Buddhist Philosophy of Consciousness: Tradition and Dialogue 9004440895, 9789004440890

Citation preview

Buddhist Philosophy of Consciousness

Value Inquiry Book Series Founding Editor Robert Ginsberg Executive Editor Leonidas Donskis† Managing Editor J.D. Mininger

VOLUME 354

Cognitive Science Edited by Francesc Forn i Argimon, University of Catalonia

The titles published in this series are listed at brill.com/vibs

Buddhist Philosophy of Consciousness Tradition and Dialogue Edited by

Mark Siderits, Ching Keng and John Spackman

leiden | boston

Cover illustration: Drawing Hands, by M.C. Escher, 1948. Image used with permission by The M.C. Escher Company. Names: Siderits, Mark, 1946- editor. | Keng, Ching (Philosophy teacher), editor. | Spackman, John (John G.), editor. Title: Buddhist philosophy of consciousness : tradition and dialogue / edited by Mark Siderits, Ching Keng and John Spackman. Description: Leiden ; Boston : Brill-Rodopi, [2021] | Series: Value inquiry book series, 0929–8436 ; volume 354 | Includes bibliographical references and index. | Summary: “Buddhist Philosophy of Consciousness brings Buddhist voices to the study of consciousness. This book explores a variety of different Buddhist approaches to consciousness that developed out of the Buddhist theory of non-self. Topics taken up in these investigations include: how we are able to cognize our own cognitions; whether all conscious states involve conceptualization; whether distinct forms of cognition can operate simultaneously in a single mental stream; whether non-existent entities can serve as intentional objects; and does consciousness have an intrinsic nature, or can it only be characterized functionally? These questions have all featured in recent debates in consciousness studies. The answers that Buddhist philosophers developed to such questions are worth examining just because they may represent novel approaches to questions about consciousness”-- Provided by publisher. Identifiers: LCCN 2020041632 | ISBN 9789004440890 (hardback) | ISBN 9789004440913 (ebook) Subjects: LCSH: Consciousness--Religious aspects--Buddhism. Classification: LCC BQ4570.P76 B825 2021 | DDC 181/.043--dc23 LC record available at https://lccn.loc.gov/2020041632 Typeface for the Latin, Greek, and Cyrillic scripts: “Brill”. See and download: brill.com/brill-typeface. ISSN 0929-8436 ISBN 978-90-04-44089-0 (hardback) ISBN 978-90-04-44091-3 (e-book) Copyright 2021 by Koninklijke Brill NV, Leiden, The Netherlands. Koninklijke Brill NV incorporates the imprints Brill, Brill Hes & De Graaf, Brill Nijhoff, Brill Rodopi, Brill Sense, Hotei Publishing, mentis Verlag, Verlag Ferdinand Schöningh and Wilhelm Fink Verlag. All rights reserved. No part of this publication may be reproduced, translated, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, without prior written permission from the publisher. Requests for re-use and/or translations must be addressed to Koninklijke Brill NV via brill.com or copyright.com. This book is printed on acid-free paper and produced in a sustainable manner.

Contents

Notes on Contributors vii

Introduction 1

Part 1 Conceptualism and Nonconceptualism

Introduction to Part 1 26

1

Knowing Blue: Ābhidharmika Accounts of the Immediacy of Sense Perception 31 Robert H. Sharf

2 Nonconceptual Cognition in Yogācāra and Madhyamaka Thought 62 John Spackman 3

Turning Earth to Gold: The Early Yogācāra Understanding of Experience Following Non-conceptual Cognition 89 Roy Tzohar

Part 2 Meta-cognition

Introduction to Part 2 114

4

Whose Consciousness? Reflexivity and the Problem of Self-Knowledge 121 Christian Coseru

5

Should Mādhyamikas Refute Subjectivity? Thoughts on What Might Be at Stake in Debates on Self-Awareness 154 Dan Arnold

6

Self-Knowledge and Non-Self 189 Mark Siderits

vi

Contents

7

The Genesis of *Svasaṃvitti-saṃvitti Reconsidered 209 Toru Funayama

8

Dharmapāla on the Cognition of Other Minds (paracittajñāna) 225 Shinya Moriyama

Part 3 Mental Consciousness in East Asian Buddhism: MSF

Introduction to Part 3 244

9

Mānasa-pratyakṣa as the Perception of Conventionally Real (prajñaptisat) Properties – Interpreting Dignāga’s mānasa-pratyakṣa Based on Clues from Kuiji 247 Ching Keng

10

Mental Consciousness and Its Objects 275 Zhihua Yao

11

Vasubandhu’s Theory of Memory: A Reading Based on the Chinese Commentaries 305 Chen-kuo Lin

Index 327

Notes on Contributors Dan Arnold is associate professor of philosophy of religions at the University of Chicago Divinity School. He is the author of Buddhists, Brahmins, and Belief: Epistemology in South Asian Philosophy of Religions (Columbia University Press, 2005), and of Brains, Buddhas, and Believing: The Problem of Intentionality in Classical Buddhist and Cognitive-Scientific Philosophy of Mind (Columbia, 2012). He is currently working on an anthology of original translations from India’s Madh­ yamaka tradition of Buddhist philosophy. Christian Coseru works in the fields of philosophy of mind, phenomenology, and cross-cultural philosophy, especially Indian and Buddhist philosophy in dialogue with Western (classical and contemporary) philosophy and cognitive science. Author of Perceiving Reality: Consciousness, Intentionality, and Cognition in Buddhist Philosophy (OUP, 2012) and Moments of Consciousness (OUP forthcoming), his research has been supported by grants from the National Endowment for the Humanities and the Australian Research Council. Toru Funayama is Professor in the Institute for Research in Humanities, Kyoto University, Japan. He specializes in medieval Chinese Buddhism in the Six Dynasties period, as well as in the scholastic tradition of the Yogācāra school of Indian Buddhism during the 6th through 10th centuries. His recent works include Rikuchō zuitō bukkyō tenkai shi (The Evolution of Chinese Buddhism during the Six Dynasties, Sui, and Tang Periods, Kyoto, 2019), Butten wa dou kan’yaku sareta no ka: Sūtora ga kyōten ni naru toki (Making Sutras into “Classics” [jingdian]: How Buddhist Scriptures Were Translated into Chinese. Tokyo, 2013), Kōsōden (a fourvolume Japanese translation of the Biographies of Eminent Monks, co-­ authored with Yoshikawa Tadao. Tokyo, 2009–2010). Ching Keng Ph.D. (2009, Harvard) is Assistant Professor of Philosophy, National Taiwan University. His fields of research include Abhidharma, Yogācāra, and Tathāgatagarbha thought in India and China. Recently, his interests have also extended to the Buddhist philosophy of consciousness. He has been part of various research projects studying Dharmapāla’s commentaries on the Viṃśikā and on the Ālambanaparīkṣā, Wŏnch’uk’s commentary on the

viii

Notes on Contributors

Saṃdhinirmocana-sūtra, and the development of the Three-Nature theory (trisvabhāva-nirdeśa) in Yogācāra. His larger goal is to explore how studies of Chinese Buddhist philosophical texts could contribute to a better understanding of Buddhist philosophy as a whole. Chen-kuo Lin is Professor Emeritus of Buddhist Philosophy at National Chengchi University. He also serves as Director of the Sheng Yen Center for Chinese Buddhist Studies. Currently three research projects are under his supervision: (1) “An Annotated Translation of Dharmapāla’s Cheng weishi baosheng lun,” (2) “Mapping the Buddhist Scholasticism during the Edo Period,” and (3) “Re-examining the Philosophical Debate between Bhāviveka and Dharmapāla in the Sino-Indic Context.” His recent research focuses on logic, epistemology and hermeneutics in East Asian Buddhism. One of his recent publications is A Collection of the Rare Manuscripts of the Commentaries on Dignāga’s Ālamabanaparīkṣā in Early Modern East Asia (in Chinese), co-edited with Kaiting Jien (Kaohsiung: Fo Guang Publishing Co., 2018). Shinya Moriyama is a Professor of Philosophy at Shinshu University (Japan). His primary research focus is Buddhist epistemology and its religious significance, specifically focusing on Dharmakīrti and his successors. He is the author of Omniscience and Religious Authority: A Study on Prajñākaragupta’s Pramāṇavārttikālaṅkārabhāṣya ad Pramāṇavārttika II 8–10 and 29–33 (LIT Verlag, 2014). The East Asian yinming/inmyō tradition of Buddhist logic and epistemology has also been a focus of his research, a result of which is found in his edited book Transmission and Transformation of Buddhist Logic and Epistemology in East Asia (Arbeitskreis für Tibetische und Buddhistische Studien, Universität Wien, 2020). Robert H. Sharf is D.H. Chen Distinguished Professor of Buddhist Studies in the Department of East Asian Languages and Cultures at the University of California, Berkeley, as well as Chair of Berkeley’s Center for Buddhist Studies. He works primarily on medieval Chinese Buddhism but has also published in the areas of Japanese Buddhism, Buddhist art and archaeology, Buddhist modernism, Buddhist philosophy, and methodological issues in the study of religion. He is author of Coming to Terms with Chinese Buddhism: A Reading of the Treasure Store Treatise (2002), and co-editor (with his wife Elizabeth) of Living Images: Japanese Buddhist Icons in Context (2001).

Notes on Contributors

ix

Mark Siderits works primarily in analytic Asian philosophy. He did his BA at University of Hawaii; his Ph.D. is from Yale. He retired from Seoul National University in 2012, having previously taught at Illinois State University. His current research interests lie in the intersection between classical Indian philosophy on the one hand, and analytic metaphysics and philosophy of language on the other. Among his more recent publications are Personal Identity and Buddhist Philosophy: Empty Persons, 2nd edition (Ashgate, 2015) and, together with Shōryū Katsura, Nāgārjuna’s Middle Way: Mūlamadhyamakakārikā (Wisdom, 2013). A collection of his papers on Buddhist philosophy, Studies in Buddhist Philosophy, was published by Oxford in 2016. John Spackman is Associate Professor of Philosophy and Director of the Program in Neuroscience at Middlebury College. He received his Ph.D. in Philosophy from Yale University, and also completed an M.A. at Columbia University in Religion. His work focuses on contemporary philosophy of mind and its intersections with Buddhist philosophy. He has published articles on such topics as debates concerning the nonconceptuality of perception and aesthetic experience, the structure of concepts, Nāgārjuna, and what contemporary work on the mindbody problem can learn from Buddhist philosophy. He is currently working on a book that develops an approach to the mind-body problem inspired by Madh­yamaka Buddhist philosophy. Roy Tzohar is an associate professor in the East and South Asian Studies Department at Tel Aviv University. He specializes in Indian, and particularly in Buddhist, philosophy of language and mind; but has also published on the philosophy of emotions and on Indian poetics. He is the author of A Buddhist Yogācāra Theory of Metaphor (Oxford University Press, 2018), winner of the Toshihide Numata Award; and co-editor (with Maria Heim and Ram-Prasad Chakravarthi) of Emotions in Classical Indian Thought (Bloomsbury Academic, forthcoming). Zhihua Yao is Associate Professor at the Chinese University of Hong Kong, Department of Philosophy. His publications include Nonexistent Objects in Buddhist Philosophy: On Knowing What There is Not (Bloomsbury Academic, 2020), Brahman and Dao: Comparative Studies of Indian and Chinese Philosophy and Religion (co-edited with Ithamar Theodor, Lexington Books, 2014), The Buddhist Theory of Self-Cognition (Routledge, 2005) and numerous journal articles.

Introduction Recent decades have seen something of a scramble to either give an account of, or else decisively refute the possibility of ownerless consciousness.1 But this sort of thing has happened before: Buddhist philosophers engaged in a centuries-long debate with their non-Buddhist opponents over whether the Buddhist doctrine of non-self could be joined with a plausible theory of cognition. Current interest in non-egological approaches to the study of the mind has been largely fueled by attempts to reconcile folk psychology with new knowledge in neurophysiology and cognitive psychology. Not so Buddhist theorizing concerning mental phenomena, which grew out of soteriological concerns. Despite this important difference, it is not just scholars of Buddhist thought but also researchers in areas like cognitive science and the philosophy of mind who are interested in finding out more about Buddhist approaches to consciousness. This volume attempts to address that interest. Buddhist philosophy of mind begins with the Buddha himself (5th century bce) and continues to develop to this day, both in Asia and in the West. Many distinct approaches have grown out of these efforts to understand the mind in a way that is compatible with the core Buddhist teaching of non-self. The scholars whose essays appear here all have expertise in some part of this long and complex tradition. But like scholars in every field, they sometimes present their findings in ways that presuppose familiarity with the disciplinary context, thereby making it challenging for those from outside the discipline to fully grasp the significance of their research. This introduction is meant to provide a framework that should help non-specialists better understand the particular theories and debates discussed in these essays, by locating those discussions in the larger project of understanding consciousness.2 Readers who are knowledgeable about the Buddhist philosophical tradition may be tempted to skip over this part and go straight to the essays. But we shall also use the framework we develop here to indicate connections between Buddhist debates and

1 See, e.g., the essays in Siderits et al. 2011; also see Gallagher 2011. 2 Needless to say, there are ongoing scholarly controversies concerning the correct interpretation of the Buddhist philosophical tradition, and some specialists will disagree with one or another facet of the reading presented here. While we stand ready to defend the framework we develop below, we judge that this is not the place to enter into those disputes. The reader interested in learning more about the Indian Buddhist philosophical tradition would do well to begin with Westerhoff 2018.

© koninklijke brill nv, leiden, ��21 | doi:10.1163/9789004440913_002

2

Introduction

recent controversies in cognitive science and the philosophy of consciousness. We hope that what we say here will prove useful to all readers regardless of disciplinary background. One more word of warning is in order concerning the essays. While all discuss some aspect of Buddhist theorizing about consciousness, some are more concerned with matters Buddhological than with general questions concerning consciousness. There is much that we still do not know about the classical Indian Buddhist philosophical tradition and its receptions in China and Tibet. Many of the scholars working today in different areas of Buddhist philosophy are mainly focused on trying to fill in the gaps in our knowledge. While many of the essays collected here attempt to bring Buddhist theorizing about consciousness into dialogue with current consciousness studies, others are principally devoted to simply trying to get clear about particular Buddhist authors or texts. The non-specialist may find the focus of such essays narrow and difficult to penetrate. This introduction is meant to help smooth the way for such readers. But we would also point out that the effort invested may be amply rewarded. Sometimes seeing familiar problems discussed in novel ways can spark important insights. Dates for major figures mentioned below: Asaṅga 4th c Vasubandhu 4th–5th c Dignāga 6th c Dharmapāla fl. mid–6th c Dharmakīrti 7th c Candrakīrti ca 700–750 Śāntarakṣita 725–88 Kamalaśīla 8th–9th c Dharmottara 8th c Kuiji 632–682 1

Early Buddhist Roots of Buddhist Views of Consciousness

Buddhism originates out of a concern to diagnose the source of existential suffering and bring it to an end. The Buddha agreed with other thinkers of his day that the sense of meaninglessness and despair that afflicts most thoughtful people grows out of a mistaken view concerning our identity. Where he disagreed with others is over what the mistake comes to. Others thought, perhaps predictably, that where we go astray is in taking the wrong sort of thing to be the self: our mistake consists in identifying with something other than our true

Introduction

3

self. The Buddha claimed instead that the mistake consists in identifying with anything at all: it is the false sense of a persisting “I,” routinely identified as subject of experiences and agent of actions, that he thought must inevitably lead to existential suffering. He then proposed a path designed to extirpate the “I”-sense and thereby bring about the cessation of suffering. The Buddha called his teachings a “middle path” between two extremes, but different pairs of extremes are named on different occasions. One pair that is important for our purposes is the pair of views called eternalism and annihilationism. The former is the view that since persons have selves that endure, death does not end the existence of the person. The latter is the view that death is the end. The middle path between these extremes is not a compromise position; it would be hard to find a middle ground between “I shall exist after my death” and “I shall cease to exist at death.” (Perhaps: “I shall be gone but live on through my children”?) Instead the Buddha articulates his middle path using what is called the 12-fold chain of dependent origination. This is a list of sequential factors, each serving as cause of the next, that are said to together constitute the continued existence of a person over several lives. What is interesting about this list is that the factors are given in completely impersonal terms. When the Buddha uses it to explain how he can reject both “I shall survive my death” and “I shall be annihilated at death,” it is clear that he is rejecting the presupposition common to both eternalism and annihilationism: that there is an “I,” something through relation to which events in this life take on meaning and value. The Buddhist doctrine of non-self is central to the project of overcoming existential suffering. The Buddha classifies the constituents of persons in three distinct ways. One such taxonomy is the skandha or aggregate classification, according to which there are five aggregates or kinds of psychophysical element: the material (rūpa), hedonic states (vedanā), perceptual identifications (saṃjñā), mental forces (saṃskāra), and the consciousness (vijñāna) aggregate. Each of the five aggregates is in turn broken down into subtypes; vedanā, for instance, is said to be made up of three kinds of hedonic state: pleasure, pain and the neutral state. In time Buddhist scholastics would compile lists of all the distinct entities, called dharmas, that were thought to make up the five aggregates. Of importance here is that consciousness aggregate consists of not one but six distinct varieties of consciousness dharma: one for each of the five external sense faculties, plus a sixth for the so-called “inner sense” (manas). (The last is the faculty of cross-modal synthesis, memory, conceptualization, and of introspection; we will have much more to say about it below, but the essay by Sharf (Chapter 1) contains a general overview.) These six kinds of consciousness are said to be distinct because each comes into existence in dependence on a

4

Introduction

different set of causes and conditions: visual consciousness arises in dependence on contact between vision and an occurrence of color-and-shape in the presence of light; auditory consciousness arises in dependence on contact between auditory faculty and sound, etc. Given the impermanence of the conditions on which consciousnesses depend, it is thought to follow that no consciousness endures for very long. Consciousness is thus not a self. The sense of an “I” as an enduring subject of experience is an illusion fostered by an unbroken succession of discrete consciousnesses, each lasting just a moment. The analogy used to illustrate this is the illusion of a wheel of fire produced when a torch is whirled rapidly. 2

The Development of a Buddhist Philosophy of Mind

What the example of the wheel of fire brings out is the extent to which early Buddhism uses causal relations among impersonal events to develop its alternative to the folk-psychological understanding of mentality. But it was left to later scholastics of the various schools known collectively as “Abhidharma” to develop these ideas in systematic ways.3 The task they set themselves was to forge an account of mental processes that was consistent with the Buddha’s teachings, and would also explain both how an illusory “I”-sense is fostered and perpetuated and how it might be extirpated. But inevitably, controversies arose. This is partly due to the fact that different schools operated with different recensions of the Buddha’s teachings (which were preserved only in oral form for some time after his death). But it is also due in part to unclarity in those teachings. It is not, for example, at all obvious what was meant by “consciousness” as the fifth element in the aggregates classification.4 One of the earliest attempts to spell this out gives the analogy of the person atop the watch-tower at the center of the city, who is said to see those approaching the city from all directions. Such a watcher is said to be aware that a person is approaching, though they cannot say who; presumably, identifying the person is the job of the guards at the city gate. The suggestion is thus that consciousness merely presents the 3 The three most important Abhidharma schools for purposes of this work are Vaibhāṣika, Sautrāntika and Theravāda. The Yogācāra school, while belonging to the later Mahāyāna movement, pursues a project very much in line with Abhidharma concerns. 4 The five aggregates or skandhas include one consisting of corporeal components, and four categories of mental events. This brings out the fact that Buddhist philosophers consistently reject physicalism. One later school, Yogācāra, is (arguably) idealist; the others have dualist ontologies, though not substance dualist.

Introduction

5

object as a bare “this”; it is a quite different mental element, one belonging to the aggregate of perceptual identifications, that is said to perform the function of basic categorization of the perceived object. But one wonders how categorization can be carried out in the absence of consciousness. The task for Ābhidharmikas (members of the different Abhidharma schools) was to sort out this and other puzzles concerning the nature and role of consciousness. For instance, one response to the early Buddhist characterization of consciousness was to take the cognition of a simple percept to be a temporally extended process, with distinct consciousnesses occurring at each moment and each accompanied by mental concomitants that perform such functions as perceptual identification and registration of hedonic valency. Consciousness is thus affirmed as a necessary component in all cognitive functioning. But if cognition of a single percept, such as the color of the mango, involves a series of successive consciousnesses, the argument for the impermanence of consciousness from its dependence on sense-object contact loses some of its force. If only the first moment in the 17 moments of consciousness said (by Theravādins) to be involved in cognition of yellow color is produced by senseobject contact, why suppose that the cognition involves 17 distinct consciousnesses rather than a single enduring state?5 And if a single enduring state can accompany many distinct mental functions, why could it not be a single enduring consciousness that serves as subject of experience across sensory modalities? These were the sorts of challenges facing Abhidharma scholastics as they tried to articulate a theory of the mental that was consistent with the Buddha’s teachings. Developments in other areas of Buddhist philosophy shaped their efforts. One result that changed the nature of the discussion is the doctrine of momentariness. The Buddha claimed that all existents are impermanent. But Abhidharma philosophers went much further, formulating arguments meant to show that everything is momentary, i.e., ceases to exist immediately upon coming into existence. This would put pressure on direct realist accounts of perception. If the object of a perceptual consciousness is among the causes of that consciousness, then it must exist before the arising of the consciousness 5 We apologize for any sense of terminological vertigo this sentence may induce by its use of “cognition” and “consciousness.” Unfortunately, there is no consistent usage of the different terms for conscious mental events across the whole tradition. One early Yogācāra text (Viṃśikā), for instance, says at the outset that it will treat four different terms as synonyms, one of which (citta) is extended to include not just consciousness (i.e., the consciousness aggregate usually termed vijñāna) but the mental concomitants (caittas) as well. In time the tradition moved toward a more consistent use of the term jñāna, which is generally translated as “cognition.” That will be our generic term for a conscious mental event.

6

Introduction

that takes it as object. And if everything is momentary, then that object must also have gone out of existence before the arising of the consciousness. But awareness of something that no longer exists would seem to count as memory, not perception. A representationalist about perception holds that what one is directly aware of in perception is not the object that impinges on the sense organ but a mental representation that is caused by that impingement. The doctrine of momentariness thus makes representationalism about perception seem more plausible. Among Abhidharma schools it was the Sautrāntikas who embraced representationalism, which they formulated as the thesis that a perceptual consciousness arises bearing the form of the object with which the sense faculty interacts. No longer is consciousness to be thought of as a bare awareness that illuminates whatever it happens to fall on. Consciousness of blue is now understood as a single state that is in some sense both blue and its illumination. A second development that played an important role in Abhidharma theorizing about consciousness is the embrace of mereological nihilism, the view that strictly speaking there are no composite objects, only the impartite entities that we think of as their parts. The Buddha had said that persons are “mere names,” and he also allegedly accepted the claim that the word “chariot” is no more than a convenient designator for the constituent parts when assembled in the right way. Generalizing from this stance, Abhidharma philosophers developed arguments to show that strictly speaking, no partite entity may be said to exist. (The strategy involves looking at all the ways in which a real whole might be related to its constituent parts.) The upshot is an ontology containing only entities that are neither composite nor qualitatively complex. Momentariness and mereological nihilism both lead to highly counter-­ intuitive results, such as that substances are merely bundles of tropes, and that persistence can at best be perdurance and not endurance. Ābhidharmikas sought to mitigate the counter-intuitiveness with their doctrine of the two truths. This involves distinguishing between two ways in which a statement may be said to be true – ultimately and conventionally – and two ways in which an entity may be said to exist – as ultimately real and as a mere conceptual construction. The ultimate reals – what the tradition calls dharmas – are those non-composite entities that enter into causal relations. Ultimately true statements are those that correctly represent how such entities are arranged; these arrangements are the truth-makers for ultimate truths. To say of an entity that it is not ultimately real but merely a conceptual construction is to say that it is conventionally taken to be real due in part to our use of a convenient designator. A convenient designator is a term that is used to denote a collection of ultimately real entities when arranged in a particular way. Our use of such terms

Introduction

7

reflects our interests and cognitive limitations; we would be hard pressed to describe how things are in the world without their employment. Statements containing such terms may be conventionally true or false, depending on how the ultimate reals happen to be arranged. The upshot is that while no statement about enduring substances, such as persons or chariots, can be ultimately true (or ultimately false), such statements can be conventionally true. Mereological nihilism rules out the existence of chariots. And momentariness rules out the endurance of any of the atoms making up the supposed chariot. Still it can be conventionally true that this is the same chariot as the one we saw earlier. What grounds this conventional truth is the fact that the collection of many atoms arranged chariot-wise that was seen earlier is part of a causal series that extends to the present collection of atoms arranged chariot-wise: each momentary atom causes a similar successor to arise in the next moment. Use of this last tool helps Ābhidharmikas answer some of the more common objections to a non-egological account of cognition. Take the objection based on appeal to cross-modal synthesis, which may be expressed in terms of the situation where one says about the mango, “I touch what I saw.” How, the opponent asks, can there be the cognition that what one is now touching is the object one first saw and then reached out for, if there is no enduring subject of experience serving to bring visual and tactile consciousnesses into coordination? The Buddhist can respond that just as our use of the convenient designator “mango” fosters the illusion that there is a single enduring object of both vision and touch, so the use of the convenient designator “person” creates the sense that there is a single enduring subject of both vision- and touch-­consciousnesses. The general strategy here is to supply causal processes among the ultimately real dharmas in order to ground conventionally true descriptions of our mental lives. Two thoroughly impersonal causal series of momentary events are usefully given the shorthand designation of, on the one hand a piece of fruit, and on the other an enduring cognitive agent. Still, the constraints imposed by mereological nihilism create challenges for this approach. Because mereological nihilism is thought to rule out qualitative complexity, it is difficult to see how there can be the simultaneous awareness of both visual and tactile data. In the case of the mango, the visual cognition is of a yellow ovoid color-and-shape occurrence, while the tangibility consciousness is of an occurrence of smoothness. There can be no sensory consciousness that takes both objects simultaneously. So how does one arrive at the awareness that the two sensory objects are somehow connected? Here is where the sixth sense-faculty, manas or “inner sense” gets brought in. This, it will be recalled, is not just the faculty of introspection – what grasps such “inner” states as emotions – but also that of conceptualization. Presumably, past experience

8

Introduction

has led to formation of a disposition, in the file labeled “mango,” that triggers certain mental states, given sensory stimulation having a certain profile. (This file will include information from all relevant sense modalities.) The perceptual identification associated with the visual consciousness outputs a “yellowovoid” categorization, which in turn triggers activation of the “mango” mental file. This in turn causes desire, which brings about action, resulting in the tangibility consciousness. And as with the example of the whirling torch that is seen as a circle of fire, the rapid succession of these states generates the illusion of a single persisting consciousness aware of all the distinct contents. 3

Some Abhidharma Controversies Involving Consciousness

This approach to cognition fosters any number of controversies among Buddhist philosophers seeking to resolve some of the issues that result from its application. For instance, there is the question whether shape is perceived by manas in the same way in which color is perceived by vision. This question grows out of the point that for some Abhidharma philosophers, shape is not ultimately but only conventionally real. Some schools claimed that since one never perceives a color without a shape, color-and-shape is a single type of material dharma. Others, though, said that since shape supervenes on an array of many atoms, shapes are mere conceptual constructions; awareness of shape is awareness of a many masquerading as a one, something that requires superimposition of a concept. Moreover, color and shape seem to be qualitatively distinct, so making color-and-shape a single ultimately real entity violates the prohibition on qualitative complexity. Yet the awareness of shape seems to be just as direct and immediate as the visual awareness of color. One proposed solution involves the assumption that manas functions simultaneously with one or another form of external sensory consciousness. This would represent an extension of the view discussed earlier that each moment of consciousness is accompanied by occurrences of various mental functions. Since manas is said to perform the function of conceptualization, this would seem consistent with that approach. But insofar as manas is understood as a sixth sense faculty, its operation is said to involve its own distinctive kind of consciousness. So invoking manas to explain cognition of shape would require that visual consciousness and manas consciousness occur simultaneously. The general question involved here is whether two or more sorts of consciousness can arise simultaneously in a single subject (or, to put it more strictly, within a single mental stream of causally connected mental events). The assumption among non-Buddhist Indian philosophers is that a subject

Introduction

9

can be in no more than a single conscious state at any moment. Attention is one of the functions said to be invariably concomitant with consciousness, and attention is generally understood to be a limited resource. Some Abhidharma schools agree. The difficulty they confront is that the Buddha seems to have held that perceptual identifications, along with certain other mental functions, occur simultaneously with sensory consciousness. This is the source of that (unhelpful) analogy of consciousness as the watcher atop the watchtower; the idea seems to be that it is other mental events that perform functions like perceptual identification (e.g., seeing the object as yellow). The difficulty is that these look like functions that involve consciousness. Perceptual identifications, for instance, seem to involve the use of concepts, and it is mental consciousness that is said to be involved in conceptualization. The claim that there can be multiple consciousnesses occurring in a single mental stream simultaneously looks like one way around the difficulty. But the overarching question behind all this is the extent to which consciousness can be thought of as a single simple event, and not as the aggregate of a variety of distinct modular processes.6 Is consciousness itself ultimately real, an impartite and qualitatively simple entity? The Buddha seems to have thought so, but can that claim be made fully coherent? This question grows particularly pressing in the Abhidharma debate over meta-cognition. Indian philosophers generally agree that at least some cognitions are themselves cognized. Not only can one cognize yellow, one can at least sometimes cognize one’s cognizing yellow: we attribute mental states to ourselves as well as to others. Various theories were developed to account for this, roughly corresponding to what are today known as first-order (reflexive) and higher-order accounts, the latter consisting of the higher-order perception (hop) and higher-order thought (hot) accounts. Having a theory to explain the possibility of meta-cognition was particularly pressing for Buddhists for the following reason. The Buddha’s path to the cessation of suffering involves extirpating the “I”-sense, and this is said to depend on coming to see that all of the aggregates are characterized by the three marks of impermanence, suffering and non-self. Suppose I were to survey all the aggregates constituting this psychophysical complex and conclude that all are characterized by the three 6 We would now say that the perceptual awareness of a yellow oval patch is the output of a series of successive types of processing of visual stimulation: first color detection, then edge detection, then shape detection. Insofar as shape detection occurs downstream from color detection, Abhidharma was right to see its cognition as involving contributions from the side of the cognizer. What they failed to consider is the possibility that the further processing involved might be cognitively impenetrable, so that the deliveries of the visual processing system might have the vividness that they took as a sign of raw unprocessed data.

10

Introduction

marks. What of the cognition that is involved in this surveying? Does it cognize its own state of having the three marks? That would seem to violate the intuitively plausible principle of irreflexivity: just as a fingertip cannot touch itself, so it would seem a cognition cannot cognize itself. This might seem like a quibble, in that the cognition that performs the survey can always be surveyed retrospectively. But this is thought to be inadequate to the soterial task of extirpating the “I”-sense. Buddhist doctrine has it that while one can have theoretical knowledge of the impermanence, etc., of the aggregates, final extirpation of the “I”-sense requires a direct and immediate awareness of all the psychophysical elements as having the three marks. (This is supposedly achieved through meditational practice.) The consciousness involved in this cognition is included in this totality. Some Ābhidharmikas bit the bullet and accepted the resulting violation of irreflexivity. (We have more to say about this below.) Others would reject the requirement of direct and immediate awareness, taking a hot approach to the question how the crucial cognition is cognized. But some took advantage of the possibility opened up by the hypothesis that there can be multiple consciousnesses operating simultaneously: the hypothesis that conscious concomitants perform their functions simultaneously with consciousness, and that these functions can be tantamount to consciousness. The idea is that one can be aware in a single moment of the impermanence, etc., of all the elements because at the same time that a single consciousness cognizes all of the elements save one (viz. itself), some of the concomitants of that consciousness take it as object of cognition. It is, once again, far from clear just what the nature and function of consciousness are thought to be in this discussion. Another issue that looms large in Abhidharma is the boundary between perception, considered to be non-conceptual, and thought, which is understood as necessarily requiring the use of concepts. Sharf’s essay (Chapter 1) discusses the Abhidharma controversy over how this line is to be drawn. But the same controversy erupts with even more force following the development of Dignāga’s system, to which we are about to turn. 4

The Yogācāra School and Dignāga

We have so far been speaking of Abhidharma schools, but there is one important development in Buddhist philosophy of consciousness that involves a school not typically classified as Abhidharma (though there is much overlap in both approach and concerns between it and classical Abhidharma). This is the Yogācāra school, which is best known for its espousal of subjective idealism or

Introduction

11

consciousness-only (cittamātra). Its idealism can be seen as the natural outgrowth of the representationalist view of sense perception discussed earlier in connection with the doctrine of momentariness.7 It is with Yogācāra that the notions of multiple simultaneous consciousnesses, and of unconscious mental states, are fully developed. The latter development comes by way of the posit of something called a “storehouse consciousness” (ālaya-vijñāna), which serves as the repository of the karmic traces that are said by idealists to be what generate sense-perception. (The storehouse consciousness thus plays one of the roles that Berkeley reserved for God.) The posit of this repository, which was originally invoked in order to account for continuity of memory and other mental functions over periods of apparent lack of consciousness (such as some meditational trance states), opened up the possibility of explaining sensory experience in wholly immaterial terms. Yogācāra took up that possibility for an interesting reason: it was felt that if it could be shown that there is no external world, then the very idea of an inner subjective realm becomes less than fully intelligible. For this reason, Yogācāra metaphysics might better be called “subjectless idealism.” But it may be worth pausing to consider the following anomaly: the “storehouse consciousness” Yogācāra uses to explain the origins of sensory experience in the absence of external objects is among the unconscious mental states accepted by that school. What this brings out is the role sometimes played by considerations of Buddhist orthodoxy in theory construction. Four of the five aggregates (viz. all but the first) are immaterial. Of these, the second, third and fourth aggregates are said to occur only in the presence of the fifth, consciousness (vijñāna). Thus when early Yogācāra resorted to the notion of mental traces, which belong to the fourth aggregate (that of the saṃskāras), to account for memory continuity over periods of unconsciousness, it felt compelled to posit another sort of consciousness in addition to the traditional six. It is not clear just what work this storehouse consciousness is doing if it functions during periods of unconsciousness. Yogācāra holds that one has no access to the karmic traces deposited in storehouse consciousness until they are activated and give rise to one of the six traditionally acknowledged kinds of consciousness. But if storehouse consciousness does not provide access to the traces, and the traces are what cause conscious experience, just what does the storehouse do? It is not clear that the word “consciousness” in “storehouse consciousness” marks anything more than the fact that the traces all belong to a 7 It should be pointed out that some Buddhologists have recently denied that Yogācāra espouses metaphysical idealism. For discussion of the controversy see Westerhoff 2018: 176–179.

