminologie grammaticale du Sanskrit I (1942), P. 123).

·INTRODUCTION 3 the so-called active categories. By the side of it rains, Lat. pluit, which expresses an event, we find Lat. itur I "going happens”, dimicatur "the (a) battle takes places”, datur "there is an act of giving”. By the side of the Dutch hij zit, Germ. er sitzt, we find in French il est assis. The case may present itself that an event affects an object, that in consequence of the occurrence of the event an object undergoes a transformation: Dutch de man wordt boos, de lucht wordt blauw, Fr. il se fâche; this idea and the state which corresponds to such an event or accident, viz. the transformation brought about or resulting from the event (Dutch de olan is boos, Eng. the inan is angry) are often denoted by the same forms as are used to express the so-called passive. To understand a passive category in other Indo-European languages, it will be useful to remember these facts. When a language possesses two or three verbal voices, as a rule one of them is more frequent than the other. In English and other modern Indo-European languages the so-called passive is only found in a minority of cases 12. The choice of this voice is mainly due to one of the following reasons: 1° it is not possible or not easy to state the active subject: now and then the agens is too indefinite to be expressly stated (Fr. on, Dutch men, which may mean "people in general"); then again it is indifferent; at times the subject is known to everyone, then again it is veiled by a mysterious haze or it remains in the background; sometimes, however, the subject is not known ("das passivum tritt dann ein, wenn der urheber der tätigkeit nicht bekannt oder so nebensächlich ist, dass er nicht genannt wird, vielleicht gar nicht zum bewusstsein kommt” 13); 2° the active subject is self-evident from the context; 3° there may be a special reason to make a person or a thing that is the centre of the interest the subject of a passive sentence; thus, the passive turn may be preferred, if the person speaking takes a great interest in the patiens; 4° the passive may be used when there are special reasons (tact, sentiment etc.) to avoid the active subject 14; 12 As to Slavic languages I refer to Fr. Miklosich, Vergleichende Grammatik der slavischen Sprachen, IV (1868-1874), p. 264. 13 W. Meyer-Lübke, Vom Passivum (Die neueren Sprachen, 6. Beiheft, 1925), P. 169. 14 A good instance is the Swedish use of the passive instead of the pronoun Ni: „Im Schwedischen, wo im allgemeinen der Höhergestellte nicht höflicherweise mit dem Pronomen Ni (2. Pers. Plur.) angeredet werden kann, sondern sein Titel gebraucht werden muss, hat man oft Gelegenheit zu beobachten, wie der schlichte Mann, falls er den Titel nicht kennt, oder dieser unbequem ist, seine Zuflucht zu passivischen Wendungen nimmt, die höchst sonderbar ausfallen können" (E. Wistrand, Über das Passivum, p. Io).

·4 REMARKS ON THE SANSKRIT PASSIVE 50 sometimes the passive turn facilitates the connection of the sentences 15; 6º the pass. constr. inay be used in order to produce a certain tension, e.g. when the person speaking wishes to suggest the unexpressed agens to his auditor 16; 7º at times it may, in certain languages, serve as a means to intensify the idea expressed by the verb 17, to accentuate the verbal form; 8º it may be used to avoid misunderstandings. - Yet, English has, like 'journalistic' German and Scandinavian, a predilection for the passive turn 18. It has more than once been pointed out 19 that it is the general rule with the passive turn' in many languages that the agens is not mentioned. The complete construction Lat. Petrus amatur ab Anna, Dutch de boom wordt door dien ouden man gepoot, is, comparatively speaking, uncommon. According to W. Nausester 20 the complete passive turn is not found in German proverbs: “Jedenfalls fand ich in Simrocks 13 Tausend Nummern enthaltender Sammlung kein einziges Beispiel". In many languages it is very rare: Ernout has pointed out that in 6-7000 lines of the Roman poet Plautus there are only two instances of the passive verb + ab) + agens 21. As a result of statistical investigations made by some pupils of Jespersen's we now know that 70-94% of the passive sentences in various English texts contained no agens. Yet, English has it comparatively often. According 13 See also O. Jespersen, The philosophy of language, D. 167 sqq.; J. Wackernagel, Vorlesungen über Syntax 12, p. 143 (with references); G. Overdiep, Stylistische grammatica van het moderne Nederlandsch (1937), § 280. 16 E. g. Anc. Fr. Thomas, Tristan v. 2081 f. Tristan i fud reconells, d'un sun ami aperceuz. „Derjenige, der ihn wiedererkennt, wird zwar genannt, aber zur Hebung der Spannung wird zunächst nur die erschütternde Tatsache mitgeteilt: „Tristan wurde erkannt"; dann erst kommen die Details” (Alice Berger, Der Ausdruck der passivischen Idee im Altfranzösischen, Diss. Berlin 1934, p. 12). 17 A. Berger, O.C., D. 15: „Es ist durchaus möglich, dass etwas aktivisch Gedachtes nur um der Betonung der Verbform willen passivisch ausgedrückt wird". E. Gamillsсheg (Historische französische Syntax) has emphasized this point. 18 See, e.g., M. Deutschbein, Grammatik der englischen Sprache für höhere Schulen (1930), p. 137; E. Einenkel, Geschichte der englischen Sprache II, Hist. Syntax, 3. Aufl. (1916) p. 136 f.; G. Wustmann, Allerhand Sprachdummheiten (Leipzig 1892). 19 See e.g. Jacob Grimm, Deutsche Grammatik IV, II; Th. Gartner, Rätoromanische Grammatik, $ 131 and other authorities quoted by W. Meyer-Lübke, Vom Passivum, p. 170. 20 W. Nausester, Das Kind und die Form der Sprache, Abh, aus dem Gebiete der pädagogischen Psychologie und Physiologie, VII, 7 (1904), p. 30 ff. 21 A. Ernout, Mémoires de la Société de Linguistique, 15, D. 329 sqq. ; see also E. Wistrand, Uber das Passivum, Göteborgs kungl. Vetenskaps- och Vitterhetssamh. Handl., sjätte Följden, A, I, I, P. & (1941).

·INTRODUCTION 5 to Ed. Schwyzer 22 there are only 5 cases of a personal agens in the first six books of the Iliad (+ 5000 lines). "Il n'y a que trois exemples sûrs dans la Chanson de Roland qui compte 4000 vers, et sur les 30 000 vers et plus dont se composent les cinq poèmes de Chrétien de Troyes que j'ai dépouillés je n'en relève qu'une trentaine" 23. Investigations made by my assistant, Miss de Zwart, have brought to light that in a part of a modern Dutch novel 24 there are found 339 active verbs, 23 reflexive constructions, 101 instances of the copula zijn, 18 passive verbs without an agens and 3 passive sentences with an agens; in another novel 25 (50 pages) these figures are: 435, 24, 83, 27 and 8; in another part of the same work (50 pages) : 476, 21, 104, 20 and 3; in a part of a third Dutch novel 26 : 407, 2, 66, 3 and in another part of this work: 545, 14, 104, I and I, in a third part: 384, 10, 64, 4 and o; in some ancient Dutch farces 27: 420, 2, 80, 7 and 3, and 226, 4, 65, 14 and 1. In two parts of a modern French novel 28 the figures are: 448, 48, 58, 8 and 3 and 473, 55, 75, 6 and 1. In a German scientific book 29 : 330, 49, 28, 77 and 9. "Es verhält sich nämlich durchgehend so, dass besonders von dem voll entwickelten Passivum ausser in der artifiziellen Schriftsprache sehr wenig Gebrauch gemacht wird" —, in this manner Wistrand 30 already formulated his view. According to Nausester 31 the turn der Schüler wird 2'011 Lehrer gelobt is only "ganz vereinzelt” to be found in the works of Schiller, Kleist, Hebbel (e.g. Agnes Bernauer I, 18: Der Herzog ist von den Bürgern so warm empfohlen worden). "Wenn ich aber ... die Aktenschreiber befrage, wenn ich die Bücher der Gelehrten und Verkehrten aufschlage, da kann ich eine reiche Fülle von solchen 22 Ed. Schwyzer, Zum persönlichen Agens beim Passiv, besonders im Griechischen, Abh. d. Preuss. Akad. d. Wiss., ph.-h. Kl., 1942, 10, p. 52. 23 H. Kallin, quoted by Meyer-Lübke, O.C., P. 170. 24 M. ter Braak, Afscheid van domineesland. 25 Vestdijk, St. Sebastiaan. 26 Den Doolaard, De druivenplukkers. 27 Taken from P. H. van Moerkerken, Het Nederlandsch Kluchtspel in de 17e eeuw I (1899). 28 A. Malraux, La condition humaine. 29 Th. Stcherbatsky, Erkenntnistheorie und Logik nach der Lehre der späteren Buddhisten, übers. von O. Strauss, 1924. 30 Wistrand, O.C., D. 6. It would appear to me that ,,die Kraft, die mächtig genug ist, eine ganz neue grammatische Kategorie (viz. the passive) ins Leben zu rufen..." does not disappear in a mysterious way, as is Wistrand's view („scheint auf eine höchst merkwürdige Weise zu verschwinden, sobald ein passives Ausdrucksmittel da ist"), but only realizes itself to a certain extent. 31 Nausester, O.C., p. 33 ff.

·6 REMARKS ON THE SANSKRIT PASSIVE Ausdrücken zusammentragen" 32. As has been explained by Ernout 33, the complete passive turn, which in ancient Latin was highly infrequent, has in that language developed itself but slowly. From these investigations we may conclude that in many languages an agentless turn is more needed than a 'complete passive construction 34. This fact agrees with the phenomenon to be found in some languages (Mandingo, Yoruba etc. 35), viz. the so-called 'Man-Passiv by the side of the active voice. In those languages there only exists a passive when the agens is not expressed or when it remains undetermined. We may also reier to the Old Irish passive, which occurs only in the third person: berir bretli foir "judgment is passed upon him”; to express the other persons the third person singular is used with an objective pronoun of the so-called logical subject, e.g., n0-11-berar one carries me" = "I am carried" 36. Old Irish berair means "there is a carrying", just like such Umbrian and Oscan forms as U. ferar, O. sakrafír mean "let there be a carrying" and "let one consecrate, let there be a consecration". When the evidence concerning these r-forms is taken as a whole, it would appear originally to have been an impersonal of the third person. In this treatise we will reserve the term passive sentence (without quotation-marks) only for those passive turns that contain an agens. In other cases the word 'passive' will be put between .... Another point to be considered is the origin of the “passive' forms in various languages. Confining our remarks to Indo-European languages we can say first, that in original Indo-European there was, apparently, no form especially appropriated to the expression of the passive voice. No passive category met with in the historical period was originally passive. As in Greek 37, Indo-Iranian, and Gothic the middle 32 I refer also to E. Löfstedt, Syntactica II, 367: „Der vollständige passivische Ausdruckstypus laudor ab aliquo ist im grossen und ganzen sowohl der einfachen, unbefangenen Umgangs- und Volkssprache als auch der dichterischen Sprache fremd... (Dieser) Typus hat in der Tat etwas Formelles, etwas Prosaisches und Papiernes..."; cf. also I, p. 71; J. Svennung, Untersuchungen zu Palladius und zur lateinischen Fach- und Volkssprache (Uppsala, 1936), p. 460. 33 Ernout, O.C. 34 Cf. also Rita Schlaepfer, Die Ausdrucksformen für „man" im Italienischen, Diss., Bern 1933. 35 See H. C. von der Gabelentz, Über das Passivum (Abh. sächs. Ges. d. Wiss., Phil.-Hist. Kl. 1867), D. 467. 36 See also H. Pedersen, Vergleichende Grammatik der keltischen Sprachen, II (1913), § 624. 37 See also Ed. Schwyzer, Zum persönlichen Agens beim Passiv Abh. Preuss. Akad. 1942, 10, p. 74.

·INTRODUCTION voice was used also as a passive, we may conclude that already in Indo-European times the middle forms, at least in the perfect tense, were used for this purpose. Other languages, however, have chosen reflexive forms to express the passive idea : the Balto-slavic family of languages, Swedish and Danish and also Gothic. As regards Sanskrit, everybody knows the intimate connection between the -ya- class and the 4th present-class, a very considerable part of which signifies a condition of mind or body or a state of feeling. It is not even possible precisely to define the limits of the two classes; in the older language there are a number of cases in which the accent wavers and changes. As this 'passive' in -ya- is found only in Indo-Iranian languages, it must be relatively young. Moreover, many a text has been interpreted and translated as if the verb were a real passive, whereas it might have been more reasonable to take it as an intransitive. But most 'passive' sentences in Sanskrit literature are formed by means of a verbal adjective in -tá- (or -ná-). This category is not a young one. In original Indo-European it must have been in frequent use as verbal and other (especially possessive, relational and characterizing) adjectives. In Greek the verbal adjectives in -tó- serve to denote a condition attained to: "Die Adjektiva sind überwiegend aus dem allgemeinen Verbalstamm gebildete Verbaladjektive aktiver und passiver Geltung zur Bezeichnung eines erreichten Zustandes (wie die Partizipia des Perfekts), aber auch der Fähigkeit, Möglichkeit (besonders negiert: &víxntos invictus), aber auch der passiven Notwendigkeit... ; sie sind in altertümlicher Weise nicht ins Verbalparadigma einbezogen, wie z.B. lat. amatus est.... In der ältesten Zeit und noch später herrscht das komponierte Verbaladjektiv vor (kennzeichnend &xívnto xai xeXlVnuéva ... Plat. Soph. 249d). Doch sind auch manche Simplizia alt (bes. adjektivische und substantivierte), wenn auch vielleicht erst zu den Komposita hingebildet («patoi te QaTOL TE, entoi t'őpontol te Hes. E. 3f.)... Solche Verbaladjektiva werden zu den meisten Verba nur gelegentlich, von sehr vielen überhaupt nicht gebildet” 38. In Latin and other languages of this family the same category has become a part. perf. pass. By the side of this the suffix -tó- is also a means of forming adjectives signifying the idea of being provided with, being supplied with, being fitted with: "le même suffix donne des dérivés de noms, indiquant la possession de telle ou telle chose" 39: lat. barbatus "(bearded)""; 38 Ed. Schwyzer, Griechische Grammatik I (1939), p. 501 sq. 39 A. Meillet, Introduction à l'étude comparative des langues indo-européennes, P. 268 sq.

·8 REMARKS OX THE SANSKRIT PASSIVE cf. also honestus "honoured" (honos); and such adjectives of participiai origin, but not ielt as part of the verbal system as gratus "pleasing”, certus "sure" (cerno). But the direct relation of the so-called passive participles (latus, lectus, potus etc.) to the verbal system developed only in the course of time. These adjectives in -tó- represent a category which is also found in other families of languages 40. Now, the authors of Sanskrit grammars have neglected to make a clear distinction between the 'passire construction' with a verbal adjective and the 'passive construction' with a personal rerbal form. In both cases the question whether the agens is expressed is of greater moment than their works would have us believe. The construction with the verbal adjective has to be studied in connection with the construction of nominal sentences in general. Like their counterparts in the sister languages the Indian verbal adjectives might be connected with personal iorms (as- etc.) and the use and development of this personal 'participial' construction is worth studying too. In addition to the above I should like to remark that the grammars dealing with the Indian verbal forms only inform us of the use of the passive in general, but do not descend to particulars as to the use made of it by individual authors. Vor do they give us any insight into a number of syntactical and stylistic details which will be discussed in this paper. Yet, some understanding of these and other questions in relation to the use of the passive voice in Indian literature will no doubt help to elucidate a number of difficulties concerning the diatheseis of the Indo-European verb in general. The more so, as the rather numerous syntactical studies in the sister languages may give occasion to conclusions based upon recent and detailed examination of the German, Latin and Greek facts and, at the same time, upon incomplete and partly antiquated inquiries into the Indian material. The present writer here undertakes to set forth what appeared to him the most essential facts in connection with the use of the socalled personal ‘passive' verbal forms in Vedic and Sanskrit and to describe their meaning or function, their syntax and the stylistic aspects of their use, in particular with reference to the question how far the so-called 'passive' is a real or complete passive. With that aim he has chosen a rather large number of texts and studied them separately in order to bring into relief as well as possible the charac- 40 See J. Gonda, Een onbevredigend behandeld punt in de Maleise grammatica, in: Bijdragen tot de Taal-, Land- en Volkenkunde van Nederlandsch-Indië 97, p. 515 sqq., and: Indonesische werkwoordsvormen, ibidem 105, p. 333 sqq.

·TEXTS 9 teristic features of the literary genre or species to which each text belongs. There was, however, no necessity to mention all particulars every time and to establish the same facts again and again. In addition to this the constructions with the verbal adjective have been attentively examined and, in order to throw light upon the function of the agens and the use of the verbal adjective as a predicate, reference has been given to the passages where these forms are used as attributes or appositions. The formal categories which together constitute the Sanskrit passive (participial constructions, -j'a-present, the other personal forms and the gerundive) have been treated separately, although in discussing them together some repetition might have been avoided. The forins without an agens and those with the agens expressed are consistently kept distinct. Furthermore, the author has tried to make a comparison between the partial results obtained in the said manner, and in doing so he has given his best attention to the relative frequency of the 'passive' forms and constructions and to some problems relating to the historical development of the passive construction in Vedic and Sanskrit. Lastly, a few remarks have been added concerning the passive turn in Indo-European languages in general. There has been no repetition of many useful and correct remarks to be found in the grammars and handbooks. Readers will be aware that the quotations from these well-known books are but few. II_TEXTS Satapathabrāhmana and other brāhmaṇas. I have examined a comparatively large part of the Satapathabrāhmana, viz. I, I, I-I, 3, 4; I, 8, I-I, 8, 3; 2, I, I-2, I, 4; 2, 2, 3-2, 3, I; 2, 4, 3-2, 6, 4, being 90 pages of Weber's edition. In this text five groups of 'passive forms are to be distinguished: the so-called passive participles (verbal adjectives) in -ta- or -na-; the so-called gerundives; the forms of the present-system, the stem of which is formed with the accented -ya-; middle forms used in a 'passive' sense the special passive 3d sg. of the aorist, ending in -i. I-THE VERBAL ADJECTIVES The verbal adjectives in -na 41 are met with in a comparatively small number of cases, whereas those in -ta- are very frequent. They are used in various ways: 41 As to their origin and early development see also J. Gonda, Zur Homonymie im Altindischen, Acta Orientalia 14 (1936), esp. p. 188 ff. GONDA, The Sanskrit Passive 2 ·IO REMARKS ON THE SANSKRIT PASSIVE A. 1) as pure attributive and descriptive adjectives. 1, 2, 4, 15 ned arena vajreņa saņšitena þrthivīm hinasānīti “lest I should injure the earth with this sharp thunderbolt"; 3, 4, 1o. I noted only a few instances, but some cases which I listed as 'appositions may also be considered as attributives. 2) Occasionally the negative element a- or an- which may appear before any adjective is prefixed to a verbal adjective: 1, 2, 1, 22 tasmād aniruktena jajuşā grhņāti “therefore he takes it with an undefined formula". 3) A few instances of the construction ‘ante Christum natum before the birth of Christ' are met with : 2, 3, 1, 8 yathā... āvasathenopak!ptenopāsīta “as...he would honour him by trimming his house”. 4) Now and then the substantive has not been expressed : 1, 1, 3, 5 itarāsu samsystam iva "with the other (water) some matter has become mixed up”; 9, I, 18 atiriktam "what is in excess"; 3, 3, 16 skannam (viz. butter) ; 2, 2, 3, 17 krtam; in fine compositi: 1, 2, 5, 5 yajñasammitam. B. In this case these forins are occasionally equivalent to substantives: 1, 1, 4, 3 pistam "flour”; 2, 1, 2 pistāni; 9, I, 5 pustam "wealth"; 8, 2, 13 yathāyathena klptena "in complete and proper order"; 2, 6, 1,6 dugdha-“milk”; 20 pariurte "in an enclosed place”; in a compound: I, 8, 3, 15 subhūtam "well-being"; 25 vaşaťkytam ; 9, I, I sūktavāka-; 2, 6, 1, 48 havirucchistam; 6, 3, I sukytam, etc.; 4, 4, 5 purohitaḥ. C. Apart from the attributive function the verbal adjectives, like other nouns, serve as predicates in nominal sentences. 1) The so-called verbum substantivum is very often not subjoined. a Many of the sentences of this Brāhmana are moulded in this shape. I have listed the following occurrences (mostly pure adj.): I, I, I, 20 evam hi mithunam klptam "for in this way a copulation is regularly accomplished" cf. 2, 5, 2, 17; 1, 2, 2, 9; 3, I, 25; 3, 2, 12; 3, 3, 16; 17;9, 2, 20; 2, I, 4, 2; 3, I, 20 parābhūtā vai tāḥ (viz. prajāh) “they (the creatures that are not allowed to take part in the sacrifice) are forlorn"; 4 3, 9 sarvatraiva svistakrd anvābhaktah "Agni S. is always allowed a share in it..."; 4, 3, 14; 5, 1, 5 prajāpatim evaitad abhyanūktam "... this is said", 5, 2, 2; 6, 1, 20; 6, 2, 7; 10; resultative: 2, 5, I, 17; (sambhūtāh after sambhavanti: 15); I, 9, 3, 6 yad vividham "what was torn"; 2, I, I, 2; etc. b have found one instance of a verb. adj. in fine compositi; 2, 3, I, 17 caturdhāvihitam hidam payaḥ. c Now and then the verb. adj. is followed by the particle iva : 1, 2, 3, 8 samtatam iva hi māmsam; 3, 3, 6; 2, 1, 4, 27. d Here too we meet ·TEXTS II with verb. adj. to which the negative particle a- has been prefixed: I, I, I, 13 anirukto vai prajāpatih ; 2, I, 4, 27; 3, I, 20; 4, 3, 7; see also 1, 8, 2, 2. The suffix of the superlative is added: 2, 4, 4, 19 ... payasyāvak! ptatamā bhavati. e The subject of the sentence is frequently left unexpressed: a In a subordinated sentence: 1, 2, 2, 16 yadā śrtah; 2, 3, 8; 9, I, 29 grad artvijye pravrtaḥ ("he becomes nonhuman) when he is chosen for the office of sacrificial priest"; 5, I, 15; see also I, 1, 1 Jatra yatrāgner vyaktan. ß in an apodosis: 1, 9, 2, 30 yad dhy urdhvam samistayajuso 'tiriktam; 2, 3, 1, 15 jan nu evaitad agne retas tena nv eva śrtam; ibid. y The verb. adj. may even do duty for a substantive (see above): (I, 9, I, 20 istam ca vittanı ca "what was searched for and found "); in a subordinate sentence: I, 6, 3, II yadi ... uddrstar "when ... (the new moon) becomes visible". f Beside the verb. adj. a noun-predicate is found: 2, 5, 4, 10 viśvam vā etat. karma krtam "that sacred work was made complete". g An object depending on a verb. adj. is rarely met with: 2, 5, 1, 5 tā ayam vāyith āvistah "these (the quarters) the wind entered", the preceding quotation from RV 8, 101, 14d has āviteśa; compare in § 4 praja ha tisro atyāyam iyuh (RV. 8, 101, 14a): tad yāḥ parābhūtās (author of the ŚBr.) "this is said regarding those who passed away" and ny anyā arkam abhito viviśre (RV. 8, 101, 14b): tā agnim abhito nivistāḥ "they settled down around the fire”. In § 1 and 2, however, where the facts are mentioned for the first time, the perfect is used: prajāh ... parābabhūvuḥ etc., in § 3, which relates the thoughts of the creator, the present: prajah ... parābhavanti. But when it is said that Prajāpati created beings that continued to exist because they could resort to female breasts, the conclusion is given in the shape of a nominal phrase with predicative verb. adj.: ta imā aparābhūtāḥ: "they are these (creatures) which have not passed away" = "... which are free from the quality of passing away": now, it is neither a (quasi-) historical fact nor an action that is related, but a state, a condition to this day extant. So, whilst interpreting (§ 4 and 5) the Rgvedic verses the author does not use a pure historical tense to express a (quasi-) historical fact or an action, but, taking up the substances within the range of his argument he subjoins nominal predicates to them and thereby he transposes the contents of his communications from the historical sphere into the timeless and therefore ever-present logical sphere. So, tad yāḥ parābhūtāḥ means “this is said regarding those that are qualified as having passed away", and tā agnim abhito nivistāh means "they (are those that) are qualified (and that are meant ·I2 REMARKS ON THE SANSKRIT PASSIVE in the Rgvedic text) as settling down around the fire". h In a limited number of cases an instrumental expressive of instrument or concomitancy has been added: I, I, I, 14 = 2, I, I, 4 adbhir vā idam sarvam āptam "all this is (qualified as being) pervaded (or obtained) by water"; 2, 1, 16 antariksena hīme dyāvāprthivi vistabdhe “by means of the atmospheric region the sky and the earth are in the state of being) firmly kept asunder"; 8, 1, 36 etābhyām evaitad āśīrbhyam sarvam aptam "through everything is obtained" (et. āś. is clearly not the agens, but the instrument); 2, 3, 10 pāpmanānusaktaḥ "affected with evil". In the two texts quoted in the first place the instr. expresses the instrument when we connect the sentence with the following part of the argumentation (e.g. 2, I, I, 4 adbhir evainan etad aptvā), but the statement means also: "... pervaded by water" and then adbhih has to be considered as an instr. agentis. 2, 1, 2, 5. An instr. agentis is found 2, 5, 1, 4 etad rşiņābhyanuktam. As for I, 2, 5, 8 and 9 see p. 19. j Under a I listed only occurrences to transitive verbs; here I add the verb. adj. derived from intransitive verbs, used in the same way: I, 2, 4, 14 asyām hime sarve lokāḥ pratisthitäh "for on it all these worlds rest” (result.); the same verb. adj.: 3, 3, 10; 8, 3, 15; 9, 2, II patnīşu vai yonau retah p-m "in the wines, in the womb the seed has is known to have) its place, it is a property of the seed to... in..."; 29 he bestows plants on this world tā imā asmim loka oșadhayah p-āh "(and so as is well-known) have a firm place in...";9, 3, 12; 2, I, I, 6 in a 'logical' conclusion which elucidates a well-known fact; 2, 5, 1, 4 parābhūtāḥ and niviştāh (see above); 6, I, II at the beginning of an argumentation devāh manusyān abhyupāvrttāh "the gods came ... to men” (it is a 'quality of them); 6, 2, 9 ya asya prajā jātāḥ (j. emphasizes the 'quality' not the historical fact). The negative a(n)- has been prefixed : 2, 3, I, 13 anupasthita-; 5, I, 4 aparabhūta-; 5, 2, 3 ajáta- 2) a To a verbal adjective used as a predicate, the verbs asti or bhavati are, however, not seldom added: asi is found only in quotations from Vāj. S., among the other forms of the verb as conjunctives and optatives are, comparatively speaking, more frequent than indicatives: 'I, 9, 3, 19, but here the verb. adj. does duty for a subst., nātra tirohitam ivästi; 2, 3, I avasthyūtā stha "you have been spat upon"; 2, 5, 4, 10 sarvam jitam devānām āsīt; 1, 9, I, 24 tat tad antarhitam iva manusyebha āsa (not antardadhe) and 25 tad vā rsīņām anuśrutam āsa, in a sense, these two sentences do not form part of the 'story' in which perfects are used (vidām cakāra, apīyāya), I should rather say ..

·TEXTS 13 built up, that the author not only relates ‘historical facts, but also his own opinion with regard to a position of affairs in the past; in the same way I, 2, 15 ilvāca upādhatta teşām al pakād evägnir asamcita āsa atha hovāca that fire of theirs wanted but little to be completely when he said”; 2, 2, 4, 3 (condition resultant on a preceding development) I, 3, I, 20 (thinking:) sarvo me yajña āhavanīye šyto 'sat; 2, 3, I, 4 śrtam asad iti; I, I, 4, 23 (he says :) supratigrhītā usan; 2, 1, 13; 21; 2, 5, 1, 18; 2, 27 (always in a formula to be thought or said, semi-nominal sentences); 20 jñātibhyo hāsyai tad ahitaın syāt (that will turn out injurious to her relatives”; 1, 3, 3, 16; 46 syātāni; 6, 2, 6 syuh (3 X), the verb. adj. followed by syāt etc. are, too, practically adjectives: the prefixes a(n)- and su- are often met with, cp. also 2, 6, 2, 6 abhimāniko ha rudrah paśünt syāt; I, 9, 3, 19 santu (words to be said). Forms of the verb bhū- are more numerous, but I have met with no other forms than bhavati, bhavanti and (once) bhavatah ; these turns are in general expressive of an opinion or conclusion of the author, of quality, result, state: I, I, 2, 21 parivrtam iva vā etad ano bhavati tad asyaitac cakṣuḥ pāpmagrhītam iva bhavati (“find itself in the position of being..."); I, 4, 24; 2, 3, 5 tathaiņām śytam bhavati “thus it becomes boiled (drinkable) for them"; 8; 3, I, I; 8, 3, 25 samsraco hy eva khalu parišisto bhavati "the residue... is that which remains"; 9, 2, 14; 2, 2, 3, 21; 5, 1, 18 (expressive of state) 5, 2, 20; 5, 4, 10; 1, 8, 3, 7 sarvā vai tat tā istā bhavanti "all (these deities) are now characterized as being offered to" i.e. "to all offering has now taken place”; 8, 3, 10; 25; 9, I, 10; 9, 2, 26; 2, 5, I, 18 prasva upasamnaddha bhavanti lit. "... are tied"; 6, 1, 5 dhānā apistā bhavanti ". remains unground", relation of a (negative) fact, but also a conclusion of the author about a state of being, the sentence tato 'rdhāḥ pimsanti which precedes expresses only a fact; 8; 9; 2, 5, 3, 5; bhavatah: 2, I, 2, 3. Besides, there is an instance of the verb sthā- subjoined to a verb. adj.: 2, 1, 2, 6 prajäḥ... ekarūpa... tasthu "... remained”, cp. my remark ad 2, 6, I, 5. 2) b The verb. adj. forms the last part of a compound: 1, 1, 2, 21 (see above), c The verb. adj. may be followed by iva: e.g. I, I, 2, 21 (see above). d Now and then a(n)- has been prefixed: 2, 6, 1, 5 (see above). e There are also a limited number of intransitives: 1, I, 4, 14 tasmin... väk pravistāsa "into him had entered a... voice' (Eggeling), it is not a "historical event, but a state of affairs resultant on a preceding event; 3 a skin of a black antelope is used "in order that nothing of the oblation may get spilt": askannamn havir asad iti..., "if any th ·14 REMARKS ON THE SAYSKRIT PASSIVE grain... should now be spilt on it...”: jad evātra tandulah ...skandāt (ask. h. asat: the oblation may continue to be askanna-); 1, 2, 15 apasyati ... vady atra kimcid āpanna.in bhavati : ap. expresses a fact, ap. bh. a state. f There are a comparatively large number of instances of the same turn, an instrumental being added: an instr, expressive of means or instrument: etc.: I, I, I, 15 tad evāszaitena sarvam āptanı bhavati ; 1, 4, 5 abhinihitam eva savyena pānina bhavati "while it is still being held down with his left hand -", but the author adds : atha ... āharati (action) "then he takes..."; the same:6; 2, 1, 7:8;2, 1, 22 etat ... jajuşaiva grhītam bhavati “it is customary to take..." (concluding and elucidating remark of the author) beside, in the same $, tad itarena vajuşā grhạāti "he takes...” (fact); 2, 2, 3, 13 rātribhir hy evāntarhitau bhavataḥ, the word r., being opposed to agnibhyām in the preceding sentence has been emphasized and so occupies the first place, it is, however, not made the subject of the sentence: parallelism, the subject need not be expressed, and a predicate expressive of state is chosen; 1, 4, 29 (2 X). g The predicate is attended by a genitive; 1, 2, 3, 5 tathaisām śrtam bhavati "thus is become boiled (drinkable) for them"; 3, I, I; 3, 18. Thus far the genitives express the person on whose behalf the process is going on, or who is interested by it; a genitive of agens presents itself 1, 8, I, 39 yajamāno vai dhruvo tad yajamānasya prāśitar bhavati; ibid., where the underlying possessive character seems to be apparent:"the dh. represents the y.: hence this is the food of the y. (i.e. is eaten by him)”; compare also 2, 5, 2, 46 y'e yajamānayor vāsasī parihite syātām "the garments, worn by the y. and his wife..." 2, I, 4, 29 asya āptam bhavati, gen. of agens < gen. of concern; compare 4; 4, 18 sarvam ha vā asya prītam bhavati sarvanı āptam yah ...; I, I, I, 15; 2, 5, 4, 10 sarvam jitam devānām āsīt etasja ... s. j. bhavati, translated by Eggeling: "to the gods all was conquered all is conquered by him”!; 1, 9, I, 25 tad vā ?šīnām anuśrutam āsa may be translated : "t became known to the ſşis", but the rșis are, anyhow, the agents of the action of hearing ; 2, 5, 2, 3 it reads yāś caivāsya prajā jātā asan "... that were born to him” (cp. 22), gen, of concern, but, of course, he is also the begetter of the children. 3) Over and above the use of the predicate verbal adjective hitherto mentioned it is not infrequently expressive of 'actuality', i.e. it expresses such acts as are simultaneous with other acts done or effects realized. But at the very moment these acts are performed and the formulas which contain the verb. adj. are recited, they come to belong ... .

·TEXTS 15 to the sphere of that which has come to an end, of the complete. So we observe the transition to its being used as a category expressive of the 'actual past, i.e. of such past acts as are so recent as not to have lost their actuality at the time of their being related, of such past acts as bring about effects or results that are of moment to the person speaking and his auditors. This is, however, not precisely the use in the texts quoted below. This use is found only in quotations from the Vāj. Samhitā, which are to be recited by a priest and in other formulas. They are directed to divine and other revered powers, metra, animals, oblations, food etc. included. a There verb. adj. is not accompanied by the verbum subst.: I, I, 2, 2 (Vāj. S. I, 7a) pratyustam raksaḥ pratyuştā arătayaḥ "he now heats (these objects) saying 'scorched is the raksas, scorched are the enemies" ; ibid. (Vāj. S. 1, 7b); 1, 2, 15 (Vāj. S. 1, ge) apahatam raksah "repelled is the r."; 1, 4,4 (Vāj. S. I, 14b); 21 (Vāj. S. 1, 16d; e); 2, 1, 14 (Vāj. S. 1, 19); 1, 3, I, 4 (Vāj. S. I, 29); 9, 3, 10 tad u tat prthivyām vişnur vyakramsta gāyatrena chandasā tato nirbhakto yo 'smān dvesti etc. "hence (he strides thrice) with the texts Vāj. S. 2, 25 a-c); 'on the earth V. strode by means of the g.: excluded therefrom is he who hates us'” (3 X); also 12 (3X); in these cases the priest performs an act and recites the appropriate words: by the act and the words the effect he has in view has been realized 42; cf. e.g. I, I, 4, 4 "he shakes it (the black antelope skin) saying: 'shaken off is the rakşas...', whereby he repels from it the evil spirits, the raksas”. An instructive text which occurs four times (1, 8, 3; 10; 9, 1, 1; 2, 5, 2, 42; 6, 1, 45) isitā daivyā hotāro bhadravācyāya preșito mänuşah sūktavākāya "the divine hotars are summoned for the proclamation of success, the human one is (by these words) called upon for the song of praise": 1, 8, 3, 10 the author adds "by these words he urges on the human h. to singing praises": so, preșitah means: "he is affected by the action of calling upon (viz. by means of this formula) > he is called upon”; in another formula (cp. TS 2, 6, 7, I; 2) 1, 8, I, 19 etc. upahūtam vīmadevyam etc. (3 X); 20; 21; 22; 23 (2 X); 24 (3 X) etc., cp. also 29 upahüte dyāväpythivi ... tad ime d. upahvayate "he thereby calls to him heaven and earth". – An interesting text, which presents this turn in a narrative section, is 2, 4, 4, 6 teno ha tata īje daksaḥ ... ime 'by etarhi dākṣāyaṇā rājyam ivaiva prāptā rājyam iha vai prāpnoti yah... “D. 42 As is well known the word of man has a magical and creative power; it is able to have the same effect as a rite. See, e.g. J. Gonda, Remarks on similes in Sanskrit literature (1939), $ 42, ·16 REMARKS ON THE SANSKRIT PASSIVE performed that same sacrifice; and even to this day these descendants of his are possessed of the royal dignity; r.d. he, therefore, here obtains, who..." I am inclined to list this occurrence under CI (see above), but some notion of the actual past may be present too: "the descendants have obtained the r.d. and so they do in these days”; prūpnoti, however, is expressive of the 'timeless present'. - An instance of this use of the verb. adj. accompanied by a personal form is found 1, 1, 3, IO (Vāj. S. 1, 13d) proksitā stheti ... atha haviḥ proksati. A pronoun expressive of agens attends on the verb. adj. in the words to be spoken by the ksatriya: 1, 3, 2, 15 vaiśya maji yat te putro nihitam tad āhara "...bring to me what you have stored away!" 4) The verbal adjective may be the predicate of an absolute locative denoting the circumstances under which the main action comes to pass. As is well known, this locative is closely connected with the locatives of time and circumstance: 1, 8, I, 4; 5; 39 ned asamsthite jajíže prāśnānīti; 2, 5, 2, 30 sammrste 'gnau; 5, 3, 17 etc. Not infrequently only the predicative locative has been expressed: I, I, I, 3 samsthite "after the completion of the sacrifice"; 6; 2, 5, 2, 31; 37 vasatkrte . juhutah "on the 'vasať being uttered..."; 3, 17 atha prätar hute vāhute vā "then, in the morning, either after or before the performance of the agnihotra"; 3, 1, 9 astain ite. The loc. abs. is, however, not often to be met with. 5) Now and then the accusative of a verbal adjective is employed as a predicate (1 X P., 8 X th.): 1, 2, 5, 16 justäm evainām etad devebhyaḥ karoti “in this way he makes it pleasing to the gods”; 2, 1, 4, 4 tat suhitam kartavai brūyāt; 3, I, 9 vichinnani agnihotram manyamahe; 13 (intrans.); the suffix of the comparative is added to the verb. adj.: 1, 3, 2, 12 attūrain evaitat parimitataran ... karoti; the substantive has been left unexpressed: 2, 2, 14 Srtam vedāni "I will ascertain whether (it) has been baked" (= 2, 3, I, 16); 8, 2, 9 suhitam kuruta; 2, 5, 26. 6) Occasionally impersonal forms are met with; 1, 2, 5, 18 etasmin devayajana istam bhavati“ the sacrifice takes place, is performed"; 19; 2, 4, 4, 17 te evaitad yajati ... asya daśapūrnamāsābhyām iştam bhavati va evam etad veda "it is to these two gods), that he offers and ... to the new and full moon offering takes place by the influence of him who..."; the same thought is first expressed by an active sentence, then by an impersonal-intransitive turn of a complementary character which emphasizes the efficacy of the action and its result; cf. § 16 "whosoever prepares the s. at full moon, renders his ...

