Language and Reality: On an episode in Indian thought

1 LANGUAGE AND REALITY On an episode in Indian thought by Johannes Bronkhorst translated from the French by Michael S. Allen and Rajam Raghunathan revised and with a new appendix (published: Leiden – Boston: Brill, 2011) 1 2 TABLE OF CONTENTS PREFACE TO THE FRENCH EDITION I. INTRODUCTION: WORDS AND THINGS 1. Aim of the Lectures 2. Early Brahmanical Literature 3. Pāṇini’s Grammar 4. A Passage from the Chāndogya Upaniṣad 5. The Structures of Languages 6. The Buddhist Contribution 7. Vaiśeṣika and Language 8. Verbal Knowledge II. THE CORRESPONDENCE PRINCIPLE 1. The Contradictions of Nāgārjuna 2. The Reactions of Other Thinkers 3. Sarvāstivāda 4. Sāṃkhya 5. The Āgamaśāstra of Gauḍapāda 6. Śaṅkara 7. Kashmiri Śaivism 8. Jainism 9. Early Vaiśeṣika 10. Critiques of the Existence of a Thing before Its Arising 11. Nyāya 12. Mīmāṃsā 13. The Abhidharmakośa Bhāṣya of Vasubandhu 14. The Abhidharmasamuccaya of Asaṅga and Its Bhāṣya 15. Bhartṛhari 16. The Problem of Negation 17. Dignāga and Verbal Knowledge 18. The Bodhisattvabhūmi 19. Prajñākaragupta 20. Indian Thinkers and the Correspondence Principle APPENDIX: The Mahāprajñāpāramitāśāstra and the Sāṃkhya tanmātras BIBLIOGRAPHY ABBREVIATIONS CITED PASSAGES INDEX 2 3 PREFACE TO THE FRENCH EDITION The following work is based on a series of lectures delivered before the Section des Sciences Religieuses at the École Pratique des Hautes Études (Sorbonne, Paris) in May 1997. I had addressed the main subject of these lectures before on other occasions, but never on this scale.1 The lectures in Paris gave me the opportunity to present the historical background of my subject, which I call the correspondence principle, as well as examples of its many manifestations. Not even a series of lectures, however, could do justice to the subject, and what follows is far from exhaustive. Ideas and intuitions about the relationship between language and reality abounded in classical India, and a full understanding of all their aspects, all their expressions, and the networks of thought connecting them together is still a distant goal. The correspondence principle is presented here as a modest contribution to such a complete understanding. Its exploration has necessarily been limited to several more or less representative cases of philosophical thought in classical India. I hope my readers will pardon the somewhat arbitrary choice of thinkers studied. I am well aware that many thinkers, and even entire currents of thought, have not been mentioned in this book, though they might indeed have shed light on certain obscure points in our investigation. There are trails that remain unexplored, thoughts still unexpressed. I can only hope that future research will help to fill in the gaps left by this exploratory study. In preparing this book and its preliminary studies, I have benefited from the critical responses of many colleagues and friends, too many to list them all. I would nevertheless like to mention here John Dunne, Vincent Eltschinger, Danielle Feller, Pierre-Sylvain Filliozat, Eli Franco, Gerdi Gerschheimer, Brendan Gillon, Lars Göhler, Minoru Hara, Masaaki Hattori, Richard Hayes, Jan Houben, Shoryu Katsura, Katsumi Mimaki, Édith Nolot, Claus Oetke, Lambert Schmithausen, Frits Staal, Ernst Steinkellner, Tom Tillemans, Gary Tubb, Tilmann Vetter, François Voegeli, and Toshihiro Wada. I have learned a great deal from the difficulties my students encountered in reading an earlier version of the text; the fault was entirely mine, and I can only hope that the causes of their confusion have now been cleared up, at least in part. I would like to thank everyone who contributed to this book in one way or another—sometimes with a pertinent criticism, sometimes by refusing to comprehend the incomprehensible, at other times with much appreciated encouragement. I am especially grateful to Marie-Louise Reiniche for organizing the lectures in Paris, and to Charles Malamoud and Lyne Bansat-Boudon, who gave me so much of their time in seminars. And I would like to thank my wife, Joy Manné, for her constant encouragement, even when she must have sometimes wondered whether my obsession with the arising of a pot had not gone too far. Lectures delivered elsewhere, but touching on aspects of the same theme, have been published separately; see. e.g., Bronkhorst, 1996b and 1997, in the bibliography at the end of this work. 1 3 4 I. INTRODUCTION: WORDS AND THINGS 1. Aim of the Lectures These lectures aim to draw attention to a belief that underlies an important part of Indian thought, one that has not yet received the attention it deserves. In brief, there was a period in classical India when most thinkers were convinced that the words of a sentence correspond rather exactly to the things constituting the situation described by the sentence. This conviction has some parallels in the history of Western philosophy, but it has not played a role of comparable importance.2 I would ask, therefore, that you leave aside comparison, at least for the time being, and instead attempt to discover with me something of the internal logic—that is to say, the raison d’être—of Indian thought. The conviction at stake, which I will refer to as the correspondence principle,3 allows us to understand several aspects of the thought of the period in question, such as, for example, the deconstructive arguments of the Buddhist Nāgārjuna, as well as the reason why his arguments, after a period of glory, came to be ignored by Indian thinkers. As we shall see, the correspondence principle bears a special relation to Buddhism, though it was never limited to it. For several centuries the problems linked to this principle occupied practically every philosophical current in India, to the point that one can say it influenced, even determined, classical expressions of Brahmanical thought. My project is ambitious. Its goal is not simply to understand Indian thought in its multiple expressions, but further to discover the reasons (or at least one of the reasons) why it took one form instead of another, why one position was chosen in favor of another. If Indian thought offers solutions to certain problems, or answers to a good number of questions, then the goal of this project is to identify at least one of these problems, one of these questions. We shall see that different positions adopted by different currents of thought, though apparently unrelated, prove to be answers to a single question, even if it is not always explicitly formulated. And it is not simply a matter of details; among the positions that seem related in this way are the satkāryavāda of the Sāṃkhyas etc., the anekāntavāda of the Jainas, the śūnyavāda of the Buddhists, and several others besides. In other words, this project aims to bring out a unity that lies hidden behind multiple forms of Indian thought, at least in one domain. For this reason, I will refer in the course of these lectures to many 2 Lars Göhler brought to my attention the following sequence from the Tractatus Logico-Philosophicus of L. J. Wittgenstein: “Der Elementarsatz besteht aus Namen. Er ist ein Zusammenhang, eine Verkettung von Namen.” (4.22) “Der Name bedeutet den Gegenstand. Der Gegenstand ist seine Bedeutung. (‘A’ ist dasselbe Zeichen wie ‘A.’)” (3.203) “Der Konfiguration der einfachen Zeichen im Satzzeichen entspricht die Konfiguration der Gegenstände in der Sachlage.” (3.21) “Der Name vertritt im Satz den Gegenstand.” (3.22) For parallels in the more restricted domain of negative sentences, see note 305 below. 3 Needless to say, my use of this expression differs entirely from other uses thereof, whether in philosophy, physics, or elsewhere. 4 5 schools of thought, belonging to the three principal religious currents of classical India: Brahmanism, Buddhism, and Jainism. The main subject of these lectures pertains to a relatively brief period in the history of Indian thought. As is often the case in Indology, it is not possible to be very precise with dating. Very approximately, I would say that the period under consideration coincides with the first half of the first millennium, during which time the correspondence principle reigned virtually unchallenged. This does not mean that earlier thinkers were not convinced of a close connection unifying words and things. The acceptance of such a connection is already characteristic of Vedic literature, the earliest literature in our possession, as well as of post-Vedic literature from before the turn of the common era. The correspondence principle, however, extends well beyond the simple belief in such a connection, and appears only beginning in the first centuries of our era. 2. Early Brahmanical Literature By way of introduction, I will trace a few ideas from early Brahmanical literature bearing on the connection between words and things. Although these ideas differ from the correspondence principle in fundamental ways, in a sense they also constitute its background. Note that I will not be addressing, at least not primarily, the powers of speech (a topic brilliantly explored by Louis Renou for the Ṛgveda4), nor the theme of Speech as goddess (which has been treated by Charles Malamoud5). These Vedic speculations reveal the idea that words and things are much more closely connected than might initially appear. It is this connection that confers efficacy on magical formulas, and that allows one to draw conclusions about the nature of things on the basis of their names. It is obviously beyond the scope of these lectures, and beyond my abilities, to offer a comprehensive study of concepts related to the connection between words and things as expressed in Vedic literature. Instead, let us examine a few passages seeking to explain, in one way or another, the presence of such a connection. Let us begin with a passage from the Bṛhadāraṇyaka Upaniṣad: “At that time, indeed, the world was undivided. Name and form divided it [or: it was divided by name and form], such that one says: it has this name and that form. Today name and form divide this same world, such that one says: it has this name and that form.”6 This passage seems to assert that in the beginning the world knew no division between names and forms. That is to say, words and the things denoted by them formed a unity; words were not distinct from their objects. This interpretation, which stays close to the original Sanskrit, is naturally inseparable from the idea of a close connection between words and things. A passage from the Taittirīya Brāhmaṇa confirms this conclusion: “Prajāpati brought forth creatures. Those brought forth were conjoined. He entered them by 4 Renou, 1955. In his contributions to Detienne and Hamonic, 1995. 6 Bṛh-Up 1.4.7: tad dhedaṃ tarhy avyākṛtam āsīt / tan nāmarūpābhyām eva vyākriyatāsaunāmāyam idaṃrūpa iti / tad idam apy etarhi nāmarūpābhyām eva vyākriyate ‘saunāmāyam idaṃrūpa iti / 5 5 6 means of form (rūpa). That is why one says: ‘Prajāpati, truly, is form.’ (rūpaṃ vai prajāpatir iti). He entered them by means of name (nāman). That is why one says: ‘Prajāpati, truly, is name.’”7 A passage from the Chāndogya Upaniṣad likewise confirms the idea of a close connection between words and things. The idea is expressed thus: “This deity thought: ‘Let me separate name and form (nāmarūpa), entering with this living soul into these three deities. Let me make each of them threefold.’ The deity, entering with that living soul into those three deities, separated name and form.”8 Context shows that the three deities mentioned are heat (tejas), water (ap), and food (anna), which figure in the section immediately preceding ours; but this does not concern us at present. Nor is it important for us to know whether the deity separated name and form—as we have translated it—or whether it divided the three deities into name and form, as has been suggested.9 In either case one is justified in concluding that prior to this activity, name and form were not separate. A passage from the Śatapatha Brāhmaṇa teaches that Brahman entered into this world with form (rūpa) and name (nāman); the passage concludes with the assertion that this world extends as far as form and name.10 Through the narrative of an original unity of names and forms—of words and the things they denote—passages such as these imply that the separation of the two does not affect their true nature, that in reality they remain closely connected. This conviction explains the almost ubiquitous use of etymologies in the Brāhmaṇas of the Veda. It is not possible to enter into details here; let it suffice to recall that Vedic etymologies establish connections, usually of a mythic order, between diverse objects on the basis of resemblances between the words used to refer to them. The god Rudra is so called because he wails (rud-; ŚB 6.1.3.10); Agni owes his name to the fact that he was produced before (agre) the other gods (ŚB 2.2.4.2); sacrificial animals are called paśu because Agni beholds (paś-) them (ŚB 6.2.1.2); etc., etc. All of these etymologies presuppose, implicitly or explicitly, that names express something of the essence of the objects they denote, that the attribution of names to things is not accidental. Certain myths elaborate on this last idea. Names are closely related to the objects they denote, because ancient sages, whose knowledge was superior to ours, were the ones who named things. One encounters this myth already in the Ṛgveda, where it is presented in the following form: Bṛhaspati! When they [the poet-seers] set in motion the first beginning of speech, giving names (nāmadheyaṃ dadhānāḥ), their most pure and perfectly guarded secret was revealed through love. (1) 7 Taitt-Br 2.2.7.1: prajāpatiḥ prajā asṛjata / tāḥ sṛṣṭāḥ samaśliṣyan / tā rūpeṇānuprāviśat / tasmād āhuḥ / rūpaṃ vai prajāpatir iti / tā nāmnā ‘nuprāviśat / tasmād āhuḥ / nāma vai prajāpatir iti / Cf. Parpola, 1979: 148; Smith, 1989: 59. Sometimes Vāc (= speech) is presented as the wife of Prajāpati, or is connected to him in some other way; the separation of the two allows for the creation of beings; see Carpenter, 1994: 26. 8 Ch-Up 6.3.2-3: seyaṃ devataikṣata: hantāham imās tisro devatā anena jīvenātmanānupraviśya nāmarūpe vyākaravāṇīti / tāsāṃ trivṛtaṃ trivṛtam ekaikāṃ karavāṇīti / seyaṃ devatemās tisro devatā anenaiva jīvenātmanānupraviśya nāmarūpe vyākarot / 9 Hacker, 1950: 259 (82). 10 ŚBr 11.2.3.3: … tad dvābhyām eva pratyavaid rūpeṇa caiva nāmnā ca … / etāvad vā idaṃ yāvad rūpaṃ caiva nāma ca / 6 7 When the wise ones fashioned speech with their thought, sifting it as grain is sifted through a sieve, then friends recognized their friendships. A good sign was placed 11 on their speech. (2) The sages thus gave names to things after having purified, or sifted, their thought. Here we have a theme that reappears regularly in Sanskrit literature, showing that the connection between words and things must be especially close. For the sages had a much deeper knowledge of reality than ordinary mortals such as ourselves. Other texts, more recent than the Ṛgveda, take up the same idea, emphasizing the unique perceptual powers of these sages. The Nirukta (1.20), followed by the Mahābhārata (12.262.8), the Mahābhāṣya of Patañjali (ed. Kielhorn, vol. I, p. 11, l. 11 ff.), and other texts besides, employs in this context the expressions sākṣātkṛtadharman and pratyakṣadharma(n): the sages had a direct perception of dharma(n), i.e., of the universal order.12 The Vaiśeṣika Sūtra remarks that those who gave names to things were superior to us (asmadviśiṣṭa), because they perceived things before naming them.13 The Yuktidīpikā makes a distinction between names (saṃjñā) that are based on (the nature of) an object and those based on its appearance (svarūpa). The second type of word is merely a means of identifying an object; it functions by way of convention, regardless of the meanings of its constituent parts. The words gajakarṇa and aśvakarṇa are examples: though literally signifying “elephant ear” and “horse ear,” respectively, they usually refer to certain plants.14 The Yuktidīpikā goes on to observe: “The supreme seer (ṛṣi) [i.e., Kapila], who through his knowledge as a seer had beheld the appearances of all things, and who gave them names with great care (prayatnataḥ), does not use words that are based on appearance. … Likewise, the masters (ācārya) who follow [Kapila’s] point of view do not admit the introduction of new names, in view of the fact that [the supreme seer] uses only names [that are based on the nature of the object].”15 The Yuktidīpikā thus excludes certain Sanskrit words from the domain of use sanctioned by the sages of old. (At this point let me pause to recall an observation made by Professor Madhav M. Deshpande here in Paris a little more than a year ago, on the occasion of a colloquium in memory of Louis Renou.16 According to Deshpande, the Vedic 11 RV 10.71.1: b haspate prathamáṃ vācó ágraṃ yát praírata nāmadhéyaṃ dádhānāḥ / yád eṣāṃ śréṣṭhaṃ yád ariprám sīt preṇ tád eṣāṃ níhitaṃ gúhāvíḥ //1// sáktum iva títaünā punáṃto yátra dh rā mánasā v cam ákrata / átrā sákhāyaḥ sakhy ni jānate bhadraíṣāṃ lakṣm r níhit dhi vācí //2//. Tr. Doniger O’Flaherty, 1981: 61. 12 See Ruegg, 1994, esp. p. 307 ff. Eli Franco (1997: 30 n. 38), following other authors, argues for the interpretation of the word dharman as “thing.” The Carakasaṃhitā (Vimānasthāna 3.24) essentially uses the second expression to characterize people “in the beginning”: ādikāle … pratyakṣadevadevarṣidharmayajñavidhividhānāḥ … puruṣā babhūvu[ḥ] … 13 VS(C) 2.1.18-19: saṃjñākarma tv asmadviśiṣṭānāṃ liṅgam / pratyakṣapūrvakatvāt saṃjñākarmaṇaḥ /. I add a verse quoted without source by Viśvanātha Miśra (1996: 204): laukikānāṃ hi sādhūnāṃ vāg artham anuvartate / ṛṣīnāṃ punar ādyānāṃ vācam artho ‘nudhāvati // 14 YD p. 5, l. 5 ff. (Pandeya) / p. 7, l. 16 ff. (Wezler & Motegi). Elsewhere (p. 26, l.4. ff. [Pandeya] / p. 60, l. 3 ff. [Wezler & Motegi]) the Yuktidīpikā characterizes the words gajakarṇa and aśvakarṇa as saṃjñāśabda. 15 YD p. 5, l. 9-12 (Pandeya) / p. 7, l. 23-27 (Wezler & Motegi): prayatnato bhagavataḥ paramarṣer ārṣeṇa jñānena sarvatattvānāṃ svarūpam upalabhya saṃjñāṃ vidadhato nāsti svarūpanibandhanaḥ śabdaḥ / … / tanmatānusāriṇām apy ācāryāṇāṃ tābhir eva saṃvyavahārān nāsty apūrvasaṃjñāvidhānam praty ādaraḥ / 16 Deshpande, 1996: 153. 7 8 texts—especially the Brāhmaṇas—distinguish between two kinds of sages, namely, the ṛṣi and the ācārya. The authors of the metrical mantras are known as ṛṣi, while the authors of the Brāhmaṇas are referred to as ācārya. This distinction shows that already in the era of the Brāhmaṇas a special role was reserved for the ṛṣis, the authors of the ancient mantras later assembled in collections such as, notably, the Ṛgveda-Saṃhitā.) The close connection between words and things is also explained in a myth from the Manusmṛti. According to the first chapter of this text, in the beginning the creator created the names, actions, and states of all things from the words of the Veda.17 In the fourth chapter of the same text, it is said that all things are determined in speech; speech is their root, they issue forth from speech.18 The idea of the goddess Vāc, Speech, as creator of the world is not, of course, absent from the Ṛgveda19 and other Vedic texts,20 but the introduction to the Manusmṛti does not present speech as a mythical being; it is a matter, rather, of the words of the Veda. The idea of word or speech—often in the form of the Veda—as the origin of things is found in other texts, too. Already the Śatapatha Brāhmaṇa says: “Everything here is speech, for everything is obtained through speech.”21 And in the Taittirīya Brāhmaṇa one encounters the idea that Prajāpati created earth, sky, and heaven by pronouncing the words corresponding to them: bhūr, bhuvaḥ, suvar.22 The Maitrāyaṇī Upaniṣad speaks of two forms of Brahman, the one word, the other wordless; to attain the supreme Brahman (i.e., the wordless Brahman), one must be versed in the Śabdabrahman, the Brahman that is word.23 The Mahābhārata, in one of its recensions, has the following verse: “In the beginning a divine, eternal word, without beginning or end, consisting of the Veda, was pronounced by Svayaṃbhū—all activities proceed from it.”24 The Nāṭyaśāstra in turn observes: “The śāstras in this world consist of speech, and in speech they are established. Thus there is nothing higher than speech, for speech is the cause of all.”25 The thinker Bhartṛhari, at the beginning of his Vākyapadīya, refers to the Veda as the organizer, or creator—the Sanskrit word is vidhātṛ—of the world. He writes: “Different sciences unfold, based on the primary and secondary limbs (aṅga and upāṅga) of the creator of the worlds [i.e., the Veda], [sciences] which are the causes 17 See Manu 1.21: sarveṣāṃ tu sa nāmāni karmāṇi ca pṛthak pṛthak / vedaśabdebhya evādau pṛthak saṃsthāś ca nirmame // 18 Manu 4.256ab: vācy arthā niyatāḥ sarve vāṅmūlā vāgviniḥsṛtāḥ / 19 Brown, 1968. 20 Holdrege, 1994: 42 ff.; cf. Smith, 1994: 70 ff. 21 ŚB 10.5.1.3: vāg ghy evaitat sarvam / vācā hy evaitat sarvam āptam. Cf. ŚB 14.3.2.20. 22 Taitt-Br 2.2.4: sa bhūr iti vyāharat / sa bhūmim asṛjata / … / sa bhuva iti vyāharat / so ‘ntarikṣam asṛjata / … / sa suvar iti vyāharat / sa divam asṛjata / 23 MaiU 6.22-23: dve vāva brahmaṇī abhidhyeye śabdaś cāśabdaś ca / atha śabdenaivāśabdam āviṣkriyate / atha tatrom iti śabdaḥ / … evaṃ hy āha: dve brahmaṇī veditavye śabdabrahma paraṃ ca yat / śabdabrahmaṇi niṣṇātaḥ paraṃ brahmādhigacchati // … yaḥ śabdas tad om ity etad akṣaram / yad asyāgraṃ tac chāntam aśabdam abhayam aśokam ānandaṃ tṛptaṃ sthiram acalam amṛtam acyutaṃ dhruvam … 24 Mhbh 12.224.55 + 671*.1: anādinidhanā nityā vāg utsṛṣṭā svayaṃbhuvā / ādau vedamayī divyā yataḥ sarvāḥ pravṛttayaḥ //. Cf. Pollock, 1985: 518. 25 Nāṭyaśāstra 15.3: vāṅmayānīha śāstrāṇi vāṅniṣṭhāni tathaiva ca / tasmād vācaḥ paraṃ nāsti vāg ghi sarvasya kāraṇam // 8 9 of the mental traces (saṃskāra) of knowledge.”26 Another verse from the same text says: “Those who know the sacred tradition know that this [universe] is a transformation of the word. In the beginning this universe proceeds exclusively from Vedic verses.”27 And the great Śaṅkara states that the world, including the gods and everything else, is produced from the Vedic word; he says this in commenting on Brahma Sūtra 1.3.28, which seems to speak likewise.28 The belief in a close correspondence between words and things appears in a different context in Pāṇini’s grammar, in a rule governing the formation of plurals. To designate, say, three trees, there is no need to use the word “tree” three times. Pāṇini dictates that only one of the three words should remain (in the plural, of course); this is the meaning of the expression ekaśeṣa in the rule in question.29 This rule shows that Pāṇini fundamentally accepted the notion of direct correspondence, one word for each object. 3. Pāṇini’s Grammar Instead of citing further texts in which the close connection between words and things is expressed in a more or less mythological way, I propose to consider briefly the workings of grammar, one of the most ancient and admired sciences of India. Grammar—I have in mind primarily Pāṇini’s grammar, the oldest to have come down to us—produces words and sentences on the basis of verbal roots (dhātu), nominal themes (prātipadika), and suffixes (pratyaya). These constituent elements are invested with meaning. In a derivation they are joined together to form words and sentences that express or correspond, obviously enough, to the collection of meanings possessed by their constituent elements. In another lecture delivered in Paris, titled “Meaning-Bearing Linguistic Elements in the Sanskrit Grammatical Tradition,”30 I drew attention to a great gulf separating Pāṇini from Patañjali, the author of the Mahābhāṣya, or “Great Commentary.” For Patañjali, in contrast to Pāṇini, the real meaning-bearers are not the constituents of words but the words themselves. It seems that Kātyāyana, the author of the vārttikas contained in the Mahābhāṣya and therefore the predecessor of Patañjali, was already in agreement with the latter on this point. The reason for this difference between the grammarians is not obvious. After all, the chronological distance between them was not so great: two hundred years separate the probable dates of Pāṇini (ca. 350 B.C.E., or even later) and Patañjali (150 B.C.E.), and 26 Vkp 1.10: vidhātus tasya lokānām aṅgopāṅganibandhanāḥ / vidyābhedāḥ pratāyante jñānasaṃskārahetavaḥ //. The preceding verses leave no doubt that these lines refer to the Veda; and the verse itself speaks of “aṅga and upāṅga,” which could only apply to the Veda. Halbfass translates vidhātṛ as “organizing principle” (1991: 5) or “Organisationsprinzip” (1991a: 126). 27 Vkp 1.124: śabdasya pariṇāmo ‘yam ity āmnāyavido viduḥ / chandobhya eva prathamam etad viśvaṃ pravartate // 28 BS 1.3.28: śabda iti cen nātaḥ prabhavāt pratyakṣānumānābhyām. Śaṅkara comments (p. 96): ata eva hi vaidikāc chabdād devādikaṃ jagat prabhavati. 29 P. 1.2.64: sarūpāṇām ekaśeṣa ekavibhaktau. Cf. Renou, Terminologie p. 115, s.v. ekaśeṣa. 30 “Les éléments linguistiques porteurs de sens dans la tradition grammaticale du sanscrit,” published in Histoire Épistémologie Langage 20 (1), 1998, pp. 29-38. 9 10 Kātyāyana could be placed around 200 B.C.E.31 In my lecture I tried to provide a solution to this puzzle, ascribing the profound change in attitude to the influence of scholastic Buddhism. This particular solution, however, does not affect our present reflections. What concerns us at the moment is the difference in the two attitudes toward grammar. For Patañjali, grammar analyzes words, thereby arriving at their constituent parts, though the latter are not true bearers of meaning. This is why a good number of modern scholars have understood the word for grammar, vyākaraṇa, in the sense of analysis. For Pāṇini, grammar proceeds differently. His grammar does not divide words into stems and suffixes. On the contrary, it combines these constituent elements in order to form words. Paul Thieme, recognizing that Pāṇini’s grammar does not analyze but on the contrary proceeds as if the constituent elements of words were naturally given units, has proposed another interpretation of the word vyākaraṇa. He translates it as “[word-]formation,” a rendering he justifies with the following gloss: vividhena prakāreṇa (or viśeṣeṇa) ākṛtayaḥ kriyante yena, “[grammar is] an instrument by which forms are created in various ways” or “specifically.”32 Thieme defends his interpretation in an excursus added to his article “Meaning and Form of the ‘Grammar’ of Pāṇini” (1982: 23-34 [1170-1201]). There he discusses, to this end, several passages from Vedic literature. Some of them support his interpretation, but others suggest a different explanation. Let us consider these passages more closely. Vedic Sanskrit quite often uses the derivatives of vyākṛ-, i.e., the root kṛ with the two verbal prefixes vi and ā, in the sense of “separation, division, differentiation,” a fact that Thieme acknowledges. He gives several examples of such use, including the following passage from the Bṛhadāraṇyaka Upaniṣad (1.4.7), which we have already seen: “At that time, indeed, the world was undivided (avyākṛta). Name and form divided (vyākriyata) it [or: it was divided by name and form], such that one says: it has this name and that form. Today name and form divide (vyākriyate) this same world, such that one says: it has this name and that form.”33 This passage is of particular interest in that it uses the verb vyākṛ- to refer to the division—in the sense of differentiation—of names and forms at the beginning 31 For the date of Pāṇini, see Hinüber, 1990: 34: “… das Datum seiner Grammatik [kann] im Lichte der Numismatik kaum lange vor etwa 350 v. Chr. angesetzt werden”; and Falk, 1993: 304: “[es] fällt … schwer, Pāṇinis Regel über geprägte Münzen der ältesten Phase der indischen Numismatik zuzuordnen. … Damit … befinden wir uns schon in den Jahrzehnten nach 350 v. Chr.” For the date of Patañjali, see Cardona, 1976: 263 ff. Kātyāyana was contemporary with Aśoka, or slightly later; see Scharfe, 1971. 32 Thieme, 1982: 11 (1178), 23 (1190) ff., and already 1957: 267 (616). See in particular the following observations (1982: 11 [1178]): “Pāṇini does not analyze. … The method of his representation is just the contrary of an analysis. He proceeds as if the elements, into which he has dissected the word forms, were naturally given units. He does not demonstrate how wordforms can be analyzed into their constituent functional elements by methodical deductions and inferences [of the kind discussed by Kātyāyana and Patañjali], rather, he presupposes these elements and shows in which, sometimes highly complicated, ways they are to be combined.” See also Deshpande, 1996: 149-151. 33 Thieme translates: “All this was unseparated (indistinguishable) [in the beginning of creation]. Then it became separated (distinguished) by name and shape [so it became possible to say]: ‘This particular one is of the name NN and of such and such a shape.’ Therefore, even to-day distinction is made by name and shape: ‘This particular one is of the name NN [and] of such and such a shape.’” 10 11 of time.34 It suggests that the function of grammar is to effect such a differentiation. We have seen that several Vedic passages mention this initial differentiation of words and things. Some of these passages, such as the ones I have already cited from the Bṛhadāraṇyaka and the Chāndogya Upaniṣad, use the verb vyākṛ- in this context. To give the latter passage again: “This deity thought: ‘Let me separate (vyākaravāṇi) name and form (nāmarūpe), entering with this living soul into these three deities. Let me make each of them threefold.’ The deity, entering with that living soul into those three deities, separated (vyākarot) name and form.” Thieme, however, cites this passage from the Chāndogya Upaniṣad in support of another interpretation of the verb vyākṛ-. He translates it as “drive asunder,” an expression still close to our “separate,” but a parenthesis adds “unfold, form in various ways.”35 According to Thieme’s interpretation, if I understand it correctly, the deity formed name and form, probably because they did not yet exist. But there is no need to accept this. The previous passage, from the Bṛhadāraṇyaka Upaniṣad, taught that names and forms, although undifferentiated, already existed in the beginning. The present passage can be understood in the same way. If one accepts the existence of names and forms from the start, there is no need to supply a new interpretation of the verb vyākṛ-. These two passages suggest, therefore, another explanation for the expression vyākaraṇa, distinct from Thieme’s interpretation. In light of these passages, the expression denotes a separation, or differentiation, not of stems and suffixes, but of linguistic elements from the objects they denote. This separation, judging from Pāṇini’s grammar, produces meaning-bearing elements, based on denoted objects, and these elements are subsequently combined, thus producing words and sentences. In this context it is important to bear in mind the role that meanings play in Pāṇini’s grammar. They are the starting-point for derivations: meanings give rise to semantic elements, which in turn are brought together to form the words and sentences of the Sanskrit language.36 Meanings are the objects denoted. They constitute the point of departure, while grammar creates the corresponding verbal expressions. Grammar thus brings together the primitive elements of the language while separating them from their objects. It is in this way, it seems to me, that one arrives at an interpretation of the word vyākaraṇa that proves satisfying from the point of view of its form as well as its cultural context. This interpretation confirms, moreover, that thinkers of the period were deeply convinced of the close connection linking words and things. Incidentally, the lack 34 Note that certain parallel passages use the expression vyāvṛt- instead, which confirms that the idea of separation is indeed intended. See, e.g., PB 24.11.2: prajāpatiḥ prajā asṛjata tā avidhṛtā asañjānānā anyonyām ādaṃs tena prajāpatir aśocat sa etā apaśyat tato vā idaṃ vyāvartata gāvo ‘bhavann aśvā aśvāḥ puruṣāḥ puruṣā mṛgā mṛgāḥ—“Prajāpati created the creatures; these, not being kept apart (and) not agreeing together, devoured each other. This pained Prajāpati. He saw these days (i.e., this fortynine-day-rite). Thereupon, this became separated (vyāvartata) (i.e., all the kinds of beings kept apart): cows (became) cows; horses (became) horses; men (became) men; deer (became) deer.” Tr. Caland, 1931: 612-13. The Bṛhadāraṇyaka Upaniṣad (1.4.3) uses the causative of pat- to recount how the initial body (ātman) was divided into two: sa imam evātmānaṃ dvedhāpātayat. 35 His full translation (p. 24 [1191]): “This divine element (devatā) considered: Well then, I shall enter these three divine elements (heat/fire: tejas-, water: āp-, food: anna-, i.e. matter) by this living Self and drive asunder (unfold, form in various ways) name and shape. … Then this divine element entered etc.” 36 Bronkhorst, 1980; Houben (1997: 84 ff.) expresses a slightly different point of view. 11 12 of interest on the part of modern scholars in the semantic aspect of Pāṇini’s grammar would explain why investigations of the word vyākaraṇa have failed to take into account the interpretation suggested by the Vedic passages we have just examined. Let us examine some further Vedic passages. The Taittirīya Saṃhitā contains the following myth: “Speech, not being separated (avyākṛtā), spoke remaining turned aside. The gods said to Indra: ‘Separate (vyākuru) this Speech for us.’ … Indra placed himself in the middle and separated (vyākarot) Speech.”37 Here it is less obvious that the intended meaning is “separation.” Some authorities interpret vyākṛ- in this passage as meaning “articulation” instead.38 Thieme cites the passage to support his interpretation, “form in various ways.”39 As we have seen, however, the formation of Speech comes about by means of a differentiation, the separation of Speech from the objects it denotes. If one again keeps to the basic sense of the verb vyākṛ-, another aspect of the passage also becomes more intelligible. Why did Indra place himself “in the middle” before separating Speech, and in the middle of what? The answer seems to be as follows: he placed himself between words and denoted objects, for the separation of Speech is its separation from the objects it denotes. This passage uses the verb vyākṛ- in a sense one might possibly translate with the verb “to form.” The reason is simple: word-formation is fundamentally a differentiation. This explains the use of the verb in other passages with a sense closer to that of “forming.” One such passage comes from the Aitareya Upaniṣad: “Which is this Self: that by which one sees, that by which one hears, that by which one smells fragrances, that by which one forms (vyākaroti) speech, or that by which one distinguishes the sweet from the non-sweet?”40 Comparison with Aitareya Brāhmaṇa 5.22 (= 24.3) suggests that one forms speech and distinguishes the sweet from the non-sweet by means of the tongue, palate, and teeth, inside the mouth. This passage reads as a sort of complement to the passage from the Aitareya Upaniṣad, in that it goes on to associate that which is in the nostrils with that by which one smells fragrances, the dark of the eye with that by which one sees, and the inside of the ear with that by which one hears.41 These are the Vedic passages, along with some others, that Thieme cites in support of his interpretation. It seems clear that the sense of “separate, differentiate” is undeniable in certain passages, and that the sense of “form” is 37 Taitt-S 6.4.7.3: vāg vai parācy avyākṛtāvadat / te devā Indram abruvann imāṃ no vācaṃ vyākurv iti … tām Indro madhyato ‘vakramya vyākarot / 38 E.g. Scharfe, 1977: 80. 39 He translates (Thieme, 1982: 23-24 [1190-91]): “[Human] speech used to speak being turned away (understandably), being unformed (unarticulated). Then the heavenly said to Indra: Do form us this speech … Then Indra formed it (gave it different forms, made it articulate), having stepped in the middle of it. Therefore this [human speech] is spoken being formed (having different forms, being articulate).” 40 AitUp 3.1-2 (= AitĀr 2.6): kataraḥ sa ātmā? yena vā paśyati, yena vā śṛṇoti, yena vā gandhāñ jighrati, yena vā vācaṃ vyākaroti, yena vā svādu cāsvādu ca vijānāti / 41 AitBr 5.22 (= 24.3): … tad yathā ‘ntaraṃ mukhasya jihvā tālu dantā evaṃ chandomā atha yenaiva vācaṃ vyākaroti yena svādu cāsvādu ca vijānāti tad daśamam ahaḥ … tad yathā ‘ntaraṃ nāsikayor evaṃ chandomā atha yenaiva gandhān vijānāti tad daśamam ahaḥ … tad yathā ‘ntaraṃ akṣṇaḥ kṛṣṇam evaṃ chandomā atha yaiva kanīnikā yena paśyati tad daśamam ahaḥ … tad yathā ‘ntaraṃ karṇasyaivaṃ chandomā atha yenaiva śṛṇoti tad daśamam ahaḥ 12 13 derivative. It is likely that Patañjali, for whom the process of the separation of words from the objects they denote was no longer topical, used the verb only in the sense of “forming.” This is what Thieme seeks to prove, successfully in my view. But Patañjali’s opinion is no longer decisive when investigating the original meaning of the term vyākaraṇa. For Patañjali, the constituent elements of words were no longer true bearers of meaning. One can conclude from this that for him the nature of grammar had changed dramatically with respect to Pāṇini. Whatever one thinks of this explanation of the original sense of vyākaraṇa, one thing is clear. We have numerous indications that the connection between words and things was conceived of as particularly close during the Vedic period. And although certain specific ideas—such as that of the separation of words from the objects they denote—perhaps did not survive very long after the end of the Vedic era, the notion of a profound correspondence remained. It is expressed in the repeated use of etymologies. It is sometimes expressed in the observation, already present in a passage from the Ṛgveda, that the sages, on the basis of their superior knowledge, gave names to things. Sometimes, the Veda itself is presented as the organizer of creation. And at a less theoretical level, the notion is expressed in the frequent use of mantras, whether Vedic or non-Vedic; mantras were believed to influence objective, non-linguistic reality, and this by reason of the connection linking words, and sometimes also sounds, to things. 4. A Passage from the Chāndogya Upaniṣad We cannot leave Vedic literature without mentioning a passage from the Chāndogya Upaniṣad, one that accepts the close connection between words and things but nevertheless denies that words give access to a deeper reality than the world of our everyday experience. The passage takes the opposite position: speech conceals, so to speak, the reality hidden behind it. This passage—as you have perhaps already guessed—is the one in which Śvetaketu receives instruction from his father. His father tells him: Just as, my dear, one can know all that is made of clay through a single piece of clay—[all that] rests on words, is a modification, is only a name; the reality is the clay— Just as, my dear, one can know all that is made of copper through a single copper ornament—[all that] rests on words, is a modification, is only a name; the reality is the copper— Just as, my dear, one can know all that is made of iron through a single nail-cutter—[all that] rests on words, is a modification, is only a name; the reality is the iron— 42 Even so, my dear, is this teaching. The position of Śvetaketu’s father is particularly interesting for our purpose, because it is rather close to the position that came to be developed within Buddhism, to which we shall turn momentarily. The presence of this unusual 42 Ch-Up 6.1.4-6: yathā somyaikena mṛtpiṇḍena sarvaṃ mṛnmayaṃ vijñātaṃ syāt / vācāraṃbhaṇaṃ vikāro nāmadheyaṃ, mṛttikety eva satyam // yathā somyaikena lohamaṇinā sarvaṃ lohamayaṃ vijñātaṃ syāt / vācāraṃbhaṇaṃ vikāro nāmadheyaṃ, loham ity eva satyam // yathā somyaikena nakhanikṛntanena sarvaṃ kārṣṇāyasaṃ vijñātaṃ syāt / vācāraṃbhaṇaṃ vikāro nāmadheyaṃ, kṛṣṇāyasam ity eva satyam / evaṃ somya sa ādeśo bhavatīti //. Cf. Kuiper, 1957; Olivelle, 1996: 346, note on Ch-Up 6.1.4-6; Houben, 1997: 68. 13 14 opinion in the Chāndogya Upaniṣad does not seem to be the result of Buddhist influence, however, since the parallel Buddhist position did not exist in early Buddhism, as we shall see. It seems much more likely that this position—according to which words are associated with modifications, and the deeper reality is not reached—is further proof of the wealth of ideas present in Brahmanical religion in the period of the early Upaniṣads.43 It is noteworthy that this passage from the Chāndogya Upaniṣad allowed Brahmins of a much later period to accept Buddhist ideas while maintaining that they remained faithful to the Veda. 5. The Structures of Languages Before turning to the main thesis of these lectures, let us briefly look at certain ideas regarding the structures of languages, as found in early, though post-Vedic, texts. Familiarity with these ideas will help us to understand later developments concerning the connection between words and things. The Vedic ideas we have just dealt with concerned the forms of words. The connection that exists between a word and the object it denotes relates the form of the word to the object. Etymologies derive their value from this fact, and the efficacy of mantras is explicable solely on this basis. Certain later thinkers, however, among them the Buddhists, considered word forms to be conventional, and therefore devoid of intrinsic meaning. For them, one would imagine, the structure of the language used—Sanskrit in the cases we shall be looking at—was bound to come into consideration, with the inevitable question of whether the structure of Sanskrit is shared by other languages. Are the conclusions we draw from the use of one language universal? Or do they have value only within the context of that particular language? To us, these questions might appear natural and inevitable. The thinkers of ancient India, however, do not seem to have been much troubled by them. This was not simply due to Sanskrit being considered the only true language. A significant number of Brahmins doubtless believed this, but others, such as the Buddhists, did not share their conviction. The lack of concern on the part of all these thinkers would seem to be due instead to their being convinced that the structure of Sanskrit does not actually differ from that of other languages. The very question of the structures of different languages did not arise. The texts speak of other idioms as if their only difference consisted in word forms.44 Let us examine the following passage from the Pāli Buddhist canon. It is found in the Araṇavibhaṅga Sutta of the Majjhima Nikāya: “How, O monks, is one not attached to the language of a region, and how does one not go beyond popular usage? Here, O monks, in different regions one calls the same object (for example, a bowl) either pāti, or patta, or vittha, or sarāva, or dhāropa, or poṇa, or pisīla. One 43 The idea that the supreme Brahman is beyond the domain of words is expressed from time to time in the Upaniṣads, as e.g. Taitt-Up 2.4 (= 2.9): yato vāco nivartante aprāpya manasā saha— “From which words, as well as the mind, turn back without reaching it.” Note that Ruben (1979: 66 ff.) interprets the position of Śvetaketu’s father as a kind of materialism/hylozoism. 44 A similar view is not unknown in Western tradition; see Umberto Eco’s observations (1995: 200) on the polygraphy of Athanasius Kircher, which was based on the supposition that all languages are directly reducible to the grammar of Latin. 14 15 expresses oneself without attachment, using the same term as the people in each region, thinking: ‘In designating this, these venerable ones express themselves thus.’”45 This passage shows an awareness of linguistic differences among the various regions of north India. But only one aspect of these differences is mentioned here, namely, differences of vocabulary. There is no question of differences of structure among these languages. We find the same exclusive interest in word forms in different languages in the Mahābhāṣya of Patañjali. At the beginning of this text we learn that besides the correct word, go (“cow”), there are many degraded words such as gāvī, goṇī, gotā, gopotalikā, etc.46 Elsewhere in the same text it is said that incorrect forms such as āṇapayati, vaṭṭati, and vaḍḍhati exist alongside the correct forms ājñāpayati, vartate, and vardhate.47 In the world, one also uses the root kas instead of kṛṣ and diś instead of dṛś.48 Other languages—so these examples seem to suggest—are identical to Sanskrit (or to Pāli), with the sole exception of the form of their words. With the grammarian Bhartṛhari—who belongs to the fifth century of the common era and thus, in all likelihood, postdates Patañjali by about six hundred years—we again find the same position regarding the diversity of languages: he speaks only of differences in words. Certain words are correct, others are not. He admits, however, that in certain communities incorrect words are directly denotative: “For those who do not know how to speak properly, erroneous forms are well known by an uninterrupted transmission, such that for them, the correct forms are not denotative.”49 The Dutch scholar Jan Houben (1993: 149) sees in this concession a reflection of the linguistic situation in India during Bhartṛhari’s era: the simultaneous presence of many languages and dialects in parallel with Sanskrit, the latter still being invested with an incomparably higher status than the others. Let us also examine a passage from the Nyāya Bhāṣya, which discusses the conventional nature of the connection between words and things. At the heart of this discussion we find the following observation, offered as an argument against a natural connection between the two. The sentence in question says: “The Ṛṣis, the Āryas, and the barbarians employ words as they will to make meanings understood.”50 What interests us here is the listing of barbarians (mleccha) alongside Ṛṣis and Āryas, thus alongside exemplary users of Sanskrit. Barbarians, however, do not use Sanskrit. They speak other languages. But what exactly is the difference between Sanskrit and these other languages? Like the preceding passages, the present passage suggests that it is the form of their words. The barbarians choose 45 MN III p. 235: kathañ ca, bhikkhave, janapadaniruttiyā ca anabhiniveso hoti samaññāya ca anatisāro? idha, bhikkhave, tad ev’ ekaccesu janapadesu pātī ti sañjānanti, pattan ti sañjānanti, vitthan ti sañjānanti, sarāvan ti sañjānanti, dhāropan ti sañjānanti, poṇan ti sañjānanti, pisīlan ti sañjānanti / iti yathā yathā naṃ tesu tesu janapadesu sañjānanti, idaṃ kira ‘me āyasmanto sandhāya voharantīti, tathā tathā voharati aparāmasaṃ / 46 Mbh I p. 10 l. 8-9: bhūyāṃso ‘paśabdā alpīyāṃsaḥ śabdāḥ / ekaikasya śabdasya bahavo ‘pabhraṃśāḥ / tad yathā / gaur ity asya gāvī goṇī gotā gopotalikety evamādayo ‘pabhraṃśāḥ / On goṇī see Thieme, 1978: 48 (905) n. 52. 47 Mbh I p. 259 l. 4-7 (P. 1.3.2 vt. 12 with his bhāṣya): bhūvādipāṭhaḥ prātipādikāṇapayatyādinivṛttyarthaḥ (vt.) / … ke punar āṇapayatyādayaḥ / āṇapayati vaṭṭati vaḍḍhatīti / 48 Mbh I p. 259 l. 14 (on P. 1.3.2 vt. 13): loke hi kṛṣyarthe kasiṃ prayuñjate dṛśyarthe ca diśim. For all these examples, see Deshpande, 1993: 26, who seems however to read dis instead of diś. 49 Vkp 1.181: pāraṃparyād apabhraṃśā viguṇeṣv abhidhātṛṣu / prasiddhim āgatā yena teṣāṃ sādhur avācakaḥ //. Translation based on Biardeau, 1964a: 191. 50 NBh 2.1.56: ṛṣyāryamlecchānāṃ yathākāmaṃ śabdaviniyogo ‘rthasampratyāyanāya pravartate. 15 16 forms that are not allowed, that are false, while the Ṛṣis and the Āryas, it goes without saying, use only correct words. The following impression emerges from these passages: fundamentally, everyone speaks the same language; the differences one notes in various regions result from the fact that many speakers use incorrect words. One may add that these incorrect words represent so many corruptions of Sanskrit words, which for their part are correct and original.51 This description does not, of course, answer literally to what we find in these early texts, but it does seem to capture its essence. We occasionally meet with the idea in early texts that Sanskrit is used throughout the world, even on other continents, which supports the picture we have just sketched. The Mahābhāṣya expresses this notion in a passage that specifies the domain of Sanskrit—literally, “the domain of the use of word” (śabdasya prayogaviṣayaḥ), but context shows that Sanskrit is meant—listing, among other things, the seven continents, the earth, and the three worlds.52 The Mīmāṃsā Bhāṣya makes a similar observation in the course of a discussion in which it tries to prove that the connection between words and things could not have been imposed by an agent: Moreover, there is no divergence: the word gauḥ applies to that animal which has a dewlap etc., in this country as in all countries, even the most distant. How could multiple agents have come together [to agree]? And a single agent would not be able [to impose his decision everywhere]. Therefore the connection has no 53 creator. It goes without saying that we must read this passage in light of another from the same commentary, which indicates that the same word gauḥ is not exactly used by everyone. On the contrary, the Mīmāṃsā Bhāṣya (on sūtra 1.3.24), following the example from Patañjali’s Mahābhāṣya that we examined, mentions the words gāvī, goṇī, and gopotalikā as variants of gauḥ, while specifying that these variants are not denotative. The fact that the first passage is borrowed from another author, the Vṛttikāra, while the second passage comes from the pen of Śabara himself, does not prevent us from taking the observation that the word gauḥ is used “in this country as in all countries, even the most distant” with some reservation. Doubtless there were many countries, even in the mind of the Vṛttikāra, where people used one of these erroneous variants instead of the correct word gauḥ. In other words, in principle Sanskrit is spoken everywhere, but many speakers infelicitously replace correct words with incorrect words. Thus different languages, each exhibiting its 51 See Bronkhorst, 1993b. Mbh I p. 9 l. 20-23 (Paspaśāhnika on vt. 5): saptadvīpā vasumatī trayo lokāś catvāro vedāḥ sāṅgāḥ sarahasyā bahudhā vibhinnā ekaśatam adhvaryuśākhāḥ sahasravartmā sāmaveda ekaviṃśatidhā bāhvṛcyaṃ navadhātharvaṇo vedo vākovākyam itihāsaḥ purāṇaṃ vaidyakam ity etāvāñ śabdasya prayogaviṣayaḥ. A verse from the Manusmṛti (10.45) seems to corroborate that Sanskrit (?; āryavāc) is spoken outside of India, while adding that other languages (mlecchavāc) are also used: mukhabāhūrupajjānāṃ yā loke jātayo bahiḥ / mlecchavācaś cāryavācaḥ sarve te dasyavaḥ smṛtāḥ //— “Those castes (jāti) that fall outside the world of those who were born from the mouth, arms, thighs, and feet [of the primordial man], whether they speak barbarian languages or Sanskrit, are all considered dasyu.” 53 MīBh 1.1.5 (Frauwallner, 1968: 46): avyatirekaś ca / yathā asmin deśe sāsnādimati gośabdaḥ, evaṃ sarveṣu durgameṣv api / bahavaḥ saṃbandhāraḥ kathaṃ saṃgaṃsyante / eko ‘pi na śaknuyāt / ato nāsti saṃbandhā /. Translation based on Biardeau, 1964: 160. 52 16 17 own structure, do not exist. Ultimately there is only Sanskrit, and other languages in principle share its structure. One must wonder to what degree this conviction was able to hold sway thanks to the prohibition, found in certain Dharmaśāstra texts, against learning the languages of the mlecchas, the barbarians.54 Be that as it may, it seems justified to conclude that those in India who considered Sanskrit to be the perfect and original language also regarded it, in a certain sense, as a universal language.55 6. The Buddhist Contribution The belief in a close connection between words and things forms the background for the developments we shall be studying in what follows. We shall see how a new tendency, originating in certain developments within Buddhism, came to replace this more or less general belief with much more specific ideas. Let us note first that Buddhism, though less obsessed with speech than Brahmanism, periodically uses etymologies, even in its earliest literature, which suggests that a connection between language and reality was implicitly accepted. If it is true, as the English scholar Richard Gombrich maintains, that some of the etymologies in early Buddhist literature had no other goal than to mock the Brahmins,56 one might conclude that the Buddha himself, or one of his disciples, attached no real value to these etymologies, and perhaps none to etymologies in general. Be that as it may, it did not take long for Buddhists of a later period to forget their master’s skepticism: the etymologies one finds in their works by no means give the impression that they are mocking anyone. The pair “name and form,” nāmarūpa, also appears in early Buddhist texts. But it is no longer a question of names in the strict sense of denominations. On the contrary, these texts use the expression rūpa, “form,” to refer to the physical aspect of a person, the first of the five groups (skandha) of factors constituting the person; the expression nāman is reserved for the other four groups—vedanā, saṃjñā, saṃskāra, vijñāna—which together constitute the non-physical, hence psychic, aspect of a person. This new use of the term nāman shows that the original division of the world into objects and their names had lost its meaning for Buddhists. They maintained the tradition of a division into nāman and rūpa, reinterpreting the term nāman in a psychic rather than a linguistic sense.57 What interests us here, however, is neither the etymologies of the Buddhists nor their use of the term nāman, but rather some developments in scholastic Buddhism. For reasons that historical research will have to make clear, Buddhism, in the centuries following the death of its founder, turned progressively toward an enumeration of the essential elements of the latter’s teaching. This resulted in lists of what are called “the dharmas.” These dharmas were classified in accordance with the five “groups” (skandha), or constituents of a person. The existence of the person as such was nevertheless not recognized by the Buddhists. The conclusion was 54 HistDh Vol. II, Part I, p. 383. This set of beliefs may be compared to the search for the perfect, universal, and original language in Medieval and Renaissance Europe; see Eco, 1995: 73 ff. 56 Gombrich, 1992: 163. 