12

Introduction

single mental stream. Recent attempts at theorizing consciousness are sometimes criticized for failing to make clear exactly what it is that is being investigated. Given that most people take themselves to have unmediated access to their own conscious states, we can see how this situation might arise: it is assumed that we all have an intuitive grasp of what it is we are talking about. In any event, the same difficulty seems to have arisen in the Buddhist tradition as well. Among the most important Yogācāra philosophers is Dignāga, founder of the Yogācāra-Sautrāntika school. The name of this hybrid school has an odd sound to it, given that Yogācāra is idealist while Sautrāntika espouses representationalist realism. But Dignāga’s aim was not to reconcile the metaphysics of the two schools. His aim was to develop an epistemology that could be used by Buddhists, regardless of their metaphysical commitments, in debates with their non-Buddhist opponents. In the Indian tradition, an epistemological theory is a theory of the number and nature of the epistemic instruments, those processes that invariably cause veridical cognitions. The central claim of Dignāga’s epistemology is that there are just two epistemic instruments, perception and inference. This follows, he claims, from the fact that there are just two sorts of cognizables: pure particulars, which are the objects of perception; and generalized objects, which are cognized by inference. The difference may be illustrated by the contrast between seeing a fire before one’s eyes, and inferring the existence of an unseen fire after perceiving smoke. Suppose that all fires are known to be red. The red that one cognizes by perceiving a fire is a particular hue, and that hue will vary in dependence on the fire seen. The red that one cognizes when one infers the existence of a fire is of a generic hue, and that hue does not vary with variation of the hue of the inferred fire. Moreover, strictly speaking there is no such thing as generic redness. If there were universals such as redness, they would be permanent, and every existing thing is impermanent. What one cognizes through inference is not ultimately but only conventionally real. Entities that are merely conventionally real are also said to be conceptual constructions. This is because all conceptualization involves the use of universals. As such, conceptualization falsifies the nature of the real, although in ways that prove by and large useful for creatures like us. This is why when I see smoke and say there is fire, what I say is conventionally true. Dignāga thus draws a bright line between perception and thought, thereby reigniting an earlier debate between conceptualist and non-conceptualist accounts of cognition in Abhidharma. Already in classical Abhidharma there were difficulties in explaining the transition from perception to thought if perception is understood as non-conceptual. Buddhist philosophers usually

Introduction

13

discuss the distinction between perception and thought in terms of the contrast between “one cognizes blue” (nīlam vijānāti) and “one cognizes that it is blue” (nīlam iti vijānāti). The iti in the second sentence is a quotation device, making it clear that the cognition in question involves the use of language and thus of concepts. The first sentence is meant to describe visual consciousness, the second to describe the achievement of manas-consciousness. The question is how one gets from the first state to the second if the first is nonconceptual. The perceptual identification accompanying a visual consciousness will output the basic category blue, which is of course a concept. There is no special mystery as to how a particular stimulus can evoke a stereotyped response. The smoke detector sounds the same alarm regardless of the specific properties of the smoke (or, for that matter, steam from the shower). There is no need to endow the senses with conceptual capacities in order to explain how perception might make its content available to manas for reflection. What the debate over the transition concerned was how veridical perception can give rise to cognitions that are thought to distort reality. For it is, after all, conceptualization that makes possible the identification of the many constituents of the causal series of psychophysical elements as an “I.” Vasubandhu’s response is that the output of perceptual identifications is a natural-kind concept such as blue, something that is not inherently falsifying. The mischief occurs downstream from this function. It is Dignāga and his commentator Dharmakīrti who change the parameters of this dispute. This is because they fully recognize the involvement of universals in all concept deployment.8 The theory of apoha they develop as a radical nominalist alternative to realism about universals brings out the dependence of concepts on the interests and cognitive limitations of the individual or species that employs them. This may still leave room for the possibility that conceptualization plays a role in guiding us to effective action in everyday life. But their soteriological project seems to demand access to how things are independently of the concepts we happen to employ. Dignāga’s account of the epistemic instruments has perception playing that role, but not inference. This is how the Dreyfus-McDowell debate comes to be anticipated in the Buddhist tradition. The non-Buddhist Nyāya school of classical Indian philosophy holds that perceptual cognitions involve two distinct phases: an initial non-conceptual phase in which the subject is aware of the individual constituents of a perceptual judgment; and a subsequent conceptual phase in which those individual 8 The point here may be put as: conceptualism is not a viable alternative to realism about universals.

14

Introduction

constituents are cognized as constituting a relational complex. Thus in the visual perception of a patch of blue color one is first aware of a blue trope, the universal blueness, and the relation of inherence (the relation whereby blueness may be said to be in the blue trope). Only after this does one come to be aware of the state of affairs, blue-inhered-in-by-blueness. This account is defended on the grounds that one can only cognize a relational complex if one has first cognized the relata. Nyāya adds that since verbal expression necessarily involves a subject-predicate form, the initial phase of perceptual cognition is inexpressible. This account is mirrored in Dharmakīrti’s theory of perceptual judgment. On his account, the perception of blue involves two distinct, successive cognitions: first a perception of a unique particular (a trope occurrence on the standard Yogācāra-Sautrāntka ontology); then a judgment that may be expressed as “blue.” The judgment is driven by the initial perception: it involves the production of a mental image that is at once a copy of the perceptual representation, and also something grasped as what is common to all members of the exclusion class not non-blue. But where Nyāya has a single two-phase perceptual cognition, Dharmakīrti has two distinct cognitions, only the first of which counts as perception. This difference is partly driven by a difference in ontologies: where Nyāya has both tropes and universals, plus inherence as the glue that holds universals to individual tropes, Yogācāra-Sautrāntika has only the unique tropes. But the spartan ontology in turn reflects the radical nominalism of the system. There being no universals, there are no real resemblances either. That the many particulars we call “blue” seem similar to us is the result of species-relative interests that generate affordances for cognitive systems with those interests.9 Perceptual judgments are inferences, and the concepts they deploy are not as benign as Vasubandhu’s notion of natural kind concepts would suggest. Still Dharmakīrti’s story does begin to explain how the transition from raw perception to concept-laden thought might be possible. The clarity does not last long though. The waters are muddied in several discussions having to do with transitions between perception and thought.10 The question of the transition from a non-conceptual state to a conceptual mode of cognition arises in connection with the soteriologically important phenomenon known as “subsequent cognition” (pṛṣṭhalabdhajñāna). What such cognition is subsequent to is a sort of trance state induced through advanced meditation practices and regularly described as non-conceptual. Entry 9 10

For an exposition of Dharmakīrti’s semantics of apoha (exclusion) in terms of Gibsonian affordances see Chatterjee 2011. There is a useful discussion of the rules covering these transitions in Kellner 2020.

Introduction

15

into this state is said to mark the initial attainment of cessation of suffering (i.e., nirvāna). The difficulty is that its being nonconceptual would seem to mean that any content it had cannot be communicated to others. In that case, how are buddhas and bodhisattvas to exercise their characteristic virtue of compassion, by teaching to others what they have thereby learned? Hence the need for a subsequent state that does involve conceptualization, making communication possible. But once again, it is unclear how a non-conceptual state might have communicable content. The difficulty is only compounded when, as in some texts, the prior trance state is identified as the cessation absorption (nirodha samāpatti), which is said to be utterly devoid of consciousness. Equally puzzling is the claim made in later Yogācāra-Sautrāntika concerning what is called habituated perception. Suppose that observation conditions are less than ideal, so that one cannot rule out either of two competing hypotheses as to what one observes: that might be water in the distance, or it might just be shimmering sunlight. On further investigation one concludes that what one saw is water, since one’s thirst is slaked. The claim is that after many such episodes, the step of gathering confirming evidence can be skipped: through a process of habituation one now sees the water in the distance. Now this may seem like a perfectly valid description of the process whereby skilled perception is developed: the trained jeweler can tell “just by looking” whether a given stone is natural or synthetic, whereas the layperson must perform various tests. The difficulty lies in fitting this phenomenon into Dignāga’s strictly dichotomous scheme. Habituated perception is classified as a kind of perception, based in part on its phenomenal character of vividness, but also on the speed with which the trained expert carries it out. This should make it nonconceptual in nature. But this is hard to reconcile with Dharmakīrti’s claim that the infant’s seeing the breast as a source of nourishment is evidence that the infant has conceptual resources gleaned from experience in prior births. For a radical nominalist like Dharmakīrti, the “seeing as” phenomena of affordances can only be explained in terms of processes of mental construction. The skilled jeweler’s ability must be understood as involving “unconscious” inference. And this is precisely what is denied when that ability is classified as habituated perception. These difficulties led another Buddhist school, Madhyamaka, to reject Dignāga’s dichotomy and insist that all perceptual cognition involves concepts. (Some later Naiyāyikas likewise questioned their school’s claim that perception involves a non-conceptual phase.) This rejection may reflect the fact that Madhyamaka is less concerned than other Buddhist schools to claim a certain sort of epistemic privilege for insights allegedly attained through meditation. The Buddhist soteriological project is crucially dependent on its promise to

16

Introduction

end our ignorance about the ultimate nature of reality. Once Dignāga had proclaimed all conceptualization to be distorting, it became tempting to look to the altered states of consciousness produced through various yogic practices as a source of purified perception. We see this not only in the case of “subsequent cognition” but also in the theory of habituated perception, which was used to explain how focused concentration might yield veridical cognition of a totality (as in the cognition that all dharmas are impermanent). Madhyamaka’s core claim, however, is not that the ultimate nature of reality transcends all conceptual capacity and is only accessible through a kind of non-dual consciousness. Its core claim is rather that because the supposedly ultimate reals can be shown to be empty of intrinsic nature (and thus fail to meet the accepted criterion of ultimate reality), the very idea of the ultimate nature of reality turns out to be incoherent. Our best theories will inevitably be built around our interests and cognitive limitations, and it is a mistake to suppose that they require grounding in something that prescinds from those facts about us. It is a mistake to suppose that consciousness is something ultimately real with its own intrinsic nature. Our best theory of perception will characterize it along functionalist lines. And that requires that perception be seen as performing basic categorization. Non-conceptual perception is a myth. There is, we might add, another way of resolving the non-conceptuality problem that is not discussed in any of the essays in this volume, but is at least hinted at in the work of the Yogācāra-Sautrāntika commentator Dharmotttara. The basic idea is that there is a levels distinction involved in the nonconceptual/conceptual dichotomy, so that it is a mistake to look for a way of understanding the transition from one sort of state to the other. More specifically, it might be that when we look at consciousness events at the scale of the momentary, none can be appropriately called conceptual. It is only when we ascend to the scale of cognitions, understood as sequences of consciousness events and so temporally thick, that the use of patterns and regularities becomes visible. It may then be that perceptual cognitions are inextricably bound up with conceptuality, but that this is so only because of the level at which it makes sense to speak of cognitions as either perception or thought. 5

Buddhist Accounts of Meta-cognition

The controversy over meta-cognition also takes on new life after Dignāga, who developed arguments for the reflexivity thesis, the thesis that every cognition cognizes itself. His commentator Dharmakīrti provides the example of a light as a way of refuting the principle of irreflexivity: just as a light may illuminate

Introduction

17

not only the other objects in a room but also itself, so a cognition may cognize not only its intentional object but also itself as cognizer. The main arguments for the reflexivity thesis are two. The first, Dignāga’s argument from memory, attempts to show that the alternative higher-order perception account of meta-cognition runs afoul of the principle that one only remembers what was previously experienced. It is agreed that after perceiving blue, one can come to cognize one’s cognizing blue: meta-cognition can be achieved through introspection. But since all cognitions are momentary, the introspective awareness must involve memory: the visual cognition no longer exists, and one is aware of things that no longer exist only by means of memory. Since we only remember that which we previously experienced, it follows that there must have been an earlier cognizing of the seeing of blue that is now being remembered. And since the only time at which that earlier cognizing would not once again involve memory is the time of the seeing of blue, it follows that the cognition of blue also involved the cognition of the cognizing of blue. The cognition of blue is also the cognizing of the cognition of blue. Cognitions are reflexive in nature. The second proof, arguably formulated by Dignāga but clearly developed by Dharmakīrti, is meant to show that any higher-order account of meta-­cognition (whether hop or hot) generates a vicious infinite regress. The idea is that the higher-order cognition posited by any ho approach will require a yet higherorder higher-order cognition in order to count as the cognition of the first-­ order target cognition. Dharmakīrti’s argument begins with the point that blue and the cognition of blue are never experienced separately. (Dharmakīrti is here presupposing a representationalist account of perception.) This may be put as: it is not the existence of blue but the cognition of blue that leads to effective action (such as successfully picking blueberries). Dharmakīrti takes this to show that one is aware of the object of a cognition only if one is aware of the cognition whereby one comes to be aware of the object. Since cognitions are by nature the means whereby one is capable of action with respect to objects, it follows that every cognition must itself be cognized. But taking the ho approach to meta-cognition would mean that it is a distinct cognition that cognizes the cognition of blue. And unless this higher-order cognition were reflexively self-cognizing, it would in turn require yet a higher higher-order cognition to perform its function. Since the series thus generated has no clear terminus, one must conclude that every cognition cognizes itself. The careful reader will have noticed that there is a slide at a crucial step in Dharmakīrti’s argument. One can agree that it is not the existence of blue but the cognition of blue that leads to effective action. But one can still reject the claim that it is not the existence of the cognition of blue but the cognition of

18

Introduction

the cognition of blue that leads to effective action. A blueberry-picking machine need not reflect on its blue-detecting representation in order to successfully pick blueberries. Only the refrigerator light illusion (the light is always on whenever I look) can justify this step in Dharmakīrti’s argument. This gap opened the door to Madhyamaka’s espousal of the rival hot account of meta-cognition. A hot account need not involve memory, so it is immune to Dignāga’s argument from memory. And the hole in Dharmakīrti’s argument returns the situation to the status quo ante, with meta-cognition once again understood as not mandatory but optional. Mādhyamikas reject the alleged counter-example to irreflexivity, the light that supposedly illuminates itself as well as other objects. As Nāgārjuna pointed out, light could properly be said to be illuminated only if it were possible for light to also be in the unilluminated state, which it cannot: there is no light in the dark. So there is reason to think that the reflexivity account cannot possibly be right. And an alternative is possible: meta-cognition might be explained as the deployment of a theory of mind that posits mental states in order to explain behavior, such as saying “blue” or reaching for a blueberry. Two additional points should be mentioned concerning the reflexivity thesis. The first is that Dignāga and Dharmakīrti were well aware that attributing two forms to cognition threatens the status of cognition as an ultimately real dharma. Mereological nihilism would dictate that an entity that has both the form of the object and the form of a cognition must be construed as a conceptual construction – a bundling together of two distinct reals for purposes of conceptual economy. They are thus quite insistent on the point that objectform and cognition-form are strictly speaking identical. They naturally recognize that it is quite difficult to say just what that would mean: how can an image or representation of blue be identical with an occurrence of illumination or “presencing”? They claim, though, that this difficulty stems from the fact that all conceptually mediated representation is structured by the subject-­ object dichotomy. Since the identity of object-form and cognition-form is affirmed as the ultimate nature of cognition, and the ultimate is held to transcend that dichotomy, it should come as no surprise that the identity of the two forms is inexpressible and so not fully comprehensible to minds still deluded by conceptual superimpositions. The second, related point concerns the question of the existence of other minds. This question was raised early in the development of Yogācāra. A realist opponent wonders how the idealist Yogācāra can defend the claim that there are other minds, given that minds are ordinarily individuated in terms of the bodies with which they are thought to be associated. Vasubandhu’s reply is that mental streams may be individuated in terms of the number and nature of

Introduction

19

causal connections among mental events. We have reason to suppose there are other mental streams to which we do not have unmediated access because under the right circumstances, events in those streams may cause events in our own, as when we hear the utterance “blue” without having a prior intention to speak. A complicating factor here is the claim that buddhas directly perceive mental events in other mental streams. Vasubandhu concedes that he cannot say what it would be like for a buddha to be aware in this intimate “first-personal” way of all mental events everywhere. But he blames this on his own cognitive limitations. The potential for embarrassment here is substantial, since solipsism is incompatible with the Buddhist ideal of a buddha (or a bodhisattva) as someone who seeks to relieve the suffering of others. In the absence of other minds, what room is there for the exercise of compassion? So one needs some reassurance that the claim that buddhas are directly aware of mental events in other mental streams does not lead to solipsism. The situation grows worse for Yogācāra after Dignāga and Dharmakīrti claim that the two forms of a cognition are ultimately identical. Dharmakīrti argues for the existence of other minds along the same lines as Vasubandhu. But a later Yogācāra-Sautrāntika, Ratnakīrti, argues that a Buddhist idealist must be a solipsist. His argument turns on the point that if a cognition’s two forms are ultimately identical, there can be no difference between first-person and third-­person attributions of mental states like pain. If the cognition of pain is just pain’s selfpresentation, what can it mean to say that there is pain that I don’t feel? 6

East Asian Buddhist Discussions of Consciousness

Our focus so far has been on developments in Indian Buddhist philosophy of consciousness. But the East Asian Buddhist tradition has interesting things to say on the topic as well. And since this part of Buddhist thought has not been as widely discussed, some historical background is necessary. We earlier described Dharmakīrti as Dignāga’s commentator, but in fact his commentaries soon supplanted Dignāga’s pathbreaking work, so that the YogācāraSautrāntika tradition in India and Tibet consists primarily of commentaries on Dharmakīrti’s work. The situation is reversed in East Asia, where Dignāga and Dharmapāla, but not Dharmakīrti, hold center stage. By examining the East Asian Buddhist reception of Dignāga’s theories, we may obtain an interesting perspective on their content. The reception of Dignāga in East Asia begins with the translation, attributed to Paramārtha (499–569), of Ālambanaparīkṣā, Dignāga’s argument for

20

Introduction

Yogācāra idealism. Xuanzang (602–664) provided a second translation of this text, as well as translating Dignāga’s Nyāyamukha together with Jīnaputra’s commentary Nyāyapraveśa. After Xuanzang, Yijing (635–713) translated Dharmapāla’s commentary on the Ālambanaparīkṣā, which is now incomplete, together with Dharmapāla’s extensive commentary on Vasubandhu’s Viṃśikā, which is also incomplete. These two commentaries attributed to Dharmapāla are extant only in Chinese translation, but they are quite conservative in the sense that Dharmapāla does not provide much original interpretation beyond what is found in the root texts. In addition to translating Dignāga’s texts, Xuanzang also compiled the Cheng weishi lun (“Establishment of [the thesis that everything is] merely consciousness and mental representations,” henceforth abbreviated as cwsl). According to the tradition, the cwsl is a selective compilation drawn from the ten commentaries on Vasubandhu’s Triṃśikā. In this text, we find very sophisticated views and debates among the Buddhist thinkers after Vasubandhu, including Dignāga, Sthiramati and Dharmapāla. This text has long been regarded as the authoritative text in the East Asian Yogācāra tradition, known as the Faxiang School (faxiang zong 法相宗) in China and Hossō School [hossō shū法 相宗] in Japan. Chief figures in this tradition include Xuanzang’s two disciples: Kuiji (632–682; aka Master Ji) and Korean monk Wŏnch’ŭk (613–696). Now a careful study of the legacy of Xuanzang could potentially contribute to a better understanding of Indian Buddhist philosophy between Vasubandhu and Dharmakīrti for two reasons. First, we have scant extant sources from the Yogācāra tradition during that period. The works of the so-called ten masters during that time, with the exception of Sthiramati, do not survive at all. Even the works of Dharmapāla are extant only in Chinese translations. We can peer into Yogācāra philosophy only through the narrow window of those translations and commentaries written in Chinese. Second, in contrast to the Indo-Tibetan epistemological tradition that was heavily influenced by Dharmakīrti, the East Asian Yogācāra tradition does not receive any influence from Dharmakīrti. Recently, scholars have increasingly come to realize the differences between Dignāga and Dharmakīrti. So the East Asian Yogācāra tradition may provide a vantage point from which we can reconstruct a pre-Dharmakīrti reception of Dignāga. A case of interest in this connection is the notion of mental consciousness (i.e., consciousness produced by the functioning of the inner sense faculty, manas) simultaneous with the five sensory consciousnesses (henceforth abbreviated as msf for Mental consciousness Simultaneous with Five). This is an intriguing notion that only rarely made its appearance in Indian and Tibetan texts. The idea is that, in addition to the mental consciousness that arises

Introduction

21

subsequently to one of the five kinds of sensory consciousness, there is another kind of mental consciousness that arises simultaneously with the arising of a sensory consciousness. Now it is intriguing to note that the commentaries written in Chinese construe this as representing what Dignāga has to say about his mysterious notion of mental perception (mānasa-pratyakṣa). In contrast, Dharmakīrti left this notion out of his epistemology, and the whole postDharmakīrti tradition actually had a hard time making sense of Dignāga’s mental perception. After Dharmakīrti, most of the commentators on Dignāga simply treated this notion as being devoid of any genuine philosophical significance. According to them, Dignāga’s stipulation was merely due to his respect for scriptural authorities. Here we see a clear case where the East Asian Yogācāra tradition and the post-Dharmakīrti Indo-Tibetan tradition give very different interpretations of Dignāga. On the other hand, however, there are also doubts about the faithfulness of the accounts offered by the East Asian Yogācāra tradition. For example, Kuiji in his commentaries often attributes certain doctrines to Sthiramati and Dharmapāla, which however are not found in currently extant works of those authors. An interesting case in this connection is the “four-fold division of cognition” schema. Dignāga had claimed that a cognition may be understood in terms of a three-fold analysis: the object-aspect, the aspect of the cognition as such (as something that grasps its content), and reflexive cognition. Insofar as Dignāga holds that the first and second aspects are ultimately identical, he also holds that the third state of reflexive awareness is actually just what we think of as the cognizing aspect apprehending what we take to be the object aspect. Kuiji attributes to Dharmapāla the claim that there is a fourth element in cognitions, the awareness of reflexive awareness, but this “four-fold division” schema is more likely to be of Chinese origin.11 If so, this would be interesting evidence concerning how the reflexivity thesis might be received in the absence of the sustained discussion of meta-cognition that took place in the Indian context. For this move looks like the beginning of precisely the infinite regress that Dharmakīrti warned would follow from a higher-order account of metacognition. The “four-fold division” schema seeks to avoid that regress by making reflexive awareness and awareness of reflexive awareness mutually dependent. Still this does look more like the sort of higher-order perception model championed by the non-Buddhist Nyāya school in India, than the first-order 11

This is why the title of Funayama’s essay (Chapter 7) contains an asterisk: the fourth of the four elements would be svasaṃvitti-saṃvitti if it occurred in Sanskrit; the asterisked “*svasaṃvitti-saṃvitti” in his title indicates that this is what the Sanskrit would have been if the Chinese usage did trace back to an Indian text. Funayama is doubtful.

22

Introduction

account of Dignāga. Perhaps the East Asian commentators found the violation of irreflexivity incurred by the reflexivity account simply too difficult to swallow. To conclude, the Yogācāra tradition in East Asia established by Xuanzang may prove useful for constructing a picture of Indian Buddhist philosophy before Dharmakīrti, one that might be very different from what we have reconstructed based largely on the currently available Indian and Tibetan sources. But reliance on the Chinese sources may also be dangerous; we have to be cautious when resorting to them. The papers in this volume aim to provide interesting case studies, as well as encouragement for further study of the Chinese sources. The essays in this collection represent the current state of play in the scholarly examination of Buddhist philosophy of consciousness. As we have tried to bring out, there are any number of connections between what the Buddhist philosophical tradition has had to say about the nature and functioning of consciousness, and recent work in the philosophy of mind. For readers new to the investigation of an Asian philosophical tradition, we hope fresh insights will be triggered from looking at the issues from a new perspective. And for readers who are already familiar with Buddhist thinking on the question of consciousness, we hope the essays included here suggest new and fruitful ways of looking at the tradition. The essays have been arranged in three clusters, each representing a particular focus, and each of the three sections is introduced by a brief preface laying out some of the common themes of the papers in that section. First is a group of three papers on the conceptualist/non-conceptualist debate in Buddhist philosophy of mind. The second section consists of five papers addressing the problem of meta-cognition. Finally there are three papers discussing the distinctive treatment in East Asian Buddhist philosophy of what is known as mental consciousness. All deal specifically with the claim found in that tradition that one kind of consciousness caused by the inner sense (manas) operates simultaneously with the consciousnesses brought about by operation of the five external sense faculties (the so-called msf thesis). Acknowledgment This book received generous support from the project "Cultivation of Young Scholars, Philosophy of Consciousness," (Grant Number, 102-2420-H-038-004MY3) funded by Taiwan's National Science Council (now, Ministry of Science and Technology). We would like to thank the project leader, Professor Timothy J. Lane.

Introduction

23

References Chatterjee, Amita. 2011. “Funes and Categorization in an Abstraction-free World.” In Apoha: Buddhist Nominalism and Human Cognition. Mark Siderits, Tom Tillemans and Arindam Chakrabarti, eds. New York: Columbia University Press. Pp. 247–257. Gallagher, Shaun, ed. 2011. The Oxford Handbook of the Self. Oxford: Oxford University Press. Kellner, Birgit, 2020, “Using Concepts to Eliminate Conceptualization: Kamalaśīla on Non-conceptual Gnosis (nirvikalpajñāna),” Journal of the International Association for Buddhist Studies vol. 43. Siderits, Mark, Dan Zahavi and Evan Thompson, eds. 2011. Self, No Self? Perspectives from Analytical, Phenomenological and Indian Traditions. Oxford: Oxford University Press. Westerhoff, Jan. 2018. The Golden Age of Indian Buddhist Philosophy. The Oxford History of Philosophy. Oxford: Oxford University Press.

Part 1 Conceptualism and Nonconceptualism

∵

Introduction to Part 1 The question of whether certain kinds of mental states are nonconceptual, independent of conceptualization, played an important role in the development of the Buddhist philosophical tradition in at least two respects, one epistemological and one soteriological. At the epistemological level, questions concerning the transition between perception and thought, and whether perception requires some conceptual content in order to provide adequate grounding for thought, troubled Buddhist thinkers as much as they did their Western counterparts. As Robert Sharf’s paper indicates, such concerns were operative from the earliest stages of the Abhidharma tradition. The second, soteriological significance of the notion of nonconceptuality is one that comes into real prominence only later, beginning perhaps in the 4th–5th centuries, in Yogācāra and Yogācāra-Sautrāntika thought. The Yogācāra school, for instance, attributes to buddhas and bodhisattvas a type of nonconceptual cognition (nirvikalpajñāna) that plays a key role in that school’s conception of liberation. For Yogācāra, all conceptualization – though it may be useful at the conventional level – falsifies reality, for conceptualization requires universals, and all things being impermanent, there are no universals. Thus, if liberation is fundamentally a recognition of the reality of non-self, then the attainment of nonconceptual cognition – through arduous practice – becomes crucial for emancipation. Corresponding to these two roles for the notion of nonconceptuality are two quite different types of arguments for viewing the mental states in question as nonconceptual. The arguments that seem to drive the notion that perception, in contrast with thought, is nonconceptual, in some respects run parallel to historical and contemporary arguments for nonconceptuality in the West. In the Abhidharma context at any rate, these arguments are predicated on the contrast between direct and indirect cognitive access. Thus as Sharf points out, for instance, the Vaibhāṣika, an Abhidharma exegetical tradition, held that each of the five sense consciousnesses grasps its object directly, without the overlay of conceptual categories. By contrast, mental consciousness, which is also involved in sensory perception, grasps the objects of the senses only as mediated by its deployment of concepts. In the later Yogācāra-Sautrāntika school, sense perception is viewed as nonconceptual because perception, unlike thought, is the cognition of the particular, and as such does not employ universals. Both of these arguments, which derive the nonconceptuality of sense perception from either its directness or its cognition of particulars, have affinities with the long tradition of Western philosophy that associates perception with

© koninklijke brill nv, leiden, ���1 | doi:10.1163/9789004440913_003

Introduction to Part 1

27

immediacy and the particular, and thought with mediation and the universal. This tradition survives in contemporary arguments for the nonconceptuality of perception, such as the fineness of grain argument, as well.1 In his contribution, Robert Sharf explores the intricacies of this type of argument in relation to Abhidharma theories of perception, and raises questions about the viability of the Ābhidharmika project of securing perception as wholly nonconceptual. As he indicates, the Vaibhāṣika recognize that although perception does not involve linguistic concepts, it does necessarily involve a minimal discriminative capacity, the ability for instance to register the difference between a blue patch and a red patch. To explain this sensory ability they posit a system of three types of discrimination (vikalpa), the thinnest of which, “inherent discrimination” (svabhāvavikalpa), is necessarily involved in sense cognition. Inherent discrimination involves only the operation of vitarka, the initial, imprecise discernment of objects, and thus purports to be a kind of discrimination that is nonconceptual. However, Sharf suggests that the Vaibhāṣika struggled to fit this account with the orthodox Sarvāstivāda position that sense cognition involves both vitarka and vicāra, a more refined form of discernment, and thus adopted the view that while both are in some sense present, only vitarka is operative in sense cognition. Sharf questions, however, the coherence of such an account. It is unclear, first, what it would mean for vicāra to be passively present but not operative. But more broadly, he suggests that it is hard to make sense of the discernment involved in sense cognition independently of some form of recollection (smṛti) and cognizance (prajñā), and if this is the case, this discernment begins to look more and more like a form of conceptualization.2 Turning to the soteriological role of nonconceptuality, it is somewhat harder to find Western analogues for the kind of argument that underlies the Yogācāra claim that the highest cognition of reality, the cognition available to buddhas and bodhisattvas, must be nonconceptual. Though Yogācāra texts raise multiple considerations in support of this idea, what is perhaps the fundamental argument is based on the notion that conceptualization is by nature dualistic. In a nutshell, the argument seems to be that conceptualization necessarily involves a distinction between the “grasper” and the “grasped,” the subject (or perhaps the subject’s act of grasping) and the intentional object; but 1 For the fineness of grain argument, see for instance Evans 1982, Peacocke 1992. For one prominent response, see McDowell 1994. 2 Sharf’s argument may find a parallel, to some degree at least, in John McDowell’s (1994) claim that perception draws passively on conceptual capacities that are used actively in thought and reflection.

28

Introduction to Part 1

that given the reality of non-self, the correct cognition of reality must be nondualistic, transcending the duality of subject and intentional object; and hence, that this cognition must be nonconceptual.3 The claim that the awareness of buddhas and bodhisattvas is nonconceptual immediately raises a challenge, however, what Spackman calls “the teaching problem.” After their liberation, buddhas and bodhisattvas resolve to remain in the world to liberate all sentient beings. But since such teaching, along with everyday interactions in the mundane world, is naturally taken to involve the use of concepts, how can the liberated state of buddhas and bodhisattvas be nonconceptual? One possible response to this problem, canvassed by a number of Yogācāra texts, is to hold that after achieving nonconceptual cognition, buddhas and bodhisattvas can enter into a different kind of state termed “subsequent nonconceptual cognition” (pṛṣṭhalabdhanirvikalpakajñāna), which allows for the employment of concepts, but in a way that is in some sense transformed by the experience of nonconceptuality. The papers by Spackman and Tzohar present different accounts of this notion of nonconceptual cognition. Spackman considers a number of different models of nonconceptual cognition, drawn from both Yogācāra and Madhyamaka texts, that seek to respond to the teaching problem in different ways. The goal of his paper, he emphasizes, is not to reconstruct the views of any particular school or schools, but rather to draw on the insights presented by these models to develop a plausible account of what nonconceptual cognition might be. Drawing attention to several challenges faced by the models, he argues for two claims. Note that the models based on subsequent cognition are “twotiered”: they solve the teaching problem by holding that only states of meditative absorption in buddhas and bodhisattvas are nonconceptual, while everyday states involving subsequent cognition are conceptual. Spackman argues first that we should instead favor a single-tiered model, which holds that at a deeper level the two types of cognition are nonconceptual in the same sense, since both nonconceptual cognition and subsequent cognition are held to be nondelusory in the same respect, in that they avoid the reification of subject and object. Second, he argues for a supervenience version of conceptualism, which rejects the notion of nonconceptual content and holds that the content of mental states supervenes on a subject’s concepts. He thus supports what he calls the nondual cognition model, according to which what makes the mental states of buddhas and bodhisattvas nonconceptual is not that they possess nonconceptual content or no content at all, nor that they avoid conceptual 3 For a fuller presentation of this argument, see the paper by Spackman below.