·TEXTS 17 > - offering palatable (stadayati) to the gods, and his o. is p. (svadate) to the g.". D. As is well known the Sanskrit verbal adjectives are also used to express attendant circumstances or other qualifications of the main action, whether temporal or concessive, causal, conditional etc. As this use of the mere participle (participium conjunctum) is far less known in modern languages such as English or Dutch, various conjunctions are to be added, when translating into a European language. In the parts of the Śat. Br. dealt with in this survey this mode of employment of the words in -ta- is often to be found. 1) Aparatactically joined intransitive verbal adjective expresses attendant circumstances or other qualifications of the main action: I, 1, 3, 2 so 'yam puruşe 'ntah praviştah prāň ca pratyari ca "on entering into man it becomes a forward and a backward one", 8, 3, 12; 2, I, 1, 6; 6, 3, 6 payo ha vai prajā jātā abhi samjánate "... when they are born"; 5, 3, 20 maruto ha vai krīdino vstram hanisyantam indram āgatam tam abhitah paricikrīduḥ; 3, I, 9 agnāv evaibhya etat pravistebhi'o juhoti; ibid. apretebhya evaibhya etaj juhoti; 2, 4, 4 tasya bhitasya SVO mahimäpacakrāma "he (Prajāpati) being terrified, his own greatness departed from him”; a noun-predicate is added : 1, 8, 3, 14 vaya evainam etad bhitam. 2) The occurrences of transitive verbal adjectives employed in the same way are more numerous: I, I, I, 9 (quotation) sa yad evāśitam anasitanı tad aśnīyād iti "let him therefore eat what, when eaten, counts as not eaten"; 2, I, 22 yad vā adistam devatāyai havir grhyate "now whatever oblation, in being taken, is announced to a (particular) deity ..."; 2, 5, 19 hatah puruso hato 'svah sete; 3, I, 12 (thought); 8, 1, 14; 8, 2, 8 (3 X); 2, 3, I, 3; 4, 3, 25; 5, 2, 20 niruktam vă enah kaniyo bhavati "when confessed the sin becomes less"; 5, 3, 3 anann eva prānan paridīrnah śiśye; I, 3, 2, 6 Gilluşmai tvā juştan grhņāmi; I, 3, 3, I; 2; 3. is a well-known device in Sanskrit literature to begin a new sentence with an absolutive derived from the same root as the verb of the preceding sentence and repeating in a succinct form the contents of that sentence. By means of this anaphorical and recapitulating use of the absolutive which is only a special case of a mode of connecting sentences often to be found in various languages, especially in primitive and popular speech, the speaker or the author makes an opportunity for himself and for his auditors or readers to pause to collect their thoughts, to gain starting-points to which he may add a new link of his 3) It ·18 REJARKS OY THE SANSKRIT PASSIVE . argument. So the author of the ŝat. Br. says, 2, 3, 3, 3 sa vat agni ādhatte, tad enam janayati; tam janayitva bibharti "when he establishes the two fires, he reproduces that (viz. Agni); and having reproduced him, he maintains him”. Now, the same device is used when the subject of the subjoined sentence is not identical with the subject of the preceding sentence, but with its object: then, however, it is not possible to use the absolutive and the author has to resort to the participium conjunctum. Thus used, the verbal adjective clearly expresses the state of being effected by a preceding action. This use, which occurs frequently, harmonizes with its character as a verb. adj. I, 1, 3, 5 tam indro jaghāna. sa hataḥ putih sarvata evapo 'bhipra susrāva "him Indra slew. He being slain flowed forth ... towards the water”. In the same manner, when the preceding sentence is subordinated: 2, 4, 1 indro ha yátra vytráya vájram prajahára, sá práhytaś catúrdhābhavat, translated by Eggeling: "when I. hurled the thunderbolt at V., that hurled one became fourfold". See also 1, 3, 4, 6 (2 X); 7 (2X);9, 2, 1, 2, 6; 3, I, 32; 33; 5, 2, 30; 5, 3, 3; 6, 1, 23. The verbal adjective used in this way may be intransitive: 2, 4, 4, 20 ... āhutibhyo 'dhijanayati; sa esa ahutibhyo jātał... When the verbal adjective is intransitive and the subject of the subjoined sentence is identical with the former subject the same construction is possible: I, 4, 27 ...tad asyām pratisthāyām pratitisthati; so 'syām pratisthita ādhatte... The subject of the second sentence is neither identical with the subject of the former sentence nor with its object: 5, 4, 7 iyam (prthivī) haivainam badhāya prati pradadāv anayā haivainam prati prattam jaghnuḥ; I, 9, 2, 26; after a -ya-form: 2, 4, 4, 22 retah sicyate; tad ytavo retah siktam imāh prajāḥ prajanayanti. Now and then the recapitulating verb. adj. forms part of an absolute locative: 2, I, 2, 16 tāın (a brick) abhipadyābabarna; tasyām ābrdhāyām agnir vyavasašāda; 5, 2, 30. There is an example of its being used without a substantive: 1, 3, 1, 3 ...nirnenekty evainā etan nirņiktābhiḥ pracaranīti "he in reality rinses them, thinking 'with these that have been rinsed I will proceed”. Only occasionally the participle sant- is added to a verbal adjective: 1, 8, 1, 17 ...samavadyati. samavattām eva satīm tad enām pratyakşam hotari śrayati "he now cuts off in the hand of the hotar. That which is cut up piece by piece he thus makes visibly enter the hotar"; here the subject remains the same and so the absolutive might have been used, but the construction chosen lays stress on the patiens, not on the action, and the addition of satim seems to point out that the action has taken place and that the result is an accomplished fact. Now and then one or ·TEXTS 19 two other sentences are found between the personal form of the verb and the verbal adjective: 1, 2, 4, 7 samsyaty evainanı ... iti yadi iti brūvāt tena sansitena nātmānam upaspršati "he makes it sharp. When he (further) says..., let him ... say.... When it has been sharpened, he must not touch himself with it". From time to time we meet with a verb. adj. used in this manner without a preceding personal form of the same verb: then the meaning has to be concluded from the context: I, I, 4, 15 "Manu, we will sacrifice for thee ... with this bull':: ‘so be iť. On his (the bull's) being killed the voice went from him": tasvālabd hasja sú väg apacakrāna. As for the verb. adj. being accompanied by an oblique case, we may point out some interesting cases: 2, 5, I, I sa prajā asrjata; tā asya prajāh systā parābabhūruḥ. Now, Eggeling has translated this text as follows: "he created living beings. The 1. b. created by him passed away". But, § 2, it reads sa dvitīyāh sasyje ta asya paraiva babhūvuh etc. and, $ 3, anaśanata yā vai me prajāḥ parābhavanti “my creatures pass away from want of food”. In § 3 the word me may be most probably considered as a possessive genitive, though a genitive of damage 43 would be possible too; in $ 2, however, the words tā asya etc. taken as a thing apart, I should like to give the preference to a genitive of damage, which would be also possible in § 1. But $3 it reads sa prajā asſjata; tá asya prajāh systāh... sambabhūvuh, and here Eggeling's translation seems to be correct: "the beings created by him tinued to exist”. So asya in § 1 may be considered too as a genitivus agentis, but its connection with other employments of this case is perfectly clear 44. 2, 5, 2, I only a gen. agentis is possible: prajāh sasrje tā asya prajāḥ systă varuṇasya yavăn jaksuh. The verb. adj. is accompanied by an instr. of instrument: 1, 2, 5, 8 (if we may take this text as an instance of a nominativus pendens) so 'yan vişnur glānah, chandobhir abhitah parigrhitaḥ ...nāpakramanam āsa. I have not met with an instance of an instrum. agentis, but 5, 2, 2 presents a tatpuruşa-compound of which the first member represents an instrum.: tā varuņo jagraha ; tā varunagrhītāḥ paridirnah "... on being seized by V., they became...". It is worth mentioning that there exists also a nominal recapitulating construction: 2, I, 2, 16 agnir vyavašaśādāgner vyavasādam anv asură vyavaseduh and along with the falling fire-altar the Asuras fell down". con- . 43 See Speyer, Sanskrit Syntax, § 131. 44 See Delbrück, Altindische Syntax, § 106.

·20 REMARKS ON THE SANSKRIT PASSIVE 4) Some of the peculiarities treated under 3 present themselves also when the verbal adjective does not recapitulate the contents of the preceding sentence. It is used without a substantive: I, I, 1, 21 no anāptah sādayet “... nor should he put (the water) down before reaching..."; 8, I, 16; 2, 4, 3, 9; 4, 4, 25 upahūta upahvayasveti; 5, 2, 29. The verb. adj. is accompanied by an instr. of means or instrument: 2, 6, 1, 6 dugdhe sakrd upamathita e kaśalākayā “(being mixed) with the milk... by stirring it once with a single splinter”; cp. also 5, 2, 14 karambha pātrāṇi kurvanti yāvanto grhyāh syus tāvanti ekenātiriktāni "... exceeded by one"; by a pure instr. agentis: 1, 9, I, 2 ...yo yajate; etena hy uktā ?tvijas tanvatc "...since it is by his order that the priests spread it”. 5) The verbal adjectives used in this manner are inore than once compounded with a preceding adverbial element: thus with the negative Q- or an-: I, I, I, 18; 8, 3, 24 yad vā anădistam devatāyai havir grhyate "when any sacrificial food is taken without being announced to any one deity”; 9, 2, 32; 2, I, 4, 26; 5, 3, 16; 17; 1, 3, I, 6, which forms at the same time a good instance of the employment of these adjectives as equivalent to our adverbs: yathānuparato yajamānasya sa patrăn kşipuyat “that it may unceasingly destroy the enemies...”. 6) The particle iva is added twice: 1, 2, 2, 16 (thinking :) ned veva magna iva mușita iva śayātai. Of these occurrences + 40 refer to a non-personal notion, 5 to an animal, 6 to a god, 4 to a dragon, 6 to 'creatures' or 'beings', 13 to a human person. E. Rarely a verbal adjective serves as an adverb: perhaps I, 3, 3, 6 praklptam haivāsya strī vijāyata iti, according to Eggeling p. means swithout a difficulty", according to Sāyana, however, "a completely formed (child)". Such cases as 1, 9, 3, 3 samtatan avyavachinnam ninayati are not to be considered as adverbs: cp. I, 3, 5, 13; 16 etc. F. There are a good many instances of a verbal adjective forming part of a compound. 1) The verb. adj. is the last member of the compound: I, I, I, I pavitrapūtaḥ; I, 2, 9 agnidagdham; 17 savit, prasūtaḥ; I, 3, 6; 2, 4, 4;2, 5, 4, 5; I, 4, 19 varşavrddham; 2, I, 6 yajuskrtah; 2, 3, I caturdhāvihitaḥ; 2, 5, 14 puruşasammitā; 3, I, 24 ajyaliptābhyām ; 8, I, 12 pancāvattā; 17 ātmamchrtayā; 9, 2, 35 sunirbhakta- (2 X); 2, 2, 4, 15 prathamadugdham; 3, I, 9 hutocchistam, 39 agnihotrocchiştam; 4, 4, 5 purohitaḥ; 5, 2, 2 varunagrhitaḥ; 47 devakrtam etc.; 6, 1, 6 agnisvättebhyah; 48 havirucchistam.

·TEXTS 21 2) The verb. adj. forins the first member: 1, 2, 3, 9 apakrāntamedhā); 2, 5, 5 yajñasammitam; 16 vimystāntarāmsā; 9, 2, 12 etc. samistayajuḥ; 2, 1, 3, 4 anapahatapāpinānalı ; 2, 3, 17 kytakarmā; 3, I, 9 uditahominām; II hutocchistasya ; 5, 2, 48 utsannayajſiah ; 6, 1, 37 āhitāgnih. 3) The verb. adj. appears in a longer compound: apart from 2, 5, 2, 42 sūktavāka-: 5, 3, 16 havirucchistāśāh. So, these compounds are, on the whole, rather simple and not elaborated. II_THE GERUNDIVES The gerundives, of which I have listed 14 cases (exclusive of the occurrences entered under E), are used in various ways, but never as pure attributives: A as predicates without the attendance of the verbum substantivum (7 X): 2, 2, 3, 3 they said: kim iha kartavyam "what is to be done here?" ; 2, 4, 8 = 18 agnihotram hotavyam; 12 tasmād esopajīvanīyā (eșā: the cow); 5, 2, 33 with an attending instrumental vat kiin ca vācā kartavyan (instr. of accomp. circumst.); with a personal subject... I, 9, 3, 2 prati tam osato yah pratyusyo 'ty u tam syjete yo 'tisyiyah “they scorch him who deserves to be scorched, and allow him to pass who deserves to pass”; also: martya- "mortal” 2, 1, 3, 9. B as predicates accompanied by the verb bhu- (2 X): 2, 6, 1, 12 harati yad dharyam bhavati "removes (the dust) which has to be removed"; with a personal subject 2, 4, 12 upajīvanīyo ha vai bhavati ya veda "he is affording a livelihood who knows...". The comparison of these two groups, A and B, seems to show that, on the whole, the gerundive accompanied by bhavati emphasizes the being in a state to be of a person or a thing 45, whereas the mere gerundive in first line expresses what must happen. C the gerundive is used as a predicate, but the subject is not expressed: 1, 2, 3, 9 tasmād eteşām pašūnām nāśitavyam "for this reason one should not eat of these animals" (partitive genitive); the same impersonal (neutral) construction, but with the attendance of an instrumental expressing the agens: 9, 3, 16 (quotation from Yājñavalkya) tad dhy eva brāhmanenaistavyam yad brahmavarcasi syād iti "for this the B. should strive, that he be...". D the gerundive has been added as a predicate to a noun or a pronoun which depends on a verb: 2, 2, 4, 16 (they ... 45 Compare Speyer, Sanskrit Syntax, S 3 Rem.: "bhavati is the proper verb, if there be laid some stress on the predicate, in other terms, if it be pointed out that the subject is invested with the dignity or possesses the quality predicated of it".

·22 REMARKS ON THE SANSKRIT PASSIVE said :) sa jasmai na idani prathamaya zotavyar vaksyati... "to whichever of us he says it shall be offered first..."; the substantive has not been expressed 2, 6, 1, 33 (quotation from Asuri) prāśitatryan tv eva vayam manyāinaha iti “but we think that some should be eaten"; 1, 2, 3, 2 the gerundive is subjoined to a genitive denoting the object of the verb vid- "to know": tasya haite 'pi badhyasja vidāſīcakruh "they also knew of his going to be killed”. E i the gerundive may be used as a substantive: havyain "oblation" I, I, 2, 23; ādya- "which is to be eaten, food" e.g. 1, 3, 2, II (beside attar- "eater"); 8, 2, 17; 8, 3,6; anūbandhyā (viz. gauh) term denoting a barren cow which "has to be bound afterwards" 2, 4, 4, 14; cp. also the term āhavanīya (e.g. 2, 3, 1, 7) "the eastern of the three sacrificial fires” (litt. "which has to accept the offerings”). 2 this substantive may appear as the second (last) member of a compound: devayajyā = "worship of the gods". - So, there are 5 examples of a personal subject and one of an instruinentalis agentis. III--THE PRESENT-SYSTEM, FORMED WITH -YA- I have listed about rio instances of personal forms of the presentsystem of this class, which belong, however, to a rather restricted number of roots, certain forms being used very frequently: kriyáte ( 30 times), grhyáte, hūyáte, etc. Only a few instances are met with of forms that are not indicatives: vyārdhyanta, atapyata, asiryata, dhriyekan, nirniucyeta, nucyātai. The forms of this present-system are hardly əZE. ised with a noun or a proper name expressing the agens and in pinoys we never find a nominal case (e.g. an instrumental) deter- -4019 paa nction expressed by the verb. But, although they serve to Suissəidxa Teitive actions, they are not restricted to the sphere of pure (ieninəu) [eutor clearness' sake the most striking modes of using Jo yea you pill be severally treated. ppuusui 6 E 'Onsitives are found: I, I, 4, 22 where an energetic form sỊ Ba|punsəSrb precedes eșa vă idam sarvam vivinakti yad idam əvət ayı semte "(it is the wind) that separates everything here (on əy7 səziseridergoes separation"; 2,4, I aśīryata (2 X): tasmäc charo 'fe47 Moušīryata "hence the designation arrow, because it was broken m2, 2, 2 ta u hy etad ubhayyah samprcyante, but in the same & the verb samgacchante is used (“mingle, unite'), the samhitā-text quoted (Vāj. S. 1, 21) has prcyantām; 3, 3, 4 ya proksanyah parisisyante "the sprinkling-water which is left”; 9, I, 2, 18 atiricyate "it remains (over)”; 2, 1, 3, I ya eväpūryate 'rdhamāsah ... yo 'pakşīyate "the ·TEXTS 23 ... half-moon which increases that which decreases"; 4, 4, 15; 18; 19; 1, 2, 2, 5 evam u haisa ātma yajñasya samdhiyate "and thus this body of the sacrifice forms itself into a unit”. The action expressed by the verb may also refer to persons (gods, demons and other divine or non-human beings of this kind included); the cases are the following: 2, 6, 2, 12 yo mrtjor inucyātai "who shall be detached (loose, severed) from death”, in the text quoted from Vāj. S. (3, 60 a) in the same § we find the middle aorist muksiya without difference of meaning; 5, 2,25 tasnāt sarvasmāt pranucyāmahe "from all that we rid ourselves"; 1, 2, 4, 10 nucyante (2 X) "escape”; 2, 3, I, 6 sa yathāhis tvaco nirmucyeta, evam rätreh nirmuc yate, ... sarvasmāt pāpmano nirmu- Cyate yo "even as a snake frees itself from its skin, etc.” (also 5, 2, 47); ibid. vi hi srivante yathārthanı "they are free". The verb dhriyáte "maintains itself; turn the mind on, be determined to" is ‘passive' by its form, although it is not used in a ‘passive' sense 2, 2, 3, 20 yadi pavamānāya dhriyeran "if they determine upon (offering to) Agni Pavamāna”; 23. The verb vyrdhyate may be dealt with under this head: 3, 1, 7 tad va asmai tad vyrdhyate "he fails in it”. Of these cases 14 refer to a non-personal 'patiens' ('subject of the passive sentence') 8 ( + 2 X dhr-) to a personal ‘patiens', three times the 'patiens' is an animal. B. Facts or actions which suppose the activity of an actor or which, at least, usually appear only resultant from the preceding activity of an actor, may be stated in such a way that only their coming to pass, not the actors, the beings which effect the process, stands out. Whereas the examples quoted under A are used without any reference to an 'agens', in the following cases, which are comparatively speaking numerous, it is possible (perhaps I should rather say: it may, on occasion, be possible) to mention the actor who has (or who had) a hand in the matter. Thus the form kriyáte occurs very often to denote "take place, come to pass, is brought about, is effected, is done", e.g. I, I, I, 18 "he (i.e. the Adhvaryu priest) pours out some water into the jug and puts it down north of the Gārhapatya fire. For water is female and fire is male, and the Gārhapatya is a house: hence a copulation productive of offspring takes place thereby in this house”: tad grhesv evaitan mithunam prajananam kriyate. It is not the Adhvaryu priest who effects the copulation, but the two elements, the female water and the male fire; the act, however, becomes only possible by the preceding actions of the priest. Other examples are: 3, I, 18 the woman looks down upon the sacrificial butter, which represents seed: hence "a pro- ·24 REMARKS ON THE SANSKRIT PASSIVE ductive union is thereby brought about”; 1, 2, I, 1; 2; 2, 5, 15; 9, 2, 6; etc. ; 2, 2, 5 the Adhvaryu places the cake over the fire, the Agnidhra puts the clarified butter on: these two acts take place simultaneously: tad vă etad ubhayam saha kriyate; 3, 3, 18 abhyadhānāya hy evedhich krijate "for the firewood is prepared for the purpose of being put upon the fire" (this firewood had been brought by the Āgnīdhra-priest, but it has not been mentioned in the preceding part of the text); 3, I, 26 Yājñavalkya said: ... katham svayam nānvāhur jatra bhūyasya ivāśisah kriyante “and why do not they (the sacrificers) themselves recite when far higher blessings are prayed for?:? 46; 8, I, 36; 2, 5. 19 sarigrānio vai krūram samgrāre hi krūram kriyate "the bloody one is the battle, for in battle bloody deeds come to pass are done)":9, 2, 14 prastarasyaiva ripan k.; 2, 4, 4, 25; I, 9, 3, + jad vai yajižasya mithyā k. "where anything is done wrongly at the sacrifice”; moreover: 2, 4, 4, 20; 5, I, 14; 5, 2, 23; 6, 3, 9. There are some standing phrases: 1, 2, I, 18 he (viz, the Adhvaryu) pours the rice on the lower stone... ity u hi havir grhyate it is with the intention ... that the rice-oblation is taken" (Eggeling); h.g. also 22; 1, 4, 9; 8, 3, 24; 2, 4, 3, 9; ajyāni grhyante 1, 3, 2, 7; 16. 2, 2, 4, 6 "Prajapati was aware that it was his own (sta-) greatness that had spoken (āha) to him. He offered it up with 'stāhā! This is why offerings complete themselves with strāhā”: sa svāhet ver'ājuhot tasmod u svähetyeva huyate; huyate 'without a subject also 3, I, 36; 3, I, 23 ahutir huyate; 29. Another ‘etymological figure' is 2, 5, 2, 23 aśanar aśyate. I, 8, 2, 7; = 9, 2, 26 Jäbhya eşa (=y: devatābhva) rajñas tāvate "for whichever deities this sacrifice is performed” (Eggeling); 27; 2, 2, 4, 13. 1, 2, 3, 5 pasur ha vă eșa ālabhraie vat juroa āśaḥ "now it is as an animal sacrifice that the sacrificial cake is an ofiering" (this cake being a substitute for the animal sacrifice); 2, 4, 4, 22 retaḥ sicyate "a discharge of semen takes place”; 5, 2, 46 sāma giyate. "Without a subject': 2, I, 2, 12 ya ichet pra me divetati "whosoever should wish that (presents) should be offered him" (Eggeling) ...yad dhastena pradījate pra haivāsmai dīvate "for whatever is a present with the hand, that indeed is a present to him"; compare such nominal sentences as 2, 4, 4, 21 tasnzad esātra payasyā bharati “therefore there is an offering of curds". vasat kriyáte: 2, 5, 2, 33; 8, 2, 15, where also devatāyai hūyate. Personal 'subjects' are seldom found. Now and then at the beginning of a certain part of the text the 46 J. Eggeling, The Satapatha-Brāhmaṇa translated I, p. 77 (and n. 2).

·TEXTS 25 active verb and the noun which denotes the actor are found, but further on the -ya-form is used and the actor remains in the background. Although the -ya-form denotes an action effected by a person who is known to the reader, the author does not take an interest in this person, but intends to throw light upon the process which comes to pass. Thus, I, I, I, 12 ff., while discussing the preparation of the offerings, the author informs us several times that the Adhvaryu 'brings forward the water': 12 apaḥ pranayati ; ibid. ; 13; 14; 16; 17; 18; then, 20, he says: tā uttarenāhavanījam pranayati "he then carries it north of the Ā.-fire”; in doing so, (21) the Adhvaryu is forbidden to carry it beyond the north side of the fire, or to put it down before reaching this side; for, if he were to put it down before gaining that side, he would not obtain by it the fulfilment of the wish yasmai kāmāya pranīyante "for which the bringing forward has taken place". 1, 8, 3, I ff. the separation of the two spoons (juhū and upabhrt) is dealt with: sa vai srucau vyūhati etc.; the reason why the priest separates them is set forth in $ 5 ... "behind the j. stands the eater, behind the u. the one to be eaten: thus he now brings the eater to the east (to the front) and the one to be eaten he drives toward the west (he drives back)"; then, § 6, it runs: tad vā etat samāna eva karman vyātikriyate "thus the separation comes to pass in one and the same act", here stress is laid on the words samāna eva karman which supply the starting-point for the next link of the argument: "hence from one and the same man spring both the 'eater' (enjoyer, husband), and the one to be enjoyed (the wife)", and it is only the separation which has come to pass in one and the same act in which the author feels interested, in § 6 the person who has brought it about is quite immaterial. Cp. also I, 1, 2, 18-19 havir grahīsyantam: ghyante; 8, 1, 38-39 nägnau juhvati: prāneşv eva hūyate ; 8, 2, 10. The 'subjects' are non-personal. C. In a restricted number of cases an instrumental attends a -ya-form to denote the accompanying circumstances or the cause by which the action happens to be. After having taught (2, I, 4, 5) that three sticks of aśvattha wood have to be put on the fire with (= while muttering) a certain set of rg-verses (...rgbhir abhyādadhati), the author deals ($ 6) with the fire unto which an oblation is made with a rg-verse etc.: yasminn agnov ?cā ... āhutim va juhuyāt, and $ 10 he quotes a question in which the same instrumental attends on a -yaform: yan narcă ... agnir ādhiyate kenādhīyate "if the fire is not set up with a rg-verse..., wherewith then is it set up" (Eggeling); in the GONDA, The Sanskrit Passive 3 ·26 REMARKS ON THE SANSKRIT PASSIVE answer the author construes brahmaṇādhīyate "with the brahman it is set up" (E.) and satyenaivādhīyate; cp. also $8 11-13; 3, I, 30 (2 X); 5, 2, 46. 2, 6, 1, 3 yad u caivāsyātrātmano 'caranena hanyate vā mīvate vā tad u caivāsyaitena punar āpyāyate "and whatever perishes or becomes lost to him (whatever injury or loss he suffers) through his own unrighteous conduct that becomes thereby complete (uninjured etc.) to him”. It is only once that I have found an instance of an instrumental denoting the persons who exerted influence to bring about a certain result, and in this case it still seems more correct to use the term instrumental of means or influencing circumstances than instrumental of agens: 1, 8, 1, 34 gas ja hi prajā bhavaty eka ātıranu bhavaty athota daśadhā prajayā havis kriyate tasmāt prajā bhiyo haviskaranam "for who possesses offspring, though he, of his own self, be one only, yet that offering comes to pass (is made) tenfold by influence of (by) is offspring: hence offspring means more abundant offering" 47. Moreover, I have to deal with two texts: 1, 9, 2, 8 tira iva vai nzithunena caryate "secretly, doubtless, union takes place": cp. 5, 2, 15 yatra haivam įtvijaḥ samvidānā vajñena caranti "where the officiating priests thus with a perfect mutual understanding deal with the sacrifice", 9, 2, 8 the construction is impersonal. 2, I, 2, 4 tā mithunena t'yārdhyanta "they were deprived of intercourse”: instrumental attending on words expressive of separation and disjunction. So, in all cases, a thing, not a living person, is 'subject of the -ya-form. D. The idiom tapas tapyate (a 'passive' with etymological object 48) is met with 2, 2, 4, 1; 5, I, I. That many a '-ya-form-sentence' can be translated by an active phrase is, of course, of no moment, cf. 2, 5, 2, 47. There are several nominal constructions admitting of a 'passive' translation, e.g. 2, 3, 1, 36 nāsya sä paricaksā "here no offence is committed". I have found only two imperatives of the -ya-form, both of which are in a quotation: 1, 2, 1, 13 (Vāj. S. I, 18) tapyadhvam; 2, 2, 2 (Vāj. S. I, 21) prcyantām. A small number of participles of the -yá-form are met with, of which one forms part of an absolute construction; most of them are used as attributes or 'appositions' expressing attendant circumstances: 2, 5, 2, 19 in a formula agnaye samidhyamānāyānubrūhi "recite to the fire that is being kindled", I, I, I, 20 nen mithunom caryamānam 47 Cf. Delbrück, Altindische Syntax, § 91. 43 See Speyer, Sanskrit Syntax, $ 319; R. 2.

·TEXTS 27 artarena saincaran iti the copulation which is taking place”; 3, 2, I esa c'ai tayamano yātān eta puruşas tāvun vidhiyate "in being spread 1: (the sacrifice is made of exactly the same extent as the man"; 6; 2, 2, +. 17 we find twice hūyamānas ya "the offered (milk)" without a substantive; 1, 2, 5, 20 sa c'ū upary-u pary eta prokşarīşu dhärjamaniisu atha sphyanı udyachati "while the sprinkling-water is being held close above the wooden sword...". The subject' is always a thing, IVOTHER FORJIS AND JAIMINTYABRĀHIAXA There are only a few middle forms which ought to be mentioned: I, 8. I, 6 "The flood swept away all these creatures" 1110nur e't'aikan parisišişe "JI. alone remained” (pure intrans.): 2, 4, 4, 20 (the moon) paścūd dad;ść "is visible" (id.) ; 5, 1, 13 “if (the Maruts) destroy these (creatures), then nothing will be left" na tata? kim cana parišekşyate. The aorist ending in -i is rare: I, 1, 1, 3 in a quotation from Vāj. S. 2, 28a tan me'rādhiti “that has succeeded for me” (intr.) and taken over into the text; I, 4, 8 esa hi vajīža ulikhale pratyasthăd cșa hi prāsāri "the sacrifice has now obtained a firm footing..., it has become diffused" (Eggeling); prāsāri, however, is rather, like the intransitive active forms, "has begun” (“in Gang kommen" Pet. Wtb.). In the Jaiminīya-brāhmaṇa the -vá-forms are rather often found: in Caland's Auswahl there are about 150 instances. Most of them are without any doubt intrans., at times we may render the Sanskrit form by a reflexive construction. Other -J'á-forms which occur beside the active forms derived from the same root may be considered as “agentless passives' etc., but it is very difficult, nay impossible, to draw a hard and fast line between these groups; as a rule the meaning of the rerb, which is an insufficient criterion, turns the scale. Among the - 1a-forms there are a large number of indicatives, imperf., however, are not rare. The following are intrans.: 17 49 acchidyata "brach"; 127 avacchidyante + abl. "sich trennen von”; 168 (p. 221, 1. 78) azacchidyāmahai; 200 tad vyacchidyata "dadurch trat eine Unterbrechung ein"; 15 asrjyanta "sie machten den Start"; 17; 203 "kamen hervor" (2 X); 207 (passim); 23 parišişyeta "übrig bleiben"; 116; 144 samseşaḥ samasispata; 88 ati-ric- "übrig bleiben"; 147b atiricyante "hinausreichen, überschüssig sein”; ibid. ; 136 (p. 159) hiyate "bleibt zurück"; ibid. ; 168 nā hīyadhvam; 177; 140 nirnucyeta “sich befreit"; ibid. ; 150, 5 pacyate; 180 "wurde gar"; 164 pra ſīryate "get torn, 49 = Caland, Das Jaiminiya-brāhmaṇa in Auswahl (1919), nº 17.

·28 REMARKS ON THE SANSKRIT PASSIVE burst”; 176 paryadyśyata "wurde sichtbar"; ibid. ; 184 parisușyate "shrinks in"; 116 avakīryate "forsakes”; 124 ksīyate "vergehen” 50; 70 sam asmai sa kama ?dhyate "das geht ihm in Erfüllung"; 78; 177; 180; 186; 192; 208 prāvlīyata "er knickte zusammen"; 179 paropyate "geht verloren"; 168 (p. 221) mīyate "abnehmt (an Lebenskraft)"; 60; 45 ākhyāyate "heisst"; 47; 54 etc.; 47 vyavrhyanta ( 2X) “got torn, burst" ; 212 nirabhidyata; 169 procyate ; 90 adriyeta 51 “heed" ; 139. An 'agentless pass.', Fr. on etc.: 23 pradīyate "wird hingegeben", "on donne"; ibid. opt.; 149; 53 suyate; 59 adhigamyate (2 X) "wird gefunden”; 102 dakşıìā nīyante; 107 pratidlīyate "man legt"; 136 anc yate "wird überliefert”; 199; 147b kriyate "findet statt"; 165; 168; 168 (p. 220) jīyate, hīyate; 167 upadhīyate; 99 āurścyate "er wird abgetrennt"; 101 yujyate "wird verwendet"; 253 siro 'cchidyata (2X) es wurde abgetrennt"; 129 retaḥ sicyate; 164; 182 asrjyanta (2 X) “kamen ins Dasein", oder "wurden entlassen, wurde erschaffen" (Caland); 30; 44; 104; 106; 164 (stanza) upadhīyate. (The agens remains in the background:) 103 vrstih pradiyate; 200. (The agens is known, but not expressed:) 9 prajāpatih prajā asījata ... agnih asrjyata ; cf. also 63 anušasyate; 128 stūyate etc. (6 X); 171 kriyate "wird angewandt”; 200; 206; 171 ājyata. At times the -ya-form is attended by an accus., which, then, depends on the praeverbium: 5 punarmstyū atimuc yate; 106 rathantaram te 'nvasrjyanta (ibid, tad anu ’srjyanta); 99; 182; 206; ādriyate + acc. 161. The turn tapo 'tapyata is found 143 52, -ya-form attended by an instr. separ.: 13 rupena vyrdhyante "sie verlieren die Schönheit" (2 X); + instr. relat.: 146 vācā samrdhyante "sie gedeihen inbezug auf die Rede"; ibid. ; + instr. instr. II tenāsmál lokān nāvacchid yate ; 144; 196 prānair u gajſas tāyate; ibid. ; 23 pavaráncna ... annādyam pradīyate; 60 etena punar āhriyate yat ...; 62 tenaiva samdhiyate. There are no examples of an instr. agentis; 107 Caland has proposed: tad vad evamvidā vajñāyajñīyam antatah kriyate "wenn also das Y. von einem solches Wissenden am Ende gemacht wird", the mss. have evamvidhim; perhaps evamvidhamn? The part. praes. pass. is an attrib. or appos.: 107 vaster bhidyamānasya; 164 onutsriyamanah; 168 (p. 222) ; 171; 176; 185 āditsyamānam; 207 (3X); + man-: 102 hiyamanāmanyata (2 X); adriyamāna- + acc. 47; as a subst. 116; loc. abs. 104 (2 X); 204. With an instr. ag. : 180 yathā rajñā brāhmano 'numantryamāna ūgacched evam. 4. 50 See W. D. Whitney. A Sanskrit Grammar, $ 761. 51 See Whitney, 0.Eng. $ 773. 52 See Panini 3, 1, 88.

·TEXTS 29 Other personal forms are rare: fut.: vicchetsyante (e conj.) “werden zunichte gehen"; an aor. 200 (they said) tyacchedi; perf.: 191 parisisise; dadhre 53 - inf.: 203 nartitum dadhre. The intrans. -'á-forms are used in the same way as the intrans. forms, which belong to the fourth class; cf., e.g., 76 azapad vate "geht verlustig"'; 86 (stanza) rathad avapadyate "stürzt hinab"; 146 0714hyanti "gehen irre” etc. Aitareva-Aranyaka In the Aitareya-Aranyaka, of which I examined 1, 1-3, I (p. 75- 132 K.), the -tá- and - ve-forms are the normal predicates of subjectless (agentless) sentences. Cp., e.g., I, I, 2 nyüne vai retah sicyate ... nyüne 'nnad vam pratisthitam. The ppp. is used as a predicate: 1, 1, 2 pratisthitam, + vad vai niskrtam (quoted) tat samskrtam; 2, 4 sarlmitālı ; 3, 2 sritāḥ (ibid. sarve häsmin kāınāļi śrayante, see also 2, I, +); 4 (after a quotation gato jajña uigrah...) ato hy esa jāta ugrah *for from it he was born who is terrible"; 3, 8 vistāh ; 2, 1, 1 nivistah (after ni ... viviśre in a quotation); āvistah; 2, 4 tad vā idam brhatīsahasram sampannam "this is produced as..."; 2, 3, 6 mitam; 8 agatam; 2, 5 samtatáh; 3, 1, 2 āyattah. In the case of several verbs the same form often recurs. An instr. instrumenti etc. is added: 2, I, 6 nāmabhir dāmabhih sarvanı sitam (cf. ibid. tantisainbaddhāh); chandobhis channah; pranenávytam ;pranena vistabdha-; 7 vācā srstar prthivi. An instr. 'agentis' is met with in the frequent turn 2, I, 4 etc. tad ... rsinoktanı a rşi says, there is an utterance of an şși”; we have to remember that a rşi is not the author of mantras, but only the ‘mouth-piece'. At times such an instr. may, in a sense, be conceived as an instr. ag.: 2, 3, 5 ātmā sarvatah sariraih pariurtah "the self is surrounded on all sides by members". See also 2, 3, 6 attam vai bhūvasā kanīyah. An interesting passage is 2, 4, 3 yadi vācābhivyährtam jadi prāạenäbhiprānitam etc. (8 X) atha ko 'ham "if speech pronounces words" (Keith), but a more accurate rendering is: "if pronouncing takes place by means of) speech..."; it is open to question who is the real author. So these predicative ppp. denote a result, a condition, a state; their adjectival character is obvious, cf. also 2, 3, 6 where riktam and pürnam "empty" and "full" are used by the side of ricyate. There are neither instances of ppp. denoting actions nor of real instr. ag. added to them. Their adjectival character is also apparent from the 53 Cf. Whitney, O.C., § 773.

·30 REMARKS ON THE SANSKRIT PASSIVE considerable number of negative forms (1, 2, 2 aniruktah "undefined"; 5, 2, anarusaktā; 2, 3, 6 amitar, etc.) and compounds (I, I, 3 rupasamrddham; 2, 4, 2 sukytanı), from such turns as 2, I, 5 ... ity etat tad uktam bhavati "this is the meaning","thus it is said", cf. also 3. I, 3 instr. ubhayam vyāptam bhavati (instr. instr.), and 1, 2, 3 madhyata udbhyte syātāızı "let them be raised in the middle", 3, 4 with an instr., from the neutral use (e.g. I, I, I ekain iti tv eva sthitam "but the established rule is one"). Gerundives. The turn tat tan nädri jam occurs several times (e.g. I, 2, 3) 54. The personal forms in -ya- are, in the complete text, rather numerous. The -já- present is either intrans.: 1, 2, 3 dyśyete "seem", 2, 3, 2 "is visible"; 3, 2, 4 "appear, seem to be"; 1, 3, 5 prasriyante "spring from"; 2, I, 4 tad aśīryata "it was decayed" (aśārīty “(men said) it had decayed" is the actual past: aor.); I, 7 jāryate “decay"; 2, 4, I sirabhidyata "burst forth": perhaps I, 4, I pratyacyante 55; 2, 3, 3 adhigamyante; 6 ricyate; 1, 2, 4 ucchidyai“to lose his hold". The agens is quite immaterial or 'general': 1, 1, 2 retał sicyate "the seed is poured out”; 4 kriyate "is done, takes place”; 5, I apyante; samiskriyante ; 2 apyate “is gained"; ibid. väg udgate; 1, 3, 7; 2, 3, i samop yate ; 5, I pratidhīvate "is appointed". The subject may be expressed by "people, man", Fr. "on" I, 2, 4 kriyate; (there is no subject) 4, 3 ripar krivate "the form is given"; 5, 3; 2, 3, 4 anusas yante ; 4, 2 havir grhgate. The subject has been mentioned or may be inferred from the context, but is left unexpressed in the sentence 1, 4, 8 apjate “is pervaded"; 2, I, 5 prātāyata. The -J'a- form is attended by an inst, instr.: 1, 3, 3 vācă hi nămadhe yāni dhīyanto "for by speech names are made" (= 6). An instr. ag. is found 2, 7, 7 sa (= Indra, the subject of the preceding sentences) tena bandhună ... hūyate. Cf. also 3, I, I samdhīyate prajayā “he gets children"; 2. A part. pr. pass. occurs 3, 2, 3 (sain-dhā-) ; 2, 3, 5 asyamānah and vyaj yamānā + instr. instr. (appos.). Other personal forms are rare: 2, I, 4 aśāri; 5 prātāyi "he is stretched out"; 2, 2, padi "he went". These aorists are intrans.; pādi appear in the narrative 56, where the style is vivid. Perfect: 2, I, 3 dadyśe ('agentiess pass.') "he is seen"; 8. 54 See Keith, The Aitareya Ary, p. 176 n. 2. 55 Keith, ibid., p. 189, 11. 11. 56 See Keith, p. 213, n. 9.