57 PTSD p. 350 s.v. nāma; cf. e.g. May, 1959: 253. Cf. also Buswell, 1997: 566-67. 55 17 18 reached that the five groups, and therefore the dharmas, were the truly existent parts of the person, the person itself not existing. This idea was then extended, at least in certain Buddhist schools, in such a way as to cover all macroscopic objects, which are wholly constituted by the dharmas. In the final analysis, only the dharmas, the ultimate constituents, exist, and not the objects constituted. At the end of these developments, the list of dharmas was presented as an enumeration of all that exists. Buddhism then had an ontology at its disposal, with the dharmas representing the only truly existent things. The objects of our experience—such as houses, chariots, or persons—in reality do not exist. But it is not so easy to rid oneself of the phenomenal world. It is possible that persons, the houses in which they live, and the vehicles they use do not truly exist. That does not change the fact, however, that everyone believes in their existence. Where does this belief in non-existent objects come from? Here another consideration comes into play: the objects of our experience do not truly exist; they are only names, they are existent by denomination. In a moment we shall explore the significance and consequences of this idea. Let us note here that it is often expressed in Buddhist texts from continental India from a certain period on. Its classical expression is in a passage from the Milindapañha, “The Questions of King Milinda,” which is often quoted, and for good reason. This text presents the idea not only in perhaps its earliest form, but also in the context of a delightful story, that of the meeting of the monk Nāgasena with the Greek king Milinda. I will present a few extracts from this story, which might date from a period close to the reign of King Milinda, who has been identified as the Indo-Greek king Menander from the second century B.C.E. The idea that interests us might therefore also date from the second century B.C.E.58 Let us now consider the passage: King Milinda asked him: “How is your Reverence known, and what is your name, Sir?” “As Nagasena I am known, O great king, and as Nagasena do my fellow religious habitually address me. But although parents give such names as Nagasena, or Surasena, or Virasena, or Sihasena, nevertheless this word ‘Nagasena’ is just a denomination, a designation, a conceptual term, a current appellation, a mere name. For no real person can here be apprehended.” When the king expresses his skepticism, Nāgasena clarifies his position with the example of a chariot: “Is the pole the chariot?” – “No, reverend Sir!” – “Is then the axle the chariot?” – “No, reverend Sir!” – “Is it then the wheels, or the framework, or the flag-staff, or the yoke, or the reins, or the goad-stick?” – “No, reverend Sir!” – “Then is it the combination of pole, axle, wheels, framework, flag-staff, yoke, reins, and goad which is the ‘chariot’?” – “No, reverend Sir!” – “Then is this ‘chariot’ outside the combination of pole, axle, wheels, framework, flag-staff, yoke, reins, and goad?” – “No, reverend Sir!” – “Then, ask as I may, I can discover no chariot at all. Just a mere 59 sound is this ‘chariot.’” 58 De Jong (1996: 383) finds this date for the text too early but does not propose another. Mil p. 25 l. 5-13 & p. 27 l. 2-14: atha kho milindo rājā āyasmantaṃ nāgasenaṃ etad avoca: katham bhadanto ñāyati, kinnāmo si bhante ti? nāgaseno ti kho ahaṃ mahārāja ñāyāmi, nāgaseno ti maṃ mahārāja sabrahmacārī samudācaranti / api ca mātāpitaro nāmaṃ karonti nāgaseno ti vā sūraseno ti vā vīraseno ti vā 59 18 19 After these questions the king admits: “[I]t is in dependence on the pole, the axle, the wheels, the framework, the flagstaff, etc., that there takes place this denomination ‘chariot,’ this designation, this conceptual term, a current appellation and a mere name.” After which Nāgasena adds: “It is just so with me. In dependence on the thirty-two parts of the body and the five Skandhas there takes place this denomination ‘Nagasena,’ this designation, this conceptual term, a current appellation and a mere name. In ultimate reality, however, this person cannot be apprehended.” There is no need to cite passages from other Buddhist texts that express the point analogously.60 The central ideas are clear. Composite entities do not exist; only their parts truly exist. The ultimate parts of these composite entities are the dharmas. In the final analysis, therefore, the dharmas constitute all that exists. Composite entities, for their part, do not have real existence and are only names, designations. They are “existent by designation” (prajñaptisat), as the texts often affirm. Obviously, these ideas profoundly affect one’s view of the world. For the Buddhists who adhere to them, the world of our experience is no longer what it appears to be. The phenomenal world is at base only an illusion. Scholastic Buddhism allows one to grasp the reality hidden behind phenomena, a reality that is quite different from what our imagination and language would have us believe. Reality consists in the dharmas and nothing else. The Buddhist scholastics took pains to list the dharmas exhaustively, so much so that their writings offer us complete enumerations of everything that exists. The following characteristics of this Buddhist ontology should be kept in mind. All that exists is contained in the lists of dharmas or elements of existence. The objects that make up the phenomenal world are constituted of these dharmas but do not truly exist. We nonetheless believe that they exist, on account of the words of our language. 7. Vaiśeṣika and Language It goes without saying that this Buddhist ontology could scarcely have seemed attractive to non-Buddhist thinkers. Why accept that the objects of our experience, including the very persons we are, do not really exist? It is hardly astonishing to find that no one in India agreed with the Buddhists.61 What is surprising is that sīhaseno ti vā, api ca kho mahārāja saṅkhā samaññā paññatti vohāro nāmamattaṃ yad idaṃ nāgaseno ti, na h’ettha puggalo upalabbhatīti / … / kin nu kho mahārāja īsā ratho ti / na hi bhante ti / akkho ratho ti / na hi bhante ti / cakkāni ratho ti / nahi bhante ti / rathapañjaraṃ ratho ti / na hi bhante ti / rathadaṇḍako ratho ti / na hi bhante ti / yugaṃ ratho ti / na hi bhante ti / rasmiyo ratho ti / na hi bhante ti / patodalaṭṭhi ratho ti / na hi bhante ti / kin nu kho mahārāja īsā-akkha-cakka-rathapañjara-rathadaṇḍa-yuga-rasmi-patodaṃ ratho ti / na hi bhante ti / kim pana mahārāja aññatra īsā-akkha-cakka-rathapañjara-rathadaṇḍa-yuga-rasmi-patodaṃ ratho ti / na hi bhante ti / tam ahaṃ mahārāja pucchanto pucchanto na passāmi rathaṃ, saddo yeva nu kho mahārāja ratho /. Tr. Conze, 1959: 147-49. 60 See Bronkhorst, 1996: 114 ff. Hattori (1977: 52 ff.) maintains that the idea of nominal existence had been developed by the Sautrāntikas; Katsura (1991: 134 n. 17), however, has rightly pointed out that the idea was accepted by the majority of the schools of Abhidharma. 61 With the likely exception, of course, of the passage from the Chāndogya Upaniṣad (6.1.4-6) discussed above. 19 20 efforts to produce a Brahmanical ontology gave rise to a system that, despite its opposition to Buddhist positions, remained extremely close to them. I have in mind the Vaiśeṣika system, which later exerted a profound influence on other Brahmanical systems of thought, in particular Nyāya and Mīmāṃsā. Recall that ancient India produced essentially two Brahmanical ontologies, two systematic accounts of what exists: Sāṃkhya and Vaiśeṣika. Sāṃkhya represents a systematization of ideas found already in earlier literature, first in the Mahābhārata, but also in certain Upaniṣads. Several slightly different forms of the classical system are known to us, one of which is expressed in the texts of classical Yoga. Though the elements of Sāṃkhya can be found in religious literature up to a much later period, Sāṃkhya did not meet with great success as a philosophical system. It suffered from ambiguities or even contradictions that perhaps contributed to its relatively rapid decline. Vaiśeṣika ontology is different. To begin with, it represented a new creation. There is no trace of it in previous literature. The Vaiśeṣika Sūtra, the root text of the system, seems to constitute its earliest record. This text is preserved in five different versions,62 each of which already contains additions from later periods, making reconstruction of the original text a hopeless task. This does not change the fact that the Vaiśeṣika system, in one form or another, appears to be a new creation and not the result of an organic development. We are thus justified in thinking that at a certain moment in Indian history, someone conceived of a system that would be the predecessor of the classical system that has come down to us. If I am keen to stress the role of an individual—or possibly several individuals—in this creation, it is because systems of thought are never the result of organic developments. Even if, as in the case of Sāṃkhya, all the elements were already present in earlier texts, their combination into a new system still required a human agent. In the case of Vaiśeṣika, the influence of an older tradition seems decidedly less important than in Sāṃkhya, and the contribution of the founder seems, consequently, greater. That we do not know the exact form of the original Vaiśeṣika, or the name and date of its actual founder, does nothing to change the fact that someone at some point in time had to have created the system in its earliest form. The motive of Vaiśeṣika’s founder is equally unknown to us. We can, however, work out a hypothesis, which might also shed light on the original form of the system. Having already treated this subject elsewhere,63 here I will limit myself to the essentials. Vaiśeṣika texts more recent than the Vaiśeṣika Sūtra present a coherent system—or at least one that aspires to coherence—demonstrating several traits that are remarkable to say the least. Vaiśeṣika enumerates a certain number of “categories”—the term most often used to translate the Sanskrit padārtha— supposed to encompass everything that exists. The number of categories varies according to the text: the Padārthadharmasaṅgraha, dating perhaps to the sixth century, has six, while later texts have seven, and the *Daśapadārthaśāstra or 62 To the versions commented on by Śaṃkara Miśra, Bhaṭṭa Vādīndra, and Candrānanda, Harunaga Isaacson has been able to add two further versions. See Isaacson, 1995: 200 ff. 63 Bronkhorst, 1992b. 20 21 *Daśapadārthī,64 which is difficult to date, lists ten categories. For now, it suffices to bear in mind that these categories cover all existent entities, and that their enumeration is thus an enumeration of everything that exists. To take a concrete example, the six categories of the Padārthadharmasaṅgraha are substance (dravya), quality (guṇa), motion (karman), universal (sāmānya), particularity (viśeṣa), and inherence (samavāya). This means that for the author of this text, everything that exists is either a substance, quality, or motion, or a universal, particularity, or inherence. Nothing exists beyond these six categories. The categories admit of internal divisions, of course. There are, for instance, nine substances, twenty-four qualities, five motions, etc.; but this does not alter the fact that Vaiśeṣika offers in its six categories a catalogue of everything that exists. We should note that Vaiśeṣika shares this trait with the scholastic developments within Buddhism that we have just discussed. Admittedly, the content of its list is completely different from that of Buddhist lists, but this should not surprise us. The lists of dharmas were, at least in principle, enumerations of important elements mentioned by the Buddha, while the categories of Vaiśeṣika have nothing to do with the Buddha. Then there is the following consideration. In the developments that concern us here, the dharmas were the ultimate constituents of the objects that make up the phenomenal world; according to the Buddhists, these composite objects do not themselves exist. The Vaiśeṣikas, for their part, were in no way inclined to deny the reality of the phenomenal world. Thus, their enumeration of all existent things had to encompass not only ultimate constituents but also composite objects. This consideration helps us to understand the nature of their categories, as well as certain other aspects of the system with which we cannot concern ourselves here. We have just mentioned one common trait between scholastic Buddhism and Vaiśeṣika: both offer an exhaustive enumeration of everything that exists. But they share other traits, too, which leads to the following hypothesis: that Vaiśeṣika was created under the influence of scholastic Buddhism, as a kind of Brahmanical response to it. An in-depth discussion of this hypothesis is not possible here, and it will not be presupposed in what follows. What matters for our purpose is one of the other traits shared by scholastic Buddhism and Vaiśeṣika: both accept that there is a close connection between words and the objects of the phenomenal world. Here, too, it is imperative to take account of the differences that accompany this common trait. For the Buddhists, the objects of the phenomenal world do not exist. We are led to believe in them because of words. For the Vaiśeṣikas, objects of the phenomenal world really do exist, and the presumed parallelism between words and things allows us to draw conclusions about the external world on the basis of language. Let us take another look at the Vaiśeṣika list of categories. It is clear that the first three—substance (dravya), quality (guṇa), and motion (karman)—constitute its core. The designation “object” (artha) is reserved for these three categories,65 with the other categories not counting as “objects.” Their role is clearly subordinate to that of the three main categories. A universal (sāmānya/jāti) groups together a certain number of objects, while particularity (viśeṣa) distinguishes them from one 64 65 See Ui, 1917; Miyamoto, 1996. See WI p. 3 §§ 14 and 15; VS(C) 2.2.24-25. 21 22 another, and inherence (samavāya) is responsible for their interaction. While the secondary role of these categories does not prove that they were added at a later time to an original list containing only substance, quality, and motion, we should not rule out this possibility too quickly. What is important for us is to note the foundational role of the first three categories in the system. These first three categories—substance, quality, and motion—correspond to the three principal types of words in Sanskrit: nouns, adjectives, and verbs. This observation was made in 1918 by the Dutch scholar B. Faddegon, and has been repeated numerous times after him.66 It would be a mistake, however, to conclude that this classification into three categories imposed itself inevitably on Indian thinkers. On the contrary, many did not accept it, including the Buddhists and the Sāṃkhyas. I am inclined instead to believe that the Vaiśeṣikas consciously accepted the classification, as well as the correspondence between these categories and the three types of Sanskrit words. Admittedly, I do not know of any passage in their texts that makes this point explicit; perhaps there is no such passage. But their belief concerning the close connection between words and things is expressed in other ways, to which we shall now turn our attention. Among the three categories of substance, quality, and motion, substance is, in a certain sense, the most important. It is substance that supports the other two, which are in turn largely determined by the substances in which they inhere. In a complete enumeration of all that exists, a specification of substances is therefore essential. This specification occurs in two stages. First, a list of nine substances is presented. Five of these are the elements (bhūta): earth (pṛthivī), water (ap), fire (tejas), wind (vāyu), and ether (ākāśa), a list reflecting then current views on the composition of matter; the other four are time (kāla), space (diś), soul (ātman), and mind (manas). Obviously, Vaiśeṣika by no means claimed that there were only nine, or even five, substantial objects in the world, so a further specification is needed. The number of substantial objects is quite considerable, but Vaiśeṣika texts make no effort to list them all. Why is this? Further, how can one claim to present a complete enumeration of everything that exists without providing a criterion for determining which objects are substances? The answer is simple: in principle, a dictionary of nouns is sufficient to delimit the domain of substances. One must be careful, of course, since some nouns denote qualities, or motions, or something else, but the principle remains sound. Indeed, the existence of certain nouns attests to the presence of substances whose existence is less than evident. The personal pronoun “I,” for instance, indicates the existence of the soul (conceived of as a substance by the Vaiśeṣikas). The fact that this pronoun does not enter into apposition with the word “earth,” etc., proves that the soul is different from the body (which is, in the case of human beings, a form of earth). The substance “time” (kāla), again according to the Padārthadharmasaṅgraha, is the cause of the origin, preservation, and destruction of all produced things, and this because language tells us so. The commentators gloss this in the following manner: we say that a certain object is produced at this or that moment, etc. This same substance “time” is also responsible for our use of words referring to various durations, such as “day,” “month,” “year,” etc. 66 Faddegon, 1918: 107; cf. already Müller, 1852: 10-11, 32. 22 23 One encounters a good number of arguments of this sort, intended to prove the existence of various qualities. Pleasure, for instance, is a quality of the soul, because we say: “I am pleased.” The qualities “distance” (paratva) and “nearness” (aparatva) are responsible for our use of the words “distant” and “near,” respectively. It is not my intention to delve more deeply into the Vaiśeṣika system. But I do wish to observe that it is language that structures reality as it is conceived of by this school. Language does this, first, by means of the semantic relationships between words. Temporarily setting aside the categories of sāmānya, viśeṣa, and samavāya, we arrive at a position according to which everything that exists is either dravya, guṇa, or karman. Substances (dravya) are divided into numerous subcategories; Vaiśeṣika, we have seen, recognizes nine. The first of these subcategories—“earth” (pṛthivī) or “that which is made of earth” (pārthiva)—is in turn divided into innumerable objects, such as trees, pots, etc. Now, trees themselves have sub-species, such as the Śiṃśapā, Palāśa, etc. The result can be schematized as follows: sat _____________|___________________________ | | | dravya guṇa karman ___|___________..... ____|__..... ___|__....... | | | | pārthiva āpya, etc. rūpa, etc. utkṣepaṇa, etc. ____|__________..... | | vṛkṣa ghaṭa, etc. __|_________..... | | śiṃśapā palāśa, etc. This scheme is imposed at least in part by language, with some support from generally accepted notions of the period. The division into three types of existent things corresponds, as we have seen, to the three most important types of words: nouns, adjectives, and verbs. The presence of the five elements among the nine types of substances is due to widely recognized views of the period. The division of earthen things into trees, pots, etc., seems natural, given that Vaiśeṣika maintained that an object can only be composed of one substance at a time, a subject we shall return to in a moment. The scheme is ontological in nature: each of its elements represents a universal that is more encompassing than those represented by the elements below it. All of the elements—i.e., all existent things—are characterized by the universal “existence” (sattā). Other universals characterize different portions of this totality. Thus, the universals sattā, dravyatva, pārthivatva, and vṛkṣatva all inhere in the element “tree.” On the other hand, we cannot say that everything that exists, or is a substance, or is made of earth, is also a tree, a bearer of the universal vṛkṣatva. The existence of all these universals, each connected to sets of objects that either 23 24 include other sets or are included in them, confers on the scheme its ontological dimension. In other words, this structure corresponds to reality and not just to the use of words. Universals by their nature are always connected to sets of objects that either include or are included within other sets. Were this not the case, there would be saṃkara, “mixture,” which authors of the school identify as “destructive of universals” (jātibādhaka).67 This indicates that reality does indeed have the structure just described, namely, that of a genealogical tree without any “contamination,” if I may borrow an expression used in establishing a stemma of manuscripts. This elegant structure brings with it some rather surprising consequences. To take an example from our diagram: trees—all trees—pertain to the element earth and not to water, because water occurs alongside earth and not directly above trees. As a result, Vaiśeṣika found itself obliged to consider trees as composed exclusively of earth. Water, which is undeniably found in trees, is therefore not actually part of a tree; it is connected to the tree without being one of its constituents. Let us also point out that the scheme as a whole could be attractive to thinkers who, accepting the close connection between words and things, relied on language to uncover the structure of reality. We shall see in what follows that the Vaiśeṣikas were not the only ones to follow this approach. 8. Verbal Knowledge Before leaving Vaiśeṣika, let us consider an important question. According to the sketch I have just drawn, Vaiśeṣikas derive their ontology at least in part from language. Language is thus a source of knowledge for them. Do their texts ever address this point directly? The answer, though uncertain, might be yes. Vaiśeṣika texts do speak of “verbal knowledge,” but it is not immediately clear what they mean. Does the expression refer to knowledge derived from language as such, or to knowledge produced by verbal communication? Verbal communication presupposes a speaker, or possibly an author, and can be either true or false depending on the trustworthiness of the speaker. Knowledge derived from language as such—for instance, the knowledge expressed in the Vaiśeṣika system—comes directly from language and not from any speaker. In both cases one might theoretically speak of “verbal knowledge.” Which of the two is intended in early Vaiśeṣika texts? The answer to this question, let me reiterate, is not immediately clear; but I believe there are certain passages that lend credence to the idea of a knowledge derived from language as such. Let us take a look at the most important of these passages. The Vaiśeṣika Sūtra does not have much to say on the subject. One sūtra remarks that inferential knowledge explains verbal knowledge.68 This suggests that verbal knowledge is included in inferential knowledge, but this is not necessarily 67 See WI p. 70 (§ 334); Ki. p. 23 l. 3-4; Shastri, 1976: 323 ff. VS(C) 9.19: etena śābdaṃ vyākhyātam. The other versions offer no variants. The preceding sūtra contains the word laiṅgikam and clearly refers to inferential knowledge. 68 24 25 what the sūtra means.69 As to whether verbal knowledge is derived from language itself or from the utterances of another person, the text remains silent. The next important text of early Vaiśeṣika to have come down to us is the Padārthadharmasaṅgraha of Praśastapāda. This text speaks more clearly of the connection between verbal knowledge and inference, but it does not answer all of our questions. The relevant passage reads: Word, etc., is also included in inference, for it proceeds in the same way. Just as, for someone who knows the convention, there is inference with respect to an object beyond the reach of [one’s] senses on the basis of a definite observation of the sign and memory of the knowledge [of the convention], so, too, [there is inference] on the basis of word, etc. The collection of transmitted texts consisting of Revelation (śruti) and Tradition (smṛti) also depend on the authority of those who uttered them, for [the Sūtra] says: tadvacanād āmnāyaprāmāṇyam (VS(C) 1.1.3); liṅgāc cānityaḥ (VS(C) 70 2.2.37); buddhipūrvā vākyakṛtir vede (VS(C) 6.1.1); buddhipūrvo dadātiḥ (VS(C) 6.1.4). The first part of this passage deals with word (śabda) as a means of knowledge. It says, in essence, that a person who knows the convention (samaya) of words—let us not forget that convention is what allows one to move from word to the comprehension of an object, according to the Vaiśeṣika Sūtra71—can infer from them the objects they designate. The text is less clear as to whether this knowledge comes directly from language, or rather from the utterance of a trustworthy person. It is perhaps significant that this part of the passage makes no mention of trustworthy persons. They are mentioned in the second part, in connection with the sacred texts, śruti and smṛti. These texts owe their authority to the authority of those who uttered them. Trustworthy speakers also feature in the discussion of analogy (upamāna), which likewise depends on the utterance of a trustworthy person.72 The absence of any mention of trustworthy persons in the lines claiming that verbal knowledge is a kind of inference might perhaps support the conclusion that verbal knowledge does not necessarily originate from the utterances of such persons. Note, however, that the word “also” in the line: “The collection of transmitted texts consisting of Revelation and Tradition also depend on the authority of those who uttered them,” goes against this interpretation. The fact that the three classical commentators on the text seem to understand verbal knowledge solely in connection with trustworthy persons is another point against the theory that verbal knowledge comes directly from language. But there remains a more definitive argument in favor of interpreting verbal knowledge as derived directly from language rather than from the utterances of trustworthy persons. The Nyāya Sūtra and its Bhāṣya are aware of this position, which they place in the mouth of an opponent who, moreover, presents verbal 69 According to Madeleine Biardeau (1964: 210 n.), “[ce sūtra] n’identifie pas forcément śabda et anumāna.” 70 WI p. 48 § 256-57: śabdādīnām apy anumāne ‘ntarbhāvaḥ samānavidhitvāt / yathā prasiddhasamayasyāsandigdha- (v.l. om. asandigdha-) liṅgadarśanaprasiddhyanusmaraṇābhyām atīndriye ‘rthe bhavaty anumānam evaṃ śabdādibhyo ‘pīti / śrutismṛtilakṣaṇo ‘py āmnāyo vaktṛprāmāṇyāpekṣaḥ (v.l. °pekṣo ‘numānam eva) tadvacanād āmnāyaprāmāṇyaṃ liṅgāc cānityo buddhipūrvā vākyakṛtir vede buddhipūrvo dadātir ity uktatvāt / 71 VS(C) 7.2.24: sāmayikaḥ śabdād arthapratyayaḥ. Cf. Houben, 1992. 72 WI p. 48 § 259: āptenāprasiddhagavayasya gavā gavayapratipādanād upamānam āptavacanam eva. 25 26 knowledge as identical to inference, as in the Padārthadharmasaṅgraha. The opponent’s position is then rejected by the Nyāya Sūtra and Bhāṣya, specifically on the grounds that words are capable of producing valid knowledge only when they come from a trustworthy person. In other words, language on its own is not a means of knowledge, contrary to what the opponent maintains. Here are the key portions of the opponent’s position: Word is an inference; it is not a distinct means of knowledge. –Why? Because the object of a word is inferred. –How so? Because it is not apprehended by means of perception. A sign-possessing object (liṅgin), though unapprehended, comes to be known through a knowledge of the sign (liṅga); this is inference. Likewise, an object (artha), though unapprehended, comes to be known through a knowledge of the word [by which it is designated]. Therefore, word is an inference. … There is also the following reason: when a means of knowledge is distinct, apprehension involves two distinct processes. For apprehension takes place one way in the case of inference and another way in the case of analogy. … But in the case of word and inference, apprehension does not involve two processes. The process involved in inference is the same as the process involved in word. Since there is no difference, word is an inference. … One grasps an object through verbal knowledge when the relation between a word and the object connected to it is well known, just as one grasps an object through inferential knowledge when the relation between an 73 inferential sign and the object possessing the sign is well known. The response of the authors of the Sūtra and Bhāṣya shows that this verbal knowledge is not the same as verbal communication. According to sūtra 2.1.52: It is in virtue of the statement of a trustworthy person (āpta) that belief in an object 74 arises from word. The Bhāṣya explains: It is not because of words on their own that one believes [in the existence of] imperceptible objects such as “heaven,” “the Apsaras,” “the Northern Kurus,” “the seven continents,” “the ocean,” or “the shape of the world”; rather, one believes [in their existence] because they have been spoken of by trustworthy people. 75 Otherwise one would not believe in them. Inference, however, is not like this. These passages show that the authors of these sūtras and their commentary were aware of the idea of verbal knowledge as something distinct from verbal communication and identical to, or at least included within, inference as a means of knowledge. But where did the idea come from? Who is the opponent whose views 73 NBh 2.1.49-51: śabdo ‘numānaṃ na pramāṇāntaram / kasmāt? śabdārthasyānumeyatvāt / katham anumeyatvam? pratyakṣato ‘nupalabdheḥ / yathānupalabhyamāno liṅgī mitena liṅgena paścān mīyata ity anumānam / evaṃ mitena śabdena paścān mīyate ‘rtho ‘nupalabhyamāna ity anumānaṃ śabdaḥ // itaś cānumānaṃ śabdaḥ / … / pramāṇāntarabhāve dvipravṛttir upalabdhiḥ / anyathā hy upalabdhir anumāne ‘nyathopamāne … / śabdānumānayos tūpalabdhir advipravṛttir yathānumāne pravartate tathā śabde ‘pi / viśeṣābhāvād anumānaṃ śabda iti // … / saṃbaddhayoś ca śabdārthayoḥ saṃbandhaprasiddhau śabdopalabdher arthagrahaṇaṃ yathā saṃbaddhayor liṅgaliṅginoḥ saṃbandhapratītau liṅgopalabdhau liṅgigrahaṇam iti // 74 NS 2.1.52: āptopadeśasāmarthyāc chabdād arthasaṃpratyayaḥ. 75 NBh 2.1.52: svargaḥ apsarasaḥ uttarāḥ kuravaḥ sapta dvīpāḥ samudro lokasaṃniveśa ity evamāder apratyakṣasyārthasya na śabdamātrāt pratyayaḥ, kiṃ tarhi? āptair ayam uktaḥ śabda ity ataḥ saṃpratyayaḥ, viparyayeṇa saṃpratyayābhāvāt / na tv evam anumānam iti / 26 27 are set forth? Madeleine Biardeau comments as follows: “It is impossible to identify the opponent here, who might well have been invented, but for the Naiyāyika he seems to represent those who wished to see a ‘natural’ connection … between the word and its object.”76 The fact that verbal knowledge is here presented as distinct from verbal communication, and thus as independent of a speaker, initially calls to mind the position of Mīmāṃsā, which maintains that the Veda has no author. But the early texts of this school, as far as I know, do not identify verbal knowledge with inference. On the contrary, “in the case of Vedic speech, one’s knowledge is direct (pratyakṣa),” as Śabara writes in his Mīmāṃsā Bhāṣya.77 If, therefore, the Buddhist thinker Bhavya, in his Madhyamakahṛdayakārikā 9.5, attributes to a Mīmāṃsaka the view that the existence of objects such as heaven is known on the basis of the Veda,78 one has to conclude that these objects are known through direct knowledge rather than through inference. The identification of verbal knowledge with inference does occur, however, in the Padārthadharmasaṅgraha of Praśastapāda, as we have just seen. If the opponent in these Naiyāyika texts is indeed Praśastapāda, or rather one of his predecessors, it seems likely that some Vaiśeṣikas must indeed have explicitly held that language allows one to gain knowledge of the world. Let us note further that the grammarian Bhartṛhari, in his Mahābhāṣyadīpikā, explicitly maintains that the words “heaven,” “apūrva,” and “deity,” once apprehended, allow one to infer the existence of radically imperceptible objects.79 Did Bhartṛhari borrow this idea from the Vaiśeṣikas? The Mahābhāṣya of Patañjali, an absolutely authoritative text for Bhartṛhari, describes grammarians as śabdapramāṇaka, or “accepting the word as authority”; it glosses the expression by saying: “What the word says is authoritative for us.”80 Is it possible that Bhartṛhari interpreted this line as evidence of a verbal knowledge independent of trustworthy speakers? We shall return to the question of a verbal knowledge derived directly from language when we come to the Buddhist thinker Dignāga, near the end of these lectures. It is worth mentioning in this context that Yoga Sūtra 1.9 defines the mental activity (cittavṛtti) known as vikalpa—which is distinct from correct knowledge (pramāṇa) and error (viparyaya)—as “resultant from verbal knowledge [but] void of substance.”81 The Yoga Bhāṣya gives several examples of this kind of “verbal knowledge” (not to be confused with verbal communication, or āgama, mentioned in Yoga Sūtra 1.7 and defined in the Bhāṣya). One example is the reification of negations: to say, for instance, that the soul (puruṣa) is characterized by the absence 76 Biardeau, 1964: 204: “Il est impossible ici d’identifier cet adversaire qui pourrait être fictif, mais il semble qu’il ait personnifié pour le Naiyāyika ceux qui voulaient reconnaître une relation ‘naturelle’ … entre le mot et son objet.” 77 MīBh 1.1.2 (Frauwallner, 1968: 20): pratyakṣas tu vedavacane pratyayaḥ. 78 Kawasaki, 1977: 6-7. 79 Mahābhāṣyadīpikā, Manuscript p. 11a l. 11; “Critical edition” Āhnika I p. 28 l. 8-9; ed. AbhyankarLimaye p. 33 l. 24 – p. 34 l. 1; ed. Swaminathan p. 40 l. 11: tatra yathaiva svargāpūrvadevatāśabdā upalabhyamānā atyantāparidṛṣṭānām arthānām astitvānumānam. Vkp 2.119 seems to say the same thing, but less clearly. 80 Mbh I p. 11 l. 1-2; also p. 366 l. 12-13 (on P. 2.1.1 vt. 5): śabdapramāṇakā vayaṃ, yac chabda āha tad asmākaṃ pramāṇam. Bhartṛhari’s Mahābhāṣyadīpikā does not comment on these words (“Critical edition” Āhnika I pp. 30-31). 81 YS 1.9: śabdajñānānupātī vastuśūnyo vikalpaḥ. 27 28 of arising does not in fact signify a positive characteristic.82 It is clear that the Yoga Sūtra and Bhāṣya warn against the danger of using language to draw conclusions that do not correspond to reality. Let us conclude this excursus with a word on the chronology of Vaiśeṣika. The Padārthadharmasaṅgraha is a relatively late text, later in any case than the Nyāya Sūtra, and probably later than the Nyāya Bhāṣya. It is therefore unlikely that the opponent in the passage cited from the two Nyāya texts was Praśastapāda, the author of the Padārthadharmasaṅgraha. Fortunately, there is no need to identify the opponent of the Naiyāyikas with the latter. We can be quite certain that the Padārthadharmasaṅgraha follows a long trajectory of development within the Vaiśeṣika school, and that Praśastapāda owes much to his predecessors. It is therefore possible, even probable, that his view of speech as a means of knowledge is an old doctrine of the school, expressed in other, earlier works. One cannot rule out the possibility that the doctrine might even be as old as the sūtra we cited, which is unfortunately too concise to be sufficiently clear. A final remark is called for before we leave—at least for the time being—the Brahmanical school of Vaiśeṣika. We mentioned the hypothesis that Vaiśeṣika was created under the influence of scholastic Buddhism, as a sort of Brahmanical response to the latter. But the influence does not seem to have gone only one way. One should not rule out the possibility that Vaiśeṣika in turn had an influence on certain texts of Buddhist scholasticism. We cannot explore this subject at length in these lectures; here I will limit myself to a few examples. Consider the Abhidharmasamuccaya attributed to Asaṅga, a scholastic text from the school known as Yogācāra. This text, likely dating to the fourth century of the common era, discusses many elements that seem to bear an undeniable resemblance to certain elements from Vaiśeṣika. In dealing with “formations not associated with mind” (cittaviprayukta saṃskāra) the text lists, among other things, union (yoga), speed (java), time (kāla), orientation (deśa), and number (saṃkhyā).83 These elements do not figure in the texts of Sarvāstivāda, at least not in this way, but they do have parallels in Vaiśeṣika. The union (yoga) of Asaṅga might correspond to the Vaiśeṣika category of inherence (samavāya); speed (java) to the quality of vega (a form of saṃskāra); time (kāla) to the substance of the same name; orientation (deśa) to the substance diś; and number (saṃkhyā) to the quality of the same name. It is true that for the author of the Abhidharmasamuccaya these elements are only designations (prajñapti), but this applies to all of the formations not associated with mind. Another section of the Abhidharmasamuccaya lists six kinds of conjunction (saṃprayoga), one of which (avinirbhāga-saṃprayoga) seems to correspond to the Vaiśeṣika notion of inherence, and another (or even two: miśrībhāva-saṃprayoga and samavadhāna-saṃprayoga) of which seems to correspond to its quality of saṃyoga.84 82 YBh 1.9: sa na pramāṇopārohī, na viparyayopārohī ca / vastuśūnyatve ‘pi śabdajñānamāhātmyanibandhano vyavahāro dṛśyate / tad yathā: caitanyaṃ puruṣasya svarūpam iti / yadā citir eva puruṣas tadā kim atra kena vyapadiśyate? bhavati ca vyapadeśe vṛttiḥ / yathā caitrasya gaur iti tathā pratiṣiddhavastudharmo niṣkriyaḥ puruṣaḥ / tiṣṭhati bāṇaḥ, sthāsyati, sthita iti gatinivṛttau dhātvarthamātraṃ gamyate / tathā anutpattidharmā puruṣa iti utpattidharmasyābhāvamātram avagamyate na puruṣānvayī dharmaḥ / tasmād vikalpitaḥ sa dharmaḥ, tena cāsti vyavahāra iti // 83 Abhidh-sam p. 10 l. 15 ff.; Abhidh-sam(R) p. 15 ff.; Abhidh-sam-bh p. 10; Gokhale, 1947: p. 18 l. 20 ff.; Kritzer, 1999: 248 ff.; Jaini, 1959: 537 ff. 84 Abhidh-sam p. 33 l. 19 ff.; Abhidh-sam(R) p. 55 ff.; Abhidh-sam-bh p. 47. 28 29 Later we meet with an element corresponding to the Vaiśeṣika notion of universals: the dharma called sabhāgatā, or “homogeneous character,” accepted by some Sarvāstivādins, and also found in the Abhidharmasamuccaya under the names sabhāga and tatsabhāga.85 An in-depth study of these correspondences has yet to be written, however, so one should be careful not to draw conclusions too hastily. 85 Abhidh-sam p. 29 l. 19 ff.; Abhidh-sam(R) p. 47 ff.; Abhidh-sam-bh p. 43; Gokhale, 1947: p. 27 l. 32. 29 30 II. The Correspondence Principle 1. The Contradictions of Nāgārjuna Let us resume our discussion of Buddhism where we left off. According to the Buddhist scholastics, objects do not exist in reality, but owe their existence to language, to the words by which they are designated. On this view, the phenomenal world is determined, even created, by words. Note that this position postulates a connection between the objects of the phenomenal world and individual words. But words are not normally used in isolation; they appear in sentences. What, then, is the connection between the phenomenal world and sentences? This question arises almost inevitably in the intellectual context of scholastic Buddhism, and the response seems equally inevitable. A sentence describes a situation; and the words of the sentence—we shall limit ourselves for the moment to nouns and verbs—correspond to the elements that together constitute this situation. A simple sentence of the sort “Devadatta reads a book” describes a situation in which there are three elements: the person named “Devadatta,” a book, and the act of reading. Given what we know about Buddhists of this period, it goes without saying that none of these three elements really exists. They exist as designations, without absolute reality. Now let us examine the sentence “Devadatta writes a book.” Following the same logic, the three principal words of the sentence should correspond to three elements, which together constitute the situation described by the sentence. What are these three elements? Devadatta, a book, and the act of writing. The situation is thus completely parallel to what was denoted by the sentence “Devadatta reads a book.” But there is a crucial difference. At the moment when Devadatta is writing a book, the book is not yet there. If the book were already there, it would not need to be written. In other words, the three elements do not coexist and therefore do not together constitute the situation described by the sentence. Confronted with such a contradiction, we might be inclined to reject the very notion of a one-to-one correspondence between the words of a sentence and the elements that together constitute the situation described by the sentence. At the very least, we might forgo the requirement that the objects and actions referred to must be present in the situation described by the sentence. To Buddhists of the period, however, this type of contradiction appeared in an altogether different light.86 Let us not forget that for them Devadatta, the book, and the act of reading do not exist. They are but entities in a phenomenal world that has no reality apart from words. A contradiction concerning non-existent things could not trouble them. Or rather, contradictions of this sort would only serve to prove that the entities of the phenomenal world cannot exist. It was the Buddhist thinker Nāgārjuna who saw that this type of argument could be used to attack the views of those who affirm the reality of the objects of 86 Not to all Buddhists. Édith Nolot brought to my attention a difficult to date passage from the Milindapañha, where the expressions “I churn buttermilk” (takkaṃ manthemi)—what one actually churns is curdled milk (dadhi)—and “I am making a non-existent thing” (asantaṃ sādhemi) are explained as being popular expressions (lokasamaññā); cf. Mil pp. 173-74, Nolot, 1995: 151-52. 30 31 the phenomenal world. These thinkers included, on the one hand, the nonBuddhists, but also those Buddhists who wished to maintain the reality of a world hidden behind the phenomenal world. The challenge presented by Nāgārjuna was thus in keeping with the opposition that emerged during this period between those Buddhists—most often adherents of the Mahāyāna—who rejected even the existence of the dharmas, and those who were not so radical. Against whom were Nāgārjuna’s arguments directed? The answer matters little; as the rest of these lectures will show, in his day practically all thinkers in India felt threatened by them. For the moment, what matters is that a good number of Nāgārjuna’s arguments cannot be understood without taking into account the foregoing reflections on the connection between language and the phenomenal world. To be more precise, these arguments are based on the presupposition, now familiar to us, that I call the correspondence principle. According to this principle, the situation described by a sentence is constituted of elements that correspond one-for-one to the words of the sentence. This formulation no doubt leaves much to be desired, but it allows us, even as it stands, to understand many of Nāgārjuna’s arguments. Let us note from the start, however, that the correspondence principle is not primarily a logical position, but rather an intuition shared by thinkers of the period. It is clear that some of the logical consequences of this principle are quite dubious; the very existence of a “situation constituted of elements corresponding to the words of a sentence” implies the simultaneous coexistence of these elements. Yet such coexistence is not always apparent; far from it, as we shall see in what follows. Primarily representing an intuition, the correspondence principle was not immediately subjected to logical analysis. On the contrary, only rarely do we find a more or less explicit formulation of it in the literature of the period. Even so, this does not warrant our rejecting it as something unexpressed and logically implausible. Philosophical thought does not proceed in an exclusively logical manner, and quite often the problems it seeks to solve have nothing to do with logic. The rest of these lectures will indeed show how a more or less vague intuition was able to influence, and to some degree even determine, the development of Indian thought. Before offering a few concrete examples from Nāgārjuna’s works, we should recall that a considerable number of texts attributed to this author have been preserved, either in their original Sanskrit or in Tibetan and even Chinese translations. The question of their authenticity is far from resolved, and the only sure way to avoid dispute is to limit ourselves to the Fundamental Verses of the Middle Way (Mūlamadhyamakakārikā), a work whose author is more or less by definition Nāgārjuna. The examples to follow are therefore all from this text. Having already devoted an article to the correspondence principle in the thought of Nāgārjuna,87 here I will limit myself to a few striking examples. Consider first the following verse: If there existed anywhere something unarisen, it could arise. Since no such thing 88 exists, what is it that arises? 87 Bronkhorst, 1997. MadhK(deJ) 7.17: yadi kaścid anutpanno bhāvaḥ saṃvidyate kvacit / utpadyeta sa kiṃ tasmin bhāve utpadyate ‘sati // 88 31 32 This is the problem we referred to not long ago. The situation in which something arises, a pot for instance, is expressed in the sentence: “The pot arises.” This situation must be constituted of elements corresponding to the words of the sentence. In other words, there must be a pot in order for it to arise. But since the pot is obviously not yet there at the moment of its arising—the pot and its arising do not exist together—what is it that arises? This example shows that the arising of things presents a problem for those who accept the correspondence principle. I repeat that this problem was not particularly serious for Nāgārjuna, who, as a Buddhist, did not consider the phenomenal world—the world of things that arise, etc.—to exist in reality. Indeed, the verse is meant as a challenge to those who attribute reality to the phenomenal world, a challenge taken quite seriously by other thinkers, as we shall see in what follows. First, though, let us take a look at a few other verses of Nāgārjuna’s, which confirm that the arising of things is not possible. One of them argues that the cause of a thing cannot exist: Neither for a non-existent object nor for an existent object is a cause possible. How can a non-existent object have a cause? And why would an existent object need 89 one? This verse has to do with the situation described in sentences of the type “A is the cause of B,” or more concretely, “the seed is the cause of the sprout.” According to the correspondence principle, the situation described here is constituted of the sprout and the seed that causes it. If the sprout does not exist in it, the seed is not the cause of anything, and thus is not a cause at all. If, on the other hand, the sprout exists in the situation that also contains the seed, what purpose does the seed serve? It is not difficult to see that if a situation must be constituted of elements corresponding to the words of the sentence describing it, every statement having to do with the past, the future, or the transition between these temporal domains is potentially contradictory. The simple observation that the future depends on the present, or on the past, presupposes a situation in which future, present, and past coexist, which runs counter to our experience. This is the situation Nāgārjuna evokes in another verse from his Mūlamadhyamakakārikā: If the present and the future depend on the past, the present and the future will be in the past. If, again, the present and the future are not there in the past, how can 90 they depend on it? To conclude our discussion of Nāgārjuna’s Mūlamadhyamakakārikā, let us look at one more verse: 89 MadhK(deJ) 1.6: naivāsato naiva sataḥ pratyayo ‘rthasya yujyate / asataḥ pratyayaḥ kasya sataś ca pratyayena kim // 90 MadhK(deJ) 19.1-2: pratyutpanno ‘nāgataś ca yady atītam apekṣya hi / pratyutpanno ‘nāgataś ca kāle ‘tīte bhaviṣyataḥ // pratyutpanno ‘nāgataś ca na stas tatra punar yadi / pratyutpanno ‘nāgataś ca syātāṃ katham apekṣya tam // 32 33 Neither the road already traveled (gata-) nor the road yet to be traveled (agata-) is presently being traveled. But the road being traveled is not traveled apart from 91 these two. The problem raised in this verse is addressed in a different form a few lines down in the same text. If one could travel the road that is being traveled, there would be two acts of traveling: the one by which the road is being traveled, and again the traveling on it. Were there two acts of traveling, there would have to be two travelers; for without a 92 traveler, the act of traveling is not possible. These verses revolve around the phrase “[the road] being traveled is traveled” (gamyamānaṃ gamyate). As this sentence—which Nāgārjuna seems to consider wellformed—contains two verbal forms, the situation it describes must, according to the correspondence principle, contain two actions, though this is not intended by the sentence: the two verbal forms obviously refer to a single action. Nāgārjuna goes further still: two actions require two agents; but our sentence obviously concerns just one agent. The Parisian scholar Kamaleswar Bhattacharya has on several occasions drawn attention to the fact that this mode of argument is closely linked to the Sanskrit grammatical tradition.93 He invokes the commentator Candrakīrti, who remarks on verse 2.6 saying: “An action (kriyā) necessarily requires a means of realization (sādhana = kāraka): an object (karman) or an agent (kartṛ). Now the act of moving resides in an agent. Therefore, it requires an agent of motion.”94 Bhattacharya comments: “Candrakīrti refers here to the grammatical theory according to which the action denoted by a verbal root resides either in the agent (kartṛ) or in the object (karman), and the root gam (‘to go, move, travel’) is one of those that denote actions residing in the agent (kartṛsthakriya). It is therefore indispensable that there be two agents of motion if there are two motions. But in the case under consideration here, there is only one agent.”95 No doubt Candrakīrti employs grammatical terms in this passage. But that does not necessarily mean that this mode of argument is distinctly grammatical. It is true that some grammatical texts speak of actions residing in the agent or in the 91 MadhK(deJ) 2.1: gataṃ na gamyate tāvad agataṃ naiva gamyate / gatāgatavinirmuktaṃ gamyamānaṃ na gamyate //. Bhattacharya (1986a: 297) notes that according to commentators, the root gam-, as used at the end of the verse, has the sense of “to know, to discern”: “À la fin du verset, la racine gam-, dans gamyate, a, disent les commentateurs, le sens de ‘connâitre, distinguer’: gamyate = prajñāyate, Candrakīrti.” 92 MadhK(deJ) 2.5-6: gamyamānasya gamane prasaktaṃ gamanadvayam / yena tad gamyamānaṃ ca yac cātra gamanaṃ punaḥ // dvau gantārau prasajyete prasakte gamanadvaye / gantāram hi tiraskṛtya gamanaṃ nopapadyate // 93 See Bhattacharya, 1980; 1981; 1984; 1985; 1986a; 1995; 1996. The 1995 piece responds to an article by George Cardona (1991). 94 MadhK(V) p. 39 l. 14-15: … avaśyaṃ kriyā svasādhanam apekṣate karma kartāraṃ vā / gamikriyā caivaṃ kartary avasthitā, ato gantāram apekṣate /. Translation based on Bhattacharya, 1984: 192. 95 Bhattacharya, 1984: 192: “Candrakīrti se réfère ici à la théorie grammaticale selon laquelle l’action que dénote une racine verbale réside soit dans l’agent (kartṛ), soit dans l’objet (karman), et la racine gam ‘aller, se mouvoir, parcourir’ est l’une de celles qui dénotent des actions qui résident dans l’agent (kartṛsthakriya). Il est donc indispensable qu’il y ait deux agents de mouvement s’il y a deux mouvements. Mais, dans le cas que l’on envisage ici, il n’y a qu’un agent.” 