Introduction to Part 1

29

acts such as reflection and deliberation, but that the subjects adopt a non-­ reifying, nondualistic attitude toward their experiences and actions. Tzohar’s contribution focuses not only on the nature of nonconceptual cognition, but also on its relationship to the subsequent cognition that follows after it, as discussed in such Yogācāra texts as the Mahāyānasaṃgraha, Vasubandhu’s Triṃśikā, and Sthiramati’s commentary on it. As Tzohar points out, these texts clearly view subsequent cognition as essential to the bodhisattva’s pursuit of their vow to save all sentient beings. For although subsequent cognition involves the application of concepts, this application is transformed by the experience of nonconceptuality in such a way that the process of conceptualization does not reify its objects, but regards all phenomena as essenceless. Subsequent cognition thus becomes the only means by which bodhisattvas can fulfill their vow. It represents, then, as we have seen, one natural solution to the teaching problem. As Tzohar indicates, one function performed by subsequent cognition, which is crucial in allowing bodhisattvas to fulfill their vow, is to provide them with an error-free knowledge of cause and effect, and thus of the aspirations and needs of sentient beings. Tzohar argues as well that this idea of subsequent cognition offers a striking inversion of standard views concerning the relation between nonconceptual states and language. For on this account, the notion of nonconceptuality is not completely divorced from language, but rather, as mediated by subsequent cognition, serves as the crucial background against which we can understand the real nature of language and thought, as functioning in a non-reifying way. In reading these essays, it may be helpful to reflect on how the issues they raise relate to contemporary controversies concerning nonconceptuality, including debates about what it means to call a mental state nonconceptual. For instance, in recent discussions about whether perceptual experience is nonconceptual, a contrast has frequently been drawn between state nonconceptualism and content nonconceptualism (Heck 2000). To say that an experience is nonconceptual in the sense of state nonconceptualism, is to say that it is not necessary, in order to have that experience, that the subject possess all of the concepts that characterize its content – that is, that the experience outruns the subjects’ concepts. To say that an experience is content nonconceptual, by contrast, is to say that its content is different in kind from the contents of paradigmatic conceptual states like beliefs and thoughts – perhaps, for instance, iconic rather than propositional. A state may be content nonconceptual, but state conceptual, and likewise state nonconceptual but content conceptual. It is not altogether clear which of these senses of nonconceptuality is operative in the views discussed by these papers, but this is a potentially fruitful area for future inquiry. Spackman maintains that the notion of nonconceptuality that

30

Introduction to Part 1

is most relevant in the Buddhist contexts considered here is in fact neither state nonconceptualism nor content nonconceptualism, but supervenience nonconceptualism. On the other hand, Tzohar suggests that the Yogācāra approach to nonconceptual cognition actually breaks down the sharp distinction between nonconceptual and conceptual states assumed in many contemporary discussions. At any rate, how the Buddhist views on nonconceptuality considered here relate to contemporary debates is a question that invites further reflection. Works Cited Evans, G. 1982. The Varieties of Reference, ed. John McDowell (Oxford: Oxford University Press). Heck, R. 2000. “Nonconceptual Content and the ‘Space of Reasons,’” Philosophical Review, vol. 109, pp. 483–523. McDowell, J. 1994. Mind and World (Cambridge, MA: Harvard University Press). Peacocke, C. 1992. A Study of Concepts. (Cambridge, MA: mit Presss).

Chapter 1

Knowing Blue: Ābhidharmika Accounts of the Immediacy of Sense Perception Robert H. Sharf Abstract What is the place of immediate or nonconceptual or unconstructed cognition (nirvikalpajñāna, wu fenbie zhi 無分別智) in the Buddhist tradition? This turns out to be a complex and contentious issue. In a previous paper, I argued that Buddhist scholiasts did not foreground nirvikalpajñāna in their accounts of cognition until the rise of the Pramāṇavāda tradition associated with Dignāga and Dharmakīrti. Prior to Dignāga, nirvikalpajñāna was generally regarded as an exotic yogic attainment, the purview of buddhas and advanced practitioners. It was only with the rise of Pramāṇavāda that nirvikalpajñāna is reconstrued as a pre-reflective and self-intimating feature of all states of cognition. In short, it is only after Dignāga that unconstructed cognition (something akin to the modern phenomenological notion of “pre-reflective consciousness”) would come to play a central role in Buddhist theories of mind. In this chapter I turn to the closest candidate we have for unconstructed cognition in earlier (pre-Pramāṇavāda) exegesis. This is the Ābhidharmika notion that the five sense consciousnesses apprehend their object-supports directly or immediately, as opposed to the sixth consciousness – mind consciousness (manovijñāna) – which alone has the capacity for conceptual discrimination. In an oft-repeated example, visual consciousness is said to know “blue” but not “this is blue”; it is mind consciousness that knows “this is blue.” In this paper I will explore Vaibhāṣika, Sautrāntika, and early Yogācāra writings on the immediacy of sense perception; do they in fact ascribe nonconceptual “content” to sense consciousness, and if so, how is it understood?

Keywords Abhidharma – Buddhist theories of perception – non-­conceptual cognition – vitarka – vicāra.

…

© koninklijke brill nv, leiden, ���1 | doi:10.1163/9789004440913_004

32

Sharf

Is there such a thing as direct, non-conceptual experience, or is all experience, by its very nature, conceptually mediated?1 Is some notion of non-conceptual sensory awareness required to account for our ability to represent and negotiate our physical environment, or is it merely an artifact of deep-seated but ultimately misguided metaphysical assumptions? Perhaps conscious experience in humans is inextricably tied to the representational or self-reflexive capacities of language; if so, does it necessarily follow that newborn infants and non-human animals are not conscious? Is the very notion of non-conceptual experience logically incoherent or unintelligible? Or perhaps the problem lies in our use of concepts to describe phenomena that lie, by definition, beyond the confines of conceptual thought. These are, of course, much debated issues in contemporary philosophy of mind, sometimes framed in terms of “first-order” (or “same-order”) theories of consciousness versus higher-order theories. (Among the higher-order theories, a further distinction is made between “higher-order perception” [a.k.a. hop, or “inner sense” theories] versus “higher-order thought” [hot] theories.)2 Meanwhile, philosophers working in the areas of cognitive science and perception have engaged in parallel debates over “non-conceptual mental content,” “nonlinguistic thought,” and even “non-linguistic conceptual thought”—subjects that bear directly on the relationship between thought, concepts, language, and perception. Here too we find a burgeoning and highly technical literature, in which cognitivists face off against phenomenologists, conceptualists take on non-conceptualists, “state non-conceptualism” is championed over “content non-conceptualism,” and so on.3 The growing complexity and sophistication of the literature on non-conceptual mental content has not brought the field closer to consensus on whether “non-conceptual experience” is intelligible in the first place. Some years back there was hope that the notion of “qualia” might contribute focus and clarity to these issues. More recently attention has turned to the distinction between “phenomenal consciousness” (p-­consciousness) and “access consciousness” (a-consciousness).4 Yet irrespective of whether one talks of qualia, or p-consciousness, or first-order consciousness, or same-order consciousness, or intransitive consciousness, or pre-reflective awareness, the 1 A longer version of this chapter appeared in Philosophy East and West (vol. 63, no. 3, July 2018), pp. 826–870. 2 The philosophical literature in this area is vast; on higher-order thought, see, for example, Rosenthal 1986, 1993, 2005; Carruthers 2011; Lau and Rosenthal 2011. On the same-order side, see esp. Strawson 2015. 3 Here too the literature is quite large. I provide references when I return to the topic in my concluding remarks. 4 See Block 1995, as well as the responses to Block’s nomenclature and Block’s rejoinder (1997) in Behavioral and Brain Sciences 1995 (18): 247–269, and 1997 (20): 144–166.

Knowing Blue

33

challenge remains the same: how to make conceptual sense of an experience that is alleged to be non-conceptual? Despite the difficulties, many philosophers continue to be drawn to these issues. The interest is sustained, in part, by contemporary research in cognitive science and artificial intelligence that promises to revolutionize our understanding of consciousness and perception. But despite the wealth of new empirical data and the profusion of increasingly sophisticated philosophical arguments, the underlying quandaries go back to the dawn of philosophical reflection, and the verdict is still out on whether the spate of new work bespeaks progress (however that might be measured) or is mere “conceptual proliferation” (prapañca, xilun 戲論, “frivolous discourse”). The notion of unmediated or non-conceptual experience has also emerged as a topic of debate among scholars of religion. One early and somewhat fashionable theory of mysticism, now known as “perennialism,” holds that all the major world religions are historically and spiritually grounded in a single ineffable mystical experience. Some exponents of perennialism have gone further, arguing that the experience is contentless: it is “pure consciousness” itself.5 Starting in the 1970s, however, this popular theory came under sustained attack by “contextualists” (a.k.a. “constructivists”), who argued, on philosophical, historical, and ethnographic grounds, that all experience, including so-called mystical experience, is socially, culturally, and conceptually mediated. The notion of a single perennial experience that lies behind the diversity of religious traditions is, according to the contextualists, a modernist conceit based on a systematic misreading of the historical record.6 Determining where the Buddhist tradition stands on these issues is no easy task. In my earlier forays into the mysticism debates, I argued that Buddhist conceptions of the path articulated in early sūtra and abhidharma literature, as well as in mārga (“stages of the path”) compilations such as the Visuddhimagga, do not construe the goal of Buddhist practice as a sort of direct or unmediated experience (Sharf 1995, 1998). Which is not to say that immediate (or non-conceptual or unconstructed) experience was unknown in the early tradition. The notion of “non-discriminative” or “non-conceptual discernment” (nirvikalpajñāna, wu fenbie zhi 無分別智) appears sporadically early on as a sort of exotic yogic attainment, the purview of buddhas and advanced

5 Of course, any attempt to describe or impart this experience will, of necessity, be culturally mediated. For a philosophical defense of the perennialist position see Stace 1960. For a defense of the “pure consciousness” hypothesis see Forman 1993, and Forman ed. 1990. 6 Leading the contextualists were Steven Katz, who edited a series of influential volumes on the controversy (1978, 1983, and 1992), and Wayne Proudfoot (1985).

34

Sharf

practitioners.7 But non-discriminative discernment did not play a central role in the analysis of mundane cognition until developments in Yogācāra and Pramāṇavāda thought in the fifth and sixth centuries (see the chapters by Keng, Spackman, and Tzohar in this volume; on the dating of Pramāṇavāda see Deleanu 2019). Dignāga was among the first to argue that there was a pre-reflective or non-conceptual aspect to all states of cognition, including workaday discursive states. This aspect, known as “self-awareness” (svasaṃvedana, svasaṃvitti), would come to play a pivotal role in later Indo-Tibetan Yogācāra epistemology, akin to the role that “pre-reflective consciousness” plays in strands of contemporary phenomenology.8 Non-conceptual states (“luminous mind,” “mirror consciousness,” etc.) are sanctioned in later Tantra, Chan, and Dzogchen as well, although such teachings often generated controversy (Sharf 2014, 2016). If I am correct in my analysis, while the seeds of “non-conceptual discernment” may have been present early on, it did not come to the fore in Buddhist thought and practice until the rise of Pramāṇavāda, Buddhist Tantra, Chan, and Dzogchen. These developments, which date from the fifth through the eighth centuries, were likely influenced by non-Buddhist Indian religious traditions. Be that as it may, the early Ābhidharmikas could not help but run into the puzzle of non-conceptual cognition; as we will see, the structure of their theories of perception made the problem unavoidable. 1

Distributed Cognition

The Dukkha-sutta of the Saṃyutta-nikāya contains the following succinct account of the arising of mind:

7 I will not deal with those rare references in the Pali suttas, such as the Aṅguttara-nikāya i. 49–52, to the effect that mind (citta) is originally pure or luminous (pabhassara), but that this purity is obscured by adventitious defilement. (See also the notion of viññāṇam anidassanam or “featureless consciousness” found in the Brahma-nimantanika-sutta, Majjhimanikāya 49, and the Kevaddha-sutta, Dīgha-nikāya 11.) Early exegetes identified the pure citta mentioned in the Aṅguttara-nikāya with the bhavaṅga-citta, which is mind in its latent or “non-arising” state, and this suggests that they didn’t quite know what to do with it. This has become a topic of some debate among contemporary Theravāda scholars; see Collins 1982: 246–247; Gombrich 2006: 43–45; Harvey 1989; Harvey 1995: 166–174; and Thanissaro Bhikkhu on accesstoinsight.org. 8 The literature on svasaṃvedana/svasaṃvitti is large and growing; see, for example, Arnold 2010 and 2012; Coseru 2012; Dreyfus 1996, 1997; Garfield 2006; Hattori 1968; Kellner 2010; Matilal 1986; Moriyama 2010; Sharf 2016; Watson 2006, 2010; Williams 1998; and Yao 2005.

Knowing Blue

35

And what, bhikkhus, is the origin of suffering? In dependence on the eye and forms, eye-cognition arises. The meeting of the three is contact. With contact as condition, feeling [comes to be]; with feeling as condition, craving. This is the origin of suffering. In dependence on the ear and sounds, ear-cognition arises. The meeting of the three is contact. With contact as condition, feeling [comes to be]; with feeling as condition, craving. This is the origin of suffering.9 The passage goes on to repeat the same formula for the nose and odors, tongue and tastes, body and tactile objects, and mind and mental phenomena. (I’ll refer to this oft-repeated formula below as the “arising-of-mind pericope.”) Its appeal lay in the succinct manner in which it touches upon all eighteen elements (dhātu) that collectively comprise the phenomenal world, as well as their causal and temporal interrelations. But the terse formulation also raised a number of problems that would preoccupy Buddhist exegetes for centuries to come. To appreciate the pericope, keep in mind that Buddhist scholiasts sought to produce, using the resources provided in the scriptures, a robust account of mind and perceptual experience that does not invoke an enduring self or subject (ātman). Rather than adducing a single overriding cogito or “witness consciousness,” scriptural accounts distribute cognition among six quasi-­ independent registers; five are associated with the material senses, and one is associated with the mind or mental sense (manas), namely, mind-cognition (manovijñāna). These six registers are structurally alike insofar as each involves the interactions of three discrete elements: the perceptual object (viṣaya), the sense faculty (indriya), and cognition (vijñāna). This yields a total of eighteen elements that collectively account for the entirety of conscious experience; there is no need for an independent observer that sits astride the process. So far so good. But it is not obvious how this model can account for what modern philosophers call the synthetic unity of apperception or cognitive binding. How is it that these six registers interact to create the semblance of a unified and integrated phenomenal domain?

9 Katamo ca bhikkhave dukkhassa samudayo? Cakkhuñca paṭicca rūpe ca uppajjati cakkhuviññāṇaṃ. Tiṇṇaṃ saṅgati phasso. Phassapaccayā vedanā. Vedanāpaccayā taṇhā. Ayaṃ kho bhikkhave dukkhassa samudayo. (Dukkha-sutta, Saṃyutta-nikāya 12.43, ii.72; trans. Bodhi 2000: 580, with changes). See also Saṃyutta-nikāya iv.33, Madhupindika-sutta: mn 18, i.108; Chachakka-sutta mn 148, iii.280; Milindapañha 56; etc. Lin (2015: 104) identifies a total of twelve parallel passages in the Saṃyuktāgama (Za ahan jing 雜 阿 含 經 ); see note below.

36

Sharf

There is, to my knowledge, only a single sutta that acknowledges and addresses this problem directly.10 The Mahāvedalla-sutta (Majjhima-nikāya 43) consists of a series of questions that Mahā Koṭṭhita poses to Sāriputta, along with Sāriputta’s responses. One of the questions involves the binding problem: given that the cognitive reach of each of the sense faculties is restricted to its own register, what integrates them? Sāriputta responds as follows: “Friend, these five faculties each have a separate field, a separate domain, and do not experience each other’s field and domain, that is, the eye faculty, the ear faculty, the nose faculty, the tongue faculty, and the body faculty. Now these five faculties, each having a separate field, a separate domain, not experiencing each other’s field and domain, have mind as their resort, and mind experiences their fields and domains.”11 In short, mind is unique among the six faculties in having access to, and serving as foundation or “resort” (manopaṭisaraṇaṁ, the Chinese parallel reads yi 依, “basis”) for, the other five. The Mahāvedalla may well be the earliest surviving work to posit a structural asymmetry between (1) the five cognitions associated with the material senses (hereafter: the five sense cognitions), and (2) mind-cognition. This asymmetry would prove crucial for the Buddhist analysis of mind and cognition, and the Ābhidharmikas develop it at length. We learn not only that the five sense cognitions have access only to their proper objects (i.e., the eye to visual objects, the ear to sounds, and so on), but also that these objects must, at least according to the Sarvāstivāda Vaibhāṣikas, coexist with cognition itself. That is to say, eye-cognition can only register a visual form if said form exists in the present. Mind-cognition, on the other hand, has access to mental objects (dharmas) as well as to sense cognitions that are transduced from the five material senses. Moreover, mind can perceive past and future objects as well as objects in the present. Finally, and most germane to this essay, Vaibhāṣika 10

11

While the Mahāvedalla is the only sutta of which I am aware that addresses the asymmetry directly, it is recognized obliquely in a distinction between two kinds of “contact” mentioned in the Mahānidāna-sutta of the Dīgha-nikāya (dn 15, ii.62; trans. Walshe 1995: 225); see also the discussion in the Abhidharmakośa (T.1558: 29.52c4–10). mn 43, i.295; trans. Ñāṇamoli and Bodhi 2009: 391. The Chinese parallel reads: 有 五 根 異 行異境界。各各受自境界。眼根耳鼻舌身根。此五根異行異境界。各 各受自境界。誰爲彼盡受境界。誰爲彼依耶。尊者大拘絺羅答曰。五 根異行異境界各各自受境界。眼根耳鼻舌身根。此五根異行異境界。 各 各 受 自 境 界 。 意 爲 彼 盡 受 境 界 。 意 爲 彼 依 (T.26: 1.791b11–17). Note that in the Chinese, it is Mahākauṣṭhila (Juchiluo 拘 絺 羅 ) who is responding to Śāriputra, rather than the other way around. On the Mahāvedalla-sutta see Anālayo 2011: 268–276.

Knowing Blue

37

­ asters hold that the five sense cognitions apprehend their objects directly, m without access to concepts or language. As such, it is sometimes said that sense cognition grasps the “inherent characteristic” (svalakṣaṇa) of the object but not its “generic characteristic” (sāmānyalakṣaṇa). Grasping the generic characteristic involves apprehending the object not merely as a one-off particular but as the token of a generic type, and this requires capacities associated with mind-cognition such as ideation (saṃjñā), recollection (smṛti), and knowledge (prajñā).12 In an oft-repeated example, visual cognition is said to know blue but not “this is blue.” Only mind-cognition can know “this is blue.” Note that the arising-of-mind pericope gives no hint of any structural difference between sense cognition and mind-cognition; the asymmetry seems to have been introduced in later texts such as the Mahāvedalla and early commentarial works so as to render the Buddhist model of distributed cognition intelligible. This is merely one of several lacunae and ambiguities in the pericope with which later commentators had to struggle. Another was how to understand the temporal and causal relationships that pertain between the perceptual object, the sense faculty, and cognition. The pericope reads, “In 12

This is, however, a simplification of the svalakṣaṇa-sāmānyalakṣaṇa distinction. The Vaibhāṣika analysis of the topic is complex, somewhat opaque, and inconsistent across sources. To render their analysis of perception coherent and filial to scripture, for example, the Vaibhāṣikas distinguished two different kinds of svalakṣaṇa: the inherent characteristic belonging to the cognitive object proper, and the inherent characteristic pertaining to the sense sphere. The Mahāvibhāṣā contains the following exchange: “Question: How is it that tactile bodily cognition takes the generic characteristic (sāmānyalakṣaṇa) as its object, given that [according to scripture] the five sense cognitions apprehend [only] the inherent characteristic (svalakṣaṇa)?” “Answer: The inherent characteristic is of two types: (1) the inherent characteristic of the entity (dravya-svalakṣaṇa), and (2) the inherent characteristic of the sense spheres (āyatana-svalakṣaṇa). If we speak of the inherent characteristic from the perspective of the entity, then [in addition to the inherent characteristic] the five sense cognitions also take the generic characteristic [of the entity] as their object. But if we speak of the inherent characteristic from the perspective of the sense spheres, then the five sense cognitions only take the inherent characteristic as the object. Hence there is no contradiction.” 問 云 何身識縁共相境。以五識身縁自相故。答自相有二種。一事自相。二 處自相。若依事自相説者。五識身亦縁共相。若依處自相説。則五識 唯 縁 自 相 。 故 不 相 違 (T.1545: 27.65a12–16). The argument seems to be that any single sense cognition grasps only the inherent characteristic pertaining to its own sense field – the eye sees only colors and forms, the ears sound, etc. – and hence it does not grasp the entity as it exists in itself. But from the point of view of the entity, it is perceived as a whole [by all the senses?], and this entails grasping both inherent and generic characteristics. See also the discussions of svalakṣaṇa and sāmānyalakṣaṇa in Chapter 6.2 of the “Miscellaneous Aggregates” (Zayun 雜 蘊 ) section of the Mahāvibhāṣā (T.1545: 27.200b16 ff.); Abhidharmakośa (T.1558: 29.3a4–11); Cox 1988: 34–35; and Dhammajoti 2007a: 101–103; 2009: 19–22.

38

Sharf

dependence on the eye and forms, eye-cognition arises.” Does this mean that the eye (the visual sense faculty, cakṣurindriya) and form (the material object, rūpa) precede the arising of visual cognition (cakṣurvijñāna), as Sautrāntikas maintain? Or is the dependence logical rather than temporal, such that the three arise simultaneously (the Vaibhāṣika position)?13 While Vaibhāṣikas and Sautrāntikas disagree over whether sense cognition coexists with its object, they concur that the full recognition of the object does not occur until the immediately following moment, when mind-cognition, shaped by ideation, recollection, knowledge, and other mental factors (caitasika), apprehends the throughput of the previous cognition. In technical terms, the preceding moment of cognition is said to function as the mana-indriya or “mind-faculty” of the subsequent arising of mind (manas). Thus in the Vaibhāṣika schema, full recognition of a sense object is a two-step affair: in the first moment, sense cognition arises along with the sense faculty and object, and in the second moment mind-cognition arises. In contrast, the Sautrāntika model involves three steps: sense faculty and object arise in the first moment, sense cognition in the second, and mind-cognition in the third. One issue on which both the two-step and three-step models concur is that mind-cognition does not apprehend the sense object (viṣaya) directly, since the object has already passed when mind-cognition arises. The question then arises: when the sense object is no longer present, what precisely is apprehended? Ābhidharmika exegetes discuss this “after image” under the rubric of the ākāra, a particularly difficult concept that is variously translated “aspect,” “form,” “mental image,” “mode of activity,” and so on.14 The variety of translations reflects the different and somewhat conflicting accounts of ākāra in Ābhidharmika sources. Putting aside such complexities, it would appear that the notion of ākāra arose to help fill the epistemic gaps between what we perceive, what we think we perceive, and what is really there. For our immediate purposes, we need simply note that all Ābhidharmika voices agree that the “object” grasped by mind-cognition – the ākāra, however it is understood – is the outcome of a multi-step process that involves, at some point along the way, conceptual discrimination. They disagree, however, as to precisely when and where conceptuality kicks in. 13

14

There is a large literature on the relationships among the exegetical schools known as the Vaibhāṣikas, Dārṣṭāntikas, Sautrāntikas, Yogācāras, and so on in Ābhidharmika compendia. For recent work on the topic see esp. Dessein 2003; Fukuda 2003; Harada 1996; Honjō 2003; Kritzer 2003, 2005; Park 2007; Yamabe 2003. On Abhidharma theories of ākāra see esp. Dhammajoti 2007b, 2009: 269–275; Kellner 2014; and Kellner and McClintock 2014.

Knowing Blue

2

39

Knowing Blue

We have seen that, for Vaibhāṣikas, the recognition of the object by mind-cognition is a two-step process, while for Sautrāntikas it involves three steps. Both agree, however, that prior to the conceptual recognition of a sense object by mind-cognition, the object is grasped non-conceptually by sensory cognition. This non-conceptual perceptual moment is described in the Vijñānakāyaśāstra (Apidamo shishen zulun 阿毘達磨識身足論), one of the earliest Sarvāstivāda compendia, as follows.15 There are six cognition bodies, namely, eye-cognition, ear, nose, taste, touch, and mind-cognition. Eye-cognition is only able to discern blue; it is unable to discern “this is blue.” Mind-cognition is also able to discern blue. As long as [mind-cognition] is unable to discern [the color’s] name, it is unable to discern “this is blue.” But should it be able to discern its name, it discerns, at the same time, both blue and “this is blue.” The colors yellow, red, white, and so on are [analyzed] in the same way as the color blue. Ear-cognition is only able to discern a sound; it is unable to discern “this is a sound.” Mind-cognition is also able to discern the sound. As long as it is unable to discern its name, it is unable to discern “this is a sound.” But should it be able to discern its name, it discerns, at the same time, the sound and “this is a sound.” 有六識身。謂眼識耳鼻舌身意識。 眼識唯能了別青色。不能了別此是青色。意識亦能了別青色。乃至未 能了別其名。不能了別此是青色。若能了別其名。爾時亦能了別青 色。亦能了別此是青色。如青色黄赤白等色亦爾。耳識唯能了別聲。 不能了別此是聲。意識亦能了別聲。乃至未能了別其名。不能了別此 是聲。若能了別其名。爾時亦能了別聲。亦能了別此是聲.16

The Vijñānakāya passage goes on to repeat the same formula for nose and odor, tongue and taste, body and touch, and concludes by saying that “mindcognition is also able to discern all dharmas” 意識亦能了別諸法. The notion that sense perception is to mental perception what knowing blue is to knowing “this is blue” would prove beguiling. Vasubandhu draws on it, for example, in his discussion of “contact” (sparśa) in his Abhidharmakośa (Apidamo jushe lun 阿毘達磨倶舍論, T.1558: 29.52c4–10). Saṃghabhadra makes similar use of the distinction in his discussion of mind-cognition in the *Nyāyānusāra 15 16

On the Vijñānakāya-śāstra, which is attributed to *Devaśarman or *Devakṣema (Tipo­ shemo 提 婆 設 摩 ), see Frauwallner 1995: 28–31. T.1539: 26.559b27–c6; cf. the discussion of the six cognitions earlier in the same text, T.1539: 26.582c20–583a13.

40

Sharf

(Apidamo shun zhengli lun 阿毘達磨順正理論, T.1562: 29.342a17–22). These texts all concur that the difference between non-­conceptual sensory cognition and conceptual mind-cognition is captured in the claim that the eye knows blue but not “this is blue.” This bare sensation of blue appears, at least on the surface, to be what analytic philosophers call a “quale” or “raw feel.” But just as with qualia, the claim raises a host of thorny questions. What, if anything, might be said about the nature or content of this non-conceptual or pre-­conceptual blue quale? As discussed above, five sense cognitions are said to apprehend the “specific characteristic” (svalakṣaṇa) of the object – a transient particular – while mind-cognition apprehends the “generic characteristic” (sāmānyalakṣaṇa) which involves the application of concepts and categories. Such a distinction is necessitated, in part, by the Vaibhāṣika doctrine that a sense faculty can only apprehend an object that is immediately present, and thus no two moments of sense cognition apprehend the same object-field (viṣaya).17 (While the objectfield grasped by sense cognition is continually changing, the perceptual object [ālambana] of mind-cognition, which is apprehended via a conceptual category, can persist from one moment to the next.) But Vaibhāṣikas are aware that, although the immediate sense perception of a svalakṣaṇa does not involve the use of concepts and language, it still entails some minimal discriminative capacity – after all, the claim, it would appear, is that eye-cognition can differentiate blue from yellow or red or white, and that this occurs in the perceptual stream prior to the application of concepts such as “blue” and “yellow” and “red” and “white.” The Vaibhāṣikas then devise a specific discriminative mechanism, namely “inherent discrimination” (svabhāvavikalpa), to account for this capacity, and they explain it with reference to two mental factors, vitarka (jue 覺, xun 尋) and vicāra (guan 觀, si 伺), which might be rendered “coarse discernment” and “fine discernment.” (For reasons that will become clear below, I will leave these two terms untranslated in this essay.) These complex notions turn out to be key to understanding what it means to know blue but not “this is blue.” 3

Vitarka, Vicāra, and the Three Kinds of Discrimination

The vitarka-vicāra pair appear frequently in early sūtras, and while their precise meaning is neither clear nor consistent, both terms initially seem to have 17

On early Ābhidharmika accounts of the succession of moments of cognition, see Takatsukasa 2014 and 2016; and Yamabe 1990.

Knowing Blue

41

been associated with thought, rumination, or the inner discursive activity that precedes speech.18 In time vitarka came to be understood as the initial and imprecise detection/discernment/recognition of a perceptual object, while vicāra is understood as the subsequent mental inspection of said object. Accordingly, many commentators, both medieval and modern, interpret the two terms as two kinds or degrees of conceptual discrimination or discursive thought. In his detailed analysis of these terms, Lance Cousins (1992: 153) summarizes the early Pali Abhidhamma understanding as follows: “vitakka is ‘thinking of’ something, whereas vicāra is ‘thinking about’ that same thing.” This formulation is intriguingly reminiscent of the distinction between knowing blue and knowing “this is blue.” Note, however, that the early Ābhidharmikas did not initially deploy vitarka and vicāra in their analyses of sense perception per se; rather, their exegetical interests lay in distinguishing different dhyānic states. That vitarka is “coarse” and vicāra “subtle” helped to differentiate the coarser quality of mind in the first stage of dhyāna from the more refined mind in the second stage.19 Numerous passages, all of which concern the nature of dhyāna, agree that vitarka and vicāra are present in sense cognition as well as in mundane mindcognition.20 As such, they would not seem immediately relevant to the asymmetry between sense and mind-cognition. However, as Sarvāstivāda writers pondered the conceptually unconstructed character of sense cognition, vitarka and vicāra were brought into play. Exegetes, drawing on earlier materials that associated the pair with the initial moments of perceptual recognition, would expand on the notion that vitarka denotes the “coarse” functioning of mind (cittaudārikatā, cuxin 麁心), while vicāra is mind operating in a more fine-grained or “subtle” manner (cittasūkṣmatā, xixin 細心).21 On the basis of this distinction, and in some tension with statements in the *Abhidharmahṛdaya, Mahāvibhāṣā, and the Abhidharmakośa that both vitarka and vicāra are 18

19 20 21

On the early meanings of the pair, and their appearances in early commentaries, see esp. Cousins 1992; Jaini 1977: 84–85; and the more extended discussion in Sharf 2018. The Pali Vitakkasaṇṭhāna-sutta (Majjhima-nikāya 20, i.118) contains five different techniques for removing distracting “thought” (vitakka). The association of vitarka and vicāra with the dhyānas was likely influenced by non-­ Buddhist traditions, notably the Yoga-sūtra; see Cousins 1992: 148–151; and Bronkhorst 1986: 29–64. See, for example, the *Abhidharmahṛdaya (or *Abhidharmasāra, Apitan xinlun 阿 毘 曇 心 論 ), T.1550: 28.810a11–17; Mahāvibhāṣā (Apidamo dapiposha lun 阿 毘 達 磨 大 毘 婆 沙 論 ), T.1545: 27.377b6; and Abhidharmakośa, T.1558: 29.8a11–18. For the articulation of this distinction in later Vaibhāṣika texts, see the *Nyāyānusāra, T.1562: 29.394a2–c3; Abhidharmāvatāra, T.1554: 28.982a24–27; and the discussion in Dhammajoti 2008: 144–146 n.107.