·TEXTS 31 .. ... Bradāran yako panisad. It is not possible to understand the use of the passive verbal forms thoroughly without studying the active forms at the same time. It is not till one has acquired a sufficient insight into the use an author nakes of the latter, that the peculiarities of the passive stand out. For this reason studied all the verbal forms in a number of texts. In the narrative parts of BArlp., 3d adhy., there are 93 forms of the peri. act. or med.: uvāca introducing an or. dir. occurs 58 X, papraccha 9 X, ipararāma at the end of an episode 8 X, babhüva i 1101 preceded by a verb. adj.) 2 X; moreover, there are 16 other forms. The perfect represents the narrative past: 3, I, 2 tán hovāca ... to ha brāhmaṇā na dadlırşuh tă hodācakāra. te ha brāhmaṇāś cukrudhuh etc. The indic. pr. occurs only once. There are two 'passive' forins: 3, I, I tatra hi ... brāhmaṇā abhisametā babhūvuh and padū ūbaddha babluruh, semi-nominal sentences, expressing a condition in the past. In the dialogues (or. dir.), however, we find the following forms: indic. pr. act. or med. : 121, nom. sent. + as- (not preceded by a v. adj.): 4, imperf.: 25, imper.: 23, conj.: 1, opt.: 23, fut.: 10, aor.: 6, inj. aor. 3, ind. perf.: 53, opt. perf.: 8. But of the 25 forms of the imperf. 14 are found in 3, 3, 1 and 2 and 10 in 3, 7, I; both times one of the seekers for wisdom who takes part in the discussion tells of an event which has happened formerly and which he has witnessed himself: madreșu carakah paryavrajāma, te ... grhān aima etc.; moreover, 3, 9, 9 yad asminn idamı sarvam adhyārdhnot, tenādhyardha iti (etymological explanation). The aorist expresses the actual past: 3,7, 2 puruşam pretam āhur Vyasyarsisatāsyāngānīti; 8, 2 tvä ... i podasthām "je me dresse devant toi" (Senart); 5; 9, 18; 19; 26. Of the perfect forms 40 are veda and vettha, 11 äha and āttha; moreover, 3, 3, 2 uvāca introducing a quotation and ibid. praśasanısa. There are 9 indic. pr. pass. and I optat., which, in general, are intrans. : 3, I, 5 kona ... atimucyate "par quoi échappe-t-il"; 2, 11 atraiva samavanīyante "ils se rassemblent"; 13 nidhiyate "se dépose"; 8, 12 mucyedhvam "acquit yourselves" ; 9, I ucyante "it says" ; 26 na hi grhyate "for it is not to be laid hold of", in the same siryate, sajyate, and vyathate, risyati, which are not passives; besides, 8, 10 tapas tapyate. It must be noted that 8, 8 where a passive verb might have been used, the author prefers two active constructions, which form a parallelism: na tad aśnāti kim cana, na tad aśnāti kaś cana "il ne mange ni ne se mange" (Senart). The v. adj. without as-or bhu- occurs 75 times, but the number of repetitions is extremely high: 3, 6, I the expression ota- ca prota- ·32 REMARKS ON THE SANSKRIT PASSIVE ca is found 12 times, cf. 9, 26 etc.; this form being the predicate of a nominal sentence, expresses a condition, a quality etc. ; 3, 9, 26 kasının 11u prānaḥ pratisthitaḥ may be translated by: "where is the location of the p.?"; 7, 23 ato 'rzyad ārtan "whatever is different from it is suffering", here the v. adj. is almost an adject. The verbs as- or bhi- are added 9 times: 3, 7, 1 etc. yasminn ayam ca lokah paraś ca lokah... samidybdhāni bhavanti, semi-nominal sentence. 3, 9, 26 pratisthitau sthah the verb as- is necessary to indicate the gramm. person; 20 only the last of a number of parallel answers has bhavanti: kasinin 31u rūpāni pratisthitānīti :: hrdaye hydayena hi rūpāni jānāti, hrdaye hy eta rupāni pratişthitāni bhavanti, which forms a rather ponderous concluding statement; cf. also 21; 22; 23. With stha- (3 X) a durative is formed: 8, 9 süryacandramasau vidhrtau tişthatah "ont leur existence propre” (Senart). 3, I tasyāsīd duhitā gandharvagrhītā, t. ... āsit means "he had” and gandh. is an apposition; cf. 2, II. 2, 2 sqq. an instr. determines the v. adj. sa gandhenātigrahena grhīta!ı. The gerundive preceded by a- occurs three times: 9, 26. In the prose narrative of another section of the same Upanişad (5) I, 1-6, 2, 15) there are 112 forms of the perf. act. or med. (35 vac-, II ah-, 35 vid-): narrative past, 75 of the indic. pr., 13 of the opt. pr., 2 conj., 5 semi-nom. sentences with bhu-, 1 X bh- + part. fut., I imperf. (5, 5, I asrjanta). The indic. pr. pass, which occurs 5 times, is an intransitivum: 5,9, I (annam) yad idan adyate "which serves as food, which is eaten"; II, I etad vai paranam tapo jad vyāhitas tapyate "the greatest pain surely is that, which one endures from disease"; 13, 2 yujyante (2 X), cf. 3 where, in parallel sentences, sainyañci "turned against each other" is used; 9, I an instr. ag. is added: yenedan annam pacyate. A perf. 'pass.' is met with 5, 14, 3 darśatam padam iti, dadyśa iva hy eṣaḥ “... parce qu'on y voit ..." (Senart). The sentences wherein the v. adj. is predicate (8) denote a condition etc.: 5, 2, 2, (2 X); 14, 4(5 X); 6, 1, 4; cf. 14. Gerundives : 5, I, I; 14,6 (+ ag. instr.). -- In the or. dir. there are 40 indic. pr., I conj. pr., 4 opt. pr., 13 imper., 2 semi-nom. sent. ; 2 imperf. (hist. past; 5, 14, 8; 6, 2, 5), 26 aor. (act. past; 5, 12 vyajñāsista "vous avez compris?"; 4 inj. aor.; 13 perf. (9 veda, vettha; 3 ättha) ; 3 fut. The ind. pr. pass., being an intransitivum, occurs: 6, 2, 2 na sampüryate ne se remplisse pas"; 7 vijñāyate "it is known"; aor. + mā 5, 14, 7. The v. adj. is predicate 5, 12, 1 (adj., in nom. sent.); 6, 2, 2 api hi na rşer vacah śrutam; ...iti "we have heard” (almost the act. past); 5 (varam dadmah ::) sa hovāca: pratijñāto ma eșa varah, ·TEXTS 33 which in an emotional and vivid manner answers to our “I accept the boon which you have promised me" ("fait-accompli-Darstellung'); 6, 2, I and instr. ag. is added and the verb. adj. almost expresses the act. past: anuśisto nv asi pitrā "are you (an) instructed (one) by (the influence of) your father?". In the stanza 5, I, I two intr. -ya-forms are found udac yate and azaśisyate. In the other parts of this text the so-called passive forms have the same functions. They form some clearly defined categories. The -yaforms are intrans. : 1, 2, I sarnahanyata “se solidifia"; 5 adhriyata "turned the mind to"; 3, 12 sqq. atyarıcyata; 17 annam adyate "se mange" (cf. 5, 2), ibid. anenaiva tad adyate "c'est grâce à lui qu'elle se mange" (cf. § 18 yad anenānnam atti); 27 giyate; 4, 7 tan nainarūpābhyān eva vyakriyate “that determined itself by nameand-form" (instr. of means) ; 15 ksīyate “s'épuise"; 5, 14 ā ca pūryate, apa ca kşīyate "he becomes full and wanes" ; 2, 1, 3; I, 5, 15 sarvajyanim jīyate "he suffers the loss of all"; 2, I, II näsmäd ganaś chidyate “his followers never part from him"; 4, I, 6 manasā striyam abhihāryate “by the mind one feels attracted towards a woman"; 3, 5 svaḥ pāṇir va vinirjižāyate ; cf. I, 2 vācā ... bandhuh prajſtāyate; 3, 8 pāpmabhiḥ samsýjjate “... se charge de maux" (Senart); 19 samlayayaiva dhriyate "be determined to go to" or "se porte vers" ; 23 etc. vidyate; 35 pramucyate + abl. "se détache de"; 4, 23 (stanza) na lipyate karmanā "he does not stain himself by ..." (cf. 1. 2 na karmanā vardhate "he is not increased by ..."). As regards the v. adj. I wish to point out the following instances: 1, 2, I mrtyunaivedam avrtan āsīt, here the patiens finds itself in the condition of being enveloped (like our wrapped in darkness etc.) (cf. 6, 3 amytam satyena channam); 3, 23; 2, 4, 5; 7 sabdo grhītah "on se saisit du son" (Senart); 4; 3, 33 sa yo manusyānāni rāddhah samrddho bhavati (semi-nom.) ; 1, 4, 7 avyakstam āsīt; 5, 17 yat ... anūktam "se récite”; 4, 2, 4 abhayam prapto 'si and 4, 23 enain prăpito 'si express the actual past. Instances of the part. pr. pass. are I, 4, Io ekasminn eva paśāv ādiyamāne (the ag. remains in the background); 5, I (stanza) adyamānāni ; 2, 4, 7 dundubher hanyamānasya; 4, 5, 8 (id.). Perf. (intr.) I, 5, 22 dadhre; 3, I, 2 prastum dadhre. I add some instances taken from the Chandogya - Upanişa d. Indic. pr. in -ya-: I, I, 6 samsrjyate; 6, 6 etc. drśyate ; 10, 2 vidyante "il y a"; 2, 10, 3 aksaram atiśisyate "une syllabe en surnombre" (Senart); 4, II, 2 ksīyante ; 16, 3 hiyate "perishes, is lost"; 5, 3, 3 sampūryate ; 24, 3 păpmānah dayante "the sins burn to ashes" ; 6, 5, 1 ·34 REMARKS OY THE SANSKRIT PASSIVE tredhā vidhiyate "se répartit en trois”; 16, I sa dahyate "il se brûle" ; 7, 10, I vyādhīyante prānāḥ "all animals would become wretched” (“tout ce qui vit s'inquiète” Senart); 8, 1, 4 atiśisyate "remains". An instr. ag. is added: 8, 1, 5 where a parallel sentence containing an intrans. verb precedes: nāsya jarayaitaj jīryati, na vadhenāsya hanyate, cf. 10, I; 6, 16, I; 10, 10 pāpmanā lipyate. 1, 6, 2 sāma giyate means "the s. is sung", fr. "on chant le s.”. As regards the v. adj. I note 1, IO, 3 ucchista}l vai rile pītail syāt, this sentence, forming part of a dialogue consisting of short sentences, is an ellipse : “(were I to take that) I would have drunk the remnant of another's drink”: emotional fait-accompli-Darstellung; 6, 12, 1 bhinddhi :: bhinnan the fait acc. is a reality ; 7, 1, 3 śrutan hy eza mne + abl. (actual past). The story of Pauşya (Mbh. 1, 3), the stanzas excepted 57. I- The verbal adjectives are very numerous. The pure attributives are rather rare (9 instances, including some beginning with a-; 58, 18 prabhūtataran). The predicative use is very common. When the v. adj. is used as predicate in a nominal sentence, it is transitive 26 times and intransitive 14 times. Nearly always the verb forms part of a dialogue, the exceptions are 56, 14 sa tayā kruddhayā tatroktah (in the alternative construction: 63, 26 tam upadhyāyaḥ pratyuvāca, which is very usual, the subject is always a subst, or a proper noun, and this is never attended by an apposition like kruddha-); 59, 29 sa Ostrībhiḥ uktah (narrative past); 63, 14 tābhir nāgaloka upadhupitah (here other I. Eur. languages prefer the passive constr.: Dutch „daardoor werd de ... uitgerookt"); 62, 7 tasya ... asannah (dramatically, fait accompli). The v. adj. is in most cases very clearly expressive of actuality: 56, 15 ... kimartham abhihatah; see also 59, 4; 5; 60, 22; 64, I, where ari historical event serves as an argument to persuade a person : āvābhyām purastād ... apūpo dattah (viz. to your guru); upayuktah sa tena ... (dramatically). Now and then, however, it expresses a state or a temporary quality: 58, 25 sa niyatam kupitah; 60, 31 na hi ... ksatriyā samnihitā; 60, 32 niyatan bhavāri ucchistah; an 'active construction is not possible here. In most cases (5: I) the agens has been expressed, cp. e.g. 57, 24 yasmāc ca tvayā madvacanam anusthitam tasmäd (a wordy and dignified turn). Its usual place is either at the very beginning of the sentence: 57, 4 mayāyam vrta upadhyāyaḥ; 15; 63, 27; 32; 33 etc., or preceded only by an anaphorical word; 63, 28 tatra ca mayā drste ... ; 30, cp. also 64, I 57 Quoted from Boehtlingk's Chrestomathie 3, p. 56 ff.

·TEXTS 35 .. tatas tasya vacanān majā u payuktan; at times another word is found at the beginning 63, 31 puruşaś capi maya distah (P. is brought into prominence); the agens is put after the v. adj. 59, 2 (inversion); 5 tr'an api tathaiza kuruşta yathā krtain itpūdhyayena (21. has been emphasized). An instr. expressing the cause is met with 60, 2. There are 6 instances of a ist pers. sg., 3 of 2d pers., I 28 of a 3d pers. and some 'impersonal constructions' such as 60, 30 na yuktanı + inf.; 61, 20; neutral constr. 61, 33 prāk te 'bhihitam; 63, 27; 59, 5. The personal saktas – inf.: 61, II; 27; 31. — There are 18 instances of a 1. adj. being used with as- or bhii-, mostly transitives. They always forin part of a dialogue and are either expressive of a state or quality: 58, 7 lubdho 'si; 59, 7 pritau svah ; 61, II sunirurtā bhava ; 61, 13; 63, II, or of actuality (II X). In 8 instances the agens has been expressed; it is placed at the beginning, 63, 25 it'ai inanūg asi inayā ila sapta), however, we find it elsewhere; an instr. expressing the cause is put at the beginning 64, 8, at the end 63, II. There are 8 instances of a ist pers. sg., 2 of a ist pers. du., 7 of a 2d pers. sg., I of a 3d pers. sg. In most (10) cases the copula is placed inimediately after the v. adj. 56, II bhrātrbhir abhihato 'smi; cf. also 17 sambhrūnto vişannaś cāsīt; the exceptions are: 56, 9 kenāsy abhihatah (interrogation); II yena a.d. (relative phrase); 60, 13 upadhyāyenāsny anujñāto grham gantum (v. adj. + inf. seem to form a group); 63, 33 ienāsmi sopacāram ukta! (adverb determines the v. adj.); 28 tenāsmi nūgalokam gatah (accus. is dependent on the v. adj.; 61, 2 asti "there is", "il y a" is placed at the beginning (according to the rule); 63, 25 tvam manäg asi mayā na saptaḥ. Neutral constr. 61, 1; 3. The formula svāgatarı te occurs thrice. The v. adj. which expresses attendant circumstances etc. of the main action is mostly transitive (+ 3 : I). In 15 cases (+ 30 %) the agens has been added: 57, 2 tenaivam ukto j-s tam pratyuvāca, once (61, 8) an instr. of cause is subjoined. The v. adj. has to be connected with the subject of the active sentence of which it forms part (56, 8 sa ... abhihato ...upāgacchat), it precedes the verbal form which serves as the principal verb. The v. adj. is used as an apposition to the subject: 61, 26 mamāpi tu śāpo na bhaved bhavatā dattah; to the object: 58, 12; 60, 16 ... kundale bhikṣitum tasya ksatriyaya pinaddhe; 61, 7 (in these instances the v. adj. follows the verb); 61, 22 there is no verb; the v. adj. precedes the verb: 64, 3 (in all these instances the agens is expressed); to an instrumental of cause 58, 22; to an instr. expressing the subject of a passive sentence:

·36 REMARKS ON THE SANSKRIT PASSIVE " 56, 14; 59, 4 (+ ag.) ; 61, I always preceding the verb); lastly 60, 20 ... apaśyad rsabhan tam adhirūdhain ca purusam. 4 X the v. adj. forms the ad part of a compound: 56, 26 (app. to subj., preceding the verb, 2 X); 63, 15. Four times a loc. abs., preceding the verb, and having a 'thing' as subject, is met with: 56, 19; 58, 23; 61, 31; 62, I. – V. adj. are rarely used as adverbs, I have found only niyatam (60, 32), drdhan (57, 31; 62, 7). - The v. adj. forms part of a compound in a restricted number of cases: first member of a bah.: 61, 4; 15; 22; 26, 5; 7 grhītamātraḥ; last member of a tatp.: 61, 16; 62, 14. II-A gerundive (12 cases) acts as an attributive 63, 2 puruşam capaśyad aśvam ca darśanīyam (emphasized), cp. 57, 20; as a predicate without a copula : 58, 26 (pers.) ; 61, 8 (P.) ; 57, 32 (thing); 59, 32 (2 X th.); id. with an instr. expressing the agens: 57, 5 (th.); constr. suśruşună ca bhavitavyaın 59, 15; na śakyā + inf. 60, 33; as a subst. : 59, 32 (2 X). III-There are only 3 personal forms of the indic. pres. in -ya-: intrans. 59, 26 parihiyate (th.); without an agens 61, 14 pătrail āsādyate; with an agens 61, 28 kim caitad bhavata na jñāyate vatha (= Dutch "is het u niet bekend?"). There are 9 (usually somewhat courteous) imperatives : ganyatām (3 X), usyatām, äsyatām, kriyatām (n. 59, 30; th. 61, 15; fig. et. karma k. 59, 22), yacyatām (p. 60, 29). 9 forms of the part. in -ya- are found, 5 are used as attributives or 'appositions', attending on a 'person': (preceding the verb) 57, II sa kliśyamāno 'paśyad upāyam; 62, 10 tam -anı apaśyat ... ; (following the verb) 59, 17; 62, 3 (2 X); id., attending on a 'thing' 63, I + ag., an animal, 63, 14. The acc. + part. constr. occurs twice: 59, 26 (th. + ag.); 61, 16 (th. + ag.). As regards the frequence of the active and middle verbal forms: in the narrative parts of this text there are 6 indic. pr., which are durative and "historical', besides, 2 indic. pr. t sma (dur. past); the imperf. (PPp. + imperf. excepted) occurs 46 times (41 act., 5 med.), the perfect 115 times (101 act., 14 med.) 58), it is often found at 58 The imperfects are: gam- (6 X), icch-, kr-, vas-, prati-pad-, pas- (13 %), yam- (2 X), grh-, pra-dru-, prcch- (2 X), raks- (5 %), dham-, cint-X, bhakş- X, vad- X (3 X), dr-X, ni-vid- X, labh- (med.), stha-, pra-sthā- (med., 2 X), bhāş-. The perfects are: vac- (24 X), prati-vac- (41 X); pra-vac-; bhū- (3 %), kr- (5 %), khan-, gam- (7 X), stu-, grah- (2 X), vis- (4 %), dis-, pat- (2 X), ah- (8 %), a-caks-, upa-sthd-, dha-, upa-kram- (2 x), vr- ("choose"), perf. periphr. 9.

·TEXTS 37 the end of a pericope, but also in other places. There are only 5 instances of an aor. (na) aśakat (2 X), avocat, anvagāt, abhüt; one -tavant- form. In the oratio recta there are 68 cases of an indic. pr., of which I 25 are durat.; the other cases indicative of the present time or timeless; besides, 7 indic. are modal; the fut. occurs 18 X, the optat. 9 X, the imper. 42 X; the perf. 2 X (historical) + 1 X ähuh (introd.). Sakuntala by Kalidasa, acts I and VI, prose. .. ... I–The verbal adjectives are very numerous. The attributive use is extremely rare: two forms ifc. Predicative use : trans. 93; intrans. 34. Generally, the v. adj. expresses actuality; there are, however, exceptions and special cases : 6, 11 +; 12 + paścād imām mudrām tadargulau niveśayatā mayā pratyabhihitā; ...; tac ca mayā mohän nāi$usthitani, a historical event is actualized; ibid. paribhrastam; the v. adj. expresses a state and an actuality at the same time: 6,11 + snamäpi kautühalenākārita esaḥ "he has been inspired with the same curiosity as mine" ; 14 +;I, 13 + api samnihito 'tra kulapatiḥ, + instr. of cause 6, 30 +; 6, 23 + pratihatan amangalam “may evil be averted", periphr. I, 21 +; 22 +; 28 +; (state or quality) I, 14 +; 6, 15 +; 6, 18 + cakitacakiteva sthitā; 1, 22 + neutral ucitam, etc.; the Pp. attended by parvain or pura is expressive of the narrative preterit: I, 22 + purā kila ... devair menaka ... preşitā; 6, 8 +. 10 trans. and 5 intrans. v. adj. are attended by the copula (12 X pers. subj.): indic. 6, 1 + anugļhito 'smi; 29 +; id. yadartham asmi harină bhavatsakāšam preșitaḥ; I, 17 +; 27 +; + instr. causae I, 30 +; 6, 27 + simile nirāso 'smi jivite samvittaḥ; 1, 22 + kim nu khalu ... ismi samurtta (emotional deliberation), etc.: indic. 9, opt. 3, impf. I, imp. I, fut. I; Ist p. sg. 8, ad p. sg. 1, 3d p. sg. 5, 2d p. pl. 1; 9 X the copula follows immediately; 3 X the ag. is expressed; besides, 1, 19 + enām vismytāsi (object). When there is no copula, the agens has been expressed † 50 X (all trans.); + 10 X an instr. rei (causae etc.) has been added; usually the ag. precedes the v. adj. (13 exceptions: 6, 22 + sarvathā pramārjitam tvayā pratyādeśaduḥkham šakuntalāyāḥ, exclamation; cp. 6, I + nirvartitam mayā; 7 + krtam bhavatā nirmakșikam; 12 + etc.). Trans.: 42 X pers. subj., 37 th., 14 neutral; intr. 23 p., 10 th., I n. The v. adj. is rather rarely used to express att circ. etc.: 6,8 +;

·38 REMARKS ON THE SA SKRIT PASSIVE . ; 19 +; ifc. 6, 6 +,8 +, usually preceding the verb.; appos. to the subj. + iva: 6, 14 +; the same but the v. adj. is put twice: 6, 18+; to obj.: 1, 16 7,6, II +; ifc. 8 + (2 X); to a gen.: 6, 4+. A loc. abs. occurs 6, 2 + (impers., preceding the verb.) ; 3 → devena pratisiddhe utsave ... kin ārabhase. A v. adj. acts as an adverb: 1, 17 + viśrabdham. Av. adj. forms the first member of a compound in a rather high number (+ 10) of cases, + 14 X it serves as second member (appos., attr., praed., subst.). II—The gerundive is used in various ways: as an attr.: 2 X e.g. I, 19 + (th.); attr. ifc. 6, I +; constr, kā tvam visrastavyasya roddhavyasya vā (I, 27 +) = visarjanasya rodhanasya ca (-tavya- bhāve); as a pred., without cop. and ag. : thing 5 X: 6, I +; 4 +19+; IO +; 16 +; the same, i.f.c. 4 X: anatikramanīyā 6, 31 +; 1, 21 + rājarakṣitavyāni; 22 + samavayoriparamanīyam; 6, 13 + ; pers. ifc. 6,1 +; animal 1, 9 + na hantavyaḥ; neutral: 4 X:1, 21 na bhelavyam; 6, 23 +; 26 +; the same, with ag.: th. I, 14 + vinītavesena pravestavyāni tapovanāni; neutr. 6, 4 t; bhavitavyam + instr. 4 +: + double instr. I +; 22 +; with a copula without an ag.: fut. pers. 6, 7 + cikitsitavyo bhavisyati; opt. neutr. II +; with asti "il y a" I, 23 + asti nah ... prastavyam; pred. ifc. 1, 7 +. III—There are 35 pers. forms in -ya-: 22 indic. pr., fut., 12 imper. Indic.: the ag. is added 4 X ; 5 X the subject is a person, 10 X a thing, 7 neutr. The form expresses an intrans. (5 X) e.g. I, 22 + prabhāvarān iva laksyate; is intrans, without a subj. (7 X) e.g. 6, I + yujyate; agentless pass, or fr. "on":6, 7 + ucvate "they say"; 22 + śruyate ; = we: I, 12 + drśyate "we see" or "you see"; 13 + akathito 'pi jñāyate; modesty: tad api kathyate "we will explain that too"; courteousness: 1, 22 + katanra äryena rījarşivaiso 'lamkriyate; emotion, the patiens is emphasized: 6, 25 + mamāpi sattvair abhibhūyante grhāh; the pat. is thema 6, 29 + j'ena ... SO 'nena abhinandyate. Fut.: 6, 22 + intr. Imper. (2 X P., 7 X th., 3 X n., I X animal) the ag. is expressed once: 6, 7 + äryena dīvatām; rather frequently neutr. śrūyatām etc.; this imper. is more or less courteous, but not infrequently a lower on the social ladder is addressed with it, e.g. 1, 7 + (king to charioteer) mucyantām... The part. praes. in -ya- occurs only: expressing att. circ. + instr. ag. 1, 22 +; as app. to the obj. + ag. I, 21 +; id. without ag. 9+ paśyainam y&pādyamānam. ...

·TEXTS 39 . Daśakuntāracarita by Dandis, ch. 8 59. 1-The verbal adjectives are numerous. Their use as attributives is, comparatively speaking, rather rare: 11 instances (259, 17 with agens), moreover 14 instances of a v. adj. acting as the 2d member of a comp. (ifc.) (259, 14 yuşmattaś cauryārjitam dhanam; 277, 3 tvadı: padistena nayena etc.). Used as pred. in a nom. sent.: 25 X trans., besides, 2 X ifc.: 270, II kumāraḥ śärdūlabhaksitah (exclamation); 272, 3; 13 X intrans. In £ 24 cases the v. adj. expresses actuality, 8 X a state or condition (260, 6 pārnāni + instr.; 8 kini idam a parjāptam...; 261, 6 spastaın; 271, 15 sa ... dārakatvenaiva kalpitā, etc.); $ 9 times the v. adj. expresses an historical event, a narrative preterit: 250, 3 1пaya paribhraniatā ... ko 'pi kumārah drstah ; 269, 4 idam tu jģiātva devyāhan ajñaptah ...; at times the v. adj., when employed in this manner, seems to express a sudden or unexpected event: 269, 20 ...ity acintayam. tävad apatitau ca... dvau inzgau ... avadhişanı ... apatat; 250, 3; elsewhere a tendency towards variation seems to be the only motive 60: 272, 9 abravit ... mayoktam ... abrūta; 269; 265, 13; in a sentence beginning with präg eva "(already) before now” raya krtam means “I had made”: 274, 9, here the v. adj. signifies the remote past 61. Sixteen times (= 40 %) the agens has been expressed; it may be found anywhere: seldom at the beginning of the sentence (or phrase) 259, 16; 269, 16 urddhenoktam; often in the second place, sometimes near the end, e.g. 269, II; at the end: 265, 13 (variation); 275, 13; the v. adj. is put at the end in most cases, 259, 17 at the beginning of the ad member of a disjunctive interrogation; 265, 13 (variation). There are 3 'instances of a ist pers., I of a 2d pers.; the V. adj. is neutral (uktanı) in 3 cases; a thing is “subj. of the pass. sent." in 21, a person in 16 cases, an animal once. na śaktal + inf. is found once. There are only 4 instances of a v. adj. + as- etc. (trans.) 252, 8 the v. adj. expresses the historical past sa ... nätyädyto 'bhūt (state of being, condition); the periph. form is a narrative preterit: 269, I; 271, 17 arthaś ca ... anuşthito 'bhūt; the v. adj. is attended by an imper.: 271, 16 atirahasyam yuşIIĀSV eva guptanı tisthatu. 2 X a pers., 2 X th, subject; agens 269, I niitraih ... bhratre ... devī darsitābhūt. The v. adj. expressing attendant circumstances etc. of the main 59 Ed. Godbole-Panşīkar 10, D. 250-278. 60 See A. Beth, Variatieverschijnselen in het Oud-Indisch, Diss. Utrecht 1943, p. 94 f. 61 Compare Speyer, Sanskrit Syntax, § 339.

·40 REMARKS ON THE SANSKRIT PASSIVE action (24 instances; 5 trans., 6 intrans., 13 ifc. ; 19 X a person subj., 3 X a thing, 2 X an animal) always precedes the princ. verb, 269, 6 (after an initial imper.) excepted. Only once an instr. ag. has been added 269, 2 nirbhartsitaś ca tayā, twice an other instr. (252, 8; 254, 7 tena hinah). The v. adj. acts as an apposition: to the subj. 258, 15 (+ instr. ag.), 253, I (+ gen. ag.); to the obj. 276, 1, mat prahitāni 261, 15, + instr. soc.; 255, 3 (+ instr. ag.); to an instr. ag. 268, II; 256, 18 to a gen. 257, 8; 269, 10 (comp.) (7 X a person, 3 X a thing). Acc. + part. construction : 272, 6; pred. app. 274, 14. Locat. abs.: 10 instances (4 P., 3 th., 3 neutral, of which 2 + instr. ag., 3 intr., 1 X 2nd m. of comp.), always preceding the verb. - V. adj. are rather rarely employed as adverbs (a.o. 6 X ifc.). - They form part of a compound in a very high number of cases. II-A gerundive (25 cases) does not serve as an attributive (the Roma capta constr.: 255, II excepted); it is used 22 X without a copula, 2 X with a copula : 259, 6 durāraksyam bhavet ; 277, I asmi bodhyah (apodosis); 16 X without an expressed agens, 6 X with an agens, once without a subject (neutral); 271, 6 sanıdeśyam; twice it forms the 2nd member of a compound; na śakyam + inf. occurs once. P. 256 ff. it serves as a means to vary the style; it is often used in or. dir. III–There are 20 pers. forms in -ja-, of which 8 are indic. pr., 5 imperf., 2 opt. The agens is added twice; 7 X the subject is a person, 6 X a thing; 2 X the form is neutral. The 'passive form' expresses an intrans, notion : 253, 3 laksyate; u prativisişgate; 274, 3 drśyate ; 265, 5 vyaśīryanta; 266, 2 adahyanta (vimānitäh manenād.); 275, 15 apriyanta; 'agentless passive' 252, 6 agan yatāmareșu; the form is neutral: 255, 2 tathā kriyate "let it be done..."; the agens is known: ? 256, 6 ācaryate (viz. tena, ácaratīty arthah Comm.); 276, 13 kim udgirycta grasyeta vā (patiens is the thema of the sentence); fr. "on": 260, 8; this constr. is probably chosen in order to keep the subject 254, I; there seenis to be no reason why the passive is used: 259, 15 tair dhanam ... bhujyate; 265, 13. There are 5 imper. in -ya-:- 250, 4 kriyatām; II śrūyatām; 258, 13 kriyantām; 259, IO anubhūyantām; 272, 8 anugrhyatām; 3 are courteous forms, 2 express fr. "on", once a person is subj., 3 X a thing, i form has no subj. I have found II forms of the part. in -ya-: 4 express attendant circumstances 254, I pers. subj. without ag.; 261, 3 id. + ag. ; 272, 17 id., following the verb, expressing the notion of fr. "on" in a nominal sentence: 256, II (thing subj.) (in these instances the passive form is due to the ·TEXTS 41 sa construction of the sentence in general, which is determined by the caraxt, the thema etc.); the part. acts as an app. to the obj. 258, II coing): 261, 14 (th., • ag.); it is predicative : 259, 18 (th. + ag., the patiens is theina); acc. + part. constr. instr. ag. 253, 7 (pers.) ; 261, 12 th.); twice it forms the ist member of a compound: 257, 5; 262, 2. Paiicatantra, Hitopadcśa, prose Vetāla-tales, etc. To get an insight into the use of the passive forms in contradistinction to the active forms in the rather easy and artless prose of later times, I studied a number of tales, fables etc. appertaining to the cycles of the Pañcatantra, the Hitopa deśa, the prose V etālatales etc. With regard to the fourth part (tantra) of the Pañcatantra 6? I found the following facts. In the dialogue the functions of the indic. pr. act. (or med.) and the verb. adj., which are equally frequent, are clearly distinct. The indic. is either a durativum: 2, 4 phalūni prayacchati nityam a form expressing facts present or represented as such: 4, 3 mamāsti ... näina vānarah, or a modal tense: 4, 20 tasmåt kathayāmi nijābhiprāyam (volunt.); 3, 13 kim karomi (dubit.) 8, 4 (possibility) etc., which may also denote a near future: tad bhayo 'pi visvāsayāmi. In subordinate sentences the ind. pr. act. (or med.) is often consecutive or final: 4, 13 tām ... anaya g'ona tasyāḥ prananı yāśārvādam grhņāmi. We find it in exclamatory sentences beginning with kathan: 9,6 katham aham tatra gacchūmi. The verb adj., on the other hand, expresses the actual past: 4, 2 tad aham tayaivam proktas tava sakāśan ägatah. Occasionally such a verb. adj. is also met with in a sentence beginning with yathā and paraphrasing the object of the verb "to say" etc. 18, 5 ... ücatur yathă gajam dystvā dūrato 'pi nastah. Moreover, it is used in ‘fait accompli’-expressions: 4, 16 ... kim vilambyate. tvaryatām aham tava prsthan ārüdhaḥ "I am already sitting on your back”. In a story or tale related by one of the persons of the fable the use of the v. adj. may bring about a dramatic effect; 16, 12 where the person speaking relates a fact that had happened to himself: atha kadācit ... nirgatah pradhāvan karparopari patitah etc. (the after-effect of the stumbling has not yet lost its actuality). When a v. adj. is the predicate of several sentences in succession that part of the text is 'dramatic': 14, 5-8. Now and then it is a pure adjective and expresses a quality or a condition: 12, 23 madbhujapariraksitaḥ sa deśaḥ; 22, 19 bhaktāsi me tvam; more than I 62 Pañcatantra, p. III of Bühler's edition, p. 1-44 (Bombay 1868). GONDA, The Sanskrit Passive ·42 REMARKS ON THE SANSKRIT PASSIVE once it expresses an actuality and a condition etc. (resulting from a preceding action) at the same time: 10, 12. In the narrative parts, on the other hand, the ind. act. or med. is either a praesens historicum: I, 18 evam uktvā tasya ºphalāni prayacchati, or a durativum: 10, 2 nityam ekaikam adišati; cf. especially such groups as I, 21 nayantau... tişthatah, which are also found in oratio directa (e.g. II, 6). This form is often historical and durative at the same time: 12, 7. Jore than once such words as nityanı "continually" precede the verb. On the other hand we often see that sentences in which the following words occur, kadācit "once", atha kadācit, anyadā "at another time", athānyatame divase "some day or other", anyedyuḥ, anyadine "one day", atrāntare "meantime" and sentences beginning with atha, which serves to introduce some new person, thing or fact, tatah, which cominonly subjoins an event which is subsequent in time, that "therefore, now” contain a v. adj.: 1, 9; 33, 9; 2, I athānyatanie divase tayā sa prstah ; 10, 6; 26, 4 aträntare ... bhūmau patita; 6, 15 atha tena cintitam; 12, 13 tataś ca ... tenābhihitah ; 37, II etc. The same may be said of sentences beginning with the loc. abs. tathānusthite "having done in this way": 13, 12 t. sygālena saha samhāntikāgatah; 38, 7 etc., after the absolutive constructions evam uktvä: 28, 15; ity abhidhāya : 37, 16; tac chrutvā: 29, 13, cf. 42, 5, after other absolutives: 29, 18; 33, 19 etc. After the expressions kadācit etc. (atha not included) another verbal form is only rarely found: 30, 9 atha sa duḥkhābhibhūtā tam ... viksitum na saknoti; 33, 20 atha ... cintayāmāsa; 37, 8 atha ... apaśyat; 34, Io atrāntare kham utpapāta. After an absolutive, however, an imperfect is not rare: 6, 19 tam drstvū bhūyo 'py acintayat; 8, 11; 9, 9; 22, 7; 36, I; 17, 17 ity uktvä tam api svastanakșirena pustim anayat; 21, 20 ... annam ... akarot; 22, 1 ... kathamcit tam divasam atyavāhayat; 33, 18 upõdravat; in these cases the verbs express ideas the realization of which extends over a certain space of time; cf. also 35, 6 iti vicintya aprcchat; but 24, 12 is an exception: päşānakhandam praksipat. Elsewhere the abs. is followed by a perfect: 32, i tad ūkarnya vānarah prahasya provāca, by a form in -tavat-; 37, 15 abhihitavān, by an aor, I, 19 agăt; 5, 8; 13, 5 avocat. It would appear to me that one verb has a predilection for the imperf. (or perf. etc.), another (gam, pat- X, dró-X, pradā-, prastha-) for the v. adj. The same difference between the personal active verb and the v. adj. can be studied when both forms occur in the same syntactical unit. A frequent construction is yāvat ... tāvat... "during which time ... in ..

·TEXTS 43 the meanwhile". Speyer has already remarked 63 that yūtat in these sentences is construed with the present, even when it denotes past facts. But he failed to notice that in the apodosis the v. adj. is very common. This form expresses a new fact and, as a rule, the action is either brought about in a moment or ingressive : 30, 31 "when she was finding herself ..., a thief entered (pravistah) the house”; 13, 13 ...palā vitun ārabdhah; 14, 2 drstah; 40, 19 gatah ; 21 sthital ; 23 inőritak ; 17, 19 ... tāvat tena śrgālasutenābhihitain. Now, in the tūvatapodosis an aorist is also possible: 34, 7 sāpi ... nadītīre yüvad upavistä tisthati tāzat taträntare ... tatrāgamat. Elsewhere we find an inperf.: 37, 10 jāvad udakamı nitvägacchati tāvat tām 11lytlīnz apaśyat, an indic. pr. 10 śruoti, a perf. 38, 12 dadarśa. The v. adj., however, is more frequently used. Next, a v. adj. is found in sentences continuing sentences introduced by atha etc, or a loc. abs., followed by a 1. adj.: 38, 7 sqq. tathānuşthite ... brāhmanah taya ... kipāntah pätitah. süpi ... nagare pravistä. tatra ca ... rājño 'gre nitü; 18 sqq.; 8, 22 sqq.; perhaps it may be said that these parts of the text are 'dramatized'. As regards the perfect, babhūva in semi-nominal sentences is found 16, 2; 22, 13; also 26, 6 tüsnīm b.; as an introductory verb before an or, dir it is frequent: äha 3, 16; 4, 15; uvāca 10, 22; 27, 13 etc.; cintayāmāsa 4, 19; 33, 20; papraccha 27, 2; sometimes, however, the author uses the aorist: avocat 12, 8; 21, 17; avādīt 31, 6; to introduce stanzas the perfect is not often used (āha 11, 8). This long and often pompous or cumbrous form serves also to conclude a pericope, to mark ideas which stand out a little above other verbal notions in the same context: 10, 8 ... kathamcit na virarāma; IO, 13 evam uktvá sa tām visarjayāmāsa; 23, 15 babhrāma; 24, 13 nipapāta; 26, 5 papāta; 34, 2 Etiam uktvā tasmā aśesam vittan arpayāmāsa; 36, 9 babhaſīja; 39, 19 it expresses a result and a condition(?): să garbham dadhāra. atha prasavasamaye prasūtā. In or. dir. the only instance is the introductory ūcuḥ 13, 2. Aorists are in narrative parts): 3, 14 iti vicintayan ... agamat; 16, 22 ajijanat; 27, 12 ayat. So there is not much room for the supposition that the use of the verbal form which at first sight seems to be capricious and disorderly, is highly a matter of arbitrariness. But we must not preclude the possibility that the tendency to vary the style (as well as the rhythmical and euphonical motives) had influence too. Perhaps the use of the 63 Speyer, Sanskrit Syntax, § 475 I.