33 34 object; Bhattacharya mentions a vārttika of Kātyāyana (no. 3) and a verse from the Kāśikāvṛtti, both on Pāṇini 3.1.87.96 The Kāśikā even adds that the action expressed by the root gam resides in the agent. But neither of these two texts specifies that two motions necessarily require two agents. This conclusion does not follow from grammar, but from the logic of Nāgārjuna. We shall see in what follows that Vasubandhu, a later Buddhist thinker and the author of the Abhidharmakośa Bhāṣya, seems indeed to attribute argumentation based on the correspondence principle to a grammarian—more precisely, to a śābdika, “one who deals with / relies on words”; the commentator Yaśomitra, followed by Louis de la Vallée Poussin in his French translation, glosses this term as vaiyākaraṇa, or “grammarian.” We cannot be certain that Yaśomitra correctly interprets Vasubandhu on this point; or rather, we cannot at all be certain that Vasubandhu’s critique is aimed only at grammarians. After all, Vasubandhu’s opponent—like Nāgārjuna and many other thinkers, whom we shall come to— whether grammarian or not, relies on words; that is to say, he relies on the correspondence principle, whereas Vasubandhu accepts only a modified version of the principle. We shall see that other thinkers, too, notably the Vedāntin Śaṅkara, employ certain grammatical expressions (chiefly kartṛ, “agent,” and kāraka, “actor [in the event described by the sentence], actant”) in the course of defending positions analogous to those of Nāgārjuna. But whatever Vasubandhu may have thought, there is no specific connection between these thinkers and Indian grammar, whether Pāṇinian or otherwise. Grammar may have supplied the terminology here, a very useful terminology at that, but Nāgārjuna’s ideas hardly owe their existence to it. Even without grammar we can conceive that an action requires an actor, especially if the sentences mentioning the actions mention their actors as well. We say “the pot arises”: do we really need grammar to tell us that the act of arising, expressed in this sentence, requires the subject mentioned in the very same sentence (the pot), just as the act of going, in the sentence “he goes,” requires a goer? That being said, I do not by any means wish to deny the importance or omnipresence of grammatical terminology in the “śāstric” literature of classical India. Moreover, it is not just a question of terminology. Grammar provided the tools of analysis for the entire classical tradition of Indian thought. In this broad sense, every Indian thinker, including Nāgārjuna, was influenced by grammar. But to repeat, I do not believe that this connection was particularly decisive in this specific case. The above examples are not enough to do justice to the challenge Nāgārjuna presented to classical Indian thought, but they should suffice to illustrate the kind of threat his work posed. As the purpose of these lectures is not to provide an indepth study of Nāgārjuna’s arguments, but rather to examine the reactions this kind of threat provoked from thinkers in the centuries following him, I will say no more on the subject of this important thinker. 2. The Reactions of Other Thinkers 96 Bhattacharya, 1980: 93 n. 25. 34 35 How did Indian thinkers react to the challenge posed by Nāgārjuna? We should note at this point that Nāgārjuna was not an isolated thinker. On the contrary, he was the founder of a school, known by the name of Madhyamaka or Mādhyamika, that survives even to the present day, mainly in Tibetan Buddhism. Nāgārjuna thus had many disciples and successors, who carried on and developed his thought. Several currents are discernible within the Madhyamaka school, which shows that it gave rise to original and innovative ideas.97 But there were other thinkers in India who did not necessarily agree with Nāgārjuna. How did they react? Buddhist or not, their view of the world was equally threatened by his deconstructive arguments. We know that there was considerable interaction among intellectuals in the centuries following Nāgārjuna, so we would expect to find, by way of reaction, thinkers defending their positions against those of Nāgārjuna. In reality, such defenses appear to be missing from the philosophical literature of the first millennium. At first sight, this absence is astonishing. It has led the American scholar Richard P. Hayes to conclude that the works of Nāgārjuna had relatively little effect on the subsequent development of Buddhist philosophy.98 According to Hayes, the explanation for this strange situation lies in the logical inadequacy of Nāgārjuna’s arguments. Under close examination, the arguments prove fallacious. Later thinkers, skilled as they were at logic, would have realized this and would not have taken the trouble to respond. Some of the arguments set forth in the Mūlamadhyamakakārikā are indeed probably not very convincing at the level of logic. Others, however, are free of logical flaws. These arguments constitute an important part of Nāgārjuna’s text, perhaps even the most important part. I am referring to the arguments we have studied, which are based on the correspondence principle. Did Indian philosophers coming after Nāgārjuna really ignore them? I believe that they did not. Or rather, the consequences of the correspondence principle were well known to them; philosophers were preoccupied with them perhaps even before the time of Nāgārjuna, and most certainly afterward. Among the problems they sought to resolve, the purported impossibility of the arising of things received special attention. To recall the problem in brief: an object, say a pot, cannot arise, for there must be a pot in order for it arise. But if the pot already exists, it no longer needs to arise. Focusing on this particular problem, we see that there are two possible solutions. First, one could maintain that the pot existed even before it arose. Strange as it might seem, this position was defended by several currents of Indian thought, to which we shall soon turn our attention. A second, quite different solution would be to reject, or at least to adapt, the correspondence principle. Several examples of this solution will also be examined. It is worth noting in advance that Indian thinkers did not sacrifice the principle without hesitation, and without trying to preserve as much of it as possible. Before we turn to these solutions, a preliminary remark is called for. As we have seen, the correspondence principle concerns words in a sentence. It does not refer exclusively to the correspondence between words in general and things in 97 98 See, e.g., Ruegg, 1981; Santina, 1986; Robinson, 1967; Ming-Wood, 1994 (developments in China). Hayes, 1994: 299. 35 36 general, as was the case with Vaiśeṣika and the currents of Buddhism we considered.99 This general correspondence is not rejected, but here the accent is placed on the correspondence between the words of a sentence and the things constituting the situation described by the sentence. The importance of sentences thus lies at the heart of the correspondence principle and the problems it raises. In a way, the importance of sentences in verbal communication is obvious. It was even more so for thinkers who considered the Veda to be the primordial expression of speech. Mīmāṃsā in particular was not as interested in the individual words of the Veda as in its sentences, primarily its injunctions. A Vedic injunction is never expressed with just a single word. As an opponent in the Mīmāṃsā Bhāṣya says: Even if words and their relations to objects are innate and permanent, nevertheless, dharma cannot be what is indicated by Vedic injunctions. For an injunction is a sentence. In the sentence “He who desires heaven should perform the Agnihotra sacrifice,” no single word (on its own) means that one attains heaven through the Agnihotra. One understands this only when the three words (agnihotraṃ juhuyāt svargakāmaḥ) have been spoken, and there does not exist a fourth word [denoting 100 “heaven-through-Agnihotra”] distinct from this group of three words. Śabara’s exact response matters little. The idea is clear: Vedic hermeneutics must take account of the sentence as a linguistic unit. Bhartṛhari accepted this idea, and drew from it some astounding conclusions. The world having been created or organized, according to him, by the Veda, it corresponds not just to the words of the Veda but also to its prescriptions. Thus, a child knows his duties even without being taught, and birds know how to make their nests. I have dealt with this subject elsewhere,101 so I will not go into further details here. What I wish to point out is simply that arguments based on the notion of the sentence as a linguistic unit had to be taken seriously by thinkers concerned with Vedic hermeneutics. 3. Sarvāstivāda Let us now take a closer look at particular currents of thought and the ways in which they tried to resolve the problems entailed by the correspondence principle. Nāgārjuna’s arguments had been directed primarily at Buddhist Abhidharma, and more specifically, it would seem, at the school known as Sarvāstivāda. Did this school respond to his attacks, and if so how? The answer is quite surprising: not 99 Wada (1990: 21) observes that “[the Vaiśeṣika view of the world] is closely connected with the view that each and every meaningful unit or morpheme in a true statement refers without fail to an entity existing in the world.” This formulation is similar to our correspondence principle. We shall see in what follows that Vaiśeṣika did not remain indifferent to this principle; as for its earliest period, however, it seems that what was accepted was a correspondence between words and things in general rather than a correspondence between the words of a sentence and the situation described by the sentence. 100 MīBh 1.1.24: yady apy autpattiko nityaḥ śabdaḥ saṃbandhaś ca tathāpi na codanālakṣaṇo dharmaḥ / codanā hi vākyam / na hy agnihotraṃ juhuyāt svargakāma ity ato vākyād anyatamāt padād agnihotrāt svargo bhavatīti gamyate / gamyate tu padatraya uccarite / na cātra caturthaḥ śabdo ‘sty anyad ataḥ padatrayasamudāyāt /. Translation based on Biardeau, 1964: 192. 101 Bronkhorst, 1996: 128 ff. On the linguistic unit in Bhartṛhari, see Houben, 1995a. 36 37 only did Sarvāstivāda prove impervious to his attacks, it appears to have been aware of the problems even before Nāgārjuna raised them. Recall that the very name of the school Sarvāstivāda expresses its most characteristic doctrine: all things—meaning the dharmas, of course, which according to the ontology of this school constitute everything there is—exist, be they past, present, or future. Here we have a ready response to the problem of the arising of the pot: the pot can arise because it already exists, namely, in its future state. (Strictly speaking, it is obviously not the pot that exists in its future state, because the pot never exists for the Sarvāstivādins, not even in its present state; only the dharmas exist.) A passage from the Mahāvibhāṣā addresses precisely the question we are interested in: When future conditioned dharmas arise, do they arise being already arisen, or do they arise being not yet arisen? There are difficulties with both views. How so? On the first view, why do they not continue to arise? On the second, how do you deny that the saṃskāras exist (or begin to exist) after not existing (abhūtvā bhāvaḥ, abhūtvā bhavanam)? Here is the reply. –One can say (asti paryāyaḥ): They arise being already arisen, because there are causes and conditions. That is to say: all the dharmas possess their own-nature already, for each future dharma resides in its essential character (svabhāvalakṣaṇa). Possessing an own-nature already, they are said to be already arisen: it is not the case that their own-nature arises from causes and conditions. Being [subsequently] produced through the coming together of causes and conditions, they are said to arise. They arise being not yet arisen, because there are causes and conditions. That is to say: future dharmas are said to be unarisen, for it is in virtue of causes and 102 conditions that they actually attain to being arisen. This passage occurs in all three versions of the Vibhāṣā / Mahāvibhāṣā,103 so it must have belonged to the original version of the text, or at least to a very early version. The Chinese translation of Xuanzang—the latest of the three versions—includes an additional passage, immediately after the passage just quoted, that addresses the problem of the arising of things even more directly: When future conditioned dharmas arise, do they arise already existing, or not existing? There are difficulties with both views. On the first, what is the point of their arising, since their being (svarūpa) already exists? On the second, one will have to say that the dharmas, after not existing, exist: the doctrine of Sarvāstivāda is overthrown. Here is the reply. –One can say that the dharmas arise already existing. You avoid the second difficulty, but how is the first to be resolved? As follows. –The own-nature (svabhāva) of the dharma exists, but not its activity. 104 Encountering causes and conditions, the dharma gives rise to activity. For comparison, let me again quote verse 7.17 from the Mūlamadhyamakakārikā of Nāgārjuna: 102 Translation based on La Vallée Poussin, 1937: 15-16. TI 1545 ch. 76, vol. 27, p. 394b l. 19-27 (the passage translated by La Vallée Poussin); TI 1546 ch. 40, vol. 28, p. 295a l. 6-11; TI 1547 ch. 7, vol. 28, p. 465c. l. 11-17. Enomoto (1996) discusses the different recensions of the Vibhāṣā. 104 TI 1545 p. 394b l. 27 – p. 394c l. 5; translation based on La Vallée Poussin, 1937: 16. 103 37 38 If there existed anywhere something unarisen, it could arise. Since no such thing 105 exists, what is it that arises? The parallelism between Nāgārjuna’s verse and the passage from the Mahāvibhāṣā is undeniably striking, and one is tempted to regard the latter as a response to the problem formulated by Nāgārjuna. A pot (or rather a future conditioned dharma) can arise because it already exists. The question of knowing what the difference is between the arisen pot and the pot that existed before its arising leads to an interesting discussion, one that need not concern us here. It involves the elaboration of a theory capable of supporting the basic position that things—again, the Sarvāstivādins are of course speaking of dharmas—exist prior to their arising. The fact that this last passage from the Mahāvibhāṣā is found only in the version translated by Xuanzang confirms that it was a later addition, perhaps owing to the influence of Nāgārjuna. One should note, however, that the passage immediately preceding it, which we looked at first, addresses basically the same problem. In that first passage, the wording is not as close to Nāgārjuna’s verse as it is in the other passage, and we have no reason to doubt its early provenance. Now the original Vibhāṣā might possibly be earlier than Nāgārjuna.106 The conclusion would then be that the Sarvāstivādins had given thought to the problem of the arising of things even before Nāgārjuna. Nāgārjuna’s originality, with regard to this particular problem, would accordingly be less considerable than one might think. The question of Nāgārjuna’s originality merits further investigation. Another text of the Sarvāstivādins, probably older than the Vibhāṣā,107 contains a passage that is quite similar to a verse from Nāgārjuna’s Mūlamadhyamakakārikā. The text is the Abhidharmahṛdaya of Dharmaśreṣṭhin. The passage reads: (Verse II-10 ab:) “All conditioned (saṃskṛta) dharmas possess arising (jāti), duration (sthiti), change (anyathātva), and destruction (vyaya).” Each of the conditioned dharmas possesses four characteristics (lakṣaṇa): arising, duration, change, and destruction. … Question: If each of the conditioned dharmas possesses these four characteristics, do the latter also possess characteristics? Reply: (Verse II-10 c:) “They also possess the four characteristics.” In these characteristics there arise at the same time four other characteristics: arising creates arising, duration creates duration, change creates change, and destruction creates destruction. Objection: If this is so, there is an infinite regress. (Verse II-10 d:) “They serve as characteristics for one another.” 105 MadhK(deJ) 7.17: yadi kaścid anutpanno bhāvaḥ saṃvidyate kvacit / utpadyeta sa kiṃ tasmin bhāve utpadyate ‘sati // 106 The Vibhāṣā dates somewhere between the first and third centuries of the common era; see Rospatt, 1995: 25-26 n. 38. Nāgārjuna’s date is also uncertain, and likewise ranges from the first to the third century; see Ruegg, 1981: 4 ff. n. 11; 1982: 505-30 (p. 507: 150-200). Walser (2002) dates Nāgārjuna toward the end of the second century. 107 See Willemen, 1975: VI ff. 38 39 Each of these characteristics acts as a characteristic with respect to others. Thus, arisings cause one another to arise, durations cause one another to endure, changes cause one another to change, and destructions destroy one another. Therefore, 108 there is no infinite regress. The problem addressed in this passage is the same as one that Nāgārjuna raises in the seventh chapter of his Mūlamadhyamakakārikā. Consider the third verse from this chapter, where the argument takes the following form: If arising, duration, and perishing themselves have conditioned characteristics (saṃskṛtalakṣaṇa), there will be an infinite regress. If they do not, they are not 109 conditioned. It is possible, however, that the Abhidharmahṛdaya is earlier than Nāgārjuna’s Mūlamadhyamakakārikā. In other words, we might have here another example in support of the idea that Nāgārjuna was not as original as one might think. These few passages suffice to show that the Sarvāstivādins did not, in any case, ignore the arguments of Nāgārjuna;110 they might even have anticipated some of them. Returning now to the problem of the arising of things, we may note that once the problem is solved, several other problems based on the correspondence principle are also solved. In other words, a solution to the problem of the arising of things considerably reduces the threat posed by Nāgārjuna. We shall therefore continue to look at how other thinkers responded to the problem. The challenge posed by Nāgārjuna, and by the correspondence principle in general, was not an exclusively Buddhist concern. Non-Buddhists—Brahmins and even Jainas—were also threatened, and we should therefore take these thinkers into consideration as well. We shall see that virtually all of them, in the period following Nāgārjuna, were concerned with the problem of the arising of things, and that they sometimes proposed quite different solutions. 4. Sāṃkhya First let us consider Sāṃkhya, a non-Buddhist school of thought that in its classical form dates back to the first half of the first millennium. Its teaching includes the “doctrine of the effect [pre-]existing [in the cause]” (satkāryavāda). This position has no clear precedents in earlier literature and seems to be a rather late development within the school.111 We encounter it, perhaps for the first time, in the works of Āryadeva, who is traditionally regarded as the student and successor of Nāgārjuna. 108 TI 1550, vol. 28, p. 811b l. 16-27. For the translation cf. Abhidh-hṛ(A) 2.10, p. 68; Willemen, 1975: 19-20. Note that neither the Abhidharmahṛdaya nor the Amṛtarasa uses the specific expressions jātijāti, utpādotpāda, sthitisthiti, etc., used by Nāgārjuna and other texts of Sarvāstivāda; see Amṛtar(B) p. 44 n. 57. 109 MadhK(deJ) 7.3: utpādasthitibhaṅgānām anyat saṃskṛtalakṣaṇam / asti ced anavasthaivaṃ nāsti cet te na saṃskṛtāḥ // 110 For arguments in the Abhidharmadīpa and the Vibhāṣāprabhāvṛtti, see Bhattacharya, 1991. 111 Liebenthal, 1934: 9 n. 11; Johnston, 1937: 25; Frauwallner, 1953: 385; Larson, 1969: 165; Franco, 1991: 127. We shall examine below a few passages sometimes considered to be early antecedents of satkāryavāda. 39 40 He briefly ridicules the doctrine in his Catuḥśataka (11.15)112 and discusses it in depth in the seventh chapter of his *Śataka.113 Satkāryavāda does not track well with what we know of early Sāṃkhya.114 According to the testimonies of a good number of authors—foremost among them are Bhartṛhari, Mallavādin, and Dharmapāla—the early school considered material objects to be constituted of qualities alone, such as “color, etc.,” or “sound, etc.,” which probably stands for color, sound, taste, smell, and touch. This view of the nature of material objects differs appreciably from the view of later Sāṃkhya, as expressed in commentaries on the Sāṃkhyakārikā. According to these commentaries, material objects are instead constituted of the five elements: earth, water, fire, wind, and ether, which in turn possess the qualities just listed. This new view corresponds well with satkāryavāda. It describes all changes in the world as so many modifications, which do not affect the substrate. The substrate endures without ever losing its essence. It is in virtue of this substrate, and in the form of the substrate, that the effect exists in its cause, as satkāryavāda claims. In the early school, on the other hand, material objects had no substrates, as they were merely collections of qualities. The question of whether satkāryavāda was a part of early Sāṃkhya thus seems fully justified. We would also be right to ask whether it was not, perhaps, the acceptance of satkāryavāda that forced the school to introduce the doctrinal changes just mentioned. The question thus inevitably arises: what led Sāṃkhya to adopt this strange viewpoint of satkāryavāda, which goes against our own intuitions, of course, but which also ran counter to those of Indian thinkers of the period, as demonstrated by the critical reactions of certain other schools? The sources themselves do no tell us, so any answer will necessarily be speculative in nature. As the doctrine of the effect pre-existing in the cause would respond quite well to the problem of the arising of things, and thus to the challenge posed by Nāgārjuna, we might imagine that such was indeed its original purpose. This conjecture was advanced by Walter Liebenthal as early as 1934.115 However, there is no need to insist too much on the connection with Nāgārjuna. If it is true that Āryadeva was his student, we would have to conclude that the doctrine of the effect pre-existing in the cause was introduced into the Sāṃkhya system during Nāgārjuna’s lifetime. This is not entirely impossible, of course, but it seems wiser to conclude more modestly that satkāryavāda was a response to the problem of the arising of things, a problem that seems to have occupied thinkers even before Nāgārjuna. How do the classical texts of the school justify satkāryavāda? The answer lies at the end of kārikā 9 of the Sāṃkhyakārikā, the base text of Sāṃkhya philosophy. There we find the words kāraṇabhāvāc ca sat kāryam, which might be translated: “and because [the cause] is a cause, the effect exists.” Here we have the same kind of argument that Nāgārjuna uses so effectively: for there to be a cause, there needs to be an effect. And the effect must be there at the same time as the cause. I insist on this “Nāgārjunian” reading of the phrase, knowing full well that from among the Sāṃkhya commentaries only the Yuktidīpikā, the most detailed 112 Lang, 1986: 106-07. Satkāryavāda seems to be refuted also in the Yogācārabhūmi; see Mikogami, 1969. Translated into English by G. Tucci (1981). 114 On this subject, see Bronkhorst, 1994: 315 ff. 115 Liebenthal, 1934: 4. 113 40 41 commentary on the Sāṃkhyakārikā, appears to support my interpretation.116 All of the other commentaries that I have seen interpret the expression kāraṇabhāvāt to mean that the effect is identical to the cause.117 The Yuktidīpikā comments on the expression as follows: “If the effect does not exist, [the cause] is not a cause.”118 From a strictly logical point of view, this explanation is, of course, equivalent to the expression being explained. But it suggests that unless the effect is present, there is no cause. We shall see in what follows that the kind of argument used by Nāgārjuna, as well as the correspondence principle that underlies it, subsequently disappeared from most philosophical discussions in India, for reasons we shall come to later. It therefore seems plausible to take the “Nāgārjunian” interpretation as the original interpretation, and the other as a later interpretation, sought out and discovered at a time when the original reasoning had lost its power to convince.119 Nevertheless, it would be desirable to have textual confirmation of our interpretation of the argument from the Sāṃkhyakārikā. We find such confirmation in other texts, to which we shall now turn. Already in Āryadeva’s *Śataka the argument from the Sāṃkhyakārikā is placed in the mouth of the defender of Sāṃkhya. Giuseppe Tucci translates it from Chinese to English as follows: 120 The effect pre-exists in the cause, on account of the existence of the cause. In this form, the remark makes no sense. The phrase “on account of the existence of the cause” might well correspond, however, to the Sanskrit kāraṇabhāvāt—the expression from the Sāṃkhyakārikā. Nothing in the Chinese version conflicts with this reconstruction. The correct translation would then be: The effect is present in the cause, because [the latter] is a cause. Or possibly: The effect is present in the cause, because [the effect] is [identical to] the cause. It is possible that “in the cause” was added by the translator, Kumārajīva: we know that he revised his first translation of the text (now lost) to make it more 116 See however below for support from other sources; further Bronkhorst, 2000, for supporting evidence from Jaina sources. 117 A third interpretation is theoretically possible: “because of the existence of the cause.” This interpretation, however, does not amount to a comprehensible argument. Note further that the Sāṃkhya Sūtra attributed to Kapila, a relatively late text of the school, contains the sūtra (1.118) kāraṇabhāvāc ca; the commentators of this text, too, interpret it to mean that the effect is identical to the cause. Muroya (1996: 53) sees a resemblance between the expression niyatakāraṇaṃ kāryam in the Nyāya Bhāṣya (on sūtra 4.1.49) and the kāraṇabhāvāt of the Sāṃkhyakārikā. 118 YD p. 54 l. 32 (Pandeya) / p. 124 l. 7 (Wezler & Motegi): ihāsati kārye kāraṇabhāvo nāsti. 119 Solomon (1974: 149, 179) considers the Sāṃkhyavṛtti (V2) to be the oldest commentary on the Sāṃkhyakārikā; she dates it, approximately and without precise evidence, to the end of the fourth or the beginning of the fifth century. The fact that the Yuktidīpikā seems to preserve an older interpretation suggests, however, that it is older than the Sāṃkhyavṛtti. For the date of the Yuktidīpikā, see note 124 below. 120 TI 1569, vol. 30, p. 177b l. 26; tr. Tucci, 1981: 61. 41 42 comprehensible to his Chinese audience. In the course of this revision, Kumārajīva might have translated the verses of the original into prose with his own additions and explanations.121 In other words, one cannot rule out the possibility that Āryadeva’s Sanskrit original—which unfortunately we do not have—was close, or even identical, to the Sāṃkhyakārikā. The commentator Vasu offers the following explanation, as translated by Tucci: If the pot does not pre-exist in earth, then earth could not become the cause of the 122 pot. Here one has the impression that Vasu is merely paraphrasing the explanation from the Yuktidīpikā, which we have already seen: “If the effect does not exist, [the cause] is not a cause.” There is, however, an important difference. While the formulation in the Yuktidīpikā is somewhat ambiguous, Vasu makes it clear that we are speaking of the cause. The cause requires the effect in order to become a cause. Without the existence of the pot, for instance, earth could not become its cause. The “Nāgārjunian” nature of the argument is now undeniable. Vasu’s commentary is probably nearer to the interpretation of the Yuktidīpikā on this point than to the interpretation of other commentaries on the Sāṃkhyakārikā, which take the line to mean that the effect is identical to the cause. (Note, however, that this other argument was also known to Āryadeva and Vasu.123) Āryadeva and his commentator Vasu thus confirm that the considerations adduced in favor of satkāryavāda included, probably from the outset, an argument of a Nāgārjunian nature, claiming that a cause cannot exist unless its effect is there at the same time. One might wonder how to account for the close parallels between Āryadeva and his commentator Vasu, on the one hand, and the Sāṃkhyakārikā and its commentary the Yuktidīpikā, on the other. In my opinion, the most likely answer is that both Āryadeva and the Sāṃkhyakārikā made use of a Sāṃkhya work no longer in existence; this work would have been one of the first, or even the very first, to offer arguments in support of satkāryavāda. Vasu and the Yuktidīpikā might likewise have drawn from a common, now lost commentary. There is also another possibility, less likely in my opinion, but important enough to merit a brief digression. I recall the American scholar Karen Lang drawing attention to the fact that Vasu’s commentary seems to be composed in the Vārttika-style, which is characterized by aphoristic sentences followed by explanations in less dense prose. This would mean that some of the so-called sūtras from the text that Kumārajīva translated might be aphorisms from Vasu’s commentary, rather than Āryadeva’s original work. If so, then the two sentences examined above might both belong to Vasu’s commentary. He might even have borrowed them directly from a work of the Sāṃkhya school. This possibility 121 Lang, 1988: 132. TI 1569, vol. 30, p. 177b l. 26-27; tr. Tucci, 1981: 61. 123 See TI 1569, vol. 30, p. 177b l. 29 - p. 177c l. 2; Tucci, 1981: 62. The Vibhāṣāprabhāvṛtti, which in its date of composition should not be too far from Vasubandhu’s Abhidharmakośa Bhāṣya, presents the second position in the following passage (Abhidh-d, introducing kārikā 310, p. 273 l. 29-30): sāṃkhyaḥ paśyati: vidyamānam eva jāyate / tad yathā kṣīre vidyamānaṃ dadhi, kāryakāraṇayor ekatvāt / 122 42 43 becomes more likely in view of Vasu’s having done the same thing, at least on Karen Lang’s thesis, with certain Nyāya texts: Vasu might have quoted several sūtras from this school, presenting them as aphorisms of his own commentary.124 This possibility certainly opens the door to the hypothesis that Vasu might have borrowed the two sentences from a Sāṃkhya work. Since Kumārajīva completed his (second) translation in 404 C.E., it seems unlikely that he knew of the Sāṃkhyakārikā and the Yuktidīpikā; the latter appears in any case to be later than Bhartṛhari,125 thus certainly later than 404. But it seems at least possible that Vasu borrowed the two sentences from an older Sāṃkhya work, such as the work of Vārṣagaṇya known, probably, as the Ṣaṣṭitantra. This hypothesis is all the more appealing in that the Ṣaṣṭitantra was likewise composed in the Vārttika-style.126 The Ṣaṣṭitantra was well aware of satkāryavāda. The text itself is lost, of course, but we have some partial knowledge of it through the works of a few later authors who critique its positions. The Dvādaśāranayacakra of the Jaina author Mallavādin is a particularly valuable source of information concerning the Ṣaṣṭitantra.127 Mallavādin analyzes the argument in question and rejects it, for reasons we shall examine below. It is worth noting here that the Ṣaṣṭitantra set forth not only the view that the effect exists in its cause, but also the complementary view that the cause exists in its effect. This is the view according to which everything is made of everything; following Albrecht Wezler, we shall refer to this position as sarvasarvātmakatvavāda.128 This complementary doctrine was not as successful as satkāryavāda; indeed, it nearly vanished without a trace. Having happily avoided this fate, however, the doctrine of sarvasarvātmakatvavāda allows us to conclude that early Sāṃkhya thinkers were interested not only in the arising of things, but also in their disappearance. This is hardly surprising, as the two problems are related. Nāgārjuna had already rejected the possibility that things— meaning dharmas, of course—that never arose should pass away.129 Sarvasarvātmakatvavāda, viewed in this light, would have developed under the influence, this time indirect, of the correspondence principle. This principle rules 124 Lang, 1988. Bronkhorst, 1985: 93 ff. To the arguments presented in this publication, a few others can now be added. Bhartṛhari knew the Sāṃkhya doctrine in an older form than that of the Yuktidīpikā; see Bronkhorst, 1994. The Yuktidīpikā probably postdates Praśastapāda (if, at least, the latter indeed introduced into Vaiśeṣika the notion of God as cause of the world, which seems likely; see Bronkhorst, 1996a); Praśastapāda, in turn, is without a doubt later than Dignāga (Bronkhorst, 1993: 158 ff.; 1993a: 705 ff.), whom the Yuktidīpikā also quotes (Wezler & Motegi, 1998: XXVII). That Dignāga came after Bhartṛhari is well known. We should view with caution the supposed citation of the Kāśikā in the Yuktidīpikā (Wezler & Motegi, 1998: XXVIII), in light of the fact that the Kāśikā quotes grammatical texts, unfortunately no longer extant, other than the Cāndravyākaraṇa and Jainendravyākaraṇa; see Bronkhorst, 2002; 2003a. 126 Cf. Frauwallner, 1958: 101 (240): “Vṛṣagaṇa (i.e. Vārṣagaṇya, JB) [gab] öfter eine entscheidende Formulierung zunächst in wenigen knappen Worten … und [erklärte] sie dann in einem weiteren Satz genauer …” 127 Note that the Dvādaśāranayacakra was partially reconstituted by Muni Jambuvijaya with the help of Siṃhasūri’s commentary, the Nyāyāgamānusāriṇī. 128 See Wezler, 1981. 129 MadhK(deJ) 1.9: anutpanneṣu dharmeṣu nirodho nopapadyate. 125 43 44 out the possibility of something non-existent arising. And without arising, there is no passing away.130 We said that Āryadeva’s *Śataka offers one of the earliest reports of the doctrine of satkāryavāda. This point obliges us to consider briefly a text that tradition attributes to Āryadeva’s predecessor, Nāgārjuna. I have in mind the Mahāprajñāpāramitāśāstra or Mahāprajñāpāramitopadeśa, which is preserved in the Chinese translation of Kumārajīva, completed in 404 or 405; a large part of it has been admirably translated into French by Étienne Lamotte. The scholar K. Venkata Ramanan, in his book Nāgārjuna’s Philosophy,131 suggests that this text knows of satkāryavāda as a doctrine of Sāṃkhya. This is not the case, however. Two passages supposedly mention and criticize the doctrine. They are: If the cause (kāraṇa) pre-exists in the cause, there is no effect (kārya); if the cause 132 does not pre-exist in the cause, there is no effect either. And again: It is illogical (na yujyate) for the effect (phala) to pre-exist in the cause (hetu); it is equally illogical for it not to. If the effect pre-existed in the cause, it would have no cause [since it would exist already]. If it did not pre-exist in the cause, what would be the point of the cause [since the 133 effect would not be there]? These passages recall arguments from the Mūlamadhyamakakārikā and make no mention of Sāṃkhya. They in no way prove that satkāryavāda was already a Sāṃkhya doctrine at the time of Nāgārjuna. We should add that the traditional attribution of this text to Nāgārjuna is not unanimous.134 Moreover, the text knows a relatively late form of Sāṃkhya, one that probably postdates Nāgārjuna. Recall that in the older form of the system, material objects are merely collections of qualities. Qualities are included in the list of the tattvas, the elements derived from primordial nature (prakṛti). The form of the system as it appears in the Mahāprajñāpāramitāśāstra, however, is closer to the classical system.135 Its description of the evolution of the elements does not include qualities, whereas it does include the tanmātras, which are part of the classical system. It is true that the tanmātras of the Mahāprajñāpāramitāśāstra, unlike those of the classical system, seem to be atoms, serving to constitute the molecules of the 130 Note that the Vaiśeṣika thinker Vyomaśiva describes the position of the Sāṃkhyas as follows (Vyomavatī II p. 126 l. 12-13): … sāṃkhyānām ayaṃ viparyayaḥ: kṛtakeṣu ghaṭādiṣu nityatvadarśanam iti, na samutpadyate bhāvo na svātmānaṃ jahātīti. The phrase kṛtakeṣu nityatvadarśanam comes from the Padārthadharmasaṅgraha (WI p. 41 § 222), but we cannot conclude with certainty that Praśastapāda attributed this position to the Sāṃkhyas; according to another interpretation of Vyomaśiva (Vyomavatī II p. 126 l. 10-11) and that of Srīdhara (Nyāyakandalī p. 422 l. 5), the phrase concerns the Mīmāṃsakas. 131 Ramanan, 1966: pp. 178-79, with p. 360 n. 18a. 132 TI 1509, vol. 25, p. 104c l. 24-26; translation based on Mppś I p. 380. The Chinese text contains without a doubt an error in translation or in transmission, putting “cause” (in its first and third occurrences) in place of “effect.” 133 TI 1509, vol. 25, p. 296b l. 18-19; translation based on Mppś V p. 2172. 134 Mppś III p. VIII ff., especially p. XL. 135 See the Appendix at the end of this volume. 44 45 elements. Nonetheless, it seems likely that by the time of the Mahāprajñāpāramitāśāstra, material objects were no longer regarded as mere collections of qualities. And one hesitates to assign too early a date to the transition (whether gradual or sudden) to the classical system, since the older form of the system was still known to Bhartṛhari, Mallavādin, and Dharmapāla. Note that all of this militates against the traditional position that would see Nāgārjuna as the author of the Mahāprajñāpāramitāśāstra. The worldview of classical Sāṃkhya, in essence, leaves no room for a real arising or destruction of things. This is exactly what the Yuktidīpikā says in the following passage: “[To say] that [something] is made, arises, is born, is but a common linguistic convention; in reality nothing arises, and nothing perishes.”136 We have already looked at the end of verse 9 of the Sāṃkhyakārikā. The beginning of the verse contains an additional argument in favor of satkāryavāda, which also merits our attention. The argument appears in the form of the compound asadakaraṇāt, which becomes intelligible when construed with the closing words of the verse, sat kāryam; the compound can then be translated as follows: “[the effect exists,] because one does not make what does not exist.” In other words, in a sentence such as “The man is making a mat,” there must be a mat in order for the man to make it. Once again we see the correspondence principle at work. Here an opponent might object that if the effect is already existent, the agent’s action is useless. The Yuktidīpikā replies: “If the action related to something that both [the opponent and the Sāṃkhya] accepted as non-existent, then one could say that the agent’s effort is useless when the effect is [already] there. But it is not possible to act on something non-existent. This [objection] is therefore inappropriate.”137 Let us now examine another source of classical Sāṃkhya thought: the text described in its colophons as pātañjala sāṃkhyapravacana yogaśāstra, or “Patañjali’s Treatise on Yoga Explaining Sāṃkhya [Thought].” It is better known as the Yoga Sūtra and Yoga Bhāṣya, and indeed consists of sūtras with commentary. The later tradition of attributing only the sūtras to an author named Patañjali, and the commentary to an author named Vyāsa, is less reliable.138 Although the attribution of the two texts is naturally an important question, we shall not concern ourselves with it here. Whatever the names of the authors, the two texts shed light on a number of questions concerning the Sāṃkhya system, and can also contribute to the present discussion. The Sūtra and Bhāṣya—the latter dating probably to the beginning of the fifth century of the common era139—together include a passage that is worth quoting in full. It reads: What does not exist does not arise, and what exists does not pass away; in view of this [truth], how do the unconscious impressions (vāsanā), which arise as real things, pass away? 136 YD p. 54 l. 3-6 (Pandeya) / p. 121 l. 20 - p. 122 l. 2 (Wezler & Motegi): kriyate utpadyate jāyata ity evamādir lokasya vyavahāraḥ pravartate … paramārthatas tu na kasyacid utpādo ‘sti na vināśaḥ. 137 YD p. 52 l. 2-3 (Pandeya) / p. 117 l. 7-9 (Wezler & Motegi): yady ubhayapakṣaprasiddhasyāsataḥ kriyāyogaḥ syāt ata etad yujyate vaktum kārye sati kartuḥ prayāso ‘narthaka iti / tat tv asataḥ karaṇam anupapannam / tasmād ayuktam etat /. The opponent’s position is first expressed p. 48 l. 32 - p. 49 l. 3 (Pandeya) / p. 111 l. 7-11 (Wezler & Motegi). 138 For more on the question of authorship, see Bronkhorst, 1984; Maas, 2006: xii f. 139 See Bronkhorst, 1985: 171. 45 46 The past and the future exist in their own nature, because properties are divided along the [three] times. (YS 4.12) That which will be manifested is future, that which has been manifested is past, that which is engaged in its own activity is present. These three things constitute the object of knowledge. And if this [threefold object] did not exist in its own nature, this knowledge, which would be without an object, would not arise. Hence [the sūtra says:] “The past and the future exist in their own nature.” Moreover, if the fruit of an action that leads to enjoyment, or [of an action] that leads to liberation, were non-existent (nirupākhya) when it is [yet] to arise, then the good conduct that has [such fruit] as its goal or cause would be useless. And the cause of an [already] existent fruit is capable of making it present, but not of creating something new. The cause, when it is established, performs a specific service for that 140 which is caused, but does not produce anything new. This position is indistinguishable from that of the Sāṃkhyas. In the final analysis, nothing arises, and the question of knowing how something can arise makes no sense. At the same time, nothing that exists passes away. In other words, the Yoga Sūtra and Bhāṣya present a position that combines satkāryavāda and sarvasarvātmakatvavāda. But the similarity of the position of the Yoga Bhāṣya to Sarvāstivāda is equally evident. Indeed, it appears elsewhere in the text. Sūtra 3.13 and its commentary, in particular, mention the different ways in which things pass from the future, through the present, into the past—ways found in the Vibhāṣā and other Sarvāstivāda texts.141 It therefore seems undeniable that at least the version of Sāṃkhya expressed in the Yoga Sūtra and Yoga Bhāṣya is similar to Sarvāstivāda, probably borrowing from it even the doctrine that gave the latter its name. It seems likely that satkāryavāda and sarvasarvātmakatvavāda likewise had historical links to Sarvāstivāda. We know nothing of the circumstances that would have made such a borrowing possible; our texts have nothing explicit to say on the subject. By contrast, we do know the results of this borrowing. They include the notion that nothing truly arises, a notion that rendered Sāṃkhya impervious to Nāgārjuna’s critiques. Before leaving Sāṃkhya, let us examine a few theories that have been advanced to explain the origins of satkāryavāda and sarvasarvātmakatvavāda. Albrecht Wezler sees in the latter a conception of being (Seinsbegriff) that is also found in the Bhagavadgītā and in the teaching of Uddālaka Āruṇi in the Chāndogya Upaniṣad. The Bhagavadgītā contains the line: “There is no existence from what 140 Yoga Sūtra 4.12 and its Bhāṣya: nāsty asataḥ saṃbhavaḥ, na cāsti sato vināśa iti dravyatvena saṃbhavantyaḥ kathaṃ nivartiṣyante vāsanā iti, atītānāgataṃ svarūpato ‘sty adhvabhedād dharmāṇām (YS 4.12) bhaviṣyadvyaktikam anāgatam, anubhūtavyaktikam atītam, svavyāpāropārūḍhaṃ vartamānam / trayaṃ caitad vastu jñānasya jñeyam / yadi caitat svarūpato nābhaviṣyan nedaṃ nirviṣayaṃ jñānam udapatsyata, tasmād atītānāgataṃ svarūpato ‘stīti / kiñ ca bhogabhāgīyasya vāpavargabhāgīyasya vā karmaṇaḥ phalam utpitsu yadi nirupākhyam iti taduddeśena tena nimittena kuśalānuṣṭhānaṃ na yujyate / sataś ca phalasya nimittaṃ vartamānīkaraṇe samarthaṃ nāpūrvopajanane / siddhaṃ nimittaṃ naimittikasya viśeṣānugrahaṃ kurute, nāpūrvam utpādayatīti / 141 See La Vallée Poussin, 1937a: 237 ff.; Chakravarti, 1951: 95 ff.; Ruegg, 1978; cf. Frauwallner, 1973: 99 ff. ≈ 1995: 186 ff. 46 47 does not exist; there is no non-existence from what exists.”142 The Chāndogya Upaniṣad has Uddālaka Āruṇi say: “This world, my dear, was existent in the beginning, one only, without a second. Some say on this subject: ‘This world was non-existent in the beginning, one only, without a second. The existent was born from this non-existent.’ But how, my dear, could this be?” he said. “How could the existent be born from the non-existent? This world, my dear, was instead existent in the 143 beginning, one only, without a second.” Wezler believes there is a historical connection between sarvasarvātmakatvavāda and the ideas expressed in these passages. Wilhelm Halbfass, for his part, considers the passage from the Chāndogya Upaniṣad to be the background of satkāryavāda.144 In another article, Wezler maintains that a short phrase from the Mahābhāṣya of Patañjali proves that the latter knew of sarvasarvātmakatvavāda.145 I do not wish to disagree with these scholars, but feel that caution is required. The scene presented by the history of ideas is often more complicated than that of various traditions existing in strict isolation. A single original idea can give rise to numerous, sometimes contradictory conceptions, and conversely, convergences of ideas are not uncommon. In Brahmanical traditions in particular, connections with Vedic sources were inevitably stressed, whether such connections were genuine or not. Even so, there remains a fundamental difference between the teaching that existence cannot come from what does not exist, and satkāryavāda. A pot is made of clay, and comes from clay. It is quite another thing to say that the pot is present in the clay before it is made.146 So while there might well be a historical connection between satkāryavāda and the passages from early Brahmanical literature just cited, their content is profoundly different. These passages therefore do not explain the presence of satkāryavāda in Brahmanical thought.147 In fact, the three passages cited by Wezler and Halbfass—one from the Bhagavadgītā (as interpreted by Wezler), another from the Chāndogya Upaniṣad, a third from the Mahābhāṣya—are all missing an essential feature: namely, the presence of the correspondence principle, even in a disguised form. This principle dominated a certain period of philosophical thought in India, a period beginning, it 142 Bhag 2.16ab: nāsato vidyate bhāvo nābhāvo vidyate sataḥ. This is clearly the way Wezler (1981: 398) understands the line (even though he does not translate it), for he describes it as “m.E. historisch nicht zu trennen” from the teaching of Uddālaka Āruṇi in the Chāndogya Upaniṣad. 143 Ch-Up 6.2.1-2: sad eva somyedam agra āsīd ekam evādvitīyam / tad dhaika āhur asad evedam agra āsīd ekam evādvitīyam / tasmād asataḥ saj jāyata // kutas tu khalu somyaivaṃ syād iti hovāca / katham asataḥ saj jāyeteti / sat tv eva somyedam agra āsīd ekam evādvitīyam //. For a discussion of the historical context of this passage, see Bodewitz, 1995. 144 Halbfass, 1992: 59. 145 Wezler, 1982. The phrase in question is: atha matam etat prakṛtyanvayā vikārā bhavantīti (Mahābhāṣya ed. Kielhorn, on P. 4.3.155, vol. II p. 325 l. 17-18); which Wezler translates (p. 159): “Wenn man … aber die Meinung vertritt, dass … bei Produkten-einer-Veränderung das Ursprüngliche ‘folgt’ (d.h. nach wie vor da ist / erhalten bleibt) ….” 146 The error is nonetheless common. Śrīvāstavya (1983: 132 ff.), for example, does not hesitate to conclude from the fact that the pot arises from a modification of clay that the pot is present in the clay even before its arising. Proceeding in this fashion, he finds satkāryavāda in the Carakasaṃhitā, which seems to make no mention of it. 147 Theoretically it is of course possible to interpret the Bhagavadgītā passage differently from Wezler. This possibility will not here be examined. 47 48 would seem, with the scholastic developments of Buddhism and culminating in the work of Nāgārjuna. The problems associated with the principle were taken up by Brahmanical thinkers, who situated them in relation to their own traditions. Sarvasarvātmakatvavāda as well as satkāryavāda thus have, perhaps, some historical connection with the passage from the Chāndogya Upaniṣad. Nevertheless, it is impossible to ignore the obvious connection between these two doctrines and Sarvāstivāda teachings on the existence of the future and the past. The two doctrines are thus additionally—or even exclusively—rooted in developments of the correspondence principle. This root, I believe, is more important and more decisive than the Vedic root, for it explains the occurrence and the very form of satkāryavāda, while the Vedic connection (if it exists) explains nothing. 5. The Āgamaśāstra of Gauḍapāda Let us now turn to a text belonging to a third strand of Brahmanical thought. The Āgamaśāstra, also known as the Gauḍapāda-Kārikā or Māṇḍūkya-Kārikā, is traditionally attributed to a certain Gauḍapāda. We cannot take this attribution at face value, however, for there is no guarantee that the work is by a single author.148 Its fourth chapter in particular seems to be a Buddhist text, or at least deeply influenced by Buddhist ideas. Still, the Āgamaśāstra in its entirety is considered an early work of Vedānta philosophy. The fact remains, however, that the parallels between the Āgamaśāstra and certain Buddhist texts are many; among the latter, Nāgārjuna’s Mūlamadhyamakakārikā figures prominently.149 First let us look at a passage dealing with the problem of the arising of things. It comes from the notorious fourth chapter: In their debates with one another, some teachers maintain the arising of what exists; other intelligent ones [maintain the arising] of what does not exist. Nothing that exists can arise—what does not exist cannot arise [either]; arguing thus, followers of non-duality teach non-arising. We approve of the non-arising taught by them; we are not in contradiction with 150 them. Listen to how there is no contradiction. Once again we recognize the familiar argument of Nāgārjuna.151 Arising not being possible, it does not exist. The third chapter likewise presents this argument in a few verses: 148 See, e.g., Vetter, 1978; Wood, 1990, with Bronkhorst, 1992; and more recently King, 1995: 15-49; 1995a. Note that the Yogavāsiṣṭha sometimes adopts positions quite close to those of the Āgamaśāstra; a study of the link between these two works was proposed by Walter Slaje (1994: 197 n. 3). In more recent studies, Slaje dates the composition of the Mokṣopāya (from which the Yogavāsiṣṭha derived) “during or soon after Yaśaskaradeva’s reign, AD 939-948” (2005: 22). On the use of the correspondence principle in the Mokṣopāya, see Bronkhorst, 2001: 207 f. 149 See especially Bhattacharya, 1943; Nakamura, 1992; Bouy, 1997. 150 GK 4.3-5: bhūtasya jātim icchanti vādinaḥ kecid eva hi / abhūtasyāpare dhīrā vivadantaḥ parasparam // bhūtaṃ na jāyate kiñcid abhūtaṃ naiva jāyate / vivadanto ‘dvayā hy evam ajātiṃ khyāpayanti te // khyāpyamānām ajātiṃ tair anumodāmahe vayam / vivadāmo na taiḥ sārdham avivādaṃ nibodhata // 151 Nakamura (1992: 243) also cites in this context a line from the Laṅkāvatāra Sūtra (Laṅkāv(V) 3.21ab p. 62: asan na jāyate loko no san na sadasan kvacit); but this line deals with the nature of the world, not the 48 49 No individual soul (jīva) arises; it has no origin. Such is this ultimate truth, wherein 152 nothing arises. And again: The birth of something existent is possible through illusion, but not in reality. For someone who thinks that [something] arises in reality, it is an arisen thing that arises. The birth of something non-existent is possible neither through illusion nor in 153 reality. The son of a barren woman is born neither in truth nor through illusion. We see that “Gauḍapāda” reaches a conclusion similar to that of Nāgārjuna, using the latter’s arguments. Unlike the Sarvāstivādins and the Sāṃkhyas, he does not conclude that the effect exists in the cause, or that things exist prior to their arising. No, for him these arguments prove that, in ultimate truth, there is no arising. Elsewhere in the third chapter he sets forth this position on the basis of Upaniṣadic statements, which shows that the impossibility of arising is here presented as a Vedāntic position. 6. Śaṅkara Having mentioned “Gauḍapāda,” we must now speak of the Āgamaśāstravivaraṇa, a commentary on the Āgamaśāstra apparently written by the renowned Śaṅkara, author of the Brahma Sūtra Bhāṣya and the Upadeśasāhasrī, as well as other works.154 A Vedāntin, the author of the Āgamaśāstravivaraṇa interprets the typically Buddhist passages of the fourth chapter in a clearly Vedāntic way. For instance, the first verse of this chapter uses two typically Buddhist terms, dharma and saṃbuddha. The verse can be translated as follows: “That one who, through a knowledge that is similar to ether and non-different from its object, is perfectly awakened (saṃbuddha) to the dharmas, themselves similar to ether—to that best of bipeds, I offer homage.”155 Read as a Buddhist verse, these lines are not difficult to interpret. Many Buddhists, primarily those belonging to the Great Vehicle (Mahāyāna), indeed believe that the dharmas, the elementary entities of Buddhism, are void like ether. The commentator Śaṅkara, by contrast, interprets the key term nature of arising. Cf. the following verse (Laṅkāv(V) 3.22): na san nāsan na sadasad yadā lokaṃ prapaśyati / tadā vyāvartate cittaṃ nairātmyaṃ cādhigacchati // 152 GK 3.48: na kaścid jāyate jīvaḥ sambhavo ‘sya na vidyate / etat tad uttamaṃ satyaṃ yatra kiñcin na jāyate // 153 GK 3.27-28: sato hi māyayā janma yujyate na tu tattvataḥ / tattvato jāyate yasya jātaṃ tasya hi jāyate // asato māyayā janma tattvato naiva yujyate / vandhyāputro na tattvena māyayā vāpi jāyate // 154 See Mayeda, 1967; Hacker, 1968: 124 (218) ff.; 1972: 116 (253) n. 2; Vetter, 1968: 407-09; Bouy, 1997: 127. Wood, 1990: 97 ff., is against such an identification. 