42

Sharf

­ resent in sense cognition, Vaibhāṣika commentators will claim that vitarka p alone is active (as opposed to simply present) in sensory perception.22 Vitarka is that rudimentary or “coarse” mode of recognition that knows blue, while vicāra is a more refined discriminatory capacity involved in knowing “this is blue.” But this immediately creates a problem, as the orthodox and widely accepted Ābhidharmika position was that the operations of the five sense cognitions were avikalpaka – non-discriminating. Yet Vaibhāṣikas could not deny that vitarka involved some form of discrimination, even if relatively rudimentary or primitive. Indeed, as we have seen, vitarka was commonly associated with thinking or reflection (saṃkalpa, siwei 思惟, etc.). This apparent contradiction was not lost on the Ābhidharmikas, and the problem is raised directly in the Abhidharmakośa. If the five sensory cognitions are always accompanied by vitarka and vicāra, why are they explained as being free from discrimination? (1.33ab) They are explained as free from discrimination as they are [free from] examination and recollection…. According to the Vaibhāṣikas, there are, in brief, three kinds of discrimination (vikalpa): (1) inherent discrimination (svabhāvavikalpa); (2) discrimination through examination (abhinirūpaṇavikalpa); and (3) discrimination through recollection (anusmaraṇavikalpa). The five sensory cognitions are explained as free from discrimination insofar as they have inherent [discrimination] but not the other two kinds. It is like a onefooted horse that is called a horse without feet. Inherent discrimination alone is vitarka. 若五識身有尋有伺。如何得説無分別耶。頌曰。説五 無分別由計度隨念 … 論曰。傳説。分別略有三種。一自性分別。二計 度分別。三隨念分別。由五識身雖有自性而無餘二。説無分別。如一 足馬名爲無足。自性分別體唯是尋.23

This is a revealing passage; pace the widespread understanding that sense cognition is “without discrimination,” the Vaibhāṣikas concede that it does in fact have some capacity to discriminate, if only in a rather crude fashion. The 22

23

The Sautrāntikas have a rather different analysis, as they do not reify vitarka and vicāra into independently existing dharmas to begin with, and while they agree that the difference is between coarse and fine discriminative activities of mind, they see these in relative, rather than absolute terms, and hence they believe that they cannot coexist. See the extended refutation of this position in the Mahāvibhāṣā, T.1545: 27.219a. T.1558: 29.8a27–b5; cf. *Nyāyānusāra T.1562: 29.350b9–10.

Knowing Blue

43

Vaibhāṣikas come up with an amusing apologia for this inconsistency, namely, that sense cognition is said to lack discrimination in the same way that a onefooted horse is said to lack feet! It would appear that the Vaibhāṣikas simply couldn’t make sense of “direct” sensory perception without bringing in some sort of discriminative capacity. After all, knowing blue must be to know, at the very least, that it is something other than yellow or red or white. The Vaibhāṣika doctrine of “three kinds of discrimination” (*trivikalpa, san fenbie 三分別) allows us to distinguish this rudimentary cognitive capacity from more complex and decidedly conceptual forms of discrimination. In what I will propose as a charitable reading of this doctrine, svabhāvavikalpa – a term intended to denote “essential” discrimination – is simply to register difference.24 Arguably, such a capacity is fundamental to, if not definitive of, sentience itself; all life forms can distinguish, in some manner, food from non-food, mates from non-mates, and so on. Given its ubiquity, this capacity must not be dependent on concepts or language or discursive thought or even, perhaps, conscious awareness.25 This is why I have been careful to translate vikalpa, when used by the Ābhidharmikas, as “discrimination” rather than “conception” or “imagining.” I have been similarly careful to translate vijñāna as “cognition,” rather than “consciousness.” Whether the cognitive capacity of direct sense cognition – the ability of eye-cognition to discriminate or register blue – entails what we could call phenomenal consciousness or whether it might be subliminal or subdoxastic is a fraught topic that we will visit briefly below. To return then to vitarka and vicāra, we saw that the *Abhidharmahṛdaya, Mahāvibhāṣā, and Abhidharmakośa all contain statements to the effect that both vitarka and vicāra exist in sense cognition. But at the end of the Abhidharmakośa passage on the subject, or at least at the end of Xuanzang’s 玄 奘 (602–664) translation, it says that svabhāvavikalpa – the minimal discriminative capacity associated with sense cognition –“is vitarka alone” (自性分別 24

25

Some understand svabhāvavikalpa to mean the “discrimination of inherent nature” – i.e., the immediate (unconstructed) apperception of the “essence” (svabhāva) of a given real. But parsing svabhāvavikalpa in this manner ignores its syntactic parallelism with abhinirūpaṇavikalpa and anusmaraṇavikalpa, and similarly ignores the fact that the early Ābhidharmikas felt obliged to enlist the notion of vitarka (and sometimes vicāra, see below) to make sense of it. In my reading, the need to engage vitarka is an acknowledgement that even in svabhāvavikalpa there is some form of “differentiation” or “construction” (vikalpa) present. For an alternative approach, see Keng 2019, who approaches the Ābhidharmika notion of svabhāvavikalpa as anticipating Dignāga’s notion of mānasapratyakṣa, i.e., it concerns the strictly non-conceptual perception of the shape or contour of an object. That is to say, my association of “sentience” with some minimal capacity to discriminate is stipulative, and as such does not in itself entail phenomenal consciousness.

44

Sharf

體唯是尋). Xuanzang’s translation, completed in 654, matches the extant San-

skrit text: tatra svabhāvavikalpo vitarkaḥ. But curiously, our earliest textual witness to the Abhidharmakośa, namely, Paramārtha’s (Zhendi 眞諦, 499–569) translation completed in 567, reads: “among the [three vikalpa], the svabhāvavikalpa is precisely vitarka-vicāra” 此中自性分別即是覺觀 (T.1559: 29.168b3–4). One might dismiss this inconsistency as an anomaly. Yet we find a similar inconsistency in the Mahāvibhāṣā discussion of the three kinds of discrimination. The Mahāvibhāṣā passage may be the earliest reference to the *trivikalpa doctrine in any extant Sarvāstivāda text, not to mention the source for Vasubandhu’s discussion in the Abhidharmakośa. In Xuanzang’s translation the passage reads as follows: There are, in brief, three kinds of discrimination (vikalpa): (1) inherent discrimination, which is to say vitarka and vicāra; (2) discrimination through recollection, which is to say the recollection [that arises] associated with mind-cognition; and (3) discrimination through examination, which is the unsettled (vyagra) knowledge of the manas. 略有三種分 別。一自性分別。謂尋伺。二隨念分別。謂意識相應念。三推度分 別。謂意地不定慧. (T.1545: 27.219b7–9)

Xuanzang’s translation of the Mahāvibhāṣā thus associates inherent discrimination with both vitarka and vicāra. But this is not the case in the earlier translation of the same passage by Buddhavarman (Futuobamo 浮陀跋摩, d.u.) and Daotai 道泰 (d.u.), completed in 437: Discrimination is of three kinds: there is inherent discrimination, discrimination through recall, and discrimination through contemplation. As for inherent discrimination, it is vitarka. Discrimination through recall is recollection. Discrimination through contemplation is knowledge. 分 別有三種。有自體分別。有憶念分別。有現觀分別。自體分別者。謂 覺是也。憶念分別者。謂念是也。現觀分別者。謂慧是也. (T.1546: 28.

169b5–7)

So in the early translation of the Mahāvibhāṣā, which is also our earliest witness to the *trivikalpa doctrine, the inherent discrimination characteristic of sense cognition is said to be vitarka alone, but Xuanzang’s later translation includes both vitarka and vicāra. This is the precise inverse of what we found in the two translations of Abhidharmakośa, in which Paramārtha’s earlier translation associates inherent discrimination with both vitarka and vicāra, while

Knowing Blue

45

­ uanzang’s later text has only vitarka. Moreover, we have already seen that X other sections of these same works, as well as the earlier *Abhidharmahṛdaya, explicitly state that vitarka and vicāra are both present in sense cognition. What is going on? These texts may provide us with a murky but nonetheless telling glimpse into the struggle that Sarvāstivāda exegetes faced in explaining the discriminative capacities of non-conceptual sense perception. They recruited vitarka and vicāra for this purpose – following scriptural precedent, they used vitarka to reference the initial non-conceptual and somewhat crude impression of the object by the senses (the initial striking of a bell, for example), and vicāra for the subsequent mental investigation of said object (the ensuing reverberations), but they were stymied by orthodox teachings that associated both vitarka and vicāra with sense cognition. This paralleled the problem posed by mental factors such as recollection (smṛti) and knowledge (prajñā). According to Sarvāstivāda orthodoxy, recollection and knowledge are both classified as “omnipresent mental factors” (mahābhūmika), and as such must be present in all cognition, including that of the five material senses. How does one square the fact that vicāra, smṛti, and prajñā are all stated to be present in non-conceptual sense cognition, yet all three seem to entail conceptuality? This is where the *trivikalpa doctrine came to the rescue. To see how this works, let’s turn to the Mahāvibhāṣā discussion of the meaning of vitarka and vicāra, which is the reference point for all later Vaibhāṣika discussions of the issue. The passage begins with a definition of vitarka. What is vitarka? Answer: when any of the [six kinds] of cognition seek out, distinguish, reveal, determine, represent, and discriminate [the object’s] nature and category, this is vitarka. While the terms, such as “seeking out” etc., for this quality of cognition may differ, there is no difference in essence, since all these terms express the essential nature of vitarka. 云 何尋。答諸心尋求辨了顯示推度搆畫分別性分別類是謂尋。諸心尋求 等名雖有異而體無差別。皆爲顯了尋自性故. (T.1545: 27.219a2–4)

This passage goes on to critique those (Dārṣṭāntika?) exegetes who hold that vitarka and vicāra are mere nominal entities, and who further maintain that, since the meanings of vitarka and vicāra are relational and mutually contradictory, it makes little sense to say that both could coexist in the same instant. The position of the Mahāvibhāṣā is that vitarka and vicāra are in fact real entities (dharmas) that can coexist in time, but at any particular moment one or the other will come to the fore. To make its point, the text uses various illustrations, such as putting vinegar and water or salt and grits in one’s mouth at the same

46

Sharf

time – some of the flavors (vinegar, salt) are sharp and others (water, grits) dull; both are present, but at any one time one will dominate over the other.26 After clarifying the relationship between vitarka and vicāra, the text goes on, as we have seen, to introduce the three kinds of vikalpa: “There are three kinds of discrimination: (1) inherent discrimination, which is to say vitarkavicāra; (2) discrimination through recollection, which is the recollection associated with mind-cognition; and (3) discrimination through examination, which is the knowing associated with mind-cognition when not absorbed in meditation.” Notice that the second kind of discrimination is associated with recollection (smṛti), and the third with knowledge (prajñā). The passage continues, In the realm of desire, the five cognitions have only [the first] kind, namely, inherent discrimination. Although they also have recollection, it is not the discrimination of recollection because they are incapable of recollection. Although they also have knowledge, it is not the discrimination of examination, for they are incapable of examination. 欲界五識身唯有一 種自性分別。雖亦有念而非隨念分別。不能憶念故。雖亦有慧而 非推 度分別。不能推度故. (T.1545: 27.219b9–11)

The *trivikalpa doctrine thus allows the Mahāvibhāṣā to distinguish recollection (smṛti) per se, from the discriminatory capacity of recollection (anusmaraṇavikalpa); and to distinguish knowledge (prajñā) per se, from the discriminatory capacity of knowledge (abhinirūpaṇavikalpa). The Vaibhāṣika can then claim that recollection and knowledge, which are counted among the omnipresent (mahābhūmika) mental factors, are indeed present in sense cognition, but not so their discriminatory capacities. The forms of discrimination associated with recollection and knowledge, which entail reflection and conceptual thought, are only available to mind-cognition. And while our sources, as we have seen, are not always consistent, they now can handle vicāra similarly. In other words, they can concede that while vicāra may be present along with vitarka in sense cognition (as scripturally mandated), only vitarka is operative. 26

Somewhat confusingly, the Mahāvibhāṣā also cites the example, which it attributes to the Prajñapti-śāstra (Shishe lun 施 設 論 ), of striking a bell or metal vessel, in which the initial sound is “coarse” and the later reverberations are “subtle.” It also mentions the ­example of the bird beginning to flap its wings to fly, where the initial movements are  coarse  and the later movements subtle. (The bird example is attributed to the Dharmaskandhapāda-śāstra, Fayun lun 法 蘊 論 .) It is not clear how such examples strengthen the case for coexistence. On the debate over the coexistence of vitarka and vicāra, see Dhammajoti 2008: 9–10, 144–146 n.107; and Jaini 1977: 83–88.

Knowing Blue

47

To return to the inconsistencies we found in the translations of the Mahāvibhāṣā and the Abhidharmakośa: both texts, following Sarvāstivāda orthodoxy, confirm that vitarka and vicāra are present in sense cognition, as are smṛti and prajñā, but they then aver to the *trivikalpa doctrine to show how such conceptualizing capacities are not operative. “Inherent discrimination” is associated with vitarka but not smṛti and prajñā. But the question remained: what about vicāra? Recall that authoritative texts invariably associate vicāra with vitarka when it comes to sense cognition. Does this mean that vicāra is active in inherent discrimination or not? That different Chinese translations provide conflicting responses to this question may reflect the fact that it took time to sort out the details of the new *trivikalpa theory. In other words, rather than being simple errors on the part of the Chinese translators or later copyists, the variant renderings may reflect real inconsistencies among early recensions of the source texts. In any case, the *trivikalpa doctrine might seem like a gerrymandered solution to a non-problem. That is to say, the question as to how conceptual dharmas can be present in a non-conceptual state might appear artificial insofar as it is occasioned by an inordinately rigid system of classifying dharmas. Had the Vaibhāṣikas simply built in more flexibility, as did the Sautrāntikas, the problem would never have arisen. Why not just reclassify vicāra, smṛti, prajñā, saṃjñā and other eminently conceptual caitasika such that they don’t end up associated with non-conceptual sense cognition?27 But this response fails to appreciate the philosophical conundrum with which they were wrestling. The Ābhidharmikas recognized that cognition of any kind must involve some minimal discriminatory capacity. That is to say, to discern or detect blue, even at its most primitive, is to distinguish it from yellow and red, and it is difficult to make sense of this unless the cognitive system has some “retention” (i.e., smṛti) and “cognizance” (i.e., prajñā) of yellow and red – the residuum of past experience – even when the percepts of yellow and red are not present. (Were we to live in a world where everything was blue, there would be no “blue.”) In short, the Ābhidharmikas had good reason to assume that any cognitive event must entail a modicum of discrimination, recollection, and knowledge. At the same time, the Ābhidharmikas were committed to a number of not-unreasonable metaphysical postulates, including: (1) the constituent elements of reality are in continual flux; (2) conceptual recognition of a percept 27

This is precisely the Sautrāntika position taken by Harivarman in his *Tattvasiddhi. Harivarman insists that mental factors such as vitarka, vicāra, and saṃjñā are not present when a sense cognition first apprehends an object, and he quotes scripture in support of this position. Thus for Harivarman, sense cognition is entirely free of vikalpa; see Lin 2015: 176–187.

48

Sharf

involves applying a generic name to a transient particular that emerges from the flux; and (3) there must be a temporal gap between the moment when the transient particular is first apprehended and the subsequent moment when it is identified with a generic name. In short, in this “empiricist” model, a sense datum must first be cognitively available as a raw feel for it to serve as the object of mental reflection. The Ābhidharmikas were thus forced to confront a problem familiar to contemporary philosophy of mind, namely, how to characterize the nature and status of this raw feel. So the Ābhidharmikas punted: they proffered the theory of three kinds of discrimination (*trivikalpa) that would explain, at one and the same time, (1) that there is indeed a modicum of coarse discriminatory activity in what was previously supposed to be a non-discriminative (avikalpaka) state, and (2) that certain mental factors (smṛti, prajñā, saṃjñā), normally deemed conceptual, are present but not yet (fully?) active, in non-conceptual sense cognition. The other two kinds of discrimination – those associated with active knowledge and recollection – are the preserve of mind-cognition, which has both vitarka and vicāra. The Abhidharmakośa explains, The essence of the diffuse knowledge belonging to mental [reflection] is recollection. “Diffuse” means not concentrated. The diffuse knowledge associated with mind-cognition is called discrimination through examination. Whether [mind] is concentrated or not, the recollection associated with mind-cognition is called discrimination through recollection. 意地散慧諸念爲體。散謂非定。意識相應散慧。名爲計度分別。若定 若散意識相應諸念。名爲隨念分別. (T.1558: 29.8b6–8)

Later texts will reaffirm the claim that smṛti, prajñā, and even saṃjñā are present in sense cognition but that they are inactive or weak. In the *Nyāyānusāra discussion of the three kinds of discrimination, for example, after Saṃghabhadra notes that “the essence of inherent discrimination is vitarka alone” 自性分別體唯是尋 (T.1562: 29.350b11), he goes on to clarify, “although the five sense faculties are associated with knowledge and memory, their capacity to discern and register [therein] is minuscule, and thus [knowledge and memory] are only grasped by mind” 五識雖與慧念相應。擇記用微。故唯取 意 (T.1562: 29.350b17–18). Accordingly, the mature Vaibhāṣika position is that the capacity of eye-cognition to register blue but not know “this is blue” does involve a minimal kind of discrimination (svabhāvavikalpa), and that this is associated with a rudimentary or coarse form of recognition (vitarka) as well as a weak or minuscule form of memory and recognition. In my “charitable” interpretation of the

Knowing Blue

49

Vaibhāṣika accounts of direct sense perception, I have been careful to disaggregate these rudimentary forms of discrimination from conceptuality per se. This is mandated by the fact that we normally associate conceptuality with language, and, for the Vaibhāṣika, the use of nominal designations is simply not available to sense cognition. But while I believe this charitable account may well capture the intent of our Vaibhāṣika authors, they have trouble making it stick. One problem is the claim that the omnipresent factors are present but operate only weakly. This is difficult to swallow when it comes to factors like knowing (prajñā) and ideation (saṃjñā), since it is unclear what functionality they retain, however minimal, once they are stripped of access to names and concepts. Take ideation for example: Ābhidharmikas agree that the nature of ideation is to recognize the inherent characteristic (nimitta, xiang 相) of the object.28 The Abhidharmakośa, for example, defines ideation as follows: The essence of the aggregate of ideation is that it apprehends appearances, which is to say it grasps characteristics such as blue, yellow, long, short, man, woman, enemy, friend, pain, pleasure, and so on. This is further divided into six kinds of ideation [associated with each of the cognitions] as was explained above in reference to feeling (vedanā). 想蘊謂能 取像為體。即能執取青黃長短男女怨親苦樂等相。此復分別成六想 身。應如受說. (T.1558: 29.4a4–6)

Commentators explain that the capacity to grasp the characteristic is tied to the application of names – associating a transient particular with a generic category or class. The *Abhidharmāvatāra (Ru apidamo lun 入阿毘達磨論) expands on the definition above. The meaning of the word “ideation” is that which comprehends by provisionally linking a characteristic to a name and meaning. With regard to visible forms such as blue, yellow, long, short, etc., or sounds such as those from a conch or a drum, etc., or scents like that of agarwood or musk, etc., or tastes like saltiness or bitterness, etc., or tactile sensations like hardness or softness, etc., or dharmas like male and female, etc., it comprehends them by provisionally linking their characteristic, name, and meaning. It is the causal basis for vitarka and vicāra, and thus it is called “ideation.” 想句義者。謂能假合相名義解。即於青黄長 短等

28 On saṃjñā see esp. Kramer 2012.

50

Sharf 色。螺鼓等聲。沈麝等香。醎苦等味。堅軟等觸。男女等法相名義中 假合而解。爲尋伺因故名爲想.29

Moreover, in the Abhidharmakośa, ideation is associated with “views” (dṛṣṭi), and we learn that ideation can be in error: “It is the force of mistaken ideation that leads to attachment to various views” 倒想力故貪著諸見 (T.1558: 29.5b14). Given such definitions, it is difficult to imagine what a minimal “non-­ conceptual” form of ideation might look like. Similarly, post-Mahāvibhāṣā compilations, including the Abhidharmakośa, *Samyuktābhidharmahṛdaya (Za apitan xin lun 雜阿毘曇心論), *Nyāyānusāra, and Abhidharmāvatāra, associate vitarka with the primitive and non-­ conceptualizing discriminative capacity of sense cognition, and vicāra with the more conceptual and discursive discriminative operations of mind-­ cognition.30 Yet they stumble over themselves in their attempts to define non-­ conceptualizing yet discriminative vitarka. Take, as one final example, the definition of vitarka in the Abhidharmāvatāra: Vitarka has the characteristic of causing mind to be coarse with regard to an object. It is also named “discriminative reflection.” Struck by the wind of ideation, it operates in a coarse manner. It is this dharma that serves as the cause of the transformations of the five cognitions. 尋謂於境令心麁 爲相。亦名分別思惟。想風所繋麁動而轉。此法即是五識轉因.31 Try as they might, the Ābhidharmikas, couldn’t seem to give a robust account of immediate non-conceptual sense perception – of what it means to just know blue – without bringing conceptuality – the “wind of ideation” – into the picture. 4

Final Thoughts

The question as to what it is that “sees” or “knows” was the subject of a protracted debate within the Sarvāstivāda world. The Mahāvibhāṣā records a range of early opinions on the issue, including: (1) it is the eye faculty in conjunction with eye-cognition that sees; (2) eye-cognition alone sees; ­ 29 30 31

T.1554: 28.981c20–23; cf. Dhammajoti 2008: 80. The Abhidharmāvatāra, attributed to Skandhila 塞 建 陀 羅 , was translated by Xuanzang in 658. Abhidharmakośa, T.1558: 29.8a12–27; *Samyuktābhidharmahṛdaya, T.1552: 28.880b2 (自 性 思 惟 者 謂 覺 也 ); *Nyāyānusāra: T.1562: 29.349a23–24. T.1554: 28.982a24–27; trans. Dhammajoti 2008: 83, with changes.

Knowing Blue

51

(3) ­knowledge (prajñā) in conjunction with eye-cognition sees; (4) the collocation (hehe 和合, sāmagrī) of object, faculty, and cognition sees; and (5) only one eye at a time sees.32 Vasubandhu picks up the debate in the Abhidharmakośa, in which he apparently prefers the Sautrāntika position. In short, the Sautrāntikas claim that the entire debate is misguided since there is nothing that sees in the first place. The Sautrāntika masters have this to say: You are all just chewing away at empty space! [According to the sūtras,] “eye-­cognition arises in dependence on the eye, forms, etc.” Given such [statements], with respect to seeing, where is a subject and where an object? It is only the cause and effect of [individual] dharmas, which don’t actually do anything (nirvyāpāra). In accord with worldly attitudes, we provisionally speak of the eye as that which “sees,” and cognition as that which “knows.” But the wise are not attached to such talk. As the World Honored One taught, one should not obstinately cling to colloquial expressions or stubbornly chase after conventional ideas. 經部諸師有作是説。如何共聚楂掣虚空。眼 色等縁生於眼識。此等於見孰爲能所。唯法因果實無作用。爲順世情 假興言説。眼名能見。識名能了。智者於中不應封著。如世尊説。方 域言詞不應堅執。世俗名想不應固求.33

According to this view, since there is nothing that sees or grasps an object, there is also nothing that is seen or grasped. Thus the whole notion of direct sense perception is wrongheaded from the get-go, as it conjures a cogito or subject of experience that apprehends an object. For the Sautrāntikas, such a cogito – a something that sees or knows – can only be the ersatz self (ātman) that the Buddhists disavow as a fiction.

32

T.1545: 27.61c7 ff. Later Sarvāstivāda compendia will return repeatedly to this debate, and extended discussions are found in the Pañcavastuka-vibhāṣa (Wushi piposha lun 五 事 毘 婆 沙 論 , T.1555: 28.991b20 ff.), *Samyuktābhidharmahṛdaya (T.1552: 28.876b12 ff.), Abhidharmakośa (T.1558: 29.10c21 ff.), and *Nyāyānusāra (T.1562: 29.364a23 ff.), among other works. On this debate see Cox 1988: 73 n.19; Dessein 2003: 305–308; Dhammajoti 2007a: 51–91; Harada 1997; Jaini 1977: 74–83; Kellner 2014: 283; Kritzer 2003: 333–335, and 2005: 32–33. On Ghoṣaka see Sastri 1953: iii–iv. 33 T.1558: 29.11b1–9. Scholars have traced this Sautrāntika argument back to the Yogācārabhūmi-śāstra. See, for example, T.1579: 30.610a19–27, where, in response to the question as to what sees, a distinction is made between ultimate and worldly ways of understanding. The ultimate understanding is that nothing sees, “since all dharmas, by their natures, arise from conditions and vanish in an instant, without doing anything 諸 法 自 性 衆 縁 生 故 。 刹 那 滅 故 。 無 作 用 故 .” See also Kritzer 2005: 32–33; Harada 1997; and Miyashita 1986.

52

Sharf

Vasubandhu returns to the same argument in Chapter 9 of the text, this time in the context of his refutation of the Vātsīputrīya notion of the person (pudgala). The scriptures state that cognition knows the object. What is cognition doing with respect to this object? It does nothing at all, but merely arises with the object. Just as the effect corresponds to its cause, although it doesn’t do anything, it nevertheless arises resembling the cause, and thus we say that it corresponds to the cause. In the same way, cognition arises without doing anything, yet it resembles the object and thus we speak of it as “knowing” the object…. Or we can compare the knowing of cognition to the movements of a flame. Why do we say that a flame is capable of movement? The continuous flickering of the blaze is provisionally called a “flame” [as if it were a single thing]. Since this “flame” continually arises from moment to moment in different places, we say the flame is “moving.” But there is no separate movement. Similarly, the successive [instances] of mind are ­provisionally called cognition, and when it arises in conjunction with different objects we speak of it as “knowing.” 經説諸識能了所縁。識於所縁 爲何所作。都無所作但以境生。如果酬因。雖無所作而似因起説名酬 因。如是識生雖無所作而似境故説名了境 。。。。 或如燈能行識能了 亦爾。爲依何理説燈能行。焰相續中假立燈號。燈於異處相續生時。 説爲燈行。無別行者。如是心相續假立識名。於異境生時説名能了.34

The Sautrāntika position, at least as articulated by Vasubandhu, is that there is nothing at all that knows the object – not the eye, not cognition, not the discernment of prajñā. Vasubandhu even rejects the Dārṣṭāntika “emergence” theory that it is the collocation of multiple interdependent entities that sees. Saṃghabhadra also weighs into this debate, and he seems to have formulated his innovative doctrine of the “three kinds of perception” (xiangliang 現 量, pratyakṣa) in order to rebut the Sautrāntika position.35 In his ­analysis of the 34 35

T.1558: 29.157b20–c2; cf. Cox 1988: 39. The three kinds of perception are: (1) perception based on the sense faculties (yigen xianliang 依 根 現 量 , *indriyāśrita-pratyakṣa); (2) perception of [inner] feelings (lingna xianliang 領 納 現 量 , *anubhāva-pratyakṣa); and (3) perception of awareness or comprehension (jueliao xianliang 覺 了 現 量 , juehui xianliang 覺 慧 現 量 , *buddhi-pratyakṣa). The topic comes up twice in Saṃghabhadra’s *Nyāyānusāra: T.1562: 29.374c13 ff., where the third category of perception is called jueliao xianliang 覺 了 現 量 , and T.1562: 29.736a9 ff., where the third category is called juehui xianliang 覺 慧 現 量 . On this theory see Cox 1988: 75 n.27; Dhammajoti 2007a: 137–139; 2009: 276–277; Yao 2005: 86–89; and Sharf 2018.

Knowing Blue

53

Vasubandhu-Saṃghabhadra debate over what, if anything, sees, ­Dhammajoti, a leading scholar of Sarvāstivāda Abhidharma, concludes that the Sautrāntikas must have held to something akin to the “self-awareness” or “self-illumination” position later associated with svasaṃvedana.36 Without it, he believes, Sautrāntikas would be unable to explain the “sense of vividness” associated with sense perception prior to the arising of mind-cognition (2009: 255). If I understand him correctly, Dhammajoti is saying that for Sautrāntikas the “subject” and “object” of cognition do not coexist – there is no intentionality, no relationship between grasper and grasped – and thus cognition can only be known if it is self-illuminating or self-intimating. But this may simply reflect Saṃghabhadra’s mischaracterization of the Sautrāntika reductionist (if not eliminativist) position. The sense that conscious awareness is a something that stands in need of explanation – that we need to explain it via a theory of self-illumination or other-illumination or what have you – is an illusion or unwarranted reification precipitated by the tremendous speed at which the causally interconnected cognitive processes that undergird the illusion transpire. Yao Zhihua takes up some of the same arguments in his study of Buddhist theories of self-cognition. Like Dhammajoti, Yao traces the svasaṃvedana idea back to earlier Mahāsāṃghika and Sautrāntika materials. But curiously, Yao believes that, in positing his doctrine of three kinds of perception, even Saṃghabhadra – arch-critic of the Sautrāntikas – comes to embrace something along the lines of self-awareness. Among other things, Yao notes that Saṃghabhadra’s category of “perceptual awareness” (jueliao xianliang 覺了現 量, juehui xianliang 覺慧現量, *buddhi-pratyakṣa) is essentially reflexive in nature.37 But it is not clear to me that Saṃghabhadra is interested in the nature of awareness per se any more than is Vasubandhu. His concern, rather, is to render intelligible the orthodox Vaibhāṣika understanding of cognition, in which (1) the full sensorial awareness associated with mind-cognition necessarily involves memory and conceptual construction, and (2) this awareness is nevertheless predicated on a prior moment in which there is a direct, unmediated perceptual encounter with an “external” object. Saṃghabhadra wants to prove that, prior to knowing “this is blue,” there must be a moment in which the eye knows blue directly. 36 37

Dhammajoti 2007a: 110, 141, 159; 2009: 252–255. Yao also notes that, in his exposition of the three perceptions, Saṃghabhadra anticipates Dignāga’s memory argument and invokes the gap between first- and third-person experience, all of which play a central role in later arguments for svasaṃvedana (Yao 2005: 87–88).

54

Sharf

Gareth Evans (1982) may have been the first to introduce the notion of “nonconceptual content” or “non-conceptual information” into contemporary analytic philosophy.38 But in working through the issues, Evans concludes that the content of perceptual informational states need not be “perceptual experiences” or “states of a conscious subject” (1982: 157). He explains: [We] arrive at conscious perceptual experience when sensory input is not only connected to behavioural dispositions in the way I have been describing – perhaps in some phylogenetically more ancient part of the brain – but also serves as the input to a thinking, concept-applying, and reasoning system; so that the subject’s thoughts, plans, and deliberations are also systematically dependent on the informational properties of the input.39 His account is similar, in some respects, to the Vaibhāṣika “two-step” scheme. And like the Vaibhāṣikas, Evans struggles to explain the relationship between the contents of our subliminal non-conceptual informational states and our conscious conceptual representations of them. While Evans argues that the conscious representations of sensory data involve a conceptual component, many who joined the discussion do not agree. The so-called non-conceptualists (or anti-conceptualists) aver to the intuition that infants and non-human animals, despite their lack of linguistic and conceptual faculties, are nevertheless conscious of their environment. (In Nagel’s terms, that it is like something to be them.) They also find it intuitively obvious that our immediate experience of color or music, to pick two popular examples, is too fine-grained to be captured fully in words and concepts. Some speak of the content of sense perception as “analog” in contrast to the “digital” content of propositional belief. As our capacity to discriminate perceptually far outstrips our capacity to discriminate conceptually, this analog perceptual capacity must be given a certain phenomenological priority. To quote Joseph Levine, “When you see a color and think ‘that color,’ the seeing is prior to the demonstrating, or else you really don’t know what you’re demonstrating…. But 38 39

The issues can be traced back to Kant, whom some identify as the first “conceptualist,” but Robert Hanna (2005, 2008) argues that Kant laid the groundwork for both conceptualism and non-conceptualism. Evans 1982: 158, emphasis in original. Evans goes on to argue that it is not necessary that conscious experience be tied to events for which a concept is ready at hand: “All I am requiring for conscious experience is that the subject exercise some concepts – have some thoughts – and that the content of those thoughts should depend systematically upon the informational properties of the input” (p. 159). See also his discussion of mental self-­ ascription (pp. 224–235).