·44 REMARKS ON TIIE SANSKRIT PASSIVE -tavant- forms which do not occur in the or. dir. is to a certain degree due to these influences: as a voluminous form it may easily serve to conclude a pericope etc.: 44, 20 before the concluding stanza of a fable ... svāśrayam sukhenadhisthitavān; 12, 12 ...grāmam äsäditavān (this form also 5, 9); 15, 8 bhaksitavān; 23, 14 samāro pitavān (with an 'aoristic sense?); 17, 4 sanar pitavān; 30, 22 älingitavati ('aor.' sense?); when introducing an or. dir. it may have been chosen for the sake of variation: 5, 22; 7, 4; 22, 2; 37, 15. As regards the indic. pr. pass., in dialogues it is either intr. (or neutr.) 2, 9 na śakyate; 4, II; 21, 2 etc. na yujyate; 28, 20 na jñāyate; 29, 4 dhriyāmahe; or a form expressing fr. on 36, 17 ...na mitrabandhavādayo ... samudre praksi pyante ; 14, 1o yat sthīyate; 29, 10; sometimes this form is less to the point than an act. verb: 27, 17 iat kasmai ... pradījate (“we”); 4, 15 bhadrażn yady evam tat kim vilambyate; 29, 8 tat kim kathyate. In the narrative parts there is no difference: intr. 1, I athedam ārabhyatc ...; 5 tad yathānusrüyate; (intr. or Fr. on:) 19, 16 nīyate (“may be" or "we") 40, 3; 47, 15; 44, 10 vidāryate "when he was about to be rent in pieces...". As introductory formula we find ucyate (Fr. on) by side of uktam ca. From the first tantra of the same text 64 we may gather similar information: the v. adj. expresses (among the notions mentioned above) the ‘remote actual past': 12, 7 äkarnitam maya nītisāram pituh pūrvam utsangam hi niveşatā; it actualizes: 29, I (cf. 28, 12); it relates a fact in a dramatic' (vivid) style of writing: 30, 15 (in the same way the -tavant-form: 33, 18); durative: 27, 12 dūrībhūtas tişthati; the ind. pr. p. is intr.: 18, 3 yujyate; 21, 8 drśyate ; 30, 2 praśasyate; an intr. aor. 18,12 samajani (narr. past); Fr. on: 26, 17 sambhavyate; 37, 19 vadhyate; modal and emotional in or. dir. 28, 13 kim kriyate; 20, 2 kim ... vijſīāþyate. In the Hitopadeśa 65 the ind. pr. act. (med.) is either a durative form: 172, 23 pratyaham khädati, a narrative historical present: 168, 22 mrgakākau nivasataḥ, a timeless present: 171, 6 asti... Ovrkşah, or expressive of a present fact: 171, 14 ärye tvām abhivande ; 184, 22 (formulas), modal etc. : 169, 33 katham etän dantaih spyśāmi; 168, 24; 178, 1 (often in a subordinate sentence: 170, 27; 186, 6; 187, 29; de conatu: 176, 15). That there can exist a striking difference between this indic. and the v. adj. becomes especially clear from texts like 188, 64 Edited by Parab and Panšīkar (Bombay 1925). 86 Book I and the fables from the other books in Boehtlingk's Chrestomathie", to which the pages and lines refer.

·TEXTS 45 .. 15-17: ... tadhu ... sarvadā ranate. athaikada samudradattenālokita. The r. adj. is often preceded by an adv. indicating a fixed moment: 169, 18 prāta) ; 170, 25 prabhāte; 171, 10 kadācit; 173, 2 ekadā. It may also be expressive of an action happening in the space of time, denoted by a preceding part. pr.: 168, 22 svecchayā bhrāmyan kenacit ... avalokitan, etc. In dialogues the v. adj. chiefly denotes the acr. past: 168, 27 idūnīril pravisto 'smi; 170, 8 ilktalil et'a inajā pīr- T'Qi?? ; 171, 12 klī hato 'sui; a more or less remote past in subord. sent. : 182, io j'ad ete saśakāḥ trajā niksīritāļ... ; without an instr. ag. the v. adj. is actual-resultative: 172, 23 j'eşan apatvāni khäditāni The 1. adj. – asti or tisthati (durat.) is very rarely met with : cf. 169, 32 drdhabandhanabaddho 'sti tātad ayahı ??l?gall; the absol. + sthitah is dur.-narr. 170, 2; 163, 10; sa ca pracchanno bhiitvā sthitah; 183, II. Imperfects express the dur. and narr. past: 159, 20 āsit; 163, 3 etc, apaśyat; 173, 15 akarot; 163, 3 etc. acintayat; 185, 26 we find rijñā cintitai, a short introductory formula, acintayat, however, forms part of more extensive sentences: 168, 23 tam drstvā sygālo 'cintayat, so instr. + cintitarn (cf. e.g. tadā tenoktam etc.) are indicative of a brief and vivid style. The perf. is used in the same way as in the Panc.: 173, 5 ... krtvá sa vyādhak... hataś chinnadruna iva papāta (but 175, 23 patito 'hazı), after this papāta the author pauses a moment and adds a stanza; cf. also 176, 30 Susvūpa (“he slept"), but the ingressive prasvap- (176, 12) appears as prasuptaḥ (after ekadā), su ptah 182, 22; 186, 32 is ingressive too. 174, 19 vanikputram sevakam cakūra, but in the next sentence tato 'sau tena nivojitah, more dramatic; 190, 15 the first communication has to impress the hearer, the second is said 'en passant, cf. 168, 18; 185, 29 marks the end of a pericope in a sentence which adds no new information to that given above, but 30 new events are expressed by drstā prstā ca; 187, 7; before a stanza 159, 21 ekadā ... ślokadvayanı Śreśrā'a; 27 cintayārrāsa; 186, 19, etc. The perf. babhūva is often found in semi-nominal sentences (as the verbal member of the predicate): 170, 30 sayatno babhūva; at the end of a fable or of a pericope: 177, 17; 178, 3 phalāsaktā babhūvuḥ; 165, 26; the forms are durative. The perf. periphr. is met with at the end of a fable: 177, 2. As regards the indic. pr. pass., it is intrans.: 183, 18 vidāryate; 184, 28 pariciyate; 187, 2 vidyate; 168, 4 yujyate; 162, 27 prastūyate ; 177, 24 śruyate "one could hear"; it is courteous, at times it is a way of avoiding addressing somebody right out: 27 deva yadi kiyaddhanopakşayah kriyate tada... ; 179, 14; 186, 28; 189, 26; 160, 31 ucyate "it SO ·46 REMARKS ON THE SANSKRIT PASSIVE says" and 177, 8 evam śruyate introduce stanzas; cf. 183, 27; with an instr. ag. 162, 7 tan mayā nitim grāhayitum sakyante; 188, 8 in an emotional exclamation, when the patiens is the thema of the sentence: aho mahad āścaryan. pakşibhyani kırınaḥ samuhyate; atha ... dadhi tūram vāram tena käkena khádyate (here this form seems to bring out the condition of the sour milk). The pass. imper. is often courteous: 171, 16 etc.; it may include the person speaking: "we": 180, 7; it is more 'general': 187, 29; it occurs in formulas: 181, 2 śrūvatāın; one may use it when speaking to an inferior: 185, 25; an imp. act. is, on the contrary, addressed to a king: 175, 5 (drsti prasüdar kurui, stereotype formula); 24. With an instr. ag. the imper. pass. seems to sound dignified or official: 175, 20; st. 191 +. mere tendency to vary the style without any other motive for using this form and not another, is rarely to be found. It would appear to me that the use of the gerundive, e.g., is sometimes rather due to emotional motives than to the above tendency: 180, 9 kim aham nirbalah, mama grhärasthitāny andāni sarnıldrenāpahartavyāni; 188, 6 kim ahain ajiłał, na kim api mayā vaktavyam. It is only in the introductory formulas that the author has freely chosen various verbal forms on a large scale : brite, u ktarzı + ag., avadat, alia, abravit etc. Vetālapañcavim satikā by Sivadāsa. The prose recension of the Vetālapañcavimsatikā by Sivadāsa 66 is, from the point of view of the verbal forms, an interesting text. The author of it has a decided predilection for the verbal adjective. In the introduction and the first story (P. 5, 1-12, 12, the stanzas left out of account) the v. adj. is used 128 X in the narrative (55 X without an instr. ag. and without as-, 1 X with as-, 25 X with an instr. ag., moreover, 8 X neutral; uktam ca etc., 39 X neutral + instr. ag.: rājñoktam etc.) and 45 X in the dialogues (or. dir.: 19 X without an instr. ag. ; 16 X with an instr. ag., I X with as-, I X neutr, with as-, 3 X neutr. without as- or instr. ag., 5 X neutr. with instr. ag.). In the narrative there are no other passive forms, in the dialogues we find: indic. pr. 6; imper. 2, gerund. II (3 without ag., 6 + instr. ag., 1 + gen. ag., I + vartate) = 19. In the narrative parts there are 15 active forms: indic. pr. 7. perf. 4, -tavant- 3, aor. I; in the dialogues 53: indic. pr. 24, imper. 14, fut. 15. In the or. dir. the v. adj. is as good as always expressive of the actual past; sästre kathitam asti 5, 66 Vetālap. ed. H. Uhle, Abh. K. Morgenl. 8, I. (1881).

·TEXTS 477 34 lowever means "in the s. it says"; 8, 18 niūrchito 'si is a semi-nom. sentence; 9, 28; 10, 40 ...anuraktā sā are nom. sent. This form is also used when the person speaking relates events that have taken place in the immediate past: 8, 20; 36; II, 20 adya ... tātan mayā drstāḥ... grhiivā pradhūvito 'ham tāsāiz madhye ekā trisulena tāditā vastrăbharanani patitāni, tada mayā grhitāni. In the other parts oi the text the v. adj. either expresses the narrative past or serves as an introductory formula before a stanza etc. In the second member of a jūtat + indic. pr., tāyat sentence the v. adj. is constantly found: 8, 9; 10; 10, 36. The active forms, to be found in the narrative parts, are: asii 5, 9 and 7, 30 at the beginning of a story (timeless present); 5. 28 pratyahan darśanam karoti (durative and periphrastical, at the end of a pericope, before ekasmin divase); asti "finds himself" in a series of short, dramatic sentences 8, 19,... idānim gaccha" sū niskāsitä. ūgatā să durmukhī, jatra rājakıııāro 'sti; moreover, 3 ind. pr. after gävat (see above). The aorist agamat, occurring in the sanie context, 9, 20 is narrative. The perfect jagāna 9, 40 is placed at the end of a pericope dadşśāte 7, 33 before a stanza; babhūva 5, 32 in a semi-nom. sent. ; 9, 45 the same form seems to serve as a means to vary the style: ...darśanaipi samjīſātan; kušalapraśnaś ca babhiiza. Like the above daltavān 5, 26; the other -tavant-forms are also narrative. In the or, dir. the indic. pr. is either dur. or mod.-fut., sometimes (6,9 kim abhilașasi) it is a real ‘present. Ind. pr. pass. 9,6; 22; 30 vidyate (intrans.); 9, 32; 33 ko 'syā dandaḥ kripate is ‘general and modal, “what punishment ought she to receive?"; cf. 9, 29. The imper. pass. is courteous 5, 34; 7, 29 śrüyatām (to a king), but the act. imper. is also used when a king is addressed (e.g. 6, 35). Friends use the latter form. Rgreda. From the Rgveda I collected the present-forms in -ya- 'pass.' mentioned in Delbrück's Syntax 67, Grassmann's Wörterbuch and Avery's Contributions 68. There are 191 indic. forms (1 pers. sg. 1, 2 sg. 38, 3 sg. 139, 3 du. I, I pl. 1, 3 pl. 11 69). They are often intrans.: 5, 54, 12 sam acyanta "crumble down"; 6, 2, 8 ajyase "lässt du treiben dich" (Grassmann, R.-V. übers.); 5, 21, 2 idhjase "flare up"; cf. 8, 6, 30; 6S 67 B. Delbrück, Altindische Syntax, $ 159, p. 268 sqq. J. Avery, Contributions to the history of Verb-inflexion in Sanskrit J.A.O.S., 10 (1880), p. 219 sqq. 69 See Avery, O.C., p. 318.

·48 REMARKS ON THE S.INSKRIT PASSIVE ava 10, 34, 6 tapyate "suffers grief, is grieved"; 1, 81, 17 lujvate "flieht": (Geldner, Übers.) ; 10, 146, 3 drśyate "es zeigt sich"; 5, 34, 7 dhrijate "hält sich” (Geldner); 1, 81, 3 dhīyate "steht auf dem Spiel" (Geldner); 9, 69, 2 pcyate "ergiesst sich” (Grassmann); cf. I, IIO, 4; 7, 21, 2 ni ... bhriyante “sie stürzen nieder” (Grassmann 70); 5, 31, 12 bhrijäte "herabdrängen"; 10, 90, 5 aty aricyata "reached beyond" ; 1, 31, 14 ucyase "du heisst"; cf. e.g. 3, 29, II; 6, 28, 6; 7, 82, 2 etc.; I, 176, 2 anu upyate "folgen" 71; 5, 44, 9 vidjate; cf. 10, 64, 2; 9, 71, I a ... srjyate "ergiesst sich"; 69, 2 sicyate "ergiesst sich"; cf. 39, 5; 62, 13; 5, 54, 7 ria sa jīyate zilaruto na han pate; cf. 10, 152, I hanyate = "get it hot”; matiḥ ... vacyate "mein Denken schwingt sich”; 9, 2, 2 a vacyasva “quill hervor”. Nor is a Fr. passive unfrequent, some instances of which we would prefer to call 'agentless passives': 10, 85, 28 patir bandheșu badhyate; 9, 66, 9 ajyase; 8, 38, 10 yabhyam gayatram ?Cyate; cf. 7, 70, 6; 3, 25, 5 agne apām sam idhyase durone may also be considered as intrans.; cf. IO, 118, 5 etc. ; 4, 15, 4; 3, 27, 13; 14, 10; 10, 112, 9 na ſte tvat kriyate kim cana "without you nothing is done”; cf. 3, 41, 3; 8, 79, 3; 9, 74, 4; 4, 51, 6 na vi jñāyante "sie sind nicht zu unterscheiden". At times such a -ja-form is accompanied by a loc., e.g. I, 53, I; 135, 2 deveșu huyate "... wird bei (unter) den Göttern geopfert” (Geldner) ; 10, 22, I kasininn adya jane mitro na śruyate "bei welchein Mann lässt er sich hören?"; 116, 3 sūyate pārthivesu; vāl. 3, 9. An instr. of circumstance etc. is found: 1, 188, 11 gayatrena sam ajyate "wird unter Gesang gesalbt"; an instr. of instrument etc.: I, 12, 6 agnināgniḥ sam idhyate ; 164, 47 ghrtena prthivī vy adgate ; 3, 24, 2; 6, 5, 6; 9, 65, 6; 68, 9 adbhiḥ ... mrjyate; 86, 47 yad gobhiḥ ... samajyase, cf. 8, 20, 8; 3, 40, 6 etc.; 10, 118, 4; 151, I śraddhayāgnih sam idhyate. In some cases the character of the instrumental may give rise to some doubt: instr. or agentis?: 9, 76, 4 yah sūryasyäsirena irjyate "who is cleaned by means of) the rays of the sun", cf. I, 12, 6 (see above). Instr. ag. 3, I, 21 jātavcdā viśvāmitrebhir idhyate; 10, 69, 8; 91, I; 3, 26, 3; 62, 7 asmābhis tubhyam śasyate ; 6, 45, nrbhiḥ suvira ucyase „bei Männern heisst du ...” (Grassmann); 7, 76, 7 ribhyate vasisthaih; 8, 4 hūyase nrbhih; 54, I; 9, 3, 3 eşa devo 6 70 In his Wörterbuch Grassmann remarks (850 f.): „Die passiven und medialen Formen haben oft eine Bedeutung, die zwischen der passiven und reflexiven schwankt, z.B. gefüllt werden" und „sich füllen". 71 Cf. Geldner, Der Rigveda übers. (1923), p. 230.

·TEXTS 49 vipan yubhih ytäyubhiḥ harir vājāga mripate ; 2, 3 nrbhir yato vi niyase; 99, 8; 27, 3; 78, 2; 86, 12. At times the agens is 'suppressed': 1, 77, I kathā dāśemūgnaye kasmai devajuştocyate bhimnine gīḥ; cf. 9, 78, 2. The part. praes. pass. in -ya- serves as an attrib. or appos.: I, II, 5; 5, 30, 14; 6, 38, 2; 55, 2; 7, 35, 14; 8, 2, 14; 10, 31, 9; 55, 7; 89, 6 etc. This part. is attended by an instr. of circumstance etc.: 4. 34, 6 namasā hīyamānäh; of instrument etc. (instr. sociat.): 2, 40, 3 manasā g'ujyamānan; instr. instr. or ag.?: 4, 58, 6 hrdā manasū pūyamānāḥ; instr. ag. : 3, 8, 9 unnījamānāļ kavibhik; 4, 58, 4 panibhir guhyamānam; 5; 3, 8; 6, 69, 2; 7, 63, 3 rebhair ad ety anumad vamānah; 9, 30, 2 etc. ; 96, 16; 10, 98, 8; cf. 9, 62, 13; gen. ag. : 6, 49, 13 tasya te sarmann u padadyazāne. The aorist in -j- is often found: there are 178 indicatives and about 40 subjunctives 72. Most of thein are without any doubt intransitives or eventive, agentless forms 73: aceti "ist erschienen", ajani "ist geboren", abodhi "ist erwacht”, aroci "ist erglommen”, asādi "hat sich niedergesetzt” etc. Although this sense is almost always evident, 'passive translations are sometimes found (e.g. 7, 73, 3 "ein eifriger Bote ist euch erweckt", abodhi 74), where I should prefer an intransitive (“he has become mentally active"). Other forms are as a rule agentless ‘passives' or impersonal: adhāvi, ayāni; but at times an intrans. meaning is found, e.g. 1o, 29, I vane na va vo ny adhāyi cakan "has set down". The agens is occasionally expressed by another verbal form at a short distance, e.g. I, 119, 6; yur'an rebham parisäter uruşyathaḥ ... 1 yuvam śayor avasam pipyathur gavi pra dirghena zandanas tāri dyuşā; 4, 12, 6. The origin of the process is expressed by an abl. + the prep. a: 8, 63, 7 iyam te navyasi matir aguc adhāyi asmad a "es ist dir dieses neue Lied ... dargebracht von uns” (Grassmann). An inst. of instrument etc. is found 3, 61, 6; 5, 1, I abodhi agnih samidhā janānām; 4, 16, 4 svar yad vedi arkaih; 6, 12, 4. An instr. instr. or ag.?: 9, 97, 23 pra raśmibhir daśabhir bhari bhūma; IO, 144, 5; 6. An instr. ag. occurs: I, 20, I stomo viprebhir ásaya akāri; 1, 63, 9 akāri ta indra gotamebhir brahmāni (cf. 61, 16; 62, 13, where active turns are used); 184, 5 (the verb akāri is, however, ... 72 See also Avery, O.C., pp. 275 and 318; Deibrück, Altindische Tempuslehre (Syntaktische Forschungen, II, 1876), P. 51 ff.; idem, Altindische Syntax, $ 157. 73 See also the survey given by A.J. Eaton, The Atmanepada in Rigveda, Diss. Leipzig 1884, p. 21 ff. 74 Delbrück, Altind. Tempuslehre, p. 16.

·50 REMARKS ON THE SANSKRIT PASSIVE 19 met with 12 times without an instr. ag.) ; 10, 45, 12 astāvi ... ?șibhiḥ; 63, 17 astāvi jano diviyo gayena; 2, 20, 8 tasmai tavasiyam anu dayi satrendrāya devebhir arasātau “diesem Indra ward von den Göttern im Kampf um das Wasser ganz und gar die Überlegenheit zugestanden"; 6, 25, 8 ani te dāyi ... anu ksatram ... devebhih (in these instances the agens is placed at a great distance); 6, 20, 2 asuriyanı devebhir dhāyi visvan; 4, 7, 1 ayam iha prathanzo dhāyi dhātybhir hotā; 8, 81, 29. The agens is expressed by an encl. pronom.: 1, 162, 7 interpreted in this way by Geldner 75; "ein schönes Gedicht ist von mir gemacht" suman mc 'dhāyi manma, because manma (stomam, dhiyaııı) dhā- means "ein Gedicht machen"; 175, I mnatsi, apāyi te madah. 6, 29, 5, however, an intrans. meaning seems to be expressed: na te antaḥ śavaso dhāyi asya. The gen. ag. eşām occurs 1, 61, 15 asmā id u tyad anu dāyi eşām "ihm ward das von ihnen zugestanden" (Geldner). As is well known, Vedic middle forms may be used in a 'passive' sense; very often, however, this so-called passive is an intransitive. RV. 7, 17, 1 barhir urviyā vi strnitām has been interpreted as a passive by Grassmann, Ludwig etc., but Delbrück 76 rightly translated "...breite sich aus”. Occasionally an instr. ag. has been added : 1, 135, 6 (ime võm somā ... iha) adhvaryubhir bharamāņā ayamsata "von den Adhvaryu's aufgetragen sind sie dargereicht" (Ge.); 1,92,7 divaḥ stave duhitä gotamebhiḥ “the daughter of heaven is praised by the G.” (but grne 77) may be interpreted as "(he, it) obtains praise”; in the same way śrnve I, 74, 7 "man hört' ; 9, 41, 3; 10, 94, 6 “es klingt" ; 6, 47, 16 śnue virah ... damāyan; 3, 55, 20; 5, 73, 7 s. ... sam tanih "schallt Getön" etc.; 4, 17, 9 ya ajisu maghavā śļņva ekah has been translated by Delbrück "der als Mächtiger in Schlachten allein berühmt ist". "Verhältnismässig häufig", Delbrück says, “ist passivischer Gebrauch bei aus der Wurzel gebildeten Participien auf -āná-, die man zum Präsens oder Aorist ziehen kann" 78. In fact these participles are comparatively often accompanied by an instr. ag.: 3, 10, 4 añjānah sapta hotybhih (2, 8,4; 9, 50, 5; 103; 2; 107, 22 instr. instr. etc. ; 4, 58, 9 without an instr.) ; 9, 64, 19 vujāna zkvabhih; 97, 28 (ag. or instr.?: 6, 49, 5 yujāna- 'in passivem Sinne 79 occurs 5 times without an in- 75 Geldner, Der Rigveda übersetzt und erläutert, I (1923), p. 201. 76 Delbrück, Altindische Syntax, $ 156 (p. 263 f.). TT See Delbrück, 0,6, p. 264, 11. I. 78 Delbrück, oc., D. 264. 19 Grassmann, Wörterbuch, 1118.

·TEXTS 51 str.); I, 130, 10 divodāsebhir indra stavānah; 7, 36, 5; 9, 97, 5 (but without an agens I, 12, II; 31, 8; 4, II, 2; 5, 10, 7 etc.); 9, 76, 2 hinvānah manīşibhiḥ; samvidāna- is accompanied by an instr. soc. (7, 44, 4 etc.), vidāna- (I, 169, 2) is not a passive. I have not examined all the instances of the perfect. med. - according to Avery's list there are 843 indicatives and 70 other forms -; it has, however, already been stated by Delbrück 80 that 2, 9, 3 pra tve havinsi juhurc samiddhe may be interpreted as "haben sich ergossen"; 7, 90, I on the other hand is a genuine passive: pra ... dadrire vām adhvaryubhir ... sutāsaḥ, As a description of the sentences with a verbal adjective acting as verbum finitum has not been given hitherto - Delbrück who remarked that "die Form auf -tá- in dieser Verwending sehr häufig (ist)" 81, mentions only a few examples I examined the suktas 1-70 of the Ist and 1-42 of the 3d mandala and the whole of the 2d mandala. In these texts the ppp. is used as predicate; 1, 25, 17 yato anc madhu. ablırtain (it is perfectly immaterial to the auditors who has brought the drink, it has been brought, it is at the disposal of the man who drinks); 32, 10 nihitam śarīram “war sein Leib begraben", Geldner, why *war": "he lies interred", the author continues : ... vi caranty apo "the waters flow over (him)”; 34, 2 trayah skambhāsah skabhitāsah "drei Stützen sind befestigt” (Ge.) (the author continues: trir naktam yāthah); 8 tredha haviş kytam ... raksethe; 36, 5 (3-5 I indic. praes., II tve viśvā samgatāni vratā) : 57, i after an adj. acting as a predicate (durdharani) : apāvītam "is open" 82; 2, 3, 4 barhir vardhamanan stirnam; II ghyte śritaḥ; 5, 2 a yasmint sapta raśmayas tatāḥ; 10, I; 14, 10 nibhrtam ma etat "it is settled, certain..."; 16, 2; 18, 6 ayam parişikto radāya; 28,8; 36, 5 (3X); 3, 1, 7; 12, 8; 18, 5; 28, 2; 29, 2; 30, 2 (2 X); 14; 35, 783; these instances represent the adjectival use of the ppp.; it denotes a state or a condition. Cf. also 2, 12, 4 yenemā viſvä сyavanā kytūni. 2, 41, 4 another interpretation might be admitted: ayam vām sutaḥ somah "has been pressed out (just now"; cf. also 3, 42, 5; whereas I, 51, 12 ã sma ratham ... tisthasi, śāryātasya prabhytah (viz, the soma-drinks) this actual sense is almost certain; in a vivid relation of past (“historical) facts: I, 33, hi te ... Sonia 80 Delbrück, O.C., p. 264 f. 81 Delbrück, O.C., p. 394. See also J. Bloch, La phrase nominale en sanskrit, M.S.L. 14 (1906-1908), p. 31. I, 63, 9 has been interpreted incorrectly by Delbrück (O.C., p. 394), okta is an appos. ppp. (see Geldner, o.c., I, p. 75). 35, 7, however, might be compared with 2, 41, 4; so might 3, 28, 2 and 30, 2. 82 83 ·52 REMARKS ON THE SANSKRIT PASSIVE 6 vrsäyudho na vadhrayo nirastāh. Another dramatic passage is 3. 4, 4 akari ... prasthita ... ni asādi; cf. also 23, I; 29, 12; 41, 2 satto hotā nah tistire barhiḥ. The neg. particle a- has been prefixed : 1, 15, 5 astytam "invincible"; 24, 10 adabdhāni varunasya vratāni "unverletzlich" (Ge.); 52, 2 acyutah "unshakable";2, 28, 9 az'yustāh... itşāsa? "not yet risen". In fine comp.: 51!-: I, 38, 12 susamskrta-; 2, 24, 9:3, 30, 13; 36, 1; the agens is expressed by the first member of the compound: I, 10, 7 trādātanı id yasaļ. The verbs as- or bhu- are added : 1, 10, 12 juştā bhavantu; 24, 7 nihitā) ... Synt}; 32, 11 II apihitam yad asit (cf. also I atisthan niruddhā); 38, 15; 2, 30, 10 jyog abhuvann amudhapitāsaḥ; all adjectives 84, as I, 38, 4; 2, 41, 16 aprašastū iva smasi; 3, 30, 6 vistan astu "es soll ein Ende gemacht werden": 31, 20; 36. I. An attendant instr. is seldom found: instr. ag.: 1, 58, 3 rudrebhir vasubhil purohito hotă nișattah ; 2, 37, 2 adhtaryubhiih prasthitarra madhu, neither example being narrative; 38, 7 (?). Although in the sūktas I, 1-70 and 2 the ppp. is used as an attrib. or an appos. about 125 times (the compounds excepted), an instr. ag. is only found in 7 cases: I, 51, 2 tavişībhir avrlam; 54, 10 nadivo vavrinā hita vista anuştāḥ; 55, 8; 2, II, 2; 17, 3; 22, 1; 38, 5. An other instr. occurs 1, 63, 9; 69, 8; 2, 3, 3(or ag.?: īļito agne manasā nah); 4; 7, 4; 10, 3 (a-) ; 17, I (?); 22, 3; 23, 18 tainasă parīvrtain; cf. also I, 54, 3 (?). I add a number of instances to be found in Macdonell's Vedic Reader: as predicate: 7, 49, 4 yāsu agnih pravistał ; 10, 129, 5 (condition etc.); 10, 90, 8 (6 āsīt, 7 praukșan, ayajanta) sambhytan (.cakre) ; 12 brāhinano 'sya mukhanı āsīd bāhī rājaniah krtah | ari tad asva vad vaisyah padbhyāni śüdro ajāyata (with a remarkable variation !); 13 candrama jātaḥ, ... süryo ajāyata; 15; 129, 6; 135, 6 (2X) (narrative past); 10, 15, 5 upahitāḥ pitara! "invited are the Fathers" (actual past); it would appear to me that the greater frequency of this use in these sūktas which form part of the ioth mandala, is due, not so much to a later origin of these parts of the Rgveda, as to their containing other matter. The instr. ag., which is seldom found, occurs e.g. 8, 65, 4 (after 2 abhinat, 3 vi airayat) in a vivid enumeration of Indra's achievements: ayan ha vena vā idamı suvar marutvatā jitam. The gerundive seldom serves as predicate: I, 14, 8 ye yajatrā ja idyāḥ; 2, 22, 4; with as- or bhū-; I, 33, 2 yaḥ stotrbhyo havyo asti yāman "der für die Sänger auf seiner Fahrt anzurufen ist"; 38, 5 7 .. 3, 29, I I consider krtam to be an appos.

·TEXTS 53 mā... bhūd ajosyah ; 63, 6; 2, 2, II sa no bodhi ... praśamsyaḥ; 4, I; with an instr. ag.: I, I, 2 agnih ... rşibhir idiyaḥ; cf. also 9, 76, I airumādyo nrbhih. Atharvaveda. in the fourth book of the Atharva veda active verbal forms serve 862 times as predicates; there are 123 nominal sentences, no -iavant-predicates and 29 'passive forms acting as predicates. The -ta-form (23 instances) denotes a condition etc., it is a real "verbal adjective”; there are, however, some peculiarities to be noted: without an ag. and without as- etc. (12 X) it is only such an adj.: 4, II, 8 yatraisa vaha āhitah "where ... is set"; 8; 24, 6; 28, 2 viditāv işubhrtām asisthau "who are known as best of ..."; 35, 4 yasmān māsā nirmitāḥ ... "out of which were fashioned the ... months"; 4; 6; it is a verbal member of a predicate: 6, 7 sarve te vadhrayah krtāḥ "they (are) all made impotent"; 7; when the v. adj. is placed at the beginning the sentence is emotional: 3, 6 mūrnā nırgasya dantā apiśīrņā tt prstayah. The instr. ag. is added: 24, 2; 25, 2 "by the influence of whom ... (a thing finds itself in a certain condition)"; (cf. 26, 7); gen. ag. 16, 5; 25, 2; v. adj. + as-: 12, 2 yat te ristam yat te dyuttam asti; 19, 2; 37, 3; 4; 5. The indic. pr. is intrans.: 9, 10 ucyase (with variation asi- u.); 21, 6; 36, 10 mucyate; once an 'agentless pass.': 24, 4 jasmai mīyante svaravah; an instr. ag. is added: 22, 5 yena jayanti na parājayante; an intr. conj. 16, 4 mucyātai. The perf. samśaśre denotes a condition from an action the agens of which is left unmentioned: 12, 7 yadi kartam patitva samśaśre "if, having fallen into a pit, he has been crushed...". In the Taitt. Sam h. (5, 1, 1-5, 5, 10; II, p. 1-88 W.) * the yá-form (indic.) is to be found rather often (about 70 occurrences beside a small number of other forms). It is an intransitive or an 'agentless passive', which may be translated by an intr. verb, by Fr. on or its equivalent or by a 'passive construction without an agens'. Cf., e.g., P. 3 kriyate "occurs", 12 pac yate "ripens"; 15 bhidyate "breaks" ; hanyate "is destroyed" (Keith), I should prefer "goes to ruin"; 29 aśīryanta; 40 jiryanti "grow old"; 42 ucyante Dutch "heten"; etc.; 'impersonal (Fr. on), the subject is unexpressed 69, I vujyate "is yoked”, agentless pass.' 36 reto dhiyate; the subject is known to everybody who is versed in the things under discussion 66 hūyate "the offering is made" (Keith); 14 kriyate ; 36; agens in the back- * Erroneously omitted at p. 34.

·54 REMARKS ON THE SANSKRIT PASSIVE ground 80 tābhyo vā esa ā vrścyatc "he is cut off”. At times the subject has been mentioned in the context: 73 vāyave . tüparanz ālablate, tejo 'gner vājus, tejasa - eșa alabhate "to V. ...he offers a hornless animal; V. is the t. of A., it is offered to t."; cf. 75; 33; cf. 15; 56; now and then the 'general subject is meant $5: 16 anibarīşe c'ā annani bhriyate "in a frying-pan food is kept"; 43 daksincnännan adyate (a 'fig. etym.' also 45 vidhā vidhiyate); 71. 51 it is immaterial who is the author of the process designated by the verb, the 'gramm. subject', however, has been mentioned in the context and is found at the beginning of the sentence: vişurīpain asya grhe drśyate yasya...; the 'gramm. subj.' follows 77 yajușă vă eşã kriyate, y. pacyate yad ukha. An instr. is seldom added: 13 chandobhir vă eșä kriyate (instr. of concomitancy); 61 vaišvakarnanena pāpmailo niramucjata (instr. instrumenti). The part. praes. pass. is found in a loc. abs.: P. 5, as an attrib. (appos.) 39. The Ppp. is seldom to be found: 29 vajreņaivāsmai pašūn parigrhạāti (action), tasmād vajreņa pašazah parigrhitaḥ (condition) (instr. instr.); 52. #.. Kathopanişad. In the Kathopanişad the ppp. serves as a predicate: 2, 4 jñātā; 4 viparīte ; 20 nihitah; 4, 9 arpitāh; 4, 8; 5, 8 śritāh ; 6, 2 nihsytam ; 14 śritāh; + acc. I, 29; 2, 3; 6, 16. The ppp., used as a predicate, is attended by an instr. 6, 9; by an instr. ag. 4, 8, where it denotes a condition: garbha iva subhrto garbhinībhiḥ; I, 21 devair ... zicikitsitam "there was enquired by the gods”; 22; 2, 10 mayā cito 'gnih "I have...”. A periphr. pred. is found I, II bhavitā pratītaḥ. The Ppp. is used as an attribute or as an apposition: 1, 15, 2, 12; 13; 3,5; 17; 4, I; 14; 15; 5, 2; 9; 6; 2; 17; with an instr. ag. I, 20 etad vidyanı anuśistas tvayā; 2, 8; the agens is expressed by the first member of a compound: I, II matprasysțah ; 25; app., when forming the second member of a compound: 2, 12; 3, 4; 5. The ppp. is app. of an accus. etc. I, II; 14; 3, 1 (+ acc.); I, 10 (ag.) tvat prasy stam mābhivadet. The ppp. in initio comp.: 1, 3; 10; 11; 2, 20. Ppp. as a subst. 2, 14; 19; 23; 5, 8. Loc. abs. 2, 8 (ag.) aranyaprokte. - Gerundives: pred. I, 21 suvijñeyam; 22; 2, 7; 8 su-; 8; 6, 13; + gen. ag. 1, 5 kim svid . • 85 The 3d pers. pl. occurs p. 9 bibhyati "people are afraid”; 26 dikşayanti. haranty, ahrta evdsyāgnik “f the waters strive to drag away his fire, verily it remains unmoved"; 86 vat samcitam ajyena proksati tena pūtas tena medhyas tena prokşitah "in that he anoints it..., thereby it is purified...". With as: 78 samyatta asan "were in conflict". Gerundives are used as predicates: p. 36; 49: 77; 14; + instr. instr, 26, ·TEXTS 55 yamasya kartavyan; 2, 23; + instr. ag. I, 25 na lanıbhanīyā manusyaih; 2,7; 4, 8; + instr. instr. 1, 27; 2,9; 23; instr, instr. or ag.? 4, II manasaivedam åptavyam; as a subst. 2, 13; 14. - Personal forms; intrans. I, 6 pachate "ripens”; 2, 1 hiyate 'rthāt "he loses..."; 3, 15 pramucyate "escapes”; 6, 14 "cease"; 4, 3 parisisyate "remains" ; 5, 4; 5, 2 vinuicyate; 8 ucyate; 6, I; 8 1721cylate; 15 prabhid yante; 'agentless pass.': 2, 18 hanyate; with an instr. instr. 2, 10; instr. instr. or ag.? 3, 12 drsyate ... buddhyā; 5, 11 (2 X) lipyate (bāhyado saih) “is sullied (spoilt) by the defects of external things". An interesting instance is 6, 12 astīti bruvato 'nyatra kathan tad upalabhyate "how could it have perceived itself at another place (> by another than him who says that it exists?” Part. pr. pass. app. 2, 8, + instr. ag. 2, 5 andhenaiva nījamāna yathāndhāh; in a loc. abs. 1, 2 (prose); 2, 18. Perf. 6, 5 pari dadrse (intr.). Mahābhārata. In the Bhagavadgitā, too, most -ya-forms are intransitive: 1, 29 parisusyati IV "becomes parched" and 30 paridahyate "is afire” are used in the same way, nor is it easy to draw precisely the limits of division in other cases; praduşyanti IV, 1, 41, e.g. may be translated by "are corrupted". Cf. also 2, 203. Intrans.: 2, 25 ucyate (also 48; 55; 56; 3, 6; 40; 6, 3; 8; 18; 8, I etc.) ; 31 vidyate "exists, occurs" (also 40; 4, 38; 8, 16 etc.); 34 atiricyate "is worse"; 3, 7 vifisyate "is superior" (12, 12); 13 mucyante 86 "rid themselves of, are freed from" (also 31; 5, 3 pramucyate ; 10, 3); 4, 22 nibadhyate "clings to the samsāra" (5, 12; 18, 9); 5, 7 lipyate (13, 31; 32; 32); 7, 2 avašisyate "remains"; II, 27 samdyśyante; 13, I abhidhiyate (17, 27; 18, 11); 17, 18 yujyate "is used”. In other cases the agens remains in the back-ground (it is, of course, impossible to draw the line): 1, 13 abhyahanyanta "were sounded"; 2, 20 hanyate “is slain"; 44 vidhīyate "is established" (cf. 67); 15, 3 upalabhyate "is comprehended” (Fr. on); 17, 18 kriyate "takes place, is performed" (19 etc.); 20 dīvate (21; 22); 26 prayujyate "'is employed”; 18, 25 arabhyate "is undertaken". (The expression tapyante ye tapo janāh is found 17, 5). The instr. ag. is added: 1, 43 dosaih ... utsädyante jātidharmaḥ (caus.); 3, 5 käryate ; 5, 5 prāpyate; ibid. tat (sthānam) yogair api gamyate; 6, 22 duhkhena ... zicālyate (14, 23); 44 hriyate ; 7, 5 dhāryate; 12, 5 3 86 See Whitney, A Sanskrit Grammar, $ 761 b.