155 GK 4.1: jñānenākāśakalpena dharmān yo gaganopamān / jñeyābhinnena saṃbuddhas taṃ vande dvipadāṃ varam // 49 50 dharma as “soul” (ātman). The perfectly awakened being of the verse becomes, in Śaṅkara’s interpretation, the supreme deity Nārāyaṇa.156 The Āgamaśāstravivaraṇa does not try to minimize the value of verses presenting “Nāgārjunian” arguments. On the contrary, the commentary does its best to elucidate them. Āgamaśāstra 4.4, which we translated earlier, says for example: “Nothing that exists can arise—what does not exist cannot arise [either]” (bhūtaṃ na jāyate kiñcid abhūtaṃ naiva jāyate). Śaṅkara glosses this line as follows: “Nothing that exists, [i.e.,] that is present as a real thing (vastu), can arise, precisely because it is already present. … Likewise, what does not exist, [i.e.,] what is not present, cannot arise, because it is not present, like the horn of a hare.”157 Śaṅkara explains the argument correctly, and nothing suggests that he disagrees with it. Śaṅkara’s commentary on Āgamaśāstra 4.22 introduces a further consideration, which is also well known from Buddhist texts: namely, the doctrine that the objects of the phenomenal world are merely words. The passage reads: Objection: A pot arises from clay, and a son from a father. [Reply:] It is true that among uninstructed people, [we find] the idea and the verbal expression “[That] arises.” This verbal expression and this idea are examined by people of discernment, [who ask:] “Do these two represent the truth, or are they 158 false?” The thing characterized as a pot, a son, etc., which is the object of the verbal expression and the idea, insofar as it is examined, is only a word. For the 159 Veda says: “a support for words.” If [a thing is] existent, it does not arise, for it exists [already], like clay, the father, etc. If it is non-existent, it still does not arise, precisely because it does not exist, like the horn of a hare, etc. If it is both existent and non-existent, again it does not arise, for it is impossible for a single [thing] to be 160 contrary [to itself]. Hence it is established that nothing arises. Things are therefore nothing other than words, and an existent thing does not arise. Note that Śaṅkara quotes in this passage from the teaching of Śvetaketu’s father, which we discussed earlier. We had remarked on that occasion that the acceptance of Buddhist ideas would later be facilitated by the existence of Vedic passages, even isolated ones, preaching more or less similar views. In fact, the Brahma Sūtra (2.1.14/15: tadananyatvam ārambhaṇaśabdādibhyaḥ) already refers explicitly to this Vedic passage, and the commentator Rāmānuja does not miss the 156 Śaṅkara on GK 4.1: dvipadāṃ varaṃ dvipadopalakṣitānāṃ puruṣāṇāṃ varaṃ pradhānaṃ puruṣottamam ity abhiprāyaḥ. 157 Śaṅkara on GK 4.4: bhūtaṃ vidyamānaṃ vastu na jāyate kiṃcid vidyamānatvād ev[a] … / tathābhūtam avidyamānaṃ avidyamānatvān naiva jāyate śaśaviṣāṇavad … / 158 It seems clear that the word viṣaya, “object,” is here, exceptionally, neuter. Cf. also YD p. 48 l. 27 (Pandeya) / p. 110 l. 22 (Wezler & Motegi): syād etat yadi tarhi pratyakṣaviṣayam evāsti. The phrase buddhiviṣayam eva śabdānāṃ paurvāparyam from the Mahābhāṣya (Mbh I p. 356 l. 11) might also be using the word viṣaya in the neuter; cf. Bronkhorst, 1987: 49. 159 Śaṅkara on Chāndogya Upaniṣad 6.1.4 glosses the expression vācārambhaṇa with vāgārambhaṇa and vāgālambana. 160 Śaṅkara on GK 4.22: nanu mṛdo ghaṭo jāyate pituś ca putraḥ / satyam, asti jāyata iti pratyayaḥ śabdaś ca mūḍhānām / tāv eva śabdapratyayau vivekibhiḥ parīkṣyete kiṃ satyam eva tāv uta mṛṣeti / yāvatā parīkṣyamāṇaṃ śabdapratyayaviṣayaṃ vastu ghaṭaputrādilakṣaṇaṃ śabdamātram eva tat / “vācārambhaṇam” iti śruteḥ / sac cen na jāyate sattvān mṛtpitrādivat / yady asat tathāpi na jāyate ‘sattvād eva śaśaviṣāṇādivat / atha sadasat tathāpi na jāyate viruddhasyaikasyāsaṃbhavāt / ato na kiṃcid vastu jāyata iti siddham /. I accept the variant parīkṣyamāṇaṃ in place of parīkṣyamāṇe, in spite of Ānandagiri’s commentary: mṛṣaiveti parīkṣyamāṇe satīti saṃbandhaḥ. 50 51 opportunity to emphasize that, contrary to the view of the Vaiśeṣikas, the existence of distinct words and notions does not imply the existence of distinct objects; the objects of this world do not exist as such; they are not different from the ultimate cause of the world, Brahman.161 Elsewhere in his commentary on the Āgamaśāstra Śaṅkara presents another point of view, at least on the question of arising.162 Consider Āgamaśāstra 2.32, which reads: “There is no destruction, and no arising; no one bound, and no practitioner; no one who desires liberation, and no one liberated. Such is ultimate truth.”163 In his commentary on this verse, Śaṅkara remarks: “For something existent can arise or be destroyed, [but] not something non-existent like the horn of a hare.”164 Note the difference between this position and the one we saw a moment ago. There, nothing existent or non-existent could arise; here, only an existent thing can arise. Note also that the position Śaṅkara presents here goes against the verse in the base text. This position appears elsewhere in his commentary, too. Āgamaśāstra 1.6 is one example. The first half of this verse might be translated as follows: “There is the position that all things arise being [already] existent.”165 This translation follows Paul Hacker’s suggestion (1972: 126) that the verse does not represent the opinion of “Gauḍapāda.”166 Śaṅkara would not agree with this interpretation, as Hacker rightly adds.167 We must conclude that Śaṅkara, contrary to what he affirms in his commentary on the fourth chapter, expresses support on several occasions for the view that an existent thing arises. This position is of course similar to that of Sāṃkhya. Which of the two positions expressed in the Āgamaśāstravivaraṇa represents the view of its author? The fact that Śaṅkara, even when commenting on a verse that maintains the impossibility of arising, sets forth the position that an existent thing can arise suggests that this is his real position. This impression is confirmed by his other works. Only the Āgamaśāstravivaraṇa seems to feature the argument that an existent thing cannot arise, while Śaṅkara’s own argument appears throughout his works. Generally speaking, in all of his writings except for the fourth chapter of the Āgamaśāstravivaraṇa, Śaṅkara adheres to satkāryavāda, the doctrine familiar to us from Sāṃkhya. He even characterizes himself as a satkāryavādin, as for example in his Brahma Sūtra Bhāṣya on sūtra 2.1.7. Elsewhere in the same text he remarks: “What is not present somewhere in a certain form 161 Rāmānuja, Śrī Bhāṣya on sūtra 2.1.15 (pp. 465-68): tatra kāṇādā prāhuḥ: na kāraṇāt kāryasyānanyatvaṃ saṃbhavati, vilakṣaṇabuddhibodhyatvāt …, śabdabhedāc ca … / atrāhuḥ: kāraṇād ananyat kāryam / … buddhiśabdāntarādibhir vastvantaratvaṃ sādhitam iti cen na: anirūpitavastvālambanānāṃ buddhiśabdāntarādīnāṃ śuktikārajatabuddhiśabdādivad bhrāntimūlatvena vastvantarasadbhāvasyāsādhakatvāt. 162 The close connection between words and things is expressed through the frequent use of the expression nāmarūpa; see Hacker, 1950: 258 (81) ff. 163 GK 2.32: na nirodho na cotpattir na baddho na ca sādhakaḥ / na mumukṣur na vai mukta ity eṣā paramārthatā // 164 Śaṅkara on GK 2.32: sato hy utpattiḥ pralayo vā syān nāsataḥ śaśaviṣāṇādeḥ. 165 GK 1.6ab: prabhavaḥ sarvabhāvānāṃ satām iti viniścayaḥ. Bhattacharya’s edition (1943: 2) has sarvabhūtānāṃ in place of sarvabhāvānāṃ; Śaṅkara seems to comment on the version with sarvabhāvānāṃ. 166 See also King, 1995: 22 ff.; 1995a: 323 ff. 167 See also Vetter, 1979: 58 ff. 51 52 cannot arise from there, just as oil [cannot arise] from grains of sand.”168 The Bhāṣya on Bṛhadāraṇyaka Upaniṣad 1.2.1 features a lengthy discussion in support of the existence of the effect before its arising. Another discussion of the same kind is found in the commentary on Chāndogya Upaniṣad 6.2.2.169 Let us examine a proof of satkāryavāda found in Śaṅkara’s commentary on Brahma Sūtra 2.1.18. We shall see that he expresses the correspondence principle in a form quite similar to the one we proposed. Here is the proof in question: If the effect does not exist before it arises, the arising would lack an agent and would be void. Arising, indeed, is an action, and that requires an agent, just like the action of going, etc. It would be contradictory to maintain that there is indeed an action and that it lacks an agent. [Otherwise] one could imagine that the arising of a pot, [though] spoken of, did not have the pot for its agent, but something else. … If that were the case, one would say, “The potter, etc., which are causes, arise,” instead of saying, “The pot arises.” In the world, when one says, “The pot arises,” one does not perceive the potter, etc., as also arising, because one perceives them as already 170 arisen. One could not ask for more. In the works of Śaṅkara we find not only an awareness of the linguistic basis of the arguments being used—which we shall study later— but also an analysis that fully coincides with our correspondence principle. Still commenting on sūtra 2.1.18, Śaṅkara makes another remark that points in the same direction: “For one who believes that the effect does not exist before its arising, the action of the actants (kāraka) would have no object; because an absence cannot be an object.”171 Śaṅkara gives no concrete example, but his remark doubtless refers to sentences such as “Devadatta makes a pot.” Devadatta’s action would have no object if the pot did not exist before its arising. Let us return to the Āgamaśāstravivaraṇa. Can an existent thing arise or not? If we accept the identity of the author of this text with the author of the Brahma Sūtra Bhāṣya, etc., nothing prevents us from thinking that Śaṅkara, even while writing the Āgamaśāstravivaraṇa, still considered arising to be possible, but only for something existent. The other passages, which hold that arising is impossible, always occur in the context of verses upholding this position. On this view, Śaṅkara did not wish to contradict the text on which he was commenting, though for his own part he disagreed with its position. Let us note here that both of the positions found in the Āgamaśāstravivaraṇa are based on the correspondence principle. Viewed in this light, the two positions, though opposed, may be compared to two enemies who face a more dangerous common enemy. The real enemy of both positions would be the rejection, in whole 168 Śaṅkara on Brahma Sūtra 2.1.16: yac ca yadātmanā yatra na vartate na tat tata utpadyate, yathā sikatābhyas tailam. 169 Cf. Alston, 1980: 95 ff. 170 Śaṅkara on Brahma Sūtra 2.1.18: prāg utpatteś ca kāryasyāsattve utpattir akartṛkā nirātmikā ca syāt / utpattiś ca nāma kriyā, sā sakartṛkaiva bhavitum arhati gatyādivat / kriyā ca nāma syād akartṛkā ceti vipratiṣidhyeta / ghaṭasya cotpattir ucyamānā na ghaṭakartṛkā, kiṃ tarhy anyakartṛkā iti kalpyā syāt / … / tathā ca sati ghaṭa utpadyate ity ukte kulālādīni kāraṇāni utpadyante ity uktaṃ syāt / na ca loke ghaṭotpattir ity ukte kulālādīnām apy utpadyamānatā pratīyate / utpannatāpratīteḥ / 171 Śaṅkara on Brahma Sūtra 2.1.18: yasya punaḥ prāg utpatter asat kāryaṃ tasya nirviṣayaḥ kārakavyāpāraḥ syāt / abhāvasya viṣayatvānupapatteḥ / 52 53 or in part, of the correspondence principle. Later we shall be interested in thinkers who acquiesced to this rejection. 7. Kashmiri Śaivism An interesting variant of satkāryavāda is found in certain texts of Kashmiri Śaivism, where problems relating to the arising of things are used to prove the existence of God, the creative consciousness of the universe. We shall take as our starting point the Īśvarapratyabhijñākārikā of Utpaladeva with the Vṛtti of the same author, a text admirably edited, translated, and commented on by the Italian scholar Raffaele Torella. The fourth Āhnika of the second Adhikāra (Kriyādhikāra) of this text is of particular interest for our purposes. The Vṛtti on the second kārikā begins with the observation that insentient things such as pradhāna (of the Sāṃkhyas), atoms, seeds, etc., do not have the power to produce something that does not exist.172 In other words, the seed cannot produce the sprout. For that, a conscious actor is required, as when a potter makes a pot. Utpaladeva concludes that to be a cause means to be an agent, and that to be an effect means to be a grammatical object.173 He then raises the central question: how can something non-existent become existent? Kārikās 3 and 4 give the following response (as translated by Torella): What is non-existent is non-existent and that is that. One cannot predicate the existence of what is non-existent, nor, on the other hand, would the existent gain any advantage from obtaining existence [which it already possesses]. If one then points out that in practical reality there is general consensus on the existence of the relation of cause and effect, [then one replies that] it consists in the fact that a thing, that is internally present (antarviparivartinaḥ), becomes, by virtue of the power of that indefinable Being (tasya kasyāpi), the object of knowledge for both 174 senses. The Vṛtti explains: Attributing the nature of existent to what does not exist is contradictory, and it is already established in what exists. [The relation of cause and effect consists in this:] a thing, already present within [the I], is “created” by the Lord, or in other words, is caused by him to become the object of knowledge for the internal and external 175 senses. It follows that to be an effect means for an object to be made external. The state of being external and the state of being an effect are such with respect to a single reality; indeed, it is with respect to the knowing subject that we speak of internal 172 Torella, 1994: 55: jaḍaṃ pradhānaparamāṇubījādi tu na śaktam asato nirmāṇe. Id., ibid.: kartṛtvam eva hi kāraṇatvaṃ karmataiva ca kāryatvaṃ na tv anyat. 174 Id., ibid.: yad asat tad asad yuktā nāsataḥ satsvarūpatā / sato ‘pi na punaḥ sattālābhenārtho ‘tha cocyate //3// kāryakāraṇatā loke sāntarviparivartinaḥ / ubhayendriyavedyatvaṃ tasya kasyāpi śaktitaḥ //4// (tr. pp. 175-76) 175 Id., ibid.: asataḥ satsvabhāvatā viruddhā sataś ca siddhā / siddhasyaivāntarbāhyāntaḥkaraṇadvayīvedyatāpādanam īśvareṇotpādanam // (tr. p. 176) 173 53 54 and external.176 Consequently, the cause is the knowing subject and none other; this subject is one, although the internal and external manifestation of the effect occur in succession.177 This is why God, and not the seed, is the cause, the only cause, of the sprout.178 And even in the case of the potter, who is of course a knowing subject, the production of a pot depends on specific operations performed on clay, etc., in accordance with the restrictive order known as “necessity” (niyati), which is created by God.179 Later in the same chapter Utpaladeva explains: “[A]n insentient reality cannot even be the agent of the action of being—‘it exists, is’—since it does not possess the freedom that is manifested through ‘wanting to be’ (bubhūṣāyogena). Thus the ultimate truth in this regard is that the knowing subject, and he alone, ‘causes’ the insentient reality ‘to be’ (bhāvayati), or, in other words, appears in various forms such as mount Himācala and so on.”180 Thus, even in the case of the verb “to exist, to be,” the subject, or the agent of the action in question, is God, the supreme “knower.” In short, the sprout exists already before it arises, not in its cause (the seed), but within the creative consciousness of God. And the pot exists in the same form before it is made by the potter; moreover, the true agent of this action is not the potter but God, who is likewise the true agent of every other action, including “coming to be.” As Utpaladeva says in kārikā 21: “Therefore causality, agency, action are nothing but the will of him who wishes to appear in the form of the universe, in the various manifestations of jar, cloth, and so on.”181 In the Vṛtti he expresses his agreement with the position of Śaṅkara, according to which an action without an agent cannot exist; he adds that we attribute action to other actors metaphorically, through the agent.182 8. Jainism So far we have been interested in early schools that accepted the existence, in some manner, of the effect in its cause. However strange this position might seem, it offered an effective solution to the problem of arising, which many thinkers of the period found challenging and indeed threatening. Another consequence of adopting the position: one could adhere without further difficulty to the correspondence principle, a principle many philosophers apparently considered plausible or even obvious. We shall see with what difficulty certain thinkers 176 Id., p. 56 (Vṛtti on kārikā 6): arthasya bāhyatāpādanaṃ kāryatvam, tato bāhyatā kāryatā caikāpekṣayā pramātāram apekṣya cāntarbahirvyavahāraḥ / 177 Id., p. 56 (Vṛtti on kārikā 7): evaṃ pramātaiva kāraṇaṃ sa ca bāhyāntarakāryābhāsakrame ‘py eka ev[a] … 178 Id., p. 57 (kārikā 8): ata evāṅkure ‘pīṣṭo nimittaṃ parameśvaraḥ / tadanyasyāpi bījāder hetutā nopapadyate //8// 179 Id., p. 57 (Vṛtti on kārikā 9): kumbhakārarūpe pramātari kāraṇe sthite ‘pi mṛdādisaṃskārāpekṣā ghaṭasyeśvarakṛtaniyatisaṃjñamaryādayā na svabhāvena. 180 Id., p. 61 (Vṛtti on kārikā 20): jaḍasyāpy asti bhavatīty asyām api sattākriyāyāṃ bubhūṣāyogena svātantryābhāvād akartṛtvam, tena pramātaiva taṃ bhāvayati tena tena vā himācalādinā rūpeṇa sa bhavatīty atra paramārthaḥ. (tr. p. 187) 181 Id., p. 61 (kārikā 21): itthaṃ tathā ghaṭapaṭādyābhāsajagadātmanā / tiṣṭhāsor evam icchaiva hetutā kartṛtā kriyā // (tr. p. 187) 182 Id., p. 61 (Vṛtti on kārikā 21): nākartṛkaṃ karmāsti karmādīnāṃ kartṛmukhenopacārataḥ. 54 55 gradually distanced themselves from this principle, prompted by the need to defend their positions. Before studying their solutions to the problem of the arising of things, we shall examine some passages from the literature of Jainism that seem to address the same problem. The situation they present is not entirely clear, however, and I can offer only a tentative interpretation. Professor Granoff of Yale University has drawn my attention to a story told in several places throughout Jaina literature. It is the story of Jamāli, who is introduced as a student of Mahāvīra, but who begins to preach a different message from that of his master. The Viyāhapannatti, a canonical text of the Śvetāmbaras better known as the Bhagavatī, contains the oldest version of Jamāli’s message available to us.183 The message, which follows, is rather enigmatic: “Mahāvīra claims that what is moving has moved; what is being ascended, has been ascended; what is being known, has been known; what is being lost, has been lost; what is being cut, has been cut; what is being divided, has been divided; what is being burned, has been burned; what is dying, is dead; what is being exhausted, has been exhausted. This is false.”184 In each case, a present participle is juxtaposed with a past participle. The position here attributed to Mahāvīra is indeed placed in his mouth elsewhere in the Viyāhapannatti,185 perhaps with the meaning that the process of the destruction of karman is, practically speaking, finished as soon as it is set in motion.186 We cannot be certain, however, that Jamāli’s protest is directed against this relatively harmless idea, since he goes on to refer to a bed his co-religionists are in the process of making: “It is obvious that the bed is being made, but that it has not been made.”187 Jamāli evidently takes Mahāvīra’s remark literally, and he extends it to include all juxtapositions of a present participle with a past participle, using two verbs Mahāvīra had not used.188 The verbs he uses are “to make” and “to spread,” which, though apt descriptions of bed-making, are ill suited to describe the 183 Deleu (1970: 41 n. 43), however, expresses the following opinion: “Probably the Jamāli episode originally belonged to Antag[aḍadasāo] 6 (cf. Ṭhāṇa 505a) and was inserted in the Viy[āhapannatti] ….” On the schism of Jamāli, see also Leumann, 1885: 98 ff. 184 Viy 9.33.228 (p. 458) in the Ladnun edition; 9.33.96 (p. 477) in the Bombay edition; vol. 4 p. 102 in Lalwani’s translation. As the Ladnun edition, in contrast to the other two editions, gives the text in a more complete form, I cite it here: jaṇ ṇaṃ samaṇe bhagavaṃ mahāvīre evam āikkhaï, [evaṃ bhāsaï, evaṃ paṇṇaveï,] evaṃ parūeï—evaṃ khalu calamāṇe calie, udīrijjamāṇe udīrie, vedijjamāṇe vedie, pahijjamāṇe pahīṇe, chijjamāṇe chiṇṇe, bhijjamāṇe bhiṇṇe, dajjhamāṇe daḍḍhe, mijjamāṇe mae, nijjarijjamāṇe nijiṇṇe, taṇ ṇaṃ micchā. The part in brackets was added on the basis of Viy 1.9.420 (Ladnun). For a summary of the story of Jamāli, see Deleu, 1970: 163 ff. 185 E.g. Viy 1.1.11 (p. 5)(Ladnun); Viy 1.1.5 (p. 3)(Bombay); vol. 1 p. 5 in Lalwani’s translation. Deleu (1970: 73) summarizes: “The action that is being performed equals the completed action. … This is true [e.g.] with such actions as moving, coming forth, becoming perceptible, decreasing, cutting, breaking, burning, killing and annihilating.” 186 See notes 19 and 20 in Lalwani’s translation, vol. 1 pp. 226-28. Deleu (1970: 73), however, remarks: “Abhay[adeva’s Vṛtti] indeed explains the different words as technical terms applying to the course of karman. It should be noted, though, that some of his equations are rather improbable … and that in other places … still other words are used to illustrate rather than formulate the tenet.” 187 Viy 9.33.228 (p. 458)(Ladnun); 9.33.96 (p. 477)(Bombay); vol. 4 p. 102 in Lalwani’s translation: imaṃ ca ṇaṃ paccakkham eva dīsaï sejjāsaṃthārae kajjamāṇe akaḍe, saṃtharijjamāṇe asaṃtharie. 188 Note, as Deleu remarks (1970: 73, cited above; also p. 25), that the same doctrine is illustrated elsewhere in the Viyāhapannatti with the help of other verbs; e.g. Viy 8.7.280 (p. 361)(Ladnun), 8.7.11 (p. 368)(Bombay), vol. 3 p. 209 (Lalwani): dijjamāṇe dinne, paḍigahejjamāṇe paḍiggahie, nisirijjamāṇe nisaṭṭhe; 8.7.291 (p. 363)(Ladnun), 9.7.23 (p. 370)(Bombay), vol. 3 p. 212 (Lalwani): gammamāṇe gae, vītikkamijjamāṇe vītikkaṃte; and elsewhere. 55 56 destruction of karman. More broadly speaking, Jamāli’s protest seems to be founded on the impossibility, in his eyes, of the co-existence of the present and the past. As a result, that which is in the process of being made is not yet made. Thus there is no pot at the moment when the pot is being made. This conclusion seems confirmed by the rest of the story. Jamāli meets his former master, Mahāvīra, along with the latter’s disciple Gautama. Gautama asks him the following questions: “Is the word eternal or non-eternal? Is the soul eternal or non-eternal?”189 Jamāli being unable to answer, Mahāvīra explains: “Jamāli! The world is eternal. Never was it not, never is it not, never will it not be. It was, it is, and it will be. It is stable, perpetual, eternal, indestructible, imperishable, lasting. Jamāli! The world is non-eternal. Having declined, it recovers, and vice versa.”190 Similar observations are made for the soul. What is the connection between Jamāli’s ideas and Mahāvīra’s brief sermon? If there is one—which is not certain, of course—it is doubtless that Mahāvīra’s sermon addresses the problem raised by Jamāli. This problem seems to concern the existence of the past and the present. If one takes literally the claim that what is being made is identical to what has been made, it seems necessary to conclude that nothing changes. The world, seen in this light, is without change. Mahāvīra’s sermon confirms that the world is stable, perpetual, eternal, indestructible, imperishable, lasting. It is true that this represents only one side of reality: the world is also non-eternal, as we have seen. Nevertheless, if, in some manner, the world is eternal, one could say the same of the pot that is to arise: it is already there, and its arising does not pose a problem. The story of Jamāli raises some very important questions. It seems entirely plausible that certain remarks of Mahāvīra (or remarks attributed to him), which initially had no philosophical bearing, were subsequently interpreted in a new, philosophical way. The position that the world does not change over time, a position accompanied, in Jaina fashion, by the alternative position that the world is in a constant state of flux, came to be attributed to Mahāvīra. If our interpretation is correct, the episode of Jamāli reminds us that the idea of a static world had some difficulty taking root in Jainism. Subsequent discussions in Jaina literature confirm our analysis of Jamāli’s ideas. The Viśeṣāvaśyaka Bhāṣya of Jinabhadra (6th century C.E.) discusses the problem, situating it precisely in the context of knowing whether a thing arises from something existent or non-existent. Here are a few excerpts from the commentary of Koṭyārya, who, while adding certain details, does not deviate from Jinabhadra’s words: Jamāli, suffering from a bout of fever, told his students to prepare a bed. Seeing that what he had ordered was not accomplished, he became angry, [saying:] “The sacred word that says ‘what is being made has been made’ is false, because it contradicts perception, like the word of someone who does not have total knowledge (?). … This 189 Viy 9.33.231 (p. 459)(Ladnun); Viy 9.33.99 (p. 478)(Bombay); tr. Lalwani vol. 4 p. 104: sāsae loe jamālī? asāsae loe jamālī? sāsae jīve jamālī? asāsae jīve jamālī? These are two of the so-called avyākṛta questions, also known to Buddhism; cf. Matilal, 1981: 19 ff. 190 Viy 9.33.233 (p. 459)(Ladnun); Viy 9.33.101 (p. 479)(Bombay); tr. Lalwani vol. 4 pp. 104-05: sāsae loe jamālī! jaṃ na kayāi nāsi, na kayāi na bhavaï, na kayāi na bhavissaï. bhuviṃ ca, bhavaï ya, bhavissaï ya. dhuve, nitie, sāsae, akkhae, avvae, avaṭṭhie nicce. asāsae loe jamālī! jaṃ osappiṇī bhavittā ussappiṇī bhavaï, ussappiṇī bhavittā osappiṇī bhavaï. 56 57 bed is visibly being made, in accordance with [my] instruction that the blanket be spread out; it has not been made at this [very] moment. From the viewpoint of the one oriented toward the action of spreading the thing, [the bed] is being made, it has not been made. Therefore, only the condition of being made, not the condition of having been made, is a visible property of the object being made, because the object has not been finished. The property of having been made is thus ruled out by the property of being in the process of being made, which is established by perception. Hence the opposition to perception [which I mentioned a moment ago]. For this reason, every object that is in the process of being made, without exception, has not been made; for us, [something] has been made when the action of making has been completed, not immediately.” … Jamāli first states his own point of view: something that has been made is not being made—this is the claim; because the thing in question is present, like an old pot. If, however, one were to maintain that what has been made is also being made, then the state of being in the process of being made would always prevail, because [the object in question] would be in the process of being made, like something made in its first moment. And the action would never be completed, because [the object] would always be in the process of being made, as at the first moment. … And the action would be fruitless, because all objects would [already] be made, like a finished pot. Therefore, 191 something not made and non-existent is being made. Jamāli is refuted in the following passage: 192 The ancients are those who attained śruta-jñāna. Their point of view [is the following:] A thing that has not been made is not being made, because it does not exist, like a flower in the sky [which does not exist either, and therefore cannot be made]. If one maintains its arising [in the following manner:] “what was previously non-existent and not made is being made,” then one comes to an undesirable [consequence]: the horn of a donkey could also be in the process of being made; it could be in the process of being made because it has not been made, like the pot admitted by you. As for the net of faults you hurl against [the idea] of making an existent thing, all that applies equally to [the idea] of making a thing that does not exist. There, too, all the faults [would still apply]; or rather, the faults—such as total incoherence, etc.—would be all the more serious if what is in the process of being made were not there, did not exist. Or consider [the following inference:] “The horn of a donkey is being made, because it did not exist previously, like [any] 193 product admitted [by you].” 191 Jinabhadra, Viśeṣāvaśyaka Bhāṣya, vol. II, p. 538 ff. (on verses 2789-93): jamālir dāhajvarābhibhūtaḥ ‘saṃstārakaṃ kuruta’ ity ādiśya śiṣyān vāksamakam aniṣpannaṃ dṛṣṭvā ruṣitaḥ siddhāntavacanaṃ ‘kriyamāṇaṃ kṛtam’ ity etad vitatham, pratyakṣaviruddhatvāt aśrāvaṇaśabdavacanavat / … saṃstārako ‘yaṃ pratyakṣaṃ kriyamāṇaś ca kambalaprastaraṇavyāpārādeśāt na cāsmin samaye kṛtaḥ / punar api vastuprastaraṇasāpekṣāt kriyamāṇa eva, na kṛtaḥ, tasmāt kriyamāṇasya dharmiṇaḥ kriyamāṇatvam eva pratyakṣam idam, na kṛtatvam, aniṣpannatvāt / tataḥ kriyamāṇatvena pratyakṣasiddhena kṛtatvaṃ dharmo ‘panīyate iti pratyakṣaviruddhatvam / tasmāt sarvam eva vastu kriyamāṇaṃ na kṛtam eva, kriyāparisamāptau naḥ kṛtam, nā’’rāt / … svamataṃ tāvaj jamālir darśayati: kṛtaṃ vastu na kriyamāṇam iti pratijñā, vidyamānatvāt, cirantanaghaṭavat / atha kṛtam api kriyamāṇa[m a]bhyupagamyate kenacit tataḥ sarvadā kriyamāṇāvasthaiva bhavatu, kriyamāṇatvāt, prathamasamayakṛtavat / na ca kriyāparisamāptiḥ, sarvadā kriyamāṇatvāt, ādisamayavat / … viphalā ca kriyā, sarvavastūnāṃ kṛtatvān niṣpannaghaṭavat / tasmād akṛtam avidyamānaṃ ca kriyate / 192 For a discussion of śruta-jñāna, see Tatia, 1951: 48 ff. 193 Jinabhadra, Viśeṣāvaśyaka Bhāṣya, vol. II, p. 539 (on verses 2795-96): sthavirāḥ śrutajñānāptāḥ / teṣāṃ matam: nākṛtaṃ kriyate, abhāvatvāt, khapuṣpavat / atha tasya janmābhyupagamaḥ: pūrvam abhūtam akṛtam eva kriyate / tato ‘niṣṭāpādanam: kharaviṣāṇam api kriyatām, kriyamāṇaṃ bhavatu, akṛtatvāt, tvadiṣṭaghaṭavat / yac ca tvayā doṣajālam upakṣipyate vidyamānasya karaṇe, tat sarvam avidyamānakaraṇe ‘pi tadavastham: sarve tatrāpi doṣāḥ, asati avidyamāne kriyamāṇe kaṣṭatarā vā doṣā bhaveyuḥ atyantāsambaddhatvādayaḥ / dṛśyatāṃ vā: kriyate kharaviṣāṇaṃ pūrvam abhūtatvād iṣṭakāryavat / 57 58 The position that a thing that has not been made is not in the process of being made is supported by demonstrating that a thing is in the process of being made only in the final moment of its arising, that is, at the moment when it has been made: The extended period [required] to make [something], which you cited as support for the arising of something non-existent, [cannot serve as such]. The making of a pot does not take up an extended period, because this extended period belongs to something else, not to the pot. … In [the course of making a pot] there arise [successively] at each moment the lump [of clay], [the stages of production known as] śivaka, sthāsaka, kuśūla, etc., which are many and are mutually distinct. If the period it takes to make [a pot] is extended insofar as these stages are many, what does this mean for the pot? At that moment [i.e., during the arising of these stages] the pot has certainly not arisen. … During the arising of one thing [namely, the stages prerequisite to the arising of the pot], a different thing (the pot) does not come into view, because it is different, like the pot during the arising of a cloth. If [one asks] how [the stages prerequisite to the arising of a pot] are different [from the pot, the answer is:] [the stages] śivaka, etc., are different from the pot, because they are mutually distinct, like a cloth. Hence, during the arising of a śivaka, etc., in the state śivaka, etc., how would the pot come into view? It is therefore arisen if the end [of the process] comes into view, [i.e.,] at the moment of its own arising. What fault results from this [description of the situation]? [Obviously none!] The purpose [of this demonstration is to show that the arising of a pot does not take] an extended period. Therefore, that which is being made only at the moment of its 194 arising has been made in the very same, present moment. Conclusion: at the moment when the sentence “the pot arises” is true, the pot has arisen. We are dealing, of course, with but a single moment, the final moment in a series of actions culminating in the arising of the pot. But the correspondence principle has been preserved unharmed, and in a manner that differs from everything we have seen in other currents of thought. Elsewhere in his Viśeṣāvaśyaka Bhāṣya, Jinabhadra explicitly criticizes the positions of the Mādhyamikas.195 A particularly interesting case is verse 2149, where the auto-commentary quotes a stanza from the Mūlamadhyamakakārikā.196 This verse of the Viśeṣāvaśyaka Bhāṣya, along with the auto-commentary, presents Nāgārjuna’s position on the problem that concerns us at the moment, the problem of the arising of things. Here is an excerpt from the auto-commentary: What is arisen does not arise, because it is [already] there, like a pot. If, on the other hand, what is arisen were also to arise, there would be an infinite 194 Jinabhadra, Viśeṣāvaśyaka Bhāṣya, vol. II, p. 539 ff. (on verses 2796-99): yad abhūtaprādurbhāve bhavatopapattir ucyate kriyākāladrāghīyastvam, tan naivāsti dīrghakālakaraṇaṃ ghaṭasya, yasmād anyadīya evāsau dīrghakālo na ghaṭasyeti / … iha pratisamayaṃ piṇḍa-sthāsaka-kuśūlādaya utpadyante parasparavilakṣaṇā bahavaś ca / teṣāṃ bahutvād yadi kriyākālo dīrgho bhavati, tataḥ kim āyātaṃ kumbhasya? tadāsau naivārabdha iti / … anyārambhe anyan na dṛśyate, anyatvāt, paṭārambha iva ghaṭaḥ / katham anyatvam iti cet? śivakādīnāṃ kumbhasya cānyatvam, parasparavilakṣaṇatvāt, paṭavat / tasmāc chivakādyavasthāyāṃ śivakādyārambhe katham iva ghaṭo dṛśyata iti? ata evāsau ārabdho yady anta eva dṛśyate svārambhakāle / tataḥ ko doṣaḥ? dīrghakālatvābhāva ity arthaḥ / tasmād ārambhakāla eva kriyamāṇaṃ tasminn eva ca vartamāne sampratikāle kṛtaṃ tad bhavati / 195 See Butzenberger, 1994. 196 The stanza quoted is MadhK(deJ) 2.1: gataṃ na gamyate tāvad agataṃ naiva gamyate / gatāgatavinirmuktaṃ gamyamānaṃ na gamyate //. The Viśeṣāvaśyaka Bhāṣya (in Malvania’s edition) gives the latter part of the stanza as gamyamānaṃ tu gamyate. 58 59 regress. What is unarisen does not [arise] either, because it does not exist, like the horn of a donkey. And if [one were to maintain that] what is unarisen arises, the action of arising would belong to a non-entity, like the 197 horn of a donkey or something similar. Jinabhadra further adds—still in the voice of the Mādhyamika opponent—that what arises cannot be arisen and unarisen at the same time, for in that case the problems bound up with both positions would accrue.198 Jinabhadra next criticizes the Mādhyamika position as it is expressed here and in verses 2180-86, where he also presents his own position. Let us focus on a few essential passages from his commentary on verses 2183-84: In this world there are things that arise being [already] arisen, and others [that arise] being unarisen; there are those [that arise] being arisen and unarisen at the same time, and still others [that arise] while arising; [finally,] there are things that do not arise at all, in accordance with what one wishes to express (vivakṣātaḥ). … For example, a pot arises in this world being [already] arisen in the form of clay, etc., because it is made of it. This same [pot] arises being unarisen with respect to its specific form, because this [form] did not exist previously. [The pot] arises being arisen and unarisen at the same time with respect to its color, etc., and its specific form, because it is not different from these things. It arises while arising because an action can take place only in the present moment, [given that] a [real] action is not possible by reason of the fact that the past has vanished and the future is not [yet] 199 present. Jinabhadra’s relativism, which is characteristic of Jainism in general, is easily recognizable here. The last option presented—the pot arises while arising—is, as we know, Jinabhadra’s response to the dilemma posed by Nāgārjuna. Let us return for a moment to Jamāli, Jainism’s first heretic. The later commentator Śīlāṅka, in his commentary on the Sūtrakṛtāṅga, outlines the same position as Jinabhadra concerning Jamāli’s error. He observes that what is in the process of being made has been made, in accordance with the vyavahāranaya (“the practical point of view”).200 Those who think like Jamāli, he continues, do not know that people use the word “pot” to refer metaphorically to actions, beginning with the action of digging to find clay. Since, in reality, the time of the actions that give rise to the pot and the time of the establishment of the pot coincide and therefore constitute one and the same time, what is in the process of being made has been made. Worldly convention, according to Śīlāṅka, confirms this position.201 197 Jinabhadra, Viśeṣāvaśyaka Bhāṣya, vol. II, p. 378 (on verse 2149): … na tāvaj jātaṃ jāyate, vidyamānatvāt ghaṭavat / atha jātam api jāyate ‘navasthāprasaṅgaḥ / tathā nājātam, avidyamānatvāt kharaviṣāṇavat / athājātam api jāyate kharaviṣāṇādyabhāvajātikriyāprasaṅgaḥ / 198 Jinabhadra, Viśeṣāvaśyaka Bhāṣya, vol. II, p. 378 (on verse 2149): na jātājātam ubhayadoṣāt. 199 Jinabhadra, Viśeṣāvaśyaka Bhāṣya, vol. II, p. 385 (on verses 2183-84): iha kiñcit jātaṃ jāyate, kiñcid ajātam, kiñcij jātājātam, kiñcij jāyamānam, kiñcit sarvathā na jāyate vivakṣātaḥ / … / yatheha ghaṭo mṛdrūpādibhir jāta eva jāyate, tanmayatvāt / sa evākāraviśeṣeṇājāto jāyate, prāgabhāvāt / rūpādibhir ākāraviśeṣeṇa ca [jātā]jāto jāyate, tebhyo ‘narthāntaratvāt / atītānāgatakālayor vinaṣṭānutpannatvāt kriyānupapattir vartamānamātrasamaya eva kriyāsadbhāvāj jāyamāno jāyate / 200 For a discussion of the vyavahāranaya, see Matilal, 1981: 44; see also Tattvārthasūtra 1.34 with the commentary of Siddhasenagaṇi, which corresponds to sūtra 1.33 with the commentary of Akalaṅkadeva (tr. Tatia, 1994: 23 f.), and below. 201 Śīlāṅka (p. 231): vyavahāranayābhiprāyeṇa kriyamāṇam api kṛtaṃ bhavati / … na cāsau jānāti varāko yathā ayaṃ loko ghaṭārthāḥ kriyā mṛtkhananādyā ghaṭa evopacarati, (tattvataḥ) tāsāṃ ca kriyāṇāṃ 59 60 The problem of the arising of things occupied Jaina thinkers in other contexts, too, apart from the story of Jamāli. We shall mention only a few examples here. Kundakunda, the celebrated author of the Pravacanasāra, puts it thus: “There is no origin without destruction, nor destruction without origin. Arising as well as destruction are not [possible] without something that endures.”202 In other words, it is necessary for there to be something in order for this thing to arise. His commentator Amṛtacandra (ca. 1000) rebels against the idea that there could be arising without something that endures. Were this the case, he observes, there would either be the absence of arising pure and simple (here one suspects the influence of Nāgārjuna; Amṛtacandra rejects this option without arguments) or the arising of something non-existent. In the second case, flowers in the sky and other impossible things would arise. In the case of the arising of a pot, he explains, there is both the disappearance of the lump of clay and the continued existence of the clay itself.203 Jaina thinkers tend to confront problems with the help of different “points of view” (naya), which allow them to describe a single situation in a number of different ways. (We have already encountered the vyavahāranaya, “the practical point of view,” in our discussion of Śīlāṅka’s position.) These different points of view, all of which are valid at the same time, prove to be quite useful in this situation. Siddhasena Divākara distinguishes essentially two points of view in his Saṃmatitarkaprakaraṇa (ca. 700), calling them dravyāstika (“the substance exists”) and paryāyāstika (“the modification exists”).204 The consequence of this distinction is stated in verse 1.11: “Things arise and pass away by constraint of the point of view of modifications; everything remains ever without origin and without destruction [by constraint] of the point of view of substance.”205 Though Siddhasena does not provide further details, it is clear that his two “points of view” allow him to benefit from maintaining, with the Sāṃkhyas, that an effect must exist before it arises, without thereby forsaking the viewpoint of common sense. Let us return once more to the story of Jamāli. While the Viyāhapannatti is not very clear as far as details are concerned, the position of Jinabhadra, later taken up by Śīlāṅka, presents a solution to the problem of the arising of things that is, in one sense, precisely the opposite of the solutions we have studied so far. For the Sarvāstivādins as well as for the Sāṃkhyas, a thing can arise because it is already kriyākālaniṣṭḥākālayor ekakālatvāt kriyamāṇam eva kṛtaṃ bhavati, dṛśyate cāyaṃ vyavahāro loke, tad yathā: adyaiva devadatte nirgate kānyakubjaṃ devadatto gata iti vyapadeśaḥ, (lokoktyā) tathā dāruṇi chidyamāne prasthako ‘yaṃ (iti) vyapadeśa ityādi /. The expression kriyākālaniṣṭḥākālayor ekakālatvāt is similar to the vākya of the Vaiśeṣika Kaṭandī, cited by Mallavādin: niṣṭhāsambandhayor ekakālatvāt (Bronkhorst, 1993: 145). Is there a connection? 202 Pravacanasāra 2.8: ṇa bhavo bhaṃgavihīṇo bhaṃgo vā atthi saṃbhavavihīṇo / uppādo vi ya bhaṃgo ṇa viṇā dhovveṇa attheṇa //. Cf. Matilal, 1981: 38. 203 Pravacanasāra pp. 124-25: … / yaiva ca mṛttikāyāḥ sthitis tāv eva kumbhapiṇḍayoḥ sargasaṃhārau … / yadi punar nedam evam iṣyeta tadānyaḥ sargo ‘nyaḥ saṃhāraḥ anyā sthitir ity āyāti / tathā sati hi kevalaṃ sargaṃ mṛgayamāṇasya kumbhasyotpādanakāraṇābhāvād abhavanir eva bhavet, asadutpāda eva vā / tatra kumbhasyābhavanau sarveṣām eva bhāvānām abhavanir eva bhavet / asadutpāde vā vyomaprasavādīnām apy utpādaḥ syāt / … / tatra mṛttikāyā asthānau sarveṣām eva bhāvānām asthānir eva bhavet. Cf. Matilal, 1981: 38. 204 Saṃmatitarkaprakaraṇa 1.3 (p. 271) with the commentary of Abhayadeva. 205 Saṃmatitarkaprakaraṇa 1.11 (p. 409): uppajjaṃti viyaṃti ya bhāvā niyameṇa pajjavaṇayassa / davvaṭṭhiyassa savvaṃ sayā aṇuppannam aviṇaṭṭḥaṃ // 60 61 there, and has always been there. Jinabhadra agrees that a thing can arise because it is there, but maintains that it is only there in its final moment, the one moment when the sentence “the pot arises” or “he makes a pot” is strictly true. We shall encounter another school of thought that likewise believes a pot can arise because it is there, but that situates this pre-existence between the two extremes just mentioned. For this other school, the pot does not exist from all eternity before its arising, nor does it exist only at the final moment preceding its arising; it exists during the period required to produce it. We shall return to this position momentarily. Let us conclude with a question. The brief sermon that Mahāvīra addresses to Jamāli in the Viyāhapannatti figures among the earliest sources of what is known as anekāntavāda, the doctrine according to which reality is manifold. One thus cannot avoid asking whether the doctrine was developed in response to the problem of the arising of things. This is the view of B. K. Matilal,206 and while we cannot go into an in-depth study of the question here, the view is certainly plausible. If it is correct, then anekāntavāda—which Matilal calls “the central philosophy of Jainism”—would turn out to be another doctrine inspired by the correspondence principle.207 9. Early Vaiśeṣika Satkāryavāda was almost a natural response to the challenge posed by Nāgārjuna’s arguments, or rather by the correspondence principle. Other schools of thought, however, accepted the opposite position of asatkāryavāda, “the doctrine that the effect does not [pre-]exist [in the cause].” Such was the case with the Vaiśeṣika school in particular. The doctrine is expressed already in the opening sūtras of the ninth chapter of the Vaiśeṣika Sūtra (at least according to its commentaries), and continued with the school throughout its later developments. How did early Vaiśeṣika respond to the question of how a pot arises? Later we shall see what subsequent thinkers of the school made of the problem. As for the early period, Mallavādin’s Dvādaśāranayacakra and its commentary, the Nyāyāgamānusāriṇī of Siṃhasūri, are valuable sources of information. In them one finds passages in which the idea that things exist in a certain manner before their arising is attributed to Vaiśeṣika as well. The authors of these two Jaina texts had access to several Vaiśeṣika texts, including especially, it would seem, the Kaṭandī of Rāvaṇa and its Ṭīkā by Praśasta, better known as Praśastapāda, author of the Padārthadharmasaṅgraha or Praśastapādabhāṣya.208 Unfortunately, we do not have the Kaṭandī and its Ṭīkā, but the critique in the two Jaina works allows us to identify several Vaiśeṣika positions such as they must have appeared in these texts. One learns, for instance, that arisen things are spoken of as “existent” by reason of a connection with the universal “existence” (sattāsambandha), which 206 Matilal, 1981: 26 ff. (ch. VI: “Anekānta as a Resolution of the Paradox of Causality”). See Bronkhorst, 2003. 208 Bronkhorst, 1993. 207 61 62 occurs at the moment of, or immediately after, their completion;209 this connection with the universal “existence” is the cause of the name and the concept of a thing.210 Prior to this connection with the universal “existence,” are things completely nonexistent? According to Mallavādin, the Vaiśeṣikas say no. Things exist in a certain manner even prior to their arising. Although they have no connection with the universal “existence” (sattāsambandha) at that time, they do have an essence (astitva, svabhāva, svabhāvasattā), in virtue of which they arise. Even in the absence of a connection with the universal “existence,” a substance (or a quality, or a motion) thus has an identity. And the expression asat, which normally means “nonexistent,” can be taken as a bahuvrīhi compound meaning “that which does not possess existence (i.e., that which does not possess the universal ‘existence’).” In other words, an object that is asat is nonetheless not totally non-existent: even in the absence of a connection with the universal “existence,” it still has an essence.211 The main debate is found in the seventh chapter (lit. “spoke,” ara) of the Dvādaśāranayacakra. The Vaiśeṣika position, according to which the effect does not exist in its causes, is attacked beginning in the first line: “If the effect does not exist [in its causes], it does not arise, for no agent of this operation is present, as [in the case of] a flower in the sky. Or [contrariwise,] a flower in the sky would arise, since no agent of this operation is present, as [in the case of] an effect.”212 The problem is a familiar one; it is based on the correspondence principle. The proponent of Vaiśeṣika recognizes the problem, and maintains that the effect indeed does not exist before its arising. But he distinguishes between two types of existence. The effect is deprived of connection with the universal “existence” (sattāsambandha) prior to its arising, but it still exists, or “is there,” in the sense that it possesses astitva. The Vaiśeṣika thus replies: “Unlike the flower in the sky, the effect, arising through its own essence (astitva) (or: through the very fact of existing by itself) even without [the universal ‘existence’] subsisting in it through inherence, becomes the support [of this universal ‘existence’].”213 Note that here the Vaiśeṣika admits that the effect exists in a certain manner before its arising, and therefore brings to the problem a solution similar to those of Sarvāstivāda and Sāṃkhya. The Vaiśeṣika’s opponent then raises the question of whether existence (sattā) is conferred on something existent, something non-existent, or something 209 See DNC p. 512 l. 2 – p. 513 l. 3: … niṣṭhāsambandhayor ekakālatvāt ity etad eva vākyaṃ sabhāṣyaṃ praśasto ‘nyathā vyācaṣṭe: sambandhaś ca sambandhaś ca sambandhau, niṣṭhāyāḥ sambandhau niṣṭhāsambandhau, tayor ekakālatvāt / niṣṭhitaṃ niṣṭhā, … tasya svakāraṇaiḥ sattayā ca yugapat sambandhau bhavataḥ / —“Praśasta offers a different interpretation of the line niṣṭhāsambandhayor ekakālatvāt (which would normally mean ‘Because completion and connection [with the universal “existence”] are simultaneous’) and its commentary; [he explains it as follows:] ‘The two connections on the part of the completion are simultaneous. “Completion” [means] “that which is completed”; … the two connections of that [which is completed] with its causes and with [the universal] “existence,” [respectively,] are simultaneous.’” See also below. 210 DNC vol. 2, p. 459 l. 8-9: … sattāsambandho ‘bhidhānapratyayahetuḥ. 211 DNC vol. 2, p. 462 l. 3-5: nanu asat ity atra naña uttarapadābhidheyanivāraṇārthatvāt satpratiṣedhārthatvāt katham asya sātmakatvam? na, anekāntāt, aputrabrāhmaṇavad aguṇaguṇavat / yathā nāsya putro ‘stīty aputro brāhmaṇaḥ nāsya guṇo ‘stīty aguṇo guṇaḥ tathehāpi nāsya sad ity asat /; cf. Siṃhasūri, DNC p. 460 l. 10-11. 212 DNC vol. 2, p. 455 l. 1-2: yady asat kāryaṃ notpadyeta asannihitabhavitṛkatvāt khapuṣpavat / khapuṣpam api votpadyeta asannihitabhavitṛkatvāt kāryavat / 213 DNC vol. 2, p. 456 l. 1-2: … āśrayisamavāyād ṛte ‘pi kāryaṃ svenaivāstitvenotpannam āśrayo bhavati khapuṣpavaidharmyeṇa … 62 63 that is both.214 In this context the Vaiśeṣika remarks that the connection with existence can be denied of substances, etc., but that their existence in virtue of their essential form cannot be denied; the universal “existence” thus does not make nonexistent things existent.215 Ignoring this remark, the opponent then asserts that if the universal “existence” enters into connection with non-existent substances, etc., making them existent, then it can enter into connection with the horns of a hare—i.e., with something that never exists—and make them existent. The Vaiśeṣika—Siṃhasūri specifies that it is Praśastamati, i.e., Praśastapāda—replies that unlike the horns of a hare, effects, which are substances, qualities, or motions, are not completely nonexistent from the Vaiśeṣika point of view.216 Siṃhasūri takes the opportunity to remind us that there are two types of existence (bhāva): existence in virtue of essence (svabhāvasadbhāva) and existence in virtue of connection [with the universal “existence”] (sambandhasadbhāva).217 Here Mallavādin introduces the analysis of the word asat, “non-existent,” as “that which does not possess existence,” which sanctions an interpretation according to which the object in question still has an essence. Mallavādin continues by attributing the following statement to the Vaiśeṣika: “And the [object that is asat] does not [on that account] lack an identity, like the horn of a hare; just as in another system pradhāna, etc., have an identity, let it be the same in ours, even without connection with [the universal] ‘existence.’”218 This remark is extremely interesting. The other system mentioned here is, without the slightest doubt, Sāṃkhya; pradhāna, the primordial nature from which the world and its inhabitants are derived, indeed plays a central role in Sāṃkhya. Pradhāna is the source of all effects, and, according to the Sāṃkhya doctrine of satkāryavāda, all effects are present in pradhāna. The comparison of Vaiśeṣika with Sāṃkhya is quite surprising, since the two systems most often hold opposing positions. It is clear, however, that if the proponent of Vaiśeṣika seeks to postulate a sort of pre-existence for the effect, then whether he wills it or not, he thereby comes close to the position of Sāṃkhya and, what is more, to satkāryavāda. One can only conclude that the Vaiśeṣika has been forced to adopt such positions under pressure from the correspondence principle. The Vaiśeṣika next recalls that “universals, etc.”—i.e., universals (sāmānya), particularities (viśeṣa), and inherence (samavāya), as Siṃhasūri correctly specifies— possess an identity, but without thereby being connected with existence.219 He also recalls that the existence of universals is proven by the fact that different things can 214 DNC vol. 2, p. 459 l. 1-2: iha prāk sattāsambandhāt satāṃ vā asatāṃ vā sadasatāṃ vā dravyādīnāṃ satkarī sattā? A similar critique is found in Bhāvaviveka’s Madhyamakahṛdayakārikā and Tarkajvālā; see Tachikawa, 1994: 898. 215 DNC vol. 2, p. 460 l. 1-2: … dravyādīnāṃ sattāsambandhaḥ pratiṣidhyate na tu svarūpasadbhāva iti sattā naivāsatāṃ satkarī. 216 DNC vol. 2, p. 462 l. 2: na, śaśaviṣāṇādivad atyantanirātmakatvānabhyupagamāt kāryadravyaguṇakarmaṇām. 217 DNC vol. 2, p. 462 l. 11-12; cf. Mallavādin DNC p. 441 l. 5: dvividho hi bhāvaḥ: svabhāvaḥ sambandhaś ca. 218 DNC vol. 2, p. 462 l. 6-7: na ca tad api nirātmakaṃ śaśaviṣāṇavat, sattāsambandhād ṛte ‘pi yathā parapakṣe pradhānādīnāṃ sātmakatvaṃ tathehāpi syāt. 219 DNC vol. 2, p. 462 l. 8 – p. 463 l. 1: tvatpakṣe dṛṣṭāntābhāva iti cet, sāmānyādivad vā, … sāmānyādivad eva sātmakaṃ na ghaṭādivat sātmakam; and p. 462 l. 26 – p. 463 l. 8 (Siṃhasūri): … vaiśeṣikapakṣe sattāsambandharahitasātmakadṛṣṭāntābhāve codite sāmānyaviśeṣasamavāyānāṃ sātmakatvavat syād … 63 64 be referred to by a single word, idea, or convention.220 This last point is interesting when one considers the position adopted by Nyāya and later Vaiśeṣika, which we shall come to later. For Vaiśeṣika as Mallavādin knew it, universals explain the application of a single word to multiple things, but they are not themselves denoted by words. We shall return to this question later. Throughout the rest of the debate, the Vaiśeṣika’s position that a thing can exist in a certain manner without being entirely existent is reiterated several times. A thing can be existent in virtue of the existence of its essence, while being nonexistent with respect to its connection with “existence.”221 We cannot discuss all of these passages here, all the more so in that they have been dealt with, on two occasions, by the German-American scholar Wilhelm Halbfass.222 (Halbfass regularly raises the question of the degree to which Mallavādin and Siṃhasūri are reliable sources for our knowledge of Vaiśeṣika. The very fact that the positions they attribute to Vaiśeṣika fit so well with the problem of arising inspired by the correspondence principle—which Halbfass does not address—justifies a certain amount of confidence in their reliability.) Let us note further that Mallavādin goes on to cite a passage concerning distinctions among things prior to their connection with existence, a passage we can identify as belonging in all likelihood to the Kaṭandī.223 An extremely important question still remains. If the effect exists in some manner before its arising, can one deduce from this that the effect is beginningless? Or does it obtain this existence, this pre-existence, at a given moment prior to its arising in the strict sense? Mallavādin and Siṃhasūri seem to have nothing to say on this point. The Yuktidīpikā, which we studied earlier in connection with Sāṃkhya, for its part contains a debate with a Vaiśeṣika on satkāryavāda, in its discussion of verse 9. In the course of presenting the argument that one cannot produce something nonexistent—an argument we saw not long ago—the text has the Vaiśeṣika say: “But the effect is produced by the agent, etc., in an intermediate period. What is this intermediate period? It is the period between the commencement of the agents’ work and the production of the effect. When the causes set to work producing the effect and the effect has not yet come to light, that is the intermediate period. During this [period] the effect is produced by the actants (kāraka).”224 This passage 220 DNC vol. 2, p. 463 l. 1-3: sāmānyādīnāṃ sātmakatvam asiddham iti cet, na, svarūpabhinnatve sati abhinnavāgbuddhivyavahāraviṣayatvād yad viśeṣaṇaṃ tat sāmānyam. 