Knowing Blue

55

if the seeing, the perceptual experience, is prior to the demonstrating, then the demonstrating can’t be what captures, or brings into existence, the content of that experience.”40 The non-conceptualist critique of the conceptualists can sound like Saṃghabhadra’s critique of Sautrāntika, insofar as they believe it impossible to make sense of our world without recourse to a notion of non-conceptual perceptual content. Levine writes, “It seems clear to me that somewhere within our representational system of thought there have to be links with the world that do not themselves employ the contents of other representations” (2010: 174). But the conceptualists refuse to give ground. They argue that it is impossible to connect percept with concept and belief unless sense-experiential states are conceptually structured to begin with, since evidential relations can only pertain among conceptual states. That is to say, what they call “sense experiential states” can underwrite empirical beliefs only insofar as they have conceptual content. As for the “richness” argument (i.e., the supposedly finegrained nature of color or music perception), conceptualists argue that we go astray in considering sense perception to be the cause of an ensuing perceptual experience. We should, rather, be looking for a constitutive account, in which case there is no necessary conflict between the fine-grained analog nature of sense data and our conceptual capacities. “On the conceptualist view, experience of a colour sample, R, just is a matter of entertaining a content in which the demonstrative concept ‘thatR shade’ figures as a constituent” (Brewer 2005: 221). To be clear, the conceptualists are not denying that we see, in all its high-resolution detail, “blue,” but this seeing is necessarily constituted, in part, by the conceptual apparatus through which “we” “know” that “we” are “seeing” “blue.” John McDowell sums up the position by saying that perceptual experience is “an actualization, in sensory consciousness, of conceptual capacities” (2009: 127). It should now be clear that the contemporary debates over the existence and nature of non-conceptual experience parallel the disputes that took place some two millennia ago among the Ābhidharmikas. The modern debates are no more and no less sophisticated, technical, scholastic, and interminable, and one wonders if any headway has been made. Indeed, I find it difficult to find a contemporary argument that is not at least anticipated in the ancient literature. If there is a lesson here, it may be that once the rupture is made between 40

Levine 2010: 191. For an overview of the literature on the conceptualist/non-conceptualist debate, see Bermúdez and Cahen 2015; Crowther 2006; and the “McDowell-Dreyfus debate” in Schear ed. 2013. On the non-conceptualist side see esp. Bermúdez 1995, 2003, 2007; Collins 1998; Coliva 2003; Peacocke 1992. On the conceptualist side see Bill Brewer 2005; McDowell 1994, 1996, 2009.

56

Sharf

knower and known, subject and object, first-person and third-person, nurture and nature, mind and world, it is impossible to bring them together again. Of course, human social life, not to mention philosophical reflection, is possible only within the space opened up by this rupture. My suspicion is that the problems encountered by both ancients and moderns in making conceptual sense of non-conceptual sense perception are due, in part, to our misunderstanding of the logical structure of the mind-world antinomy. Mind and world – knower and known – do not denote autonomous domains or frames of reference, or even interdependent perspectives. Rather, mind and world logically enfold one other; mind is only possible within the world, and the world only possible within the mind. Which is to say that the logical relationship among these antinomies is bound in paradox, and this, I believe, is key to understanding why the problem of non-conceptual experience seems so intractable. But the exploration of this topic will have to await another occasion. Acknowledgements This chapter is an abbreviated and slightly revised version of a longer article entitled “Knowing Blue: Early Buddhist Accounts of Non-conceptual Sense Perception” (Sharf 2018). My thanks to Ching Keng and Chen-kuo Lin who organized the workshop in which the paper was first presented, as well as to the editors of this volume who helped cut the paper down to a manageable size for this collection. I would also like to express a particular debt of gratitude to Katsura Shōryū, Lin Qian, and Yamabe Nobuyoshi, each of whom took particular care in going through the text and sharing their extensive expertise. Phyllis Granoff, Jowita Kramer, Don Lopez, Alexander von Rospatt, Elizabeth Horton Sharf, and Evan Thompson also provided invaluable comments and suggestions on earlier drafts. References Note: texts in the Taishō shinshū daizōkyō 大 正 新 修 大 藏 經 (edited by Takakusu Junjirō 高 楠 順 次 郎 and Watanabe Kaigyoku 渡 邊 海 旭 , Tokyo: Taishō issaikyō kankōkai, 1924–1932) are indicated by the text number (“T.”) followed by the volume, page, register (a, b, or c), and line number(s). Anālayo. 2011. A Comparative Study of the Majjhima-nikāya. Taipei: Dharma Drum.

Knowing Blue

57

Arnold, Daniel Anderson. 2010. “Self-Awareness (svasaṃvitti) and Related Doctrines of Buddhists Following Dignāga: Philosophical Characterizations of Some of the Main Issues.” Journal of Indian Philosophy 38 (3): 323–378. Arnold, Daniel Anderson. 2012. Brains, Buddhas, and Believing: The Problem of Intentionality in Classical Buddhist and Cognitive-scientific Philosophy of Mind. New York: Columbia University Press. Bermúdez, José. 1995. “Non-conceptual Content: From Perceptual Experience to Subpersonal Computational States.” Mind and Language 10 (4): 333–369. Bermúdez, José. 2003. Thinking Without Words. Oxford and New York: Oxford University Press. Bermúdez, José. 2007. “What Is at Stake in the Debate about Non-conceptual Content?” Philosophical Perspectives 21 (1): 55–72. Bermúdez, José, and Arnon Cahen. 2015. “Non-conceptual Mental Content.” The Stanford Encyclopedia of Philosophy (Spring 2015 ed.), Edward N. Zalta (ed.), url = . Block, Ned. 1995. “On a Confusion about a Function of Consciousness.” Behavioral and Brain Sciences 18: 227–287. Block, Ned. 1997. “Biology Versus Computation in the Study of Consciousness.” Behavioral and Brain Sciences 20: 159–166. Bodhi, Bhikkhu. 2000. The Connected Discourses of the Buddha: A New Translation of the Saṃyutta Nikāya. Somerville MA: Wisdom. Brewer, Bill. 2005. “Does Perceptual Experience Have Conceptual Content?” In Matthias Steup and Ernest Sosa, eds. Contemporary Debates in Epistemology, 217–230. Oxford: Wiley Blackwell. Bronkhorst, Johannes. 1986. The Two Traditions of Meditation in Ancient India. Stuttgart: Franz Steiner Verlag Wiesbaden GmbH. Carruthers, Peter. 2011. “Higher-Order Theories of Consciousness.” The Stanford Encyclopedia of Philosophy (Fall 2011 ed.). Edward N. Zalta, ed. url = . Coliva, Annalisa. 2003. “The Argument from the Finer-Grained Content of Colour Experiences: A Redefinition of Its Role within the Debate between McDowell and Non-Conceptual Theorists.” Dialectica 57 (1): 57–70. Collins, Arthur W. 1998. “Beastly Experience.” Philosophy and Phenomenological Research 58 (2): 375–380. Collins, Steven. 1982. Selfless Persons: Imagery and Thought in Theravada Buddhism. Cambridge and New York: Cambridge University Press. Coseru, Christian. 2012. Perceiving Reality: Consciousness, Intentionality, and Cognition in Buddhist Philosophy. Oxford and New York: Oxford University Press. Cousins, Lance S. 1992. “Vitakka/vitarka and Vicāra: Stages of Samādhi in Buddhism and Yoga.” Indo-Iranian Journal 35 (2): 137–157.

58

Sharf

Cox, Collett. 1988. “On the Possibility of a Nonexistent Object of Perceptual Consciousness: Sarvāstivādin and Dārṣṭāntika Theories.” Journal of the International Association of Buddhist Studies 11 (1): 31–87. Crowther, Timothy M., 2006. “Two Conceptions of Conceptualism and Non-conceptualism.” Erkenntnis 65 (2): 245–276. Deleanu, Florin. 2019. “Dating with Procrustes: Early Pramāṇavāda Chronology Revisited.” Bulletin of the International Institute for Buddhist Studies 2: 11–47. Dhammajoti, Kuala Lumpur. 2007a. Abhidharma Doctrines and Controversies on Perception. 3rd ed. Hong Kong: Centre of Buddhist Studies, University of Hong Kong. [First ed. 1997]. Dhammajoti, Kuala Lumpur. 2007b. “Ākāra and Direct Perception (Pratyakṣa).” Pacific World, 3rd series, 9: 245–272. Dhammajoti, Kuala Lumpur. 2008. Entrance into the Supreme Doctrine: Skandhila’s Abhidharmāvatāra. Hong Kong: Centre of Buddhist Studies, University of Hong Kong. Dhammajoti, Kuala Lumpur. 2009. Sarvāstivāda Abhidharma. 4th revised ed. Hong Kong: Center of Buddhist Studies, University of Hong Kong. [1st ed. Colombo, 2002]. Dreyfus, Georges. 1996. “Can the Fool Lead the Blind? Perception and the Given in Dharmakīrti’s Thought.” Journal of Indian Philosophy 24 (3): 209–229. Dreyfus, Georges. 1997. Recognizing Reality: Dharmakīrti’s Philosophy and Its Tibetan Interpretations. Albany: SUNY Press. Evans, Gareth. 1982. The Varieties of Reference. Oxford: Oxford University Press. Forman, Robert K.C. 1993. “Mystical Knowledge: Knowledge by Identity.” Journal of the American Academy of Religion 61 (4): 705–738. Forman, Robert K.C., ed. 1990. The Problem of Pure Consciousness: Mysticism and Philosophy. Oxford and New York: Oxford University Press. Frauwallner, Erich. 1995. Abhidharma Literature and the Origins of Buddhist Philosophical Systems. Sophie Francis Kidd and Ernst Steinkellner, trans. Albany: SUNY Press. Fukuda Takumi. 2003. “Bhadanta Rāma: A Sautrāntika before Vasubandhu.” Journal of the International Association of Buddhist Studies 26 (2): 255–286. Garfield, Jay L. 2006. “The Conventional Status of Reflexive Awareness: What’s at Stake in a Tibetan Debate?” Philosophy East and West 56 (2): 201–228. Gombrich, Richard F. 2006. How Buddhism Began: The Conditioned Genesis of the Early Teachings. 2nd ed. London and New York: Routledge. Hamilton, Sue. 1996. Identity and Experience: The Constitution of the Human Being. London: Luzac Oriental. Hanna, Robert. 2005. “Kant and Nonconceptual Content.” European Journal of Philosophy 13 (2): 247–290.

Knowing Blue

59

Hanna, Robert. 2008. “Kantian Non-conceptualism.” Philosophical Studies 137 (1): 41–64. Harada Wasō 原 田 和 宗 . 1996. “‘Kyōryōbu no “Tansō no” shiki no nagare’ to iu gainen e no gimon (1) < 経 量 部 の [ 単 層 の ] 識 の 流 れ > と い う 概 念 へ の 疑 問 (i).” Indogaku Chibettogaku kenkyū 1: 135–193. Harada Wasō 原 田 和 宗 . 1997. “‘Kyōryōbu no “Tansō no” shiki no nagare’ to iu gainen e no gimon (2) < 経 量 部 の [ 単 層 の ] 識 の 流 れ > と い う 概 念 へ の 疑 問 (ii).” Indogaku Chibettogaku kenkyū 2: 22–59. Harvey, Peter. 1989. “Consciousness Mysticism in the Discourses of the Buddha.” In Karel Werner, ed., The Yogi and the Mystic: Studies in Indian and Comparative Mysticism, 81–99. Richmond, Surrey: Curzon. Harvey, Peter. 1995. The Selfless Mind: Personality, Consciousness and Nirvāṇa in Early Buddhism. Richmond, Surrey: Curzon. Hattori Masaaki. 1968. Dignāga, On Perception, Being the Pratyakṣapariccheda of Dignāga’s Pramāṇasamuccaya from the Sanskrit Fragments and the Tibetan Versions. Harvard Oriental Series no. 47. Cambridge: Harvard University Press. Honjō Yoshifumi. 2003. “The Word Sautrāntika.” Journal of the International Association of Buddhist Studies, 26 (2): 321–330. Jaini, Padnamabh Shreevarma. 1977. Abhidharmadīpa with Vibhāṣāprabhāvṛtti. 2nd ed. Patna: Kashi Prasad Jayaswal Research Institute [1st ed. 1959]. Katz, Steven T., ed. 1978. Mysticism and Philosophical Analysis. Oxford and New York: Oxford University Press. Katz, Steven T., ed. 1983. Mysticism and Religious Traditions. Oxford and New York: ­Oxford University Press. Katz, Steven T., ed. 1992. Mysticism and Language. Oxford and New York: Oxford University Press. Kellner, Birgit. 2010. “Self-Awareness (svasaṃvedana) in Dignāga’s Pramāṇasamuccaya and -vṛtti: A Close Reading.” Journal of Indian Philosophy 38 (3): 203–231. Kellner, Birgit. 2014. “Changing Frames in Buddhist Thought: The Concept of Ākāra in Abhidharma and in Buddhist Epistemological Analysis.” Journal of Indian Philosophy 42 (2-3): 275–295. Kellner, Birgit, and Sara McClintock. 2014. “Ākāra in Buddhist Philosophical and Soteriological Analysis: An Introduction.” Journal of Indian Philosophy 42 (4): 427–432. Keng, Ching. 2019. “What is Svabhāva-vikalpa and with Which Consciousness(es) is it Associated?” Journal of Indian Philosophy 47 (1): 73–93. Kramer, Jowita. 2012. “Descriptions of ‘Feeling’ (vedanā), ‘Ideation’ (saṃjñā), and ‘the Unconditioned’ (asaṃskṛta) in Vasubandhu’s Pañcaskandhaka and Sthiramati’s Pañcaskandhakavibhāṣā.” Rocznik Orientalistyczny 65 (1): 120–139. Kritzer, Robert. 2003. “Sautrāntika in the Abhidharmakośa-bhāṣya.” Journal of the International Association of Buddhist Studies 26 (2): 331–384.

60

Sharf

Kritzer, Robert. 2005. Vasubandhu and the Yogācārabhūmi: Yogācāra Elements in the Abhidharmakośabhāṣya. Studia Philologica Buddhica Monograph Series 18. Tokyo: The International Institute for Buddhist Studies. Lau, Hakwan, and David M. Rosenthal. 2011. “Empirical Support for Higher-order Theories of Conscious Awareness.” Trends in Cognitive Sciences 15 (8): 365–373. Lin Qian. 2015. “Mind in Dispute: The Section on Mind in Harivarman’s *Tattvasiddhi.” Ph.D. diss., University of Washington. Levine, Joseph. 2010. “Demonstrative Thought.” Mind and Language 25 (2): 169–195. McDowell, John. 1994. “The Content of Perceptual Experience.” Philosophical Quarterly 44 (175): 190–205. McDowell, John. 1996. Mind and World. Cambridge: Harvard University Press. McDowell, John. 2009. “Conceptual Capacities in Perception.” In John McDowell ed., Having the World in View: Essays on Kant, Hegel, and Sellars, 127–144. Cambridge, MA: Harvard University Press. Matilal, Bimal Krishna. 1986. Perception: An Essay on Classical Indian Theories of Knowledge. Oxford and New York: Oxford University Press. Miyashita Seiki 宮 下 晴 輝 . 1986. “Kusharon ni okeru honmu kon’u ron no haikei: Shōgi kūshōkyō no kaishaku o megutte 『 倶 舎 論 』 に お け る 本 無 今 有 論 の 背 景 : 『 勝 義 空 性 経 』 の 解 釈 を め ぐ っ て .” Bukkyōgaku seminā 44: 7–37. Moriyama Shinya. 2010. “On Self-Awareness in the Sautrāntika Epistemology.” Journal of Indian Philosophy 38 (3): 261–277. Ñāṇamoli, Bhikkhu, and Bhikkhu Bodhi. 2009. The Middle Length Discourses of the Buddha: A Translation of the Majjhima Nikāya. 4th ed. Boston, MA: Wisdom. [1st ed. 1995]. Park, Changhwan. 2007. The Sautrāntika Theory of Seeds (bīja) Revisited: With Special Reference to the Ideological Continuity Between Vasubandhu’s Theory of Seeds and Its Śrīlāta/Dārṣṭāntika Precedents. Ph.D. dissertation, University of California, Berkeley. Peacocke, Christopher. 1992. A Study of Concepts. Cambridge: mit Press. Proudfoot, Wayne. 1985. Religious Experience. Berkeley: University of California Press. Rosenthal, David M. 1986. “Two Concepts of Consciousness.” Philosophical Studies 49 (3): 329–359. Rosenthal, David M. 1993. “Thinking that One Thinks.” In Martin Davies and Glyn W. Humphreys, eds., Consciousness: Psychological and Philosophical Essays, 197–223. Oxford: Blackwell. Rosenthal, David M. 2005. Consciousness and Mind. Oxford: Oxford University Press. Sastri, Shanti Bhikshu. 1953. Abhidharmāmṛta of Ghosaka. Santiniketan Visvabharati. Schear, Joseph K., ed. 2013. Mind, Reason, and Being-in-the-world: The McDowell-Dreyfus Debate. London and New York: Routledge.

Knowing Blue

61

Sharf, Robert H. 1995. “Buddhist Modernism and the Rhetoric of Meditative Experience.” Numen 42 (3): 228–283. Sharf, Robert H. 1998. “Experience.” In Mark C. Taylor, ed., Critical Terms for Religious Studies, 94–116. Chicago: University of Chicago Press. Sharf, Robert H. 2014. “Mindfulness and Mindlessness in Early Chan.” Philosophy East and West 64 (4): 933–964. Sharf, Robert H. 2016. “Is Yogācāra Phenomenology? Some Evidence from the Cheng weishi lun.” Journal of Indian Philosophy 44 (4): 777–807. Sharf, Robert H. 2018. “Knowing Blue: Early Buddhist Accounts of Non-conceptual Sense Perception.” Philosophy East and West 63 (3). Stace, Walter Terence. 1960. Mysticism and Philosophy. Los Angeles: Jeremy P. Tarcher. Strawson, Galen. 2015. “Self-intimation.” Phenomenology and the Cognitive Sciences 14 (1): 1–31. Takatsukasa Yūki. 2014. “The Problem of the Simultaneous Arising of Six Vijñānas: In the Pañcavijñānakāyasaṃprayuktabhūmi and the Manobhūmi.” Indogaku Bukkyōgaku kenkyū 62 (3): 1248–1252. Takatsukasa Yūki. 2016. “The Sequential Arising of Vijñānas in the Early Yogācāra and the Sarvāstivāda.” Indogaku Bukkyōgaku kenkyū 64 (3): 1222–1226. Walshe, Maurice. 1995. The Long Discourses of the Buddha: A Translation of the Dīgha Nikāya. Boston: Wisdom Publications. Watson, Alex. 2006. The Self’s Awareness of Itself: Bhaṭṭa Rāmakaṇṭha’s Arguments Against the Buddhist Doctrine of No-self. Vienna: Sammlung de Nobili, Institut für Südasien-, Tibet- und Buddhismuskunde der Universität Wien. Watson, Alex. 2010. “Bhaṭṭa Rāmakaṇṭha’s Elaboration of Self-Awareness (svasaṃvedana), and How It Differs from Dharmakīrti’s Exposition of the Concept.” Journal of Indian Philosophy 38 (3): 297–321. Williams, Paul. 1998. The Reflexive Nature of Awareness: A Tibetan Madhyamaka Defence. London and New York: Routledge Curzon. Yamabe Nobuyoshi. 2003. “On the School Affiliation of Aśvaghoṣa: ‘Sautrāntika’ or ‘Yogācāra’?” Journal of the International Association of Buddhist Studies 26 (2): 225–249. Yamabe Nobuyoshi 山 部 能 宜 . 1990. “Zenrokushiki no kaidōe ni kansuru ‘nandatōgi’ no kōsatsu 前 六 識 の 開 導 依 に 関 す る 「 難 陀 等 義 」 の 考 察 .” In Yuishikironsho ni okeru kanbun yōgo no chūjakuteki kenkyū 唯 識 論 書 に お け る 漢 文 用 語 の 註 釈 的 研 究 (ii), Bukkyō bunka kenkyūjo kiyō 28: 26–31. Yao, Zhihua. 2005. The Buddhist Theory of Self-Cognition. London and New York: Routledge.

Chapter 2

Nonconceptual Cognition in Yogācāra and Madhyamaka Thought John Spackman Abstract The claim that meditative states are nonconceptual plays a central role, both epistemologically and soteriologically, in much Buddhist philosophy. In its simplest form, this claim is subject to important objections. I follow Forman in distinguishing, in principle, between meditative experiences that have representational content (e.g. meditative experiences of everyday objects) and those that do not (e.g. “pure consciousness experiences”). If enlightened beings maintain their meditative practice throughout everyday life, this type of practice will be contentful. But if this practice is wholly nonconceptual, how can these beings make the perceptuallybased discriminations necessary to daily life (e.g. between wholesome and harmful food), since such discriminations are usually thought to require the application of concepts? One response to this challenge would be to say that enlightened beings have special access to a level of perceptual discrimination that is nonconceptual. But I argue against such a view, and in order to develop a different response I distinguish two different senses of nonconceptuality. What is primarily at stake in recent debates about nonconceptualism, I suggest, is what I call supervenience nonconceptuality. To say meditative experiences are nonconceptual in this sense is to say that they do not supervene on the subject’s conceptual capacities, that is, that it is metaphysically possible for there to be differences in the subject’s meditative experiences without any differences in their conceptual capacities. Against the above response, I argue that any experience with representational content must be conceptual in this supervenience-based sense, so we must accept that contentful meditative experiences are conceptual in this sense. However, I argue that contentful meditative experiences are nonconceptual in another, “occurrent,” sense. To say such experiences are nonconceptual in this sense is to say that the subject does not actively employ concepts in thought while undergoing them, which I argue must ultimately be cashed out in terms of the fact that the subject does not identify as the thinker of those thoughts. I argue further that if there are such things as pure consciousness experiences, they would be nonconceptual in both senses.

© koninklijke brill nv, leiden, ���1 | doi:10.1163/9789004440913_005

Nonconceptual cognition in Yogācāra and Madhyamaka Thought

63

Keywords Yogācāra – Madhyamaka – nonconceptual – meditation – ­nonduality – Asaṅga – Kamalaśīla – Candrakīrti

…

1 Introduction The notion that, by virtue of extensive meditation and other forms of practice, advanced Buddhist practitioners can achieve a state of nonconceptual cognition (nirvikalpajñāna) that transcends the delusions of ordinary conceptual activity is one that is extensively developed in many Mahāyāna texts.1 As Robert Sharf points out in his contribution to this volume, while the claim that advanced meditative states are nonconceptual did occasionally appear in early sutras and the Abhidharma literature, it was not extensively developed there. Nonconceptual cognition does, however, come to take a central place later, in the soteriology of Yogācāra and Madhyamaka thought, where it is closely connected with prajñā, wisdom, and it also plays a central role, as Sharf notes, in the later traditions of Tantra, Chan, and Dzogchen. There are, however, a variety of different, often incompatible, notions of nonconceptual cognition that appear in different Mahāyāna texts, and there is also commonly disagreement among interpreters as to what model any particular text espouses. This paper examines a number of these different models of nonconceptual cognition in order to see what we can learn from them about how best to understand the sense in which there might be something conceptually distinctive about the mental states of advanced Buddhist practitioners.2 My approach to this issue will be broadly akin to what Mark Siderits has called “fusion” or “confluence” philosophy (Siderits 2015). That is, my central goal here is not to offer an interpretation of any particular Buddhist text or texts, though some interpretive efforts will be necessary. My main aim is, rather, to draw both on Buddhist models of nonconceptuality, and on resources developed in recent 1 While the Sanskrit term nirvikalpajñāna is sometimes translated as “nonconceptual awareness,” I here render it as “nonconceptual cognition.” 2 For further discussion of the notion of nonconceptual cognition in Yogācāra texts in particular, see the paper by Roy Tzohar (Chapter 3). I will not address here the notion of mānasapratyakṣa or mental perception discussed by Ching Keng (Chapter 9), since although it is sometimes viewed as a kind of nonconceptual or direct perception, it raises a somewhat different set of issues.

64

Spackman

discussions of nonconceptual content in analytic philosophy, in order to shed light on a problem of general interest, namely in what sense, if any, certain mental states in advanced Buddhist practitioners should be viewed as nonconceptual. There has in recent years been a surge of interest in empirical studies of meditation and mindfulness in psychology and neuroscience, and in some of these there has been an assumption that these practices are in some sense nonconceptual, though there has been little empirical investigation of this particular issue.3 If we are to understand better whether and how meditative practices might foster nonconceptual mental states, we need a better grasp of the various senses in which certain states have been viewed as nonconceptual in the Buddhist tradition, and of the critical challenges facing these different notions of nonconceptuality. My approach in this paper will also be a naturalistic one. In the texts we will consider, buddhas and bodhisattvas are often regarded as supernatural beings, capable of miraculous deeds such as healing the sick and granting wishes. My interest here is what we can learn from these texts about nonconceptual awareness insofar as it is something potentially available to advanced practitioners lacking such supernatural status. One reason why different notions of nonconceptuality emerged in Mahāyāna thought is that different texts sought to deal in different ways with a central problem that faces the idea that the cognition of buddhas and bodhisattvas is nonconceptual. Let’s call this the teaching problem. It is central to the Mahāyāna understanding of buddhas and bodhisattvas that after achieving liberation, these beings remain in the world, teaching with expedient means in order to liberate sentient beings at different levels of spiritual development. And yet recognizing these different levels of development, crafting different types of teaching, using language to convey those teachings – all of these seem to require high-level abilities to discriminate types of things, and such abilities are the hallmark of concepts. So if the liberated state of buddhas and bodhisattvas is wholly nonconceptual, how is it possible for them to engage in such teaching? In what follows, I will distinguish several different models of the cognition of buddhas and bodhisattvas that various Mahāyāna texts have been seen as presenting, which deal with the teaching problem in different ways. These models are of two broad types, which might be called two-tiered models and 3 To take just one example, the psychologist Amishi Jha and her colleagues have defined mindfulness in this way: “Mindfulness is a mental mode characterized by attention to present moment experience without conceptual elaboration or reactivity” (Morrison et al., 2013). See also Jha et al., 2010, p. 54.

Nonconceptual cognition in Yogācāra and Madhyamaka Thought

65

single-tiered models. The two-tiered models solve the teaching problem by holding that only states of meditative absorption of buddhas and bodhisattvas are strictly nonconceptual. They allow that while these beings are engaged in teaching and other mundane activities, their mental states are conceptual. I will distinguish three quite different versions of the two-tiered model, which I  will call the contentless absorption model, the nonconceptual content model, and the no active thought model. Single-tiered models, on the other hand, view all of the mental states of buddhas and bodhisattvas as nonconceptual, including those involved in teaching and mundane activities; they avoid the teaching problem by maintaining that there is a way of engaging in apparently conceptual activities like teaching that is, at a deeper level, nonconceptual. I will focus in particular on what might be called the no content model. On the basis of a consideration of the challenges facing these models, I will argue that we have good reason to adopt a single-tiered model rather than a two-tiered one. Though there is of course a sense in which one might view states of meditative absorption as nonconceptual and the everyday states of buddhas and bodhisattvas as conceptual, there is a deeper sense in which both types of state are properly regarded as nonconceptual. In particular, I will propose a single-tiered account which draws on the insights of some of the other models, what I will call the nondual cognition model. On this account, contrary to what one might expect, I will suggest, the nonconceptuality of the experience of advanced Buddhist practitioners requires neither that the content of their experience, nor the mental acts involved in it, be independent of the subject’s concepts; it is, rather, a matter of taking a certain attitude toward one’s actions and experience, a receptive, nondual attitude. Though this specific model is not to my knowledge espoused by any Buddhist text, it offers what I take to be a plausible philosophical account of nonconceptual cognition that both avoids the difficulties facing the other models under discussion, and is consistent with fundamental Mahāyāna commitments. Before turning to the models of nonconceptual cognition it will be helpful to consider more fully different possible meanings of “nonconceptual,” and to clarify the notion of nonconceptuality with which I will be working here. 2

Nirvikalpajñāna and Nonconceptuality

The Sanskrit word most widely translated as “nonconceptual,” nirvikalpa, has a range of meanings that is slightly different from the English “nonconceptual,” and it is important to bear this difference in mind in our discussion here. As Paul Griffiths suggests, the term vikalpa as used by Indian Buddhist intellectuals

66

Spackman

is best thought of as “the constructive, conceptual, classificatory activities of the mind,” but there is a strong emphasis on the constructive aspect of this activity (Griffiths 1990, 85–87). For instance the Bodhisattvabhūmi, as Griffiths points out, registers eight different types of vikalpa, constructive activity, including the construction of svabhāva (intrinsic natures or essences), distinctions (viśeṣavikalpa), and I (aham iti vikalpa).4 It might, perhaps, be argued that all of these varieties of vikalpa have their root in the construction of concepts themselves, so that at least part of the sense of nirvikalpa is “nonconceptual.” But it is important at any rate to bear in mind in what follows that the term places a strong emphasis on human constructive activity. In discussing Buddhist models of nonconceptual cognition it will be helpful to situate these models in relation to several different accounts of nonconceptuality that have been distinguished in recent debates about whether perceptual experience has nonconceptual content. Early participants in these debates, beginning with Gareth Evans (1982), operated with a definition of nonconceptuality that has more recently been termed “state nonconceptualism” (snc) (Heck 2000). According to this account: snc A subject S’s perceptual experience at time t is nonconceptual if and only if it is not necessary, in order for the subject to have that experience, that she possess at t the concepts used in a correct characterization of all of the constituents of its presentational content.5 State conceptualism (sc) would then be the denial of snc. The idea of snc is, in effect, that a perceptual state is nonconceptual if and only if its content contains constituents that cannot be captured by the concepts available to the subject. Evans argued, in particular, that perceptual states are nonconceptual in this respect because their contents are more fine-grained than the repertoire of concepts their subjects possess. On this view, for example, it is frequently

4 It is worth emphasizing that although the segments of the Buddhist tradition that we will be investigating here viewed vikalpa mainly as objects of suspicion, some Buddhist schools – for instance, arguably, the non-Yogācāra Abhidharma schools – viewed the distinctions drawn by means of some types of vikalpa as potentially beneficial, at least at the level of conventional truth. 5 In contemporary analytic philosophy of mind, the content of mental states is typically ­spoken of as “representational content.” Because the notion of representation of a mind-­ independent world is inimical to many schools of Buddhism, though, I will here speak of “presentational content,” content that presents a world as being a certain way. I intend this notion to be neutral on the question of whether a mind-independent world exists.

Nonconceptual cognition in Yogācāra and Madhyamaka Thought

67

the case that no concept possessed by the perceiver can adequately capture the specific shade an object looks to her.6 More recently, a number of philosophers have argued that the state definition of nonconceptuality does not capture the sense of nonconceptuality that is really at issue in debates about perceptual experience. What nonconceptualists about perception are really seeking to demonstrate, on this view, is what has been called content nonconceptualism (cnc):7 cnc Perceptual contents are nonconceptual if and only if they are different in kind from the contents of paradigmatic conceptual states like belief. Perhaps, for instance, perceptual contents are iconic while belief contents are propositional, or perhaps they are composed of Russellian propositions while beliefs are Fregean. Proponents of this definition have argued that content nonconceptualism is logically independent from state nonconceptualism. Even if a subject could have a certain perceptual content without possessing the corresponding concepts, this is compatible with the perceptual content being of the same kind as conceptual contents, e.g. propositional. Similarly, even if state conceptualism is true, this does not entail that perceptual contents are of the same kind as conceptual contents; even if subjects possess ­concepts for every element of perceptual content, these concepts might not reflect the intrinsic nature of the content itself, which might be nonconceptual. It is unclear, however, that the content definition of nonconceptuality is appropriate for all dialectical contexts. I have argued elsewhere that it would be misguided to think of the notion of nonconceptuality in which we are interested in many contexts as turning on any very specific account of the nature of experiential content, as cnc maintains (Spackman, manuscript). It may, of course, be of considerable interest for some purposes in cognitive science to discover that perceptual contents are iconic rather than propositional. But in some contexts, what we are interested in is not the nature of experiential content itself, but whether it depends closely on the concepts possessed by the subject. Even if it turned out that perceptual contents are iconic while belief contents are propositional, if we discovered that the specific perceptual contents a subject can have depend in a strong way on the concepts she 6 See Evans 1982, 229. This “fineness of grain” argument has elicited a vigorous debate. For an opposing view, see McDowell 1994, Lecture iii. 7 The notion of content nonconceptualism originates with Heck 2000. Others who adopt this distinction include Tye 2006, Raftopoulos and Müller 2006, Bermudez 2007, and Hanna 2008. For a variety of proposals concerning how perceptual contents must differ from conceptual contents in order to be nonconceptual, see Kelly 2001, Heck 2007, Hanna 2008, Tye 2006.