·56 REMARKS ON THE SANSKRIT PASSIVE 2 avāpjate ; 17, II ijyate (in 12 the verb is repeated without the instr.); 25 kriyante. Other instr. (or, if so, the character of the instr. ag. is less pronounced): 3, 38 dhūmenādriyate vahnih ; 4, 14 karnabhih ... badhyatc; 5, 10 lipyate ... pāpena; 6, 35 vairügyena ca grhyate; 10, 7 yogena yujyate. The future (vi)mokşyase (4, 16; 32; 9, I), which is usually translated "thou shalt be freed" etc., is intr. ("you shall be, become, free", or "you shall rid yourself of" etc.). In Mbh. 5, a. 186-192 the indic. pr. act. and med. is, in general, either 'timeless' or a durative (191, 20); 186, 3 it expresses ‘modality', 189, 23 a future (emotively used); the perfect is the chief narrative tense. 186, 12 ff. the imperf. expresses such acts as have been witnessed by the person speaking, in most cases, however, it expresses the durat. parokşa; 186, 37 it is a simple narrative. Whereas the perf. (veda and äha excepted) and the imperf. occur only in the narrative (87 and 46 cases resp.), the aor: is found in the narrative (agāt, agamat, prāyāt, prādāt, il cases) and in the dialogue (5 cases), where it expresses the past, which has not yet lost its importance and actuality. Now, 192, 46 ff. a reprimand begins with the aor., which is very much to the point: yasmád adās ... cāgrahīh, but 47 the god speaking goes on with the pass. constr. ...apravrttam sudurbuddhe pasmād etat tvayā kytam. The ppp.-constr., however, is chiefly found in the dialogues (40: 15), where it is expressive of the actual' past, of a state or condition resulting from a past act, of a dramatic representation (e.g. 189, 12, where the text goes on with an aor.); at times it must be rendered by a plqpf. (192, 20; 26, in subordinate sentences). Of course, there are some verb. adjectives with the syntactical value of pure adjectives, e.g. 186, 15 vjathito dino gatacetā ivābhavam. In the narrative the use of this construction is found, when the author represents the events in a dramatic way, e.g. 190, II sa hi prakrtyū vai bhitaḥ... bhayan tivram anuprāpto drupadaḥ; 192, 63 or when a condition etc. is expressed. The imp. act. is used in a menace, e.g. 189, 23 phalam präpnuhi durmato, in a supplication addressed to a god: 192, 30 etc.; the imp. pass. is courteous, "let us..."; e.g. 179, 7; impersonal: 192, 44. The ind. praes. pass. is impers. (and intrans.): 186, 4 vidyate "es gibt"; intrans. 30 kliśyase; śakyate + inf. 186, 7; imper. intr. 187, 16. The story of Nala 1-19 (= Mbh. 3, a. 52-70). 1-Not all categories will be treated. V. adj. acting as predicates are very frequent. Unaccompanied by as- or bhu- and trans., agentless :

·TEXTS 57 ... 14 occurrences; 3 a personal pat., II an inanimate pat.; 12 in oratio recta. llore than once a pure nominal sentence is formed in this way; 3, 21 suraksitam hi me veśma; II, 7 nākāle z'ihito mnrtyur martyānām; 8 paryāptah parihāso 'yam; other forms and constructions to express the same thoughts would be quite uncommon; personal fornis of the verb pari-ap- are, for instance, extremely rare; cp. also 13, 10; 39; 8, 5; 5. Elsewhere however, the turn is equivalent to a verbal sentence: II, 6 śak yasa tā giraḥ samyak kartını izlayi ..., yās kathitah pura ? = "that you have spoken formerly", here the partic. constr. is very suitable to a gentle manner of speaking: the agens is leit unexpressed. 13, 23 the ag. has been mentioned before (actuality), cp. 17, 4 (modesty?:) te vayam damayantyarthe carāmah ... sejam āsāditā bālā; 15 bhīmasya sā dattā expresses the past which is still important. The ag. is of no moment 18, 25. 13, 15 the patiens is emphasized and the stanza, while summarizing the contents of the preceding portion of the text, describes the resultant impression of events which have taken place : evam prakārair bahubhiḥ ... c'inihatan sartam sam;ddham sārthanandalam. With an instr. agentis: 19; 31 pers. pat., 18 inanim. pat. ; 40 or. dir. Sometimes the nominal character of the constr. is clear, or, at least, perceptible: 9, 3a sista te damayanty ekā (nom. sent., s. qualifies D., the next pāda is on a par with a:) sartan anyaj jitam mayā; 4, 31; 13, 38. At times it is not easy to ascertain whether a verb. adj. acts as an appos. or as an independent predicate: 15, 16-17 sū tu tam puruşam ... anugată ... tyakta tena duskaram yadi jīvati, both anug. and ty. might be considered as appos., although I prefer to disconnect d.y.j.; 17, 50; cp. 12, 68. For the most part this constr. expresses the actual past: 4, 24 kaccid drsta tvaya ... d-i... ; 3, 25; 4, 3; 19; 27 (2 X); 5, 19; 20; 6, 4; 7; 12, 29; 34; 53; 55; 98; 13, 22; 26; 38; 39; 14, 14; 17, 6; 9; 18, 3; 19, 17; 20. Now and then the past is more remote but still actual: 12, 15; 16; 16 (cp. 18c uktavān asi yat) ; 83; 13, 42; 14, 5; 15, 19; 17, 7. Sometimes it expresses the narrative past 5, 28 (jagrāha, asrjat, varayāmāsa,) tato hāheti sahasā muktaḥ śabdo narādhipaih; 29 (here the v. adj. has an actualizing effect: the action which happened suddenly is presented to the reader in a vivid manner) ; 10, 7; cp. 17, 15; 12, 71 tair uktā (the subj. being known, is left unexpressed) is a short indication of the speaking person, perhaps originally an apposition, but, as it presents itself in this context, a concise and vivid turn of expression. 13, 12, in a narrative stanza the pred: expresses a qualification or condition of the patiens which, appearing in an anaphorical repeti- GONDA, The Sanskrit Passive 5 ·58 REMARKS ON THE SANSKRIT PASSIVE ... ; tion, is emphasized : kecid dantaiḥ karaih kecit kecit padbhyam hatā gajaiḥ, cp. II, where the dynamic te tam mamarduḥ expresses action. 9, I the patiens consists of a series of words expressive of totality: puskarena hrtam rājyam yac cānyad vasu kimcana the author dwells upon the vastness of the possessions robbed from N. Incidentally, an apposition accompanies the agens: 18, 2 naişadham mrgayānena ... mayā nalam ... bhāngāsurir upasthitaḥ; 10; cp. 12, 29 (absol.); 82; 15, 17 tyaktā tenälpapunyena; 19. The type asmābhir drstapúrvah is found 1, 29, without an ag. 14. The ag. is expressed by an encl. pronoun: 8 X, pers. pat. 2, inanim. pat. 6; actual past: 5, 14; 17; 13, 37; 40 (remote past); 69; 14, 17; 17, 43; cp. also 16, 10 yatheyam me purā drstā tathārūpeyam anganā. With gen. ag. : 3, 2 yasyāham dūta āpsitah; 12, 19 aham istā ... tava, nominal sentences. Intrans. v. adj.: 36, 25 X pers. subj., II inanim. subj. ; 29 or. dir. Nominal sentences, expressing a state or condition : 4, 22; 5, 4 āsanesu āsīnāḥ prthiviksitah ; 6, 4 nivyttaḥ sa svayamvarah; 9; 7, 13 part. pr.; 15 paurajano ... avasthitah “is standing"; 9, 3; II, 8 bhītāham "I am afraid”; 13, 24 grahā na viparītāḥ; the actual, (or result.) past: 3, 24 pravisto 'ham; 4, 25 (+ obj.); 6, 3; 8, 10; 9, 20 (or appos. in a nom. sent.); 11, 4; 30; 12, 73; 88; 119 ( + obj.); 13, 16; 24; 26; 33; 41; 37: 57; 16, 37; 17, 37; 41; 42; 18, 13; the narrative past: 5, 39; 9, 10 trirātram usito rājā; 12, 96 tapasāntarhitāḥ sarve; 13, 14; 21 (in a subord. sent., being practically ye + appos.). Accompanied by as- or bhu-: 9 X, 8 pers. pat., I inanim. pat., 9 or. dir. ; 2 instr. caus. etc., I instr. ag.; trans.: 3, 21 b, being a semi-nominal sentence (a and d are nominal, c nom, with v. adj.) katham cāsi na lakṣitaḥ, the gramm. person must be expressed; cp. 10, 24d; 14, 15 (rel. sent.) asi niketah (cf. 17 nikytas tvam); emotionally II, 3 hatāsmi; 3; actual past: II, 9 drsto 'si; 14, 5, in these instances the original function of the v. adj. is still quite clear. An instr. instrumenti or causae occurs 10, 24 dharmeņāsi samövrtā; 14, 15; an instr. ag. (+ appos.) 5 tena ... sapto 'smi. Opt. + v. adj.: 19, 4 (interr.) bhavet ... cintitah; 32. Intrans. 14 X, 13 pers. subj., I inanim. subj.; II or. dir.; semi-nom. sent. 4, 27 vismitāś căbhavan sarvah; 12, 97; II, 3; 12, 118; 14, 7 na sakto 'smi + inf.; 17, 35; 18 asmy ușită (after a part. pr. p.); act. past: 3, 20; 12, 13 kva rou rājan gato 'si 8; 18, prapannāsmi (+ 2 acc.); narr. past: 5, 9; 7, 10. Opt. + v. adj.: 12, 87; 19, 28. Thus, trans, and intrans.: asmi 7, asi 6, smah 2, abhavat 2, abhavan I, bhavet 4, abhūt I. II–Gerundives used predicatively. Without bhū- or as-: 18 instances, 7 pers. pat., I animal, 8 inanim. pat., 2 neutr.; 18 or. dir.; 10 X an ·TEXTS 59 ag. is added: 3, 2 kim ca tad vo mayā kāryam; 13, 27 nātra kāryā vicāraṇā; 69; 14, 24; 15, 3; 17. 40; 41; 41; 44; 45; 46 tathā kāryam atandritaih; 47; 18, 15; 16; 25. The agens is not expressed, when there is no definite person acting: 13, 27 "there is no room for doubt”; 15, 3 aham prastavyah “I am 'consultable", when we may conclude it from the context: 17, 40; 44; 45; 47; when the person speaking intentionally passes it over in order to express himself courteously or to address his desire, although it is pronounced to one single person, to an indefinite plurality of agents': 18, 15 ayam artho na samvedyo bhima mātah kathamcana. With as- or bhi-: 2 instances; I, 20; 13, 69. 11g sabyữ - instr, } infin. IO, I4. III—The present stem in - j'a-, 'passive': indic. 16 X, 11 pers. pat., 5 inanim. pat.; 13 or. dir.; 3 X ag. Most of the instances are intrans, : 2,7 kimartham duhitā me 'dya nätisvastheva laksyate; 20 na dřśyante ("on ne les voit pas"); 5; 5; 8, 15; II, 9 "to be visible”; 19, 26; 7, 9 jīyate sma "lost the game", Dutch "verloor het”; 9, 29 vidyate "es gibt"; 13, 40; 17, 5; 15, 5 buddhir dhriyate me + loc., Dutch "zich richten op"; 18, 26 jñāyate "it is known". With an ag. : (two instances are intrans. accompanied by an instr. expressive of "by influence of...") 8, 14 nrpatih puşkarenaiva jīyate; 13, 50 jarena kliśyate ; 10, 26 sauhrdenāvakrsyaie (clearly 'passive'); 9, 32 the 'agentless pass.” seems to have been used for the sake of indulgence: yadi māiil no cihātum ihecchasi, tat kimartham vidarbhāņām panthāh samupadiśyate. Imperfect: 7 X, 4 pers. pat., 3 inanim. pat.; 2 or. dir.; o ag. Intrans.: 7, 18 ajīyata ; 9, 4 vyadiryateva hydayam "burst"; 19; 3 ; 14, 12 antaradhīyata "disappeared”; 26; 15, 13 sa vai kenacid arthena taya ... vyayujyata "parted from her"; 17, 3 na präjñāyata karhicit "there has never been news of him". The imper. (5 inst.) is a courteous form: 2, 9 anubhūyatām ayanı vīrāḥ svayamvaraḥ; 4, 12; neutral: 12, 69 asyatām; 4, 17 tatha bhadre vidhiyatām; 7, 13. The pass. future is found 14, 22 (intr.) śreyasā yokşyase. We shall now go through the indicatives in -ya- as they occur in the story of Sāvitri (Mbh. 3, a. 293-299): indic. pres. 14; 9 pers., 4 inanim., I neutr.; 2 ag. Intrans. 4, 8 laksyate; 5, 14 śruyate; 5, 75; 78 drśyate "is visible"; 84 samtapyete "are afflicted”; 6, 13 dhriyate "is alive"; Fr. on: 2, 13 ucyate "he is called”; 26 kanyā pradīyate ; 5, 21 me nīyate bhartā; cp. also 2, 28 (2 X); agentless passive for the sake of indulgence 5, 54 tvayaiva datto hriyate ca; with instr. ag. 2, 20 (in a recapitulation); 5, 41 ucyase budhaih (cf. Nala 8, 14 etc.). ... Re ·60 REMARKS ON THE SANSKRIT PASSIVE .. Rāmāyana: the texts in Boehtlingk, Chr.3 and 3, 11-15 87. The verbal adjective. Attrib.: 33 cases (I X ati-). Subst.: 2. Predicate: without as- or bhu-, trans., without ag.: 18 X, 4 pers. pat., 12 inanim. pat., 2 animal, expressing the actual past or actuality in general in or. dir.: B. 107, 13; 19; 108, 3; 17; 109, 26; 112, 29; 113, 5 + i. caus. aham caikesunā hatah ; 18; II, 43; 12, II; 12; 13, 8; the v. adj. is a real adj. 15, 5;13, 19 viviktaḥ ... punyo ramyas tathaiva ca; + instr. rei 15, 12. With an instr. ag. : 26 X, 7 pers., 15 inan., 4 n., expressing the actual past in or. dir.: B. 107, 12; 14; 112, 24 isunābhihatah kena; 24. kasya vāpakrtam mnayā; 113, 6 (-actual result) kena sma nihatāḥ sarve; 15; 16; II, 31; 53; 54; 15, 28 (2 X); 13, 16 vijñātam... mayā "is known to me"; in a narrative: II, 15 sqq. sarvadevair niyojitāḥ "apsarasah apsarobhiḥ ... munih munih ... nito madanavaśyatvam ... tāś ca ... muneh patnītvam āgatāḥ. tatāke nirmitam grham. tatraivāpsarasah ramayanti (durat. pres.); 61; 67; II, 60; 18; cf. 12, 33; the v. adj. expresses a condition or a state: II, 83 yadā prabhrti cākrāntā dig iyam punyakarmaṇā; the plqpf. 12, 7 yathoktam lakṣmanenaiva ; 11. With a gen. ag.: 13, 15 vidito ... mama "is known to me". The v. adj. is intrans. 23 X, 14 P., 8 inan., I neutr., expressing the actual past in or. dir.: B. 104, 13; 109, 20; 112, 23; II, 9; 49; 62; 12, 2; 3; 10; 30; 13, 3; expressing the past in a narrative: B. 112, 2; II, 13 pravyathitāḥ (sc. abhūvan Comm.); 17; 66; 92; 14, 30 I jātāh (II jajñire); in the poet's narrative: II, 68 astam gataḥ sūryah; 15, 24; the v. adj. is a real adj. II, 78 ksānta-; 83 nirvairāh praśāntā”; with an instr. expressing the person by whose influence... 15, 29 cf. II, 48 jñātāḥ (: jñāyamānā vartante Comm.). The v. adj. is trans. and accompanied by as- or bhū-; 8 X, 7 P., I inanim.: asmi: the v. adj. is practically an adj.: 112, I anūdhā tvam abhavah; 13, I; 15, 28 prīto 'smi: 13, 10 dhanyo 'smy anugrhīto 'smi; with a pred. adj. 107, 13 aphalo 'smi krtah; + fut. 107, 4 pūtā bhavisyasi; + opt. 109, 26 tarpitāś ca syuh ; 25 bhavet. Intrans.: 8 X, pers. subj., Ist pers. sg. or plur.: 7; actual past: 11, 72 sukham asmy usito niśām; 94; 15, 2; the v. adj. is a real adj. 106, 15 paritusto 'smi; 13, I; cf. 14, 35 samnato 'bhavat; the periphr. constr. is resultative 114, 8 sūdrāyām asmi vaišyena jātah (= ... putro 'smi). Locat. absol, intrans.: 8 cases, e.g. 105, 12 abhyudite sürye; 108, 2 pranaşte tu pasau; with one member intrans. 11, 40 prabhāte; P. IIO, 17 sadasyānumate ; loc. abs. trans. 113, 17. Aec. + part. constr.: II, 94 nivedaya mām prāptam (act. 87. Ed. Pansikar*, Bombay 1930.

·TEXTS 61 ... past); 12, 9; 15, 26 etc. (= 7 inst.) $ 85-90 times a verb. adj. is used to express attendant circumstances, an appos. condition; it is very oiten accompanied by an instr. ag. 11, 61 agastyena tadă devaih prärthitena ... bhaksitaḥ sa; 64; 70 pratigrhītas tu muninā... nyavasat; 73 iti tenokio jaguma; 76 pādapān hastihastair vimrditān sakunisanghaiś ca ... pratināditân; 93 ; 12, 32; 13, 9; 15, 6 etc.; 106, 27 gautamenaitan uktasya; an instr. causae etc. II, 79; 12, 34; 13, 3; 14, 16 etc.; - abi. II, 18 tapoyogān munin vauvanom asthitam; su-: 15, 22 parnain suparicchaditân; parama-: II, 46 paramasamhrstah. I found only one adverb. Verb. adj. in compounds : iic. 49, ifc. 70, imc. 12. Gerundires acting as pred. without an agens and without as- or bhu-: 13 X, 4 pers., 8 inanim., I neutr.: 104, 16; III, 10; II; 13; 15; 31; 31; 32 na bhetaz'yam; 12, 30; 30; 13, 7; 7; 18; with an agens: III, 7; 9; 17 (3 X inanim.), neutr. ag. 13, 21 gantavyam nyagrodham api gacchatu; ger. subst. 7; ger. app. 3; ifc. 2; imc. 3. Constr. avikreyam sutanı āha 108, 23; 29; 30. The present stem in -ya-. Indicatives: 19, 4 pers. subj., 14 inanim. subj., I neutr. Intrans.: 105, 8 laksjate; 16:26; 110, 25 abhidhiyate; + i. instr. 112, 26 śastrena vadho vidhiyate; 113, 31 kliśyate; 11, 7 dyśyate (or Fr. on); 47; 50 drsyante ; 50; 51 ; 79; 15, II; I4; II, 19 śrlīvate ; 13, 21 älaksvate; + instr. ag.: II, 82 drsyate nopabhujyate (pass.);111, 5 supyate ... śramakhinnena "es wird ...” (“impersonal constr.'). Imper.: 9 X, I pers., 2 inan., 6 n., intr. or without an ag. ; 114, 6 apanīyatām; II, 9 kathyatām; 40 gamyatām; 73; 12, II; 4 nivedyatām ; 12 praveśyatām; 26 asyatām; 15, 4 drśyatām. Other passive forms are rare: part. pr. attrib. I X, appos. 3 X : intr. or 'agentless pass.' 113, 24; + ag. 15, 31 sitayā ... anväsyamāno nyavasat; II, 23. Perfect: 112, 5 intr. dadyśire; II, 7 śuśruve (2 X inan.). In the chapters 1-6, 9, 13 of the 5th book of the same epic the v. adj. is rather rarely used as an attrib., e.g. 9, 1 ayatam; 51 supta-. Pred., without as- or bhū-, trans., agentless: infrequent, expressing the actual past, e.g. I, 123; 189; the same, with an agens (28 X, 16 P., II inan.), expressing the actual past in or. dir.: 1, 107; 124; 133; 134; 189; 3, 30 (neutr.) ; 45; 13, 3; 5; 5; 11; 21; 53; 62; a condition that, having been brought about as a result of an action, still lasts: 1, 87 aham ... sägarena vivardhitaḥ; 90; 105; 119; 143 mama bhakşyah pradişpas tvam īśvaraih; cf. 3, 46; the past: 1, 144 vara eșa purā datta mama dhätrā; 151 ; 146 (in a narrative); in a recapitulation in the poet's narrative: 9,69 tāḥ sarvā rāvanena hrtaḥ striyah; 70. Instr.

·62 REMARKS ON THE SANSKRIT PASSIVE 7 > caus. I, 118. Intrans. : actual past: 1, 108; 3, 23; 48; 9, 42; narrative past: I, 144; adject. 9, 73 višiştā gunato hi sītā; + instr. caus. 9, 32, etc. V. adj. + as- or bhū-, trans. : the v. adj. is practically an adj.: 1, 123 prito 'smi; act. p. 176 samāksipto 'smi sahasā; imper.: 13, 47 kāmam astu hrta sētā; fut.: 1, 174 bhavisyāny aham āśitā; 13, 12 i. ag. bhakṣitā sā bhavisyati; neutr. 47; 1, 58 the v. adj. is an appos.: samdhyayā samabhisprstam yathā syāt süryamandalam. Intr.: adj. 1, 130; 132; 136 babhūvāvasthitaḥ; 158; 3, 15 prasiddheyam bhaved bhūmih; 13, 7. Acc. C. part. constr. I, 188 etc. Gerundives: neutr. I, 106 pratikartavyam; 124; 151; 13, 16; + i. ag. 1, 114 tasmāt tvam pūjaniyo me; 2, 32; 3, 47; 11. 1, 88. na śakya- + inf. + i. ag.: 2, 25 na hi yuddhena vai lankā sakyā jetum surair api; 31; 3, 14; neutr.: 2, 41; 3, 24 na śakyam ... lankā praveștum ... tvayā; 36 etc.; 3, 29 na śakyam mām avajñāya praveştum nagarīm imam. Ger. + as- + gen. ag. : I, 120 mānyo 'si maina. Present-stem in -ya-. Indic.: intrans.: 1, 63 drśyate ; 72; 80; 2, 27; 13, 5; I, 66 laksyate ; 2, 27 abhigamyate; + i. ag. I, 22 bhidyate 'yan girir bhutaih [; + i. caus. 15, 49 caturbhiḥ tapyate; intr. 20, 15 cakşur mama nibaddhyate: baddham jāyate C.; 27 upalabhyate: samartho bhavati C.]. Imperf.: intr. I, 50 vyavasīryanta; 53 adrśyata. Imper.: I, 105 gamyatām; 109; 123 manyur eşopaniyatām. Perf.: intr.: 1, 32 dadyśe garudeneva hriyamāņo mahoragaḥ; 54; 55. Saundarānanda of Ašvaghoșa, c. 1-6, 10, (14, 15). I-The verbal adjective acts as an attributive 4, 27; 28; 39 etc. ; 3, 23 nava ivodito ravih "like the newly-risen sun" (19 X); the negation a- is prefixed 4 X; the v. adj. is attended by an instr. ag. I, 39 tatra tajjñair upākhyātān avāpur mahato nidhin; an instr. rei 3, 24 pradipta iva sandhyaya ghanaḥ; the v. adj. ifc. 15 X. 15 X the v. adj. is used as a subst., e.g. 4, 5 cestita- (I include 2, 10 jāto vā vipule kule etc.), the same, compounded with a- 5 X ; + instr.: 5, 48 mayoktam "what I have told you". 8 X it acts as an adverb. Acting as predicate the v. adj. is = 19 X trans., 15 X intrans. 24 X as- etc. are not added: 7 X the agens is expressed: 6, 16 priyeņa drstā niyatan tato 'nya ... mām agamad vihāya : actuality, op. 23; 41; 10, 55; cp. 61 na dattam anyena "no other can give this" (Johnston); at the same time expressing a state or quality: 10, 16 kā cărutarā matā te; 6, 40 (act., qual.); without an agens 5, 16 ('actual state', the author addresses his readers) samkleśapakso dvividhaś ca distah, ep. 14, 23; expressive of a property: 6, 39; 5, 38;

·TEXTS 63 a state in the past: I, 24 tasmād iksvākuvamśyās te bhuvi śäkyä iti smytāḥ; intrans. 4, 30 bhagavān praviștaș (actual.) ; 6, 15 priyasya zlīnam hrdayam viraktamı “has become estranged", cp. 10, 16d; 48; 5, 8; 2. 3f. vikräntah etc. (quality). Ifc. 2, 3.9 X as- etc. are added: the form expresses a quality: 6, 6 nirurtah syāt; 5, 41 sa bhavet pramattah ; 4, 15 bhavec ca rustā; 4, 42; periphrast.: 4, 45 katham nu grāto na gurur bhavet; actuality: 4, 19 katham kyto 'si; the agens is expressed: 4, 6 systāv abhūtāın it'a bhutadhātrā "they seemed as if produced by the Creator..."; 6, 29 dhrtah priyenā vain abhūt "it was held by ..."; 2, 45 an instr. soc. is found: evanādibhir atyakto babhu- 2'A. 23 X a person is subject, I do X an inanimate thing. There are 14 instances of a loc. abs., 4 times it follows the predicate or the verb. The acc. c. part. constr. occurs 1, 37; 2, 20; 21; 26; 6, 3. Aśv. has also used these forms to vary his style, e.g. Bc. 4, 16. The v. adj. expressing attendant circumstances of the main action and serving as an apposition to the subject. There are 55 instances of the type 2, 25 sthitaḥ kārtayuge dharmne ... näsrasat, besides 6 X with instr. ag., e.g. I, 31 tataḥ sa tair anugatah pariciksepa, 3 X with instr. rei (causae etc.), e.g. 1, 38 man yunā taptāḥ ... nišasvasuh, 12 X at least ifc., e.g. 2, 7, where iva is added śiśyire ... pitur ankagata iva ; with a- 3, 4 anadhigataḥ. iva also 1, 13; 16; 3, 34; 1, 12 virejur krtā iva. Several times the v. adj. is used as an apposition to an acc. (e.g. I, 37), then too, it may be attended by an instr. (e.g. 2, 6 yah pūrvai rājabhir yātān yiyāsur dharmapaddhatim). V. adj. as adverb: 5, 33 dhrstam. The ppp. is very often found ii c.: in cantos 1-4, 10 = 32 X, cp. also in such compounds as 1, 35 vyāyatābaddhavāsasah; 42 spastāſīcitamahapatham, etc.; very frequently too imc.: in the same cantos I 28 X. Most cases of a ppp. ifc. are already mentioned; in canto 4 e.g. 7, 8 (4 X), 18, 22, 23, 31, 33. II—The gerundives are not numerous: this form is attrib. 3, 7 padam śivam ahāryam; substant. 5, 42 vadhya-; (1, 3 śisya-, 2, 27 and 33 bhrtya-); predicate: 5, 27 etat trayam khalv avaśena sevyam; 29 soko niyatam nişevyah; 42 socyah; na śakya- 10, 60; 14, 48; the ger. appears ifc. 3, 8 krtakāryam; i m c. 5, 17 prayatnaneyabuddhih; iic. 2, 45- III--Personal forms in -ya- are rather numerous: cantos 1-6, 10, 14, 15: indic. pr. without an ag. 20, for the greater part intrans., II X a person, 10 X a thing is subject; most intransitives have an intelligible syntactical function: 5, 17 nirmucyate "obtains salvation"; 14, 44 duhkhān na vimucyate; 14, 10 sajyante "attach themselves”; 15 ·64 REMARKS ON THE SANSKRIT PASSIVE ... and no vidhīyate "is intended for”; 43 smytir vasya na vidyate "who lacks attention”; 15, 44; 38 dyśyate (2 X); compare also 14, 47 sa kşanj'ate carann ivorvyān bahukantakāyān “is hurt like a man walking on very thorny ground"; 15, 16 dahyate; 10, 53; 15, 47 gato yatra na badhyate "by going along which one can avoid affliction" (the subject has not been expressed); I, 58 śrüyante "are called”. Elsewhere, however, the use of the passive voice is due to a tendency to an artificial manner of writing which exaggerates the application of devices also found in natural speech. Cp. I, 15 yatra sma mīyate brahma kaiścit kaiścin na mīyate kāle niinīyate solilo na cākāle pramīyate "there some contemplated the Absolute, no one did hurt (another), soma was produced (measured out) at the proper time, one died untimely" (mi- "perceive”, nii-"injure”, mā- "measure", pra-mi- "die"; the passive constr. is chosen in order to obtain a jingle). A tendency to variation is met with 14, 12 rathākṣo 'bhyajyate yathā; but II yatha dhatte ... vrani, 13 yatha ... khādetānı; qp. also 14, 8 and 9, in 9, however, the subj. of the pass. sentence has been emphasized: tasmăd duşyati nāhāro vikalpo 'tra tu vārgate "therefore to take food is not wrong, but a choice of foods is prohibited". An instr. rei is added to a passive form expressing Fr. "on": 15, 16 dustena ceha manasa badhyate vā paro na vā "one may or may not cause hurt to another by ..."; 10, 52 rägāgninänena tathābhidahye. Opt. + instr. rei 10, 52; 15, 48 Germ. es, Dutch er + pass, might be a correct translation: Jatra klešair vidahyate "es brennt"; the same constr. in order to vary the style 14, 18 socata ... dījate satrave yathā, cp. 16 active voice, etc. 5, 31, where the ag. is added, the patiens is emphasized or, rather, it is the thema of the sentence. yathausadhair hastagataih savidyo na daśyate kaścana pannagena tathānapekso ... na d. śokabhujangamena, we too usually say: "hij is door een slang gebeten, he is bitten by a snake”, and not "een slang heeft hem gebeten", unless the snake must be the thema. The constr. na sakyate + inf. occurs 15, 61. So, 4 times the ag. is added. The pass. aorist is found 9 X in the verses 2, 40-44 which are constructed in an artificial manner, cp. 42 na tenābhedi māryādā kāmād tena satsu api bhogeşu nāsevīndriyavrttitā "he never offended against the rules ..., he did not indulge his senses"; the last time only the agens has been suppressed. It is well known that Aśvaghoşa has a special liking for series of stanzas in which the same verbal category constantly occurs. The pass. perfect is used I, II Suśruve yatra nisvanah: Fr. on, "where the sounds were to be heard"; intr. 2, 8 dadrse; II dadyśire; 4, 3 .