221 DNC vol. 2, p. 463 l. 21-22: ekayā svabhāvasattayā sat tad evāsat sambandhisattayā; these are Siṃhasūri’s words, which are meant to explain Mallavādin’s more laconic idam eva tad ekasattāsad asad api … 222 See Halbfass, 1986; 1992: 169 ff. 223 Bronkhorst, 1993: 151-52. 224 YD p. 52 l. 16-21 (Pandeya) / p. 118 l. 9-15 (Wezler & Motegi): āha, nanu ca madhyame kāle kartrādibhiḥ kāryaṃ kriyate / kaḥ punar asau madhyamaḥ kāla iti? āha: ārambhāya prasṛtā yasmin kāle bhavanti kartāraḥ / kāryasyāniṣpādāt taṃ madhyamaṃ kālam icchanti // iti yadā hetavaḥ pravṛttārambhā bhavanty uddiśya kāryaṃ na ca tāvan naimittikasyātmalābhaḥ saṃvartate sa madhyamaḥ kālaḥ / tasmin kriyate kārakaiḥ kāryam iti /. Cf. Motegi, 1994: 815 ff.; Motegi rightly notes that the reading kāryasyāniṣpādāt in the verse is a correction and constitutes a departure from the manuscripts. The critical edition of Wezler and Motegi has kāryasyā niṣ<ṇ>ātās, which is difficult to interpret. 64 65 allows us to conclude—cautiously, of course—that the pre-existence of the effect is not beginningless. But neither is it momentary. The mat one is in the process of making is halfway existent, so to speak, during its arising, i.e., during the period it takes to make it. This should not surprise us too much. There has to be a mat, in some form or another, for the entire period in which the sentence “he makes a mat” obtains; the correspondence principle demands it. The Vaiśeṣika, it appears, submits to this demand, but unlike the Sāṃkhya and the Sarvāstivādin, he goes no further. Effects exist prior to their arising, it is true, but only for the interval of time needed for them to arise. It is interesting to note that in later Vaiśeṣika texts, the idea of pre-existence on the part of effects disappears completely. Discussions of the connection with existence (sattāsambandha) also tend to disappear, but more slowly. Dignāga’s Pramāṇasamuccayavṛtti alludes to the idea in a passage doubtless based on lost Vaiśeṣika texts.225 Vyomaśiva’s Vyomavatī, a commentary on the Padārthadharmasaṅgraha of Praśastapāda, includes a discussion of the connection with existence, part of which reads as follows: The aforementioned [dilemma] of knowing whether existence [arises], on the basis of connection with [the universal] “existence,” for existent things or for nonexistent things, is not [a real dilemma]. For arising and connection are simultaneous. In this way, the connection of objects with their causes and with [the universal] “existence” (svakāraṇasattāsambandha) itself constitutes their arising. Prior to this there is no existence, for satkāryavāda is rejected. And in admitting that there is connection with their own causes and with [the universal] “existence” on the part of [hitherto] non-existent things, it does not follow that such is the case for [totally non-existent things] such as the horns of a donkey, etc.; for the absence of causes for their arising is certain, insofar as such things do not exist at all. In the case of eternal things, the dilemma does not apply, for the relation of prior and 226 posterior does not apply. The influence on this passage of certain ideas attributed to the Vaiśeṣikas by Mallavādin and Siṃhasūri is undeniable. The phrase “For arising and connection are simultaneous,” in particular, recalls an early discussion within the school, a discussion that left its traces in the Dvādaśāranayacakra. This text says there are many who think that the connection with the universal “existence” occurs after the connection with causes. The author of the vākya and the bhāṣya—who seems to be Rāvaṇa, the author of the Kaṭandī—maintains instead that connection with the universal “existence” occurs at the same time as arising. Finally, Praśastamati, also known as Praśasta and Praśastapāda, believes that the connection with causes and the connection with the universal “existence” arise in an already established 225 See the Tibetan translation of Kanakavarman (Hattori, 1968: 235): bye brag pa rnams ni ‘bras bu skye ba’i raṅ gi rgyu la ‘du ba ‘am yod pa la sogs pa daṅ ‘phrod pa ‘du bar ‘dod do, which Hattori (p. 69) translates as follows: “The Vaiśeṣikas hold that the ‘rise’ (janman) of a result is either the inherence (samavāya) [of the result] in its own cause (kāraṇa) or the inherence of being or some other [characterizing] property in it.” Cf. Nozawa, 1993: 9. 226 Vyomaśiva’s Vyomavatī, vol. 1, p. 36 l. 13-17: tatra yat tad avocāma, kiṃ satāṃ sattāsambandhāt sattvam? atha asatām iti? tad asat / niṣpādasambandhayor ekakālatvāt / tathā hi, padārthānāṃ svakāraṇasattāsambandha evotpattiḥ / na pūrvaṃ sattvam, satkāryavādapratiṣedhāt / na cāsatāṃ svakāraṇasattāsambandhābhyupagame kharaviṣāṇādiṣu tathābhāvaprasaṅgaḥ, tadutpattikāraṇābhāvasya tadabhāvenaiva niścayāt / nityeṣu pūrvāparabhāvānupapatter vikalpānupapattiḥ /. Cf. Halbfass, 1986a: 77; 1992: 253. 65 66 object.227 But none of these thinkers seems to reject, as Vyomaśiva explicitly does, the pre-existence of the effect prior to its arising. Uddyotakara adopts the same position as Vyomaśiva. Connection with the universal “existence” takes place neither for something existent nor for something non-existent; as soon as the thing is there, it stands in relation to existence.228 Uddyotakara thus seems to align himself with a later position than that of Praśastapāda. One might be tempted to conclude from this that Uddyotakara himself belonged to a later period than Praśastapāda.229 We shall return to this question in a moment. The Japanese scholar Masanobu Nozawa (1993) has shown that Vyomaśiva’s definition of effect (svakāraṇasattāsambandha; cf. the passage quoted above) was not followed by other Vaiśeṣika authors, either contemporary or subsequent. The Vyomavatī might therefore be the latest Vaiśeṣika text to preserve traces of the early discussion; but certain texts from other schools mention this definition and even seem to be aware of the related notion of pre-existence as it appeared in the early discussion. Witness the following passage, which comes from the Prameyakamalamārtaṇḍa of the Jaina author Prabhācandra (ca. 1000): Concerning what was said—namely, that the connection [of an object] with its causes and with [the universal] “existence” (svakāraṇasattāsambandha) constitutes its arising, etc.—[the reply is] no, for if “arising” were synonymous with “connection with its causes and with [the universal] ‘existence,’” it would be eternal; and if arising were eternal, the effect would not pass away. Moreover, would this [connection] be the inherence of [the universal] “existence” in existent [objects], or rather in non-existent [objects]? To begin with, [it is] not [inherence] in nonexistent [objects], for then there would be [inherence of the universal “existence”] even in lotuses in the sky, etc. And if [you say] that in such as these there is not [inherence of the universal “existence”] because they are totally non-existent, how would there not be total non-existence in the case of qualities, quality-possessors, etc.? If [you say] because of inherence, there will be mutual dependence (itaretarāśraya): for these [qualities, quality-possessors, etc.] are not totally nonexistent when inherence [of the universal “existence”] is established; and there is inherence [of this universal] when they are not so. Nor is there [inherence of the universal “existence”] in existent [objects]. For would their existence prior to the inherence be due to another inherence, or to themselves? If [you say that] their existence would be due to another inherence, [the reply must be] no, because this 230 [inherence] is held to be unique. And even [if one were to grant] multiple [inherences], prior to this [other inherence], their existence would be due to [yet] another inherence, [thus] giving rise to an infinite regress. But if one maintains an 231 existence due to themselves, the very idea of inherence becomes useless. 227 DNC vol. 2, p. 516 l. 3 – p. 517 l. 2: tattvopanilayanāt sadādyabhidhānārthaṃ kāraṇasamavetasya vastuna uttarakālaṃ sattāsambandha iti bahūnāṃ matam / vastūtpattikāle eva iti tu vākyakārābhiprāyo ‘nusṛto bhāṣyakāraiḥ / siddhasya vastunaḥ svakāraṇaiḥ svasattayā ca sambandha iti prāśastamato ‘bhiprāyaḥ / 228 Uddyotakara, Nyāyavārttika 2.2.64 (p. 669 l. 15-16): na sataḥ sattāsambandhaḥ, nāsataḥ / yadaiva tad vastu, tadaiva sattayā sambaddham …. For further examples, see Halbfass, 1992: 190 ff.; also Halbfass, 1989. 229 Halbfass (1989: 555; 1992: 191-92) mentions several conflicting views on the relative chronology of these two authors. 230 According to Vaiśeṣika, there exists but a single inherence; cf. WI p. 87 (§377): … sarvatraikaḥ samavāya iti—“Everywhere there is one inherence.” 231 Prabhācandra’s Prameyakamalamārtaṇḍa on sūtra 4.10, pp. 618-19: yac coktam: svakāraṇasattāsaṃbandha evātmalābha ityādi / tan na, ātmalābhasya svakāraṇasattāsamavāyaparyāyatāyāṃ nityatvaprasaṅgāt, tannityatve ca kāryasyāvināśitvaṃ syāt / kiñ ca, 66 67 What is striking in this passage is the attention devoted to the various ways one could imagine an object to exist before its arising. In the context of a debate with Vaiśeṣika, such attention would seem wasted, since one of the characteristic teachings of the school—at least in the texts that have come down to us—is precisely its rejection of the existence of an object before its arising. It seems likely that Prabhācandra thus still had access to Vaiśeṣika texts in which this possibility was envisioned, probably the Kaṭandī of Rāvaṇa and the Ṭīkā of Praśasta. Why is it that the discussion of the pre-existence of things, though clearly occupying a position of prime importance in the Vaiśeṣika texts known to Mallavādin,232 is absent from earlier (Vaiśeṣika Sūtra) and later texts of the school? The answer must be that the problem to which this discussion offered a solution arose at a particular time, and that the problem was addressed in a different way in later texts. The more or less sudden appearance of the problem of the arising of things can be explained without too much difficulty in light of the fact that it was invented, or in any case systematically exploited, by Nāgārjuna. Its appearance within the Vaiśeṣika school thus seems due to the impact of the ideas of this Buddhist thinker. Its disappearance was doubtless the result of the discovery, at a later time, of much more satisfying solutions; we shall turn to these momentarily. One might even venture to hypothesize that the survival of Praśastapāda’s Padārthadharmasaṅgraha was made possible insofar as it does not discuss the problem of the arising of things, while the loss of the Ṭīkā by the same author could be the result of the stress it places on a solution later viewed as outdated. This hypothesis might shed light on the following question. Uddyotakara lived close to the time of Praśastapāda.233 But Praśastapāda accepted the preexistence of that which is to arise, while Uddyotakara rejected this possibility, as we have already noted. What is the reason for this difference? The answer is not difficult to guess. Uddyotakara was the author of the Nyāyavārttika, a commentary on the Nyāya Bhāṣya. Now in the Nyāya Bhāṣya we meet with a completely different solution to the problem of the arising of things. We shall come to the relevant passages soon; for now, let us go ahead and observe that the Nyāya school solves the problem by having words denote genera or universals as well as individuals. The word “pot” in the sentence “the pot arises,” and the word “mat” in the sentence “he makes a mat,” thus already refer to something before the pot arises or the mat is asau satāṃ sattāsamavāyaḥ, asatāṃ vā syāt? na tāvad asatām, vyomotpalādīnām api tatprasaṅgāt / athātyantāsattvāt teṣāṃ na tatprasaṅgaḥ, guṇaguṇyādīnām atyantāsattvābhāvaḥ kutaḥ? samavāyāc cet, itaretarāśrayaḥ: siddhe hi samavāye teṣām atyantāsattvābhāvaḥ, tadabhāvāc ca samavāyaḥ / nāpi satām, samavāyāt pūrvaṃ hi sattvaṃ teṣāṃ samavāyāntarāt, svato vā? samavāyāntarāc cet, na, asyaikatvābhyupagamāt / anekatve ‘pi ato ‘pi pūrvaṃ samavāyāntarāt teṣāṃ (the edition has: pūrva(vaṃ)samavāyantarāt teṣā) sattvam ity anavasthā / svataḥ sattvābhyupagame tu samavāyaparikalpanānarthakyam /. A quite similar passage appears in the same author’s Nyāyakumudacandra (on 1.7, pp. 220-21). Cf. also DNC vol. 2, p. 459 l. 1-4: iha prāk sattāsambandhāt satāṃ vā asatāṃ vā sadasatāṃ vā dravyādīnāṃ satkarī sattā? na tāvad asatāṃ satkarī śaśaviṣāṇādīnām api satkaratvaprasaṅgāt / nāpi satāṃ bhūtatvāt … / satāṃ ca punaḥ sattāsambandhāt sattvād anavasthāprasaṅgāt / 232 Principally the Kaṭandī and Praśasta’s Ṭīkā (assuming, of course, that the theory advanced in Bronkhorst, 1993, is correct), but other texts, too: DNC pp. 516-17 contrasts the view of “many” Vaiśeṣikas (iti bahūnāṃ matam) with the those of the vākya- and bhāṣya-kāra and of Praśastamati. 233 Halbfass, 1992: 191 ff. with p. 202 n. 95. 67 68 made. For in Nyāya, as in Vaiśeṣika, universals are eternal. Uddyotakara accepts this solution, and therefore has no need to posit the obscure pre-existence of objects such as the pot and the mat. He is then free to reinterpret the position of the Kaṭandī, in such a way that the question of pre-existence no longer arises. Vyomaśiva, as we have seen, agrees with Uddyotakara on this matter: he, too, believes that arising and the connection with [the universal] are simultaneous, and he, too, does not breathe a word about the pre-existence of things. Vyomaśiva had also found a solution to the problems posed by sentences such as “the pot arises” and “he makes a mat.” But although he accepts the position of the Nyāya school in regard to the denotation of words—the object of a word is an individual qualified by a universal234—he presents a different solution to the problem of the arising of things. We shall return to it momentarily. 10. Critiques of the Existence of a Thing before Its Arising Before we leave the realm of positions that uphold the existence of a thing before its arising, let us take a look at a few passages from post-Nāgārjunian literature that express a more or less critical attitude toward these positions. We shall begin with a passage from the Āgamaśāstravivaraṇa of Śaṅkara, parts of which we have already seen. The Āgamaśāstra of “Gauḍapāda,” in a series of verses (4.14 ff.), takes up an argument already found in the Vigrahavyāvartanī of Nāgārjuna (verse 49). The cause, observes the Āgamaśāstra, is the origin of the effect, and the effect the origin of the cause. The author is evidently referring to the following type of chain: the father produces a son, and the son, becoming a father himself, thus produces a father. This is not possible, the Āgamaśāstra says, because it would follow that the father is born of the son. If the cause and the effect arise, their order must be determined; for if both arise at the same time, they do not stand in a causal relation, like the two horns of a cow.235 This argument, however unconvincing, can be understood easily in light of the correspondence principle. In particular, the argument is based on the fact that the sentence “the cause is the origin of the effect, and the effect the origin of the cause” uses just two words—“cause” and “effect”—to refer to a potentially infinite series of things. It takes account, in any case, of just two objects, asking which of the two comes first, or whether they are simultaneous. The argument is therefore based solely on words, paying no attention to the things they designate. This is precisely the observation of the opponent to whom Śaṅkara gives voice in his introduction to verse 20 of the fourth chapter of the Āgamaśāstra: “Basing yourself only on our words ‘there exists a causal connection between the cause and the effect,’ you offer [but] a semblance of reasoning when you say, ‘like the father born from the son,’ and ‘they have no causal connection, like the two horns of a cow,’ etc. For we did not maintain that the effect is established on the 234 See, e.g., Vyomaśiva’s Vyomavatī, vol. 2, p. 241 l. 23-24: yatra yatra gakāraukāravisarjanīyānām itthambhūtānupūrvīm upalabhate, tatra tatra gotvaviśiṣṭo ‘rthaḥ pratipattavyaḥ pratipādayitavyaś ceti. 235 GK 4.14-17: hetor ādiḥ phalaṃ yeṣām ādir hetuḥ phalasya ca / hetoḥ phalasya cānādiḥ kathaṃ tair upavarṇyate // hetor ādiḥ phalaṃ yeṣām ādir hetuḥ phalasya ca / tathā janma bhavet teṣāṃ putrāj janma pitur yathā // saṃbhave hetuphalayor eṣitavyaḥ kramas tvayā / yugapat saṃbhave yasmād asaṃbandho viṣāṇavat // 68 69 basis of a cause that is itself unestablished, or that the cause is established on the basis of an unestablished effect. On the contrary, we maintain that there is a causal connection like [the one that exists] between the seed and the sprout.”236 Note the sentence: Basing yourself only on our words “there exists a causal connection between the cause and the effect,” you offer but a semblance of reasoning. It reveals unequivocally that the inner workings of the argument were perfectly clear to the interlocutors in this debate, at least in Śaṅkara’s day. The fact that Śaṅkara, though aware of the objection, nevertheless maintained the validity of this kind of argument shows, if proof were needed, that an understanding of the exact nature of the argument is not sufficient to invalidate it. We come across an interesting reaction to the satkāryavāda of the Sāṃkhyas in the Dvādaśāranayacakra of Mallavādin. Recall that the Sāṃkhyas had argued that the effect exists in the cause precisely because the cause is a cause; without an effect, the cause cannot exist. Later representatives of the Sāṃkhya school replaced this argument with a different one, but the Yuktidīpikā, as we saw, still comments on the older form. Mallavādin, for his part, is aware of the argument, which he criticizes. He attacks the idea that there is no cause without an effect. According to him, a cause sometimes has an effect, and sometimes does not. There is no rule about the existence or non-existence of an effect when there is a cause, because the effect might or might not exist when the cause exists, and likewise when the cause does not. If one persists in holding that a cause must produce something, not even seeds would be causes, because sometimes they do not produce anything. A cause would thus be a non-cause, because the connection between cause and effect is not free of exceptions.237 Note that Mallavādin’s response to the satkāryavāda of the Sāṃkhyas inevitably has to do with the denotation of words, in particular the word kāraṇa, “cause.” He grants that a cause is called kāraṇa because it produces (karoti) something, but adds that this production is a general rule, to which there are numerous exceptions. Precisely because there is no rule concerning the existence or non-existence of an effect when there is a cause, it follows that the question of whether a cause does or does not produce an effect can be asked only after the cause has been identified as such. The purpose of saying “the cause produces something” is indeed to communicate something unknown. In this situation, the exceptions deal with cases in which the cause does not produce its effect.238 236 Āgamaśāstravivaraṇa introducing GK 4.20: nanu hetuphalayoḥ kāryakāraṇabhāva ity asmābhir uktaṃ śabdamātram āśritya cchalam idaṃ tvayoktaṃ putrāj janma pitur yathā, viṣāṇavac cāsaṃbandha ityādi / na hy asmābhir asiddhād dhetoḥ phalasiddhir asiddhād vā phalād dhetusiddhir abhyupagatā / kiṃ tarhi? bījāṅkuravat kāryakāraṇabhāvo ‘bhyupagamyata iti / 237 DNC I p. 36-37: tathā ca kāraṇe kāryasadasattāniyamaḥ, kāraṇe saty eva bhāvābhāvābhyām asati ca … / … bījādīnām apy akāraṇataiva kvacid akaraṇād iti kāraṇam apy akāraṇam eva, kāryakāraṇāvyabhicārābhāvāt / 238 DNC I p. 37: kāryasadasattvāniyamāt tu kāraṇe kāraṇatāyām eva karaṇākaraṇe kāryasya, aviditavedanārthavidhiparatāyāṃ vākyapravṛttes tasyām avasthāyām anupajanitaviṣayatvād apavādasparśasya … /. The commentator Siṃhasūri—on the basis of whose Nyāyāgamānusāriṇī it has been possible to reconstruct, at least partially, the Dvādaśāranayacakra, itself lost—comments on the enigmatic expression anupajanitaviṣayatvād apavādasparśasya as follows: tadā hi “karotīti kāranam” iti kāraṇatvavidhānamātraṃ kriyate deśakālādiviśeṣāviśeṣaṇād asati svaviṣaye kam artham apavādaḥ spṛśet: kiṃ karoty eṣa na karoty api kvacit kadācid iti? 69 70 Mallavādin implicitly grants that the Sāṃkhyas’ argument in favor of satkāryavāda would be correct if the word kāraṇa denoted only those causes that give rise to effects. In that case, the cause would be accompanied by its effect without exception, and the effect would have to exist in the cause. Mallavādin therefore accepts the correspondence principle. He avoids the consequences the Sāṃkhyas had drawn from it by modifying the way in which the word kāraṇa denotes its object. A cause is then no longer inseparably tied to its effect. 11. Nyāya Having already stated the solution of the Nyāya school to the problem of the arising of things, we may now turn our attention to the relevant passages. We shall see that the Naiyāyikas were not alone in their efforts to find such a solution, several variants of which we shall be examining. They all have in common the postulation of objects of denomination besides just individuals. By showing that the word “pot,” in the example “the pot arises,” refers to something other than an individual pot, one can still believe that the sentence correctly describes a situation, but without having to accept the existence of things before their arising. These solutions choose the path of semantics, but they do not abandon the correspondence principle. On the contrary, they are based on it. They show that the principle is correct, that there is a direct connection between words and the things they denote. The only error would be to believe that these denoted things must be individuals. So far we have been interested in thinkers who sought an “ontological” solution to the problems connected with the correspondence principle. All of them were confronted with the problem of the arising of things, which they sought to resolve in various ways. Nāgārjuna, perhaps the most radical among them, concluded that the phenomenal world is contradictory and hence non-existent. “Gauḍapāda” and his commentator Śaṅkara (the latter in a limited number of passages only) drew the conclusion that in ultimate reality, nothing either arises or ceases to exist. The majority of the remaining thinkers we have so far studied preferred to assert that things pre-exist or at least exist at the moment of their arising. The Jainas, for their part, accepted this last solution, but without rejecting the opposing point of view. These thinkers thus chose an “ontological” solution. Others preferred a “semantic” solution. Recall that the problem resulting from the correspondence principle consists essentially—to continue using the example of the pot and its arising—in the fact that there is no pot in the situation described. In other words: nothing in this situation corresponds to the word “pot.” The problem is solved as soon as one can show that the word refers to something existent. If one denies that the individual pot already exists when it arises, one must find some other object for the word “pot” to designate, an object that does exist when the pot arises. Finding such an object was not especially difficult. Patañjali’s Mahābhāṣya, a text dating in all likelihood to the second century before the common era, thus well before Nāgārjuna and all discussion of the arising of things, already contains, in its opening pages, a discussion of words and their objects.239 According to 239 For a more thorough discussion of Patañjali’s views, cf. Scharf, 1996: 21-148. 70 71 Patañjali (who seems to deviate from the opinion of Kātyāyana, whose vārttika he is commenting on), the object designated by a word is eternal. The question then arises: what object is eternal? Patañjali presents two possible answers. First he proposes that words refer to forms (ākṛti), for if forms are eternal, individuals (dravya) are not. He then offers a second answer, playing on the ambiguity of the word dravya, which sometimes means “individual,” sometimes “substance.” Though individual objects are not eternal, the substance of a thing, Patañjali assures us, is. His second answer thus consists in saying that dravya, i.e., substance, is the object denoted by a word.240 The discussion hinges on the two expressions ākṛti, which we have translated as “form,” and dravya, which can be translated as either “individual” or “substance.” The reason for this choice becomes clear elsewhere in the Mahābhāṣya, where we learn that two thinkers who preceded both Patañjali and Kātyāyana had designated these two elements as objects of words. Vājapyāyana had maintained that form (ākṛti) is what is signified by a word.241 Vyāḍi, for his part, considered dravya to be the object of a word.242 Patañjali himself, in the introduction to his work, expresses the opinion that words signify both form and dravya.243 Later thinkers thus had little difficulty finding an eternal object for a word, which might thereby solve the problem of arising. We shall see that the two positions presented in the Mahābhāṣya were indeed used to this end, as were others. It is also important to emphasize that this solution was not accepted lightly. Recall the position of early Vaiśeṣika, which we spoke of earlier. Nothing would have been easier for those thinkers than to have said that universals, which figure among the categories of Vaiśeṣika, constitute the objects designated by words. But the school chose a different solution instead, probably to avoid the “semantic” solution we shall now address. The “semantic” solution denies that words refer, or refer exclusively, to individuals. In the sentence “the pot arises,” the word “pot” is taken to refer not to the individual pot that is in the process of arising, but rather to something more general, something that covers the individual pot, of course, but that is not reducible to it. We shall see that this modified version of the correspondence principle found favor with most later thinkers, Brahmins and Buddhists alike. The words of a sentence, understood in this way, must correspond to some existent thing that is different from an individual. What object can fulfill this function? Such an object is easily found in the case of schools that accept the existence of universals. A universal, as we saw in our discussion of Vaiśeṣika, is eternal and therefore certainly present when one utters a sentence such as “the pot arises.” If one accepts that the word “pot” refers, not to an individual that does not yet exist, but rather to an eternal universal belonging to all pots, the sentence no longer poses a problem. This solution was indeed accepted by several Brahmanical schools in classical India. Let us closely examine a few passages referring to it. The Nyāya Sūtra and its Bhāṣya devote a section (NS 2.2.59-66) to the question of what words designate.244 The Bhāṣya specifies from the outset that the 240 Mbh I p. 7 l. 8-25. P. 1.2.64 vt. 35 (= Mbh I p. 242 l. 10): ākṛtyabhidhānād vaikaṃ vibhaktau vājapyāyanaḥ. 242 P. 1.2.64 vt. 45 (= Mbh I p. 244 l. 8): dravyābhidhānaṃ vyāḍiḥ. 243 Mbh I p. 6 l. 8 (Paspaśāhnika): kiṃ punar ākṛtiḥ padārtha āhosvid dravyam? ubhayam ity āha / 244 Cf. Scharf, 1996: 151-193; Meuthrath, 1996: 82 ff. 241 71 72 discussion has to do with nāmapada, which is to say, primarily, nouns such as “cow” (go). The two texts first present a series of arguments in favor of the individual as the object designated by a word (NS 2.2.60). This position is then confronted with certain objections; in the context of this discussion, the following example is mentioned: in the sentence “he makes a mat” (kaṭaṃ karoti) one uses the [word “mat”], even though the [mat] is not present (atadbhāve ‘pi tadupacāraḥ), because—as the sūtra enigmatically says—it is for that (tādarthya). The Bhāṣya explains: “‘Because it is for that’ means: when the grass used to make a mat is being arranged, one says, ‘he makes a mat.’”245 In such cases, the Bhāṣya continues, “such and such a word designates something it does not [normally] designate.”246 What is being proposed here is that the word “mat” in the sentence “he makes a mat” designates the grass used to make the mat, rather than the mat itself. The expression that appears in the Bhāṣya is upacāra, which refers to “use,” or more specifically “metaphorical use.” Note how close these remarks come to the Vaiśeṣika position we examined via the Dvādaśāranayacakra of Mallavādin. The Vaiśeṣika had asserted, in the case of the sentence “he makes a mat,” an obscure sort of existence (we spoke of “preexistence”) on the part of the mat, in order to find an object for the word “mat.” The Nyāya Bhāṣya seems to avoid this position, suggesting that here the word “mat” designates, exceptionally, the causes of the mat. This passage from the Nyāya Bhāṣya shows in any case that the Nyāya Sūtra and its Bhāṣya were well aware of the problem of the arising of things. The passage also proposes a solution to the problem, but it is not definitive. The definitive solution comes in sūtra 2.2.66, which says: “The object of a word is the individual, the form, and the universal” (vyaktyākṛtijātayas tu padārthaḥ). The essential element in this list is clearly the universal (jāti). Thanks to it, the word “mat” in “he makes a mat” refers to something, namely, to the universal that inheres in this mat as in all mats, even before the mat is made. It seems beyond doubt that this second solution is the preferred solution of the Nyāya Sūtra and its Bhāṣya. First, universals offer a solution that applies invariably in every similar case, unlike the solution of the grass and the mat. More importantly, sūtra 2.2.66 comes at the end of the discussion in question, concluding it, so to speak. The other solution offered by the Sūtra and Bhāṣya is nevertheless not without interest. The grass that is to constitute the mat constitutes the cause of the mat. To say that the word “mat” refers to its cause is close to the Sāṃkhya position, according to which an object, in our case the mat, exists in its cause. Nyāya, of course, agrees neither with this position nor with the Sāṃkhya solution to the problem of the arising of things. It is therefore all the more surprising to find this 245 NBh 2.2.62: tādarthyāt: kaṭārtheṣu vīraṇeṣu vyūhyamāneṣu kaṭaṃ karotīti bhavati. The Mahābhāṣyadīpikā of Bhartṛhari seems to offer another example of the same principle, this time concerning the use of a verb rather than a noun (Manuscript p. 13a l. 4-5; “Critical edition” Āhnika I p. 33 l. 24-26; ed. Abhyankar-Limaye p. 40 l. 17-19; ed. Swaminathan p. 48 l. 18-19): yathā tādarthyād adhiśrayaṇamātraṃ dṛṣṭvā pacaty ayam iti prayuñjānaṃ prati vacanam ucyate nāyaṃ pacaty adhiśrayati kiṃ tu udakaṃ saṃpādayatīti—“For example, someone uses the expression ‘he cooks,’ having seen merely the act of putting [something] onto the fire, because [he thinks] it is for that (tādarthyāt). We reply: ‘He does not cook; [certainly] he puts [something] onto the fire, but he prepares the water.’” 246 NBh 2.2.62: atacchabdasya tena śabdena abhidhānam. 72 73 solution, in an adapted form, in the texts of Nyāya: the mat does not exist in its cause, but in certain circumstances the word “mat” refers to this very cause. Another discussion relevant to our problem is found in the Nyāya Sūtra and its Bhāṣya 4.1.11-18. The Bhāṣya opens the debate with the question of how arising is possible (katham utpattir iti cet). The Sūtra and Bhāṣya both answer that manifest things (vyakta) arise from other manifest things. A pot, for instance, arises from shards. Having declared this fact to be indisputable, the Sūtra and Bhāṣya go on to present another position, according to which something existent arises from something non-existent (asataḥ sad utpadyate). The sprout, in the example given by the Bhāṣya, arises having destroyed the seed (upamṛdya bījam aṅkura utpadyate).247 This position comes in for immediate critique with the observation that if such were the case, the thing that destroys the seed would no longer need to arise, since it would clearly be there already.248 The opponent defends himself in sūtra 4.1.16 and its commentary, and this defense is of particular interest to us. The opponent argues that the nouns in a sentence (literally: the words that express a syntactic function in relation to a verb; kārakaśabda) can refer to things past and future (nātītānāgatayoḥ kārakaśabdaprayogāt). The Bhāṣya explains: “The nouns in a sentence can refer to what is past and to what is future, [and therefore] to what is not there. One frequently observes metaphorical applications (bhāktāḥ prayogā[ḥ]), as [in the sentences] ‘the son will be born,’ ‘he rejoices at the son who will be born,’ ‘he names the son who will be born,’ ‘there was a pot,’ ‘he is sorrowful over the broken pot,’ ‘the shards of the broken pot,’ ‘sons not born cause a father pain.’”249 The Sūtra and the Bhāṣya go on to reject the opponent’s position, but without touching on the principle he had invoked. According to this principle, once again, the nouns in a sentence can refer to things past and future. Note that this formulation is exactly the opposite of the correspondence principle, which instead implies: the nouns in a sentence refer to things present. The authors of the Sūtra and the Bhāṣya were obviously familiar with a third response to the problems connected with the correspondence principle. The existence of the effect prior to its arising is again the object of discussion in Nyāya Sūtra 4.1.47-49, as interpreted by the Bhāṣya. Sūtra 47 presents the position of an opponent who maintains that the effect, prior to its arising, can neither be non-existent, nor existent, nor existent and non-existent. The next sūtra replies that it is non-existent, the proof being that its arising and perishing are perceptible. Sūtra 49 is of particular interest, saying: “This non-existent [effect] is, however, established by the mind (buddhi).”250 The Bhāṣya explains: “Prior to its arising, the effect, the cause of which is determined, is established by the mind in 247 The same position is mentioned in Śaṅkara’s Bhāṣya on Bṛhadāraṇyaka Upaniṣad 1.2.1, and rejected there, too, though for different reasons: ghaṭādikāraṇasyāpy asattvam evānupamṛdya mṛtpiṇḍādikam ghaṭādyanutpatter iti cen na / mṛdādeḥ kāraṇatvāt / mṛtsuvarṇādi hi tatra kāraṇaṃ ghaṭarucakāder na piṇḍākāraviśeṣaḥ / etc. See also the Bhāṣya on Chāndogya Upaniṣad 6.2.2: yady api bījopamarde ‘ṅkuro jāyamāno dṛṣṭo ‘bhāvād eveti tad apy abhyupagamaviruddhaṃ teṣām / katham / ye tāvad bījāvayavā bījasaṃsthānaviśiṣṭās te ‘ṅkure ‘py anuvartanta eva na teṣām upamardo ‘ṅkurajanmani / etc. 248 NBh 4.1.15: yad upamṛdnāti na tad upamṛdya prādurbhavitum arhati vidyamānatvāt / yac ca prādurbhavati na tenāprādurbhūtenāvidyamānenopamarda iti / 249 NBh 4.1.16: atīte cānāgate cāvidyamāne kārakaśabdāḥ prayujyante / putro janiṣyate, janiṣyamāṇaṃ putram abhinandati, putrasya janiṣyamāṇasya nāma karoti, abhūt kumbho, bhinnaṃ kumbham anuśocati, bhinnasya kumbhasya kapālāni, ajātāḥ putrāḥ pitaraṃ tāpayantīti bahulaṃ bhāktāḥ prayogā dṛśyante / 250 NS 4.1.49: buddhisiddhaṃ tu tad asat. 73 74 the form ‘This [alone] is capable of giving rise to that, not just anything,’ because one sees that the arising is determined.”251 Note how the mind is called on to play a role in the debate here; later we shall see other examples of this. Let me summarize the four responses found in the Nyāya Sūtra and its Bhāṣya: 1. The authors of the Sūtra and the Bhāṣya themselves choose the position according to which words, and nouns in particular, refer to universals. Words also refer to individuals and to forms, but it is clearly universals (possibly with forms) that allow one to avoid the problem of the arising of things. 2. They mention another position, according to which nouns can designate the causes of the things they normally designate. We have already shown how this position is similar to that of classical Sāṃkhya. 3. According to a third position known to these Naiyāyika authors, a noun can denote its object even when that object is located in the future or the past. This third position, needless to say, recalls that of the Sarvāstivādins, who maintained precisely that objects exist in the past and the future as well as in the present. But let us note the difference between this third position and that of the Sarvāstivādins. For the latter, a past or future object exists. For the Naiyāyika, or for the opponent he has appear in his texts, the past or future object does not exist, but a word can designate it. 4. The fourth position is perhaps only a variant of the third. A non-existent object can be “established by the mind.” This position might come close to that of Vyomaśiva and other Vaiśeṣikas, as we shall see in what follows. 12. Mīmāṃsā Let us now consider the position of Mīmāṃsā on this matter. Śabara’s Bhāṣya on Mīmāṃsā Sūtra 1.3.30-35 discusses the problem of whether words refer to individuals (vyakti) or to forms (ākṛti).252 The accepted position (siddhānta) is first presented in sūtra 1.3.33 and its commentary. Here are a few excerpts: Rather, it is the form [that constitutes the object of the word], because [only it can be] the object of an action. (MīS 1.3.33) … The form is the object of the word. Why? Because [only it can be] the object of an action. The [Vedic] saying “One should pile a falcon-pile [i.e., construct a falcon altar]” is possible for a form, if the word “falcon” has the form as its object. If, however, the individual were denoted, [the saying] would be meaningless, for it would express something impossible, given that one cannot produce the individual “falcon” through piling. Therefore, the word denotes the form. [Objection:] But the piling will be carried out with individual falcons. [Reply:] The object of the word “falcon” is not the means par excellence (sādhakatama), for this word “falcon” refers to that which is most desired (īpsitatama). For this reason, [the sense of the saying is:] “A falcon ought to be 253 produced through piling.” This is possible if the word denotes the form. 251 NBh 4.1.49: idam asyotpattaye samarthaṃ na sarvam iti prāg utpatter niyatakāraṇaṃ kāryaṃ buddhyā siddham utpattiniyamadarśanāt. 252 In Mīmāṃsā the term ākṛti becomes a synonym of sāmānya, “universal”; see Harikai, 1997: 398. 253 MīS + MīBh 1.3.33: ākṛtis tu kriyārthatvāt //33// … ākṛtis śabdārthaḥ / kutaḥ? kriyārthatvāt / śyenacitaṃ cinvīta iti vacanam ākṛtau saṃbhavati yady ākṛtyarthaḥ śyenaśabdaḥ / vyaktivacanatve tu na 74 75 This passage requires a word of explanation, since the Vedic example chosen by Śabara might be confusing. The example is śyenacitaṃ cinvīta, a line from the Taittirīya Saṃhitā (5.4.11) we have translated literally as “One should pile a falconpile.” As the passage goes on to show, Śabara interprets this to mean “A falcon ought to be produced through piling” (cayanena śyeno nirvartayitavyaḥ). We thus have a sentence of the type “He makes a mat”—or, if you prefer, “A mat ought to be made”—a problematic sentence for thinkers of the period, since there is not yet a mat when the sentence applies. The passage from Śabara should probably be interpreted in this way.254 The passage is certainly not resistant to such an interpretation. The sūtra itself argues that the reason words must refer to forms is that they sometimes refer to the object of an action. The end of the passage makes it clear that we are dealing with grammatical objects: the expression “that which is most desired (īpsitatama)” refers to Pāṇini’s sūtra kartur īpsitatamaṃ karma (P. 1.4.49), which is precisely how a grammatical object is defined. Śabara rejects the opponent’s interpretation (“the piling will be carried out with falcons”), in which the falcons would be the instrument of the action. Note that the expression “means par excellence (sādhakatama)” refers to a sūtra from Pāṇini (P. 1.4.42: sādhakatamaṃ karaṇam) that defines the instrument. It seems likely that Śabara, as well as the author of the Mīmāṃsā Sūtra, here offers his solution to the problem of the arising of things, a problem based on the correspondence principle. Another interpretation of the passage is nevertheless possible, an interpretation that seems to be followed by Kumārila Bhaṭṭa, the author of the commentary known as the Tantravārttika.255 In the Vedic sentence “One should pile a falcon-pile,” it is obviously not a real falcon that is meant, but the form of a falcon; one prepares an altar that has the form of a falcon. The word “falcon” in this sentence thus does not designate any individual of the species falcon, but rather the form, and only the form, of the bird. This interpretation—although suggested by the Vedic example selected, “One should pile a falcon-pile”—has obvious weaknesses. For instance, it is totally incomprehensible in this case that Śabara, on the basis of the possibility of using the word “falcon” to refer secondarily to its form, should have thereby been led to conclude that all words (and it is no doubt chiefly a question of nouns) refer to forms, and only indirectly to individuals. Another weakness is that the wording of the sūtra, which Śabara repeats, loses its meaning on this interpretation. Śabara explains the words of the sūtra as follows: “The form is the object of the word. cayanena śyenavyaktir utpādayituṃ śakyata ity aśakyārthavacanād anarthakaḥ / tasmād ākṛtivacanaḥ / nanu śyenavyaktibhiś cayanam anuṣṭhāsyate / na sādhakatamaḥ śyenaśabdārthaḥ, īpsitatamo hy asau śyenaśabdena nirdiśyate / ataś cayanena śyeno nirvartayitavyaḥ, sa ākṛtivacanatve ‘vakalpyate / 254 It is worth noting that MīS 1.1.8 and 14, as well as Śabara’s Bhāṣya thereon, consider the use of karoti, “he makes,” to indicate that the object of the verb is produced, and therefore not eternal; in the case of śabdaṃ kuru, “make a sound,” where the object (sound) is eternal, the sentence has to be understood in the sense of śabdaprayogaṃ kuru, “make use of a sound.” 255 Cf. Tantravārttika p. 246: sā tv ākṛtyantaravilakṣaṇaśyenākṛtyā syān na tu vyaktyantaravilakṣaṇaśyenaviśeṣeṇāsādhāraṇena vā—“This [injunction (codanā)] ought to be [carried out] with a falcon form, as distinct from other forms, not with a particular falcon, as distinct from other individuals, or with something specific.” See also the passages from Kumārila quoted and translated in notes by Scharf (1996: 279 ff.). 75 76 Why? Because [only it can be] the object of an action.” The key expression in the sūtra and in the commentary is kriyārthatvāt, lit. “because [that] is the object of ‘doing.’” In the context of a sentence such as “He makes a mat,” this expression poses no problem. The object of the action “making” cannot be an individual, because the individual in question does not yet exist. We have already encountered this argument on numerous occasions. In the other interpretation, this expression cannot, as far as I know, be analyzed in any comprehensible way. Usually, individuals are the objects of actions, not forms. It is perhaps significant that Kumārila completely ignores the expression kriyārthatvāt in his commentary.256 The fact that Kumārila did not know, or no longer accepted, the sense that Śabara intended can, I believe, be explained without great difficulty. The problem of the arising of things constitutes—as we have already seen in other contexts—an episode in the history of Indian thought. For several centuries it attracted the attention of all thinkers, leading to many attempted solutions. Afterwards, the problem disappeared and was given hardly any thought. Confronted with vestiges of the problem in the literature of their schools, later thinkers endowed them with new meanings. We have seen how commentators of the Sāṃkhya school found a novel interpretation for an expression of this type. Here we see Kumārila apparently misunderstanding a passage from the Śābara Bhāṣya. This does not mean that Kumārila was a fraud. Not being a historian, he was doubtless unprepared to find in the classic text of his school an argument that, in his day, had lost its value. He consequently sought—and found—an interpretation of Śabara’s words that seemed acceptable to him. 13. The Abhidharmakośa Bhāṣya of Vasubandhu Let us now turn our attention to the Abhidharmakośa Bhāṣya, a work that probably dates to the very beginning of the fifth century,257 written by an author known as Vasubandhu. The work presents the viewpoint of the Sautrāntikas (or one similar to theirs), a viewpoint most often defined in opposition to that of the Sarvāstivādins. We have already noted that the central doctrine of Sarvāstivāda— according to which the past and the future exist as well as the present—made the system virtually impervious to Nāgārjuna’s critiques. Vasubandhu’s Sautrāntika, by contrast, does not accept this doctrine and is consequently open to attack. One can therefore expect to find other strategies of defense in this work. Consider, first, a passage in which Vasubandhu rejects the existence of three dharmas the Sarvāstivādins had included among the cittaviprayukta saṃskāra, or 256 This interpretation of Śabara’s words involves another difficulty, too, which has been studied by Peter M. Scharf (1993). The word “falcon” in the example “A falcon ought to be produced through piling” (Śabara’s paraphrase of the Vedic saying) cannot, according to Scharf, designate a universal (Scharf uses the term “class property”) without designating an individual. See also Scharf, 1996: 218, 277 ff. Scharf fails to recognize, however, that in Mīmāṃsā a word cannot designate two things at once (see Bronkhorst, 1997a). He is also unaware of the interpretation we consider the original interpretation, and is thus led to disapprove of Śabara’s argument. 257 See Schmithausen, 1992. Eli Franco draws my attention to the fact that, in the light of Schmithausen’s observations, Frauwallner’s dating of the Younger Vasubandhu (i.e., the author of the Abhidharmakośa Bhāṣya), namely 400-480, has to be advanced by at least some seventy or eighty years. 76 77 “formations not associated with mind.”258 The three are nāmakāya, padakāya, and vyañjanakāya, which represent (at least according to the definitions given in the Abhidharmakośa Bhāṣya) words, sentences, and phonemes, respectively. These three dharmas are distinct from voice (vāc), hence also from sound (śabda) and vocal sound (ghoṣa), none of which has the ability to produce an awareness of a designated object; it is nāman, or nāmakāya, that does this. Vasubandhu, as we just noted, rejects the existence of these three dharmas. For him, there is no need to posit such entities in addition to vocal sounds, which would be required, in any case, to produce (utpādayati) or manifest (prakāśayati) them. The Sarvāstivādin retorts that word (nāman) arises together with its object (arthasahaja). To this Vasubandhu replies: if that were the case, there could be no present word for a past or future object.259 This reply obviously indicates that for Vasubandhu there do exist present words for past and future objects. Concretely speaking, in a sentence such as “the pot arises,” one can take the word “pot” to refer to a future object that is in the process of arising. From this point of view, the sentence “the pot arises” no longer presents a problem. We encountered this position earlier in the Nyāya Sūtra and Bhāṣya, where we noted that it resembles the Sarvāstivāda position, but without having to suppose the existence of past and future objects, as Sarvāstivāda does. Elsewhere, the Abhidharmakośa Bhāṣya includes a debate with a grammarian (at least according to Yaśomitra’s commentary) over the interpretation of the expression pratītyasamutpāda, “dependent arising,” or, according to the analysis given by Vasubandhu, “appearance having attained.” The grammarian objects to this interpretation, and Vasubandhu responds: This interpretation is not possible; the word pratītyasamutpāda is no good, says the 260 grammarian. Indeed, when two actions have a single agent, the first action is 261 marked off by the gerund: snātvā bhuṅkte = “having bathed, he eats.” But there exists nothing that, before its arising, having first attained, then arises. And an action without an agent does not exist either. The grammarian’s objection is without value. Let us ask him whether that which arises is present or future: “Do you say that a present thing arises? If it is not already arisen, how is it present? If it is already arisen, how could it arise again without having to do so indefinitely? Do you say that a future thing arises? How can you attribute agency, in this action of arising, to that which is future and non-existent? Or how can you admit an action without an agent?” The grammarian’s theory, and the way that he opposes agent and action, none of that holds. For him, there is an agent here who is “the one who arises,” and an action here that is the action of arising. But one does not observe the action of “arising” as distinct from “the one who arises.” The passage concludes with the words: tasmād acchalaṃ vyavahāreṣu: “There is thus no deception in conventional 262 expressions.” 258 Abhidh-k-bh(P) p. 80 l. 12 ff.; cf. Abhidh-k(VP) vol. I p. 238 ff. Abhidh-k-bh(P) p. 81 l. 19: evaṃ saty atītānāgatasyārthasya vartamānaṃ nāma na syāt; cf. Abhidhk(VP) vol I p. 242. 260 Vasubandhu uses the word śābdika in this context; Yaśomitra’s commentary uses vaiyākaraṇa (Abhidh-k-bh(D) p. 454 l. 17). 261 According to P. 3.4.21: samānakartṛkayoḥ pūrvakāle. 262 Abhidh-k-bh(P) p. 138 l. 4 ff.: na yukta eṣa padārthaḥ / kiṃ kāraṇam / ekasya hi kartur dvayoḥ kriyayoḥ pūrvakālāyāṃ kriyāyāṃ ktvāvidhir bhavati / tad yathā snātvā bhuṅkta iti / na cāsau pūrvam utpādāt kaścid asti yaḥ pūrvaṃ pratītyottarakālam utpadyate / na cāpy akartṛkāsti kriyeti / … / naiṣa doṣaḥ / idaṃ tāvad 259 77 78 The final sentence is important: “There is no deception in conventional expressions.” The line reads as if aimed directly at Nāgārjuna’s arguments, which seek to show precisely that conventional expressions are deceptive. The way this passage tries to defend the reliability of conventional expressions is somewhat technical: it operates by rejecting the distinction between agent and action in the case of verbs such as bhavati, “it becomes,” and utpadyate, “it arises.” Vasubandhu rejects the idea of an action existing alongside an agent. His position is in keeping with the Buddhist—or rather Abhidharmic—vision of things according to which reality is constituted of a series of momentary elements, the dharmas, which exhaust the realm of what exists. Most of the objects that inhabit the world of our daily experience do not really exist. Actions, for instance, do not exist. There are only successions of dharmas, the locations of which can sometimes be slightly set off from one another, thus giving the impression of movement. Thus, no verb can refer to an object in the world, and most nouns are unable to do so either. Verbs as well as nouns can at most refer to dharmas. And in the sentence “such and such dharma arises,” the part about arising must also refer to the dharma in question. Note that such sentences are different from “the pot arises,” insofar as the word “pot” does not refer to any corresponding object; such an object does not exist in the ontology of Abhidharma. A pot is only a collection of dharmas, and the discussion must necessarily focus on these dharmas. Interestingly, the kind of argument Vasubandhu employs here is similar to arguments Nāgārjuna had used in support of his view of the world. Both raise the question of whether, in the sentence “such and such dharma arises,” the dharma is present or future. For both thinkers, the question is used to demonstrate the inadequacy of an opponent’s position. For Vasubandhu, the opponent here is a grammarian, according to the commentator; Vasubandhu himself refers to him as a śābdika, or “one who holds to words.” The śābdika’s position does not stand because it can give no answer to the questions raised. Vasubandhu thus recognizes the relevance of these questions and the contradictions to which they give rise at the level of phenomenal reality. The solution he proposes is situated at a different level, however, a level at which the action and the object it brings about are not distinct from one another. As soon as one grants that subject and verb refer to the same object, the problem is resolved. Once again, then, we have a solution that resorts to a novel way of understanding the denotation of words. In the other cases we have seen, thinkers had examined and adapted the way that nouns refer to objects; here we are interested in verbs. Vasubandhu thus does not abandon the correspondence principle, at least not at the level of phenomenal reality. We have already noted that for those unwilling to abandon the principle, the search for a ayaṃ praṣṭavyaḥ śābdikaḥ / kimavastho dharmaḥ utpadyate vartamāna utāho ‘nāgata iti / kiṃ cātaḥ / yadi vartamāna utpadyate / kathaṃ vartamāno yadi notpannaḥ / utpannasya vā punar utpattāv anavasthāprasaṅgaḥ / athānāgata utpadyate katham asataḥ kartṛtvaṃ sidhyati (the edition has siddhaty) akartṛkā vā kriyeti / … aniṣpannaṃ cedaṃ yad uta śābdikīyaṃ kartṛkriyāvyavasthānaṃ bhavatīty eṣa kartā bhūtir ity eṣā kriyā / na cātra bhavitur arthāt bhūtim anyāṃ kriyāṃ paśyāmaḥ / tasmād acchalaṃ vyavahāreṣu /. Translation based on Abhidh-k(VP) vol. II pp. 78-79. As La Vallée Poussin’s French translation was made on the basis of Tibetan and Chinese translations of the text, before the original Sanskrit was accessible, it renders the Sanskrit only approximately; nonetheless, it has been followed as far as possible here. On this passage, see also Radicchi, 1997: 400-01. 