68

Spackman

­ ossesses, this dependency would specify one important sense in which those p contents might be called conceptual. If for instance the perceptual contents of two organisms A and B were both iconic, but A’s contents differed from B’s (e.g. presenting temporally persistent objects as opposed to merely light and dark patches) in a way that depended on their different conceptual abilities, in one important sense it would be natural to view these contents as conceptual. In order to capture the way in which experiential contents can depend on concepts, I will here take up a proposal by Philippe Chuard (2007), that we construe the question of whether perceptual content is nonconceptual in terms of whether it supervenes on the subject’s concepts. On this view, the important issue is whether it is metaphysically possible for there to be a difference in the content of a subject’s experiences without a difference in the conceptual capacities exercised in those experiences. According to supervenience nonconceptualism (spnc): spnc A subject S’s perceptual content is nonconceptual if and only if it is not necessarily the case that, for any two distinct constituents c1 and c2 in S’s perceptual content, S possesses distinct concepts of c1 and c2 which are exercised in distinct conceptual capacities in relation to them. spnc captures a notion of nonconceptuality that is relevant in many contexts, and I would suggest that it is an appropriate notion to use in the context of many of the Buddhist debates concerning nonconceptual cognition The crucial issue in these debates is arguably an epistemological one. In those parts of the Buddhist tradition in which the notion of nonconceptual cognition was emphasized, concepts were viewed, at least in part, as sources of cognitive error, and the central goal of this notion was thus to identify a state independent of those errors, a state of correct apprehension of reality. A natural way of representing what is at issue in these debates is thus in terms of whether the content of that state can vary independently of concepts, and this is what spnc captures. In considering below what we can learn from the Buddhist models about the states of advanced practitioners, then, it will be helpful to appeal to this supervenience account of nonconceptuality. 3

Models of Nonconceptual Cognition

Let’s turn now to the different Buddhist models of nonconceptual cognition  Some of the most fully developed discussions of nonconceptual cognition in the Buddhist canon present models which are two-tiered in the

Nonconceptual cognition in Yogācāra and Madhyamaka Thought

69

sense I have specified. The discussions in the Mahāyānasūtrālaṃkāra and the Mahāyānasaṃgraha, both traditionally attributed to the fourth century Yogācāra philosopher Asaṅga, are prime examples.8 In the Mahāyānasaṃgraha, for instance, we find a distinction between three kinds of nonconceptual cognition: preparatory nonconceptual cognition (prāyogikanirvikalpakajñāna), fundamental nonconceptual cognition (mūlanirvikalpakajñāna), and subsequent nonconceptual cognition (pṛṣṭhalabdhanirvikalpakajñāna). The preparatory and subsequent nonconceptual cognitions are, however, called “nonconceptual” merely by courtesy, as it were. The preparatory cognition is what a bodhisattva gains in hearing talk of “nonconceptual knowledge” and striving to attain it through practice. It is thus clearly conceptual; 8.14Bh says it is called nonconceptual only because it “gives rise to fundamental nonconceptual knowledge” (Lamotte 1938, 332). The subsequent cognition which follows after fundamental nonconceptual cognition is what allows a bodhisattva to take rebirth and engage in teaching; it is explicitly said at 8.16Bh to involve concepts (Lamotte 1938, 336), and is called nonconceptual only because it “comes from fundamental nonconceptual knowledge” (Lamotte 1938, 333). As to the fundamental nonconceptual cognition it is said to be “nonconceptualizing” because it is “without concept (vikalpa)” and “nonconceptual” because “it is not conceived of by others” (Lamotte 1938, 8.16bh, 335). As Tzohar emphasizes (Chapter 3), the ultimate soteriological goal presented in these texts is actually not fundamental nonconceptual awareness, but rather, because of its salvific role, subsequent awareness. Clearly, however, the causal source of subsequent awareness is fundamental nonconceptual awareness. There are, however, different ways of interpreting the nature of nonconceptual cognition presented in these texts. One is along the lines of what I called the contentless absorption model, which holds, as the name suggests, not that nonconceptual cognition has nonconceptual content, but that it contains no appearances and no presentational content at all. Dan Arnold, for instance, appears to offer such an account in his reading of the Mahāyānasaṃgraha. Arnold points out that verse 8.2 of that work says that nirvikalpakajñāna avoids citrīkāra, objectifying reality or “making a picture,” and indeed the Upaniban­ dhana commentary9 asserts that this is in fact the essence (svabhāva) of nirvikalpakajñāna (Arnold 2003, 16). This might seem to contradict the very nature of cognition since it is generally accepted in the Buddhist philosophical tradition that any type of cognition (jñāna) must have an intentional object 8 For an account of the role of nonconceptuality in the Mahāyānasūtrālaṃkāra, see D’Amato (2009). 9 This commentary is attributed to Asvabhāva (translated in Lamotte 1938).

70

Spackman

(ālambana).10 However, verse 8.5 clarifies that the cognitive object (ālambana) of nirvikalpakajñāna is “the ineffability of dharmas, that is to say, essenceless suchness” (Arnold 2003, 23). Arnold thus concludes that nonconceptual cognition does have an intentional object, but that this object “turns out to be something like the abstract state of affairs of there being no ultimate objects that could be intended” (Arnold 2003, 24). Every intentional object has, according to the text, an aspect (ākāra), which Arnold interprets as equivalent to “appearance or phenomenal aspect.” Verse 8.6 states more specifically that “The phenomenal aspect of the nonconceptual awareness of bodhisattvas is the absence of appearances (nimitta) with respect precisely to that intentional object which is to be known” (Arnold 2003, 25). For Arnold, the implication is that what is unique about nonconceptual cognition is that it lacks precisely what is often taken to be one of the hallmarks of cognition – appearances, or phenomenal characters, themselves. To put it in terms familiar from contemporary philosophy of mind, on this model there would not be anything it is like to be a buddha or a bodhisattva, at least in moments of nonconceptual cognition. Notice that the contentless absorption model is not committed to viewing the cognition of buddhas and bodhisattvas as nonconceptual in any of the senses described earlier. Each of those accounts defines nonconceptuality in terms of the presentational content of experience. But on the contentless absorption model, there are no appearances or content in nonconceptual cognition so a fortiori such cognition has no nonconceptual content. The second two-tier model to be considered, the nonconceptual content model, differs on precisely this point. The nonconceptual content model maintains that states of nonconceptual cognition do contain appearances, and do have content. But it holds that this content is nonconceptual in the sense ­specified by snc – that the subject’s concepts do not adequately capture this content – and holds that this is so, more specifically, because concepts involve a duality between subject and object, and nirvikalpajñāna is nondualistic in this sense. It might be argued, as well, that for this model the deeper source of this state nonconceptuality is that this state is content nonconceptual, for its content differs in kind from the content of conceptual states in being nondual. And it also seems clear that on this view nirvikalpajñāna would be supervenience nonconceptual. Two interpreters who seem to adopt this kind of model in their interpretation of the above texts, the Mahāyānasūtrālaṃkāra and the Mahāyānasaṃgraha, are Paul Griffiths and Mario D’Amato.

10 This claim is made, for instance, in the Abhidharmakośa of Vasubandhu. See Abhidharmakośabhāṣyam ad 5.25 (Pradhan 1975: 295).

Nonconceptual cognition in Yogācāra and Madhyamaka Thought

71

Griffiths, for instance, explicitly rejects the view that nonconceptual cognition is “without phenomenological attributes or content” (Griffiths 1990, 87). This he does because, for instance, at Mahāyānasaṃgraha/Upanibandhana 8.2, nonconceptual cognition is said not to consist simply in the absence of thinking (amanasikāra), since then dreamless sleep and unthinking drunkenness would count as nonconceptual cognition. What differentiates nonconceptual cognition from such states, on this view, is that it has an ālambana, a cognitive object, and an ākāra, which Griffiths interprets to mean content. This ākāra, as we have seen, is said to be the absence of nimitta, but for Griffiths nimitta are not simply appearances, but “those things which our constructive intellect develops in order to divide and classify objects in the world of experience” (Griffiths 1990, 88). What ultimately distinguishes the appearances of nonconceptual cognition from ordinary appearance, for Griffiths, is that they transcend the subject-object dualism that dominates ordinary experience. As he puts it, nonconceptual cognition “never confronts objects of awareness, in that it does not function in accordance with the usual divisions of objects of awareness into such things as physical form; this, in turn, is because in mirrorlike knowledge there is no difference between apprehension and that which is to be apprehended” (Griffiths 1990, 89). Griffith’s model draws, it seems, on certain passages in the Mahāyānasūtrālaṃkāra and the Mahāyānasaṃgraha that seem to attribute the nonconceptuality of nirvikalpakajñāna to its nondual nature. The most explicit statement appears at Mahāyānasaṃgraha 8.18Bh: it is because it is not distinguished from its object (grāhya) and is completely identified with it (samasama), that the fundamental nonconceptual knowledge is called nonconceptual: this knowledge does not set up the duality of object and subject (grāhyagrāhaka). (Lamotte 1938, 339) I will have more to say later about how we might understand this notion of the nonduality of nonconceptual cognition and its object. The third model of nonconceptual cognition I will consider, the no active thought model, takes a quite different tack from the first two. This model accepts that nonconceptual cognition includes appearances and content, but holds, contra the nonconceptual content model, that this content is conceptual in the sense of sc, that it can always in principle be captured by the subject’s concepts. Such cognition is nonetheless nonconceptual in that at the time a subject is in such a state, she is not actually engaging in conceptualization about any intentional objects. It thus locates nonconceptuality not at the level of presentational content, but at the level of the subject’s attitude toward

72

Spackman

that content. The model also seems to accept that this cognition is content conceptual; it is the kind of thing that can in principle be captured by concepts. And given that content conceptualism is a stronger doctrine than supervenience conceptualism, it would count as supervenience conceptual as well. To see how such an account might work, consider the view of the 8th century Indian Madhyamaka philosopher Kamalaśīla. In his Bhāvanākrama, Kamalaśīla offers an extended response to certain teachers who maintain that “those who think of nothing and do nothing are liberated from saṃsāra” (Demieville 1952, 348).11 He argues, in particular, that the mere absence of recollection and reflection (smṛtimanasikārābhāva) could never bring about a state of nonconceptuality, because the nonexistence (abhāva) of something can never be a cause (Demieville 1952, p. 350). Given that humans are accustomed to believe in the reality of form or matter (rūpa), what is necessary to penetrate into nonconceptual cognition is correct analysis (bhūta-pratyaveksā), which alone makes possible the knowledge of the emptiness of all dharmas (Demieville 1952, 349–351). There is controversy among commentators as to whether this conceptual analysis serves merely as a necessary condition for the attainment of nonconceptual cognition or is capable of capturing the content of that cognition itself. I will consider the account of one interpreter who takes the latter view, Paul Williams. On Williams’ reading, in his debate or debates with the Chinese monk Mahāyāna, Kamalaśīla puts forth the view that, while the state of cognition in absorption itself is nonconceptual in that it does not involve active thought, its content is nonetheless conceptual in the sense of sc, in that this content could in principle be captured by concepts possessed by the subject. As Williams puts it, “It follows that for Kamalaśīla while the mental state itself may be non-conceptual in that a person enjoying such a state is not engaging in conceptualizing, it is possible for others, or the yogin after his gnosis, to adequately explain what the state was, and its content” (Williams 1992, 196). Whether Williams takes Kamalaśīla’s account to count as content conceptualist as well is less clear. But at any rate it seems that since the yogin is said to be able to capture all distinctions within the content of nonconceptual cognition verbally, this state is portrayed as supervenience conceptual. It might be doubted that the kind of nonconceptual mental state envisioned by Kamalaśīla is really possible – after all, how could the content of a mental state be conceptual if the subject of that state is not actively engaged in thought about that content? Kamalaśīla does not offer any guidance on the subject, but for a model of how to make sense of such a state we might look to John McDowell’s account in Mind and World of how a subject’s conceptual 11

The renderings from Demieville’s French are mine.

Nonconceptual cognition in Yogācāra and Madhyamaka Thought

73

abilities are drawn on passively in constituting her perceptual content as conceptual. As McDowell puts it, “In experience one finds oneself saddled with content. One’s conceptual capacities have already been brought into play, in the content’s being available to one, before one has any choice in the matter” (McDowell 1994, 10).12 Finally, let’s consider a model of nonconceptual cognition that differs from those we have so far discussed in that it is single-tiered, that is, it holds that all of the mental states of buddhas and bodhisattvas, including those involved in teaching, are nonconceptual. Candrakīrti, at least on the reading of interpreters such as John Dunne and Mario D’Amato, adopts a model of this kind, which could be called a no content model. As both of these interpreters emphasize, unlike our first two models and more akin to that of Kamalaśīla, Candrakīrti holds in his Madhyamakāvatāra that all experiential content, including the barest sense-data, is conceptual in the sense of sc. And yet he also affirms that buddhas and bodhisattvas are, as he puts it, entirely “beyond conceptual mind” (Mipham and Candrakirti 2002, 105). As Dunne understands it, the implication, since all experiential content is conceptual, is that for buddhas and bodhisattvas, there are no appearances whatsoever – as he puts it, “they do not see tables, chairs, people, places” (Dunne 1996, 545). The claim here is analogous to the contentless absorption model, but in this case there are no appearances or content either in absorption or in teaching and everyday activities. Just as for the earlier model, then, this model cannot be said to adopt snc, cnc, or spnc in regard to nonconceptual cognition The obvious question however is how teaching and everyday activities are possible for buddhas and bodhisattvas, if they have no experience. As Dunne understands it, Candrakīrti’s view is that “A buddha teaches in a magical, transcendent way” (Dunne 1996, 549). It is in fact the dharmakāya, the formless reality-body of a buddha, that “causes a didactic sound to emit from” the buddha’s emanation body (nirmāṇakāya) in a manner appropriate to the context; but throughout this process, the buddha remains without appearances, without cognitive activity, and without effort. 4

Challenges for the Models of Nonconceptual Cognition

As I have said, my main goal here is not exegetical, but rather, to draw on the models of nonconceptual cognition we have discussed in order to consider in what sense the meditative states of advanced Buddhist practitioners should be 12

As a conceptualist, of course, McDowell might not be willing to endorse Kamalaśīla’s conception of the nonconceptual awareness of buddhas and bodhisattvas.

74

Spackman

viewed as nonconceptual. With this goal in mind, I will now consider some difficulties facing these models. In particular I will raise two challenges, one concerning any model that accepts supervenience nonconceptualism, and one raising what I see as a more general issue facing any two-tiered model. Let’s consider first the difficulties facing supervenience nonconceptualism, which among the views we have discussed, is adopted by the nondual content model. There has in recent years been a great deal of debate over the viability of various forms of nonconceptualism, and I cannot hope to do justice to that debate here.13 I do want at least to suggest, however, the height of the hurdles faced by supervenience nonconceptualism, by presenting in brief a line of argument I have developed more fully elsewhere, which I call the matching argument (Spackman, manuscript). The aim of this argument is to support supervenience conceptualism (spc), for any being “conceptually like us,” in regard not only to the states of buddhas and bodhisattvas but to the presentational content of all mental states. By a being “conceptually like us,” I mean any being with most of the conceptual capacities we have, e.g. capacities for discriminating types of things, for consciously recognizing that a thing is of a certain type, for learning new concepts from experience, and for using concepts in inference. It is not part of my argument, then, that the experiential content of all animals or all possible beings is supervenience conceptual. As John McDowell has pointed out, it is open to conceptualists to accept that the perceptual content of many animals is nonconceptual, but to hold that in humans and other beings that have conceptual abilities, perceptual experience becomes suffused with conceptual content (McDowell 1994). However, the argument for spc in relation to beings conceptually like us should cover, in addition to humans and perhaps some other animals, buddhas, bodhisattvas, and Buddhist practitioners generally, since they have capacities for discrimination, conscious recognition, learning, and inference. The core idea behind the argument is that in beings conceptually like us, any distinction between constituents in experiential content necessarily brings with it a distinction in abilities to recognize the types to which those constituents belong, which are quintessentially conceptual abilities. The argument might be formalized as follows, with S referring to any being conceptually like us:

13

See for instance Peacocke 1992, McDowell 1994, Heck 2000, Chuard 2006, Kelly 2001, Dokic and Pacherie 2001.

Nonconceptual cognition in Yogācāra and Madhyamaka Thought

75

1.

It is a necessary condition for a token c of a perceptible property f to be a constituent in the experiential content of S that S have the ­ability to ­consciously recognize some other actual or possible property-token c* as being of the same type as c (i.e., f). 2. If S has the ability to consciously recognize two property-tokens c and c* as being of the same type f, and on some views also satisfies certain general conditions on concept possession, she satisfies the minimal conditions on possessing a concept of f. 3. It is a necessary condition for a token c of a perceptible property f to be a constituent in the experiential content of S that S possess a concept of f. (1, 2) 4. Necessarily, for any two tokens c1 and c2 of distinct property-types in the perceptual content of S, S possesses distinct concepts of c1 and c2 which are exercised in distinct conceptual capacities in relation to them. 5. The experiential content of S is supervenience conceptual. (3, 4, spc) Let’s consider the central premises of this argument. The first premise seems to be supported by many everyday experiences. For instance, when we see an equilateral triangle, we can generally easily match it with another equilateral triangle as opposed to an isosceles triangle. But there are a number of objections that might be raised against this claim. One objection might come from Evans’ “fineness of grain” argument, mentioned above. If it were true that there are many discriminable shades of red for which subjects have no distinct concept, it could be argued that they would have no matching ability of the kind specified in premise 1. However, McDowell has argued, persuasively to my mind, that even though we do not possess a distinct word for each distinct item of perceptual content, nonetheless we have the ability to form demonstrative concepts corresponding to any item we see, when we see it. For instance, presented with a specific shade of red, we can form the concept THAT SHADE,14 which we can then subsequently use to pick out shades of the same type in our visual experience. It has been claimed, however, that there is substantial empirical evidence that ordinary perceivers often do not retain particular shades and shapes in memory for long enough to recognize them even just after the initial stimulus has disappeared (Tye 2006, Kelly 2001, Dokic and Pacherie 2001). The worry is that if this is the case, then it is implausible to attribute demonstrative concepts of these shades and shapes to perceivers, since a concept must be capable 14

In what follows, words in capitals refer to concepts.

76

Spackman

of being correctly used on multiple occasions. But while this constraint on concept use is reasonable, it does not entail that in order to possess a concept a subject must be able to use it after the initial stimulus has disappeared. It would be sufficient to justify attribution of a demonstrative concept THAT SHADE in relation to a given perceived shade if a subject could recognize a second shade visible at the same time as the first as being of the same type.15 Of course, it is unusual for us to perceive two samples of the same shade at the same time, so it might be thought that this response doesn’t help much. But even if no second shade is actually present, the matching ability still exists, and it is the ability that justifies attribution of the demonstrative concept. This kind of ability is precisely the recognitional ability specified in premise 1, and it seems to me that subjects do always have abilities of this kind in relation to the constituents of their experiential content. In comparison to the first premise, the second premise of this argument should be relatively uncontroversial. A plausible case could be made for this claim whether one adopts a less demanding conception of the conditions on concept possession or a more demanding one. Less demanding accounts, which often credit non-linguistic animals with concepts, typically focus on some such requirement as that in order to possess a concept F, the subject must be able to respond differentially in a regular way to multiple items of the relevant kind.16 Premise one entails that the subjects in question do have such abilities, and so less demanding accounts would count them as having the relevant concepts. More demanding accounts of concept possession, which often reserve concepts for human beings, emphasize additional requirements, for instance that subjects who possess a concept F must be capable of reflectively recognizing that something is an F, or forming beliefs involving F for the reason that their perceptual experience has a certain content.17 Since my claim for supervenience conceptualism was restricted to beings conceptually like us, and beings conceptually like us will possess these more sophisticated conceptual abilities, premise two should be acceptable regardless of one’s views on concept possession. Premise four maintains that any subject conceptually like us necessarily possesses distinct concepts matching every distinct property in perceptual content. This might seem to fly in the face of our sense that we often see more distinctions than we can recognize. One way of bringing this objection into 15 See Chuard 2006 for a similar argument. 16 See for instance Millikan 1998, Dretske 1999. 17 The former of these requirements is advocated for instance by Peacocke 1992, the latter by McDowell 1994.

Nonconceptual cognition in Yogācāra and Madhyamaka Thought

77

focus is by means of a consideration of cases of associative agnosia. In this condition, subjects with normal visual capacities lack the ability to recognize via visual perception the kind of thing a perceived object is, though they may recognize the object by for instance touching or smelling it. One such patient, for instance, when presented with a stethoscope, described it as “a long cord with a round thing at the end,” and suggested it might be a watch.18 But cases of associative agnosia do not represent counterexamples to premise four. The case just mentioned, for instance, only shows that for the patient in question, since he cannot recognize this and other similar objects as being of the kind “stethoscope,” being a stethoscope is not among the properties in his perceptual content. On the other hand, he still does perceive properties such as long and round, but presumably he could match the present tokens of these properties with others of the same kind. So the claim that subjects necessarily possess distinct concepts corresponding to the distinct properties in their perceptual content remains unchallenged. While this brief sketch of the matching argument cannot hope to build a thorough case against supervenience nonconceptualism, it may perhaps suffice to indicate one central challenge it faces. I will thus take it that whatever model of nonconceptual cognition we develop, it will have to allow that the content of this cognition is conceptual in the sense of spc. The second challenge I will raise is a general epistemological concern about any two-tiered model of nonconceptual cognition, a concern which gives us, I will argue, good reason to favor a single-tiered model. All of the two-tiered models considered above portray the nonconceptual tier of the cognition of buddhas and bodhisattvas, the level in which they are deeply absorbed in meditation, as nondelusory, a veridical perception of reality. But what about the conceptual tier of their cognition, the level of teaching and everyday activity? Is it capable of being nondelusory too? Unless proponents of these models are willing to accept that buddhas and bodhisattvas are generally deluded about this level of reality, this conceptual tier will presumably have to be viewed as in some sense nondelusory as well. The Mahāyānasaṃgraha, at least, clearly ­recognizes this in its discussions of the second tier. It speaks of this second tier,  the “subsequent” cognition (pṛṣṭhalabdhanirvikalpakajñāna), as including, among other things, the recognition at the time of penetration that “I am penetrating (I understand),” preaching to others about what one has understood, and recognizing that all dharmas have the same nature of emptiness (Lamotte 1938, 8.19Bh, 341). Clearly, this cognition involves complex conceptual activity of the kinds involved in reflective awareness, intentional action, 18

Quoted in Farah 2004, 70.

78

Spackman

and deliberation. But this use of concepts is also said to be transformed by the ­influence of the fundmental nonconceptual cognition that produces it: “Subsequent nonconceptual knowledge is free of blemishes (anupalipta) like space” (Lamotte, 8.14, p. 332). In some sense, then, the use of concepts within subsequent awareness must be nondelusory. Presumably, what brings about this transformation is that by virtue of the insight into suchness attained in fundamental nonconceptual cognition – that all dharmas are empty of intrinsic nature – the subsequent cognition, despite its employment of concepts, recognizes all dharmas to be empty.19 Assuming that what is recognized to be empty here is both apparent objects in the world and the subject who perceives them, this recognition of emptiness can be seen as fundamentally equivalent to the notion of nonduality which is emphasized in both the Mahāyānasaṃgraha and the Mahāyānasūtrālaṃkāra. If this is the case, however, it points the way toward a uniform account of what makes both tiers of cognition nondelusory, and so, toward a single-tiered model of nonconceptual cognition rather than a two-tiered one. This point can be made in relation to all of the two-tiered models noted above, though it is most easily seen, perhaps, in regard to the nonconceptual content model. This model holds that the absorbed states of buddhas and bodhisattvas are nonconceptual in that these states transcend the duality of subject and object, and so cannot be grasped by concepts, which presuppose such a duality. But if, as we have seen, subsequent cognition itself involves a recognition of the emptiness, and hence the nonduality, of the grasper and the grasped, then it would seem that subsequent cognition should count as nonconceptual in the same sense as nonconceptual cognition itself. Both tiers of awareness would be nonconceptual in the same nondual sense. The same argument can be made, a bit more indirectly, in regard to the contentless absorption model. On this model, the absorbed cognition of buddhas and bodhisattvas is nonconceptual in that it involves no appearances and no content at all. But it seems implausible that it is just the fact that these mental states lack appearances and content that makes them nondelusory. For the model allows that subsequent cognition is nondelusory, and yet it manifestly contains appearances and content. The absence of appearances in the absorbed state must play at most a contributory role to the state’s being nondelusory, perhaps for instance by allowing the practitioner to recognize that dharmas have no intrinsic reality or are nondual. But on this view, what makes both levels of ­cognition nondelusory would be the recognition of these latter facts, not the 19

For a similar account of the transformation of subsequent cognition see Tzohar (Chapter 3).

Nonconceptual cognition in Yogācāra and Madhyamaka Thought

79

absence of appearances itself. Similar points could be made about the no active thought model. The claim that both the absorbed cognition and the everyday or subsequent cognition of buddhas and bodhisattvas are nonconceptual in the same sense might seem to fly in the face of a more natural account of nonconceptual awareness, namely that absorbed cognition is nonconceptual for the simple reason that it does not involve the subject’s use of concepts, while subsequent cognition is conceptual in that it does. In Chapter 3 of this volume, for instance, Roy Tzohar adopts a view of this kind, adding that what distinguishes subsequent cognition from deluded everyday cognition is that it involves a non-reifying use of concepts, and suggesting that for this reason the model of nonconceptual cognition that I propose is best seen as a type of subsequent cognition The view I am suggesting accepts, however, that in the sense just specified, absorbed cognition is nonconceptual and subsequent cognition is conceptual; it is thus not in disagreement with such an account so far as it goes. What my view claims, though, is that there is a deeper sense in which both types of cognition should be seen as nonconceptual, in that both adopt a nondualistic attitude toward experience. There is, on this view, a deeper unity behind the apparent differences between these states in regard to whether they have content and involve the active deployment of concepts. There is good reason, I have suggested, to hold that this account applies to both states, and not merely to subsequent cognition. It might be objected that even if we accept, as I have maintained, that subsequent cognition must in some sense be nondelusory, the texts we are considering suggest that it is nondelusory in a different sense from the sense in which fundamental, absorbed cognition is nondelusory.20 On this account, at the level of ultimate truth, the employment of concepts is always delusory; fundamental cognition thus counts as nondelusory in this ultimate sense. Subsequent cognition by contrast, would be nondelusory in a merely conventional sense, in that it promotes the attainment of fundamental nonconceptual cognition by various sentient beings; but since it employs concepts, at the ultimate level it is delusory. However, I do not think that an account of this kind can give us a complete explanation of what makes subsequent cognition nondelusory. If the only sense in which subsequent cognition is nondelusory were that it can be used to promote fundamental nonconceptual cognition in others, then first, we would have no explanation for why it promotes fundamental nonconceptual 20

I am grateful to Ching Keng and Mark Siderits for pressing this objection, and to Mark Siderits for the Santa example below.

80

Spackman

­cognition. It might be said that there is simply an inexplicable causal relation between these two states. But surely the texts we have discussed point to an explanation of this relation, namely that subsequent cognition contains an insight into emptiness or nonduality gained from the subject’s prior experience of fundamental nonconceptual cognition Furthermore, if subsequent cognition were nondelusory only in that it can help promote nonconceptual cognition in others, there would be nothing precluding buddhas and bodhisattvas from being in a state of subsequent cognition while still being fundamentally deluded about the nature of concepts, and about the subjects and objects in the world they pick out. But presumably, in reentering the world of saṃsāra, buddhas and bodhisattvas do not simply reenter the delusion they once left behind. The Mahāyānasaṃgraha represents bodhisattvas as unstained by the world. Presumably, just as a mother might assure her child that Santa is coming tonight in order to promote the child’s happiness without being deluded about the existence of Santa, buddhas and bodhisattvas must be able to use concepts without being deluded by them. If this is so, however, subsequent cognition must be characterized by a nondelusory attitude toward concepts, subjects, and objects themselves. The natural suggestion is that in subsequent cognition both concepts and the subjects and objects they pick out are regarded as empty, or in a nondual way, which once again points toward a singletiered account of nonconceptual cognition. 5

The Nondual Cognition Model

I have argued that we have good reasons to adopt a single-tiered, supervenience conceptualist model of nonconceptual cognition. What should this model look like? The natural place to begin a sketch of such a model is with the idea that what makes both fundamental nonconceptual cognition and subsequent cognition nondelusory is that they involve a nondualistic attitude. To see how a nondualistic model might hold that everyday activity involving intentional thought and deliberation can be viewed as nonconceptual, we need to clarify a bit further what is meant by nondual cognition As we noted above, the Mahāyānasaṃgraha holds that nirvikalpajñāna is nondualistic in the sense that this cognition is identical with its intentional object. But what does this mean? There are two fundamentally different ways in which this nonduality might be understood. According to one view, any employment of concepts is dualistic, because any use of concepts involves a distinction between thought, or perhaps the subject of thought, and its intentional object: the very

Nonconceptual cognition in Yogācāra and Madhyamaka Thought

81

fact of a thought being directed toward an intentional object creates an illusory dualism. On this view, to say that nirvikalpajñāna is nondual would be to say that because it involves no thought, it involves no distinction at all between thought (or its subject) and its intentional object. Perhaps it is possible to make sense of this notion of nonduality in relation to experiences of complete absorption in which there is (hypothetically) no thought; but it is harder to make sense of it in relation to cognition which contains thoughts, since it is hard to conceive of a thought which is not, at least in some minimal sense, directed toward something. It is thus hard to see how, if nonduality is interpreted in this way, a nondualistic model could provide a foundation for a single-level account of nonconceptual cognition. A second way of understanding the nonduality of nonconceptual cognition is more conducive to a single-level model. On this view, what makes a state dualistic is not that it involves a distinction between thought (or its subject) and intentional object, but rather that the relevant thoughts, subjects, and intentional objects are regarded by the person in question as having an enduring, independent existence, rather than as ultimately having the same nature, emptiness. On this account, then, nondual cognition would be cognition in which the (conventional) subject regards neither itself, nor its experiences and actions, nor any of its intentional objects as independently existing entities. Nonduality in this second sense is an attitude that can be adopted, or not adopted, toward any content of experience and any mental act including thought and deliberation, even if these experiences and acts involve concepts. It can thus form the basis for a single-tiered model of nonconceptual cognition Consider the contents of experience. It seems perfectly possible that a subject could adopt a nondualistic attitude toward any experiential content, even if that content is not independent of the subject’s conceptual abilities, as I have argued is always the case. Even if my perception of a rectangle, for instance, is dependent on my possession of the concept of a rectangle, this is no obstacle to my viewing that content nondually in this sense, since the fact that this content is supervenience conceptual does not entail that I must regard the rectangle as an independently existing object related to an independently existing subject. A similar point can be made in regard to mental acts like intentional thought and deliberation. Suppose I am asked to decide whether George is trustworthy enough to be given a certain job, and after some reflection and deliberation, I decide that he is. My reflection and deliberation are of course conceptual activities, but it certainly seems open to me to regard them nondually in the sense discussed, that is, not to regard them as acts performed by me as an enduring, independent self.

82

Spackman

Even if it is accepted that in both absorbed cognition and subsequent cognition the subject adopts a nondualistic attitude toward their experience, why should this attitude be regarded as rendering their cognition nonconceptual? On the first understanding of nonduality noted above, there is a clear connection between nonduality and nonconceptuality, since any employment of concepts is by its nature dualistic. It might be less clear what this connection is on the second understanding of nonduality, for which even mental acts and contents that involve concepts may be nondualistic, as long as the subject does not construe these acts and contents as enduring, independently existing entities. But on this account, nonduality and nonceptuality are still closely linked. Whatever their other differences on the nature of concepts and nonconceptuality, all the texts we have considered would agree that any use of a concept brings with it a powerful disposition to regard both the intentional object of that use, and the subject who uses it, as enduring, independently existing entities. On the account I have suggested, nirvikalpajñāna is nondual precisely in that it does not succumb to this disposition. It resists the inclination, deeply rooted in our conceptual capacities, to view subjects and objects as enduring, independently existing entities.21 In short, on this account, nondual cognition is nonconceptual in that it resists the pull toward dualism that makes concepts delusory. There are, no doubt, a number of concerns that might be raised about the nondual cognition model. First, it might have been noticed that in one way or another, each of the models considered above presents nirvikalpajñāna as in some respect a passive, or receptive, or effortless state. This is most explicit perhaps in Kamalaśīla’s no active thought model, which views nirvikalpajñāna as nonconceptual in the sense that it does not involve active thought, even though the concepts possessed by the subject are drawn on passively in order to constitute the content of the cognition. It might thus be argued that the nondual cognition model is not consistent with the spirit of nirvikalpajñāna, since on this view nonconceptual cognition may involve intentional thought and deliberation, and it is hard to see how these could be regarded as receptive states rather than as the acts of a subject.

21

Though it might be possible to characterize nirvikalpajñāna as not adopting the belief that subjects and objects are enduring, independent entities, I do not mean to suggest that extirpating this belief is as easy as saying, or thinking, that one does so. The “I-sense,” from the Buddhist perspective, is notoriously difficult to overcome. One way of accommodating the difficulty in overcoming such beliefs would be to adopt a dispositional account of them – to view them as consisting in deep-seated, emotionally laden dispositions to reify oneself and objects.

Nonconceptual cognition in Yogācāra and Madhyamaka Thought

83

In the Buddhist context, however, there is a straightforward sense in which even these mental acts should be regarded as receptive; indeed I view the idea that nirvikalpajñāna is in a certain sense receptive as a corollary of the claim that it is nondualistic. From the Buddhist perspective, what gives rise even to intentional thoughts and deliberations is not an independently existing self, but rather simply previous events within the five skandhas, previous thoughts, desires, bodily states, and so on. To return to the earlier example, the account of what gives rise to my deliberation about George’s trustworthiness need not include any reference to an independently existing self, but only to such things as my intention to deliberate, my recollections of George, and so on, which have a conventional reality quite apart from the illusion of an independent self. On the Buddhist account, however, we ordinarily tend to regard our ­actions as in part the product of just such an independently existing self, as actions that we, as independent beings, undertake. In contrast with these ordinary states of  mind, nonconceptual cognition involves the recognition that even intentional thought and deliberation do not originate in such an independent self, but simply arise from previous events in the five skandhas, and it is in this sense that it will be experienced as a receptive state. Notice that the content  of  this  receptive attitude, what it receives, will thus be the products of pratītyasamutpāda, dependent arising, the conventionally real causal pattern of events in the five skandhas. A related, but broader, objection to the nondual cognition account is that if we allow that nonconceptual cognition may include intentional thought and deliberation, it is hard to see what difference there is between nonconceptual cognition and ordinary, deluded states of mind. Given what has been said, however, it should be clear that there is all the difference in the world. Our ordinary experience is pervaded by dualistic attitudes. On the view I have proposed, to have such an attitude is not simply to engage in conceptualization, tout court, but it can be viewed as getting caught up in thought, deliberation, or imagination beyond what is called for by one’s circumstances. By contrast, we can think of nonconceptual cognition, given its receptive nature, as a state in which the practitioner is simply aware, nondually, of whatever arises through pratītyasamutpāda in her particular circumstances. If what is called for by the circumstances is thought and deliberation, she thinks and deliberates attentively. But if she is meditating, for instance, or eating, or playing tennis, often no thought will be called for, and in such circumstances, she is simply absorbed in her activity and whatever is happening in her environment, without the distraction of unnecessary thoughts. On the present model nonconceptual cognition is best construed as a kind of absorbed cognition, cognition that is absorbed in experience and activity.