·TEXTS 65 inababhāse trividhena namnā "she bore three names" (2, 8 etc. pers. subj.). In these cantos I have found no form of the pass. imper. 88. The participle in -ja- occurs 14 X, as an attrib. + ag. 5, 4 hadbhil sampūjya mānāya tathāgatāya kartum praņānam na šašāka nandah ; constr. Roma capta: I, II agninām hūyanānānām ... urist'anar "the sound of the turning of oblations in the fires, the s. of the o. burning in the f."; 5, 41 intr. attrib. sarvato vesnani dahyamane "a house everywhere in flames”; 8 X as an appos., after the verb only 6, 36 dahramānāh (the waiting women) and 4, 440, without an agens but clearly 'passive': 5; 40 ārobyamāṇasya tam eva mārgam bhrastasya sārthād iva sārthikasya; 42; with a personal ag. 6, 49 juratijanena säntuyamānā; pers. subj. and inanimate cause or agens 6, 10 kamena kopena ca dah yamānā; 33; 10, 41 manorathena jehriyamārah; 4, 44 the form is used thrice in an artificial way: sa kamarāgena nigrhyamaño dharmānurāgena ca krsyamāṇaḥ | jagama duhkhena nivartyamānah ... Loc. absol. 5, 52 pass.' without an ag. pravasyamānesu siroruheșul. So, 3 X the 'patiens' is inanimate. Meghadüta of Kalidāsa. The v. adj. acts 18-19 X as pred.: 12-13 trans., 6 intrans., 4 a pers. subj., 13-14 an inanimate subj., I neutr., 2 X the verbs as- or bhūare added, 4 X the v. adj. is attended by an instr. caus. or instr., 5-6 X by an instr. ag. This form is expressive of a condition, a quality, a state etc.: 16 tvayy avattam krsiphalaın, 75 raktāśokaḥ kesaraś ca ... pratyāsannau + gen.; 58; of a condition brought about by a preceding action : 71 ārambhaḥ vibhramair ... siddhah; cp. 105. An interesting instance is 72 yasyopante ... kāntayā vardhito me ... Onamito balamandāravykşah, app. expressing att. circ. or particulars or pred. expressing a condition resultant on an action which has not yet lost its actuality? The v. adj. expresses actuality: 6; 30; 65 nītā at the beginning, vividly; 85; 90; 104; cp. 105 cetah ... krtam aśaranam + instr. caus.; 108d; 110; elsewhere the v. adj. denotes a past fact which is distinctly present to the mind of the person speaking: 108 I tvani api kim api rudati viprabuddhā; c kathitam prcchataś ca tvayā me; 88c; 27d vañcito 'si expresses ‘future actuality': "you will be deceived": an instance of the 'fait-accompli-Darstellung' which made Spitzer 89 remark that the person speaking "unter dem 88 It is used in counteous speech to an inferior: Bc. 3, 47 the person speaking is included; 62, to a superior: 4 31; variation: 46 ff., impers. 92. 9 L. Spitzer, Stilstudier. I, p. 258. See also my paper "De Javaanse denvormen', Bijdragen tot de Taal-, Land- en Volkenkunde, vol. 103, p. 283 sqq. 89 ·66 REMARKS ON THE SANSKRIT PASSIVE Eindruck seiner Erregung mit seiner Phantasie vorauseilt, erst zu erwartende oder mögliche Ereignisse antizipiert, die Grenzen zwischen Gedachtem und Geschehenem einreisst und Definitives, Unausweichliches ausdrückt wo er in Wirklichkeit noch weit vom fait-accompli entfernt ist". The v. adj. is found in a relative sentence: 85; 88; 104. Loc. absol.: 10 instances, 6 intr., 4 trans., 6 inanimate subj., 3 animate subj., I neutr. The v. adj. is used as an apposition etc.: 20 instances : 10 trans., 8 intr., 2 ifc., 13 X animate subject (the cloud included), 6 X inan., I X an animal; an instr. rei occurs 73, 76, 93, an instr. ag. 76; a verb. adj. with the prefix a- 98, an intrans. v. adj. followed by san 26 viśrāntaḥ san vraja ...; an appos. to an acc. 8, 55, 89, to an acc., i f c. 89, to a gen., ifc. 67, constr. jātam varse jānāni tvām. The v. adj. follows the verb or pred. 98, 110, 73, 76. The adv. drutam is found 22. The v. adj. is used ifc. attrib. 6, 93; i i c. 55 etc. I have found 6 gerundives: 5 trans., I intrans.; instr. ag. 29, 30, 85; 7 gantavyā te; 34 sthātavyam te (neutr.), instr. rei 77; 5 X an inanimate patiens, 1 X neutr.; moreover, gejam (subst.) 83. The ger. is expressive of the future or a possibility 77 laksanair laksanīyam bhavanam. Besides, a ger, acts as an appos. 100 locanānăni agam yah; if c. 2, 70, 74 prekşanīya-. There are 4 pers. forms in -g'a-: 70 and 102 with instr. rei asrais tāvan drstir ālipyate me; 2 X instr, ag. : 56 (where the pass. voice seems to be due to a tendency to vary the style) tripuravijayo giyate kimnaribhiḥ, cp. also 70 mārgah sūcyate and 104 alingyante ... mayā te. The part. pr. pass. is met with 5 times; always as an appos. + instr. ag. or instr. rei, the patiens is an inanimate object or the cloud. Raghuvamsa of Kalidāsa, c. I, 3, 5, 15, 16, 19. I–The verbal adjective. Used as an attrib. e.g. 5, 39 āptah dütah; 40 rddhām rajadhānīm ; 62 janaḥ sametaḥ; 63 °þuruşaiḥ pranataih etc. Used as predicate: 48 instances; 24 trans., 24 intrans., 19 an anim. subj., 29 an inanim. subj. (or animal); as-or bhū- added 4 X, 5 X with instr. rei, 14 X with instr. ag. This form being practically a predicative adjective expresses a condition, a quality or property : I, 28 dvesyo 'pi sammataḥ śiştah "a correct man, though unfriendly, was acceptable”; 60; 3, 40; 49; 15, 72; 18 nätiparyāptam; it represents a tenseless verbal form, a timeless present': 5, 71 yāvat ... ākramate na bhānur ... tävad arunena tamo nirastam; it expresses ·TEXTS 67 an actuality, the actual past 16, 76 ( 2 X); 5, 66 rāfrir gatā; 15, 43; 16, 14; 63 ; 10; 1, 87 (the v. adj. is placed at the beginning) ; 3, 48; 5; 14; 24; it expresses the past: I, 12 tadanvaye ... prasūtah suddhimattarah ; 14 sthitaḥ ... krāntvā; 27; 36; 3, 40; 61 (in a vivid description); 15, 46; 76 upasthitah ("on vit alors ... saluer" Renou); ‘dramatical present 3, 45 sa cet svayam ... tvam antarāyo bhavasi cyuto vidhih "but if you yourself stand in the way..., all ceremonies must be at an end!". With an instr. ag.: actuality etc. 3, 50; 63 (intensifying?) na me tvad anyena visodham ayudham; 5, 4 tvayā jñānam ... āptam "you have acquired"; 21; 33; 66; 16, 76; the past: 1,78 sa śāpo na tvayā ... śrutah (here there is also actuality); 16, 19; the past: 5, 39; 16, 52; the remote past: 19, 3. With an instr. rei: 5, II trpta- (condition); 16, 59 II (act. past); I (in a relat. sent., remote past); 19, 7 and 56 (narrative past). With as- it expresses the act. past.: 5, 11 prāpto 'si "you have come"; 15, 18 tvam asi dhātrā ... upapăditaḥ; the past in a communication: 5, 20 mayā maharşir vijñāpito 'bhūt, and, 1, 75, after purā: p. āsīt ... asritā. When expressing the actual past (actuality) the v. adj. finds itself in an oratio directa. When there is an agens, the ag. as a rule precedes the v. adj., while the pat. may be found everywhere; 16, 19, however, which is a rel. sent., the ag. is put after the v. adj. The loc. abs. is met with 1, 87 after the verb, 5, 71 before the verb, 3, 18 and 16, 87 bef. + ag., 15, 82 it is 'impersonal (neutral) tag. The acc. c. v. adj. construction is found 5, 29 vrstim śaśamsuh patitām; 16, 24; 80 etc. When acting as an apposition etc., the trans. v. adj. (+ 50 inst.), is very often attended by an instr. ag. (21 X) or another instr. (causae etc., 12 X); there are 24 intrans. Particulars are of little interest. Adverbs are scarce. There are a good many instances of a PPP. iic., imc., ifc. II-Gerundives: 17 occurrences; attrib. 4, pred. without as- or bhū-6, pred. with as-or bhū- 5 (I, 16; 16; 16;3,6; 16; 15, 43), constr. amocyam aśvam yadi manyase 3, 65, app. I, II; gen. ag. 1, 11; 3,6; 5, 50; instr. ag. 5, 18;7 X the 'patiens' is a person, 7 X an inanimate thing, 2 X an animal. Besides, rather often iic., i mc., ifc. I, 20; subst. I, 28; 5, 34. III–There are 26 occurrences of pers. forms of the pres. in -ya- (20 indic., 4 imperf., 2 imper.); indic.: an inanimate thing is ‘patiens' 4 X a person, 8 X an instr. ag. is added, 5 X another instr. The indic. form expresses an intrans.: I, 10 hemnah samlaksyate hy agnau vi- ·... 68 REMARKS ON THE SANSKRIT PASSIVE śuddhih “appears, comes to light"; 16, 7; 1, 70 duyase "to endure affliction"; 16, 21; 1, 27 mucye "to become free" (mukto bhavāmi, Mall.); 5, 69 samsrjyate sarasijaih "comes in contact with lotusflowers" (samgacchate, Mall.); + part. praes. Pass. 16, 62 sīryamāṇaḥ sanılaksyate; a fr. on-form 19, 49 nivāryate; a passive: sometimes the author emphasizes the patiens: 19, 24 maya ... bhāgyam api tasya kāń ksyate (kāńksati no manah, Mall.); the same 3, 44 makhārśabhājām prathamo manīşibhis tvain eva devendra sadā nigadyase, but the use of the pass. produces also rhyme; also 16, 152; 12 the road mentioned in the first half of the stanza is thema of the sentence which fills up the second half: sa vāhyate ... śivābhiḥ; 3, 62 padam hi sarvatra gunair vidhīyate (adage, perhaps the passive voice emphasizes the universality) “virtues set footstep everywhere”; sometimes the active might be expected: 1, 67 payah pūrvaih upabhujyate; 5, 9 (probably on account of a tendency to variation); 16, 19 is an intrans. attended by an instr. expressing the agens: vānarais tāḥ kliśyante. The -ya-form of the causative represents a passive: 5, 5 (+ instr. caus.); 8 (the same, why used?); 16, 20 the -ya-form accompanied by an instr. may be considered as passive: tiraskriyante krmitantujālaih gavāksäh; cp. I, 61 mantraih ... pratyādiśanta iva ... sarah. The imperf. is intrans. : 3, 2 alaksyata; 24 paryacīyata “increased” (vav;dhe, Mall.); 34 adyśyata; 15, 77 may be considered as an 'agentless pass.': anvamīyata śuddheti "it was inferred...". The imper. is a courteous form: 3, 65 yujyatānı; 66 vidhīyatār. The part. in -ya- is used as an appos. to an acc. I, 39, to a gen. I, 69; 15, 27; to an abl. 16, 79, to a nom. I, 38; 16, 27; 58; 64; 5 it is attended by an instr., 5 an inanimate 'patiens', 3 a person. It forms part of the predicate 16, 62 siryamanah samlakşyate na ... hāraḥ. It appears i i c. 3, 7; 59. I have found no aorists, but there is a rather considerable number of perfects: the pas- (why pass.?); 3, 35; 19, 16 (artificial, why pass.?); 18 tena ... suśruve ... vacah (artificial); intrans. 5, 27 na jaghne "met no impediment" + abl. "by virtue of ..."; 3, 40 dadyśe "appeared"; 15, 54 samyuyuje + instr. soc. "joined" (samgataḥ, Mall.), the form must be considered as passive because the present is samyujyate, e.g. Manu 9, 22; pipriye (19, 30) is intrans.; 3, 20 svayam mumuce ... bandhanāt, here Mallinātha remarks karma kartari lit "reflexive perf.". Kirätärjuniya of Bhārazi, c. 7-9. ::I-The verbal adjectives. In these 175 stanzas the attributives are ·TEXTS 69 rather rare: 7, 4; 13; 19; 28; 33; 8, 9; 23; 34; 36; 9, 22; 46; 48; 53; 75; 78; if c. 9, 6 saudhajālapatitā ravibhāsah; + instr. 9, 9. Substantivized v. adj. are very numerous: 7, 5 kāntā-; 10; 15; 39 brohitāni; 8, 8 dayita-; 14; 50; 29 gataiḥ; 46 īhitaih; 9, 17 cũrna°; 20 dhvántam; 37; 37 skhalitan; 38; 41; 46; 50 sitkrtāni; 52; 58; 59; 62; 67 vrīditasya (: vridāyāḥ Mall.) ; 68; 71 viviktam; 71; 73. The words kāntā- and dāritā- however, occur many times. An abstract noun derived from a ppp. 8, 17; 9, 64. The v. adjs. are rarely used as predicate of a nominal sentence: 9, 15 in a series of short sentences: raſījitā nu vividhās taruśailā nāinitam nu gaganam sthagitam. rue puritā nu z’işameșu dharitrī samhytā nu kakubhas timirena; in the same way 25; the v. adjs. describe a state of affairs in the past as a part of a narrative; 37 it expresses the narrative past: kāntaześnia ... Jātan ... ramanibhih; 49; 66 vridayā ... na sthitam na vigatam hrdayeșu (an artificial turn). na yuktam 40. In the whole of the poem this form occurs 78 X (22 X in the narrative; 56 X or. dir.; 4 X as-or bhu-). The loc, abs. is found 8, 29; 38 etc. (+ 6 instances). The acc. part. constr. occurs 7,23 (+ iva). There are about 60 occurrences of the v. adj. used to express attendant circumstances etc.; they are often accompanied by an instr. ag. Adverbs are rare: 7, 4; 9, 58; 63. There is an exceedingly high number of v. adj. iic.; 145, as a poet of kāvya the author has a special liking for bahuvrīhi-compounds; imc. $ 55, ifc. + 35. Not rarely a v. adj. i ic. is accompanied by an instr., e.g. 9, 37 II manmathena pariluptamatīnām prāyaśaḥ skhalitam apy upakāri. Gerundives are not frequent: pred. 9, 30 asahyam; 39 anuneyah (short sentences); (+ asti 13, 7; neutr. 13, 46); appos. 8, 34; attrib. 48 sakhișu nirvācyam samślesam; subst. 7, 29 sevya- (: prabhu- Mall.); 39 vacanīyam; a -tā-stem built on a ger.: 8, 39 socanīyatā-; 45 vilokanīyátā-; 47 tulyatā-; 9, 78; ifc. 8, 4 karapraceya-, cf. 9, 78a. The pass. Present-stem is rarely met with: in a short sentence expressing an opinion of the author or stating a truth 9, 13 langhyate na khalu kälaniyogah; + instr. ag. 58 prāpyate gunavatäpi guņänām vyaktam āśrayavaśena višeşaḥ; stating a fact, intr. 13, 39 dyśyate ; 42 smaryate gilt bei den ... Männern", Cappeller ; 48 ucyate (svagunah) "one mentions his own merits", the subj. is known; neutral, with a periphrasis: 54 bhavatā tapasyate; the agens remains in the background: 58 näbhiyoktum anrtam tvam isyase (viz, asmābhih); 62 gīyate + "es gibt ..." (C.); so, the indic. pr. pass. is more frequent in oratio directa; the same + sma and instr., narrative past: 9, 14 nīyate sma; 53; 7, 6. The imp. pass. is used for the sake of varying ... nrbhih" ·70 REMARKS ON THE SANSKRIT PASSIVE the style 8, 8 jahihi kopam dayito 'nugamyatām; 9, 39 (; 13, 70, cf. 71; not in this manner 12, 36). The part. in -ya-, attended by an instr. ag. or caus., is found 8, 13; 55; 9, 20; intrans. 9, 31. In the narrative portions of this text the pass. perfect is extremely frequent: in c. 7-9 there are 31 instances, amongst which 28 are accompanied by an instr. ag., which is, however, often an inanimate object conceived as a person; 2 X a person, 29 X an inanimate thing is patiens. The chief reasons why this form has been employed seem to be the desire to vary the style of writing and the custom of the poets of kāvya to show their learning. Generally, perf. act. or med., present-forms with sma alternate with these passive perfects. More than once highly unnatural turns of expression are the outcome of the author's wish to combine the use of the narrative perfect with the avoidance of monotony and series of stanzas constructed upon the same scheme. See 7, 2 nāsede caritaguņatvam ātapatrai”; 4; 7 gandharvaiḥ ... pratīye ... vicitrata vidhātuh; 17; 22 mayüraih dhranir upasuśruve ; 34 nācene ... tāri väranena; 37; 8, 4 puro 'bhisasre... ºjanaiḥ; 6 yastih ... dadrse vadhūjanaiſ; 16 runude kaścij jaghanena kāntayā; 18; 30 āpaḥ ... ºjanaih prapedire; 32; 40; 43; 9, 6; 10 samdhyayānuvidadhe Omaitri; 12; 29; 30; 32; 34; 36; 40; 48; 54; 56 ādade ... utpalaih ... kampah; 58. Intrans. or Fr. on: 9, 12 upalebhe "es war zu merken"; 42; 63 dadyśe. [13, 51 there is an instance of the aor. pass. bhavatā nirūsi ma ... krtajñată, which represents a quite uncommon expression; 54 the fut. präpsyate phalam tvajā]. In the whole of the poem 1or instances of the perf. pass. are to be found 90. The other middle forms are not used in a passive sense. Kathāsaritsāgara 75, 21 sqq.; 76; 77; 83; 86; 87; 88; 98; 99, 1-41 (= Vetālapancavimšatikā Introd., Ith, 2d, 3d, gth, 12th, 13th, 14th, 24th, 25th story). The verbal adjective does not often act as an attribute : 86, 62 tām dastāın divyakanyakām; 93 samrddhe 'pi pure; 87, 31 hrdayaiḥ sphuțitaih; 98, 8 milita- etc. (+ 14 cases, besides + 8 ifc., e.g. 86, 66 tvaduktena pathā; 96 svepsitāın, aiganām; 75, 118 hastiruddhena pathä; 76, 21 pathitasiddhasya mantrasya). Subst. are rare: 75, 123 istam; 12, 103 āptād; cf. also 137. Verbal adjectives serving as predicates : without as- or bhu-, trans., agentless: 9 instances, always in or. dir.: actual past: 75, 47 rājan kytah prasādaś cet; 81a nyastam yad . 90 See C. Cappeller, Bhāravi's poem Kirātārjuniya (H.O.S., 15, 1912), p. 180 ff.

·TEXTS 71 ut palam karne (tayā, the ag. is found 8od and 8ıb); 82; 105; 137; 77, 33; timeless: 76, 39 sa suto matah; 77, 42 ayam muktaś cauraih (instr. soc.), etc. With an agens: (62, 20 P., 38 th., 4 n.) actual past in or. dir.; 75, 28 and 29 mayā yāni nityam phalāni te || haste dattāni tāni kva sthāpitāni sada tvayā; 30; 80; 80; 81; 82; 83; 98; 98; 100; 101; 104; III; 113; 114; (114); 123; 135; 136; 136; 137; 148; 165; 169; 170; 174; 176; 177; 180; 76, 35; 40; 77, 34; 71; 83, 47; 59; 61; 61; 75; 82; 88, 35; 98, 31; 68; 73; 99, 6; 32; the narrative past: 77, 50 sā ca tena dattā vanik putrāya ; 83, 6; 45; 56; 79; 13, 51; the remote past: 75, 140 Śaśamsa ca kathāmadhye tat tasmai yat ... ° kathanam krtam tena"... that he had made mention of ..."; 77, 40; adject. (timeless) 86, 19 dharno hi prepsitaḥ sa me; + gen. 75, 86 viditam mama; 98, 72 va tasya vāñchita. Intrans. (25 instances, 21 P., 4 th., always in or. dir.): actual past in or. dir.: 75, 46 eso ham āgato bhikso; 98; 102; III; 76, 4; 77, 60; 61; 86, 83; 107; 87, 11; 98, 43; the narrative past: 86, 142; 143; adject. (timeless) 75, 48 tasminn ullambitamstah ko'by ekah puruṣaḥ sthitaḥ“finds himself”; 113 rajasvalā niśās tisrah sthitāham "I find myself in my monthly periods”; 132; 12, 33; 160; 161; 87, II; 99, 23; 23; the past 75, 193; + instr. caus. 98, 66 tadāścaryena dhairyena tusto 'ham. A verb. adj. + as- or bhū- and instr. ag. is found 8 X (7 pers., 7 or. dir.): act. past: 75, 94 dhātrī tasyāsmi sthāpitā tena; 77, 33; 65; 87, 14; 99, 22; the narrative past 98, 8 unmūlito 'bhūt ... rāstrabhedibhih; cf. 99, 33 ; with a gen.: the past: 99, 21 bhikṣor āsīd yad ipsitam. Without an (instr.) ag.: 86, 67 sāpito 'si mamāsubhih; 98, 33, cf. Pet. Wtb., 7, 61 "in dieser Bedeutung nur das partic. pass. sāpita- zu belegen"; 99, 25 sarvāh khyātā bhavantu. An intrans. v. adj. + as- or bhū-: (13 pers., 7 or. dir.): the actual past: 76, 20; 77, 34 präptāsmiha; 86, IOI ; 87, 22; 88, 59; 98, 42; the narrative past: 77, 63 gatvā bāhyam pravištābhūd udyānam tatrāpaśyac ca ... ; 95 skandhāt sa tasya ... nasto 'bhūt; 76, 17 pravrtto 'bhūt praroditum; the remote past in a narrative: 76, 27 dadarśa cātra tam ... yah sa ... gato 'bhavat. The v. adj. is a timeless adject.: 86, 134 asit kruddhas tu mām prati; 88, 2 śrānto 'si; 46. Locat. abs. with intrans. v. adj. are, comparatively speaking, numerous: $ 20 cases; the pers. subj. is a little more frequent than the inan. subj., locat. abs. with trans. v. adj. are rare: 77, 82; 75, 191 hamsair jagdhesu śālisu; 196, (cf. 77, 78 appos.); an impers. loc. abs. + instr. is not unfrequent, but only udite and ukte are met with : 75, 39 karisyāmīty ukte tena mahābhuja; 77, 46 iti śārikayodite.

·72 REMARKS ON THE SANSKRIT PASSIVE In the acc. c. part. constr. (e.g. 12, 86, but not very frequent) the acc. is sometimes accompanied by an instr. ag.: 75, 182 upāyam mantripretrena tam sambhāvya tathā kytam ; 185 tām kravyādair bhakṣitām sutām matvā; 86, 140; cf. also 77, 64, where the instr. ag. must be connected with an appos. 'Impersonal' v. adj.: 86, 15 videśam te gantuni yuktam; 87, 21 iha sthitai iha snātam ... vihrtam ca tayātreti; the same with an instr. ag.: 88, 58 caurena ... ruditam hasitam ca kim tena; 59; 75, 83. Adverbs are rare, e.g. 86, 45; 87, 27 etc. Among the verbal adjectives which express attendant circumstances or act as appositions etc. (e.g. 75, 31; 76, 108) there are a good many that are accompanied by an instr. ag.: 75, 67 so 'nvito mantriputrena; 75; 80 ity ukto rājaputrena (a very usual formula); 87; 125; 126; 127 dāsibhih ... rajjūtkşiptaḥ; 166; 77, 7; 32; 56; 57; 98, 14; 99, 25 etc.; by an instr. rei: 86, 42; 73; 89; 99, 3 etc.; an abl. 14, 56 cauryan nivyttah. Gerundives (24 instances, 12 P., 10 th., 2 n.). Predic.: 77, 83; 83, 33 kasmai deyā; 36; 38; with instr. ag.: 75, 138 maya samacāraḥ kartavyah; 75, 190; 83, 18; 86, 14; 65; 66 nivõranīyo nāham te nānugamyaś ca; 98, 70; 86, 80; neutral or 'impersonal': 86, 146 saptāhāni na gantavyam prasīda sumukhi tvayā; 75, 38; with gen. ag. : 75, 177 sā... vikreyā mama vartate ; 159; (subst.) 87, 23 aprapyam nāma nehāsti dhīrasya; with bhū- or as-: 99, 27 pūjyä ... bhavisyati; 75, 112 + i. ag.; appos. 88, 45; na śakya-: 86, 45; 79; loc. abs. with a ger. 98, 8 pradeyāyānı ca tasyām. The present-stem in -ya-. The indic. without instr. ag. occurs + 17 X (8 p.): intrans.: 75, 47 vidyate "there is"; 86, 112; 76, 40 ucyate ; 86, 50 drśyate; 159 dyśyase "you look"; 83, 36 tat kvopayujyate "is useful"; 86, 168 naitat ... yujyate; to avoid a form of the 1st pers. sg.: 77, 44 kathoparodhataḥ śāntanı avācyam api kathyate. The agens, left unexpressed, is "they", Dutch ze, viz. they that are charged with a certain task, they that usually perform a certain action etc.: 88, 35 vadhāya niyate yo 'yam. The -ya- form expresses a certain vagueness: 75, 79 yasya na nāma ... avabudhyate sä katham präpyate... ; 83, 35 Śūdrāya ... ksatriya diyate katham "how is it possible to give...” Fr. on 86, 21, The -ya- form is accompanied by an instr. ag.: 88, 12 laksyante te ca nāsmabhih; 87, 22 kim ātmaivam hanyate tvaya (is the pat. emphasized?); 88, 12 the -ya- form lays stress upon the condition resultant on the action of the agens nityam musyāmahe cauraih "we constantly fall victims to thieves", the citizens do not know who are the thieves; 17 I after having seen a certain man the king says ayam ·TEXTS 7 sa nanam cauro me muşnāti; cp. also 99, 7 parārtho yena sádhyate; 8 Imperf. intr. 87, 50 prāņair vyayujyata; "agentless passive' 88, 14 purī sānusvataiva ca; the constr. of the sentence (the patiens is thema) has exerted influence: 77, 41 tatas ... tatra nirbhayaḥ pravista! śvašurābhyām sa harşād drstvābhyanasıd yata. The original intransitives are used in the same way: 98, 46 dīvate hrdayam hi nah; 86, IOS tathārajyata să rājižā vacobhiḥ. Imper. pass. : 76, 29; 77, 27; 47; 85 83, 25; 26; 35. In general, these imperatives are more or less courteous once we find an instr. ag. added, a turn which appears to be ceremonious: 5, 28 tad eșā tanayā rājanis tvayā mahyam vitīryatām. Partic pres, pass. app.: 76, 18; 77, 78; 88, 33 ("while being led away" Dutch "terwijl men (ze) hem wegbracht(en)”); instr. ag. : 75, 181 87, 20 stajanaih sāntv'yamano 'pi; 44; 56; 88, II; 40; 41; acc. c. part constr. 75, 36 sāhāyyam kriyamānam tvayārthaye. Fut. pass. 99, 25 g'atraisā kīrtayisyate "where this (work) will be recited”. Aor. pass. instr.: 75, 185 vyapādi sokena; 77, 29; 76, 6 ajani; 83, 12 + instr. ag., pass.: 83, 2 tena ... abhyadhāyi sah (the subj. of this sentence is the same as in I); still more artificial is 13 ity anasigaratir nämnā pitrā tena ziyadhāyi sā. Perf. pass., intr.: 75, 152 janārataḥ Śltśrut'e ; 87, 31 dadyśire; instr. ag. : 75, 68 tena ... kanyā dadyśe (variation: 67 apaśyat, 70 jahāra); 167 tathābhītaś ca jagyhe sa drstvā purarak sibhih (sa is thema, and by putting t. at the head of the line, the author dwells upon the thema); 168 ninye ca (the same constr. is continued); cp. 77, 57 II; 80; why pass. 77, 42 mahotsatas tena cakre? Total amount of the personal forms and the p. pr.: 32 P., 20 inan., 2 anim. III-CONCLUSIONS AND ADDITIONAL REMARKS We now come to the following discussions and conclusions. First the term passive is, in my opinion, not adequate 91. As to the -ya-present, in Vedic prose this category is in most cases, and often almost exclusively, used in a way which for the sake of brevity I have hitherto called 'intransitive': the verb is used without an object, and it would be impossible to add such an element: ricyáte, mucyáte etc. The num- 91 Apart from the fact that the term 'passive' should not suggest the notior. of suffering or enduring, I refer to E. Schwyzer, Zum persönlichen Agens beim Passiv Abh. Preuss. Akad., 19.42, IO, p. 10. GONDA, The Sanskrit Passive E ·74 REMARKS ON THE SANSKRIT PASSIVE ber of roots forming a -ya-present is rather restricted. Nearly always the processes denoted by these forms bear upon movements, experiences, transformations with reference to a state or condition which the subject experiences, reactions of the subject etc. These processes may affect the subject in the core, in the very essence of his personality or only concern it in its periphery; "es geht etwas mit dem Subjekt vor" 92. Some scholars are perhaps inclined to speak of 'medial sense'. The use of this term, however, involves risk 93 : the formal category called 'middle' voice' has no separate, well-defined, uniform, clean-cut idea and function of its own 94. But the use of the middle endings, prescribed for the passive' by the grammarians, is not a matter of chance 95. Nor will the term 'impersonal do, because we usually mean by it a verb which is always put in the 'third person'96, granted that a considerable number of the -ya-presents, which are nearly related to the -ya-forms, may easily be translated by impersonal constructions: třsyati: Old Dutch mij dorst etc. I feel inclined to avoid the term 'reflexive', which some people will use in special cases, because we may safely speak of a reflexive verb only in those cases in which the object of the verb is a reflexive pronoun. Moreover, the starting-point of the process expressed by the -ya-form, is, as a rule, not the subject: we have to be mindful of the difference between such expressions as the Eng. I pride myself; the Fr. je me baigne, which are reflexives, and Skt. avašisyate "he remains, he is left” 97. As a rule, the starting-point of the process is not expressed. One might, perhaps, suppose that this starting-point is the 'verbal notion (Verbalbegriff) which is neither expressed nor present in the author's mind: such a supposition would call into being a hypothetical parallel 92 Cf. also A. Niederstenbruch, Das Verhältnis von Passiv, on und passivisch gebrauchtem Reflexiv im Französischen, Diss. Bonn, 1927, p. 13. 93 I refer to O. Jespersen, The philosophy of grammar, P. 168. 94 See also Joh. Wilde, Die passivischen und medialen Ausdrucksweisen objektiven Geschehens, Diss. Leipzig 1913, D. 36 ff. 95 Yet, the so-called passive, even in the earlier language, occasionally assumes active endings, in the epics, where several forms may be explained as active forms of the fourth class (see also A. Holtzmann, Grammatisches aus dem Mahābha ta, Leipzig 1884, $ 774), these endings are not rare, in the Prākrits they are normal. There seems to be no reason to doubt that they occurred from the beginning (as one might say), leaning on the intransitive verbs. 98 However strange it may be to an unsophisticated mind to be taught so; see Jespersen, O.C., p. 212. 97 I refer also to J. Vendryes, Le langage, p. 122; J. Wackernagel, Vorlesungen über Syntax, I12 (1928), p. 90 ff.

·CONCLUSIONS AND ADDITIONAL REMARKS 75 to an active turn of the rujati rogah-type ("the disease afflicts”, see Pān. 2, 3, 54). Suffice it to say that in many cases the starting-point is altogether out of the question and does not come into the mind of the author; the process is not brought on by an agens; but it occurs 'automatically', it happens, it comes to pass, that is all. In many modern languages, e.g. English, Dutch etc. an active verb must often be used: arašisyate "he remains; hij blijft (achter, over)", in other languages the so-called objective middle voice 98 may serve as an equivalent: Lat. labor, nascitur, videor. The use of the -ya-forms under discussion has been considered by Meillet 99 to be the "vrai rôle du passif”: “La possibilité d'employer les verbes actifs absolument et le fait que les désinences moyennes soulignent au besoin cette valeur absolue ont dispensé l'indo-européen commun d'avoir un passif. On croit souvent que le passif est une forme du verbe où le sujet du verbe est indiqué comme subissant une action exercée par un agent: Paul est battu par Pierre; les expressions de ce genre se rencontrent en effet; mais ce sont des tours souvent artificiels, et en tout cas relativement rares. Le vrai rôle du passif est d'exprimer le procès la où l'agent n'est pas considéré. Le latin a dicit "il dit”, quand on pense à quelqu'un qui parle, et, à côté, dicitur "il est dit", pour signifier "on dit"....Si le passif n'était qu'un renversement de l'expression active, il serait au fond superflu. Ce qui donne au passif son utilité, c'est que, au lieu de présenter le procès comme résultant de l'intervention d'un agent, il le présente en lui-même, sans aucune notion étrangère". I concur in Meillet's cpinion, on the understanding that by 'passive' are meant the formal categories which are usually called 'passive' and that this function does not necessarily belong to a 'passive': it is often expressed by other categories, e.g. by an intransitive active 100. The category under discussion has been called by Meyer-Lübke 101 'Vorgangspassivum'. Wistrand 102 speaks of 'Sekundärhandlung': arbor cadit as distinct from agricola caedit arborem and arbor caeditur ab 98 See Wilde, O.C., D. 37. 99 A. Meillet, Linguistique historique et linguistique générale (1921), p. 195 f. 100 E. Wistrand, Über das Passivum (Göteborgs kungl. Vet.- och Vitt.-Samh. Handl., 6, A I (1941)), p. 9 has given expression to the same criticism, as I see after having formulated my views. 101 W. Meyer-Lübke, Die neueren Sprachen, 6. Beiheft (1925), p. 158 ff. 102 Wistrand, o.C, D. 13.

·76 REMARKS ON THE SANSKRIT PASSIVE agricola. I would suggest the term 'eventive'. At all events, the term ‘passive' has, for clearness' sake, to be avoided 103. In the other cases -ya- is added to roots denoting processes that are involuntarily supposed to be brought about by an agens, p.e. hanyáte "is killed”. It is quite immaterial whether the agens is expressed or even known. In Dutch the same shade of meaning may be expressed by turns containing the particle er: er wordt gebeld: "there is a ring, there is the front door bell”: when there is a ring, we tacitly admit the presence or existence of somebody who has rung; but, as a rule, we do not know who is the agens; sometimes we leave the author out of account or avoid mentioning him. Thus, Sanskrit kriyatc may be translated by "er wordt gemaakt, er wordt gebouwd" (cf., p.e., er wordt een huis gebouwd). In French the expression for the so-called 'generic person' 104 on may often be used to translate kriyáte, hanyáte etc.; it is worth mentioning that the on-turn is often chosen when one wants to emphasize the durative aspect of the verb, which is likewise expressed by the -ya-verbs 105. In a book which has been published in 1941 Wistrand 106 has tried to establish the essential characteristics of the passive voice': "passive ist eine Verbform, wenn sie eine Handlung bezeichnet, die nicht vom Satzsubjekte ausgeführt wird sondern das Subjekt trifft, sodass das grammatische Subjekt in Wirklichkeit Objekt ist. Bei passiven Verba kann das logische Subjekt in Form eines Agens hinzugefügt werden, aber notwendig ist dies nicht .... Die Ursache dazu, dass ein Agens nicht ausgesetzt werden darf, kann nur eine sein, nämlich dass die Verbalhandlung als vom grammatischen Subjekt ausgehend aufgefasst wird, oder jedenfalls als ihren Mittelpunkt im Subjekt habend in solcher Weise, dass die Verbalhandlung von keinem anderen Ausgangspunkt herrühren kann." I am under the impression that this argument will not do. In adopting the views of Paul 107 Wistrand is in the wrong -- he had done better to profit by the criticism of Jespersen, though this is, in general, too negative. I do not believe that in the 103 Alice Berger (who is a pupil of Gamillsсheg and Wechssler), Der Ausdruck der passivischen Idee im Altfranzösischen, Diss. Berlin 1934, p. 2 says: “Passivische Idee ist also: die Auffassung oder Vorstellung eines Geschehens der Wirklichkeit als Handlung, die an einem Seienden vollzogen wird und deren Urheber vorstellbar sein muss". 104 See Jespersen, 0.c., p. 204; 215 ff. 105 See A. J. Fehr, Medium-passivum, Amsterdam 1937, p. 23. 106 Wistrand, O.C., P. 12. 197 H. Paul, Prinzipien der Sprachgeschichte+ (1909), p. 278.

·CONCLUSIONS AND ADDITIONAL REMARKS 77 types of sentences under discussion the “Verbalhandlung (I would prefer the process) von keinem anderen Ausgangspunkt herrühren kann". Neither in a so-called 'eventive turn, nor in the hanyáte type does the starting-point of the process lie in the subject. In the first case the starting-point is left out of account. In the hanyáte type it finds itself outside the subject: sometimes one wants to keep the agens in the background, sometimes one wants to suppress any word denoting the agens or one is not able to mention it; be that as it may, the subject is affected by a process, of which the starting-point is found outside the subject itself. As distinct from the Dutch de bel gaat which is (or: may be) eventive (cf. de kraan loopt, de pijp lekt etc.), the turn er wordt gebeld is used when the speaker wants to denote a process, which is brought about by an agens, though this agens is N.N., "one", Dutch men, French on (on sonne). Perhaps the difference between Lat. occidit and occīditur may be described in the same way. Thus occidit = Skt. náśyati, ksiyate ; occīditur = Skt. hanyáte. It must, however, be granted, that cases are not rare in which the person using such forms as hanyáte, kriyáte does not knowingly think of a definite agens; if so, even the existence of an agens may be quite immaterial to him. It often depends on the idea expressed by the verb whether a -vá-present must be interpreted as an 'eventive' or as an 'agentless passive'. The 'eventive' use, which, to a greater or less extent, prevails in Vedic prose, and the 'agentless passive' which in the ancient languages is well-known too, are also often to be found in younger texts 108. In the Vedic metrical texts the eventive is far from being rarely used, in the Mahābhārata the so-called passive present is in the first case an 'eventive'; in Kāvya this use is found to a greater (e.g. Asvaghosa) or less extent. This force came to be felt as passive only by expressing an agens and by contrast with active forms of transitive meaning. To my mind, in a real passive turn the agens is expressed. Consequently, in this particular Vedic prose is on a par with Arabic, Lettish etc., which do not allow the addition of an agens to the 'passive' 109. It is true that in the prose of the Brāhmanas the -ya-forms may be attended by an instr. denoting accompanying circumstances 108 See also L. Renou, Grammaire sanscrite, § 342. 109 Cf. Wistrand, o.C., D. 12: "Wenn aber eine Sprache wie das Lettische oder Arabische das Hinzufügen eines Agens gar nicht erlaubt, kann das nur so gedeutet werden, dass diese Sprache kein voll ausgebildetes Passivum besitzt".

·78 REMARKS ON THE SANSKRIT PASSIVE or the cause by which the action happens. But I have not found a real instr. agentis in the parts of Taitt. S., Śat. Br., and Jaim. Br. which I examined. In order to form an idea of the development of the 'complete passive turn' in post-Vedic literature I have compiled a statistical record; the numbers figuring in it are, however, like other numbers to follow, correct only by approximation: Pausya episode -ya-present without an agens expressed 2: complete pass. turn 1; Kālid. Sak. I and VI prose: 18:4; Daśakum. II, III, VIII: II : 3; Hitop. (20 pages) : 17 : 5. In ślokas: Mahābh. Sāv.: 13 : 3; Nala (I-XIX): 13: 3; Bhagavadg. = 68 : II; Mbh. V, a. 186-192: 4:0; in a number of texts selected from the Rām. the first turn is approximately 7 times as frequent as the complete pass. construction; Manu II: 32 : 0; Kūrmap. Iśvarag. 18 : 4. Consequently, in these texts the complete turn is found only in a minority of cases. In Kāvya texts the instr. ag. is used on a larger scale: Asv. Saund. I-VI, X: 20 : 4; Kāl. Ragh. (6 cantos) 12 : 8; selections from Bhār. Kirāt.: 8:3; sel. from Siśup. 6: 7. It is an interesting fact that the whole of the Rgveda beside 191 -ya-forms without an agens expressed, contains about 15 examples of the complete turn. We nay, in my opinion, easily conclude that, with regard to the use of the complete pass. constr., the stylistical factor is of greater moment than that of chronology. In the Vedic period the complete construction was really known, but it was, in a general sense, not used unless in a certain class of literary works, particularly in verse originating in environments where the art of making poetry followed a traditional technical practice and where poets consciously forged their stanzas and coined the expressions they used to make their work fit for the special ends they had in view. The fact that the complete passive turn is found in the Rgveda which is generally admitted to be the oldest collection of Indian literary productions, cannot, on the other hand, preclude us from considering the agentless construction or (let us be guarded in our statements !) the state of affairs found in the ancient prose texts and in the traditional language of the epic śloka, to be more original; in the first place because the complete construction, on the whole, is not much liked normal speech 110, and secondly because of the presumable origin of the -ya-present. It is quite certain that the -ya-present had this foriginal function already in the Indo-Iranian period: it occurs in both Avestan 111 and 110 See above p. 4 ff., 104 H Reichelt, Awestisches Elementarbuch (1909), S 614-616.

·CONCLUSIONS AND ADDITIONAL REMARKS 79 ... Ancient Persian 112. Cf. G. Y. Av. bairreinte (Y. 32, 15): Ved. + bhriyate. According to Reichelt, in the Avesta 38 examples are found (of which some are uncertain), the verb zan- "be born" (us.zayata etc.) included, which corresponds with Skt. járate IV 113. Sometimes the Arestan -'a-present expresses an erent (Vorgang): G. Sruyè “I obtain hearing" (Y. 33: 7); Y. pairyete "es wird verglichen"; the 'agentless passive' is, comparatively speaking, often found: Y. kiryeiti (a Ved. + kriyáte) "is executed etc."; Y. avi Xeairyeite "is eaten" V. 2, 26); etc. The middle present and the -ya-forms of the verb vaz- may be used intransitively: “to drive, go or travel by any vehicle; fly; succeed etc." (cf. Lat. chor), now, Yt. 14, 43 ads: yat ... vaštåyho almya nõit zasyånte fatůyho almya nõit janrånte, which has been translated by Bartholomae 114 : „wenn ...wenn die erfolgreichen nicht weiter (zum endgiltigen Sieg) kommen, die geschlagenen keine (entgiltige) Niederlage erleiden” and bij Lommel 115 ... die vordringenden(?) dabei nicht vorankommen(?) und die geschlagenen dabei nicht unterliegen": the verbs are real eventives. Now and then the -3a-present is accompanied by an instr. denoting the means: V. 3, 9 yat ... paiti ... daxma uzdaēza kirveinte "where burial-places are made by means of) raising up (i.e. by heaping up earth)"; V. 4, 50 ngay haēnūiš karatăiš asdibīš paiti ava.karabyāt "with iron knives he must be cut down to the bones" 116; Yt. 8, II etc. yedi zī nā anašyāka ...yasna vazayanta pala ang'e gazatánho vasna yazinti; 23 etc.; an instr. denoting the person by whose influence, assistance etc. the process takes place 117: Yt. I, 29 spantayå ārmatöiš dõibrābya avāstryata mairyo "durch die Augen (Blicke) der heiligen Frommergebenheit wurde der Schurke niedergestreckt" 118. Yt. 13, 91 we find a loc. + paiti denoting the person by whom a holy text is heard: gara uštrahe ... vahmi paiti ... ašan.Syavo visruyata "Z. to whose ears came the ašəm word"; the verb visruyata has no passive function. According 112 A. Meillet-E. Benveniste, Grammaire du vieux-perse (1931), § 195. 113 See further on. 114 Chr. Bartholomae, Altiranisches Wörterbuch (1904), 1387. 115 H. Lommel, Die Yašt's des Awesta übersetzt usw. (1927), p. 140. 118 Instr. respectus; I do not agree with Bartholomae, Wtb., 454, n. 3. 117 The difference between this auctor and an agens is felt very well by the genoral public: the other day I heard two boys, about 15 years old, discussing the question, whether the cat had escaped because one of them had opened the door, or because the door was open through his fault or negligence, “if it nad not been for you she wouldn't have run off”. 118 Tamamat ·80 REMARKS ON THE SANSKRIT PASSIVE to Bartholomae and Reichelt 119 a gen, agentis is found Yt. 13, 50 kahe no ida nama āyairyāt kahe võ urva frāyezyāt "von wem wird nun unser Name gepriesen etc.", Lommel 120, however, translates: "Wessen Namen aus unserer Mitte wird er anrufen", in my opinion, the verb is an 'agentless pass.' in as much as the agens (= "who?" which has to be taken from the preceding sentence) has been suppressed. According to Meillet-Benveniste 121 there are, apart from the Old Pers. -ya-presents à valeur active', two forms denoting "l'état" 122 : maniya- (= Skt. mányate) “think”; mariya- (Skt. mriyáte) “to die”. "A ce dernier groupe se rattachent les formations de présent passif, dont on a deux exemples sûrs, d'une même racine: kariya (cf. Skt. kriyate) kunavaya ...". The first form, akariyantā (B. 3, 92) 123 is an 'agentless pass'. A propos of akunavayata Meillet-Benveniste remarks : "Cette formation nouvelle atteste à la fois la vitalité du passif et...", but I prefer to adopt the opinion expressed by Bartholomae 124 and Schaeder 125 who consider this form (B. I, 20; 24) as a middle present. Another ‘Vorstufe of a passive function is represented by Oahyāmahiy 126 (B. 1, 7) "our name is”, Dutch "wij heten”, Fr. "nous nous appelons", which form, beside which the active Batiy "announce, speak” is found several times, has an eventive sense. Consequently, in the ancient Iranian languages the eventive and agentless passive functions are in evidence. I have not found a passive turn containing an indisputable agens. It is, on the other hand, an interesting fact that there is a number of middle forms in the Avesta, which are without any doubt accompanied by an instr. agentis 127. Are we allowed to conclude from these facts that the complete construction arose at an earlier moment when the verb was a middle form? Or is 119 Bartholomae, O.C., 512, ?gar-, n. 2; Reichelt, O.C., p. 259, n. I. 120 Lommel, 0.c., p. 118. 121 Meillet-Benveniste, O.C., § 193 ff. 122 Ibidem, $ 194: "Deux thèmes en -ya- appartiennent au groupe des présents en -ya- qui expriment l'état". 123 See also W. Hinz, Altpersischer Wortschatz, Abh. K. Morgenl. (1942), P. 48. 124 See Bartholomae, in: Grundriss der iranischen Philologie, I, I (1895-1901), $ 145 and 327. 125 H. H. Schaeder, Beiträge zur iranischen Sprachgeschichte, in: Ungarische Jahrbücher, 15 (1935), p. 561. 120 See Meillet-Benveniste, O.C., $ 196, cf. also the first edition of this book (1915), $ 196. 127 I refer to Reichelt, O.C., $ 450.