78 79 solution to the ensuing problems must inevitably take account of the question of the denotation of words. Here, then, are some responses of Buddhist Abhidharma to the problems posed by Nāgārjuna. For some Ābhidharmikas—in particular the Sarvāstivādins— there is no problem at all, since things exist even before they arise. For the rest, who reject the existence of past and future things, the words of such problematic sentences as “the pot arises” refer only to dharmas. At the level of the dharmas, the problem disappears, since at that level there are neither actions nor macroscopic objects. Does this mean that these Buddhists were uninterested in universals, which play such a decisive role in this context in Brahmanical schools? Certain passages from Vasubandhu’s Abhidharmakośa Bhāṣya—which takes the position of the Sautrāntikas and critiques the views of the Sarvāstivādins—as well as from the *Nyāyānusāra and the *Abhidharmasamayapradīpikā of his Sarvāstivādin rival Saṅgabhadra (the two works are preserved in Chinese translation) show that such was not the case. I have in mind the discussion of the dharma named sabhāgatā, or “homogeneous character.” This discussion shows that the notion of universals as an object of denotation proved attractive to Buddhists of the period, even if, unlike the Vaiśeṣikas, they refused to grant that a single universal subsists in all of its individuals. Saṅgabhadra—who clearly follows the earlier Sarvāstivādins critiqued by Vasubandhu—admits that the dharma sabhāgatā is not so different from the Vaiśeṣika categories of “universal” (sāmānya) and “specific universal” (sāmānyaviśeṣa). The similarity of “homogeneous character” to these Vaiśeṣika categories is indeed one of the objections Vasubandhu raises against accepting this dharma. Saṅgabhadra replies that he would adopt the Vaiśeṣikas’ position if they would grant that “universals” and “specific universals” are not singular, that they are momentary and impermanent, that they have no support, and that they are distinguishable from the objects to which they relate.263 In other words, a dharma named “homogeneous character” accompanies every object belonging to the category in question, one dharma for each object; like all other dharmas, this dharma is momentary and independent—i.e., it is a substance (dravya), as the texts put it. “Homogeneous character” is the cause of the mutual resemblance among things belonging to the same category, which makes possible the notion (buddhi) and designation (prajñapti) of the things in question.264 Note here that the role attributed to the dharma of “homogeneous character” was not, even in the Sarvāstivāda school, its original role. As the American scholar Collett Cox remarks, early texts of the school present this dharma as a factor determining the exact conditions for rebirth after living beings die.265 The dharma thus belongs primarily to living beings, not to inanimate objects. Vasubandhu, however, distinguishes between the “homogeneous characters” of living beings (sattvasabhāgatā) and those of the skandha, āyatana, and dhātu, which is to say, all the dharmas. Without homogeneous character belonging to living beings, the notion and designation of this or that living being would not be possible. Likewise, the notion and designation of the dharmas would not be possible without the 263 *Nyāyānusāra (TI 1562, vol. 29) p. 400c l. 3-7; see Abhidh-k-bh(P) p. 67 l. 25 ff.; Cox 1995: 234. *Nyāyānusāra (TI 1562, vol. 29) p. 400b l. 28-29; Abhidh-k-bh(P) p. 67 l. 20; Cox, 1995: 233. 265 Cox, 1995: 107. 264 79 80 homogeneous characters that pertain to them. This is the position Vasubandhu attributes to the Sarvāstivādins, even if his sources are unknown to us; Saṅgabhadra seems to restrict “homogeneous character” to the dharmas that relate to living beings.266 One thing emerges clearly from these discussions and differences of opinion: the Sarvāstivāda school, beginning at a time it is difficult to date precisely, tended to reinterpret one of its dharmas—the dharma of “homogeneous character” (sabhāgatā)—in such a way that it came to resemble the universal (sāmānya) or genus (jāti) of Brahmanical thinkers. This dharma explains, among other things, the designation of distinct objects by a common name. It is in virtue of this dharma that the word “Brahmin,” to take an example, does not refer to just one specific Brahmin but also to every other Brahmin, or in other words, to everyone characterized by the “homogeneous character” of Brahmins. Note, however, that as far as problems connected with the correspondence principle are concerned, the dharma of “homogeneous character” is incapable of playing the same role that universals do for the Naiyāyikas. For them, the word “Brahmin,” in a sentence such as “the Brahmin is born,” refers to the universal characterizing all Brahmins, a universal that is eternal and omnipresent. For the Sarvāstivādins, however, the word “Brahmin” in this sentence refers to the individual who is going to be born, and not to anything supra-individual. The dharma sabhāgatā therefore does not solve the problems raised by Nāgārjuna; but we have already seen that the Sarvāstivādins were in no way threatened by his arguments. They felt no need to find solutions to problems that were, for them, ultimately non-existent. Let us note further that this development within the Sarvāstivāda school would imply that the phenomenal world is not altogether determined by language, as earlier Buddhists had thought. The chariot of King Milinda was only a name; the words of language merely superimpose forms and a structure on a reality that does not really have them. The phenomenal world is thus created by language. For Sarvāstivādins such as Saṅgabhadra, however, the forms and structure of at least one part of the phenomenal world do not result from linguistic superimposition. One does not believe in the existence of Brahmins or other living beings simply because of words such as “Brahmin.” Even apart from words, Brahmins constitute a distinct group, for they possess an objective, non-linguistic trait, namely, the sabhāgatā of Brahmins, which distinguishes them from others. This sabhāgatā is what makes it possible for us to use the word “Brahmin.” This is to say that a Brahmin, in contrast to Milinda’s chariot, is more than just a word. The phenomenal world is no longer altogether determined by language. On the contrary, it is objectively such as to allow for the use of certain words. 14. The Abhidharmasamuccaya of Asaṅga and Its Bhāṣya Let us examine another system of Abhidharma, that of the Yogācāra school. In fact, we cannot be certain that this system really differs from that of Vasubandhu, the author of the Abhidharmakośa Bhāṣya, who is more commonly considered a Sautrāntika. Vasubandhu’s views so often coincide with Yogācāra Abhidharma, 266 Cox, 1995: 111. 80 81 even on points where the latter differs from Sautrāntika, that the theory has been advanced—provisionally, it is true—that Vasubandhu was already a Yogācāra when he wrote the Abhidharmakośa Bhāṣya.267 Whatever the case may be, the central text of Yogācāra Abhidharma is the Abhidharmasamuccaya, attributed to Asaṅga, and its Bhāṣya, the authorship of which is unknown, but which might be the work of Sthiramati. These texts include a few passages that seem to reveal an awareness of the problem of arising. They occur in the discussion of the profundities (gāmbhīrya) of dependent arising (pratītyasamutpāda), in the midst of which comes the profundity of arising (utpattigāmbhīrya). On this subject the Abhidharmasamuccaya Bhāṣya remarks: “The profundity of arising [must be understood] with two meanings (namely, the meaning of dependence [paratantrārtha] and the meaning of inactivity [nirīhakārtha]),268 inasmuch as effects, though arisen on the basis of conditions, are not made by them.”269 A few lines later, the same text asserts: “A dharma is not made by itself, because that by which the [dharma] would be made, [being itself] unarisen, does not exist; it is not made by another, because the conditions [on the basis of which it arises] are not agents; it is not made by both, for the same two reasons; it does not arise by its own activity or by the activity of another, nor without a cause, on account of the efficacy of causal conditions with respect to the arising of effects.”270 The problem of arising, if I understand these passages correctly, thus becomes a mystery—the text speaks of “profundity”—something that defies comprehension. How so? Consider the situation described by the sentence “conditions produce the dharma.” This situation goes against the correspondence principle, insofar as it does not contain elements corresponding to every key word in the sentence. When the conditions are there, the dharma is not present, and vice versa. The problem is solved by postulating that the conditions are not agents, that they produce neither dharma nor anything else. The sentence “conditions produce the dharma” therefore does not hold and must be abandoned. The end of the second passage suggests, however, that conditions are “efficacious with respect to the arising of effects.” In other words, the sentence “conditions give rise to effects” (hetupratyayāḥ phalāni utpādayanti) would be correct. In this case the Yogācāras seem to have been willing to tolerate a contradiction with the correspondence principle. But to tolerate such a contradiction, even in the name of a “profundity,” amounts to admitting that the principle is not always valid. 15. Bhartṛhari An important thinker has not yet figured in our investigations. Bhartṛhari, according to a few verses added at the end of the second book of his principal work, 267 Kritzer, 1999: 19 ff.; 193 ff.; 2005. The words in parentheses follow Kritzer. 269 Abhidh-sam-bh p. 34 (§ 40 H (1) (iii)): utpattigāmbhīryaṃ pratyayebhyaḥ phalotpattāv apy atatkṛtatayā dvābhyām arthābhyām. Cf. Kritzer, 1999: 58. 270 Abhidh-sam-bh p. 34 (§ 40 H (3) (i)-(iv)): na svayaṃkṛto dharmo ‘nutpannasyābhāvād yenāsau kriyeta / na parakṛtaḥ pratyayānām akartṛkatvāt / nobhayakṛta etenaiva kāraṇadvayena / na svayaṃkāraparakārāhetusamutpannaḥ hetupratyayānāṃ phalotpattau sāmarthyāt /. This reading follows Kritzer, 1999: 60 n. 134. 268 81 82 the Vākyapadīya (“Treatise on Sentences and Words”), played an important role in the resurgence of the Mahābhāṣya in the fifth century of the common era. The Vākyapadīya was aware of the problems relating to the correspondence principle, particularly the problem of the impossibility of the arising of things. Bhartṛhari proposes several solutions. We should not be surprised to find him making greater use, in comparison with other thinkers we have discussed, of the ideas found in the Mahābhāṣya. A verse from the third book of the Vākyapadīya articulates the central problem as follows: “What we call origination is the fact of attaining one’s own nature, and only something existent attains what is to be attained. If [this thing] exists [already], why does it arise? But if it does not exist, how does it arise?”271 The problem is easily recognizable: for something to arise, it has to exist; but if it already exists, why would it arise? To quote once again the words of Nāgārjuna: “If there existed anywhere something unarisen, it could arise. Since no such thing exists, what is it that arises?”272 In my opinion, this articulation of a problem first identified by Nāgārjuna— as we shall see, other passages reveal a familiarity with him as well—indicates that Bhartṛhari was influenced, whether directly or indirectly, by this Buddhist thinker.273 Bhartṛhari is aware of the problem Nāgārjuna had raised, and he responds with several solutions, all of them taking as their starting point various reflections on the true referent of a word. At the beginning of the first chapter of the third book of the Vākyapadīya (the “Jātisamuddeśa”), Bhartṛhari mentions two possibilities in this regard. According to the first, the referent of a word is a universal (jāti); according to the second, it is an individual (dravya). Bhartṛhari probably borrowed this pair from the grammarian Patañjali, substituting the word jāti, or “universal,” for Patañjali’s ākṛti, “form.” He retains the word dravya, profiting from its ambiguity. Though Patañjali employed it primarily in the sense of “individual,” Bhartṛhari prefers the sense of “substance.” In typical fashion, Bhartṛhari does not tell us which of the two, universal or substance, is the true referent of a word. He is content to discuss both, showing that his essential point concerning the nature of ultimate reality is unharmed in either case. He also shows that in both cases words refer to something immutable. Let us look briefly at his use of universals to solve the problem of arising. If the word “pot” does not refer to an individual in the process of arising, but rather to the universal that existed already before the arising of the particular pot, the sentence “the pot arises” no longer poses a problem. The correspondence principle is thereby safeguarded, provided of course that we accept a universal as the referent of the word “pot.” Bhartṛhari goes so far as to present the universal as an element that contributes to the arising of things, in our case, the pot. He describes the process in a series of verses worth citing: “Nothing arises that has no universal; the universal incites the causes to manifest itself. Universals, having entered into 271 Vkp 3.3.43: ātmalābhasya janmākhyā satā labhyaṃ ca labhyate / yadi saj jāyate kasmād athāsaj jāyate katham // 272 MadhK(deJ) 7.17: yadi kaścid anutpanno bhāvaḥ saṃvidyate kvacit / utpadyeta sa kiṃ tasmin bhāve utpadyate ‘sati // 273 There are, of course, other indications, which I have discussed elsewhere (1992a). Some portions of what follows have been drawn from this article. 82 83 the causes, whether eternal or non-eternal, manifest themselves again and again somewhere in the effects. The universal is also the means of realization (sādhana) for a grammatical object (karman) that is being produced; it incites the action to bring about the object in which it resides.”274 In the case of the sentence “he makes a pot,” where the word “pot” is the grammatical object (karman), the word “pot” refers to the universal “pot,” which already exists; this universal, for its part, incites the causes, i.e., the elements that will make up the pot or contribute to its arising, to manifest the pot.275 The universal, as another verse from the same chapter adds, is eternal (3.1.34). The second chapter of the third book of the Vākyapadīya, the “Dravyasamuddeśa,” adopts the position that the referent of a word is a substance (dravya). Substance, we learn, does not change; only the forms associated with it change: “Gold, for example, is associated with transient forms, but terms such as “necklace,” etc., express only its pure state; the distinction introduced by these forms prevents [the terms “necklace,” etc.] from having all of these meanings. … [Likewise,] the word that expresses such forms expresses only what is eternal [in them], for it has the nature of essential reality.”276 Reality, says Bhartṛhari a few verses later, is not comprised of things that transform, even though it is such that it appears to be so.277 The chapter concludes with the assertion that reality, which is without origin, without before and after, is nonetheless perceived as if it did have an origin, etc., even though these things are contradictory: “Likewise, one perceives that reality, which is [nevertheless] without origin, eternal, and devoid of prior and posterior, has an origin, etc., although this is contradictory.”278 There is no concrete allusion to the problem of the arising of things in the “Dravyasamuddeśa,” for the position adopted here clearly shows that it is not a problem for Bhartṛhari, even when the referent of a word is accepted as being a substance. On this view, the word “pot” in the sentence “the pot arises” refers to the substance that will make up the pot. This substance exists prior to the arising of the pot, and the sentence is thus intelligible. Once again, there is no need to abandon the correspondence principle. In the third chapter of the third book of the Vākyapadīya, one finds the verse we looked at earlier, in which the problem of the arising of things is stated directly: if, during the arising of a thing, the thing already exists, why does it arise? And if it does not exist, how does it arise? This chapter is called the 274 Vkp 3.1.25-27: na tad utpadyate kiṃcid yasya jātir na vidyate / ātmābhivyaktaye jātiḥ kāraṇānāṃ prayojikā // kāraṇeṣu padaṃ kṛtvā nityānityeṣu jātayaḥ / kvacit kāryeṣv abhivyaktim upayānti punaḥ punaḥ // nirvartyamānaṃ yat karma jātis tatrāpi sādhanam / svāśrayasyābhiniṣpattyai sā kriyāyāḥ prayojikā // 275 Bhartṛhari thus attributes to the universal an activity of inciting causes, which is totally absent from Vaiśeṣika texts. He was not alone, however, in modifying the role of universals. The Carakasaṃhitā, Sūtrasthāna 1.44-45 does the same: see Comba, 1990. One might wonder to what degree this more dynamic understanding of universals was inspired by a rather strange consequence (emphasized by Gerdi Gerschheimer) of the claim that the word “pot,” in the sentence “the pot arises,” refers to an eternal universal: namely, that the sentence no longer describes a case of arising. 276 Vkp 3.2.4-6: suvarṇādi yathā yuktaṃ svair ākārair apāyibhiḥ / rucakādyabhidhānānāṃ śuddham evaiti vācyatām // ākāraiś ca vyavacchedāt sārvārthyam avarudhyate / … // teṣv ākāreṣu yaḥ śabdas tathābhūteṣu vartate / tattvātmakatvāt tenāpi nityam evābhidhīyate // 277 Vkp 3.2.10: tathā vikārarūpāṇāṃ tattve ‘tyantam asaṃbhavaḥ / tadātmeva ca tat tattvam atyantam atadātmakam // 278 Vkp 3.2.18: ajanmani tathā nitye paurvāparyavivarjite / tattve janmādirūpatvaṃ viruddham upalabhyate // 83 84 “Saṃbandhasamuddeśa.” As the title indicates, it deals with the connection between words and their referents.279 The “Saṃbandhasamuddeśa” offers yet another solution to the problem raised in this verse. According to this solution, words do not refer to absolute reality, whether in the form of a universal or a substance, but to a metaphorical existence. One verse puts it thus: “In linguistic usage, it is another existence, of a metaphorical nature, that belongs to the objects of words; [this metaphorical existence] shows the form of all [things] in all states.”280 The expression “in all states” doubtless refers to past, present, and future. We can immediately see that if words refer to their objects in the future or the past, the central problem of the arising of things is no more. All one has to do is say that the word “pot,” in “the pot arises,” designates the future pot that does not yet exist, and one rids oneself of all contradictions. Having introduced the idea of metaphorical existence, Bhartṛhari goes on to show its usefulness in several verses. One of them is the verse mentioned already, which articulates the problem of the arising of things. Another verse explains that metaphorical existence is responsible for the fact that negations can be expressed.281 This claim should not be too surprising, for the correspondence principle obliges us to admit that in order for a sentence such as “Martians do not exist” to be true, there must be Martians. Metaphorical existence fulfills this requirement without completely rejecting the correspondence principle. Other verses in this section282 contrast sentences such as “the pot arises,” where the subject does not exist at the time of the action, with sentences such as “the traveler travels the path,” where the subject really does exist during the action described by the sentence. In these verses, Bhartṛhari deals with the problems raised by the correspondence principle in a more explicit manner than elsewhere in his Vākyapadīya. Bhartṛhari has one more solution to the problem of arising. It comes in the “Sādhanasamuddeśa,” which is the seventh chapter of the third book of the Vākyapadīya. The subject of this chapter is the “means of accomplishment.” Recall Bhartṛhari’s remark about universals: “The universal is also the means of realization (sādhana) for a grammatical object (karman) that is being produced.” In other words, in the sentence “he produces a sound,” the word “sound,” which is the grammatical object, refers to a universal, which is eternal. A verse from the “Sādhanasamuddeśa” offers a solution to the problem this sentence poses when one wants the word “sound” to refer to an individual: “On the assumption that the meaning of the word is the individual, it is established that the grammatical object that is being produced, for example sound, is the means (sādhana); this is done on the basis of mental form.”283 In other words, in the sentence “he produces a sound,” the word “sound” refers to a mental reality, which, unlike the sound itself, exists when the sentence is spoken. Elsewhere in the same chapter, another verse expresses the solution just as clearly: “Non-existence prior to arising is based on a 279 An in-depth discussion of this chapter can be found in Jan E. M. Houben’s book (1995). Vkp 3.3.39: vyapadeśe padārthānām anyā sattaupacārikī / sarvāvasthāsu sarveṣām ātmarūpasya darśikā // 281 Vkp 3.3.42. 282 Vkp 3.3.44-46. 283 Vkp 3.7.7: vyaktau padārthe śabdāder janyamānasya karmaṇaḥ / sādhanatvaṃ tathā siddhaṃ buddhirūpaprakalpitam // 280 84 85 condition of the mind (buddhi). [The subject] (“pot” in the sentence “the pot arises”) becomes the agent of its arising, just as another subject, which [really] exists, [is the agent of an action].”284 It is possible, but not necessarily the case, that in Bhartṛhari’s view this last solution, which appeals to a mental reality corresponding to words, does not differ from the solution according to which words refer to a metaphorical reality. This question need not concern us here. What matters instead is that Bhartṛhari finally offers what might be considered the most obvious solution to the problem. Could not all of the questions surrounding simple sentences such as “the pot arises” have been resolved straightaway by maintaining that the pot, in order to be part of such a statement, must be an idea in our heads, and not something existing in the external world? Why does this solution not appear in the other texts we have studied?285 The question becomes even more intriguing when we realize that Patañjali, author of the Great Commentary on Pāṇini’s grammar, had already introduced the notion of ideas in his discussion of grammatical procedures. How, he asks for example, can one replace the element as with the element bhū, as Pāṇini’s grammar prescribes? The answer is simple, but significant: the idea of as is replaced by the idea of bhū.286 Patañjali predates Nāgārjuna and the problem that concerns us here. But he was probably well known to all of the authors we have discussed. Why did they not expand on the notion of ideas corresponding to words? I can offer only one hypothesis that might answer the question. The notion of ideas as the primary objects of words began to play a role in Buddhism even in its early period. We know that for many Buddhists, the objects of the phenomenal world had no real existence and were but names and designations. We have already pointed out that the term prajñapti, which means, precisely, “verbal designation,” was frequently employed in this context. Alongside prajñapti, however, another term acquired great importance for certain Buddhists: vijñapti. The term, which might be translated as “information,”287 refers to a mental entity. What is more, the schools that used the term tended to develop toward a form of idealism. Most of the non-Buddhist schools we have studied, by contrast, reject all idealist tendencies. Is it possible that this rejection of Buddhism, and of idealism in particular, hindered the acceptance, within Brahmanical schools of thought, of words as designating or corresponding to ideas? A more in-depth study might one day yield a satisfactory answer to this question, which falls outside the scope of these lectures. We can simply note that the notion of the referent as a mental entity seems to have appeared in non-Buddhist schools at a relatively late date. Let us return to Bhartṛhari, who, we have seen, offers several solutions to the problem of the arising of things, without abandoning the correspondence principle.288 Sentences such as “the pot arises” and “he produces a sound” are 284 Vkp 3.7.105: utpatteḥ prāg asadbhāvo buddhyavasthānibandhanaḥ / aviśiṣṭaḥ satānyena kartā bhavati janmanaḥ // 285 NS 4.1.49 (buddhisiddhaṃ tu tad asat; discussed above, § II.11) is of course a noteworthy exception. 286 Mbh I p. 137 l . 24 (on P. 1.1.56 vt. 14): … astibuddhyā bhavatibuddhiṃ pratipadyate. 287 See May, 1959: 263, with n. 952. 288 Jan Houben, in a personal communication, wonders whether the priority accorded to the sentence (to the detriment of word) in Bhartṛhari’s thought might be interpreted as yet another 85 86 possible because the words “pot” and “sound” refer to something even when the pot and the sound do not yet exist. All of his solutions are variants of the position that words do not necessarily refer to corresponding individuals, or that the individual corresponding to a word might exist only as a mental reality. It is also important to recognize that Bhartṛhari did not take full advantage of these solutions to free himself from Nāgārjuna’s arguments. He clings to the idea that behind appearances there exists a reality that does not change. Bhartṛhari could have preserved the world of appearances by appealing to his new proposals on the referents of words, but he does not do so. On the contrary, he patently asserts that causality and arising are not possible. By way of illustration, let us take a brief look at the second, concluding section of the “Saṃbandhasamuddeśa.” We have already quoted a few verses from the first section of this chapter; the second section covers verses 52 to 88, which consider the connection between words and things from the point of view of absolute reality.289 Several of these verses bear directly on the problem of the arising of things. Bhartṛhari remarks from the outset that usage, in his words, conforms to only one aspect of a part of reality, or to an elucidation on the basis of something else, or to an error, or even to non-existence.290 In other words, the connection between words and things is far from perfect. The following verses go further, comparing the idea to which a statement gives rise, on the one hand, to the cognition that results from a defective sense organ, on the other.291 A word expresses an object, not insofar as the object is determined by its own form, but by another form.292 Other verses observe that a non-existent thing does not arise as something existent, and vice versa. The cause of a non-existent thing is inefficacious, precisely because the thing does not exist. The cause of an existent thing, on the other hand, has no function, because the thing already exists.293 One important verse claims that the division into the three times (past, present, future) is not possible in non-existence; this being the case, the three times do not reside in existence either.294 Still other verses address causality and the unreality of the world that language presents. One can say in general that this section never strays far from the question of the impossibility of arising and causality. I have drawn your attention to this section for another reason as well: the great Buddhist logician Dignāga borrowed it virtually in toto, compiling a work that has survived only in (poor) Tibetan translation, which bears the name Traikālyaparīkṣā (or Trikālaparīkṣā), “Examination of the Three Times.” The work includes one original verse, all but seven of the verses from the present section of solution to the problems raised by the correspondence principle. This suggestion merits further consideration, but must wait for another occasion. 289 According to Houben, 1995: 142. 290 Vkp 3.3.52: pradeśasyaikadeśaṃ vā parato vā nirūpaṇam / viparyayam abhāvaṃ vā vyavahāro ‘nuvartate //. The meaning of parato … nirūpaṇam is obscure; Rau (2002: 184) translates “Festlegung durch ein Zweites”. 291 Vkp 3.3.53: yathendriyasya vaiguṇyān mātrādhyāropavān iva / jāyate pratyayo ‘rthebhyas tathaivoddeśajā matiḥ // 292 Vkp 3.3.54cd: artham āhānyarūpeṇa svarūpeṇānirūpitam. 293 Vkp 3.3.61ab, 62: nābhāvo jāyate bhāvo naiti bhāvo ‘nupākhyatām / … // abhāvasyānupākhyatvāt kāraṇaṃ na prasādhakam / sopākhyasya tu bhāvasya kāraṇaṃ kiṃ kariṣyati // 294 Vkp 3.3.69: abhāve triṣu kāleṣu na bhedasyāsti saṃbhavaḥ / tasminn asati bhāve ‘pi traikālyaṃ nāvatiṣṭhate // 86 87 the Vākyapadīya (sometimes in slightly modified form), two verses also found in the Vṛtti of the Vākyapadīya, and nothing else.295 It seems likely, as Frauwallner maintains, that the Traikālyaparīkṣā was a work of Dignāga’s youth. It shows in any case that Dignāga was concerned, perhaps at the start of his career, with the problems Bhartṛhari sought to resolve, and that he accepted, at least at one stage in his life, the solutions the latter offered, with some minor changes. More specifically, it seems certain that Dignāga accepted, at that time, Nāgārjuna’s arguments about the impossibility of causality and arising. We shall soon look at how Dignāga reacts to Nāgārjuna’s arguments in his other works. For now, a few further points remain. Bhartṛhari, as we have seen, upholds the correspondence principle, while suggesting various possibilities as to the objects that correspond to words. But the correspondence principle, even in its strictest form, leaves some things unclear. It is easy to say that the objects of the phenomenal world correspond to words, but to which words exactly? There are objects that correspond to nouns, to adjectives, and even to verbs, but what about other types of words? Nāgārjuna and other thinkers had not addressed this point. Bhartṛhari raises the question and arrives at the conclusion that certain types of words are not expressive (vācaka), but rather suggestive (dyotaka). Such words suggest a meaning that is expressed instead by the verb or the noun that accompanies them. The idea that certain words suggest rather than express meanings was not new. Yāska’s Nirukta, which predates the beginning of the common era, attributes to a certain Śākaṭāyana the view that verbal prefixes (upasarga) do not express any meaning; rather, they suggest a modification of the meaning of the noun or verb.296 The author of the Saṅgraha—who must have preceded Bhartṛhari—apparently made the same point.297 Bhartṛhari takes up the position of these two authors, and extends it to cover not only verbal prefixes but also certain particles (nipāta).298 The advantage of this position seems to consist in obviating the need to find objects in the world corresponding to particles such as “and” and “or.” One will note that Bhartṛhari nevertheless seems to admit that certain particles do have meaning. The question merits an in-depth study, which we cannot take up at this time. We have seen that among Bhartṛhari’s three or four solutions to the problems raised by the correspondence principle, one of them proposes that words refer to a metaphorical reality. If it is possible that Bhartṛhari was the first to present this solution, he was certainly not the last. Consider the Vibhāṣāprabhāvṛtti, a commentary on the Abhidharmadīpa, a Buddhist text from the Sarvāstivāda school. It contains a passage in which the following position is attributed to the Vaiśeṣikas: The Vaiśeṣika reasons [as follows]: The substance “pot,” which is not present in its shards, and the substance “cloth,” which is not present in its 295 See Frauwallner, 1959; Houben, 1995: 273 ff. Nirukta 1.3: na nirbaddhā upasargā arthān nirāhur iti śākaṭāyanaḥ / nāmākhyātayos tu karmopasaṃyogadyotakā bhavanti / 297 The Vṛtti on Vkp 2.187 (ed. Iyer p. 234) says: śabdāntaropagraham antareṇa sambhavī sann alabdhaniyamo yo ‘rthas taṃ dyotako niyamayan vācakatām atikrāmatīti saṅgrahakāra āha. On the date of the Saṅgraha, cf. Bronkhorst, 1983: 394; Wujastyk, 1993: II: xxiv ff. 298 Vkp 2.192 ff. 296 87 88 threads, arise on account of the conjunction of the shards and of the threads [respectively]. And through a secondary thought [gauṇyā kalpanayā], one speaks of the existence of the agent of arising, [an existence] that has as its object a state [of the pot] that is opposed [to the 299 present]. The pot therefore exists, in virtue of a “secondary thought,” before its arising. The text gives no further details. The position here attributed to Vaiśeṣika is confirmed in the Vyomavatī of Vyomaśiva, a commentary on the Padārthadharmasaṅgraha, also known as the Praśastapādabhāṣya. In a critique of the Sāṃkhya position, Vyomaśiva makes the following observation: One should consider the designation [“sprout” in] “the sprout arises,” as well as [“pot” in] “make a pot,” to be metaphorical, because there exists an obstacle to primary [denotation]. To explain: if one regarded the already existent pot as the object [of the act of making], and the [already existent] sprout as the agent [of the act of arising], arising would be contradicted, and the action of the actants (kāraka) would be useless, because [the pot and the sprout] would already exist. … It follows that, as something existent cannot be the thing that arises, the sprout is metaphorically the agent of its own arising, and [the pot] the object [of the act of making]. In 300 this way [these] designations are differently established. Can we conclude that the Vibhāṣāprabhāvṛtti knew the Vyomavatī, or an earlier text proposing the same explanation? The first option seems highly unlikely, given that the Vyomavatī dates to around 900 C.E.;301 the Vibhāṣāprabhāvṛtti, though of uncertain date,302 can hardly be so late. 16. The Problem of Negation Let us note that the question of the designation of particles, and of the negative particle na in particular, posed a problem not just for Bhartṛhari but for other thinkers as well. In my opinion, it would be safe to say that this very problem was what led Vaiśeṣika—doubtless under the influence of Nyāya—to introduce “absence” (abhāva) as a seventh category. Recall that the classical presentation of the system in Praśastapāda’s Padārthadharmasaṅgraha knew only six categories. 299 Abhidh-d on kārikā 310, p. 274 l. 5-7: vaiśeṣiko manyate: kapāleṣv avidyamānaṃ ghaṭadravyaṃ tantuṣu cāvidyamānaṃ paṭadravyaṃ kapālatantusaṃyogād utpadyate / gauṇyā ca kalpanayā viprakṛtāvasthāviṣayā janikartṛsattā vyapadiśyata iti /. The word viprakṛta is obscure. The editor, Padmanabh S. Jaini, suggests correcting it to viprakṛṣṭa, “removed, distant,” but this does not help. Apte’s dictionary gives, among other meanings for viprakṛta, the sense of “opposed,” which more or less seems to fit, both here and a few lines later, when the word is used a second time. 300 Vyomaśiva’s Vyomavatī, vol. 2 p. 129 l. 19-27: yac cāyaṃ vyapadeśo ‘ṅkuro jāyate, ghaṭaṃ kurv iti ayam api mukhye bādhakapramāṇasadbhāvād bhākto draṣṭavyaḥ / tathā hi, yadi vidyamānasyaiva ghaṭasya karmatvam aṅkurasya ca kartṛtvam iṣyetotpattir vyāhatā syāt, vidyamānatvād eva kārakavyāpāravaiyarthyañ ca / … / tasmād vidyamānasya utpattyarthāsambhavād upacaritam aṅkurādeḥ svajanikartṛtvaṃ karmatvañ ceti vyapadeśasyānyathāsiddhatvam / 301 Slaje, 1986: 245-46; 274. 302 Abhidh-d, Introduction, p. 129 ff. 88 89 Later Vaiśeṣika texts add absence as a seventh category. The author Śivāditya, for example, discusses it in a work titled, perhaps by way of a challenge, the Saptapadārthī, or “Treatise on the Seven Categories.” For practical reasons, the category of “absence” is divided into four sub-categories: prior absence (prāgabhāva), absence in destruction (pradhvaṃsābhāva), utter absence (atyantābhāva), and mutual absence (anyonyābhāva).303 Prior absence precedes the arising of an object; absence in destruction follows its passing away; utter absence comes into play when one says, for instance, “there is no knowledge in the pot”; mutual absence is illustrated in the sentence “the pot is not a cloth.” Some remarks made by the commentator Jinavardhana (15th c.) remind us of the close connection between the Vaiśeṣika categories and language. He considers an objection that mutual absence is superfluous, that the quality pṛthaktva, or “separation,” would suffice to explain the same situations. Jinavardhana disagrees. There is mutual absence in the case of a substance and a quality inhering therein, but the two are not separate, he notes. He adds that the quality pṛthaktva applies only in cases where one would use the word pṛthak, “separate,” as in the sentence “this is separate from that” (ayam asmāt pṛthak).304 In other words, absence corresponds to the particle na, while the quality of separation corresponds to pṛthak. The acceptance of absence as a category, and of prior absence (prāgabhāva) in particular, offers a further advantage to the Vaiśeṣika system. Recall the criticism to which thinkers of the school had to respond. Mallavādin had put it thus: “If the effect does not exist [in its causes], it does not arise, for no agent of this operation is present, as [in the case of] a flower in the sky. Or [contrariwise,] a flower in the sky would arise, since no agent of this operation is present, as [in the case of] an effect.” We know that the early Vaiśeṣikas sought to respond to this problem by postulating a kind of pre-existence of the object in the course of its arising; but how did later Vaiśeṣikas, who rejected this solution, explain the situation? It seems likely that for them, prior absence as a real entity would have proved extremely useful. It is a pot that will arise in a certain material, and not something else, because there is the prior absence of a pot in the material. From this point of view, an effect is characterized, or even defined, precisely by the fact of having a prior absence. This is exactly the definition that Śivāditya offers: “An effect is that which has a prior absence.”305 This definition is not as insipid as one might first think, for it establishes a relationship between two real entities. It moreover explains why a pot, and not just anything, arises in a given situation. Nevertheless, the acceptance of one or more kinds of absence corresponding to the negative particle does not solve every problem. One particularly delicate 303 Saptapadārthī 9 p. 10. Note that the Nyāya Bhāṣya on 2.2.12 (p. 587) distinguishes between just two kinds of absence: absence prior to origination, and absence after destruction (abhāvadvaitaṃ khalu bhavati prāk cotpatter avidyamānatā, utpannasya cātmano hānād avidyamānatā). 304 Saptapadārthī 9 p. 11: nanu pṛthaktvenaivāsya pratyayasya siddhatvāt kim anyonyābhāvakalpaneneti cet; na, guṇa-guṇinoḥ ‘guṇo guṇī na bhavati’ ‘guṇī guṇo na bhavati’ iti svarūpabhedenaiva parasparaṃ bhede anyonyābhāvasya sambhavāt / pṛthaktvaṃ tu pṛthag āśrayāśritayoḥ ‘ayam asmāt pṛthag’ iti pṛthagvyavahāre saty eva bhavati nānyathā, ato ‘sya kalpanā / 305 Saptapadārthī 179 p. 61: prāgabhāvavat kāryam. 89 90 problem arises in the case of existential negation.306 If one says that Martians do not exist, what is the object designated by the word “Martians”? The problem has a direct bearing on the debate between Buddhists and their critics over the existence of the soul. Buddhists deny the existence of the soul, but is it really possible, their critics ask, to say “the soul does not exist”? The problem is addressed in a passage from Uddyotakara’s Nyāyavārttika.307 Uddyotakara denies that one can say “the soul does not exist”: according to him, the sentence is self-contradictory. Here are a few relevant excerpts: Here, to begin with [one must say] that the two (!) words (pada) “the soul does not exist” (na asti ātmā) contradict [each other]: the word “soul” (ātmā), [even when placed] in syntactic agreement with the words (?) “does not exist” (na asti), does not communicate the non-existence of the soul. For with the word “soul,” one expresses [its] existence, and with “does not exist,” its negation; but the [very] thing denied in one place appears in the other. For example, the word “pot,” in syntactic agreement with “is not [there]” (na asti), cannot communicate the non-existence of the pot; rather, it denies [its presence] in a particular place or at a particular time. [The sentence] “the pot is not” (nāsti ghaṭaḥ) denies [its presence] in a particular place, [as in the example] “it is not in the house”; or it denies [its presence] at a particular time, [as in the examples] “it is not [there] now,” “it was not [there] before,” “it will not be [there] later.” None of these negations are possible for someone who does not accept the existence of the pot. Likewise, [in the sentence] “the soul does not exist,” is the soul denied with respect to a particular place or with respect to a 308 particular time? Uddyotakara goes on to show that the sentence “the soul does not exist” is not in fact meant to deny the presence of the soul with respect to a particular place or time, but rather to assert its simple negation, which is impossible, for: He who denies the soul [altogether] must specify the object of the word “soul.” 309 Indeed, we do not see a word without an object. Uddyotakara next attacks the most obvious solution to the problem: might not the word “soul” correspond to something imaginary, the objective existence of which is denied? Uddyotakara says no, showing by means of an analysis of imagination that such a solution is not possible.310 What if one says that some words—such as “void” 306 Existential negation was, and still remains, a difficulty for some Western philosophers as well. Brendan Gillon, in a personal communication, mentions in particular Peter Abelard, Alexius Meinong (cf. Passmore, 1966: 179 ff.), and the logical atomism of Russell and of Wittgenstein. On the latter two, and for a discussion of the “negative fact” and its importance for understanding certain aspects of Indian thought, see Gillon, 1997. 307 Cf. Oetke, 1988: 363 f. 308 Uddyotakara, Nyāyavārttika introducing 3.1.1 (p. 699 l. 1 ff.): tatra nāsty ātmeti pade tāvad vyāhanyete, nāstiśabdasamānādhikaraṇo ‘yam ātmaśabdo nātmano ‘sattvaṃ pratipādayati / kiṃ kāraṇam? ātmeti sattvam abhidhīyate, nāstīti tasya pratiṣedhaḥ, yac ca yatra pratiṣidhyate tat tasmād anyatrāsti, yathā nāstinā samānādhikaraṇo ghaṭaśabdo na ghaṭābhāvaṃ pratipādayituṃ śaknoti, api tu deśakālaviśeṣe pratiṣedhati / nāsti ghaṭa iti deśaviśeṣe vā pratiṣedho gehe nāstīti, kālaviśeṣe vā pratiṣedha idānīṃ nāstīti prāṅ nāsti ūrdhvaṃ nāstīti / sarvaś cāyaṃ pratiṣedho nānabhyupagataghaṭasattvasya yuktaḥ / tathā nāsty ātmeti kim ayaṃ deśaviśeṣe pratiṣidhyate atha kālaviśeṣe iti / 309 Ibid. p. 701 l. 5-6: ātmapratiṣedhaṃ ca kurvāṇenātmaśabdasya viṣayo vaktavyaḥ / na hy ekaṃ padaṃ nirarthakaṃ paśyāmaḥ / 310 Ibid. p. 701 l. 9 ff. 90 91 (śūnya) or “darkness” (tamas)—do not necessarily have an object? The answer is again no.311 Examples such as “hare’s horn” (śaśaviṣāṇa), “sky-flower” (khapuṣpa), etc., do not prove the contrary, for in such cases it is the connection between the things that is denied.312 Regardless of the exact arguments Uddyotakara uses to defend his position, what matters is his insistence on the fact that every word, including the word “soul” in the sentence “the soul does not exist,” must refer to an object. We thus have an instance of a sentence which, though correctly formed, is considered meaningless, and that by reason of adherence to the correspondence principle. 17. Dignāga and Verbal Knowledge Let us now turn to Dignāga. As we learned recently, in his youth Dignāga accepted Nāgārjuna’s arguments about the impossibility of causality and arising. This observation is worthy of our attention. As far as I know, no argument of this type appears in his other surviving works. Dignāga is one of the Buddhist authors who have led modern scholars to believe that Nāgārjuna was neglected, even ignored, by his later co-religionists. But the very existence of the Traikālyaparīkṣā seems to indicate that this great thinker was at one time deeply interested in this type of argument. What change in his thought could explain why all traces of Nāgārjuna and his arguments disappeared from his other works, as well as from those of his successors? Consider again the case of Bhartṛhari. As we noted earlier, Bhartṛhari had found an answer to the problems raised by Nāgārjuna: the words in a sentence do not necessarily refer to individuals. This simple assertion is enough to neutralize the most threatening arguments of Nāgārjuna. What then do words refer to? Bhartṛhari had proposed a number of possibilities. Two of them are worth recalling: universals and mental realities. Either of these potential referents would have sufficed to refute Nāgārjuna’s arguments. Bhartṛhari, however, did not use them in this way. What of Dignāga? Note that Dignāga could not accept universals as the referent of a word. Buddhist ontology simply does not leave room for such a category.313 Universals were the best candidate, however. With universals, one can easily explain why and how we are able to use a single word, such as “cow,” to refer to many different objects, in this case the set of all cows. Cows obviously have something in common, which we recognize and call a “universal.” If, on the other hand, one maintains that the word refers to a mental reality, it becomes more difficult to explain how the mental reality seems to cover the same set of individuals. Dignāga, I have just said, could not accept universals as the referent of a word. His approach was instead to create a concept that corresponds to universals without transgressing the bounds of Buddhist ontology. I am referring, as you have no doubt realized, to Dignāga’s great innovation of apoha or anyāpoha, “exclusion of other [things].” According to this theory, the referent of a word is a conceptually 311 Ibid. p. 702 l. 3 ff. Ibid. p. 703 l. 19 ff. 313 Prs 2.16, 5.17-18 + PrsV; Hayes, 1988: 183 ff., 246, 282 ff. 312 91 92 constructed object. This is the view attributed to Dignāga by his celebrated successor, Dharmakīrti,314 whom we have no reason to doubt on this score. This means that sentences such as “he makes a pot” do not pose a problem, since the relationship between the pot and the action of making exists only at the level of language. Nothing in reality corresponds to this relationship, and the problems ensuing from their relationship do not correspond to anything in reality either. Let us consider Dignāga’s theory of denotation more closely. It is presented in the fifth and final chapter of his Pramāṇasamuccaya.315 A universal term (jātiśabda), he says in the second verse, is not expressive (vācaka) of individuals. This observation is already enough to answer Nāgārjuna’s critiques. The word “pot” in “the pot arises” does not express the individual that arises, and thus does not presuppose its presence. This observation explains certain aspects of the behavior of words that go against the idea of a simple correspondence between words and things. For instance, it frequently happens that several words are used to designate a single thing, as with the words “tree” (vṛkṣa), “made of earth” (pārthiva), “substance” (dravya), and “existent” (sat).316 If these terms designated individuals, they would be synonyms. As they do not designate individuals, the problem does not arise. But if generic terms do not express individuals, what is their relation to things? Richard Hayes (1988: 257) underscores, with good reason I believe, the distinction Dignāga draws between “expressing” (vācaka) and “applying” (vṛtti/pravṛtti). Words do not express anything, but they are nevertheless applicable to objects. According to Dignāga, words refer to things by exclusion. They exclude everything that does not belong to their objects, but without expressing the objects. It is neither possible nor necessary to reproduce here the various considerations that conduce to showing that a discourse relying on exclusions can replace, in an adequate if unwieldy manner, a discourse based instead on universals as the referents of words.317 Dignāga emphasizes that the characteristics of universals, namely that they are one (eka), eternal (nitya), and subsistent in each individual (pratyekaparisamāpti), are present in exclusion.318 A discourse of this kind allows Dignāga, and other Buddhists with him, to benefit from the advantages that universals offered to the Naiyāyikas, but without having to accept their existence. What interests us here is that Dignāga rids himself of the threat of Nāgārjuna’s arguments without paying a price at the level of ontology. 314 Prv I.212: śabdārthaḥ kalpanājñānaviṣayatvena kalpitaḥ / dharmo vastvāśrayāsiddhir asyokto nyāyavādinā //. The lack of connection with an external object (according to Dharmakīrti) is underlined in Prv I.66: yad yathā vācakatvena vaktṛbhir viniyamyate / anapekṣitabāhyārthaṃ tat tathā vācakaṃ vacaḥ // 315 Edition in Hattori, 1982; partial translation in Hayes, 1988. A reconstruction of the original Sanskrit, using the newly recovered codex unicus of the original Sanskrit version of the commentary by Jinendrabuddhi, is being prepared by Ole H. Pind (Copenhagen). See http://ikga.oeaw.ac.at/Pr/Pr_dignaga.html. 316 Hattori, 1980: 62; Katsura, 1979. 317 See, e.g., Hayes, 1988: 188 ff. 318 PrsV 5.36d: sarvatrābhedād āśrayasyānucchedāt kṛtsnārthaparisamāpteś ca yathākramaṃ jātidharmā ekatvanityatvapratyekaparisamāptilakṣanā apoha evāvatiṣṭhante. Cf. Hattori, 1980: 62, 71 n. 5; 1982: 137 n. 34. 92 93 The theory of apoha resolves—perhaps one should say avoids—the problems Nāgārjuna had located within the phenomenal world, the world subject to and created by language. Let us not forget that the majority of Buddhists distinguish clearly between ultimate reality on the one hand and the deceptive phenomenal world on the other. Dignāga does the same, but he introduces an important innovation: for him, the distinction between reality and the unreality of the phenomenal world coincides with the distinction between objects of perception and objects of language. The two are completely separate, to the point that perception cannot be verbalized. Perception bears on the particular (svalakṣaṇa), language on the general (sāmānyalakṣaṇa). Only the objects of our perception are real; the realm of language, for its part, is not. Although we have seen that the realm of language is not riddled with contradictions as Nāgārjuna believed, it still remains, for Dignāga as well as for Nāgārjuna and other Buddhists, essentially the product of language. This aspect of Dignāga’s thought does not seem to have received the attention it deserves from modern interpreters. Dignāga accepts the phenomenal world—or to be more precise, the world of discourse—as correspondent to, or even created by, language. In so doing, he finds himself in a situation quite similar to that of Vaiśeṣika thinkers. There remains a fundamental difference, of course. Vaiśeṣikas do not accept the existence of a reality apart from the world of language and its referents; for them, reality corresponds perfectly to language. For Dignāga, ultimate reality eludes the grasp of language, while the world of discourse, which in the final analysis is not real, is determined by language. We shall see that the Vaiśeṣikas’ reality and Dignāga’s world of discourse have much in common. First let us examine the opening verse from chapter 5 of the Pramāṇasamuccaya: “Verbal [knowledge] does not constitute a means of knowledge distinct from inference, for it [proceeds] in the same way. … It refers to its object by means of the exclusion of other [things].”319 This verse has left modern commentators perplexed. What is this verbal knowledge, and how could it not be different from inference? The expression “verbal knowledge” is generally taken in the sense of knowledge obtained through verbal communication, what one learns from other people or possibly from books.320 The way we handle such communication would be, on this view, similar to logical inference. This interpretation seems to find support elsewhere in the Pramāṇasamuccaya, in the chapter on inference for oneself, where verse 5 says: “The statement of a trustworthy person is an inference, for they have in common not being false. [This feature] is present in an inference and in what is similar to it, and absent when the latter is absent.”321 319 The original Sanskrit of this verse has been preserved (Hattori, 1982: 107 n. 1; Hayes, 1988: 300 n. 1): na pramāṇāntaraṃ śābdam anumānāt tathā hi tat / kṛtakatvādivat svārthaṃ anyāpohena bhāṣate // 320 The translations offered for śābdam [pramāṇam] prove as much: “Verbal communication” (Hayes, 1988: 252), “[Knowledge] based on verbal communication” (Dreyfus, 1991: 339), “Mitteilung” (Frauwallner, 1959: 98-99 [774-75]). More cautious are M. Hattori, 1980: 61 (“[means of cognition] which is based on word”); R. Herzberger, 1986: 145 (“Knowledge derived from words”). 