84

Spackman

A final challenge focuses on the notion of nonduality in the proposed model. It might be argued that this model offers an implausible account of subjective experience and agency as wholly devoid of any sense of “I” or “mineness.” Within the phenomenological tradition, it has sometimes been suggested that there is an ineliminable sense of “mineness” that accompanies each of our experiences and actions at a pre-reflective level, prior to the reflective differentiation of self and other – what Dan Zahavi calls the “first-personal self-­givenness” of experience (Zahavi 2011, 59). Indeed it might be argued that if this sense of mineness is missing from an experience, the result can be various forms of pathology. For instance, one characteristic of schizophrenia is the experience of thought insertion, in which a subject believes that certain thoughts are not her own but belong to someone else. Likewise, in cases of somatoparaphrenia, subjects claim that a part of their body does not belong to them but is someone else’s.22 It might thus be argued, contra the nondualistic cognition model, that a mode of cognition in which aspects of subjective experience are not regarded as “mine,” far from being a valuable and nondelusive perception of reality, is in fact pathological, and a potential source of suffering. In response, I would make two points. First, it seems quite possible for the nondual cognition model to acknowledge the importance and value of ordinary experiences of subjectivity and agency. It has recently been argued, in response to Zahavi’s claim that the Buddhist doctrine of “no self” cannot account for the subjectivity or “mineness” of experience, that the Buddhist view can in fact accommodate at least a perspectival sense of subjectivity, since all that is denied by this doctrine is that there is a self in the sense of an enduring entity that is ontologically distinct from the stream of experience (e.g. Dreyfus 2011, Krueger 2011). Similarly, what the nondual cognition model views as characteristic of nonconceptual cognition is not that it lacks a sense of subjectivity and agency entirely, but just that this subjectivity and agency are not regarded as belonging to an enduring, independently existing self. I would also point out that the examples of pathology just mentioned are not in fact cases in which the view that one’s experience belongs to an independently existing self has been overcome. The experience of inserted thoughts in schizophrenia, for example, is one in which the thoughts in question are viewed as not mine in

22

For a fascinating discussion of the implications of these and other cases for our understanding of subjectivity, see Gallagher 2012.

Nonconceptual cognition in Yogācāra and Madhyamaka Thought

85

contrast with those aspects of experience which are mine. These pathological cases are thus quite different in kind from the nondual cognition envisioned by the model. 6 Conclusion If the nondual cognition model that I have suggested here is right, it tells us something about the conceptual status of states that might be promoted in advanced Buddhist practitioners by meditation and other forms of practice. There is clearly one sense in which states of meditative absorption might be seen as nonconceptual, while the everyday mental states of advanced practitioners are conceptual. But at a deeper level, there is a sense in which both kinds of state should be seen as nonconceptual, not in that they are wholly devoid of intentional thought and deliberation, nor in that their contents are wholly independent of concepts, but rather because of the subject’s nondualistic attitude toward these mental acts and contents. Even states that involve thought and deliberation, and states whose content supervenes on the subject’s concepts, can be nonconceptual in this sense. As I noted at the outset, this account of nonconceptual cognition may also have implications for empirical studies of meditation in psychology and neuroscience. There are important questions about what exactly is revealed by studies of neural correlates of meditative experience, but at any rate if the proposed model is on the right track, it would be wrong to assume that, in order for a state to count as nonconceptual in the relevant sense, the neural activity with which it is correlated must be completely independent of areas of the brain implicated in conceptual processing. Works Cited Arnold, Dan (2003). “Verses on Nonconceptual Awareness: A Close Reading of Mahāyānasaṃgraha 8.2-13,” Indian International Journal of Buddhist Studies 4: 9–49. Bermudez, J.L. (2007). “What Is At Stake in the Debate Over Nonconceptual Content?” Philosophical Perspectives, vol. 21, pp. 55–72. Brewer, Bill (1999). Perception and Reason. Oxford: Clarendon Press. Chuard, P. (2006). “Demonstrative Concepts Without Reidentification,” Philosophical Studies, vol. 130, pp. 153–201.

86

Spackman

D’Amato, Mario (2009). “Why the Buddha Never Uttered a Word.” In Pointing at the Moon: Buddhism, Logic, Analytic Philosophy, ed. Mario D’Amato, Jay L. Garfield, and Tom J.F. Tillemans. Oxford: Oxford University Press. Demieville, Paul (1952). Le Concile de Lhasa: Un Controverse Sur le Quiétisme Entre Bouddhistes de l’Inde et de la Chine au viiie Siècle de l’Ère Chrétienne. Paris: Imprimerie Nationale de France. Dokic, J., and Pacherie, E. (2001). “Shades and Concepts,” Analysis, vol. 61, pp. 193–202. Dretske, F. (1999). “Machines, Plants, and Animals: The Origins of Agency,” Erkenntnis, vol. 51, pp. 19–31. Dreyfus, Georges (2011). “Self and Subjectivity: A Middle-Way Approach.” In Self, No Self? Perspectives from Analytical, Phenomenological, and Indian Traditions, ed. Mark Siderits, Evan Thompson, and Dan Zahavi. Oxford: Oxford University Press. Dunne, John D. (1996). “Thoughtless Buddha, Passionate Buddha,” Journal of the American Academy of Religion, 64:3, pp. 525–556. Dunne, John D. (2004). Foundations of Dharmakirti’s Philosophy. Boston: Wisdom. Evans, G. (1982). The Varieties of Reference, ed. John McDowell. Oxford: Oxford University Press. Farah, M. (2004). Visual Agnosia. Cambridge, MA: mit Press. Gallagher, Sean (2012). “First-Person Perspective and Immunity to Error Through Misidentification.” In Consciousness and Subjectivity, ed. Sofia Miguens and Gerhard Preyer. Frankfurt: Ontos Verlag. Garfield, Jay (1995). The Fundamental Wisdom of the Middle Way: Nāgārjuna’s Mūlamadhyamakakārikā. New York and Oxford: Oxford University Press. Griffiths, Paul (1990). “Pure Consciousness and Indian Buddhism.” In The Problem of Pure Consciousness: Mysticism and Philosophy, ed. Robert K.C. Forman, pp. 71–97. Oxford: Oxford University Press. Hanna, R. (2008). “Kantian Nonconceptualism,” Philosophical Studies, vol. 137, pp. 41–64. Heck, R. (2000). “Nonconceptual Content and the ‘Space of Reasons,’” Philosophical Review, vol. 109, pp. 483–523. Heck, R. (2007). “Are There Different Kinds of Content?” In Contemporary Debates in Philosophy of Mind, ed. B.P. McLaughlin and J. Cohen, pp. 117–38. Oxford: Blackwell Publishing. Jha, A.P., E.A. Stanley, A. Kiyonaga, L. Wong, L. Gelfand (2010). “Examining the Protective Effects of Mindfulness Training on Working Memory Capacity and Affective Experience,” Emotion 10:1, pp. 54–64. Kelly, Sean (2001). “Demonstrative Concepts and Experience,” Philosophical Review, vol. 130, pp. 397–420.

Nonconceptual cognition in Yogācāra and Madhyamaka Thought

87

Krueger, Joel (2011). “The Who and How of Experience.” In Self, No Self? Perspectives from Analytical, Phenomenological, and Indian Traditions, ed. Mark Siderits, Evan Thompson, and Dan Zahavi. Oxford: Oxford University Press. Lamotte, Étienne (1938). La Somme du Grand Véhicule d’Asaṅga (Mahāyānasaṃgraha). Vols. i–ii. Louvain-la-Neuve: Université de Louvain. Trans. from the French by Gelongma Karma Migme Chodron. Retrieved from https://www.scribd.com/ document/245933020/Asaṅga-Mahayanasamgraha-tr-Lamotte-English-FromFrench-Ed. Martin, M. (1992). “Perception, Concepts, and Memory,” Philosophical Review, vol. 101, pp. 745–764. McDowell, J. (1994). Mind and World. Cambridge, MA: Harvard University Press. Millikan, R. (1998). “A Common Structure for Concepts of Individuals, Stuffs, and Real Kinds: More Mama, More Milk, and More Mouse,” Behavioral and Brain Sciences, vol. 21, pp. 55–65. Mipham, Jamgön, and Candrakirti (2002). Introduction to the Middle Way: Chandrakirti’s Madhyamakavatara, with Commentary by Jamgön Mipham. Translated by the Padmakara Translation Group. Boston: Shambhala. Morrison, A.B., Goolsarran, M., Rogers, S.L., & Jha, A.P. (2013). Taming a wandering attention: short-form mindfulness training in student cohorts. Frontiers in Human Neuroscience, 7, 897. http://doi.org/10.3389/fnhum.2013.00897. Peacocke, C. (1992). A Study of Concepts. Cambridge, MA: mit Press. Pradhan, Prahlad (ed.) (1975). Abhidharmakośabhāṣyam of Vasubandhu. Patna: K.P. Jayaswal Research Institute. Raftopoulos, A., and V.R. Müller (2006). “The Phenomenal Content of Experience,” Mind and Language, vol. 21, pp. 187–219. Siderits, Mark (2011). “Buddhas as Zombies: A Buddhist Reduction of Subjectivity.” In Self, No Self? Perspectives from Analytical, Phenomenological, and Indian Traditions, ed. Mark Siderits, Evan Thompson, and Dan Zahavi. Oxford: Oxford University Press. Siderits, Mark (2015). “Comparison or Confluence Philosophy?” In The Oxford Handbook of Indian Philosophy, ed. Jonardon Ganeri. doi: 10.1093/oxfordhb/9780199314621 .013.5. Spackman, John (2014). “Between Nihilism and Anti-Essentialism: A Conceptualist Interpretation of Nāgārjuna,” Philosophy East and West 64:1, pp. 151–173. Spackman, John (manuscript). “Conceptualism and the Richness of Perceptual Experience.” Tye, M. (2006). “Nonconceptual Content, Richness, and Fineness of Grain,” in Perceptual Experience, ed. T.S. Gendler and J. Hawthorne. Oxford: Oxford University Press.

88

Spackman

Williams, Paul (1992). “Non-Conceptuality, Critical Reasoning and Religious Experience: Some Tibetan Buddhist Reflections,” Royal Institute of Philosophy Supplement 32: pp. 189–210. Zahavi, Dan (2011). “The Experiential Self: Objections and Clarifications.” In Self, No Self? Perspectives from Analytical, Phenomenological, and Indian Traditions, ed. Mark Siderits, Evan Thompson, and Dan Zahavi. Oxford: Oxford University Press.

Chapter 3

Turning Earth to Gold: The Early Yogācāra Understanding of Experience Following Nonconceptual Cognition Roy Tzohar Abstract According to the early Indian Yogācāra, following the attainment of non-­conceptual cognition (nirvikalpajñāna) the Bodhisattva attains another kind of insight – the “cognition obtained subsequent to it” (tatpṛṣṭhalabdhanirvikalpajṇāna). This cognition appears to involve a unique kind of conceptual activity, very different from the ordinary one, which allows one to remain and operate effectively within Samsara. In the Triṃśikā-bhāṣya Sthiramati correlates this cognition to the understanding of the Dependent nature as the causal interconnectedness of all essenceless phenomena; and the Mahāyānasaṃgraha and its commentaries present intriguing descriptions of what such knowledge entails in terms of the Yogic perception of the external world. Examining these texts as well as passages from other Yogācāra treatises, the paper will unpack the phenomenological and conceptual aspects of the School’s conception of the “cognition obtained subsequent to it,” exploring its relevancy to contemporary philosophical discussion of qualia; its implications for the adequacy of the shared/private distinction with respect to experience and for the understanding of perception given the possibility of non-conceptual experiences.

Keywords Yogācāra – non-conceptual experience – conceptualism – skillful action – bodhisattva – nirvikalpa – tatpṛṣṭha-labdha- jñāna – McDowell-­Dreyfus debate

...

According to the early Indian Yogācāra, after the attainment of non-­conceptual cognition (nirvikalpajñāna) the advanced bodhisattva attains another kind of insight – the “cognition obtained subsequent to it” (tatpṛṣṭhalabdhajñāna). This cognition appears to involve a unique kind of conceptual (!) activity, very

© koninklijke brill nv, leiden, ���1 | doi:10.1163/9789004440913_006

90

Tzohar

different from ordinary conceptual activity, and it allows her to remain, operate and communicate effectively within saṃsāra. Furthermore, Yogācāra thinkers appear to identify this cognition as the state in which ultimate knowledge of causality is attained. In the Triṃśikā-bhāṣya (TriṃśBh), for instance, as we will soon see, Sthiramati correlates this cognition with the understanding of the “dependent nature” (paratantrasvabhāva) as the causal interconnectedness of all essenceless phenomena; and the Mahāyānasaṃgraha (ms) and its commentaries present intriguing descriptions of what such knowledge entails in terms of the yogic perception of the external world. This essay draws on various Yogācāra treatises to explicate the school’s understanding of this type of cognition, chiefly to account for its unique relation to non-conceptual cognition.1 Based on this analysis I argue that while nonconceptual states do not allow for any intentional or verbal and cognitive content, for the Yogācāra this does not undermine the relevance of such states to philosophical inquiry, but rather invites us to rethink the nature and operation of ordinary language and perception. For the Yogācāra, therefore, non-conceptual experiences, rather than serving merely to mark the limits of language and promote an ineffable ontology, serve to constitute and inform us of the way in which an enlightened mind abides in phenomena.

...

The issue of the cognition subsequent to non-conceptual cognition is taken up by Sthiramati in several places in his commentaries, of which the most pertinent for present purposes is his commentary on verse 22 of Vasubandhu’s Triṃśikā (Triṃś) dealing with the relations between the dependent and the perfected natures:2 Hence, this [the perfected nature] is neither different nor the same as the dependent nature. It should be expounded as impermanence and so on. When this [the perfected] is not seen that [the dependent] is not seen.3 (Triṃś 22)

1 In presenting the Yogācāra notion of subsequent cognition as well as the school’s account of the role of vikalpa in the construction of the life world, I closely follow Tzohar (2018, Chapter 6, Sections 6.1 and 6.3 respectively). 2 For a more detailed account of Sthiramati’s understanding of the relations between the three natures, see Tzohar (2018, Chapter 6). 3 ata eva sa naivānyo nānanyaḥ paratantrataḥ | anityatādivad vācyo nādṛṣṭe ’smin sa dṛśyate || Triṃś 22,( Buescher 2007, 124, 126).

Turning Earth to Gold

91

Our main concern is with the latter part of the verse, which Sthiramati understands as responding to an implicit objection that points out the difficulty of knowing the dependent as such: If the [perfected nature is] the dependent nature free from the graspergrasped relation, how is it [itself] comprehended? Or, not being grasped, how it is known to be existent? For this reason [Vasubandhu] stated: “when this is not seen that is not seen.”4 According to Sthiramati, the difficulty is that if indeed the perfected is the ­dependent free of the imagined (and hence of all duality of subject and object), in what sense can it still be said to be an object of knowledge? And if one assumes that it is not cognized, then in what sense can it be said to exist at all? Sthiramati proposes to read verse 22d as implicitly offering a division of labor b­ etween two types of cognition, logically and chronologically ordered. He explains that insofar as it is free of duality, the dependent as the perfected by definition can only be cognized through transcendent cognition that is free of conceptual d­ ifferentiation (nirvikalpalokottarajñāna, henceforth “non-­conceptual cognition”), and in this sense the dependent is indeed not an ­object of knowledge. However, this kind of cognition is understood as a pre-condition of a worldly pure cognition subsequent to it (tatpṛṣṭhalabdhaś uddhalaukikajñāna, ­henceforth “subsequent cognition”) which does allow us to conceive of the dependent as an object of knowledge. But in what sense? Drawing on scripture,5 Sthiramati points out that through this cognition all things are seen to be non-existent like a magical creation, emphasizing on the one hand that the objects of this cognition are multifarious and distinct but on the other hand essenceless insofar as they are the outcome of causes and conditions.6 4 yadi grāhyagrāhakabhāvarahitaḥ paratantraḥ katham asau gṛhyate agṛhyamāṇo vā katham astīti vijñāyate | ata āha | nādṛṣṭe ’smin sa dṛśyate || TriṃśBh 22d (Buescher 2007, 126 line 7–9). 5 On the Nirvikalpapraveśadhāraṇī/Avikalpapraveśadhāraṇī, see Matsuda (1996, 170). 6 This in contrast to the uniform experience of the perfected under non-conceptual cognition, which is said to be only of “one taste” like space: nādṛṣṭe ’sminn iti pariniṣpannasvabhāve sa dṛśyata iti paratantraḥ svabhāvaḥ | nirvikalpalokottarajñānadṛśye pariniṣpanne svabhāve adṛṣṭe apratividdhe asākṣātkṛte tatpṛṣṭhalabdhaśuddhalaukikajñānagamyatvāt paratantro ’nyena jñānena na gṛhyate | ataḥ pariniṣpanne ’dṛṣṭe paratantro na dṛśyate | na punar lokottarajñānapṛṣṭhalabdhenāpi jñānena na dṛśyate |tathā nirvikalpapraveśāyāṃ dhāraṇyām uktam | tatpṛṣṭhalabdhena jñānena māyāmarīcisvapnapratiśrutkodakacandranirmitasamān sarvadharmān pratyetīti | atra ca dharmāḥ paratantrasaṃgṛhītā abhipretāḥ | pariniṣpannaś

92

Tzohar

To figure out exactly what Sthiramti’s understanding of “subsequent cognition” entails in terms of knowledge of causality, we need to first consider the question of its relation to non-conceptual cognition. On the face of it, Sthiramati’s characterization of these two types of cognition as “mundane” and “transcendent” respectively suggests that they are distinctly different from each other.7 However, this seemingly clear-cut distinction is complicated by other Abhi­ dharma and Yogācāra sources, such as Vasubanhdu’s Abhidharmakośabhāṣya (AKBh), which groups both types of cognition under the category of the absolute truth outlook on phenomena (Dhammajoti, 2007, 80), or the ms, which gives a rather paradoxical characterization of the subsequent cognition as both free of and engaging with conceptualization (discussed in detail below). Furthermore, Sthiramati himself, in the Madhyāntavibhāgaṭīkā, presents a division between mundane and supermundane knowledge that does not at all align itself neatly with the two types of cognition.8 This apparent lack of conceptual homogeneity, which in itself is unsurprising considering that these texts were composed in different times and mark different theoretical stages, can be resolved once the nature of the relation between these two cognitions is examined more closely. For this purpose we need to go back and consider the way in which the understanding of the “subsequent cognition,” as a knowledge of causality above all else, is treated in Yogācāra sources other than Sthiramati’s TriṃśBh. Here we turn to the ms, which addresses this issue in several different chapters. Consider, for instance, the following reference to the subsequent cognition in the third chapter of the work:9 cākāśavad ekarasaḥ | jñānañ ca yathoktaṃ nirvikalpena jñānenākāśasamatayā tān sarvadharmān paśyatīti paratantradharmāṇāṃ tathatāmātradarśanāt || TriṃśBh on 22d, Buescher (2007, 126 line 10–22). 7 The view of these as distinctly different from each other is endorsed also by Arnold’s analysis (2003) of non-conceptual cognition in the eighth chapter of the ms and in Vasubandhu’s Viṃśikāvṛtti on 17c-d: “Vasubandhu’s characterization of nirvikalpajñāna and tatpṛṣṭhalabdhajñāna as, respectively, lokottara (“transcendent”) and laukika (“mundane”) helps make clear the essential philosophical problematic at issue in all of this; for these very characterizations seem to have it that the two kinds of awareness are radically incommensurable, in which case it is unclear how they could ever be related to one another…” (Ibid, 31–32 n.54). 8 “It is said by others that the natures are seen as threefold in order to demonstrate that they are the objects of mundane cognition, supramundane cognition and the direct intuition attained subsequently to it, respectively.” Iti laukikalokottaratatpṛṣṭhalabdhajñānaviṣayatvapradarśanārthaṃ yathāsambhavaṃ trayopādānaṃ svabhāvatvam ityanye. Madhyāntavibhāgaṭīkā on 3.3.a, Yamaguchi and Lévi (1934, 112, lines 12–14). 9 For the ms I am using Lamotte’s 1973 edition and translation. In this presentation, if not otherwise indicated, translations from the latter (vol. 2) into English are by Gelongma Karma Migme Chödrön (2012).

Turning Earth to Gold

93

…This subsequent cognition (pṛṣṭhalabdhajñāna), which considers every creation (prabhava) coming from the store-consciousness and every object of concept (vijñaptinimitta) as a magic show (māyā), etc., is, in its essence, free of errors (prakṛtyaviparīta). Thus, in the same way that the magician (māyākāra) is free of doubt about the things produced by magic (māyākṛtadharma), so this bodhisattva is always unmistaken (viparyāsa) when he speaks of cause (hetu) and result (phala). (ms iii.12, Lamotte (1973, vol. 1, 53–54, vol. 2, 168–169); and Chödrön (2012, 226–227)) This is further clarified by the Upanibandhana (U)10 commentary on this verse, which serves also to illuminate the function and purpose of the subsequent knowledge: This subsequent cognition, etc.: the author explains the rôle of subsequent cognition. Every creation coming from the store-consciousness: in speaking of creation, the author mentions the cause. In speaking of every object of concept, he mentions result. It is a matter of grāhaka and grāhya. The subsequent cognition that takes them as a magical show, etc., is free of error by essence, because by considering, in accordance with the truth, the ­dependent nature as a magical show, etc., one avoids error (cf. Chap. ii, § 27). In the same way that a magician is free of all error about magical things: the magician is free of error about the plants (tṛṇa), trees (vṛkṣa), etc., that are the causes of magic, because he sees them correctly; the magician is free of error about the elephant (hastin), the horse (aśva), etc., because he sees them correctly. In the same way the bodhisattva who has seen the truth (dṛṣṭasatya) correctly sees the non-existence of grāhya and grāhaka. When he attains the truth, he produces this subsequent cognition (prṣṭhalabdha), generator of speech (vāksamutthāpaka), of the conventional order (saṃvṛti), but pure. At the time when he knows cause and result and when he preaches the law, he is free of error. His listeners (śrāvaka) are subject to error, but when the propensity of the teachings (śrutavāsanā) has perfumed their mental series (prabandha), they are freed of error progressively (anukrameṇa) and by degrees (śanais) and thereby accomplish what they had to do (kṛtyānuṣṭhāna). This subsequent cognition is likewise free of concepts (nirvikalpa) and is unstained (akliṣṭa). (U. on ms iii.12, Lamotte (1973, vol. 2, 168–169); Chödrön (2012, 227–228))

10

Ascribed to Asvabhāva. “U” stands for Lamotte’s translation from the Chinese translation by Hiuan-tsang (Taishô 1598). See Lamotte (1973, vol. 1, vi).

94

Tzohar

This ms passage and its commentary raise several noteworthy points. Again, we see that non-conceptual cognition is not the end point of the spiritual development of the bodhisattva, but a mediatory stage, a precondition for knowledge of phenomena in causal terms and as essenceless and non-dual. As for the purpose of this subsequent cognition, over and above the function of nonconceptual cognition, the commentary makes clear that its role has to do with maintaining the efficacy of the bodhisattva’s salvific activity within saṃsāra. For the bodhisattva, the perspective of subsequent cognition seems to establish the outlines of a conventional and communicative order in which she operates, which however – in contrast to the perception of ordinary beings— is pure, unerring with respect to causality, free of defilement, and also – and this last point is rather confusing – free of concepts (nirvikalpa). What are we to make of the latter statement? In what sense is the subsequent cognition nonconceptual, and how is this claim compatible with the characterization of this cognition as the “generator of speech” etc.? A promising place to seek an explanation is the eighth chapter of the ms, which as part of its focus on non-conceptual cognition, discusses the similarities and differences between this type of cognition and cognition subsequent to it. There, in verse 14 (Lamotte 1973, vol. 1, 77), non-conceptual cognition is said to consist of three logically and chronologically consecutive stages – ­preparatory non-conceptual cognition (prāyogikanirvikalpakajñāna), fundamental non-conceptual cognition (mūlanirvikalpakajñāna), and subsequent non-conceptual cognition (pṛṣṭhalabdhanirvikalpakajñāna). Vasubandhu’s Bhāṣya (Bh)11 on the same verse (Lamotte 1973: vol. 2, 242–243) provides an explanation of the differences between these three kinds, clarifying that only the second stands for “proper” non-conceptual experience, while the first and third kinds are termed so only by proxy, insofar as they are a preliminary condition and the outcome of such an experience, respectively. This seems to address the difficulty presented above by proposing that the subsequent cognition is non-conceptual only by name (or, the commentator may add, by way of a metaphor due to proximity) and in that sense it constitutes a kind of knowledge that is altogether different from – and perhaps even incommensurable with – the fundamental non-conceptual cognition. This solution seems to be in line with other sections of the same chapter that explicitly take the subsequent, so-called “non-conceptual” cognition to serve as the object of conceptualization. For instance, the Bh on ms viii.16 (Lamotte 1973, vol. 2, 245), provides an account of how the fundamental and subsequent non-conceptual cognition differ in terms of the conceptual activity 11

Bh refers to Lamotte’s translation of the Chinese translation by Hiuan-tsang (Taishô 1597).

Turning Earth to Gold

95

and content they entail. According to the Bh, the fundamental non-conceptual cognition is of “one taste,” like space, and is characterized as both “not conceptualizing and non-conceptual,” unpacked respectively as being itself “without concept (vikalpa),” and as “not conceived of by others,” i.e. as necessarily private. The subsequent cognition on the other hand is said to be marked by both conceptualizing objects and being itself the object of conceptualization. While there appears to be no way around the fact that subsequent cognition involves conceptualization, still, a close reading of the latter half of the eighth chapter indicates that this conceptualization is of a different kind, a conceptualization that is tightly connected to a fresh understanding of causality and that presents a more nuanced and intricate account of the way in which these types of cognition arise as an extension of – rather than in contrast to – one another. In what follows I would like to look briefly at how the subsequent cognition is further characterized in the eighth chapter of the ms. A good place to begin is by noting the way in which the differences between these types of cognition are accounted for in terms of their different intended functions. ms viii.15 presents a set of intriguing analogies that deal with this topic: 15 (a) The three knowledges (jñāna) are like a mute person (mūka) trying to understand an object (artha), a mute person who has understood the object, a non-mute person who has understood the object. (b) The three knowledges are like an idiot (mūḍha) trying to understand an object, an idiot who has understood the object, an intelligent person (amūḍha) who has understood the object …. (Lamotte (1973, vol. 1 77, vol. 2 243–243); Chödrön (2012, 333)) The verse clearly posits the subsequent cognition at the apex of this tripartite division. This fact, together with the use of rather unflattering analogies for fundamental non-conceptual cognition – chosen perhaps with the intention of downplaying these experiences as the end-point of the bodhisattva’s path – seems to emphasize the importance of the subsequent cognition in the salvific ideology of the Mahāyāna. Non-conceptual experience, these analogies seem to tell us, is inaccessible – like a mute person who cannot communicate what he knows, or a fool who fails to make sense of what he knows – if it is not eventually used for the salvific activity of the bodhisattva, which requires verbal communication and interaction. It should be noted, however, that the fundamental non-conceptual cognition, while not intentional, is not altogether ineffective. Rather, its efficacy – and in particular the kind of agency it involves – is of a different kind. This issue is discussed in the ms viii.17 and its commentaries:

96

Tzohar

17. Just as precious gems (maṇi) and celestial music (tūrya) accomplish their own activity (svakarman) without thought, so it is always without thought that the Buddhas accomplish their kinds of activity (nānākarman). Lamotte (1973a, vol. 1, 78); Chödrön (2012, 336) [Bhāṣya:] Let us accept that this nonconceptual knowledge realizes the state of buddhahood (buddhatā); but if it is free of effort (ābhoga) and free of thought (manasikāravikalpa), how could it accomplish its activity of working for the benefit and happiness of beings (sattvahītasukhakriyā)? … Even though it is without concept, it does fulfill its rôle (kṛtyānuṣṭhāna). In the stanza, the examples (dṛṣtānta) of the precious gem (maṇi) and celestial musical instrument (tūrya) show this. A cintāmaṇi [a precious gem that satisfies all the wishes of its possessor], although without concepts (vikalpa), is able to fulfill the rôle desired by the beings who possess it. A celestial musical instrument (tūrya), even without being struck (aghaṭṭita), can produce all kinds of sounds (śabda) according to the aspirations (āśaya) of those close by. In the same way, the nonconceptual knowledge of the buddhas and bodhisattvas, although without concepts, can accomplish all kinds of rôles. (Lamotte (1973, vol. 2, 245–246); Chödrön (2012, 336–337)) According to the Bh, the issue at stake is the need to explain how abiding in a non-conceptual state is compatible with the bodhisattva working at all times for the benefit of all beings. The solution offered is that the non-conceptual state involves a unique kind of interaction, explained through several analogies that illustrate the rather nonvolitional and spontaneous nature of the bodhisattva’s activity. The U commentary states this in even stronger terms, emphasizing what the non-conceptual state entails in terms of the agency of a bodhisattva: Nonconceptual knowledge realizes the state of Buddha (buddhatā); but it is without concepts (vikalpa); how can it procure the interests of beings (sattvārtha)? … In the stanza, the author brings in the examples of the precious gem and the celestial musical instrument to prove that, although without concepts, the nonconceptual knowledge is able to accomplish all kinds of rôles effortlessly (anābhoga). Thus the cintāmaṇi and the celestial musical instrument do not say: “I am going to emit light, I am going to produce a sound”, because they are both without thought (acetana). But, by the power of the meritorious actions (puṇyakarman) and aspirations (āśaya) of beings who have been born where they occur,

Turning Earth to Gold

97

and independently of being struck (ghaṭṭhana-nirapekṣan), they emit all kinds of light (prabhā) and produce all kinds of sounds (śabda). It is the same for the nonconceptual knowledge of the buddhas and bodhisattvas: although without concepts, it fulfills its role which is to render all kinds of services (nānārthakriyā) in conformity with the merits (puṇya) and aspirations (āśaya) of beings to be converted (vineyasattva) by it. (Lamotte (1973, vol. 2, 246); Chödrön (2012, 337)) So even in non-conceptual states the bodhisattva is still active, but in the manner of a wish-fulfilling jewel – that is to say, in a non-volitional and spontaneous way. According to the passage above, such action, while efficacious, does not entail agency in the sense of deliberation or intentional cognitive content. The notion of Buddhahood that emerges from this account is very similar to what Mario D’Amato has dubbed as the model of the “mindless Buddha,” according to which, behind appearances, a Buddha is seen to be ultimately without any cognitive conceptual content and without any utterances (2009, 49–51).12 As we have seen, however, the ms (and according to D’Amato also the msa) proceeds to present yet another way of understanding the activity of the bo­ dhisattva and thus another notion of efficacy involved in the subsequent cognition. In this respect, ms viii.19 presents five fundamental “types” of subsequent cognition, which the Bh understands as knowledge claims (and hence propositional by nature) arrived at following a corresponding set of analyses (vicāra): …(iii) Subsequent intuitive knowledge (pṛṣṭhalabdhanirvikalpaka-jñāna) is of five kinds, according to whether it is analysis (vicāra) of (1) penetration (prativedha), (2) memory (anusmaraṇa), (3) preaching (vyavasthāna), (4) combined (saṃsarga), or (5) success (saṃṛddhi). (ms viii.19, Lamotte (1973a, vol. 1, 78); Chödrön (2012, 340)) [Bhāṣya:] … (iii) Subsequent knowledge is of five types: (1) analysis of penetration, (2) analysis of memory, (3) analysis of preaching, (4) analysis of synthesis, and (5) analysis of success. (1) Analysis of penetration (prativedhavicāra). At the time of penetration (prativedha), one makes the following analysis: “I am p ­ enetrating 12

D’Amato contrasts this model of the Buddha with that of the “mindful Buddha,” which is presented for instance in the msa and aligns more closely with the subsequent cognition.