·CONCLUSIONS AND ADDITIONAL REMARKS 81 it due to the incompleteness of the texts handed down to us that these things happen to be so? Now, when Reichelt 128 says: "Ein mit dem Instr. des Mittels (ai. karana) verwandter Gebrauch ist der als Agens (ai. karts) in der passivischen Konstruktion”, I would prefer to hold that the Indo- Iranian instr. ag. is a development from the instr. denoting the karana. Both uses of the instr. (and we may also include the instr. denoting the person by whose influence etc. a process comes to pass 129) express or explain the how of the process, expressed by the verb or verbal noun it depends upon. In Sanskrit persons, when being a karana or instrument, are likewise put in the instr.; a Sanskrit instr. does not only answer to Lat. a(b), but also to per. An inanimate object, when denoting 'an agens' is, on the other hand, put in the instr.: satena bandhitah "a debt of a hundred has committed him to prison" (Pān. 2, 3, 24). We may compare the use of the Slavic instrumental: "Der instr. bezeichnet bei den verba passiva, bei den verba reflexiva, die als passiva verwendet werden, und bei jenen verba intransitiva, die als passiva aufgefasst werden, das bewirkende, dieses mag eine person oder eine sache sein .... Am genauesten entspricht er dem griech. dat. in Sätzen wie ovne õßovios dovais BrosúetL" 130. Imperfects and optatives built on the -ya-stem are, comparatively speaking, rare. In the Jaim. Br., however, there are many examples, which is due to the subject-matter of this text. Daśakum. (3 chapters): 5 (ind. 8); the whole of the Rgveda 7 (indic. Í 206); Nala (I-XIX) 7 (indic. 16); Sāvitri 2 (indic. 16); Kālid. Ragh. 4 (indic. 20). As far as I know, the imperative does not occur in the parts of the ancient prose which have been examined; in the story of Pausya (Mbh.) there are 9 examples (imp. act.: 42), in the prose portions of Kālid. Śak. I and VI: 12 (act.: 78); in three chapters of the Daśakum.: 5 (act.: 20); in Pañcat. IV: 27 (act.: 77); Sukas. XXI-XXVII: 2 (act.: 42); Mbh. V, a. 186-192: 6 (act.: 33), selected portions from the Rām.: I (act. 36); Manu II: 1 (act.: 8); Rgveda: 10; Ašvagh., Bc. II-IV: 9 (act.: 13); Kāl. Ragh. (6 cantos): 2; Bhār., Kirāt. VII- IX: 4; Budhasv., Bșh. XIV, 1-XV, 107 (232 verses) : 25 (act.: 32). The instr. agentis is rarely added (Siśup., Bệhatkathām., Kathās.); in 128 Reichelt, o.C. 129 See e.g. Av. Y. 41, 4; 44, 3 (Reichelt, o.C., D. 235). 130 Fr. Miklosich, Vergleichende Grammatik der slavischen Sprachen, IV, Syntax (1868-1874), p. 704.

·82 REMARKS ON THE SANSKRIT PASSIVE many texts it is never found. As to the use of this imperative I refer to the handbooks 131. At this point I should like to make a circumstantial digression to discuss the 'original meaning of the element -ya- in Vedic and Sanskrit. When consulting the books and papers on comparative philology and Indian grammar it strikes us that the authors, generally speaking 132, agree in two axioms with regard to the Vedic and Sanskrit so-called passive present-stems in -ya-, viz. 1° from the point of view of Indo- European linguistics the '-ya-passive', like the other special passive formations in Greek, Latin etc., represent a younger development 133, 2° this formation and the radically accented ya-class (the 4th presentclass) belong together 134. As to the second point, Delbrück 135, while treating the 'yá-passive' as an addendum to the 4th pr.-class, remarked that "die innige Verwandtschaft mit dem Medium der ya-Classe sich nachweisen lässt" 136, and argued 137 that there existed no original difference between passive and medium. "Der Grund aber, warum 131 Speyer, S. S. § 10; idem, VSS. § 192; Renou, Gramm. Sanscr., § 293. 152 I pass over in silence the antiquated opinions of Bopp, Pott, Benfey etc., which are enumerated by K. Brugmann, Die arische passivbildung mit suffix -ya-..., to be found in H. Osthoff und K.B., Morphologische Untersuchungen auf dem Gebiete der indogerm. Sprachen I (1878), p. 187 ff. B.'s own view, expounded in this paper, has with good reason not been adopted by most other scholars. L. H. Gray, Foundations of Language (New-York 1939), p. 217 says that in all Indo-European there is only one distinctively passive formation, which is confined to Indo-Iranian and Armenian both apparently denominative formations. All other passives may be traced directly to the middle. 133 See e.g. K. Brugmann in Brugmann-Delbrück, Grundriss der vergleichenden Grammatik der indogermanischen Sprachen? II, 3 (1916), p. 703, especially § 622, 2; idem, Kurze vergleichende Grammatik (15193), S 691; 802 f.; A. Meillet, Introduction à l'étude comparative des langues indo-européennes (1937), p. 245. „Le verbe i.-e. ne comporte guère une formation à valeur passive" ; J. Wackernagel, Vorlesungen über Syntax, I?, (1926), p. 136; (cf. Hirt, Indogermanische Grammatik IV (1928), p. 133 f.); H. Krahe, Indogermanische Sprachwissenschaft (1943), P. III; J. Mansion, Esquisse d'une histoire de la langue sanscrite (1931), p. 75; J. B. Hofmann, in Stolz-Schmalz, Lateinische Grammatik", (1928), D. 542 f.; Ed. Schwyzer, Griechische Grammatik I (1939), p. 639; C. D. Buck, Comparative Grammar of Greek and Latin (Chicago 1942), p. 237 etc. 134 See e.g. A. A. Macdonell, Vedic Grammar (1910), p. 331; W. D. Whitney, A Sanskrit grammars (1924), esp. $8 761 and 774. 135 B, Delbrück, Das altindische Verbum (1874), p. 166, cf. p. 167, 2. 136 See also A. J. Eaton, The Atmanepada in Rigveda, Diss. Leipzig, 1884, D. 20. (This author has not formulated his views in the right way; his interpretations are not always correct). 13T Delbrück, ibidem, p. 168. . 1 ! I b ·CONCLUSIONS AND ADDITIONAL REMARKS 83 der allmählich entstehende Passivbegriff sich gerade an das Medium der j'a-Classe anlehnt, ist schon längst in dem Umstande gefunden, dass gerade diese Classe zum grössten Teil aus Intransitivis besteht, also aus solchen Verben, welche am wenigsten deutlich eine Handlung des Subjects ausdrücken". So, Delbrück by his way of imagining the development of the passive category found it was an 'Anlehnung' whereas other scholars consider it as a special use of the middle form of the 4th class: "c'est la spécialisation du présent en -J'a- dans sa forme morenne" 138 or as a development from that class 139 : "zum Passivum hat sich im Altindischen das Medium der 4. ind. Pr. kl. entwickelt”. Speyer, however, who emphasized the close connection of the two ya-groups preferred to admit one original conjugation in -yati -yate with intransitive function, whence both the 4th class and the 'passive' have sprung 140. This clever supposition of the Dutch scholar has not, it seems to me, been sufficiently considered in discussions of the Indian passive, and so we find, apart from the resignation, expressed by Brugmann 141, inter alia an ingenious, but, I think, not irrefutable theory of Diels' 142, which starts from an opposite view. According to the German scholar the 'yá-passive' is, "wenigstens vom vedischen Standpunkt aus, eine ganz für sich stehende Formenkategorie” 143; the arguments adduced to establish the connection between the -ya- and -ya-classes are not "irgendwie genügend". Either the two classes have no connection with each other or they are fundamentally identical ("im Grunde identisch"), but if we want to support this hypothesis, we are forced to have recourse to a "gänzlich ausserhalb aller Erfahrung liegenden Mittel", viz. the argument drawn from the accentuation. Setting aside this question, which, he thinks, is of no consequence, Diels then proceeds to show by some examples from Greek and Slavonic that the “ya-passive' is not an Indo-Iranian innovation, but an ancient formation : Ved. chidyáte = ("vielleicht ohne weiteres identisch" 144) Gr. OXITETAL, Ved. pacyáte = πέσσεται etc. Now, 138 L. Renou, Grammaire sanscrite (1930), p. 463. 139 In this way e.g. M. Leumann, Zur Stammbildung der Verben im Indischen, Indogerm. Forsch. 57, whose words (p. 229) I quote. J. S. Speyer, Sanskrit Syntax (1886), p. 240. 141 Brugmann, K. V. G. $ 691, p. 528. 14 P. Diels, Über das indogermanische Passivum, in Jahresbericht 91 (1913) der schlesischen Gesellschaft für vaterländische Cultur, IVb, p. I. 143 Diels, O.C., p. 2. 144 Diels, O.C., D. 7.

·REMARKS ON TUE SANSKRIT PASSIVE fris thesis is that the Indian passive forms of some transitive verbs existed already in the "voreinzelsprachliche Zeit”: *skhidyétai pass. by the side of skhinetti act. An attempt is made to explain the disappearance of this passive category in the European languages 145: "Als Mittel zum Ausdruck der Passivfunktion konnte sich diese Bildung... nicht behaupten, einfach deshalb, weil dort teils ... der charakteristische Akzentunterschicd unterging, teils ... die mediale Flexion". A restricted number of forms survived, beside which and from which an active (oxicw, Toow etc.) sprang, and here the ancient active paradigm was ousted by the passive' stem. It is almost needless to say that, from the stand-point of linguistic method, Diels' argument must be considered as rash. Wackernagel 146 soon detected a serious shortcoming: Diels did not account for the fact that TéCCETAL pass. occurs in younger texts than téace!, that oyibw and oxicoual appear in the same period etc. Besides, he unduly omits examining the connection of the -ya-class with the 4th present class, he does not make any inquiry into the real value of the forms under discussion etc. For all that, Hirt 147 has defended the theory by pointing at Skt. dhiyate: Osc. fizet "fiunt", Lat. fitur, which, however, even if this etymology were not a highly uncertain argument 148, is only a feeble attempt. Let us make a closer inquiry into the verbs which make up the two Indian ya-classes. By way of preliminary observation I wish to point out that in the 4th present-class the root, though accented, very often appears in its weak form 149. At first sight this peculiarity seems to indicate that the suffix was originally accented. But as a large number of the -ya-presents have, without any doubt, arisen at a late date 150, we may not emphasize this point too much, the more so, as in all probability a considerable part of these forms are denominatives. Now, it has been argued 151 that the accentuation as is well known the 145 Diels, O.C., P. 6. 146 Wackernagel, Sprachliche Untersuchungen zu Homer (1916), P. 133. 147 H. Hirt, Indogermanische Grammatik IV (1928), p. 133 f. 148 See e.g. A. Walde-J. B. Hofmann, Lateinisches etymologisches Wörterbuchu (1938), p. 505; Leumann, I. F. 57, p. 231 f.; A. Ernout et A. Meillet, Dictionnaire étymologique de la langue latine (1932), p. 311. 149 Cf. also Macdonell, O.C., $ 437; the roots are enumerated by Whitney, Die Wurzeln, Verbalformen und primären Stämme der Sanskrit-Sprache (1885), p. 218 f. Such forms as tányate, pácyate, are, of course, no 'Gegenbeispiele'. 150 See further on. 151 Meillet, Mém, de la Soc. de Ling. II, 309 f. 1 ·CONCLUSIONS AND ADDITIONAL REMARKS 85 Vedic denominatives are nearly always formed from a nominal stem to which -ya- was added - cannot prevent us from considering a large number of present stems of the 4th class as denominatives, because there is no evidence of the Vedic accent (-yé-) in the other I. E. languages; “l'accentuation du type skt. pásyati, yudhyati est d'accord avec celle des dénominatifs helléniques et balto-slaves et d'une partie des nominatifs sanskrits i mantrávate)...”. Meillet fails, howerer, to explain the origin of the j'á-accentuation of amitraráti "act like an enemy." etc. Nor does he account for the difference, with regard to the accent, between a pasyati "be active" 132 and pásyati "see" (which for him is likewise a denom.). Further, even if the accent of vådhyati "fight" might be elucidated by the rules of Greek accentuation, this would not explain the accent of the non-denominatives. Moreover, as the causatives which are accented on the a which precedes the element ja hare the I.E. 0-vowel of the root syllable, which proves that the root was already early accentless, it is, a priori, not a matter of course that the weak root vowel and the accentuation of the verbs of the 4th class bear, genetically speaking, no relation to each other. I venture, therefore, to suppose that the - zá-class and part of the - ya-class were identical in origin with the accent on the suffix; the more so as the 4th pr.-class consists largely of intransitive verbs. The accent occasionally fluctuates: in the Rgveda- and Atharvaveda-samhitās mucyáte is found accented on the root, in Brāhmana passages of the Taitt. Samh. lúpyate occurs beside lupyáte (AV.) etc. 153. So it will be wise to study both groups of -va-presents together. Like the Greek presents in -je--20-(vícw <I.E. nigaio, eyeipw < *dyspiw etc.) the Indo-Iranian -ya-presents are from the first thematic. It is, therefore, needless to dwell upon the intricate question as to the distinction between the stems in -ž-l-(i)żo- and those in -je-/-20, -eįe-/-ejo-, which, according to most scholars, existed in Indo-European 154, the present forms of the -7-class ending in -jo -Zsi -žti -112--te -zonti, those of the -je-l-30- class in -20 -zesi -jeti -jom- -jete -¿onti. Suffice it to quote Bruginann's remark 155: "Wo die Klassen äusserlich zusammengefallen sind, muss für eine grosse Anzahl von Präsentia um so mehr dahin gestellt bleiben, zu welcher von beiden Gruppen sie zu 152 A. A. Macdonell, Vedic Grammar (1910), $ 567e. 153 I refer to Macdonell, O.C., D. 331 (and n. 2); Delbrück, Altind. Synt., p. 267. 134 I refer to Brugmann, Grundriss, II, 3, p. 178 ff; Meillet, Introductions, P. 219 f.; Schwyzer, O.C., P. 713 f. (with further reference). 155 Brugmann, O.C., D. 182.

·86 REMARKS ON THE SANSKRIT PASSIVE ziehen sind, als vielleicht die Mehrzahl der in der geschichtlichen Zeit überlieferten Formen erst nach dem Ineinanderrinnen der beiden Bildungsklassen überhaupt geschaffen worden ist”. We have, however, to revert to this point in the following pages. The I.E. suffix -24-/-40-, from which comes, as to its form, I. Ir. -ya-, had no semantic value of its own, it was nothing but a derivative element; however, the presents formed with it generally were of an indeterminate aspect 156. It was used on a large scale, and, as it soon shared in the existing tendency of the thematic type of conjugation to spread at the expense of others 157, in several I. Eur. languages the original distribution of the two -jo-classes has been disturbed at an early period. Already in common I. Eur. this thematic group was about to become the predominating present formation and the number of 'younger forins', denominatives etc. is, generally speaking, extremely large 158. We shall recur to the productivity of the -ya-formation in Sanskrit. So we are not surprised at finding -jo-presents beside other formations derived from the same root: Gr. hom., poet., younger øñoow, SMTTW, 'break etc.": ényvuul; pów"roast": páxw; tetalow and Titelow "sneeze”: Totápvuual (c Lat. sterruo 159); teiva "stretch": TÁVOLKL "to be stretched", tavÚw (* cówoul) 160; Av. harazya- "discharge": haraza- etc. ; yuğya- "yoke etc.": yaog- yug-, yung- (@ Skt. yunákti, Lat. iungo etc.). In Vedic and Sanskrit such pairs are very frequent: ksubhyati -te "tremble": ksobhate; trasyati -te "be afraid": trasati -te etc. 161; see further on. Occasionally it is possible to suggest a hypothesis as to the analogical processes which gave rise to a -io-conjugation 162. The -1o-forms are, however, not always younger: Gr. Nitouch . 168 See especially B. Delbrück, Vergleichende Syntax der indogermanischen Sprachen II (= Grundriss?, IV) (1897), p. 26 ff. 157 See also P. Buning, De indogermaansche athematische conjugatie in het Slavisch, Diss. Leiden 1927, p. 39 ff. 158 See Brugmann, Grundriss II, 3, p. 204 ff. 159 This word has to be considered as a younger form which has taken the place of an original non-thematic formation. 160 Cf. also accouat "seize, grasp": traßov, außávo etc. (see Schwyzer, O.C., p. 698 and E. Boisacq, Dictionnaire étymologique de la langue grecque (1923), P. 553). 161 See also Lanman, Proc. Amer. Or. Soc. 1885, p. XXXVI (26 examples) ; Renou, Gramm. sanscrite, p. 435, 1. II. For a certain number of verbs this coexistence has been mentioned by Pānini 3, I, 70. 182 See, e.g., Brugmann, Grundriss, II, 3, P. 189: "(lat.) inquio -iunt (Cic. usw.) ist angeblich zu in-quit aus *en-squet griech. ÉVL-otte und in-seque hinzugebildet worden nach aio = *agio neben aüt." Cf. also Ernout et Meillet, O.C., 465.

·CONCLUSIONS AND ADDITIONAL REMARKS 87 "pray, beg' is a present formed, with shifted accent, from the aorist Atéchou (:looou! < *ATLOUCL). An -o-conjugation in one I. E. language may alternate with ocher present ciasses in other languages: Gr. TELLTEM.Ouevos "go round" ! : Téw, témou el "become, be”): Skt. carati etc., Lat. colo etc.; Gr.yébw "cacare”: Skt. hadati. The suffix -o-163 may occur in two languages or more: Gr. Baivo "go"; Lat. c'entio "come”: Skt. gacchati, gamati, Av. jasaiti, jamaiti; Gr. Ion. Oraiva (<*sten-jo) "sigh, groan", Ch. Si. stenje "groan etc.", Gr. Aeol. Tévvs! (: ctéve!, Bouzera! Hesych.), Skt. túnjati "roar": Gr. Hom. GTÉVW, A. Sax. stenan, Lit. stenù ; Gr. ortalow "gasp, quiver (of a dying fish", Lit. spirii (inf. spirii) “to kick” 164: Skt. spluráti "kick, quiver”, Lat. Sperno "drive away, scorn" etc. As the thematic jo-stenis represent a very productive formation, we generally cannot tell, whether or not the present formed from it existed already in the common Indo-European period'. Thus the younger Gr. eńcow (see p. 86) and Lit. réžiu "cut" do not prove I.E. urēg-jo. As Lit. jungiu "yoke" has taken the place of an original non-thematic present (cf. Skt. J'llrákti, Gr. Çeúywoul; Lat. iungo) 165 it does not show that Ved. etc. yujyáte belonged to the common I. Eur. language'. May we assume that *nigy-je- to wash oneself" is I. Eur., because we find in Greek viw "wash”, vibqual "to wash one's hands etc." and in epic Sanskrit nijyate (beside Ved. ninikta and the intens. nénekti) 166 ? Sometimes the thematic ¡o-present is the only one to be found: Gr. pela "dye", Skt. rájyati "be coloured, be red etc."; Skt. paśyati "see", Av. Spasyeiti, Lat. specio and (if <*OXETT QUCL Gr. OKÉTITOLOL 167. Even in these cases we must reckon with the possibility of derivation in post-common-I. Eur. times. But these instances, which might easily be multiplied, clearly show that already at an early period there existed a considerable number of -jo-presents beside other present-stems derived from the same root. We may consider this co-existence as the starting-point for the development in Sanskrit, which soon tended to combine a -ya- present and other stems to form one conjugation. 163 L.E. -Ž- 1 -i(i)ó- and -¿e-/-jo-. 164 See also R. Trautmann, Baltisch-Slavisches Wörterbuch (1923), p. 275 f. 165 See e.g. Brugmann, Grundriss, II, 3, p. 285; 382. 106 As to the cognate words in Celtic I to H. Osthoff, IF. 27, p. 177; R. Thurneysen, Handbuch des Altirischen I, p. 135; H. Pedersen, Vergleichende Grammatik der keltischen Sprachen I (1908), p. 108 f.; II (1913), p. 585. 167 Uncertain. See, e.g., Brugmann, 0.c. II, 3, P. 195; 365 f.; Ernout et Meillet, O.C., p. 921.

·88 REMARKS ON THE SANSKRIT PASSIVE With the -ž-/-(i)io-stems the root very often appears in its weak form: Ch. Slav, kypljo kypiši inf. kypěti “seethe, bubble", Skt. kúpyati "boil up, become agitated", Lat. cupio 168; Ch. Slav. minjo miniši inf. minēti "think, be of opinion", Got. munaip "intend", O. Ir. do-muin-iur "I believe”, Gr. raivouar (<*mn-jó-) "rage", Skt. mányate “think”. The other type, with full grade vowel in the radical syllable, is, however, not rare 169; Umbr. heriest “(Lat.) volet", Skt. háryati -te "desire etc.” 170; Lit. periù peréti “brood”. The original distribution of the radical vowels has admittedly been considerably disturbed. With the -ie-l-10-stems the root appears in its weak form in Gr. ortidíow, Lit. spirin; Gr. oxán "stir up" and in many other stems 171, in its strong form in Lit Overiż "open; shut”, Lat. operio "shut” (<Puerio), aperio "open" 172 Like the -ēi- and -ējo-formations, to which they are related, the -7-4-20-present stems are mainly intransitive 173: Skt. trsyati "be thirsty"; Gr. raivouar; but there are also transitives : Ch. Sl. viždo viděti "see", O. Sax. hebbiu 'have” etc. The other class consists of trans. (Lat. specio; sāgio "perceive acutely”) and intrans. verbs (Skt. śúsyati "dry, wither”, Ch. SI. Sášetb, words denoting a sound: Lit. krokiù "rattle” etc.). Many intransitives have active endings: Gr. xoipw; rxcipu "skip, frisk"; Quíw, Bów "to be inspired”; Lat. fio "become etc."; the Sanskrit occurrences will be dealt with further on. An active form may correspond to a middle present in another language: lat. salio "leap" : gr. boucl. Occasionally the same -jo-present could be used as a trans, and as an intrans. verb: Gr. xhala “lament, weep for, cry, wail"; Barra "throw; fall”; a “make sit, place; sit"; valu "settle etc.; dwell”. Similarly Ved. rísyati "be hurt, receive harm" and "harm, injure"; Ved. rádhyati "be or make subject to"; Ved. dámyati "be or make tame"; Ved. + búdhyate "wake, awake” and "notice, heed" etc. In the other cases the active voice expresses the transitive meaning, 168 Cf., however, also Ernout et Meillet, O.C., p. 235. 169 The etymological combinations, given by Brugmann, O.C., p. 182 ff. may not be relied upon in all respects. See also Meillet, Le slave commun (1924), p. 184. 170 See, however, Meillet, M.S.L., II, D. 306. 171 I refer to the enumeration of Greek presents, given by Schwyzer, O.C., D. 714. 172 Further particulars concerning the ablaut will be given in the following pages. 173 "...zeigen ganz vorzugsweise intransitive Bedeutung", Brugmann, Kurze vergl. Gramm. (1903), p. 527.

·CONCLUSIONS AND ADDITIONAL REMARKS 89 the middle forms the intransitive: Gr. Salu "kindle": falouat "blaze" ; xala "kindle”: xaloucu "burn". Sometimes the middle voice may, of course, express the 'medial notion': Gr. víla "wash off”: vicouar "wash off one's skin etc.”. Now, a restricted number of Greek middle -20-presents of intrans. or 'eventive meaning, beside which exists a trans. active, correspond with Indian -ya-forms of eventive meaning: Gr. hom. &Teo "to be bewildered, - distraught with grief” 174 (the active & W "strike with terror" does not occur until the Hellenistic period) 175: Ved. tujyáte "to flee frightened"; SETOL (e.g. Herodotus 2, 17 ( Nos) OXÍCETOLL "branches into three channels)", 8, 314, OTPUT!) cox[Lero): Ved. + chidyáte "get torn, get split, be broken"; TÉTTETX! "become cooked, - done" (cf. Arist. Eccl. 843 Tuoteva TéTTETXU): Ved. + pacyate 176. In the same way the 'agentless passive' Detvollot (Hom. 174 The identity of these verbs, admitted by Sonne (Kuhn's Zeitschr. 12, p. 297), has been denied by Uhlenbeck and other scholars (cf. also Boisacq, O.C., p. 99: "&- ambigu"; Walde-Pokorny, Etym. Wtb., II, 616 "vielleicht"). The meaning of the Indian verb has, however, been insufficiently recorded (skt. tuijóti tiját: 'frapper, pousser, blesser', (Boisacq), tujyáte is passed over). In Homer's poems drucouat means "to be bewildered" (cf. Z 41 &TUSÓLEVOL Qogéovto) and "flee bewildered" : cf. RV. I, 84, 17 ká işate tujyáte kó bibhāya, correctly translated by Geldner (Der Rigveda übers. u. erläutert, I (1923), p. 98): "Wer weicht, (wer) flieht, wer hat Furcht?”; ibid. I, II, 5 tvím devă ábibhjuşas tujjámānāsa āvisu? "to thee the gods have rendered assistance, frightened for him who does not fear" (according to Sāyana, ab. is an irregular nom. plur.), tujyámānāsaḥ corresponds with ÅTUCOLEVO!; ibid. I, 61, 14 the stem tujá- expresses the same meaning: dyávā ca bhúmā tujete "heaven and earth are frightened" (“fahren erschrocken zusammen" Geldner), but 6 this form means "to strike with terror": tujánn iśānas tujata (viz. vájrena) ("als er mit der losfahrenden (Keule auf ihn) losfuhr" Geldner), cf. also 9, 91, 4 tujată vadhéna ". coming on frightfully". But the present tuñj- (VII) has to be translated by "to make active, instigate, urge on etc.": RV. I, 131, 2 víśveşti hí två sávaneşzu tunjáte "at all Soma sacrifices they impel you (Indra)" ("in Eifer versetzen", Grassmann, Wtb. zum RV., 539); 9, 57, 2 tuſījāná ayudha: “discharging, hurling weapons" ; I, 105, 2 (and 9, 15, 3; 79, 5) it is used to express the notion of "pressing out (forcibly), causing to spout out": tunjäte vrşniyam páyah "ausspritzen" (Grassmann) (Geldner duly compares TS. 3, I, II, 8); 9, 87, 6 the same meaning is found with object rayin "wealth". Cf. also the meaning of the perfect-stem: 1, 143, 6 codáh kuvit tutujyât sätáye dhiyah ("to inspire thoughts"); the part. perf. med. is used to express the meaning "hastening, rushing on": 1, 3, 6 indrå vāhi tútujāna; 61, 12; 6, 29, 5 ("eifervoll"); 37, 5 etc. (of Indra); 7, 67, 6 and 84, 5 this part. means "eager for". 175 See also Wackernagel, Sprachl. Unters. zu Homer, p. 133. 176 I do not write an accent (according to Delbrück, Vergl. Syntax II, 37 the Greek word corresponds with pácyate, according to Diels, O.C. with pacyáte); see p. 98.

·90 REMARKS ON THE SANSKRIT PASSIVE 1 459) may be compared with hanyáte "meet one's death (by some one's hand)”. For obvious reasons we do not mention douuoooual "blow one's nose”: Ved. + inucyáte "become free, become rid of etc." on account of the middle sense of the Greek verb, Ch. Slav. znajo znati "YlyV6CXELV" etc.: Ved. + jñāyáte, because of the active voice of the Slavic forms. The active verbs, of transitive-ergative meaning, found beside these Indian -ya-forms, are non-thematic: tuj- VII (3 p. pl. tuñjánti, tuñjáte); chinátti VII; hánti II, pácati I excepted. Letting alone the question as to the chronology of the transitive active -joforms discussed by other scholars, we may conclude that already at an early period -1o-presents of intrans.-eventive meaning occurred beside trans.-active non-thematic present-forms 177. There are, in any case, some corresponding words in the Iranian languages and Sanskrit that cannot be ascribed to chance: G. Av. sruyê (Y. 33, 7) is an ‘agentless passive' (“auf dass man mich höre” Bartholomae) 178 "to obtain a hearing": Ved. + srüyáte beside Av. surunaoiti srunaoiti, Ved. śrnoti "to hear"; O. Pers. Oahyāmahiy "we are called" (Dutch "wij heten") 179: Ved. + śasyáte "to receive recognition”; Av. sanhaiti (V. 18, 1-5) "bezeichnet werden als, genannt werden" "könnte auch als präs. (*sahya-) genommen werden” 180: beside the thematic act. pr. to be found in Ved. + : śámsati and Av.: G. sanghaiti, Y. sayhaiti there are some non-thematic forms in Av.: sahyāt etc. In Vedic we have e.g. br. mīyáte "perish, be lost" beside mināti "lessen, diminish, damage"; ksīyáte "be exhausted, perish, cease": kşināti "destroy"; diryáte “break asunder, split open” (intr.): dynāti trans.; pūryáte "become full, fill one's self": prnāti "fill, sate etc."; ſīryáte -ti "go asunder, decay, perish”: senāti "crush, break"; šisyáte "remain, be left”: finasti "leave", etc. 181. Sometimes a -ya-form occurs beside the -yá-present: mấyate, ksiyate; see further on. We may safely admit that -ya-presents with intr.-eventive meaning have often joined activeergative (and middle) presents derived from the same root. On the other hand, a large number of younger -ya-forms will have taken the place of other active and middle presents. 177 See also Meillet, MSL. II, P, 305. 178 Chr. Bartholomae, Die Gatha's des Awesta (1905), p. 37. 179 See A. Meillet-E. Benveniste, Grammaire du Vieux-perse (1931), $ 196, P. 114 180 Chr. Bartholomae, Altiranisches Wörterbuch (1904), 1579. 181 See also F. B. J. Kuiper, Die indogermanischen Nasalpräsentia (Amsterdam 1937), p. 78. Compare Meillet, B.S.L. 26, p. 5 f.

·CONCLUSIONS AND ADDITIONAL REMARKS 9I Let us now consider the -ja-stems which make up the Indian 4th present-class. These presents are, according to Whitney 182, 133 in number; nearly half of them (64) have forms in use in all periods of the language, 41 occurring only in the earlier, 28 only in the modern period. As to the meaning and, I think, the history of these verbs, we may distinguish between some groups or types. I direct attention only to the main points. A very considerable part (over 50) signify a condition of mind or body or a state of feeling, thus: ksudh- "be hungry.", nuh- "be confused", śus- "be dry" etc. As a rule the root appears in its weak form 183. Some exceptions exist; these forms, however, which I exclude from this first group, will be dealt with further on. A number of presents of the ksudh-type correspond with 18+ -10-presents in one or two related languages: epic + kupyati 185 "to become moved or agitated, to boil up, to be angry”: Lat. cupio "desire (eagerly)" 186, Ch. Sl. etc. kypljo "seethe etc."; Ved. + grdhyati "to be greedy, eager; speed etc.": cí. Balto-Sl. gildið "rerlange" admitted by Trautmann 187; Ved. + týpyati "be content, satiate one's self”: Gr. TÉCROLLKU "enjoy, delight oneself”, Técrow "gladden"; Ved. trsyati "be thirsty": Got. þaúrsjan; Ved. + pádyate "go, fall etc.": Av. paid geiti etc. "move (down)"; Ved. + mányate "think etc.": Gr. alvoucı "rage”; Lit. miniù minėti "remember, think etc."; Ch. Si. mbujo mbuěti "voui(ev, doxety”; ep. + sz"idyati "sweat" : 0. H. Germ. swizzen. As there is, in my opinion, no reason to consider these presents, some of which may have existed in pre-Indian times, as essentially different from the verbal forms of the type tujyáte, which we have already enumerated, we have arrived at another argument in favour of the thesis that both - 10-groups were identical in origin. I would suppose that the accent of the type grdhyati has been influenced by the accentuation of some other group, especially of the pásyati and glāyati types. As the starting-point of the analogical process we may consider the rather numerous verbs of these types the meaning of which did not differ from the above presents: Ved. + mådyati "be **, P. 218 f. 182 Whitney, Die Wurzeln 183 See also Renou, Grammaire sanscrite, $ 312. 184 This does not mean 'etyniological identity. 185 In this enumeration only the active or middle forms are given, even when the other conjugation exists. 188 sai. kúpyati cupio", A. Walde-J. B. Hofmann, Latein. etymol. Wtb.3 (1938), p. 312; but see Ernout et Meillet, Díct. étymol., p. 235. 187 Trautmann, 0.c., p. 87. GONDA, The Sanskrit Passive 7 ·Meillet op: 600 92 REMARKS ON THE SANSKRIT PASSIVE agitated, be pleasantly excited etc.”; Ved. + śrámyati “be weary, take pains” 188; Ved. + gláyati "be loth"; Ved. + dhyâyati “think" etc., to which we may add Ved. médyati "be fat”, which notwithstanding the modern etymology (meda- "fat" <*mazda-, i.e. mad-do- or *mad[e]z-do-: O. H. Germ. mast "Mästung'' 189) for the grammatical instinct of the Indians contained a root mid- 190. In all probability the many co-existent verbs belonging to the first present class 191 have also been an element in the process: Upan. ksóbhate: ep. + kşubhyati (4th cl.) “tremble”; rgved. jásamāna-: jasyata “be exhausted"; Ved. + trásati: ep. + trasyati "tremble"; Ved. + bhrámsate : younger Ved. + bhraśyate "fall, sink, be lost"; Ved. + bódhati: búdhyate "wake etc."; rgved. ránati: rányati "be pleased"; Ved. roșati: ep. + rusyati "be cross or angry”; upan. svédate: ep. + svidyati “sweat" etc. In a few cases the first pres.-class appears only in a later period of the language (to my knowledge): Ved. + třpyati "be satisfied etc.”: ep. tarpanti 192 ; Ved. + śrámyati "be weary”: ep. + śramati. This is only to be expected. It would therefore appear to me that the accentual differentiation of the -ya-verbs was attended by a partial leaning towards other thematic root-accented presents. A peculiarly interesting pair of presents are "das mit einem Passiv bildungsgleiche” 193 jayate "be born” and mriyáte "die”. jấyate, which belongs to the Indo-Ir. period 194, may be considered as an innovation instead of the present built on the stem ĝi-gn-0- 195. "Es bildet ... ein Paar mit dem ererbten mriyáte (lat. morior)”, Leumann argues, "vermutlich wurde in urarischer Zeit *gājatai zu *bāta- (Skt. jāta-, ...) gebildet, als neben mrta- noch ein *mrjatai (Av. miryeite gegen Skt. mriyate) stand”. This reasoning does not account for the accentuation of jayate. Although it is, in my opinion, open to question, whether *m7-10- belongs to Common-I. Eur. 196, there can scarcely be any doubt that the two verbs formed a pair and influenced each other; we may compare, e.g., the similar influence of the part. fut. moriturus, 188 See Walde-Pokorny, 0.c., p. 498. 189 P. von Bradke, K.Z. 28, p. 295; Walde-Pokorny, II, p. 230 f. 190 See e.g. Pāṇini 7, 3, 82. 191 See above, p. 86 and n. I6I. 192 Cf. Kuiper, O.C., D. 80, n. 3. 193 M. Leumann, I. F. 57, p. 232. I refer to Bartholomae, Grundriss der iranischen Philologie I, p. 83 (8 148, 9). 195 See Leumann, 0.C. 198 Walde-Pokorny," 0.c, II, p. 276, upholds the identity, but see Ernout et 194 ·CONCLUSIONS AND ADDITIONAL REMARKS 93 "dont la forme s'est étendue à tous les verbes désignant la naissance par opposition à la mort: nasciturus, oriturus, pariturus" 197. But mriyáte retained its accent, because, I fancy, dying was considered as an 'event'; in the philosophical and psychological conceptions and the linguistic feeling of the ancient Indo-Iranians and Indians mrijate presumably was on a par with such ideas and presents as chidyáte "get split or torn, be broken", tujyáte "to flee frightened", vidyáte "be found, exist”, tyhyáte "to be broken to pieces, to be crushed”; dabhyáte "suffer damage or loss"; nid yáte "to be despised, held up to scorn" etc. According to Whitney 198, Macdonell 199 and other scholars mriyáte, though passive in form, is not so in sense; I should prefer to say that mriyáte is apparently a pure 'eventive'. The present jáyate has, on the other hand, been transferred to the -ya-class, because, I think, the idea it expresses was considered to be related to the concepts denoted by the ksudh-group; cf. pádyate "move (down), go, fall, fall out, sink down, perish": ut-padyate "come forth, be born, arise, appear etc."; the other compounds of pad- likewise express .conditions of body or person': ni-padyate "lie down, rest”; nis-padyate "fall out, come forth, arise (, ripen)”, etc. The present jájate is not an 'altered passive' 200, it is an intransitive 'eventive'. In Greek, too, 'be born' was not a passive verb: yiyvouci, yevraouai, šyevóun, yeye mual, réyova, and in many languages this idea is expressed by intransitives, viz, verbs denoting a state or condition etc.: Javanese wětu "come forth, come into the world, be born” 201. In Sanskrit we find sainbhavati "be born etc." Another no less effectual factor in determining the accent of jáyati has, I think, been the co-existence of such pairs of verbal forms as Ved. pácati "ripen" (trans.): pácyate "ripen" (intr.) 202 ; Ved. † sádhati “put in order, effect, accomplish”: sídhyati (the root in its weak form!) “succeed, be accomplished"; cf. also Ved. śúndhati "purify": Ved. + śúdhyati "become pure". The relation jánati "generate, bring forth": jayate has apparently been felt as similar. There also existed a number of 'causative' verbs in -áya-, which presumably derived from common Indo-Eur. and served to express the transitive meaning which corresponds with the meaning 197 Ernout et Meillet, o.C., p. 599. 198 Whitney, A Sanskrit Grammar, $ 773. 199 Macdonell, 0.c., 4442. 200 Whitney, 0.c., $ 761b. 201 Fränkel, I. F. 49, p. 235; Gonda, I. F. 50, p. 273 f. 202 See also further on, p. 98.