321 Here, too, the original Sanskrit has been preserved (Hayes, 1988: 250 n. 5): āptavākyāvisaṃvādasāmānyād anumānatā / anumāne ‘tha tattulye sadbhāvo nāstitāsati //. For the translation, cf. Hayes, 238. A translation of this verse closer to Dharmakīrti’s interpretation would be: “Since the words of an authority have the general character of trustworthiness, they constitute 93 94 A broader interpretation of the expression “verbal knowledge” is nonetheless possible. On this reading, verbal knowledge would include not only what one learns from other people, or from books, etc., but also, and chiefly, what language itself teaches us. Recall that Vaiśeṣika had little trouble deriving its system of reality at least in part from language, rather than from any particular tradition or communication. We have also seen, in the discussion that concludes the introduction to the present work, that the idea of a verbal knowledge derived directly from language did exist, perhaps even in Vaiśeṣika. Would it not be conceivable that Dignāga had the same idea as the Vaiśeṣikas? The fifth chapter of the Pramāṇasamuccaya indeed contains no indication, as far as I know, that verbal communication is at issue in this chapter. On the contrary, it does not have a single word to say about the source, or sources, of this supposed verbal communication. Does it concern the utterances of trustworthy people, or the Buddhist tradition? Dignāga says nothing on this key subject, which suggests that in reality he is not speaking (at least not exclusively) of verbal communication, but rather, as was just proposed, of a verbal knowledge arising primarily on the basis of language itself. From this point of view, it is language as such—not the utterances of trustworthy people, or Buddhist texts—that plays a determining role in the constitution of the phenomenal world; language can therefore also inform us about the structure of this world. The rest of the fifth chapter seems to confirm this interpretation. What, then, is the structure of the phenomenal world that Dignāga deduces from language? According to the Japanese scholar S. Katsura (1979), the structure is similar to that of the Vaiśeṣika system. Dignāga arrives at it in the following way. Words form a hierarchy among themselves. The word vṛkṣa, or “tree,” for example, excludes everything that is not a tree. However, it does not exclude either the object of a synonym or the object of a word with a broader extension. A tree being a substance, the word vṛkṣa thus does not exclude the object of the word dravya, “substance.” In the resulting hierarchy, some terms will be situated below the word vṛkṣa, as for example śiṃśapā, which designates a species of tree; others will be situated between vṛkṣa and dravya, or again above dravya. Between “tree” and “substance” there is pārthiva, “made of the element earth”; above dravya there is sat, “existent.” At the summit of the hierarchy is jñeya, “knowable.” One thus arrives at a sort of genealogical tree, which assigns a position to everything that occupies a place in Dignāga’s phenomenal world. Katsura’s article includes the following diagram, constructed on the basis of the text of the Pramāṇasamuccaya with the Vṛtti of Dignāga himself and the Ṭīkā of Jinendrabuddhi:322 jñeya _________|_________ | sat | asat an inference”; see Eltschinger, 1998. One will note that Dharmakīrti was already aware of the weakness of the position that the statement of a trustworthy person is an inference. 322 Katsura, 1979: 492 (17). I have not had the opportunity to verify the results of his research; a reexamination of his sources would be desirable, and might have a bearing on the reflections that follow. 94 95 ___________|_____________________________ | | | dravya guṇa karman ___|___________..... ____|__..... ___|__....... | | | | pārthiva āpya, etc. rūpa, etc. utkṣepaṇa, etc. ____|__________..... | | vṛkṣa ghaṭa, etc. __|_________..... | | śiṃśapā palāśa, etc. ____|___________..... | | puṣpita phalita, etc. This genealogical tree resembles the scheme of reality accepted by the Vaiśeṣikas. Under the element of sat, or “existent,” Dignāga presents the three categories of dravya, guṇa, and karman. These are also the foundational categories in Vaiśeṣika, which correspond, as we have already indicated, to the three types of words: nouns, adjectives, and verbs. As far as the sub-divisions of these three categories are concerned, Dignāga does not provide us with enough elements to be certain that he follows the Vaiśeṣika system, but the details he does provide do not, in any case, contradict it. We should be cautious, however, before accusing Dignāga of simply borrowing. Recall, rather, that he is basically pursuing the same goal as the Vaiśeṣikas. He seeks to discover the structure of the phenomenal world on the basis of language. The fact that he arrives at a result that is similar, at least in part, to that of Vaiśeṣika should not suprise us. For him, as for the Vaiśeṣikas, the three principal types of words—nouns, adjectives, and verbs—seem to have corresponding categories in the world. He and the Vaiśeṣikas would no doubt also agree as to the existence of the five elements of earth, water, wind, fire, and ether, as well as the five qualities corresponding to the five sense-faculties. From this point of view, a partial identity between the two systems is only natural. Note that Katsura, in a more recent article (1991: 131 n. 9), expresses a different view as to the source of the structure ascribed by Dignāga to the phenomenal world. Instead of Vaiśeṣika, Katsura proposes Bhartṛhari. He supports this view with a reference to a study by Radhika Herzberger (1986: 35 ff.). Herzberger believed she had found a hierarchical structure of universals in Bhartṛhari’s Vākyapadīya. As this conclusion was erroneous, there is no need to follow Katsura in his change of view.323 The hierarchical structure of Dignāga’s scheme merits our attention a moment longer. This structure entails some strange consequences, which we mentioned already in our discussion of Vaiśeṣika. A tree, according to the Vaiśeṣika scheme, is composed of earth but not of water, even though the presence of water, 323 See Bronkhorst, 1991: 6 ff. 95 96 in the form of sap, is not denied. Dignāga’s scheme appears to support the same conclusion. Why were both the Vaiśeṣikas and Dignāga led to accept such bizarre conclusions? Does this parallel allow us to determine the direction of the borrowing, if there was one? Recall that for Dignāga, verbal knowledge is a form of inference. As an inference, it proceeds from a term covering a relatively restricted domain to another term covering a larger domain. The restricted domain is necessarily included in the larger domain. One therefore proceeds, in this way, from the word śiṃśapā (which designates a type of tree) to the word vṛkṣa, “tree,” from vṛkṣa to pārthiva, “made of earth,” and so on. The path ascends from terms of a less extended application to terms of a more extended field of application, the latter encompassing the terms that precede them. Here one can ask whether the path that starts with a given term is necessarily unique. Is there only one way to ascend from śiṃśapā to jñeya? Or are there several possible paths existing alongside each other? Dignāga and the Vaiśeṣikas seem to accept but a single path from each starting term, which is why a tree is made of earth, rather than of earth and water at the same time. But why did Dignāga not accept that a tree could be made of both? I do not believe that any of his positions, such as the theory of apoha, necessitate this limitation of only one possible path for every term in the hierarchy. One is therefore tempted to conclude that the requirement was borrowed from the Vaiśeṣikas, for whom universals, as we have seen, are objective entities that cannot cut across one another.324 That being said, it seems clear that the presumed parallelism between verbal knowledge and inference—more precisely, the fact that verbal knowledge is supposed to proceed in the same manner as inference—facilitated this way of seeing things. An inference moves from a more restricted domain, such as the domain of all places possessing smoke, to a larger domain that encompasses the former, in this case the domain of all places possessing fire. That a cow is an animal results from the fact that the domain of things denoted by the word “cow” is included within the domain of things that one calls “animal.” Dignāga would thus have favored a hierarchical conception of the relations between words that does not allow for more than one “superior” for each word, regardless of the level in question. Note further that the hierarchical scheme of Vaiśeṣika occurs elsewhere, too. Utpaladeva, the tenth-century Kashimiri author we encountered when discussing satkāryavāda, basically accepts the Vaiśeṣika categories, except that in his view all of them are, fundamentally, identical to the knowing subject, i.e., God.325 The object of valid knowledge (pramiti), as he says in his Īśvarapratyabhijñākārikā Vṛtti, is denoted each time by a single word.326 One might think that various words—as, for example, the words “existent” (sat), “substance” (dravya), “golden” (kāñcana), “pot” (ghaṭa), and “shining” (ujjvala)—refer to one and 324 Their position resulted from their opposition to the Buddhists, according to Shastri, 1976: 326. (“As a result of conflict with the Buddhist, more and more emphasis was laid by the Nyāya-Vaiśeṣika on the objectivity of the universal. It was realized that the ranges of two objective realities could not be crossing or cutting each other.”) 325 Torella, 1994: XXII ff.; 45 ff. and 157 ff. (= Īśvarapratyabhijñākārikā & Vṛtti II.2). Strictly speaking, Utpaladeva seems to choose the Vaiśeṣika categories that suit him. 326 Torella, 1994: 48 (tr. p. 163): °ekaikaśabdavācye … sāmānyarūpe vastuni … pramitiḥ … 96 97 the same thing, but Utpaladeva tells us that they are different manifestations (ābhāsa) endowed with different capacities.327 It is clear that these “manifestations” constitute a hierarchy, the same one we encountered in Dignāga and, of course, Vaiśeṣika. There is no doubt that Utpaladeva was deeply influenced by Buddhist thinkers, and one indeed finds in his thought the theory of apoha, which emerges in his definition of vikalpa, or “analytic conceptualization”: “The ascertainment ‘pot,’ which is brought about by the free knowing subject alone, within whom lie the manifestations of this [pot] and that which is not [the pot], and which takes place through the exclusion (apohana) of that which is not [the pot]: this is the activity called vikalpa.”328 For Utpaladeva, as for the Buddhist theorists of apoha, the reality to which words refer is not absolute reality. But for Utpaladeva, in contrast to the Buddhists, absolute reality is the knowing subject, God. The theory of apoha rejects the correspondence principle. Therein lies its revolutionary aspect. We have seen that most thinkers in classical India were hesitant to reject the principle: most often they found ways to salvage it, maintaining, for instance, that words refer to things other than individuals, or even holding that individuals exist, in one way or another, before their arising. For Dignāga, words do not refer to anything, whether in absolute reality or in the phenomenal world. Even if, in the final analysis, his phenomenal world is still greatly indebted to language, insofar as it derives its essential structure therefrom, Dignāga nonetheless dismisses the correspondence principle. 18. The Bodhisattvabhūmi Dignāga, however, was not the first to reject the correspondence principle. A chapter from the Bodhisattvabhūmi, which belongs to the Yogācārabhūmi, does the same thing. The chapter is called the Tattvārthapaṭala, or “Chapter on Reality.” The Yogācārabhūmi is attributed to Asaṅga, a Buddhist thinker from the fourth or fifth century, but this attribution is certainly incorrect. The Yogācārabhūmi is instead a collection of independent parts, and multiple authors were doubtless involved in their composition. We do not know the name of the author of the Tattvārthapaṭala, which is unfortunate, since he explores a strategy of defense against Nāgārjuna’s arguments that proves to be completely original in its cultural context. The Tattvārthapaṭala distinguishes four ways of seeing reality: (1) that accepted by ordinary beings; (2) that accepted by reason (yukti); (3) the domain of knowledge purified of the obstructions of impurity; and (4) the domain of knowledge purified of obstructions to the knowable. The third and fourth ways of seeing reality naturally belong to more or less advanced Buddhists on the path to liberation. In developing its argument for the ineffability of the dharmas, which is the proper object of the fourth way of seeing reality, the text sets forth some 327 Torella, 1994: 49 (kārikā II.3.5; tr. p. 164): … sadghaṭadravyakāñcanojjvalatādayaḥ ābhāsabhedā bhinnārthakāriṇa[ḥ] … 328 Torella, 1994: 28 (Vṛtti on kārikā I.6.3): pramātur eva svatantrasyāntarlīnatadatadarthābhāsasya atadapohanena ghaṭa iti niścayo vikalpo nāma vyāpāraḥ. The translation above is a modified version of Torella, p. 131. 97 98 extremely interesting observations.329 Before turning to them, let us note that the four ways of seeing reality bear on the connection between words and things. More or less advanced followers of the path know that things derive their nature entirely from verbal designations. Ordinary people, by contrast, believe that this nature actually exists. It is thus in the context of the domain of knowledge purified of obstructions to the knowable that our author launches an attack against an opponent we can identify as a follower of the school of Nāgārjuna. The opponent is described as one who, in negating both the basis that serves as the cause of verbal designations (prajñaptivādanimittādhiṣṭhāna) and the support that serves as the cause of verbal designations (prajñaptivādanimittasanniśraya), rejects an inexpressible (nirabhilāpyātmakatayā), ultimately existing object (paramārthasadbhūta vastu) and concludes that absolutely nothing exists (sarveṇa sarvaṃ nāstīti). He negates even the bare object (vastumātra) with respect to the dharmas, i.e., rūpa, etc.330 Insofar as this person denies designation as well as reality, he is the worst of the nihilists.331 Our author does not share his opponent’s point of view. According to him, the truly existent nature of all the dharmas is accessible only to an awareness that is free of all conceptualization (nirvikalpa jñāna) and that does not agree with language (abhilāpa); it is, rather, inexpressible (nirabhilāpya). The author now engages in an extremely interesting hypothetical argument: if, he says, the truly existent nature of things agreed with language, then dharmas and things would conform to language; more precisely, he speaks of an identity of nature (tādātmya, tanmayatā, tatsvabhāvatā). In that case, every object would have many natures, since different words (abhilāpa) create many denominations (prajñapti) and many metaphorical expressions (upacāra) for every object. As it is impossible to determine that a given object has an identity of nature with one particular denomination as opposed to any other, the object has no identity of nature with any denomination.332 329 Vincent Eltschinger and Tom Tillemans, in a French translation of the text that has not yet been published, draw attention to a parallel passage in the Mahāyānasaṃgraha of Asaṅga; see Lamotte, 1973: II: 118-19. 330 Bodhisattvabhūmi (ed. Dutt) p. 30 l. 26 - p. 31 l. 5: dvāv imāv asmād dharmavinayāt pranaṣṭau veditavyau / yaś ca … / yaś cāpi prajñaptivādanimittādhiṣṭhānaṃ prajñaptivādanimittasanniśrayaṃ nirabhilāpyātmakatayā paramārthasadbhūtaṃ vastv apavadamāno nāśayati sarveṇa sarvaṃ nāstīti / … / yathā punā rūpādikeṣu dharmeṣu vastumātram apy apavadamānaḥ sarvavaināśikaḥ pranaṣṭo bhavaty asmād dharmavinayāt tathā vakṣyāmi / “It should be understood that these two views have fallen away from our Dharma-Vinaya: (1) that one which … ; and also (2) that one which, with respect to a given thing (vastu), denies the foundation for the sign of verbal designation, and the basis for the sign of verbal designation, which exists in an ultimate sense owing to its inexpressible essence, saying ‘absolutely everything is non-existent.’ … Likewise, denying the bare given thing, which is a universal denial (sarvavaināśika), has fallen away from our Dharma-Vinaya.” Tr. Willis, 1979: 160-61. 331 Bodhisattvabhūmi (ed. Dutt) p. 31 l. 17-18: prajñaptitattvāpavādāc ca pradhāno nāstiko veditavyaḥ. 332 Bodhisattvabhūmi (ed. Dutt) p. 30 l. 1-16: tatra kayā yuktyā nirabhilāpyasvabhāvatā sarvadharmāṇāṃ pratyavagantavyā / … na svabhāvo dharmāṇāṃ tathā vidyate yathābhilapyate / na ca punaḥ sarveṇa sarvaṃ na vidyate / … sa punaḥ pāramārthikaḥ svabhāvaḥ sarvadharmāṇāṃ nirvikalpasyaiva jñānasya gocaro veditavyaḥ / sacet punar yathaivābhilāpo yeṣu dharmeṣu yasmin vastuni pravartate tadātmakās te dharmā vā tad vastu syāt / evaṃ sati bahuvidhā bahavaḥ svabhāvā ekasya dharmasyaikasya vastuno bhaveyuḥ / tat kasya hetoḥ / tathā hy ekasmin dharme ekasmin vastuni bahuvidhā bahavo bahubhir abhilāpaiḥ prajñaptaya upacārāḥ kriyante / na ca bahuvidhānāñ ca bahūnāṃ prajñaptivādānāṃ niyamaḥ kaścid upalabhyate yad anyatamena prajñaptivādenaikena tasya dharmasya tasya vastunaḥ tādātmyaṃ tanmayatā tatsvabhāvatā 98 99 The text presents yet another argument. If an object derived its nature from the designation one gives it, then before receiving its designation it would be without its own nature. But as one cannot designate an object unless it has a nature of its own, the designation would have no object.333 This section of the Bodhisattvabhūmi seems remarkable to me for the following reasons. First, it seems to be familiar with the philosophy of Nāgārjuna. It characterizes this philosophy correctly, I believe, when it says that the Mādhyamika negates the ground of verbal designations and thereby concludes that absolutely nothing exists. The author of our passage approves of the first part but rejects the conclusion. He, too, believes that the nature of things does not agree with language, but he does not therefore conclude that nothing exists. He also has ideas about the connection between words and the world of our experience. The close correspondence between words and things was the point of departure for syān nānyair avaśiṣṭaiḥ prajñaptivādaiḥ / tasmāt sakalavikalaiḥ sarvaprajñaptivādaiḥ sarvadharmāṇāṃ sarvavastūnāṃ nāsti tādātmyaṃ nāsti tanmayatā nāsti tatsvabhāvatā / “Now by what philosophical reasoning is the inexpressible character (nirabhilāpya) of all dharmas to be understood? … the essential nature of dharmas is not found in the way in which it is expressed. But further, neither is absolutely nothing found. … Moreover, one should understand that only the sphere of cognitive activity which is completely freed of discursive thought is the domain of knowledge of the supreme essential nature of all dharmas. “Again, if with regard to any dharma or any given thing it is assumed that it becomes just like its expression, then those dharmas and that given thing would be that expression itself. But if that were the case, then for a single dharma and a single given thing there would be very many kinds of essential nature. And why? It is like this: to a single dharma and to a single given thing, various men will attach many different designations by virtue of numerous expressions of various kinds. That dharma and that given thing ought to have identity with, be made up of, and have the essential nature of some one verbal designation, but not of the other remaining verbal designations. But there being no fixed determination, which of the very many kinds of verbal designation would hold as the correct one? Therefore, the use of any and all verbal designations, however complete or incomplete, for any and all given things does not mean that the latter are identical to, made up of, or receive essential nature through those verbal designations.” Tr. Willis, 1979: 158-59. 333 Bodhisattvabhūmi (ed. Dutt) p. 30 l. 16-23: api ca saced rūpādayo dharmā yathāpūrvanirdiṣṭāḥ prajñaptivādasvabhāvā bhaveyuḥ / evaṃ sati pūrvaṃ tāvad vastu paścāt tatra chandataḥ prajñaptivādopacāraḥ / prāk prajñaptivādopacārād akṛte prajñaptivādopacāre sa dharmas tad vastu niḥsvabhāva eva syāt / sati niḥsvabhāvatve nirvastukaḥ prajñaptivādo na yujyate / prajñaptivādopacāre cāsti prajñaptivādasvabhāvatā dharmasya vastuno na yujyeta / sacet punaḥ pūrvam eva prajñaptivādopacārād akṛte prajñaptivādopacāre sa dharmas tad vastu tadātmakaṃ syāt / evaṃ sati vinā tena rūpam iti prajñaptivādopacāreṇa rūpasaṃjñake dharme rūpasaṃjñake vastuni rūpabuddhiḥ pravarteta / na ca pravartate / “Now, to view it in another way, suppose the dharmas themselves, of form and so forth as previously expounded, should become the essential nature of their verbal designations. If this were the case, then, first there would be just the given thing alone, i.e., completely disassociated from names, and only afterward would there be the desire to attach to that given thing a verbal designation. But this would mean that before a verbal designation was attached, at the time just prior to attaching the designation, that very dharma and that very given thing would be without essential nature. But if there were no essential nature, there would be no given thing at all; and hence, a designation would not be called for. And since no verbal designation would be attached, the essential nature of the dharma and of the given thing could not be proved. “Again, suppose that just prior to the attaching of a verbal designation, that dharma and that given thing should be identical with the designation. This being the case, even without the verbal designation ‘form,’ the idea of form would occur whenever there was a dharma with the name ‘form,’ and whenever there was a given thing with the name ‘form.’ But such does not occur. ” Tr. Willis, 1979: 159-60. 99 100 Nāgārjuna’s reflections. The author of our passage emphasizes, rather, the absolute impossibility of such a correspondence, as there are always many words for every thing. Note that this observation does not concern ultimate reality but phenomenal reality, the world of our experience. Though it is not proposed in the context of refuting the nihilist opponent, it nonetheless undermines, I believe, the basis of many of Nāgārjuna’s arguments. Indeed, the fact that the Tattvārthapaṭala was familiar with this nihilist position suggests that the observation it makes about the impossibility of a close correspondence between words and things is grounded in the difference of opinion that separated our author from the Madhyamaka tradition. 19. Prajñākaragupta Before concluding our discussion of the role of the correspondence principle among Buddhist thinkers, we must briefly consider a passage from the Pramāṇavārttikabhāṣya or Pramāṇavārttikālaṅkāra, a commentary on the Pramāṇavārttika of Dharmakīrti composed by Prajñākaragupta around 800 C.E. Eli Franco drew attention to this passage in a lecture delivered in autumn 1997 at the International Institute for Buddhist Studies in Tokyo, a copy of which he was kind enough to send me.334 The passage uses the two sentences aṅkuro jāyate (“the sprout arises”) and ghaṭaṃ karoti (“he makes a pot”) to support the author’s position. What is this position? More to the point, does it prove that Prajñākaragupta accepted the correspondence principle? I believe not. The general context of the two sentences consists in a defense of the curious position that a future object can be the cause of a present thing. But rather than analyze the broader context, which has already been admirably done by Franco, let us examine the immediate context in which the sentences appear. We read: “If [you ask] how a non-existent thing [can] be an actant (kāraka), [I reply with another question, namely:] how, in the sentences ‘the sprout arises’ and ‘he makes a pot,’ [can the sprout and the pot] be a subject and an object, respectively? If [you reply] that [the two are] actants in virtue of being present in the mind, not [even] crows would eat that (i.e., the position is completely unacceptable).”335 As the beginning of this passage shows, at issue are non-existent actants. This part of the objection is not called into question, and the reply suggests that there are indeed non-existent actants: the sprout in “the sprout arises,” and the pot in “he makes a pot,” do not exist at that moment. Other remarks from the same discussion support this conclusion. Prajñākaragupta asks, for instance, what the difference is between prior non-existence (prāgabhāva) and posterior non-existence (pradhvaṃsābhāva), thus indicating that a prior cause (which everyone accepts) is no more existent than a future cause; this observation implicitly assumes that neither one is present. We must conclude that Prajñākaragupta does not accept, either implicitly or explicitly, the idea that the words of a sentence and the things that together constitute the situation described by the sentence correspond to one another, 334 See now Franco, 2007. The same subject has been addressed by Shinya Moriyama (1998). PrvBh p. 68 l. 32-33 (on verse 2.50): kārakatvam evāsataḥ katham iti cet / katham aṅkuro jāyate, ghaṭaṃ karotīti kartṛkarmabhāvaḥ? buddhisthatayā kārakatve nātra tasya kākair bhakṣaṇam / 335 100 101 which is to say, the correspondence principle. He does not seem troubled by the fact that there is no sprout in the situation described by the sentence “the sprout arises,” and no pot in the situation described by “he makes a pot.” Prajñākaragupta’s aim is different, and his thought is free of the problem that so puzzled his distant predecessors. He uses, of course, the same type of sentences that they did—“the sprout arises,” “he makes a pot”—but he does so in order to prove something different. For him, these sentences remind us that the actants in a sentence can be situated in the future, and that in this way there can be future causes of a present situation. He is thus concerned with a different question, interesting in its own right, but without direct connection to the correspondence principle and its consequences, which in various ways so occupied earlier thinkers. 20. Indian Thinkers and the Correspondence Principle An important question should remain for us at the end of these lectures. We have suggested the role of the correspondence principle in the development of Indian thought. The influence this principle exerted seems to have been considerable. There is, if I may venture to believe at least one of my conclusions is correct, no current of Indian thought that was not affected by the consequences of this principle.336 The important question we can now no longer avoid is the following: were Indian thinkers aware of the principle? Did they consciously and deliberately accept it? Did they ever submit it to explicit critical examination? What is the Sanskrit expression for “correspondence principle”? All valid and justified questions. How to answer them? I have already had occasion to point out that the correspondence principle was principally an intuition shared by thinkers of the period, rather than an explicitly held logical position. I have presented many passages that lend themselves to an interpretation based on the correspondence principle, and I am convinced that a good number of them would be difficult to interpret in any other way. But none of these passages names the correspondence principle as such, and I know of no explicit discussion of its value. That being said, I must at once add that some of the passages we have seen speak rather directly on the subject. Take the passage from Śaṅkara’s Brahma Sūtra Bhāṣya, which we translated earlier. It reads: “[Otherwise] one could imagine that the arising of a pot, [though] spoken of, did not have the pot for its agent, but something else. … If that were the case, one would say, ‘The potter, etc., which are causes, arise,’ instead of saying, ‘The pot arises.’” (ghaṭasya cotpattir ucyamānā na ghaṭakartṛkā, kiṃ tarhy anyakartṛkā iti kalpyā syāt / … / tathā ca sati ghaṭa utpadyate ity ukte kulālādīni kāraṇāni utpadyante ity uktaṃ syāt.) One could hardly say it more directly: the words of a sentence correspond to the elements of the situation described, and if it seems otherwise to us, too bad. And the opponent in the Nyāya Bhāṣya, who proposed that nouns in a sentence can refer to things past or future, was clearly aware of the aspect of the correspondence principle according to which nouns in a present-tense sentence necessarily refer to 336 An exception may have to be made for the Cārvākas. It is clear from a passage in the Śābara Bhāṣya presenting their position that the Cārvākas refused to draw ontological conclusions from verbal usage. See Bronkhorst, 2007: 363 ff. 101 102 things that are present. Such examples convince me that at least some thinkers were well aware of the principle we have tried to bring to light. The difference between us and them is that the correspondence principle is not obvious to us; reading their texts in light of the principle demands intellectual effort on our part. But for Indian thinkers of the period, I believe, the obviousness of the principle was beyond question. This hypothesis would explain both the lack of explicit discussions of the principle and the marked resistance to abandoning it. We have certainly not said everything there is to say about the correspondence principle. The preceding pages have been able to offer only a sampling of passages dealing with the subject, even within the relatively short historical period to which we have restricted ourselves. Other aspects of the correspondence principle, such as the role of negation and the meaning of particles, have only just been broached. I believe, however, that our discussion already shows how important it is to realize that encounters with philosophy from another time period, possibly also from another culture, cannot be approached in the same way one would approach philosophical debates among contemporary thinkers. First and necessarily, there is what ancient Indian thinkers state directly. But second, and no less important, is what they do not say. The things that seemed obvious to them will not always coincide with what seems obvious to us. The correspondence principle does not seem obvious to us and, I do not hesitate to add, is false. The words of a sentence do not each at the same time necessarily correspond to an element in the situation they describe. But the belief to the contrary could have been responsible, among Indian thinkers, for more than one fundamental aspect of classical Indian thought. One can thus maintain, as I do readily, that the study of the implicit preconceptions or intuitions of Indian thinkers is an integral part of the effort required to understand Indian philosophy. *** We have avoided drawing comparisons with European thought, but now that we have come to the end of our study, we can make a small exception. European thinkers of the sixteenth and seventeenth centuries strove to recover the original language of Adam, which they believed to be the only language to confer immediate access to reality. The failure of these efforts opened the door to a search for new languages capable of offering privileged access to reality, such as the language of mathematics.337 The thinkers of classical India, by contrast, considered Sanskrit to be the language that most closely approximates reality, and they felt no need to construct other languages. Views concerning the degree and nature of this proximity varied from thinker to thinker, of course, with Mīmāṃsakas and Buddhists falling at opposite ends of the spectrum. Those who outright rejected the proximity in question were altogether few in number. This shared conviction manifests itself in various ways in the history of Indian thought; a detailed analysis of all of its manifestations and consequences remains a desideratum for the future. The present study is intended as little more than a modest contribution toward this end. 337 See, e.g., Bono, 1995. 102 103 APPENDIX: The Mahāprajñāpāramitāśāstra and the Sāṃkhya tanmātras338 In a different publication (Bronkhorst, 1994) I pointed out that classical Sāṃkhya derives the five elements from the five tanmātras, and has no place among its fundamental tattvas for the five qualities. Several sources, on the other hand, make us acquainted with an earlier form of Sāṃkhya which included the qualities among its tattvas — as derivatives of the elements — but did not mention the tanmātras. These qualities were even looked upon as the fundamental constituents of material objects. I made the suggestion "that a major change took place in Sāṃkhya doctrine, perhaps some time in the 5th century of our era". Before this change material objects were looked upon as collections of qualities, which were the final elements of the Sāṃkhya evolutionary scheme of things. After it, the qualities disappeared from the list of tattvas, and the tanmātras took their place, with this important difference that the tanmātras precede the elements, and are not final evolutes as were the qualities. It is possible, but not certain, that the tanmātras were originally looked upon as being qualities themselves.339 In addition to the arguments presented in my earlier article, the view that five qualities were still looked upon as full-fledged tattvas at the time of the Ṣaṣṭitantra may find some further support in a passage of this text regained by Steinkellner (1999: 670-71, 675 no. 8): śabdasparśarūparasagandhāḥ pañca trayāṇāṃ sukhaduḥkhamohānāṃ sanniveśaviśeṣāḥ. The value of this testimony is unfortunately somewhat reduced by the fact that the five tanmātras have the same names as the five qualities; however, if tanmātras were meant, one would have expected an indication to that effect. It is further to be kept in mind that the Ṣaṣṭitantra was familiar with the notion of tanmātra.340 Two articles by Imanishi dealing with some passages in the Mahāprajñāpāramitāśāstra, or Mahāprajñāpāramitopadeśa, complement the above information in an interesting manner. The Mahāprajñāpāramitāśāstra is traditionally ascribed to Nāgārjuna, and has been preserved in the Chinese translation of Kumārajīva, which he made in the very early years of the fifth century C.E.341 It appears to describe the evolution of Sāṃkhya tattvas in the following order:342 from prakṛti arises buddhi, from buddhi arises ahaṃkāra, from ahaṃkāra the five tanmātras, from the five tanmātras the five mahābhūtas, from the five mahābhūtas the five buddhīndriyas. A closely similar enumeration is found in Vasu's commentary on Āryadeva's *Śataka, also translated into Chinese by Kumārajīva.343 It deviates in one important respect: instead of merely mentioning the five buddhīndriyas it mentions all of the eleven indriyas at the end. The uncertain element in these enumerations is the term tanmātra. The 338 On the Sāṃkhya tanmātras, see further Bronkhorst, 1999b. See also Motegi, 1986; Kent, 1982: 267 f. Note that the Yoga Bhāṣya on YS 2.19 states quite explicitly that there are no further tattvas beyond the viśeṣas (na viśeṣebhyaḥ paraṃ tattvāntaram astīti viśeṣāṇāṃ nāsti tattvāntarapariṇāmaḥ), i.e. beyond the elements (bhūta) and the organs (indriya). 340 See note 345, below. 341 Ramanan, 1966: 16 and 335 n. 6. 342 TI 1509, vol. 25, p. 546c l. 18-29; cp. Imanishi, 1968: 70 f. 343 TI 1569, vol. 30, p. 170c l. 13-18; cp. Imanishi, 1970: 67. 339 103 104 Chinese expression normally translates aṇu, Imanishi observes,344 but here he proposes the translation tanmātra. This interpretation seems plausible, for the transition from tanmātras to mahābhūtas takes places in a manner reminiscent of the Yuktidīpikā: sound gives rise to ether; sound and touch to wind; sound, touch and colour to fire; sound, touch, colour and taste to water; and sound, touch, colour, taste and smell to earth. However, there is a vital difference between these texts and the Yuktidīpikā. To show this, I will cite the relevant passage of the Yuktidīpikā:345 From the tanmātra [called] ‘sound’, which has sound as quality, ether [is born,] which has [that] one quality. From the tanmātra [called] ‘touch’, which has sound and touch as qualities, wind [is born,] which has [these] two qualities. From the tanmātra [called] ‘colour’, which has sound, touch and colour as qualities, fire [is born,] which has [these] three qualities. From the tanmātra [called] ‘taste’, which has sound, touch, colour and taste as qualities, water [is born,] which has [these] four qualities. From the tanmātra [called] ‘smell’, which has sound, touch, colour, taste and smell as qualities, earth [is born,] which has [these] five qualities. To contrast our two texts, we consider the production of the element earth. According to the Yuktidīpikā, it is born from one tanmātra, which is called ‘smell’, and which has the five qualities sound, touch, colour, taste and smell. According to the Mahāprajñāpāramitāśāstra, on the other hand, earth is born from the five tanmātras sound, touch, colour, taste and smell. Water is similarly born from one tanmātra according to the Yuktidīpikā, and from four of them according to the Mahāprajñāpāramitāśāstra; and analogously for the remaining elements. The difference may look small, but it is vital.346 To this observation I would like to add Kumārajīva's use of a Chinese expression that normally translates aṇu ‘atom’. Kumārajīva, by using this translation, offers us an insight into his understanding of the term tanmātra. He thought of tanmātras as atoms which together constitute the elements. It is of course possible that Kumārajīva was mistaken. He may have been influenced by the conception of the constitution of matter current among his Buddhist coreligionists. But even without his special choice of translation for the term tanmātra 344 Cp. Motegi, 1986: 954 (5); Hirakawa, 1997: 458 (1137/14). YD p. 118 l. 14-16 (Pandeya) / p. 225 l. 15-19 (Wezler & Motegi): śabdaguṇāc chabdatanmātrād ākāśam ekaguṇam / śabdasparśaguṇāt sparśatanmātrād dviguṇo vāyuḥ / śabdasparśarūpaguṇād rūpatanmātrāt triguṇaṃ tejaḥ / śabdasparśarūparasaguṇād rasatanmātrāc caturguṇā āpaḥ / śabdasparśarūparasagandhaguṇād gandhatanmātrāt pañcaguṇā pṛthivī /. Cited and translated in Bronkhorst, 1994: 311. The Yoga Bhāṣya appears to represent the same point of view; see e.g. YBh 2.19: ṣaḍ aviśeṣāḥ / tad yathā śabdatanmātraṃ sparśatanmātraṃ rūpatanmātraṃ rasatanmātraṃ gandhatanmātraṃ ceti / ekadvitricatuṣpañcalakṣaṇāḥ śabdādayaḥ pañcāviśeṣāḥ /. Vijñānabhikṣu interprets this passage somewhat differently, maintaining that earlier tanmātras in this enumeration are causes of succeeding ones (YV 2.19 (ed. Rukmani p. 110): [u]ttarottaratanmātreṣu pūrvapūrvatanmātrāṇāṃ hetutvāc chabdatanmātraṃ śabdamātradharmakaṃ tatkāryatayā sparśatanmātraṃ śabdasparśobhayadharmakam, evaṃ krameṇaikaikalakṣaṇadharmavṛddhir ity arthaḥ. 346 It would appear that the Yuktidīpikā had not yet forgotten the idea of "simple" tanmātras. Cp. YD p. 91 l. 7 (Pandeya) / p. 187 l. 10-11 (Wezler & Motegi): ekarūpāṇi tanmātrāṇīty anye / ekottarāṇīti vārṣagaṇyaḥ /. The term ekottara refers here to the fact that each next tanmātra has one more quality than the preceding one; this is the position of Vārṣagaṇya and of the author of the Yuktidīpikā. The expression ekarūpa, on the other hand, refers no doubt to tanmātras that each have one quality, or consist of one quality each. 345 104 105 one is likely to arrive at such an understanding. Unlike the Yuktidīpikā, the Mahāprajñāpāramitāśāstra depicts a situation where five different tanmātras somehow jointly produce the element earth. No doubt the easiest way to visualize this is by assuming that the five tanmātras are five ‘atoms’ which jointly constitute a ‘molecule’ of earth. Seen in this way, the introduction of tanmātras into the Sāṃkhya philosophy took place when the qualities — which were believed to jointly constitute material objects — came to be conceived of in a more substancelike manner, as constituents of the smallest parts of matter. Later on the tanmātras themselves came to be considered as having qualities. This is what we find in the above passage of the Yuktidīpikā and elsewhere. A confirmation of the interpretation of the tanmātra as atomic is provided by the Yoga Bhāṣya. Here we read:347 "The tanmātra is the cause of the element. The single part of the [latter] is an atom (paramāṇu) which is itself a collection of different component parts which do not exist separately, consisting of a sāmānya and a viśeṣa. All tanmātras are like this."348 This seems to mean that the tanmātra is an atom, the single part of an element (bhūta). All, or some, elements may be composed of various tanmātras, as in the situation described in the Mahāprajñāpāramitāśāstra; the Yoga Bhāṣya is not however clear about this. But unlike the Mahāprajñāpāramitāśāstra which associates each tanmātra with a single quality, the Yoga Bhāṣya does not look upon the tanmātra as a single quality, but as a collection of a sāmānya and a viśeṣa. The viśeṣas are the normal five qualities, sound etc. The sāmānyas are the five elements, but conceived of as generic qualities; they are corporeality (mūrti; which is earth), viscosity (sneha; which is water), heat (uṣṇatā; which is fire), moving forward (praṇāmitā; which is wind), going everywhere (sarvatogati; which is ether).349 In introducing these generic qualities the Yoga Bhāṣya deviates from the other sources on Sāṃkhya that we are considering.350 The fact that the Yoga Bhāṣya, in spite of this difference, preserves the idea of the tanmātra as an atom, is no doubt significant. Another passage in the Yoga Bhāṣya can be interpreted along the same lines:351 "The single modification as sound-tanmātra of the constituents of nature (guṇa), which here take the form of something to be grasped (grāhya), is sound as 347 YBh 3.44: tanmātraṃ bhūtakāraṇam / tasyaiko 'vayavaḥ paramāṇuḥ sāmānyaviśeṣātmā 'yutasiddhāvayavabhedānugataḥ samudāya ity evaṃ sarvatanmātrāṇ[i]. 348 The translation "consisting of a sāmānya and a viśeṣa" for sāmānyaviśeṣātmā, rather than "consisting of sāmānyas and viśeṣas" or the like, seems confirmed by YBh 3.47 sāmānyaviśeṣayor ayutasiddhāvayavabhedānugataḥ samūho dravyam indriyam "the sense-organ is the substance which is an aggregate whose parts do not exist separately, of a sāmānya and a viśeṣa"; here the singular number of sāmānya and viśeṣa is guaranteed by the dual ending of their compound. 349 See Bronkhorst, 1994: 319. YBh 3.47 (sāmānyaviśeṣātmā śabdādir grāhyaḥ / viṣayaḥ) seems to suggest that the tanmātras have the same names as the qualities, also in the opinion of the author of the Yoga Bhāṣya. 350 Note however that the Yuktidīpikā under SK 38 enumerates (in ślokas) a great number of characteristics of the five elements, which includes the ones given in the Yoga Bhāṣya, though sometimes different expressions are used (YD p. 118 l. 21 ff. (Pandeya) / p. 225 l. 24 ff. (Wezler & Motegi)). 351 YBh 4.14: grāhyātmakānāṃ [guṇānāṃ] śabdatanmātrabhāvenaikaḥ pariṇāmaḥ śabdo viṣaya iti / śabdādīnāṃ mūrtisamānajātīyānām ekaḥ pariṇāmaḥ pṛthivīparamāṇus tanmātrāvayavaḥ / teṣāṃ caikaḥ pariṇāmaḥ pṛthivī gaur vṛkṣaḥ parvata ity evamādi / bhūtāntareṣv api snehauṣṇyapraṇāmitvāvakāśadānāny upādāya sāmānyam ekavikārāmbhaḥ samādheyaḥ / 105 106 object. A single modification of sound etc. when they are of the same kind as corporality (mūrti) is the earth-atom, which is constituted of tanmātras. A single modification of those [atoms] is such a thing as the earth, a cow, a tree, a mountain. Also in the case of the other elements, by taking up viscosity (sneha), heat (auṣṇya), moving forward (praṇāmitva) or giving space (avakāśadāna) as generic quality, a single modification is to be produced." The crucial word tanmātrāvayava˙ must, in view of the context which speaks of ever more composite entities, be understood as a bahuvrīhi compound: "the parts of which are tanmātras".352 Also Utpaladeva's (much later) Īśvarapratyabhijñākārikā Vṛtti (on 3.1.10-11) would seem to consider the tanmātras to be atomic. We read here:353 "There are two kinds of effect. The following five — sound, touch, colour, taste, and smell — are called tanmātras because they are subtle. The same five, when gross because they are arranged together, are called earth etc." Utpaladeva, too, appears to look upon the tanmātras as the building blocks of the elements, and therefore most probably as their constituent atoms. If the above account of the introduction and development of tanmātras in the Sāṃkhya philosophy is correct, it becomes understandable why they were introduced in the first place. The position of the school to the extent that material objects are constituted of qualities was hard to reconcile with the doctrine of satkāryavāda which the school felt compelled to adopt. Moreover, this same position no doubt led — under the influence of Buddhism or otherwise — to reflections about the basic constituents of matter. Both factors combined led to the tendency to conceive of the original qualities as of the nature of substances: the qualities became tanmātras, the ultimate constituents of the molecules of matter. However, classical Sāṃkhya was apparently not ready to abandon its qualities altogether, and took the contradictory step of attributing qualities to the tanmātras. At this point it was no longer necessary to maintain that the elements are constituted of various tanmātras. From now on an element like earth "arises" out of one single tanmātra, which however has five qualities. A peculiar passage in Vyomaśiva's Vyomavatī, the earliest known commentary on the Padārthadharmasaṅgraha better known by the name Praśastapādabhāṣya, confirms the idea that the tanmātras were at one point the ultimate constitutents of the molecules of matter. This passage discusses and explains the Vaiśeṣika position according to which a body is made up either of earth, or of water, or of fire, or of wind, but not of any combination of these elements. The objection is raised that bodies might consist of various elements at the same time. In this connection the following passage occurs:354 352 Hattori (1968: 154 n. 5.31) concludes from this passage that “[t]he Sāṃkhyas hold that the five kinds of tanmātras are composed of their respective atoms”. This interpretation may have to be revised. 353 Utpaladeva's Īśvarapratyabhijñākārikā Vṛtti on 3.1.10-11 (Torella, 1994: 65): kāryāṇi dvidhā / śabdasparśarūparasagandhāni sūkṣmatvena tanmātrasaṃjñāni pañca, anyonyavyūhena sthūlāni pṛthivyādisaṃjñāni tāny eva pañca / 354 Vyomavatī I p. 81 l. 13-21: athāvayavasyāpy anekabhūtair ārambhaḥ / tathāhi, pārthivāpyābhyāṃ paramāṇubhyāṃ dvyaṇukam, punar āpyataijasābhyām, tathā taijasavāyavīyābhyām ārabdham iti / evaṃ vāyuśabdatanmātrābhyām / etāni dvyaṇukāny utpannāni tryaṇukādiprakrameṇa śarīram ārabhanta ity abhyupagame ... 106 107 But if you accept the following: The constitution of a part, too, [can take place] with various elements. For example, a dvyaṇuka is constituted of an atom of earth and an atom of water, or again of an atom of water and an atom of fire, or of an atom of fire and an atom of wind. In the same way it [can be constituted] of wind and the tanmātra of sound. These dvyaṇukas, once arisen, constitute, passing through [the stages] tryaṇuka etc., a body. This passage presents a position that is not accepted by Vyomaśiva, who points out that according to Vaiśeṣika doctrine the resulting dvyaṇukas and more complex entities cannot possess the qualities inhering in the constituent atoms. All this does not concern us at this moment. What does concern us is that the śabdatanmātra — the tanmātra of sound, or the tanmātra which is sound — is here presented as a constituent of a potential dvyaṇuka, and therefore as some kind of atom, beside the atoms of earth, water, fire, and wind. It takes the place of what should be the atom of ether; but obviously, ether being one and omnipresent, there can be no atom of ether in Vaiśeṣika. This passage is enigmatic, because it is not quite clear who the opponent is. One may however guess that Vyomaśiva took this discussion, and therefore the position of the opponent, from an earlier work. Indeed, the same portion of the Vyomavatī ends with a long citation from a work which is identified as asya sūtrasya bhāṣyam "the Bhāṣya on this sūtra". The sūtra concerned (bhūyastvād rasavattvāc codakaṃ rasajñāne prakṛtiḥ) cannot be identified with certainty, but appears to have belonged to the Vaiśeṣika Sūtra.355 The Bhāṣya on the Vaiśeṣika Sūtra was not, of course, the Praśastapādabhāṣya, but appears to have been the Kaṭandī, composed after Vasubandhu but before Dignāga, it seems.356 It is therefore possible, or even likely, that the discussion about dvyaṇukas constituted of wind and śabdatanmātra occurred already in this earlier text, which may, in its turn, have been acquainted with an earlier work of Sāṃkhya, in which tanmātra was still known in the sens of "constituent of a molecule". It seems however clear that the position described by Vyomaśiva — and perhaps taken by him from the Kaṭandī — represents some hybrid between Sāṃkhya and Vaiśeṣika: whereas the notion of tanmātra appears to be Sāṃkhya, that of dvyaṇuka and tryaṇuka is decidedly Vaiśeṣika. How did the thinkers of classical Sāṃkhya conceive of their tanmātras? Strangely enough, the commentaries on the Sāṃkhya Kārikā contain practically no information that would allow us to answer this quesion. One might expect that the "classical" tanmātra was still thought of as something like an atom. Is this correct? The answer has to be negative, for the Yuktidīpikā rejects the very existence of atoms.357 355 Vyomavatī I p. 82 l. 20 ff. Two slightly different sūtras with their Bhāṣyas are cited Vyomavatī I p. 85 l. 17 ff. (bhūyastvād rūpavattvāc ca rūpajñāne prakṛtiḥ kāraṇaṃ tejaḥ) and I p. 90 l. 4 ff. (bhūyastvāt sparśavattvāc ca sparśajñāne prakṛtir vāyuḥ). 356 Bronkhorst, 1993. 357 YD p. 68 l. 32 ff. (Pandeya) / p. 154 l. 30 ff. (Wezler & Motegi): ucyate: yat tāvad uktaṃ paramāṇūnām apratiṣedhād vipratiṣedhaprasaṅga ity atra brūmaḥ: tadanupapattir astitvānabhyupagamāt / astitve hi paramāṇūnām abhyupagamyamāne sati satyam evaṃ syād iyam āśaṅkā kiṃ paramāṇupūrvakam idaṃ viśvam atha pradhānapūrvakam iti / na tu teṣāṃ sadbhāvo niścitaḥ / tasmād ayuktam etat / .... etc. Note however that Gauḍapāda on SK 42 refers to the tanmātras as atoms (paramāṇu). 107 108 A passage in the Yuktidīpikā appears to say that the tanmātras, far from being somehow atomic, are quite the opposite of that. The passage occurs in the middle of a discussion regarding the question whether atoms (paramāṇu) can be looked upon as the cause of the world, a position which the author of the Yuktidīpikā does not accept. The opponent then claims that a connection with atoms can be observed in the manifest world, because of the presence of colours etc.358 The Sāṃkhya replies that this can be accounted for differently, namely, with the help of the tanmātras.359 A little later the opponent makes a surprising proposal: what would be wrong in accepting that the tanmātras are atoms?360 The Sāṃkhya rejects this proposal in a few lines that merit careful attention:361 ucyate: na śakyam evaṃ bhavitum / kiṃ kāraṇam? vṛddhimatyas tanmātralakṣaṇāḥ prakṛtayo 'smābhir abhyupagamyante / kasmāt? svakāryād dhi prathīyasī prakṛtir bhavatīti ca naḥ samayaḥ / mahānti ca pṛthivyādīni mahābhūtāni / tasmāt teṣāṃ tadatiriktayā prakṛtyā bhavitavyam / paricchinnadeśāś ca paramāṇavaḥ / tasmān na tanmātrābhyupagamāt teṣām abhyupagamaḥ / This means: The answer is that it is not possible [that the tanmātras are atoms]. We accept that the prakṛtis that are the tanmātras have greater size (vṛddhi) [than their derivatives], for it is our doctrine that a prakṛti is larger than its effect. The great elements earth etc. are of macroscopic size. It follows that their prakṛti must exceed them [in size]. Atoms, on the other hand (ca), have limited size. It does not therefore follow from the fact that we accept tanmātras that we accept those [atoms]. In order to understand these lines it is necessary to recall that in the Sāṃkhya evolutionary scheme prakṛtis give rise to vikṛtis, and that the tanmātras in particular are prakṛtis that give rise to the vikṛtis called mahābhūtas "great elements". We now learn that the tanmātras are larger than the great elements, this because prakṛtis are always larger than their vikṛtis.362 In a way this is not even surprising, given that classical Sāṃkhya looks upon the original prakṛti, i.e. avyakta or pradhāna, as being omnipresent.363 358 YD p. 68 l. 26 (Pandeya) / p. 154 l. 22 (Wezler & Motegi): paramāṇvanvayo 'pi hi vyakta upalabhyate rūpādisattvāt. 359 YD p. 69 l. 2-5 (Pandeya) / p. 155 l. 2-6 (Wezler & Motegi): yat tu khalv idam ucyate pṛthivyādiṣu rūpādyupalambhād anvayadarśanād aṇūnāṃ sadbhāvaḥ ... kalpayitavya ity etad api cānupapannam / kasmāt? anyathāpi tadupapatteḥ / tanmātrapūrvakatve 'pi hi pṛthivyādīnāṃ kalpyamāne rūpādisattvād ato na yuktam etat / 360 YD p. 69 l. 10 (Pandey) / p. 155 l. 14-15 (Wezler & Motegi): āha: yadi punas tanmātrāṇām eva paramāṇutvam abhyupagamyate ka evaṃ sati doṣaḥ syāt? 361 YD p. 69 l. 10-14 (Pandeya) / p. 155 l. 16-21 (Wezler & Motegi). We follow Wezler & Motegi’s suggested reading prakṛtyā against all the mss which have pṛṭhivyā (?). 362 Curiously, the Yuktidīpikā calls the tanmātras ‘elements’ (bhūta), and their derivatives, which are smaller, ‘great elements’ (mahābhūta). This is clear from the following passage, which comments on the use of bhūta in the Sāṃkhya Kārikā (22) to refer to mahābhūtas (YD p. 92 l. 1-3 (Pandeya) / p. 188 l. 10 (Wezler & Motegi)): pūrvapadalopenātra mahābhūtānīti vaktavye bhūtānīty ucyate / bhūtasaṃjñā hi tanmātrāṇāṃ na pṛthivyādīnām atra tu sāṃkhyācāryāṇām avipratipattiḥ /. 363 See further Bronkhorst, 1999a; 2007a. 108 109 BIBLIOGRAPHY Alston, A.J. (1980): Śaṃkara on the Creation. Reprint 1983. London: Shanti Sadan. (A Śaṃkara SourceBook, 2.) Bhartṛhari: Mahābhāṣyadīpikā. 1) Edited by K.V. Abhyankar and V.P. Limaye. Poona: Bhandarkar Oriental Research Institute. 1970. (Post-Graduate and Research Department Series No. 8.) 2) Partly edited by V. Swaminathan under the title Mahābhāṣya Ṭīkā. Varanasi: Banaras Hindu University. 1965. (Hindu Vishvavidyalaya Nepal Rajya Sanskrit Series Vol. 11.) 3) Manuscript reproduced. Poona: Bhandarkar Oriental Research Institute. 1980. 4) ‘Critical edition’. Poona: Bhandarkar Oriental Research Institute. Āhnika 1 by J. Bronkhorst (1987), Āhnika 2 by G.B. Palsule (1988), Āhnika 3 by G.B. Palsule (1983), Āhnika 4 by G.V. Devasthali and G.B. Palsule (1989), Āhnika 5 by V.P. Limaye, G.B. Palsule and V.B. Bhagavat (1984), Āhnika 6 part 1 by V.B. Bhagavat and Saroja Bhate (1986), Āhnika 6 part 2 by V.B. Bhagavat and Saroja Bhate (1990), Āhnika 7 by G.B. Palsule and V.B. Bhagavat (1991). Bhattacharya, Kamaleswar (1980): "Nāgārjuna's arguments against motion: their grammatical basis." A Corpus of Indian Studies: Essays in honour of Professor Gaurinath Sastri. Calcutta. Pp. 85-95. Bhattacharya, Kamaleswar (1981): "The grammatical basis of Nāgārjuna's arguments: some further considerations." IT VIII-IX, 1980-1981 (Sternbach Commemoration Volume), pp. 35-43. Bhattacharya, Kamaleshwar (1984): "La signification du Madhyamaka: Reconsidération des arguments de Nāgārjuna contre le mouvement." Saṃskṛtasaṃskṛti˙, Cultura Sánscrita: Mémoire du Premier Symposium International de la Langue Sanskrite. México: Universidad Nacional AutĀnoma de México. Pp. 189-199 Bhattacharya, Kamaleswar (1985): "Nāgārjuna's arguments against motion." JIABS 8(1), 7-15. Bhattacharya, Kamaleswar (1986): The Dialectical Method of Nāgārjuna: Vigrahavyāvartanī. Second edition. Translated from the original Sanskrit by Kamaleswar Bhattacharya, text critically edited by E. H. Johnston and Arnold Kunst. Delhi: Motilal Banarsidass. Bhattacharya, Kamaleswar (1986a): "Notes bouddhiques." JA 274, 291-300. Bhattacharya, Kamaleswar (1991): "Réalisme ontologique et opportunisme grammatical dans l'Abhidharmadīpa et la Vibhāṣāprabhāvṛtti." Buddhist Studies, Present and Future. IABS Tenth International Conference, UNESCO, Paris, France, 18-21 July 1991. Pp. 74-76. Bhattacharya, Kamaleswar (1995): "Back to Nāgārjuna and grammar." ALB 59, 178-189. Bhattacharya, Kamaleshwar (1996): "Sur la base grammaticale de la pensée indienne." Langue, style et structure dans le monde indien: Centenaire de Louis Renou. Actes du Colloque international (Paris, 25-27 janvier 1996), édités par Nalini Balbir et Georges-Jean Pinault, avec la collaboration de Jean Fezas. Paris: Honoré Champion. Pp. 171-186. Bhattacharya, Vidhushekhara (1943): The Āgamaśāstra of Gauḍapāda, edited, translated and annotated. Reprinted with a Foreword by Christian Lindtner: Motilal Banarsidass, Delhi, 1989. Biardeau, Madeleine (1964): Théorie de la connaissance et philosophie de la parole dans le brahmanisme classique. Paris - La Haye: Mouton. (Le Monde d'Outre-mer, passé et présent, première série, études XXIII.) Biardeau, Madeleine (1964a): Bhartṛhari Vākyapadīya Brahmakāṇḍa avec la Vṛtti de Harivṛṣabha. Traduction, introduction et notes. Paris: E. de Boccard. Bodewitz, H. W. (1995): "Mahābhārata XII and ṚV. X. 129." In: Modern Evaluation of the Mahābhārata. Prof. R. K. Sharma Felicitation Volume. Ed. Satya Pal Narang. Delhi: Nag Publishers. Pp. 40-47. Bono, James J. (1995): The Word of God and the Languages of Man. Interpreting Nature in Early Modern Science and Medicine. Vol. 1: Ficino to Descartes. Madison: The University of Wisconsin Press. Bouy, Christian (1997): "La Māṇḍūkya-Upaniṣad et l'Āgamaśāstra. Concordances externes et citations." WZKS 41, 119-158. Bronkhorst, Johannes (1980): "The role of meanings in Pāṇini's grammar." IL 40(3), 1979 [1980], 146-157. Bronkhorst, Johannes (1983): "On the history of Pāṇinian grammar in the early centuries following Patañjali." JIP 11, 357-412. Bronkhorst, Johannes (1984): Patañjali and the Yoga sūtras." StII 10, 191-212. Bronkhorst, Johannes (1985): "On the chronology of the Tattvārtha Sūtra and some early commentaries." WZKS 29, 155-184. Bronkhorst, Johannes (1987): Three Problems pertaining to the Mahābhāṣya. Poona: Bhandarkar Oriental Research Institute. (Post-graduate and Research Department Series No. 30; Pandit Shripad Shastri Deodhar Memorial Lectures [Third Series].) 109 110 Bronkhorst, Johannes (1991): "Studies on Bhartṛhari, 3: Bhartṛhari on sphoṭa and universals." AS 45(1), 5-18. Bronkhorst, Johannes (1992): Review of Wood, 1990. AS 45(2), 1991, 324-333. Bronkhorst, Johannes (1992a): "Études sur Bhartṛhari, 4: L'absolu dans la Vākyapadīya et son lien avec le Madhyamaka." AS 46(1) (tome de mélanges offerts à Jacques May), 56-80. Bronkhorst, Johannes (1992b): "Quelques axiomes du Vaiśeṣika." Les Cahiers de Philosophie 14 ("L'orient de la pensée: philosophies en Inde"), 95-110. Bronkhorst, Johannes (1993): "The Vaiśeṣika vākya and bhāṣya." ABORI 72-73 (1991-1992), 145-169. Bronkhorst, Johannes (1993a): Review of Aus gutem Grund by C. Nenninger. AS 47(4), 703-707. Bronkhorst, Johannes (1993b): "Buddhist hybrid Sanskrit: the original language." In: Aspects of Buddhist Sanskrit. Proceedings of the International Symposium on the Language of Sanskrit Buddhist Texts, Oct. 1-5, 1991. Ed. Kameshwar Nath Mishra. Sarnath, Varanasi: Central Institute of Higher Tibetan Studies. (Samyag-vāk Series, VI.) Pp. 396-423. Bronkhorst, Johannes (1994): "The qualities of Sāṃkhya." WZKS 38, 309-322. Bronkhorst, Johannes (1996): "Sanskrit and reality: the Buddhist contribution." Ideology and Status of Sanskrit: Contributions to the history of the Sanskrit language. Ed. Jan E. M. Houben. Leiden etc.: E.J. Brill. Pp. 109-135. Bronkhorst, Johannes (1996a): "God's arrival in the Vaiśeṣika system." JIP 24(3), 281-294. Bronkhorst, Johannes (1996b): "The correspondence principle and its impact on Indian philosophy." Indo-Shisōshi Kenkyū / Studies in the History of Indian Thought (Kyoto) 8, 1-19. Bronkhorst, Johannes (1997): "Nāgārjuna's logic." Bauddhavidyāsudhākara˙. Studies in Honour of Heinz Bechert On the Occasion of His 65th Birthday. Ed. Petra Kieffer-Pülz and Jens-Uwe Hartmann. Swisttal-Odendorf. (Indica et Tibetica, 30.) Pp. 29-37 Bronkhorst, Johannes (1997a): "Philosophy and Vedic exegesis in the Mīmāṃsā." Beyond Orientalism: The Work of Wilhelm Halbfass and its Impact on Indian and Cross-Cultural Studies. Ed. Eli Franco and Karin Preisendanz. Amsterdam - Atlanta: Rodopi. (Poznań Studies in the Philosophy of Science and the Humanities vol. 59.) Pp. 359-371. Bronkhorst, Johannes (1998): "Les éléments linguistiques porteurs de sens dans la tradition grammaticale du sanskrit." Histoire Épistémologie Langage 20(1) ("Les grammaires indiennes"), 29-38. Bronkhorst, Johannes (1999): "Nāgārjuna and apoha." Dharmakīrti's Thought and Its Impact on Indian an Tibetan Philosophy. Proceedings of the Third International Dharmakīrti Conference, Hiroshima, November 4-6, 1997. Ed. Shoryu Katsura. Wien: Verlag der Österreichischen Akademie der Wissenschaften. (Beiträge zur Kultur- und Geistesgeschichte Asiens, 32; Denkschriften der phil.-hist. Klasse, 281.) Pp. 17-23. Bronkhorst, Johannes (1999a): "The contradiction of Sāṃkhya: on the number and the size of the different tattvas." AS 53(3), 679-691. Bronkhorst, Johannes (1999b): "Studies on Bhartṛhari, 8: prākṛta dhvani and the Sāṃkhya tanmātras." JIP 27(1/2) (Guruvandana: Essays in Indology in Honour of K. Bhattacharya), 23-33. Bronkhorst, Johannes (2000): "The last reason for satkāryavāda." Harānandalaharī. Volume in Honour of Professor Minoru Hara on his Seventieth Birthday. Ed. Ryutaro Tsuchida and Albrecht Wezler. Reinbek: Dr. Inge Wezler. Pp. 53-62. Bronkhorst, Johannes (2001): "Pour comprendre la philosophie indienne.” La rationalité en Asie / Rationality in Asia. Études de Lettres 2001, 3, pp. 195-221. Bronkhorst, Johannes (2002): "The Cāndra-vyākaraṇa: some questions." Indian Linguistic Studies: Festschrift in Honor of George Cardona. Ed. Madhav M. Deshpande and Peter E. Hook. Delhi: Motilal Banarsidass. Pp. 182-201. Bronkhorst, Johannes (2003): "Jainism's first heretic and the origin of anekānta-vāda." Jainism and Early Buddhism: Essays in Honor of Padmanabh S. Jaini. Ed. Olle Qvarnström. Fremont, California: Asian Humanities Press. Pp. 95-111. Bronkhorst, Johannes (2003a): Review of Wezler & Motegi, 1998. ZDMG 153(1), 242-247. Bronkhorst, Johannes (2007): Greater Magadha. Studies in the Culture of Early India. Leiden – Boston: Brill. (Handbook of Oriental Studies, Section Two: India, 19.) Bronkhorst, Johannes (2007a): “On the nature of pradhāna.” Expanding and Merging Horizons. Contributions to South Asian and Cross-Cultural Studies in Commemoration of Wilhelm Halbfass. Ed. Karin Preisendanz. Vienna: Austrian Academy of Sciences Press. 2007. (ÖAW, 110 111 philosophisch-historische Klasse, Denkschriften, 351. Band; Beiträge zur Kultur- und Geistesgeschichte Asiens Nr. 53.) Pp. 373-381. Brown, W. Norman (1968): "The creative role of the goddess Vāc in the Rig-Veda." Pratidānam: Indian, Iranian and Indo-European Studies presented to Franciscus Bernardus Jacobus Kuiper on his sixtieth Birthday. The Hague - Paris: Mouton. Pp. 393-397. Buswell, Robert E. (1997): "Buddhadeva: materials towards an assessment of his philosophy." JIP 25(6), 561-587. Butzenberger, Klaus (1994): "Jainism and Madhyamaka Buddhism: a survey of the Gaṇadharavāda, Section 4." Festschrift Klaus Bruhn zur Vollendung des 65. Lebensjahres dargebracht von Schülern, Freunden und Kollegen. Ed. Nalini Balbir and Joachim K. Bautze. Reinbek: Inge Wezler. Pp. 225-254. Caland, W. (tr.)(1931): Pañcaviṃśa-Brāhmaṇa. The Brāhmaṇa of twenty five chapters. Reprint: Sri Satguru Publications, Delhi, 1982. (Sri Garib Das Oriental Series, 7.) Carakasaṃhitā. The following edition has been used: The Charakasaṃhitā by Agniveśa, revised by Charaka and Dṛidhabala, with the Āyurveda-Dīpikā commentary of Chakrapāṇidatte, edited by Vaidya Jādavaji Trikamji Āchārya. Third edition. Bombay: Nirṇaya Sāgar Press. 1941. Cardona, George (1976): Pāṇini. A survey of research. Reprint Delhi: Motilal Banarsidass, 1980. Cardona; George (1991): "A path still taken: some early Indian arguments concerning time." JAOS 111(3), 445-464. Carpenter, David (1994): "The mastery of speech: Canonicity and control in the Vedas." Authority, Anxiety, and Canon. Essays in Vedic Interpretation. Éd. Laurie L. Patton. Albany: State University of New York Press. Pp. 19-34. Chakravarti, Pulinbehari (1951): Origin and Development of the Sāṃkhya System of Thought. Second edition: Oriental Books Reprint, New Delhi, 1975. Comba, Antonella (1990): "Universal (sāmānya) and particular (viśeṣa) in Vaiśeṣika and in Āyurveda." Journal of the European Āyurvedic Society 1, 7-32. Conze, Edward, (1959): Buddhist Scriptures, selected and translated. Harmondsworth etc.: Penguin Books. Cox, Collett (1995): Disputed Dharmas. Early Buddhist Theories on Existence: An annotated translation of the section on factors dissociated from thought from Saṅghabhadra's Nyāyānusāra. Tokyo: International Institute for Buddhist Studies. (Studia Philologica Buddhica, Monograph Series, XI.) de Jong, J.W. (1996): Review of Nolot, 1995. BSOAS 59(2), 382-383. Deleu, Jozef (1970): Viyāhapannatti (Bhagavaī). The fifth Aṅga of the Jaina canon. Introduction, critical analysis, commentary & indexes. Brugge: "De Tempel". Reprint: Motilal Banarsidass, Delhi, 1996. Deshpande, Madhav M. (1993): Sanskrit & Prakrit. Sociolinguistic issues. Delhi: Motilal Banarsidass. (MLBD Series in Linguistics, 6.) Deshpande, Madhav M. (1996): "The Vedic traditions and origins of grammatical thought in ancient India." Langue, style et structure dans le monde indien: Centenaire de Louis Renou. Actes du Colloque international (Paris, 25-27 janvier 1996), édités par Nalini Balbir et Georges-Jean Pinault, avec la collaboration de Jean Fezas. Paris: Honoré Champion. Pp. 145-170. Detienne, Marcel & Hamonic, Gilbert (ed.)(1995): La déesse parole: quatre figures de la langue des dieux. Série d'entretiens entre Georges Charachidzé, Marcel Detienne, Gilbert Hamonic, Charles Malamoud, et Carlo Severi. Paris: Flammarion. Doniger O’Flaherty, Wendy (1981): The Rig Veda. An anthology. Harmondsworth: Penguin Books. Dreyfus, Georges Bernard Jacques (1991): Ontology, Philosophy of Language and Epistemology in Buddhist Tradition. 2 tomes. University of Virginia: Department of Religious Studies. (Doctoral thesis, printed in 1993 by UMI Dissertation Services, Ann Arbor, Michigan.) Dutt, Nalinaksha (ed.)(1966): Bodhisattvabhumi (Being the XVth section of Asangapada's Yogacarabhumi). Patna: K. P. Jayaswal Research Institute. (TSWS, 7.) Eco, Umberto (1995): The Search for the Perfect Language. Oxford UK & Cambridge USA: Blackwell. (Translated from the original Italian (1993) by James Fentress.) Eltschinger, Vincent (1995): Relativisme et conventionalisme en théorie du langage. Autour du Tattvārthapaṭala de la Bodhisattvabhūmi et de Pramāṇavārttika I. Mémoire de licence (nonpublié). Faculté des lettres, Université de Lausanne. 111 112 Eltschinger, Vincent (1998): "Bhāvaviveka et Dharmakīrti sur āgama et contre la Mīmāṃsā (1)." AS 52(1) Enomoto, Fumio (1996): "A Sanskrit fragment from the Vibhāṣā discovered in Eastern Turkestan." Sanskrit-Texte aus dem buddhistischen Kanon: Neuentdeckungen und Neueditionenen III. (SWTF Beiheft 6.) Bearbeitet von Gregory Bongard-Levin et al. Göttingen: Vandenhoeck & Ruprecht. Pp. 133-143. Faddegon, B. (1918): The Vaiçeṣika-System described with the help of the oldest texts. VKNAWL XVIII,2. Reprint: Sändig Reprint, Vaduz, 1969. Falk, Harry (1993): Schrift im alten Indien. Ein Forschungsbericht mit Anmerkungen. Tübingen: Gunter Narr. (ScriptOralia, 56.) Franco, Eli (1991): "Whatever happened to the Yuktidīpikā?" WZKS 35, 123-137. Franco, Eli (1997): Dharmakīrti on Compassion and Rebirth. Wien: ATBS. (WSTB, 38.) Franco, Eli (2007): “Prajñākaragupta on pratītyasamutpāda and reverse causation.” Pramāṇakīrti˙. th Papers dedicated to Ernst Steinkellner on the occasion of his 70 birthday. Ed. Birgit Kellner, Helmut Krasser, Horst Lasic, Michael Torsten Much & Helmut Tauscher. Part 1&2. Wien: ATBS. Pp. 163-185. Frauwallner, Erich (1953): Geschichte der indischen Philosophie. I. Band. Salzburg: Otto Müller. Frauwallner, Erich (1958): "Die Erkenntnislehre des klassischen Sāṃkhya-Systems." WZKSO 2, 84-139 = KlSchr p. 223-278. Frauwallner, Erich (1959): "Dignāga, sein Werk und seine Entwicklung," WZKSO 3, 1959, 83-164 = KlSchr p. 759-841 Frauwallner, Erich (1968): Materialien zur ältesten Erkenntnislehre der Karmamīmāṃsā. Wien: Hermann Böhlaus. (SAWW 259. Bd., 2. Abhandlung) Frauwallner, Erich (1973): "Abhidharma-Studien. V. Der Sarvāstivāda˙: Eine entwicklungsgeschichtliche Studie." WZKS 17, 97-121. Frauwallner, Erich (1995): Studies in Abhidharma Literature and the Origins of Buddhist Philosophical Systems. Translated from the German by Sophie Francis Kidd under the supervision of Ernst Steinkellner. Albany: State University of New York Press. Gautama: Nyāya Sūtra. For the edition, see Nyāya Sūtra. Ghosh, Manomohan (ed.)(1967 f.): The Nāṭyaśāstra ascribed to Bharata-Muni. Edited with an introduction and various readings. Calcutta. Gillon, Brendan S. (1997): "Negative facts and knowledge of negative facts." Relativism, Suffering and Beyond. Essays in memory of Bimal K. Matilal. Ed. P. Bilimoria and J. N. Mohanty. Delhi etc.: Oxford University Press. Pp. 128-149. Gokhale, V. V. (1947): "Fragments from the Abhidharmasamuccaya of Asaṃga." JBBRAS 23, 13-38. Gombrich, Richard (1992): "The Buddha's book of Genesis?" IIJ 35, 159-178. Hacker, Paul (1950): "Eigentümlichkeiten der Lehre und der Terminologie Śaṅkaras: Avidyā, Nāmarūpa, Māyā, Īśvara." ZDMG 100, 246-286 = KlSchr pp. 69-109. (For an English translation of this article, see Halbfass, 1995: 57-100.) Hacker, Paul (1968): "Śaṅkara der Yogin und Śaṅkara der Advaitin: einige Beobachtungen." WZKSO 12-13, 1968-69 (Beiträge zur Geistesgeschichte Indiens, Festschrift für Erich Frauwallner, ed. G. Oberhammer), pp. 119-148 = KlSchr pp. 213-242. (For an English translation of this article, see Halbfass, 1995: 101-134.) Hacker, Paul (1972): "Notes on the Māṇḍūkyopaniṣad and Śaṅkara's Āgamaśāstravivaraṇa." India Maior: Congratulation Volume presented to J. Gonda. Ed. J. Ensink and P. Gaeffke. Leiden. Pp. 115132 = KlSchr pp. 252-269. Halbfass, Wilhelm (1986): "Mallavādin and early Vaiśeṣika ontology." ALB 50 (Golden Jubilee Volume), 271-286. Halbfass, Wilhelm (1986a): "Vyomaśiva on sattāsambandha." Studies in Indian Culture. S. Ramachandra Rao Felicitation Volume. Bangalore. Pp. 65-80. Halbfass, Wilhelm (1989): "Observations on sattāsambandha and the history of Vaiśeṣika ontology." JAOS 109, 553-557. Halbfass, Wilhelm (1991): Tradition and Reflection. Explorations in Indian thought. Albany: State University of New York Press. Halbfass, Wilhelm (1991a): "Tradition und Reflexion: Zur Gegenwart des Veda in der indischen Philosophie." Beiträge zur Hermeneutik indischer und abendländischer Religionstraditionen. Ed. Gerhard Oberhammer. ÖAW. Pp. 123-146. 112 113 Halbfass, Wilhelm (1992): On Being and What There Is. Classical Vaiśeṣika and the history of Indian ontology. Albany: State University of New York Press. Halbfass, Wilhelm (ed.)(1995): Philology and Confrontation. Paul Hacker on traditional and modern Vedānta. Albany: State University of New York Press. Harikai, Kunio (1997): "Kumārila's acceptance and modification of categories of the Vaiśeṣika school." Beyond Orientalism: The Work of Wilhelm Halbfass and its Impact on Indian and CrossCultural Studies. Ed. Eli Franco and Karin Preisendanz. Amsterdam - Atlanta: Rodopi. 1997. (Poznań Studies in the Philosophy of the Sciences and the Humanities vol. 59.) Pp. 395-415. Hattori, Masaaki (1968): Dignāga, on Perception, being the Pratyakṣapariccheda of Dignāga's Pramāṇasamuccaya from the Sanskrit fragments and the Tibetan versions. Translated and annotated. Cambridge, Massachusetts: Harvard University Press. (HOS, 47.) Hattori, Masaaki (1977): "The Sautrāntika background of the apoha theory." In: Buddhist Thought and Asian Civilization. Essays in honor of Herbert V. Guenther on his sixtieth birthday. Ed. Leslie S. Kawamura and Keith Scott. Emeryville, California: Dharma Publishing. Pp. 47-58. Hattori, Masaaki (1980): "Apoha and pratibhā." In: Sanskrit and Indian Studies. Essays in Honour of Daniel H.H. Ingalls. Ed. M. Nagatomi et al. Dordrecht etc.: D. Reidel. Pp. 61-73. Hattori, Masaaki (ed.)(1982): "The Pramāṇasamuccayavṛtti of Dignāga with Jinendrabuddhi's commentary. Chapter five: Anyāpoha-Parīkṣā. Tibetan text with Sanskrit fragments." Memoirs of the Faculty of Letters, Kyoto University, 21, 102-224. Hayes, Richard P. (1988): Dignāga on the Interpretation of Signs. Dordrecht: Kluwer. (Studies of Classical India, 9.) Hayes, Richard P. (1994): "Nāgārjuna's appeal." JIP 22, 299-378. Herzberger, Radhika (1986): Bhartṛhari and the Buddhists. An essay in the development of fifth and sixth century Indian thought. Dordrecht etc.: D. Reidel (Studies of Classical India, 8). Hinüber, Oskar v. (1990): Der Beginn der Schrift und frühe Schriftlichkeit in Indien. Stuttgart: Franz Steiner. (AAWL Jg. 1989 Nr. 11.) Hirakawa, Akira (1997): A Buddhist Chinese-Sanskrit Dictionary. Tokyo: Reiyukai. Holdrege, Barbara A. (1994): "Veda in the Brāhmaṇas: Cosmogonic paradigms and the delimitation of canon." Authority, Anxiety, and Canon. Essays in Vedic Interpretation. Ed. Laurie L. Patton. Albany: State University of New York Press. Pp. 35-66. Houben, J. E. M. (1992): "Bhartṛhari's samaya / Helārāja's saṃketa: a contribution to the reconstruction of the grammarians' discussion with the Vaiśeṣikas on the relation between śabda and artha." JIP 20, 219-242. Houben, Jan E. M. (1993): "Bhartṛhari en de structuur van het Sanskrit: ‘God's truth’ of ‘hocuspocus’?" Nieuwe Wegen in Taal- en Literatuurwetenschap. Handelingen van het Eenenveertigste Filologencongres. Ed. J. Goedegebuure. Tilburg University Press. Pp. 143-158. Houben, Jan E. M. (1995): The Saṃbandha-samuddeśa (Chapter on Relation) and Bhartṛhari's Philosophy of Language. A Study of Bhartrhari Saṃbandha-samuddeśa in the context of the Vākyapadīya, with a translation of Helārāja's commentary Prakīrṇa-prakāśa. Groningen: Egbert Forsten. xvi, 460 pp. (Gonda Indological Studies, vol. II.) Houben, Jan E. M. (1995a): "Bhartṛhari's perspectivism (2): Bhartṛhari on the primary unit of language." History and Rationality: The Skövde Papers in the Historiography of Linguistics. Ed. Klaus D. Dutz, Kjell-Āke Forsgren. Münster: Nodus. (Acta Universitatis Skodvensis, Series Linguistica, vol. 1). Pp. 29-62. Houben, Jan E. M. (1997): "The Sanskrit tradition." The Emergence of Semantics in Four Linguistic Traditions: Hebrew, Sanskrit, Greek, Arabic. By Wout van Bekkum, Jan Houben, Ineke Sluiter, Kees Versteegh. Amsterdam / Philadelphia: John Benjamins. (Amsterdam Studies in the Theory and History of Linguistic Science, Series III, Vol. 82.) Pp. 49-145. Imanishi, Junkichi (1968): "Ryūju ni yotte genkyū sareta sānkya shisō (The Sāṃkhya philosophy as referred to by Nāgārjuna)." Hokkaidō Daigaku Bungakubu Kiyō (Annual Report on Cultural Science, Faculty of Letters, Hokkaido University) 16(2), 35-96. Imanishi, Junkichi (1970): "Daiba Basō ni yotte genkyū sareta sānkya shisō (The Sāṃkhya philosophy as referred to by Āryadeva and Vasu)." Hokkaidō Daigaku Bungakubu Kiyō (Annual Report on Cultural Science, Faculty of Letters, Hokkaido University) 18(2), 63-101. Isaacson, Harunaga (1995): Materials for the Study of the Vaiśeṣika System. Doctoral thesis, Leiden University (unpublished). 113 114 Iyer, K. A. Subramania (ed.)(1983): Vākyapadīya of Bhartṛhari, Kāṇḍa II. Containing the Ṭīkā of Puṇyarāja and the Ancient Vṛtti. Delhi etc.: Motilal Banarsidass. Jaini, P. S. (1959): "The development of the theory of viprayukta-saṃskāras." BSOAS 22, 531-547. Jinabhadra: Viśeṣāvaśyaka Bhāṣya, with auto-commentary. Ed. Dalsukh Malvania (& Bechardas J. Doshi, pour le tome 3). 3 tomes. Ahmedabad: Lalbhai Dalpatbhai Bharatiya Sanskriti Vidyamandira. 1966-68. (Lalbhai Dalpatbhai Series, 10, 14 & 21.) Johnston, E. H. (1937): Early Sāṃkhya: An essay on its historical development according to the texts. Reprint Delhi: Motilal Banarsidass, 1974. Kapila: Sāṃkhya Sūtra. The following edition has been used: Maharṣi-Kapila-praṇītaṃ Sāṃkhyadarśanam, Aniruddha-kṛtā Vṛtti˙, Vedānti-Mahādeva-kṛto Vṛttisāra˙, Vijñānabhikṣukṛtaṃ Bhāṣyam, Nāgeśabhaṭṭa-kṛto Bhāṣyasāraś ceti vyākhyācatuṣṭayopetaṃ Devatīrtha-svāmikṛtena Sāṃkhyataraṅgeṇa, Sarvopakāriṇī-sahitena Tattvasamāsasūtreṇa, vividhapariśiṣṭaiś ca samupabṛṃhitam. Sampādaka˙: Janārdana-śāstrī Pāṇḍeya˙. Delhi: Motilal Banarsidass. 1989. Katsura, Shōryū (1979): "The apoha theory of Dignāga." IBK 28(1), (16)-(20) Katsura, Shōryū (1991): "Dignāga and Dharmakīrti on apoha." Studies in Buddhist Epistemological Tradition. Proceedings of the Second International Dharmakīrti Conference. Vienna, June 1116, 1989. Wien: ÖAW, Philosophisch-historische Klasse, Beiträge zur Kultur- und Geistesgeschichte Asiens 8. Pp. 129-146. Kawasaki, Shinjō (1977): "The Mīmāṃsā chapter of Bhavya's Madhyamaka-hṛdaya-kārikā — text and translation — (1) Pūrva-pakṣa." Studies 1976, Institute of Philosophy, The University of Tsukuba, pp. 1-16. Kent, Stephen A. (1982): "Early Sāṃkhya in the Buddhacarita." PEW 32, 259-278. King, Richard (1995): Early Advaita Vedānta and Buddhism. The Mahāyāna context of the Gauḍapādīyakārikā. Albany: State University of New York Press. King, Richard (1995a): "Early Advaita Vedānta: the date and authorship of the Gauḍapādīyakārikā." IIJ 38, 317-355. Kritzer, Robert (1999): Rebirth and Causation in the Yogācāra Abhidharma. Wien: ATBS. (WSTB, 44.) Kritzer, Robert (2005): Vasubandhu and the Yogācārabhūmi. Yogācāra elements in the Abhidharmakośabhāṣya. Tokyo: The International Institute for Buddhist Studies. (Studia Philologica Buddhica, Monograph Series, 18.) Kuiper, F. B. J. (1957): "Vācārambhaṇam." IIJ 1, 155-159. Kumārila Bhaṭṭa: Tantravārttika. (ASS 97) Kundakunda: Pravacanasāra. See Upadhye, 1964. Lamotte, Étienne (1973): La somme du grand véhicule (Mahāyānasaṃgraha). Tome I: Versions tibétaine et chinoise (Hiuan-tsang). Tome II: Traduction et commentaire. Louvain-la-Neuve: Institut Orientaliste, Université de Louvain. (PIOL 8.) Lang, Karen (1986): Āryadeva's Catu˙śataka. On the Bodhisattva's cultivation of merit and knowledge. Copenhagen: Akademisk Forlag. (Indiske Studier, 7.) Lang, Karen (1988): "On Āryadeva's citation of Nyāya texts in the *Śataka." WZKS 32, 131-140. Larson, Gerald James (1969): Classical Sāṃkhya: An interpretation of its history and meaning. Delhi: Motilal Banarsidass. La Vallée Poussin, Louis de (1937): "Documents d'Abhidharma: La controverse du temps." MCB 5 (1936-37), 7-158. La Vallée Poussin, Louis de (1937a): "Le bouddhisme et le yoga de Patañjali." MCB 5 (1936-1937), 223242. Lesimple, Em. (tr.)(1981): Māṇḍūkya Upaniṣad et Kārikā de Gauḍapāda. Paris: Librairie d'Amérique et d'Orient. (Les Upanishad, Texte et traduction, 5.) Leumann, E. (1885): "Die alten Berichte von den Schismen der Jaina." IS 17, 91-135. Liebenthal, Walter (1934): Satkārya in der Darstellung seiner buddhistischen Gegner: Die prakṛti-parīkṣā im Tattvasaṃgraha des Śāntirakṣita zusammen mit der Pañjikā des Kamalaśīla übersetzt und ausführlich interpretiert. Stuttgart - Berlin: W. Kohlhammer. Maas, Philipp André (2006): Samādhipāda. Das erste Kapitel des Pātañjalayogaśāstra zum ersten Mal kritisch ediert / The First Chapter of the Pātañjalayogaśāstra for the first time critically edited. Aachen: Shaker Verlag. (Indologica Halensis; Geisteskultur Indiens: Texte und Studien, 9.) Matilal, Bimal Krishna (1981): The Central Philosophy of Jainism (Anekānta-vāda). Ahmedabad: L.D. Institute of Indology. 114 115 May, Jacques (1959): Candrakīrti Prasannapadā Madhyamakavṛtti. Douze chapitres traduits du sanscrit et du tibétain, accompagnés d'une introduction, de notes et d'une édition critique de la version tibétaine. Paris: Adrien-Maisonneuve. Mayeda, Sengaku (1967): "On the author of the Māṇḍūkyopaniṣad- and the Gauḍapādīya-Bhāṣya." ALB 31-32 (1967-68), 73-94. Meuthrath, Annette (1996): Untersuchungen zur Kompositionsgeschichte der Nyāyasūtras. Altenberge: Oros. (Religionswissenschaftliche Studien, 36.) Mikogami, Esho (1969): "A refutation of the Sāṃkhya theory in the Yogācārabhūmi." PEW 19, 443-447. Ming-Wood, Liu (1994): Madhyamaka Thought in China. Leiden etc.: E.J. Brill. (Sinica Leidensia, 30.) Miśra, Viśvanātha (1996): "Vācam artho 'nudhāvati." Śrījñānāmṛtam. A Memorial Volume in Honour of Prof. Shri Niwas Shastri. Ed. Vijaya Rani and V. K. Goyal. Delhi: Parimal Publications. Pp. 202204. Miyamoto, Keiichi (1996): The Metaphysics and Epistemology of the Early Vaiśeṣikas. With an appendix Daśapadārthī of Candramati (a translation with a reconstructed Sanskrit text, notes and a critical edition of the Chinese version). Pune: BORI. (BOS, 28.) Moriyama, Shinya (1998): "Can the next life be proved from the present life? Prajñākaragupta's peculiar theory of causality (bhāvikāraṇavāda) in the context of his proof of the paraloka." Studies in Indian Philosophy and Buddhism (Tokyo University) 5, 44-57. Motegi, Shūjun (1986): "On tanmātra." IBK 68, 958-953. Motegi, Shujun (1994): "Some Vaiśeṣika thoughts referred to in the Yuktidīpikā." AS 48(2), 807-817. Müller, Max (1852): "Beiträge zur Kenntnis der indischen Philosophie, I: Kaṇâda's Vaiçeshika-Lehre." ZDMG 6, 1-34. Muroya, Yasutaka (1996): "Satkāryavāda ronshō ni okeru pariṇāma to abhivyakti (Pariṇāma and abhivyakti in the Sāṃkhya argument for satkāryavāda)." Indo-Shishōshi Kenkyū / Studies in the History of Indian Thought 8, 42-63. Nāgārjuna: Vigrahavyāvartanī. See Bhattacharya, 1986. Nakamura, Hajime (1992): "Buddhist influence as is noticed in the fourth chapter of the Gauḍapādīya-Kārikās." Philosophy, Grammar and Indology: Essays in Honour of Professor Gustav Roth. Ed. Hari Shankar Prasad. Delhi: Indian Books Centre. Pp. 241-258. Nolot, Édith (1995): Entretiens de Milinda et Nāgasena. Traduit du pâli, présenté et annoté. Paris: Gallimard. Nozawa, Masanobu (1993): "The definition of kārya in the Vaiśeṣika school." Hokkaido Journal of Indological and Buddhist Studies 8, 1-23. Nyāya Bhāṣya. For the edition, see Nyāya Sūtra. Nyāya Sūtra. The edition used occurs in the following publication: Nyāyadarśanam with Vātsyāyana's Bhāṣya, Uddyotakara's Vārttika, Vācaspati Miśra's Tātparyaṭīkā & Viśvanātha's Vṛtti. Chapter I, section I critically edited with notes by Taranatha Nyaya-Tarkatirtha and chapters I-ii-V by Amarendramohan Tarkatirtha, with an introduction by Narendra Chandra Vedantatirtha. Second edition: Munshiram Manoharlal, New Delhi. 1985. Oetke, Claus (1988): “Ich” und das Ich. Analytische Untersuchungen zur buddhistisch-brahmanischen Ātmankontroverse. Stuttgart: Franz Steiner Verlag Wiesbaden. (ANISt, 33.) Olivelle, Patrick (1996): Upaniṣads. Translated from the original Sanskrit. Oxford University Press. Parpola, Asko (1979): "On the symbol concept of the Vedic ritualists." Humanitas Religiosa. Festschr. Haralds Biezais. Stockholm: Almqvist & Wiksell. Pp. 139-153. Passmore, John (1966): A Hundred Years of Philosophy. Second edition. Harmondsworth: Penguin Books. Pollock, Sheldon (1985): "The theory of practice and the practice of theory in Indian intellectual history." JAOS 105(3), 499-519. Prabhācandra: Nyāyakumudacandra, a commentary on Akalaṅka's Laghīyastraya. Vol. 1. Edited by Mahendra Kumar. Bombay. 1938. Prabhācandra: Prameyakamalamārtaṇḍa, a commentary on the Parīkṣāmukhasūtra of Māṇikyanandin. Edited with Introduction, Indexes etc., by Mahendra Kumar. Third edition. Delhi: Sri Satguru. (Sri Garib Dass Oriental Series, 94.) Prajñākaragupta: Pramāṇavārttikabhāṣya. Ed. Rāhula Sāṅkṛityāyana. Patna: Kashi Prasad Jayaswal Research Institute. 1953. 115 116 Radicchi, Anna (1997): "Cāndravyākaraṇa e buddhismo." Bandhu: Scritti in onore di Carlo Della Casa in occasione del suo settantesimo compleanno. Ed. Renato Arena et al. Alessandria: Edizioni Dell'Orso. Pp. 389-411. Ramanan, K. Venkata (1966): Nāgārjuna's Philosophy as presented in the Mahā-Prajñāpāramitā-Śāstra. Delhi: Motilal Banarsidass. 1978. Rāmānuja: Śrī Bhāṣya. Edited, with his commentary Bhāṣyārtha Darpaṇa, by Uttamur T. Viraraghavacharya. Part I-II. Madras: Sreevathsa Press. 1963-1964. Rau, Wilhelm (2002): Bhartṛharis Vākyapadīya. Versuch einer vollständigen deutschen Übersetzung nach der kritischen Edition der Mūla-Kārikās. Ed. Oskar von Hinüber. Stuttgart: Franz Steiner. (AAWL, Einzelveröffentlichung Nr. 8.) Renou, Louis (1955): "Les pouvoirs de la parole dans le Ṛgveda." Études védiques et pāṇinéennes I. Paris: E. de Boccard. Pp. 1-27. Renou, Louis (1956): Hymnes spéculatifs du Véda. Gallimard / UNESCO. Robinson, Richard H. (1967): Early Mādhyamika in India and China. Reprint: Motilal Banarsidass, Delhi, 1978. Rospatt, Alexander von (1995): The Buddhist Doctrine of Momentariness. A survey of the origins and early phase of this doctrine up to Vasubandhu. Stuttgart: Franz Steiner. (ANISt 47.) Ruben, Walter (1979): Wissen gegen Glauben. Der Beginn des Kampfes des Wissens gegen den/das Glauben im alten Indien und Griechenland. Berlin: Akademie-Verlag. (Abhandlungen der Akademie der Wissenschaften der DDR, Abt. Gesellschaftswissenschaften Jg 1979 Nr. G 1.) Ruegg, David Seyfort (1978): "Mathematical and linguistic models in Indian thought: the case of zero and śūnyatā." WZKS 22, 171-181. Ruegg, David Seyfort (1981): The Literature of the Madhyamaka School of Philosophy in India. Wiesbaden: Otto Harrassowitz. (HIL vol. VII, fasc. 1.) Ruegg, David Seyfort (1982): "Towards a chronology of the Madhyamaka school." Indological and Buddhist Studies. Volume in honour of Professor J. W. de Jong on his Sixtieth Birthday. Ed. L. A. Hercus et al. Canberra: Faculty of Asian Studies. Pp. 505-530. Ruegg, D. Seyfort (1994): "pramāṇabhūta, *pramāṇa(bhūta)-puruṣa, pratyakṣadharman and sākṣātkṛtadharman as epithets of the ṛṣi, ācārya and tathāgata in grammatical, epistemological and Madhyamaka texts." BSOAS 57, 303-320. Rukmani, T.S. (ed., tr.)(1981 ff.): Yogavārttika of Vijñānabhikṣu. Text with English translation and critical notes along with the text and English translation of the Pātañjala Yogasūtras and Vyāsabhāṣya. 4 vols. New Delhi: Munshiram Manoharlal. (vol.1, 1981; vol. 2, 1983; vol. 3, 1987; vol. 4, 1989) Śaṅkara: Āgamaśāstravivaraṇa. 1) Māṇḍūkyopaniṣad, Gauḍapādīya Kārikā, Śāṅkara Bhāṣya tathā Hindī-anuvādasahita. Gorakhpur: Gītā Press, Saṃvat 2030. 2) Sa-Gauḍapādīya-Kārikā[A]tharvavedīya- Māṇḍūkyopaniṣad, Ānandagiri-kṛta-ṭīkā-saṃvalita-Śāṃkarabhāṣyasametā, Śaṃkarānanda-bhagavat-kṛtā Māṇḍūkyopaniṣaddīpikā ca. ... ‘Kāthavaṭe’ ity upāhvai˙ Viṣṇu-tanujai˙ Ābājī-Śarmabhi˙ saṃśodhitam. Poona: Ānandāśrama, 1936. (ASS 10.) Śaṅkara: Brahma Sūtra Bhāṣya. In: Brahmasūtra with Śāṅkarabhāṣya. Delhi etc.: Motilal Banarsidass (Works of Śaṅkarācārya, vol. III). 1964. Santina, Peter Della (1986): Madhyamaka Schools in India. A study of the Madhyamaka philosophy and of the division of the system into the Prāsaṅgika and Svātantrika schools. Delhi: Motilal Banarsidass. Scharf, Peter M. (1993): "Assessing Śabara's arguments for the conclusion that a generic term denotes just a class property." JIP 21(1), 1-10. Scharf, Peter M. (1996): The Denotation of Generic Terms in Ancient Indian Philosophy: Grammar, Nyāya, and Mīmāṃsā. Philadelphia: American Philosophical Society. (Transactions of the American Philosophical Society Held at Philadelphia For Promoting Useful Knowledge, Vol. 86, Pt. 3.) Scharfe, Hartmut (1971): "The Maurya dynasty and the Seleucids." KZ 85, 211-225. Scharfe, Hartmut (1977): Grammatical Literature. Wiesbaden: Otto Harrassowitz. (HIL vol. V, Fasc. 2.) Schmithausen, Lambert (1992): "A note on Vasubandhu and the Laṅkāvatārasūtra." AS 46(1) (tome de mélanges offerts à Jacques May), 392-397. Shastri, D.N. (1976): The Philosophy of Nyāya-Vaiśeṣika and its Conflict with the Buddhist Dignaga School. Delhi: Bharatiya Vidya Prakashan. 116 117 Siddhasena Divākara: Saṃmatitarkaprakaraṇa. Edited, with the Tattvabodhavidhāyinī of Abhayadevasūri, by Sukhalāla Saṃghi and Becaradāsa Dośi. Reprint: Rinsen Book, Kyoto, 1984. Śīlāṅka: Commentaire sur le Sūtrakṛtāṅgasūtra. In: Ācārāṅgasūtram and Sūtrakṛtāṅgasūtram, with the Niryukti of Ācārya Bhadrabāhu Svāmī and the commentary of Śīlāṅkācārya. Ed. Sāgarānandasūrijī Mahārāja. New edition by Muni Jambūvijayajī. Delhi: Motilal Banarsidass. 1978. Slaje, Walter (1986): "Untersuchungen zur Chronologie einiger Nyāya-Philosophen." StII 11/12, 245278. Slaje, Walter (1994): Vom Mokṣopāya-Śāstra zum Yogavāsiṣṭha-Mahārāmāyaṇa. Wien: ÖAW. (SAWW, 609. Bd.) Slaje, Walter (2005): “Locating the Mokṣopāya.” The Mokṣopāya, Yogavāsiṣṭha and Related Texts. Ed. Jürgen Hanneder. Aachen: Shaker Verlag. (Geisteskultur Indiens, Texte und Studien, 7.) Pp. 21-35. Smith, Brian K. (1989): Reflections on Resemblance, Ritual and Religion. New York - Oxford: Oxford University Press. Smith, Brian K. (1994): "The Veda and the authority of class: Reduplicating structures of Veda and varṇa in ancient Indian texts." Authority, Anxiety, and Canon. Essays in Vedic Interpretation. Ed. Laurie L. Patton. Albany: State University of New York Press. Pp. 67-93. Solomon, Esther A. (1974): The Commentaries of the Sāṃkhya Kārikā — A Study. Ahmedabad: Gujarat University. Śrīdhara: Nyāyakandalī. Ed. J. S Jetly and Vasant G. Parikh. Vadodara: Oriental Institute. 1991. (GOS, 174.) Śrīvāstavya, Santanārāyaṇa (1983): Carakasaṃhitākī dārśanika pṛṣṭhabhūmi = The philosophical background of Charaka-Saṃhitā. Ilāhābāda: Pīyūṣa Prakāśana. Steinkellner, Ernst (1999): “The Ṣaṣṭitantra on perception, a collection of fragments.” AS 53(3), 667677. Tachikawa, Musashi (1994): "The concept of universal in Bhāvaviveka's writings." AS 48(2), 891-902. Tatia, Nathmal (1951): Studies in Jaina Philosophy. Varanasi: P.V. Research Institute. Tatia, Nathmal (tr.)(1994): Tattvārtha Sūtra: That Which Is, [by] Umāsvāti/Umāsvāmī with the combined commentaries of Umāsvāti/Umāsvāmī, Pūjyapāda and Siddhasenagaṇi. San Francisco etc.: HarperCollins. Tattvārthasūtra. 1) Tattvārthādhigamasūtra ... by ... Śrī Umāsvāti ... together with ... his own gloss elucidated by Śrī Siddhasenagaṇi, edited with introduction in English & Sanskrit by Hiralal Rasikdas Kapadia. Part I. Bombay: Jivanchand Sakerchand Javeri. 1926. 2) Tattvārtha-Vārtika of Śrī Akalaṅkadeva, edited with Hindi translation, introduction, appendices, variant readings, comparative notes etc., by Mahendra Kumar Jain. Vol. 1. Third edition. New Delhi: Bharatiya Jnanpith. 1989. Thieme, Paul (1957): "Pāṇini and the pronunciation of Sanskrit." Studies Presented to Joshua Whatmough on his Sixtieth Birthday. 's-Gravenhage. Pp. 263-270. = KlSchr pp. 612-619. Thieme, Paul (1978): "Über einige dualistische Bezeichnungen von Himmel und Erde im Veda." KZ 92, 32-49. = KlSchr II pp. 889-910. Thieme, Paul (1982): "Meaning and form of the ‘grammar’ of Pāṇini." StII 8/9, 3-34. = KlSchr II pp. 1170-1201. Torella, Raffaele (1994): The Īśvarapratyabhijñākārikā of Utpaladeva with the Author's Vṛtti: Critical edition and annotated translation. Roma: IsMEO. (SOR, LXXI.) Tucci, Giuseppe (1981): Pre-Diṅnāga Buddhist Texts on Logic from Chinese Sources. Second edition. Madras: Vesta. Uddyotakara: Nyāyavārttika. For the edition, see Nyāya Sūtra. Ui, H. (1917): The Vaiśeshika Philosophy according to the Daśapadārtha-Śāstra. Chinese text with introduction, translation and notes. Edited by F. W. Thomas. Second edition: Chowkhamba Sanskrit Series Office, Varanasi, 1962. (ChSSt, 22) Upadhye, A. N. (ed.)(1964): Śrī Kundakundācārya's Pravacanasāra (Pavayaṇasāra) ... critically edited with the Sanskrit commentaries of Amṛtacandra and Jayasena and a Hindī commentary of Pāṇḍe Hemarāja, with an English translation of the text, a topical index and the text with various readings, and with an exhaustive essay on the life, date and works of Kundakunda 117 118 and on the linguistic and philosophical aspects of Pravacanasāra. Agas: Shrima Rajachandra Ashram. Vātsyāyana: Nyāya Bhāṣya. See Gautama. Vetter, Tilmann (1968): "Zur Bedeutung des Illusionismus bei Śaṅkara." WZKSO 12-13 (1968-69; Festschr. Erich Frauwallner), 407-423. Vetter, Tilmann (1978): "Die Gauḍapādīya-Kārikās: zur Entstehung und zur Bedeutung von (A)dvaita." WZKS 22, 95-131. Vetter, Tilmann (1979): Studien zur Lehre und Entwicklung Śaṅkaras. Wien. (Publications of the De Nobili Research Library, 6.) Viyāhapannatti. 1) Bhagavaī Viāhapaṇṇattī, ed. Muni Nathamal, Ladnun: Jain Viswa Bhārati, V.S. 2031 (Anga Suttāni, 2). 2) Viyāhapaṇṇattisuttaṁ, ed. Bechardas J. Doshi and Amritlal Mohanlal Bhojak, Bombay: Shrī Mahāvīra Jaina Vidyālaya, 1974-1982, 3 parts (Jaina-Āgama-Series, 4). 3) Bhagavatī Sūtra, Prakrit text with English translation and notes based on the commentary of Abhayadeva Sūri, by K.C. Lalwani 4 vols. (Śatakas 1-11), Calcutta: Jain Bhawan, 1973-1985. Vyomaśiva: Vyomavatī. Ed. Gaurinath Sastri. Varanasi: Sampurnanand Sanskrit Vishvavidyalaya Press. 2 tomes. 1983-1984. Wada, Toshihiro (1990): Invariable Concomitance in Navya-Nyāya. Delhi: Sri Satguru Publications. (Sri Garib Dass Oriental Series, 101.) Walser, Joseph (2002): “Nāgārjuna and the Ratnāvalī: new ways to date an old philosopher.” Journal of the International Association of Buddhist Studies 25(1-2), 209-262. Wezler, A. (1981): "Studien zum Dvādaśāranayacakra des Śvetāmbara Mallavādin, I. Der sarvasarvātmakatvavāda." Studien zum Jainismus und Buddhismus: Gedenkschrift für Ludwig Alsdorf. Ed. Klaus Bruhn and Albrecht Wezler. Wiesbaden: Franz Steiner. (ANISt 23.) Pp. 359408. Wezler, Albrecht (1982): "Paralipomena zum sarvasarvātmakatvavāda (I): Mahābhāṣya zu Pāṇ. 4.3.155 und seine einheimische Erklärer." WZKS 26, 149-166. Wezler, Albrecht & Motegi, Shujun (1998): Yuktidīpikā. The most significant commentary on the Sāṃkhyakārikā, critically edited. Stuttgart: Franz Steiner. (ANISt 44.) Willemen, Charles (1975): The Essence of Metaphysics: Abhidharmahṛdaya, translated and annotated. Bruxelles (Publications de l'Institut des Hautes Études Bouddhiques, Série "Études et Textes", N. 4). Willis, Janice Dean (1979): On Knowing Reality: The Tattvārtha Chapter of Asaṅga’s Bodhisattvabhūmi. New York : Columbia University Press. Wood, Thomas E. (1990): The Māṇḍūkya Upaniṣad and the Āgama Śāstra: An investigation into the meaning of the Vedānta. Honolulu: University of Hawaii Press. Wujastyk, Dominik (1993): Metarules of Pāṇinian Grammar. Vyāḍi’s Paribhāṣāvṛtti critically edited with translation and commentary. Groningen: Egbert Forsten. 2 tomes. (Groningen Oriental Studies, 5.) Yoga Bhāṣya. In: Pātañjalayogadarśana of Maharṣi Patañjali, along with Vyāsabhāṣya. Edited with Yogasiddhi Hindi commentary, by Suresh Chandra Shrivastava. Varanasi: Chaukhamba Surbharati Prakashan. 1988. (Chaukhamba Surbharati Granthamala, 140.) Yuktidīpikā. 1) Edited by Ram Chandra Pandeya. Delhi: Motilal Banarsidass. 1967. 2) Critically edited by Albrecht Wezler and Shujun Motegi. Vol I. Stuttgart: Franz Steiner. 1998. (ANISt 44.) 118 119 ABBREVIATIONS AAWL Abhidh-d Abhidh-hṛ(A) Abhidh-k(VP) Abhidh-k-bh(D) Abhidh-k-bh(P) Abhidh-sam Abhidh-sam-bh Abhidh-sam(R) ABORI AitĀr AitBr AitUp ALB Amṛtar(B) ANISt Apte AS ASS ATBS BBhS Bd. Bhag BM BORI BOS Bṛh-Up BS BSOAS BST Ch-Up ChSSt DNC GK GOS HIL HistDh HOS IBK Abhandlungen der Akademie der Wissenschaften und der Literatur, Mainz, Geistes- und Sozialwissenschaftliche Klasse Abhidharmadīpa with Vibhāṣāprabhāvṛtti, ed. P. S. Jaini, Patna 1959 (TSWS 4) I. Armelin (tr.), Le coeur de la loi suprême, Traité de Fa-cheng, Abhidharmahṛdayaśāstra de Dharmaśrī, Paris 1978 Vasubandhu, Abhidharmakośa, traduit et annoté par Louis de La Vallée Poussin, 6 vols., Paris 1923-1931 Abhidharmakośa and Bhāṣya of Ācārya Vasubandhu with Sphuṭārthā Commentary of Ācārya Yaśomitra, pts. 1-4, ed. Swami Dwarikadas Shastri, Varanasi 1970-1973 (BBhS 5, 6, 7, 9). Vasubandhu, Abhidharmakośa Bhāṣya, ed. P. Pradhan, rev. 2nd ed. Aruna Haldar, Patna 1975 (TSWS 8) Asaṅga, Abhidharmasamuccaya, ed. P. Pradhan, Santiniketan (VBS 12) Abhidharmasamuccaya Bhāṣya, ed. N. Tatia, Patna 1976 (TSWS 17) Asaṅga, Le compendium de la superdoctrine (Abhidharmasamuccaya), tr. W. Rahula, Paris 1971 (PEFEO 78) Annals of the Bhandarkar Oriental Research Institute, Poona Aitareya Āraṇyaka Aitareya Brāhmaṇa Aitareya Upaniṣad The Brahmavidyā, Adyar Library Bulletin, Madras José van den Broeck, La saveur de l'Immortel (A-p'i-t'an Kan Lu Wei Lun), La version chinoise de l'Amṛtarsa de Ghoṣaka (TI 1553), Louvain 1977 (PIOL 15) Alt- und Neu-Indische Studien, Hamburg V. S. Apte, Sanskrit-English Dictionary, 3 vols., Poona 1957-1959. Asiatische Studien, Études Asiatiques, Bern Ānandāśrama Sanskrit Series, Poona Arbeitskreis für Tibetische und Buddhistische Studien, Universität Wien Bauddha Bharati Series, Varanasi Band Bhagavadgītā Bibliothèque du Muséon, Louvain Bhandarkar Oriental Research Institute, Poona Bhandarkar Oriental Series, Poona Bṛhadāraṇyaka Upaniṣad Brahma Sūtra Bulletin of the School of Oriental and African Studies, University of London, London Buddhist Sanskrit Texts, Darbhanga Chāndogya Upaniṣad Chowkhamba Sanskrit Studies, Varanasi Dvādaśāraṃ Nayacakram of Mallavādin, with the commentary Nyāyāgamānusāriṇī of Siṃhasūri, 3 parts, ed. Muni Jambuvijaya, Bhavnagar: Sri Jain Atmanand Sabha (Śrī Ātmānanda Jaina Granthamālā no. 92, 94, 95), 1966, 1976, 1988. Āgamaśāstra of Gauḍapāda, ed. Bhattacharya, 1943 Gaekwad's Oriental Series, Baroda A History of Indian Literature, ed. J. Gonda, Wiesbaden 1973 f. History of Dharmaśāstra, by Pandurang Vaman Kane, second edition, Poona 1968 f. Harvard Oriental Series, Cambridge Mass. Indogaku Bukkyōgaku Kenkyū, Journal of Indian and Buddhist Studies, Tōkyō 119 120 IIJ IL IS IsMEO IT JA JAOS JBBRAS Jg. JIABS JIP Ki. KlSchr KZ Laṅkāv(V) LDS MadhK(deJ) MadhK(V) MaiU Mbh MCB Mhbh MīBh Mil MīS MN Mppś NBh NS ÖAW P. PB PEFEO PEW PIOL Prs PrsV Prv PrvBh PTSD PTS Renou, Terminologie RV Saptapadārthī Indo-Iranian Journal, Den Haag, Dordrecht Indian Linguistics Indische Studien, Beiträge für die Kunde des indischen Altertums, ed. A. Weber, vol. 1-18, Berlin, Leipzig 1850-1898 Istituto Italiano per il Medio ed Estremo Oriente, Roma Indologica Taurinensia, Torino Journal Asiatique, Paris Journal of the American Oriental Society, New Haven Journal of the Bombay Branch of the Royal Asiatic Society Jahrgang Journal of the International Association of Buddhist Studies, Madison Journal of Indian Philosophy, Dordrecht Praśastapādabhāṣyam with the commentary Kiraṇāvalī of Udayanācārya. Ed. Jitendra S. Jetly. Baroda: Oriental Institute. 1971. (GOS 154.) Kleine Schriften [in der Serie der Glasenapp-Stiftung], Wiesbaden, Stuttgart Zeitschrift für vergleichende Sprachforschung auf dem Gebiet der Indogermanischen Sprachen, begründet von A. Kuhn, Göttingen (Saddharma)laṅkāvatārasūtra, ed. P. L. Vaidya, Darbhanga 1963 (BST 3) Lalbhai Dalpatbhai Series, Ahmedabad Nāgārjuna, Mūlamadhyamakakārikā˙, ed. J. W. de Jong, The Adyar Library and Research Centre, Madras 1977 Nāgārjuna, Madhyamakaśāstra, ed. P. L. Vaidya, Darbhanga 1960 (BST 10) Maitrāyaṇī Upaniṣad Mahābhāṣya of Patañjali, ed. Kielhorn, 3 vol., third edition by K. V. Abhyankar, Poona: Bhandarkar Oriental Research Institute, 1962-1972. Mélanges chinois et bouddhiques, Bruxelles Mahābhārata, crit. ed. V. S. Sukthankar et al., Poona 1933 f. (BORI) Mīmāṃsā Bhāṣya de Śabara (ASS 97) Milindapañha, ed. V. Trenckner, London 1880 Mīmāṃsā Sūtra (ASS 97) Majjhima-Nikāya, ed. V. Trenckner, R. Chalmers, 3 vols. London 1888-1899 (PTS) Étienne Lamotte, Le traité de la grande vertu de sagesse de Nāgārjuna (Mahāprajñāpāramitāśāstra), T. 1-5, Louvain 1944-1980 (BM 18, PIOL 2, 12, 24). [translation of TCTL]. Nyāya Bhāṣya of Vātsyāyana Nyāya Sūtra of Gautama Österreichische Akademie der Wissenschaften, Wien Aṣṭādhyāyī de Pāṇini Pañcaviṃśa Brāhmaṇa (= Tāṇḍya Mahā Brāhmaṇa) Publications de l'École Française d'Extrême-Orient, Paris Philosophy East and West, Hawaii Publications de l'Institut Orientaliste de Louvain, Louvain Dignāga, Pramāṇasamuccaya Dignāga, Pramāṇasamuccayavṛtti Dharmakīrti, Pramāṇavārttika (ed. R. Gnoli, SOR 23) Prajñākaragupta, Pramāṇavārttikabhāṣya The Pali Text Society's Pali-English Dictionary, ed. T. W. Rhys Davids, W. Stede, London 1921 Pali Text Society, London L. Renou, Terminologie grammaticale du sanskrit, Paris 1942-44 (reprinted in one volume, Librairie Champion 1957) Ṛgveda Śivāditya's Saptapadārthī with a commentary by Jinavardhana Sūri, ed. J. S. Jetly, 1963 (LDS 1) 120 121 SAWW ŚBr SK SOR StII SWTF Taitt-Br Taitt-S Taitt-Up TCTL TI TSWS VBS Viy VKNAWL Vkp VS(C) WSTB vt. WI WZKM WZKS WZKSO YBh YD YS YV ZDMG Sitzungsberichte der Akademie der Wissenschaften in Wien, Phil.-hist. Kl., Wien Śatapatha Brāhmaṇa, ed. A. Weber, Berlin-London 1855 Sāṃkhya Kārikā Serie Orientale Roma, Roma Studien zur Indologie und Iranistik Sanskrit-Wörterbuch der buddhistischen Texte aus den Turfan-Funden, begonnen von Ernst Waldschmidt, ed. Heinz Bechert, Göttingen 1973 f. Taittirīya Brāhmaṇa Taittirīya Saṃhitā Taittirīya Upaniṣad Ta chih tu lun, (Mahāprajñāpāramitopadeśa), TI 1509 Taishō Shinshū Daizōkyō ou Taishō Issaikyō, 100 vols., Tōkyō 1924 f. Tibetan Sanskrit Works Series, Patna Vishva-Bharati Studies, Santiniketan Viyāhapannatti Verhandelingen der Koninklijke (Nederlandse) Akademie van Wetenschappen, Afd. Letterkunde, Amsterdam Bhartṛhari, Vākyapadīya, ed. W. Rau, Wiesbaden 1977 Vaiśeṣikasūtra of Kaṇāda with the commentary of Candrānanda, ed. Muni Śrī Jambuvijayaji, reprint: Oriental Institute, Baroda, 1982 Wiener Studien zur Tibetologie und Buddhismuskunde, Wien vārttika Word Index to the Praśastapādabhāṣya: A complete word index to the printed editions of the Praśastapādabhāṣya, by Johannes Bronkhorst & Yves Ramseier, Delhi: Motilal Banarsidass, 1994 Wiener Zeitschrift für die Kunde des Morgenlandes, Wien Wiener Zeitschrift für die Kunde Südasiens, Wien Wiener Zeitschrift für die Kunde Süd- und Ostasiens, Wien Yoga Bhāṣya. Three editions have been used: 1) Pātañjalayogadarśanam, ed. Nārāyaṇa Miśra, Vārāṇasī: Bhāratīya Vidyā Prakāśana, 1971. 2) Pātañjalayogasūtrāṇi, ed. Kāśīnātha Śāstrī Āgāśe, Pune: Ānandāśrama (ASS 47), 1904. 3) see Rukmani, 1981 ff. Yuktidīpikā: 1) Edited by Ram Chandra Pandeya. Delhi: Motilal Banarsidass. 1967. 2) Critically edited by Albrecht Wezler and Shujun Motegi. Vol I. Stuttgart: Franz Steiner. 1998. (ANISt 44.) Yoga Sūtra Yogavārttika of Vijñānabhikṣu. Two editions have been used: see YBh 1 and 3. Zeitschrift der Deutschen Morgenländischen Gesellschaft 121 122 CITED PASSAGES Abhidharmahṛdaya of Dharmaśreṣṭhin (= TI 1550) p. 811b l. 16-27 Abhidharmakośa Bhāṣya of Vasubandhu (= Abhidh-k-bh(P)) p. 80 l. 12 f. p. 81 l. 19 p. 138 l. 4 f. Abhidharmasamuccaya Bhāṣya (= Abhidh-sam-bh) p. 34 (§ 40 H (1) (iii)) p. 34 (§ 40 H (3) (i)-(iv)) Āgamaśāstra of Gau∂apāda, ed. Vidhushekhara Bhattacharya 1.6ab 2.32 3.27-28 3.48 4.1 4.3-5 4.14-17 Āgamaśāstravivaraṇa of Śaṅkara under GK 2.32 under GK 4.1 under GK 4.4 introducing GK 4.20 under GK 4.22 Aitareya Āraṇyaka 2.6 Aitareya Brāhmaṇa 5.22 (= 24.3) Aitareya Upaniṣad 3.1-2 Aṣṭādhyāyī of Pāṇini 1.2.64 1.4.42 1.4.49 3.4.21 Bhagavadgītā 2.16ab Bodhisattvabhūmi, ed. Nalinaksha Dutt p. 30 l. 1-16 122 123 p. 30 l. 16-23 p. 30 l. 26 - p. 31 l. 5 p. 31 l. 17-18 Brahma Sūtra 1.3.28 Brahma Sūtra Bhāṣya of Śaṅkara 1.3.28 2.1.16 2.1.18 Bṛhadāraṇyaka Upaniṣad 1.4.3 1.4.7 Bṛhadāraṇyaka Upaniṣad Bhāṣya of Śaṅkara 1.2.1 Carakasaṃhitā Vimānasthāna 3.24 Chāndogya Upaniṣad 6.1.4-6 6.2.1-2 6.3.2-3 Chāndogya Upaniṣad Bhāṣya of Śaṅkara 6.1.4 6.2.2 Dvādaśāraṃ Nayacakram of Mallavādin (= DNC) p. 36-37 p. 37 p. 441 l. 5 p. 455 l. 1-2 p. 456 l. 1-2 p. 459 l. 1-2 p. 459 l. 1-4 p. 459 l. 8-9 p. 460 l. 1-2 p. 462 l. 2 p. 462 l. 3-5 p. 462 l. 6-7 p. 462 l. 11-12 p. 462 l. 8 - p. 463 l. 1 p. 463 l. 1-3 p. 463 l. 21-22 p. 512 l. 2 - p. 513 l. 3 p. 516 l. 3 - p. 517 l. 2 Īśvarapratyabhijñākārikā + Vṛtti of Utpaladeva, ed. Torella p. 28 p. 48 123 124 p. 49 p. 55 p. 56 p. 57 p. 61 p. 65 Laṅkāvatārasūtra (= Laṅkāv(V)) 3.21ab 3.22 Mahābhārata (= Mhbh) 12.224.55 + 671*.1 Mahābhāṣya of Patañjali (= Mbh) I p. 6 l. 8 I p. 7 l. 8-25 I p. 9 l. 20-23 I p. 10 l. 8-9 I p. 11 l. 1-2 I p. 137 l. 24 I p. 259 l. 4-7 I p. 259 l. 14 I p. 356 l. 11 I p. 366 l. 12-13 II p. 325 l. 17-18 Mahābhāṣyadīpikā of Bhartṛhari Manuscript p. 11a l. 11; ‘Critical edition’ Āhnika I p. 28 l. 8-9; ed. AbhyankarLimaye p. 33 l. 24 - p. 34 l. 1; ed. Swaminathan p. 40 l. 11 Manuscript p. 13a l. 4-5; ‘Critical edition’ Āhnika I p. 33 l. 24-26; ed. AbhyankarLimaye p. 40 l. 17-19; ed. Swaminathan p. 48 l. 18-19 Mahāprajñāpāramitāśāstra/°upadeśa (= TI 1509) p. 104c l. 24-26 p. 296b l. 18-19 p. 546c l. 18-29 Maitrāyaṇī Upaniṣad 6.22-23 Majjhima Nikāya (= MN) III p. 235 Manusmṛti 1.21 4.256ab 10.45 Milindapañha (= Mil) p. 25 l. 5-13 & p. 27 l. 2-14 Mīmāṃsā Bhāṣya (= MīBh) 124 125 1.1.2 (Frauwallner, 1968: 20) 1.1.5 (Frauwallner, 1968: 46) 1.1.24 1.3.33 Mīmāṃsā Sūtra 1.3.33 Mūlamadhyamakakārikā of Nāgārjuna (= MadhK(deJ)) 1.6 1.9 2.1 2.5-6 7.3 7.17 19.1-2 Nāṭyaśāstra, éd. tr. M. Ghosh 15.3 Nirukta 1.3 Nyāya Bhāṣya 2.1.49-51 2.1.52 2.1.56 2.2.12 2.2.62 4.1.15 4.1.16 4.1.49 Nyāyāgamānusāriṇī of Siṃhasūri (= DNC) p. 462 l. 26 - p. 463 l. 8 p. 463 l. 21-22 p. 37 l. 23-24 Nyāyakumudacandra of Prabhācandra, ed. Mahendra Kumar Nyāya Śāstri p. 220-21 *Nyāyānusāra of Saṅghabhadra (= TI 1562) p. 400b l. 28-29 p. 400c l. 3-7 Nyāya Sūtra 2.1.52 4.1.49 Nyāyavārttika of Uddyotakara, ed. Taranatha Nyaya-Tarkatirtha et al. p. 669 l. 15-16 p. 699 l. 1 f. p. 701 l. 5-6 125 126 Padārthadharmasaṅgraha of Praśastapāda (= WI) p. 41 § 222 p. 48 § 256-257 p. 48 § 259 p. 87 §377 Pañcaviṃśa Brāhmaṇa (= Tāṇ∂ya Mahābrāhmaṇa) 24.11.2 Pramāṇasamuccaya of Dignāga 2.5 5.1 Pramāṇasamuccayavṛtti of Dignāga 5.36d ed. Hattori p. 235 Pramāṇavārttika of Dharmakīrti (= Prv) 1.66 1.212 Pramāṇavārttikabhāṣya or Pramāṇavārttikālaṅkāra of Prajñākaragupta p. 68 l. 32-33 Prameyakamalamārtaṇ∂a of Prabhācandra, ed. Mahendra Kumar Shastri p. 618-19 Prasannapadā of Candrakīrti (= MadhK(V)) p. 39 l. 14-15 Pravacanasāra of Kundakunda, ed. Upadhye 2.8 p. 124-125 Ṛgveda 10.71.1 Sāṃkhya Kārikā 9 Sāṃkhya Sūtra of Kapila 1.118 Saṃmatitarkaprakaraṇa of Siddhasena Divākara 1.3 (p. 271) 1.11 (p. 409) Saptapadārthī of Śivāditya (= Saptapadārthī) p. 10 p. 11 l. 7 f. p. 61 126 127 *Śataka of Āryadeva (= TI 1569) p. 170c l. 13-18 p. 177b l. 26 p. 177b l. 26-27 p. 177b l. 29 - p. 177c l. 2 Śatapatha Brāhmaṇa 10.5.1.3 11.2.3.3 Śrī Bhāṣya of Rāmānuja 2.1.15 (p. 465-68) Sūtrakṛtāṅga Bhāṣya of Śīlāṅka p. 231 Taittirīya Brāhmaṇa 2.2.4 2.2.7.1 Taittirīya Saṃhitā 6.4.7.3 Taittirīya Upaniṣad 2.4 (= 2.9) Tantravārttika of Kumārila Bhaṭṭa p. 246 Vaiśeṣika Sūtra (= VS(C)) 2.1.18-19 7.2.24 9.19 Vākyapadīya of Bhartṛhari (= Vkp) 1.10 1.124 1.181 3.1.25-27 3.2.4-6 3.2.10 3.2.18 3.3.39 3.3.42 3.3.43 3.3.44-46 3.3.52 3.3.53 3.3.54cd 3.3.61ab, 62 3.3.69 3.7.7 127 128 3.7.105 vārttika of Kātyāyana (= Mbh) P. 1.2.64 vt. 35 P. 1.2.64 vt. 45 P. 1.3.2 vt. 12 Vibhāṣā (= TI 1545, 1546, 1547) TI 1545 p. 394b l. 19-27 TI 1545 p. 394b l. 27 - p. 394c l. 5 TI 1546 p. 295a l. 6-11 TI 1547 p. 465c l. 11-17 Vibhāṣāprabhāvṛtti (= Abhidh-d) p. 273 l. 29-30 p. 274 l. 5-7 Viśeṣāvaśyaka Bhāṣya of Jinabhadra, ed. Dalsukh Malvania II p. 378 (under verse 2149) II p. 385 (under verse 2183-84) II p. 538 f. (under verse 2789-93) II p. 539 (under verse 2795-96) II p. 539 f. (under verse 2796-99) Viyāhapannatti (= Bhagavatī; for the editions, see the bibliography; L = Ladnun, B = Bombay, Lal = Lalwani) 1.1.11 (L); 1.1.5 (B); vol. 1 p. 5 (Lal) 8.7.280 (L), 8.7.11 (B), vol. 3 p. 209 (Lal) 8.7.291 (L), 8.7.23 (B), vol. 3 p. 212 (Lal) 9.33.228 (L); 9.33.96 (B); vol. 4 p. 102 (Lal) 9.33.231 (L); 9.33.99 (B); vol. 4 p. 104 (Lal) 9.33.233 (L); 9.33.101 (B); vol. 4 p. 104-5(Lal) Vṛtti on Vākyapadīya, ed. Iyer p. 234 (under 2.187) Vyomavatī of Vyomaśiva I p. 36 l. 13-17 I p. 81 l. 13-21 I p. 82 l. 20 ff. I p. 85 l. 17 ff. I p. 90 l. 4 ff. II p. 126 l. 12-13 II p. 129 l. 19-27 II p. 241 l. 23-24 Yoga Bhāṣya (= YBh) 1.9 2.19 3.44 3.47 4.12 4.14 128 129 Yoga Sūtra 1.9 2.19 4.12 Yogavārttika of Vijñānabhikṣu (ed. Rukmani) 2.19, p. 110 Yuktidīpikā (= YD; P = ed. Pandeya, WM = ed. Wezler & Motegi) p. 5 l. 9-12 (P); p. 7 l. 23-27 (WM) p. 48 l. 27 (P); p. 110 l. 22 (WM) p. 52 l. 2-3 (P); p. 117 l. 7-9 (WM) p. 52 l. 16-21 (P); p. 118 l. 9-15 (WM) p. 54 l. 3-6 (P); p. 121 l. 20 - p. 122 l. 2 (WM) p. 54 l. 32 (P); p. 124 l. 7 (WM) p. 68 l. 26 (P); p. 154 l. 22 (WM) p. 68 l. 32 ff. (P); p. 154 l. 30 ff. (WM) p. 69 l. 2-5 (P); p. 155 l. 2-6 (WM) p. 69 l. 10 (P); p. 155 l. 14-15 (WM) p. 69 l. 10-14 (P); p. 155 l. 16-21 (WM) p. 91 l. 7 (P); p. 187 l. 10-11 (WM) p. 92 l. 1-3 (P); p. 188 l. 10 (WM) p. 118 l. 14-16 (P); p. 225 l. 15-19 (WM) p. 118 l. 21 ff. (P); p. 225 l. 24 ff. (WM) 129