·94 REMARKS ON THE SANSKRIT PASSIVE of -ya-presents : Ved. + damáyati "make tame" (: Got. ga-tamjan, O.H. Germ. zemmen, O. Isl. temia, Lat. domo <*domā-jo) 203: Ved. dámyati intr. "to be tame"; Ved. + nāśáyati “make disappear, destroy etc.", Lat. noceo (which originally must have meant "kill, destroy, ruin”): náśyati "be lost, disappear"; Ved. tarsáyati "cause to thirst" (: Lat. torreo, O.H. Germ. derren). In a similar way Ved. + janáyati "generate, bring forth" (: A. Sax. cennan “to generate”) served as a trans. to jáyate. Lastly the accent of such forms as Ved. + śrámyati, Ved. + madyati, Ved. + trayate "protect etc.", which have à in the root-syllable, may have played a part. There can be no doubt that the verb yudhyati “fight, combat”, which does not denote a condition of mind or body, originally did so and, consequently, belonged to the ksudh-group. Although another meaning is very rare in Old Indian, AV. 12, 3, 29 it reads (subj.: boiling water) úd yodhanty abhi valganti taptáh "it rises in bubbles, it dances on, being heated" and Pañc. br. 7: 5, 2 the same compound means: “to fly out, to fire up" (of the subjects of a king). In Av. and O. Pers. the verb usually means "in aufregung geraten, sich aufregen" (of water and turbulent provinces), in Lit. we have judì, judéti “to move trembling" and jundit, justi "to begin to tremble" etc.: the original meaning of the root must have been : "to be moving (or to begin to move) in a restless, excited, turbulent manner". Hence: “to fight etc.”. As a rule the causative-stem serves as a transitive to these verbs : Ved. + tarpáyati "satisfy, please"; Ved. + śosáyati "make dry, afflict"; cf. also Ved. + socáyati "burn, afflict" beside týpyati, śúsyati, śúcyati. The causative may appear in an earlier text than the ya-present: cf. rajayati A.V., árajyate in the same text is questionable. A ‘passive' of the verbs which belong to the ksudh-type does, as a rule, not occur. We may account for the exceptions in the following way: At times a passive which is identical with the intrans. -ya-form appears in post-Vedic or even late texts, e.g. puşyate "be reared" (Bhāg. Pur. 3, 31, 25 pușyamāno janena), budhyate "be noticed, perceived" (Kathās. 41, 14 indrena tad abudhyata) (mark the meaning). Here we have, beside the presents of the 4th class, other presents of trans. meaning, to which the -ya-form corresponds as a passive. pusyati is a curious instance: in Vedic literature it is trans, and intrans., in epic Sanskrit etc, puşnāti, which is trans., appears. In classical Sanskrit See also Ernout-Meillet, O.C., p. 270.

·CONCLUSIONS AND ADDITIONAL REMARKS 95 there is a 'passive' rajyate, beside which we find ep. + rañjati "colour" (and the 4th pr.-stem rajyate, -ti A.V.? ep. +) 204. There are some verbs of the 4th class beside which a transitive present of a nasal-class may be quoted: ep. + kliśyate "suffer, be distressed”: ep. + kliśnāti “torment, molest"; Ved. rígate “flow, run” = rinīte: Ved. rináti "let loose, cause to run". Further, we find Ved. + sidhyati "become pure, clear” beside Ved. śundhati "purify" and śrındhate "purify one's self, become pure" (ātmānan; mukham; daivyāya karmane sundhadhvam); ?gved. dýhya, d'hyasva “be firm" beside drmhati (I; 4.V. also VI) "make firm" 205, Without going into further details, we may conclude that the relations kliśyate : kliśnāti etc. are similar to the relations chidyáte: chinatti etc., which have been recorded above. There is, however, a notional difference between the -ya- verbs used without a personal agens on one hand and the kşudh-group on the other. Let us consider a number of agentless passives or intransitively used ‘passive' forms such as are to be found in Vedic texts: idhyáte "flare up": inddhé; proyáte "ergiesst sich, becomes mixed": prnákti; ricyáte "be emptied, be deprived": riņákti; sicyáte "ergiesst sich": siñcáti (VI); ucyáte "heissen": vívakti; (vi)ricyate "undergo separation": vinákti; pūryáte "increase”: prnäti, píparti; (pari)-sisyate "is left”: sinásti, etc. etc. These -ya-words denote something happening or befalling a subject, they are 'eventives': a process takes place, by which the subject is, in some way or other, affected. The same processes may also be conceived in a transitive sense; then they are denoted by the active verb. The co-existence of a transitive active has apparently been a factor in intensifying the idea of external action, motive, cause, which is often more or less implicitly present. The kşudh-type, on the other hand, is apparently the outcome of a tendency toward a restriction to a certain semantic group. The kşudh-verbs express a condition of mind or body etc. : "to be hungry, to be angry" etc. A process takes place in the subject. A minority of cases excepted, only the causative verb (Ioth class) is in actual use to express the bringing about of the process. Of course this is only a subtle difference 206. A number of roots make their fourth present from their strong 204 As to this verb see Kuiper, O.C., p. 147 f. 205 Cp. also Kuiper, O.C., p. 79 sqq. and 215. 206 kliśyate (ep. +) secondarily in the IVth class, or is kliśnāti a 'young' form?

·96 REMARKS ON THE SANSKRIT PASSIVE form. Thus those roots which never distinguish the vowel of the 'weak forms' from that of the strong forms: páśyati "see": verb. adj. Ved. + spasta-; jasyata (Rgveda) "be exhausted": red. aor. ajījasata; ásyati “throw": verb. adj. Ved. + asta-; tápyati (br. +, beside tápati Ved. +) "heat”: verb. adj. Ved. + taptá-; dasyati “waste”: dasta-; náśyati "be lost”: nasta-; pádyate “fall, go": panna-, etc. Some of these forms inay have had the same history as the ksudh-type. Other -ya-verbs, however, have, in all probability, been accented on the root from the outset: Ved. + páśyati "see”: cf. Av. spasyeiti "see", Lat. specio. In spite of the remarks and observations made by Ernout and Meillet 207 (en indo-ir. comme en latin, le présent du type en *-je-l-j- a ici l'air dérivé de formes nominales" etc.), I admit a form *spék -jo or, at least, I.Ir. spás-jo. However we must appreciate the etymologies given for náhyati 208, which possibly is a denominative, its accent has, I think, always been on the root. The etymology of Ved. + as- "throw" is obscure 209, if a < *e, the accent is probably original. These verbs are transitive. The root trā- "rescue”, which is a monosyllabic form of the dissyllabic base terox-, accents the radical syllable of its root-class presentforms throughout, e.g. träsva, trádhvam; such forms in -ā-, which represents the strong grade of the second base-syllable, commonly appear without gradation 210. It seems, therefore, almost certain that Ved. + tráyate (cf. Av. pres.-st. Orāya-) was accented on the first syllable from the beginning. The verb dhyâyati "think” probably belongs here 211. Compare also Ved. + pyáyate "swell, overflow" (pya-) beside páyate "swell, fatten” (Rgveda) (pi-, pī-, verb. adj. pina-) 212. Beside these evident 'extensions' of monosyllabic roots by the addition of a long vowel also such a present as kşayati "burn" may be considered as a ya-present derived from an ā-root (ksāti- "heat”, kşāma- "burned" etc.). Some presents, however, which by the Indian grammarians and 207 Ernout et Meillet, O.C., p. 921. 208 See Walde-Pokorny, 0.C., II, p. 328 f. and especially Ernout et Meillet, O.C., p. 630. 209 See also K. Pedersen, Hittitisch und die anderen indoeuropäischen Sprachen (Det Kgl. Danske Vidensk. Selbskab, Hist.-fil. Medd. 25, 2, 1938, p. 15I, who compares Hitt. pi-eš-ši-ja-az-zi "throws away, throws off”. 210 Compare also Whitney, Grammar, § 628. 211 According to P. Persson, Beiträge zur indogerm. Wortforschung (Uppsala 1912), p. 700. 212 Cf. Walde-Pokorny, O.C., II, p. 73.

·CONCLUSIONS AND ADDITIONAL REMARKS 97 the modern scholars who followed them, are treated as belonging to the fourth class, are really or may be thematic stems derived from roots ending in y213 or forms in -ayati: Ved. gájati "sing": verb. adj. gīta-, Lit. pra-gýstu "begin to sing" 214, base găi-gi-; váyati "weave" (fut. Rgveda vayisyati; pass. br. sū. Ūyate) may be derived from uei- as vyáyati "envelop"; Wackernagel, howerer, preferred to regard it as an -éìo-present: v-áyati 215; mláyati "relax, wither" (an aorist mlāyāt? 216): base mnel-, mela-, melāi-, melēi- 217; Ved. br. śráyati (beside śrīnáti) "boil" (verb. adj. śrita-) ? To whichever group they may belong, there seems to be no good reason to deny the robability that a considerable number of presents reckoned to the fourth class were always root-accented 218 (compare also glâyati (AV. +, -te sū. ep.) "be loth, be exhausted" beside ep. glāti just like Ved. + tráyate: ep. + trāti). In other I.-E. languages there are also 'primary' verbs from monosyllabic stems ending in a long vowel 219. These belong in part to the root class (Gr. Dor. pauí, Lat. fôri), or in part to parallel ìopresents, as Lat. no "swim" < *snā[i]0. Some scholars 220 hare, however, probably been too rash in considering the original I.-E. character of sna-leo- etc. as certain or, at least, as plausible: whereas snati occurs already in Vedic, snāyate is epic, and vāyate "blow (wind)", which has often been quoted as a close parallel to Got. waian, Ch. Slav. věją etc. 221 does not appear with this meaning before the Mahābhārata. Although it is not possible to draw a hard and fast line between the so-called 'primary' verbs and the denominatives, the statement of Brugmann's 222 that the forms derived from bases in a "sich grundsätzlich nicht von der Art unterschieden, wie seit uridg. Zeit 213 The Vedic stems in -aya-, part of which are not clear, are, however, accented after the 1'; as to grbhāyati see further on. 214 Cf. W. Schulze, K. Z. 27, p. 425. 215 Wackernagel, Altindische Grammatik, I (1896), p. 94; cf. also Walde- Pokorny, O.C., I, p. 16 and 223. 216 Whitney, Roots, p. 128. 217 Walde-Pokorny, o.C., II, p. 284 ff. 218 Brugmann, who in the first edition of his Grundriss (1892), II, p. 1085 assumes the -jé-accentuation, does not mention this point in the second edition. 219 I refer to Brugmann, Grundriss”, II, 3, P. 198 ff. 220 See e.g., C. D. Buck, Comparative grammar of Greek and Latin3 (1912), P. 268. 221 Brugmann, O.C., P. 201, S. Feist, Vergl. Wörterbuch der gotischen Sprache3 (1939), p. 541 etc. 222 Brugmann, Grundriss, II, 3, p. 198.

·98 REMARKS ON THE SANSKRIT PASSIVE Denominativa mittels -ị0- von fem. a-Stämmen aus geschaffen worden sind” perhaps lacks precision as far as the accent is concerned. For the Indian denominatives in -ya- have in general the accented suffix: gopāyati "plays the herdsman”, priyāyáte "holds dear", apasyati "is active" etc. 223. There are, however, denominatives which, like the causatives, have the -áya-accent 224, and examples of this class already appear at the earliest period of the language: mantráyate "speaks, takes counsel”: mantra- "(sacred speech), counsel" etc. Let us now consider the -ya-forms of which the accent fluctuates. A particularly striking instance is pacyate: RV. I, 135, 8 pácyate yávaḥ means "the corn ripens”, but 6, 29, 4 paktih pacyáte "the (cooked) food becomes done”; 8, 58, 14 odanán pacyámānam is “porridge which becomes done, which is prepared", cf. I, 162, II; A.V. 5, 19, 4 "the brahman's cow, being cooked": pacyámānaḥ. So, the process which takes place in the ripening corn, which-as one might say—takes place automatically 225, is denoted by pácyate, the process of cooking sufficiently, which supposes technics, by pacyáte. As to jāyate, RV. 9, 55, 4 yo jináti ná jiyate has correctly been translated 226: "der stets du siegst, nie unterliegst"; 10, 152, I ná vásya hanyáte sákhā ná jīyate kádā caná I should like to interpret in this way: whose friend is never killed nor loses the day" ("went to the wall”, or other 'intrans.' expressions); in the same way 3, 59, 2, where I do not adopt Geldner's translation: "...wird er nicht erschlagen noch ausgeraubt" 227; 5, 54, 7; but A.V. 5, 19, 6 brāhmaṇó yátra jīyáte is "where a brahman is vanquished" 228 (cf. e.g. st. 2, where the insults offered to the brahman are enumerated with active verbs). In the same way we are tempted to distinguish between an 'automatic' and an 'eventive' intransitive and to try to establish an appreciable difference of meaning. Thus T.S. 7, 2, 1, 4 yadaiņām pramiyeta yada vā jấyeran “if one of them perish or they go to the wall”, and not ... they be oppressed". But in this way we presumably stress the point too much. Whitney 229 says that ksiyate is found in Vedas 66 64 - 228 In the same way Ved. grbhāyati "seize, grasp" (Whitney, Grammar, § 1066b), see Kuiper, O.C. p. 232 (derived from grbha-); compare also śrathāyati "become loose", (hima)śratha- "the melting (of snow)", like priyāyáte. 224 See Whitney, Grammar, $ 1056. 295 But RV. 10, 88, 10 it reads sá (agnih) oşadhih pacati. 226 Grassmann, Rig Veda übersetzt, II (1876), p. 216. 227 Geldner, Der Rigveda I, p. 367. 228 Whitney-Lanman, Atharva-veda Samhitā. I, p. 253. Whitney, Die Wurzeln ..., p. 29.

·CONCLUSIONS AND ADDITIONAL REMARKS 99 and Brāhmanas and kṣīyáte in Vedic and post-Vedic literature. In the RV. the accented form ksiyate occurs only 1, 62, 12 rayah... ná k. "the riches do not lessen"; 2, 9, 5, where the subject is vasaviyam "wealth” and 6, 45, 3 (ütayaḥ "help etc.") the same shade of meaning occurs; these forms are, however, unaccented. But AV. 12, 5, 45 we find ksīyáte, which obviously means: "he is destroyed": cp. (39) “Her (the cow's) slaying is magical potency (41) she eats him, (43) she makes perish his maternal connection, (45) without abode ... she makes him”, ksīyáte. A number of -ja-presents (4th class) may be considered as young. vrid yati (Dhātup. 26, 18) beside ep. + vrīdate "be ashamed"; cf. nabhyati (ibid. 26, 130) beside Ved. nábhate "burst"; ep. + trasyati: Ved. + trásati; trutyati (Pān. 3, I, 70; cl.): cl. trutati "get broken, fall asunder”; younger Ved. + bhraśyate ep. + -ti "fall, sink”: Ved. + bhrámśate. In all probability these roots have adopted a present of the - j'a-class because of their meaning; compare also ep. asyati, -te "sit"; Ved. cl. iyate "go"; ep. + drāyate "sleep", etc. There are some -ya-presents which may be called "Schallwörter": Ved. + vásgate, ep. -ti "low, bleat, roar, cry (of birds)": for the formation cp. Lat. vägio, -ire ("formation expressive: "faire wā" " 230) "cry, scream”; Ved. ráyati "bark”, cf. Gr. Zalety poényectal (Hes.) "utter sounds, speak”, Ch. Slav. laje, lajati "bark" etc. 231; cl. ghusyant beside Ved. + ghoșati "sound, cry, call out"; ep. lapyant beside Ved. + lapati "chatter, lament etc."; the two last mentioned presents evidently being young. In other I.-Eur. languages words denoting sounds etc. are often formed with the jo- (-:e-110-)-suffix 232 : Lit. krankiu, krañkti and kraukiu, kraūkti “croak”; Gr. xpów "croak” etc.; cf. also Bhoow "cough” etc., Lat. crocio "croak”, etc. The personal endings of the 'passive' -ya-class have repeatedly been discussed in a manner which I cannot consider satisfactory 233. These -ya-forms are no complete passive; and even if they were -: a Greek pass. aor. can teach us that there exist passives without middle endings. The -ya-class occasionally assumes active instead of middle endings 230 Ernout et Meillet, Dict. étym., D. 1028. See also Walde-Pokorny, Vergl. Wtb., I, p. 214 f. 231 See Walde-Pokorny, O.C., II, p. 376. 292 See, e.g., Brugmann, Grundriss, II, 3, D. 194; Schwyzer, Griech. Gramm., p. 716. 233 See already Brugman[n] in Osthoff und B., Morphologische Untersuchungen, I, p. 204 ff.

·IOO REMARKS ON THE SANSKRIT PASSIVE which are usual. In the epics active endings even are by no means infrequent 234. According to several authors they are due to a later development. Whitney explains the frequency of the active endings in the epics as a part of the general confusion of personal endings in these texts 235. To a certain extent he is right. This view has also been adopted by Schaeder 236 who stated that Veda (i.e. in the samhitā’s) and Gālā's "ausschliesslich mediale Endungen aufweisen". With the presents of the kșudh-type, however, the active endings are very usual: br. kýśyati, Ved. + kșudhyati, Ved. + gdhyati; Ved. trsyati; Ved. +d. - pyati "rave, be foolish”; Ved. + lubhyati etc. have only active endings, of other verbs the active forms are already found in Vedic texts, the middle only in post-Vedic literature: Ved. + krudhyati, ep. + -te; br. + tâmyati "become dull or stiff, faint away", ep. + -te; y. Ved. + tusyati, cl. -te; Ved. + tipyati, ep. t -te; Ved. + dúsyati "be or become bad, impure”, ep. + -te; Ved. + púşyati: ep. + -te; y. Ved. + bhrásyate, ep. + -ti; Ved. + múhyate, up. + -te; br. + śúdhyati, ep. + -te; Ved. + súsyati, ep. + -te; Ved. + hrsyati, ep. + -te; cf. also Ved. + nýtyati: ep. + -te "dance etc.” The present mányate (Ved. +) forms an exception: -ti up. +; cf. also the trans. Ved. + mrsyate "forget”, ep. + -ti. With reference to the other remarks made in this book with regard to the connection between the two -ya-classes, the frequency of active endings with the ksudh-verbs convinces that they originally could have occurred with the -ya-forms as well. Of the so-called passive aorist in -; 237 there are 178 indicatives and 42 subjunctives in the Rgveda. Only 13 times the agens is added. In this text this type is derived from trans, and intrans. verbs and we cannot consider it to be prevailingly a 'passive formation'. These statements are in accordance with the rules given by Pāṇini, who admits i-forms of a 'middle' meaning (apādi, 3, 1, 60) and karmakartari (usually rendered by reflexive'), which presents itself when sub- -- 234 See e.g. Whitney, Gr. $ 774; Ad. Holtzmann, Grammatisches aus dem Mahabharata (1884), p. 25 f. 285 Similarly D. Andersen, Om Brugen og Betydningen af Verbets Genera i Sanskrit, Diss. Copenhagen 1892, p. 118 f. See also Renou, Gramm., p. 465. 238 H. H. Schaeder, Beiträge zur iranischen Sprachgeschichte (Ungarisches Jahrbücher 15, 1935), P. 260. From this Schwyzer, Griech. Gramm., D. 714, n. 3 concludes: ''in geschichtlicher Zeit ist freilich die aktive Flexion jünger als die mediale See Renou, Grammaire sanscrite, $ 318. 2347 ·CONCLUSIONS AND ADDITIONAL REMARKS IOI 97 ject and object of the action are identical (abhedi kästham svayam eva, 3, I, 87) 238 beside forms denoting bhāva (aśāyi bhavatā, 3, I, 66) and karman (passive' (akāri kato devadattena, 3, 1, 66). The Iranian equivalents of this category, Av. väči etc. 239 and O.P. akariy etc. 240 are 'on-forms', agentless ‘passives': Y. 43, 13 yā wahmi xša Brõi väčī "von dem gesagt ist, dass es in Deinem Reiche sei” (Bartholomae) ; Yt. 8, 48 aidi Germ. "heissen" 241; Y. 53, 1 vahišta ištiš srāvī zaraQuštrahē "das beste Gut, das man kennt, ist das des Z.” (B.); Y. 45, 10 yā ... Piazd å stāvi ahuro; Y. 32, 8; 9, 3; 4 kā ahmãi ašiš aranāvi "welches Los wurde ihm zu teil?" (B.) (eventive); 44, 18; Yt. 19, 92 yat ažiš dahāko jaini "was killed”; 93 etc.; Y. 32, 14 mraoi (mravī) "dass es heisse”; Dar. Beh. II 75; 90 basta adāriy "was kept in prison”, etc. Now, in other Indian texts this form is, generally speaking, seldom found. In the prose of the Ait. Br. there are 20 instances from 4 roots (ajani, apādi, asāvi, vāci) 242, in Taitt. S., Ait. Ar., Jaim. Br. it is rare; in Mbh. I-XI it is, according to Holtzmann 243, unknown and the only instance he found in the great epic (12, 29, 89), abhedi, is clearly 'eventive': dyaur a. (bhinnā Nīlakantha). In Sukas. 21-27 there are 4 forms (2 with agens). In artificial literature the type revives : Aśv. Saund. 1-6 and 10: 9, Siśup. 1-2: 4 (with ag.), Kathās. Vet. 6 (2 with ag.), Dašak. 2-3: 5. So, it is clear that this type is in origin an agentless ‘on'-form or eventive, which could develop into a real passive in extraordinary speech 244. One of the most interesting results of my inquiries concerns the frequency of the passive use of those forms which the middle and the passive voices have in common. I need not dwell upon the expression of the so-called passive by means of middle present-forms, this matter being repeatedly dealt with 245. In the Rgveda the “passive use of 238 See also Renou, Terminologie, I, p. 122. 239 Reichelt, Awestisches Elementarbuch, § 232. 210 Meillet-Benveniste, Grammaire du Vieux-Perse, 207. 241 See, however, H. Lommel, Die Yašt's des Awesta (1927), p. 56. 242 Cf. J. Avery, Contributions to the history of verb-inflection in Sanskrit, J.A.O.S. IO, P. 294. 243 A. Holtzmann, Grammatisches aus dem Mahābhārata, § 844. 244 See also A. J. Eaton, The Atmanepada in Rigveda, Diss. Leipzig 1884, P. 20-24 245 Eaton, O.C., P. 19 ff.; D. Andersen, Om brugen og betydningen af verbets genera i Sanskrit, København 1892; B. Delbrück, Altindische Syntax, p. 263 f. I refer to my critical remarks, see above, p. 50 f.

·102 REMARKS ON THE SANSKRIT PASSIVE middle forms is often met with, and real passives with agens) are not unknown, in Vedic prose, however, it is rare and after that period it disappears. As to the middle future used in a passive sense, Eaton 246 and Delbrück 247 mention only one text taken from the Rgveda, viz. 8, 70 (= 679), 14 rşibhiḥ stavisyase. In fact, the passive sense of this form is far from being common, the gerundive to a great extent taking the place of a passive future. From a statistical account of the middle forms which are liable to be used in a passive sense 248 it appears that the ‘passive future' is very seldom found: Nala: 1; Bhagavadg.: 3 (intr.); Kālid. Śak. I and 6 (prose): I; only in the Kathās. it is comparatively frequent. In epic language the future of muc-: moksyase, -si forms an exception, if at least Mbh. 3, 279, 4 na hi me moksyase jivan (cf. 3, 284, 16) me expresses the agens 249. It may be a dative of concern denoting the person who is interested by the action. In the expression sarvapāpaih pramoksyase (3, 135, 6), quoted by Holtzmann, the verb does not represent a passive proper 250. Of the many middle forms used for the passive voice, noted by Eaton 251, there are only a few real passives. A passive aorist is RV. I, 77, 5 evágnir gótamebhir ?távā víprebhir astosta "was praised by the inspired Gotamas"; cf. also I, 124, 13 ástodhvam stómyā bráhinanā me. A passive perfect is 1, 100, 10 vidé víśvābhih krstíbhih “is known by all races”, and also 7, 90, I dadrire vām adhvaryúbhir ... sutásah. The same forms are found without an agens: 1, 132, 3 sá ghā vide anv indro gavéşano ... "Indra ist ja bekannt als der Rindersucher" (Geldner) ; 4, 34, 4 dadé vo máhi ... sávanam mádāya; 37, 3. Most forms are agentless passives', no real passives, cf. I, 135, 3 and 6, where, however, an agens, which must be connected with the participle, is present: adhvaryúbhir bháramānā ayansata; 8, 32, 16; 9, 2, 5; eventives: I, 80, 3, resultatives denoting a condition or state of being: 9, 29, 5; 10, 31, 3; cf. also the figura etymologica divó ná sárga asasygram 9, 97, 30 etc., “reflexive' intransitives: 6, 48, 6 támo dadyśe; I, 95, I. Thus, 10, 87, 19 må ... hetyá muksata should be translated: "must not escape the missile". An interesting case of opposition to 246 Eaton, O.C., p. 25. 247 Delbrück, O.C., D. 265- 248 As required by Speyer, Sanskrit Syntax, § 314 note. 249 See Holtzman, O.c., § 998b. 250 See above (Bhagavadgitā), p. 55. 361 Eaton, a. D. 25.

·CONCLUSIONS AND ADDITIONAL REMARKS IO3 active forms is 6, 69, 8 ubhá jigyathur ná párā jayethe ná páră jigye kataráś canainoh "you both have been victorious: you both conquered, neither of you has been conquered". The participles are seldom attended by an agens: 1, 135, 3 and 6 (see higher up); 3, 10, 4 agnir devébhir á gamat / aſījānáḥ saptá hótybhir ...; 1, 62, 5 grnānó ángirobhih, but 35, 10 the same form occurs without an agens. The participles of the intensive-stems of the verb mrj- "to clean" are rather frequently attended by an agens, cf. e.g. 9, 91, 2 prá vó nybhir ... marmrjānó. As to epic language, Holtzmann 252 already remarked that "sich die Praxis darauf beschränkt, einige wenige Wörter, welche "sehen, hören, bemerken" bedeuten, im reduplicirten Perfect des Mediums auch in passivem Sinne zu verwenden. So besonders suśruve "er wurde gehört”, immer mit einem "Lärm, Tumult" bedeutenden Subjecte: ...; es scheint, dass weder andere mediale Formen von śru- im passiven Sinne gebraucht werden noch auch suśrure selbst anders als in jener formelhaften stehenden Redensart". The forms quoted by him are, however, only eventives, on-forms, cf. e.g. 6, 119, 17 teşan apatatām śabdah suśruve phälgunam prati. In the same way dadyśe I, 56, 13, which followed by iva means "looked like”, cf. 7, 14, 71 dadyśāte sapakşāv iva parvatari. In Kavya the complete passive turn of middle perfect-forms are, however, far from being rare: Kāl. Ragh. 19, 18 tena niședusa prsthataḥ śuśruve vacah "by him, sitting, behind them..., were heard the words” 253. And the artificial character of this turn is quite clear from such texts as Bhār. Kir. 7, 17 ásede daśaśatalocanadhvajinya ... indrakilah, where äsede means praptah. So, we are not surprised at finding that in Kirātārjuniya 7-9 there are 28 middle perfects used in passive senses and attended by an agens beside only 3 agentless forms, that in Siśupālavadha 1-2 all (4) perfects of the passive are real passives, that even in many chapters of the Kathāsaritsāgara the number of passive perfects with an agens exceeds that of agentless forms. Whereas Kss. 75, 152 äravah śuśruve is a case of the common epic and narrative use of this perfect, 167 and 168 jagshe sa dystvä purarakşibhiḥ and ninye ca taih represent the Kavya-style. The grammarians admit the impersonal use of the verbal adjectives in -ta- (Pāņini 1, 2, 21 dyutitam or dyotitam anena; ihāhina syptam #. .. .. #.. 252 Holtzmann, O.C., § 531. 253 I refer to the texts discussed in part II of this book.

·104 REMARKS ON THE SANSKRIT PASSIVE "here a serpent has crawled") 254, which is much liked by later authors. Such constructions as gantavyam mayā occur already in the Rāmāyaṇa, and there are other impersonal turns: “La prédominance de l'expression passive entraîne la constitution d'un impersonnel (bhāva...), qui se propage de façon inégale et tend à s'adjoindre un instr. régime...” 255; thus we find such imperatives as śrūyatām and impersonal aorists in -i. But in Kāvya-works impersonal turns are coming into fashion. According to Thomas 256 and Walter 257 Kumāradāsa has been the first Kavya-poet to use the third pers. sg. of the middle perfect as an impersonal with the subject in the instrumental. Bhāravi and Māgha are fond of this turn: Kir. 3, 30 iti bruvānena mahendrasūnum maharşiņā tena tiro babhūve; 16, 35 sphutam visasre savitur mayukhaih; Siś. 3, 61 ūse makaradhvajena; 6, 39 nanyte navapallavaih 258. The author of the Bhattikāvya also has it: 14, 101 cakhadire cakhādus ca 259. As is well known, the case may present itself in which the verbal adjective in -ta- or -12a- is nothing more or less than a simple attributive adjective. The remark, made by Delbrück 260, "Eine Anzahl von Formen auf -tá- übersetzen wir durch Adjectiva. Doch ist die Grenze fliessend" is correct, but it is quite a common phenomenon that one language expresses by what we call an adjective what another denotes by a participle. The habits of a given language are, however, not exhaustively defined by stating that it makes use of such and such particular forms, there remain the questions of their frequency and of their stylistic value. In the prose of the Brāhmanas and in the story of Pauşya attributive verbal adjectives are not numerous. In other texts their frequency is not always the same, but they are often rather rare. An exhaustive inquiry into the matter might be in place, and then we might go on examining which verbal adjectives are particularly used in this way. In the Rgveda, where attributive Ppp.'s are comparatively numerous, we find such forms as prītá-: vāji ná prītáh, śuddha-: .. 254 See also Renou, Grammaire, p. 195 and 198. 255 Renou, O.C., p. 498; see also pp. 205; 314. 256 F. W. Thomas, The Jānakiharana of Kumāradāsa, J.R.A.S. 1901, D. 260. 257 O. Walter, Übereinstimmungen in Gedanken bei den indischen Kunstdichtern..., (Leipzig 1905), p. 36. 258 I refer to Walter, O.C., p. 36 f.; and to Cappeller's transl. of the Kir. 259 See also Renou, O.C., p. 459 f. (Ašv. Bc. 2, 21 see Johnston's edition, where ācakrire is transitive; Siś. 14, 35 is not a complete passive). 260 Delbrück, Altindische Syntax, p. 385.

·CONCLUSIONS AND ADDITIONAL REMARKS 105 ... I, 164, 40 píba śuddhám udakán ; śruta-: 1, 101, 7 indran abhý arcati śrutám, etc. 261. These facts correspond on one hand with the origin of the category, on the other hand with the use of the verbal adjectives in Old-Iranian, which I have, however, not thoroughly investigated. Meillet-Benveniste 262 unjustly does not differentiate between attributive and other employment and what has been entered by Reichelt 263 in his short paragraph on this topic should at least in part be dealt with under the heading 'verbal adjectives expressing attendant circumstances': V. 18, 46 yat nā mapto zšudrå fräraodayeite "when a man, after falling asleep, ejaculates semen". As we have already seen above 264, the v. adj. is often considerably more frequent when used as an apposition: in Paușya we find = 50 cases beside 9 attributes. As such it serves to express attendant circumstances or other qualifications of the main action (whether causal, concessive, temporal etc.) 265. It may be attended by an instrumental, whether this case expresses concomitancy, an instrument, or the agens. In Vedic texts this instr. is rather rare 266. In RV. I, 1-770, II, III, 1-42 I found 7 examples beside 125 attrib. and appos. without an agens. In Pausya 15: $ 50 appos. without an agens. Nor is it frequent in epic ślokas (in selected parts from the Rām. Í 1:312). As to Kārya, in Aśv. Saund. (cantos 1-6 and 10: 6:55), Kāl. Megh. 1:19), Budhasvāmin the addition of an agens is comparatively rare, but in Kāl. Ragh. (cantos 1, 3, 5, 15, 16, 19:21: † 30), Bhār. Kir., Kathās. it is often found. The difference between the two poems of Kālidāsa seems easily explained by the different character of their subject-matter. It strikes us that in the Daśakum, the agens is seldom added, and that, on the other hand Manu, Kūrmapurāna Išv. and śukasapt. have it comparatively often. . As has been shown by Meillet 267 in a fundamental paper, which rectified the older views on this subject, there existed already in the Indo-Eur. period two kinds of nominal sentences: "la phrase nominale pure” (which did not contain any personal verb-form), being the 261 See also Renou, O.C., § 153b. 262 Meillet-Benveniste, O.C., § 217. 263 Reichelt, o.C., § 670. 264 See p. 17 ff. 265 Its use has sufficiently been described by Speyer (Sanskrit Syntax, § 362). 266 Renou, Grammaire, & 152: „non fixée à date ancienne". See also Wecker, B.B. 30, 196, whose 'agens' includes too much. 267 A. Meillet, La phrase nominale en indo-européen, MSL. 14 (1906), p. I ff.

·106 REMARKS ON THE SANSKRIT PASSIVE normal turn when the third person sg. indic. was expressed, or, in a more general way, "un présent de l'indic. si la personne est suffisamment indiquée sans l'intervention d'une forme verbale". In other cases the 'copula' was usually added. Although in most languages the second turn has to a large extent driven out the former, in the ancient Indo- Iranian languages the construction without as- or bhū- has remained quite common. In treating these sentences it will be convenient to adhere to the distinction made by Delbrück and other scholars: a) sentences containing a verbal adj.; b) sentences without a verbal adj. Avery 268 collected from the Ait. Br. 6736 personal forms and 174 verbal adjectives in -ta- and -na- used as predicates without a copula derived from 53 roots. The total number of verb. adjs. is 858, derived from 120 roots. These figures show that in Ait. Br. the turn is not frequent. From RV. I, 1-70, II, III, 1-42 (which contain † 1650 stanzas) I have collected about 35 instances of the agentless construction and 3 with an agens expressed; in Ath. V. IV these figures are 12 and 4 or 5, in Kathop. 3 and 5. According to Bloch 269, who made an inquiry into the verbal and nominal sentences in all prose parts of the Mbh., the development the participial sentences (-ta-, -tavant-, adj. necess.) have undergone since the Vedic prose is remarkable. The total number of nominal sentences in the said parts is about 315, that of the 'phrases participiales' more than 200, more than 160 containing a 'part. perf. pass.' used predicatively. In the Pausya episode the agens is added in 4 cases. The verbal sentence, however, which occurs 1033 times, is the most frequent construction. As to epic ślokas, in the Sāvitrī episode the figures are: (all sentences containing a ppp.) ppp. (verb. adj.) used as a pred. without agens expressed 13, with agens 30 (attrib. 26, act. and middle personal forms 348), in Nala (sentences without a copula) I-19: transitives 14 (12 in oratio recta) and 50 (41 in oratio recta), intrans. 36 (29); according to Avery 270 the total number of personal forms of Nala and Bhagavadgītā is 1282, of which 224 are used as predicate without a copula; in Mbh. 5, a. 186-192 incl.: 21 (10 or r.) and 23 (21 or r.) (act. and middle pers. forms 284); in Manu 2: 22 and 4 (a. and m. pers. forms 265). In some parts selected from the Rām, this construction seems to be almost as numerous, but the cases of trans. PPPs, in which the agens has been added outnumber the agentless in- 1968 Avery, JAOS, IO, D. 294 f. ; 319 f. 289 Bloch; La phrase nominale en sánscrit, MSL. 14, p. 64. CE ·CONCLUSIONS AND ADDITIONAL REMARKS 107 stances; the construction is especially liked in oratio recta; there are, of course, many intransitives. In the prose of the first and sixth acts of the Sakuntalā, which contain about 250 act. and m. pers. forms, 34 passive pers. forms and 5 instances of the -facant-turn, there are about 75 pred. verb. adjs. without ag. and $ 50 with ag. expressed. In Dasak. 8 these figures are: about 220, 20, 3, 22, 16, there are about 70 other complete nominal sentences. In Sukas. 21-27, they are, respectively, 290, 7: 23, 6, 5, 33. From 20 pages of the Hitopodeśa I collected 218, 22, 6, 18, 52, 57. In Sivadāsa's Vetālap. (ed. Chle) this use of the verb. adj. is very common. Bloch 271 arrived at the conclusion that "dans le Vetāla nous ne trouvons plus que 790 verbes contra plus de 1750 phrases nominales. Ce changement est dû surtout à la substitution des formes participiales aux temps passés du verbe: en face de 38 formes verbales, le Vetāla offre 1115 adjectifs en -ta- comportant cette valeur". From these figures we may conclude that in later prose in general this participial turn is much liked, but also that the texts are divergent with reference to its relative frequency 272. It is worth while to examine also some metrical texts. In Asvaghosa, Bc. 2-4 there are 324 act. and m. pers. forms, 13 pass. pers. forms, 15 PPPs. predicatively used without an agens and 9 with an agens. In Kūrmapur. Iśvaragitā 1-4; 6 these figures are 211, 22, 16, 10; Budhasv. Bșhatkathāśl. 14, 1-15, 107 they are 184, 39, 38, 52 trans, and 2 and 25 (!) intr. In Kālid. Ragh. 1, 3, 5, 15, 16, 19 there are 32 and 12 Ppps, used in this way; from 20 pages of the Kathās. I collected 12 and 36 trans, cases besides 269 act. or m. personal forms and 23 other nominal sentences. Now, there are 'semi-nominal' sentences in which a copula is added. As a rule this type is less numerous than the complete nominal sentence: in selected parts of the Mbh: + 3: 22; Manu I 4: II, Rām. 1: 2, Isvaragītā 3: + 130!, Dasakum. + 1: 40!, Budhasv. Brhatk. 2: 25, Sukasapt. + 1:6. When the predicate is a ppp. the cases in which there is no agens expressed usually outnumber the other cases : Pauşya episode 10:8, Kālid. Śak. I and 6 prose 12: 3, Dašak. (8) 3:1, Ašvagh. S. 7: 2 etc. The conclusion, then, is that although it is, with good reason, the prevailing opinion that these sentences have much contributed to the 271 Bloch, O.C., p. 93. 272 In a different manner Bloch, O.C., D. 28.

·LIBRAR** TYDERABAD. 108 REMARKS ON THE SANSKRIT PASSII . preponderance of the so-called passive construction in Sanskrit 273, it is not advisable to call a large number of these ppp. predicates real or complete passives, because the agens which is comparatively seldom added in Vedic prose and metrical texts continues to appear, as a rule, only in a minority of cases. There are, however, exceptions: the Hitopadeśa, epic texts and those kāvya writings which combine a certain simplicity with an explicitly narrative character (Kathās., Budhasv. Bệhk.). As the verbal adjective originally indicated a person or object as being in a situation resulting from the process expressed by the verb 274, there is no reason to be surprised at this state of affairs. 273 See Renou, Gramm., § 367, D. 497 f.: "Un trait de la phrase sanscrite est la faveur dont jouit l'expression passive : le fait, en tant que générique va de pair avec le progrès du style nominal qui multiplie les adjectifs passifs en fonction de verbes". 274 H. Kallin, Etude syntactique sur l'expression du rapport de l'agent dans les langues romanes (Diss. Uppsala, Paris 1923), p. 3 f. emphasizes the durative aspect' predominating in the verbal adjectives; with these words speakers do 110t consciously express a 'rapport d'agent': Fr. le jardin est entouré d'un anır, but Engl. the house was surrounded with trees; Cl. Lat. minora castra inclusa inaioribus. "Le verbe, bien que passif au point de vue de la forme, ne l'est pas au point de vue du sens. C'est que le participe exprime ici un état, sans que cet état se présente à l'esprit comme amené par une action antérieure. La préposition marque simplement que le sujet se trouve par rapport au complément dans une situation donnée".

·No text on this page.