98

Tzohar

(2) (3) (4)

(5)

(I understand).” Here penetration has the sense of understanding (avabodha). Analysis of memory (anusmaraṇavicāra). Emerging from (vyutthāna) this penetration, one remembers and says to oneself: “I have penetrated absence of concepts (nirvikalpa).” Analysis of preaching (vyavasthānavicāra). This is to preach to others what one has penetrated. Analysis of combination (saṃsargavicāra). The knowledge of the combined object (saṃsṛṣṭālambanajñāna) sees that all dharmas have one and the same nature (ekalakṣaṇa) and leads towards the transformation of support (āśrayaparāvṛtti). When this support has been transformed, this knowledge is reproduced. Analysis of success (samṛddhivicāra). Everything succeeds (samṛddhyate) according to aspirations (yathāśayam). By means of this analysis, earth, etc., is transformed (pariṇam-) into gold (cf. Chap. I, § 60; Chap. ii, § 14). In order to obtain this success, this analysis is made; that is why the analysis of success is spoken of. Some say that it is because of this analysis that one obtains success. (Lamotte (1973, vol. 2, 250); Chödrön (2012, 341))

It should be noted that while all these knowledge claims are arrived at as part of the subsequent cognition, they are all directed at the preceding fundamental non-conceptual knowledge. These claims include being aware (in ­retrospect) of one’s having had a non-conceptual experience, committing this cognition to memory,13 and then communicating it to others – culminating in a kind of knowledge that is said to lead to no less than the understanding of the sameness of all objects and to the transformation of the basis.14 13 14

The text does not provide an explanation of how memory becomes operative once again once one emerges from the non-conceptual experience. The notion of the “transformation of the basis,” a fundamental concept in the Yogācāra soteriology, receives a variety of doctrinal explanations in the school’s system (most of which, however, share its understanding as the culmination of a path structure, in which the model of the eight consciousnesses is overturned and replaced by gnosis and enlightenment). A concise definition of this notion is provided in the ninth chapter of the ms, in which it is explained in terms of the three natures model, as the purification of the dependent from the imagined: “(iv) The transformation (parāvṛtti) of the support consists of the expulsion (tchouan che = ldog = vivartana) of the defiled portion (saṃkleśabhāga) of the dependent nature when its antidote (pratipakṣa) arises and it is reduced (tchouan tö = gyur pa = pariṇāma) to its pure portion (vyavadānabhāga)” ms. ix.1(d), (Lamotte 1973, vol.1, 81).

Turning Earth to Gold

99

There are several noteworthy points here: first, the striking difference between this framework and the characterization of the Buddha in the fundamental non-conceptual state. The characterization of the Buddha’s actions as spontaneous and nonvolitional by nature in that state is contrasted here with a cognitive world that involves reflexivity, memory, verbal representation and explicit intention and deliberation. Second, note that all this cognitive activity nonetheless emerges from, builds upon, and in some sense is directed at manifesting the implications of the fundamental non-conceptual experience. To recapitulate so far: We saw that while non-conceptual cognition is conceived of by the Yogācāra as a core experience of the bodhisattva, its characterization as ineffable and free of constructions by definition precludes any kind of intentional mental content or any propositional knowledge and the possibility of communication. Within this framework, efficacious actions, while conducive to liberation, appears as a kind of spontaneous and non-volitional activation rather than a volitional reaction. For an enlightened bodhisattva, however, fulfilling his salvific role within saṃsāra – namely, acting and communicating effectively within the phenomenal world – requires an additional kind of subsequent cognition, which is characterized on the one hand as mundane, conceptual and intentional, and on the other hand as pure and free of constructions and concepts. This paradoxical way of characterizing the subsequent cognition, however, should be understood not as plainly incoherent but rather as emphasizing the way in which the subsequent cognition stands in relation to, indeed as an extension of, the preceding non-conceptual cognition. Within the career of the bodhisattva, the subsequent cognition appears to The U commentary clarifies that this is the conception of the dependent as being free of the imagined (and by implication, as the perfected), first and foremost by reason of the destruction of any kind of grasper-grasped duality: “[T]he transformation (parāvṛtti) concerns the dependent nature (paratantrasvabhāva); the dependent nature of the mind and of the mental factors (cittacaitta) is the basis (āśraya) of the transformation that destroys the defilements (saṃkleśa); it is also the basis of all the buddha attributes (buddhadharma) […]What is the transformation of the basis? It concerns the dependent nature. When its antidote arises, i.e., when the nonconceptual knowledge arises, the dependent nature expels its defiled part: it destroys its disturbed part (skhalitabhāga), which is formed by all the objects and subjects of consciousness (grāhyagrāhaka), and is reduced to its purity part: rejecting this imaginary nature which is formed by objects and subjects of consciousness (grāhyagrāhaka), it is reduced to its very pure part (suviśuddhabhāga), completely free of subjects and objects of consciousness (grāhyagrāhakavivarjita), cognizable by introspection (pratyātmavedya) and avoiding all empty proliferation (prapañcocchedaka)” U 434c16–435a10, Lamotte (1973, vol. 2, 260–261); Chödrön (2012, 354–355).

100

Tzohar

serve as the norm rather than the exception. It is not seen as a fall from (nonconceptual) grace but rather as the continuation and application of its insights in the phenomenal world. The knowledge of causality and the use of conceptualization that it presents are therefore not ordinary, but radically transformed. Our earlier (pp.3–6) examination of the subsequent cognition, as described by both Sthiramati and the ms and its commentaries, showed that insofar as it stands for an exhaustive perspective on causal relations, such cognition necessarily involves conceptualization, but does so without reifying either conditions or their effects, and without regarding them in terms of the grasper-grasped distinction. It is, as the texts tell us again and again, a view of the dependent as the perfected – a causal process that is however not bound to its unreal conditions or end-products. As such, it marks a realm of agency and freedom – in the sense that the bodhisattva’s perception and speech under the subsequent cognition, stemming from non-conceptual cognition in which there exists no intentional object as a locus for conceptualization, does not align with an allegedly given objective world (in the sense that it forces itself on the mind) but with her “aspirations,” understood here as reality-forming.15 The parallels drawn by the Yogācāra between the understanding of non-conceptuality and the three-natures scheme suggest an important sense in which the Yogācāra operates under a different set of presuppositions than most contemporary discussions of non-conceptuality. Plainly put, within c­ ontemporary philosophical discussion of the possibility and role of n ­ on-­conceptual states, 15

ms viii.20 makes clear that no objects or conceptualization exist in the non-conceptual cognition: “…(vi) When non-conceptual knowledge is functioning (carati), not a single object appears. Thus we know that there is no object and, the object being absent, there is no concept (vijñapti)” Lamotte, (1973, vol. 1, 79); Chödrön (2012, 342). The U commentary indicates that the bodhisattva’s ability to transform and manipulate external phenomena is quite concrete, and stems from the initial non-existence of phenomena. In this respect the bodhisattva’s powers of transformation stand as evidence of the unattainability of the existence of mind-independent external objects: “…(iv) Furthermore, this object has no reality. Why? For the bodhisattvas in possession of the masteries; for the bodhisattvas who have attained the great masteries; by the power of convinced adhesion, i.e., by the power of aspiration (āśayabala), earth is transformed into whatever substance they wish: they change (pariṇam-) earth into gold (suvarṇa), etc. And also for the ecstatics: for persons other than the bodhisattvas, e.g., for the śrāvakas in possession of dhyāna” Lamotte (1973, vol. 2, 252d); Chödrön (2012, 343). As for the term “aspiration” (āśaya) – in his analysis of the semantic and doctrinal range of the term “adhimukti” in early Buddhist, Abhidharma, and Yogācāra sources, Dhammajoti (2019) points out this body of literature’s frequent equivocation between this term and such terms as āśaya, kṣānti, etc, which inhabit a common semantic field and indicate (receptivity) an inclination or a disposition that takes an active part in conditioning our experience and perception of existents.

Turning Earth to Gold

101

proponents of non-conceptual content find themselves, at some point, needing to account for how this content interacts and is integrated with our conceptual knowledge of the world. These accounts, however diverse and intricate, typically operate under the assumption that non-conceptual content is a primitive experiential level into which second-order conceptual content is later integrated (there are of course some exceptions of theories that do not rely on this assumption).16 The Yogācāra, however, seem to operate under a different set of presuppositions about the relations that hold between non-conceptual and conceptual content.17 As we have seen, the Yogācāra sources seem to understand possible experiences under a tripartite model that consists of: (1) ordinary conceptual experience; (2) non-conceptual experience, as a distinct and independent experiential mode that, insofar as it has no content and by definition, cannot come into contact with or function as the foundation of second-order conceptual experiences; and (3) a state in which – following non-conceptual experience and under the subsequent awareness – insight regarding the causal and  interdependent nature of phenomena allows for non-reifying use of language. So, like contemporary non-conceptualists, the Yogācāra thinkers were keenly aware of the problematics of considering a certain state to be both beyond the realm of language and thought and at the same time able to present itself to knowledge and communication. Their solution to this issue, however, was not to identify non-conceptuality with a unique non-reifying way of using concepts (as Spackman seems to propose in his contribution to this volume).18 16 See Tzohar, 2020, for a discussion of the relevance of the Yogācāra to contemporary debates regarding non-conceptuality, and especially to conceptual inclusivism. 17 This is not, however, the case for all Buddhist schools. Sharf, in his contribution to this volume, points out the presence of similar assumptions in Indian late Ābhidharmikas’ debates regarding the nature and status of non-conceptual awareness, an issue that found its way into Abhidharma theories of perception via the assumption of a kind of raw sense datum as a precondition for its own objectification by reflection and conceptualization. In this regard see also Keng’s analysis, in his contribution to this volume, of the relations between mental perception and sensory perception according to Dignāga. 18 Spackman (Chapter 2) argues for a notion of supervenience non-conceptuality as the most relevant to Buddhist engagement with this issue, and looking also at the Yogācāra sources discussed here, argues for an interpretation by which non-conceptuality is understood not as non-conceptual content or no content at all but as the subjects’ “nondualistic and non-reifying attitude toward the contents of their mental states.” While Spackman’s interpretation is designed to address certain philosophical difficulties (mostly how nonconceptual states and conceptual articulation may be in contact), his solution involves lowering the bar, so to speak, on what non-conceptuality entails and conflating non-­ conceptuality and subsequent awareness. Spackman points out, as possible support for

102

Tzohar

Rather, their response was to introduce the subsequent awareness as an additional intermediate category and point of view (doing similar philosophical work as that of the dependent nature and the three-natures scheme in improving on the binary opposition underlying the doctrine of the two truths). The subsequent awareness is intermediary because it is an experience of the ramifications of non-conceptuality within ordinary existence. It is made possible by the fact that non-conceptual experience – by means not of its content (since it has none) but of its sheer occurrence and of the sudden absence of vikalpa as a reality-constructing activity and a causal force – informs and enables, both theoretically and experientially, a new mode of being in the world. This mode is characterized by comprehensive knowledge of causality (no longer hinged to a false conception of an objectively given and essential reality), and by an attendant non-reifying use of language. This outlook also presents an intriguing inversion of the typical view of how non-conceptual states stand in relation to the overall understanding of language. Insofar as non-conceptual states do not allow for any intentional or verbal cognitive content, they are necessarily private, but at the same time preclude the possibility of a private language. In this sense, rather surprisingly, the Yogācāra account presents some interesting affinities with contemporary contextualist (or constructivist) approaches to non-conceptual “mystical” experiences in the realm of religious studies (see in this respect Sharf (Chapter 1)). Like the contextualists, the Yogācāra would argue that any experience, insofar as it is content-full, must be conceptual. However, unlike the contextualist, the Yogācāra – as prescribed by its soteriology – would not see this as entailing the incoherence of non-conceptual experiences, but as allocating all accounts of such experiences to the realm of the subsequent cognition. Indeed, framed in this way, we find here another way in which the notion of the subsequent cognition is operative: it allows for the coherence of a discourse of non-conceptual experiences. Thus, under this framework, non-conceptual experiences do not represent a complete break with language; rather, through the mediation of the subsequent cognition, they serve – both in theory and in praxis – as the framework his interpretation, that both states are “nondelusory;” however, it should be noted that they are so in different senses: while non-conceptuality is nondelusory because it has no content whatsoever, subsequent awareness is nondelusory because it conceives its object in a correct manner. This interpretation, it seems to me, is incompatible with the Yogācāra tripartite framework described above. The issue is not so much whether this is what the Yogācāra account indeed said (Spackman does not claim to provide an accurate textual exegesis), but rather getting its basic premises right, and especially capturing correctly what a concept means for the school.

Turning Earth to Gold

103

against which language and perception are understood and measured for what they truly are. (In this they replace objective phenomena in filling the role of serving as the grounding for language.) Here the ineffable is not merely that which lies outside the boundaries of language, but a necessary key to understanding the true function of language. The intermediary function of the subsequent cognition, however, is not limited merely to the understanding of language, but also serves an important epistemic role in resolving one of the difficult problems with which Mahāyāna soteriology grapples – namely, how to bridge the deep epistemic divide between the bodhisattva’s outlook and the ordinary perspective so as to explain how they may successfully interact within saṃsāra. As we will see below, the subsequent cognition serves to explain the interaction between these strikingly discrepant perspectives by accounting for both in terms of common underlying causal mental phenomena. The possibility of a co-habitation of the bodhisattva’s non-conceptual outlook and the ordinary discriminative perspective is discussed, along very similar lines, in several Yogācāra sources, such as the ms and the Viniścayasaṃgrahaṇī (vs). In both sources, the issue is raised mostly in relation to the question of the status of external reality once constructive thought (vikalpa) is eliminated by the adept. In the vs, this discussion is presented against the backdrop of Asaṅga’s arguments about the possibility of knowledge of inexpressible reality, arguments that characterize such knowledge negatively as the elimination of (figurative, upacāra) designations and discursive thought, but also positively – as the object perceived by noble wisdom.19 The opponent finds it hard to see how the elimination of discursive thought and figurative speech does not necessarily entail the total annihilation of the phenomenal realm: If one says: if you abandon figurative designations [produced through] discursive thought, since they set forth existents and sign-sources [rgyu mtshan, nimitta*], then by means of any single noble wisdom, there will be the unwanted consequence of the cessation of all living beings and inanimate things which are held together by [the combination of] discursive thought and the sign-sources. Just as, for instance, in the example of a magician and [his] illusory acts.20 19 20

For a detailed presentation of Asaṅga’s full arguments regarding the inexpressibility of reality in the vs, see Tzohar (2018, Chapter 3). If not otherwise indicated, all translations from the vs are my own and reproduced from that work. gal te rgyu mtshan gyi dngos po rnam par gzhag pa’i phyir rnam par rtog pa’i nye bar ’dogs pa spong ba yin na/ de lta na ’phags pa’i ye shes gang yang rung ba gcig gis rgyu mtshan dang rnam par rtog pas bsdus pa’i sems can dang/_sems can du bgrang ba ma yin pa’i

104

Tzohar

Underlying this passage is the presupposition that the phenomenal world in its entirety is dependent on discursive thought and verbal activity. The opponent points out that the elimination of discursive thought and subsequently of all phenomena is unthinkable for a bodhisattva, who is supposed to remain in saṃsāra and attend to the well-being of others. Asaṅga’s reply appeals to the Yogācāra understanding of the mental-causal grounding of language and conceptualization, and its role in the construction of the intersubjective experience of the lifeworld: [To this] we reply: since an entity arises from an unshared (thun mong ma yin pa, asādhārana) cause [through] discursive thought,21 and from a shared cause [through] discursive thought, [therefore] any imagined22 thing that arises from an unshared cause through discursive thought will cease, [while] anything that is not imagined [due to noble wisdom] and arises from a shared cause through discursive thought will not cease, because it is grasped by the discursive thought of others. If it were not so, [the existence of] the discursive thought of others would be pointless. And even though it [that which arises from a shared cause] does not cease, it should be understood as the purified vision of one whose nature is completely pure, just like, for example, in the case of many yogis [who] through knowledge [that should be born] of equipoise with respect to one object, orienting [their attention]23 to a variety of things, will apprehend a variety of visions.24 d­ ngos po thams cad ’gog par thal bar ’gyur te/ ’di lta ste/dper na sgyu ma mkhan dang sgyu ma’i las bzhin no zhe na/ vs td. 4038 Zi 13a1–2. 21 In accordance with other Yogācāra discussions of this issue, I understand “un-shared” (thun mong ma yin pa) as qualifying the “cause” (rgyu) rather than “discursive thought” (rnam par rtog pa), thereby adding an implicit instrumental case ending to the latter. The alternative interpretation of this compound would read: “(a thing) … which has unshared discursive thought as its cause.” 22 Rnam par brtags pa. A negation is added both in the Peking and the Narthang editions of the text, rendering it “un-imagined” (rnam par ma brtags pa), which however seems incompatible with the overall argument. Bstan ’Gyur 1994: v. 74 zi 1048. 23 “sna tshogs su mos par byed pa” literally, exercising “adhimukti”*. For a treatment of this term in the context of this very passage, see Dhammjoti (2019). 24 smras pa/ dngos po ni rnam par rtog pa thun mong ma yin pa’i rgyu las byung ba dang / rnam par rtog pa thun mong pa’i rgyu las byung ba yin pas/ rnam par rtog pa thun mong ma yin pa’i rgyu las byung ba rnam par brtags pa gang yin pa de ni ’gog par ’gyur la/ rnam par rtog pa thun mong ba’i rgyu las byung ba rnam par ma brtags pa gang yin pa de ni gzhan gyi rnam par rtog pas zin pa yin pas mi ’gog ste/ gzhan du na gzhan gyi rnam par rtog pa don med par ’gyur ro/ /de ma ’gags na yang rnam par dag pa’i bdag nyid can gyi mthong ba rnam par dag par ’gyur bar khong du chud par bya ste/ ’di lta ste/ dper na rnal ’byor pa

Turning Earth to Gold

105

To make sense of Asaṅga’s reply we need to turn first to the Buddhist view, presented for instance by Vasubandhu in the AKBh iv.1a (Pradhan 1975, 192: 3–5), that the arising of the “sentient and insentient receptacle world” (sattvabhājana-loka) is caused by the accumulated actions (karma) of its beings. This idea gained prominence and was further developed in various Yogācāra treatises which usually explained it in terms of the activity of the store-house consciousness.25 For instance, in the first chapter of Asaṅga’s ms (Lamotte 1973, vol. 1, 22–23, verses 59–60) it is argued that our common surrounding “receptacle world” (bhājana-loka) and personal sense sphere (prātyātmikāyatana) – i.e., our shared intersubjective experiences and private experiential content – are traced, respectively, to the maturation (vipāka) of similar and dissimilar karmic seeds (bīja) and impressions (vāsanā) in the storehouse consciousness. Within this framework, whatever causal mental activity is shared at any given moment by our respective mind-streams (saṃtāna)26 will appear as intersubjective, and whatever causal mental activity is not shared will be experienced privately.27 We can all therefore have a simultaneous perception of the external world because of our shared karmic seeds and impressions, but the discrepancies between our perceptions, as well as the fact that we do not share

25

26



27

rab tu mang pos dngos po gcig la mnyam par gzhag pa’i shes pas sna tshogs su mos par byed pa dag la mthong ba sna tshogs dag dmigs par ’gyur ba de dang ’dra’ o/ (vs td. 4038 Zi 13a2–5). An example of an exception is the Abhidharmasamuccaya ascribed to Asaṅga, which refers to the construction of the lifeworld by common and uncommon karma without any explicit reference to the store-house consciousness. See Pradhan (1950, 55) and Walpola (2000, 118). That is to say, a particular mental causal continuum. A definition of saṃtāna is provided by Vasubandhu in the ninth chapter of the AKBh, which is dedicated to a refutation of the existence of persons. As translated by Kapstein (2001, 374): “That which, preceded by deeds, is an on-going coming-to-be of mental events, is a continuum [saṃtāna R.T]. Its arising otherwise is transformation. And moreover, that potency, which immediately produces the fruit, being distinct from [that involved in] other transformations, is the distinctive feature of the transformation….” This definition, which equates saṃtāna with a causal mental process, indicates that the distinctiveness of each such stream is accounted for by the particularity of the specific causal chain that constitutes it. Similarly, the particularity and distinctiveness of any experience is accounted for by a certain “restriction” or “delimitation” (niyamaḥ) imposed upon it, so to speak, by the particularity of the causal chain of mental events that constitutes it. The categories of “private” and “shared” experience require further qualification in this context, because according to the Yogācāra, all of our experiences, even the most private, are pervaded by conceptual categories, which are by definition shared and constitutive of our general life form. See in this respect Waldron (2003, 158–61, 164–169) and Schmithausen (1987, 71–74:4.4.2–4.5.2).

106

Tzohar

our private thoughts, are both accounted for by those portions of our mindstreams that are not shared. It is important to note that this framework not only accounts for experiential agreement but also provides the Yogācāra with a more phenomenologically sophisticated description of intersubjective experiences, one that accounts as well for the discrepancies in different perceivers’ experiences of the same object. Explaining both the shared and private aspects of experience through the same function of causal patterns (their similarity or dissimilarity respectively), the Yogācāra account treats intersubjective agreement as indicating nothing more than the fact that some experiences were commonly and causally constructed. Accordingly, the discrepancies between the experiences of different perceivers with respect to a single object are seen simply as the flip side of the same inherent aspect of this process of construction, that is, they are seen as phenomenologically “built into” our shared experiences. It should be made clear that in this respect the Yogācāra approach to intersubjectivity is radically different from the contemporary phenomenological approach insofar as it does not privilege the first-person perspective. While the school does understand intersubjectivity in terms of the self and others’ shared engagement in a common world, it takes the first-person perspective to be a product rather than the enabling condition of this engagement.28 This account has important ramifications for the understanding of the school’s approach to other issues related to personal identity, like the problem of other minds (in both its epistemic and conceptual versions). This is because this account – by which the subject’s experience is understood to be constructed with and with respect to another’s experience – does not see us as having privileged access to the content of our own mental states relative to our access to the mental states of others, or more accurately, holds that this kind of privileged access is merely apparent. As such, this framework undercuts the main difficulty that gives rise to the problem of other minds, namely the seeming asymmetry in the access we have to the content of our own mind and to the minds of others. This asymmetry, which perplexed many later Buddhist thinkers and is discussed at length in the contributions to this volume by Siderits29

28 29

For a detailed account of how and why the early Yogācāra arrives at this unique understanding of intersubjective experiences, and its implications for the school’s understanding of the lifeworld and otherness, see Tzohar (2017). Siderits (Chapter 6) observes the way in which the asymmetry and the attendant problem of other minds seriously challenges the philosophical enterprise of Dignāga-Dharmakīrti, and is adequately met only in the later work of Ratnakīrti.

Turning Earth to Gold

107

and Moriyama,30 did not pose any real difficulty for the early Yogācāra. Under their account, the subject’s privileged access to her cognitive content neither attests to nor reflects any real underlying ontological conditions, but merely circumstantial causal ones.31 This framework allows the Yogācāra to present a mentally constructed realm of shared experiences, yet without undermining the irreducible differences between each individual experience, tracing them all to an underlying causal nexus. Elsewhere (Tzohar 2016) I discuss in detail how this account addresses what can be seen as a Buddhist version of the problem of linguistic incommensurability, by explaining how beings at different levels of advancement on the Buddhist path can converse in a meaningful way despite the epistemic abyss that lies between them.32 Here, however, I focus on the ­epistemic-qua-perceptual version of this problem as raised by Asaṅga above – that is, on the question of 30

31

32

The focus of Moriyama (Chapter 8) is on the problem of other minds, and more particularly its distinctively Buddhist version, which is concerned with the Buddha’s direct knowledge of the contents of other minds (and hence with the difficulty of how to settle the Yogic ability to cognize the mind of another, given the latter’s alleged ultimate nonexistence). As indicated above, the early Yogācāra rejection of the asymmetry as not grounded in any ontological conditions but merely in causal circumstantial ones serves to solve this issue: in the ultimate absence of any real distinction between the self and other, the “mind of another” as well as “my mind” are only constructs. The buddhas/yogis know them, but not as independently existing phenomena. For the Buddha, knowledge of other minds is just like knowledge of one’s own mind (and both are unreal). This does not deny the asymmetry as an experiential fact, but points out that it is ultimately unreal. This is because the distinct identity of each “mind” – in the Yogācāra lore, a saṃtāna, i.e., causal mental stream – as well as the dichotomy between “this mind stream” and “other mind stream” are a conceptual construction. They are not grounded in a distinction of ontological kind, but just in what is causally connected to what. In other words, for the early Yogācāra, the distinct identity of a saṃtāna is constituted solely by certain causal patterns and histories, causal pathways that are “carved” out of the overall human matrix of causal mental events. This “carving,” however, is not determined by any teleological or ontological conditions, but is understood as governed by ignorance and conceptual discrimination from time immemorial – which is the Yogācāra way of indicating that this process has no temporal starting point, primary cause or teleological end. It is just the way things are; the way in which the causal matrix in its totality, which is karma, gives rise to certain causal patterns, dependent phenomena that by power of habit and conditioning are repeated and become ossified, and then are erroneously conceived as essentially real. This analysis focuses on Sthiramati’s figurative theory of meaning as presented in the Triṃśikābhāṣya, pointing out features that this theory shares with contemporary analytical causal theories of reference – in particular, that it allows for the reference of a term to remain fixed even as its meaning changes from one speaker to another. This feature may be utilized to overcome incommensurability, and in the case of the Yogācāra, the distinction between a fixed reference on the one hand and meaning-variance on the other allows one to explain how a simpleton and an advanced bodhisattva may still communicate successfully, using the same words albeit with very different meanings.

108

Tzohar

how an enlightened being (for whom discursive thought has been eliminated) and an ordinary person can share a world while having deeply discrepant perspectives. Let us return to Asaṅga’s reply to the objection that the elimination of discursive thought by the bodhisattva also entails for him the annihilation of phenomena. Asaṅga addresses this issue by treating discursive thought along two recursive axes: on the one hand, it is the unified cause of all mentally constructed phenomena, and on the other hand it is the product of either shared or unshared constructive processes (traced back to common and uncommon karma). The bodhisattva’s elimination of discursive thought is understood as applying to the unshared kind alone. Therefore, while she is still active in the intersubjective constructed realm of saṃsāra (maintained by the shared discursive thought of others), she no longer actively constructs it for herself or others. Asaṅga concludes by noting that bodhisattvas see for what it is even what remains of the discursive thought of others, viewed by bodhisattvas as a purified realm, free of construction. This is supported by an analogy designed to illustrate a perspectival view of experience. The scenario involves a meditative practice in which the adept focuses his attention on one particular prescribed object; though several such practitioners meditate on the same object, they nonetheless see different things in accordance with their different dispositions. By analogy, we are to understand, the enlightened mind and the defiled mind, though they see the very same things, perceive them differently. This discussion is reiterated in strikingly similar terms in the ms: When it is shared (sādhāraṇa), the store-consciousness is the seed of the world-receptacle (bhājanaloka); when it is unshared (asādhāraṇa), it is the seed of the individual bases of consciousness (prātyātmikāyatana). Shared, it is the seed of the insensate (nirveditotpattibīja) world; unshared, it is the seed of the sensate (saveditotpattibīja) world. When [the truth of the Buddhist path (mārgasatya)] counteracting (pratipakṣa) the store-consciousness arises, the unshared store-consciousness which is counteracted (vipakṣa) disappears (nirudhyate); on the other hand, the shared store-consciousness which continues to exist, grasped by the concepts of others (paravikalpa), becomes the object of a purified (viśuddha) vision (darśana). Thus, by reason of diverse aspirations (bhinnādhimukti), the yogins have different visions (bhinnadṛśya) of one and the same thing (dravya)….33 (ms. i. 60, Lamotte (1973, vol. 1, 23); Chödrön (2012, 110)) 33

A similar argument used to refute the existence of external, mind-independent objects appears also in ms ii.14.

Turning Earth to Gold

109

In its gloss, presented below, the U commentary further clarifies what is at stake with respect to the adept’s experiences under such circumstances. It points out that being pure and free of construction, the content of such experiences – for instance their intentionality and diversity – is guided and limited by no other constraint than the adept’s “aspirations:” Among the yogins, conceptualizations (vikalpa) coming from the individual bases of consciousness are destroyed. Nevertheless, they have only a purified view (darśanaviśuddhi) of the receptacle-world grasped by the conceptualizations of others (parasaṃtānavikalpa). They see this world as a pure thing, like space (ākāśa) that is not stained (pūtīkṛta) by water (āpas), or in the grip (gṛhīta) of earth (pṛthivī) or burned (dagdha) by fire (tejas), or shaken (prakampita) by wind (vāyu). – But how does one acquire a purified vision (darśanaviśuddhi) of something real (sadartha)?34– Fearing some such objection from his adversary, the author says: Thus the yogins, by reason of diverse aspirations, etc.: Thus they have a different vision of one and the same thing according to such and such an aspiration: they see gold, etc. (suvarṇādiviśeṣa bhinnadṛśyam upalabhyate). Depending on whether they see gold (suvarṇa) or silver (rajata) or grass (tṛṇa) or wood (kāṣṭha) there, the object they perceive (ālambana) is multiple. (U. on ms i.59, Lamotte (1973, vol. 2, 83); Chödrön (2012, 111–112)) These sections in fact give us a rather neat summary of the topic. They describe the possibility of a shared experience – a co-habitation – between strikingly discrepant perspectives on reality, represented by the ordinary discriminative conceptual outlook, and the non-conceptual one. The meeting between these perspectives is rendered coherent by explaining them both in similar terms – namely, as a function of their knowledge of an underlying common and causal mental phenomenon. Within this shared realm, however, the advanced bodhisattva is not a mere non-participating guest, a “mind-less” Buddha cocooned in his non-conceptual cognition, but an interacting and communicating

34

The rest of the passage makes clear that the question at stake concerns not the procedure of attaining such knowledge but rather its experiential content and determining factors. If reality is non-conceptual and inexpressible, what then serves as a delimiting factor, a constraint, that may bring about for the yogi a certain (pure) cognitive appearance of one particular object and not another? Anticipating this question, the commentary on ms states that the only factors involved are the yogi’s aspirations, i.e., his mental effort and deliberation, and not an external objective given.

110

Tzohar

­ ember. While she occupies a world that is in fact maintained by the concepm tual thinking of others, seeing it for what it is she moves freely within this realm, which she fashions – conceptually and intentionally – according to her aspirations, or in other words, according to her will and vocation. Here, the “space”-like experience of non-conceptual cognition, rather than marking the ultimate annihilation of intentionality and verbal activity, is seen as the backdrop against which objects may newly manifest and the tool with which a bodhisattva can turn earth to gold. Subsequent cognition is therefore not a different kind of knowledge from non-conceptual cognition, rather it is non-conceptual cognition brought into the world of construction. It is, we should remember – and here we return full circle to Sthiramati’s understanding with which we ­began – a view of the dependent nature as the perfected nature, an exhaustive perspective on an underlying causal reality. And it is the possibility of conceiving of this underlying reality from a number of perspectives simultaneously (from the side of the imagined nature in essentialist terms, and from the side of the perfected without essentializing or reifying it under the subsequent cognition) that enables us to bridge the deep epistemic divide and to resolve the apparent incommensurability between the bodhisattva’s experience and ordinary experience. Finally, to bring to closure the issue of the subsequent cognition and its relation to non-conceptual cognition: above I mentioned that insofar as nonconceptual states do not allow for any intentional or verbal cognitive content, they are necessarily private, but at the same time preclude the possibility of a private language. Conversely, we should note that the private language argument does not preclude the possibility of non-conceptual experiences but rather indicates that they are meaningful only within a shared linguistic realm. This picture can yield an approach that is dismissive of the relevance of nonconceptual experiences to inquiry, or it can be taken, as the Yogācāra indeed seem to take it, as an invitation to rethink the nature and operation of ordinary language through the prism of this possibility of non-conceptual experience. The subsequent cognition, as a view on causality, seems to do just that. Glossary adhimukti aspiration, wish (see āśaya) āśaya aspiration, wish (see adhimukti) nirvikalpa non-conceptual nirvikalpalokottarajñāna transcendent cognition that is free of conceptual differentiation/non-conceptual cognition

Turning Earth to Gold

111

paratantrasvabhāva dependent nature tatpṛṣṭhalabdhajñāna cognition obtained subsequent to it vikalpa conceptual/conceptual discrimination Bibliography Arnold, Dan. 2003. “Verses on Nonconceptual Awareness: A Close Reading of Mahāyānasaṃgraha 8.2-13.” Indian International Journal of Buddhist Studies 4: 9–49. Buescher, Hartmut. 2007. Sthiramati’s Triṃśikāvijnaptibhāṣya: Critical Editions of the Sanskrit Text and Its Tibetan Translation, Beiträge Zur Kultur-Und Geistesgeschichte Asiens Nr. 57. Wien: Verlag der Österreichischen Akademie der Wissenachaften, öaw. Chödrön, Gelongma Karma Migme. (tr) , Mahāyānasaṃgraha: (La Somme du Grand Véhicule d’Asaṅga)by Étienne Lamotte, Translation and Commentary. Vol 2. Cape Breton, CA: Gampo Abbey. D’Amato, Mario. 2009. “Why the Buddha Never Uttered a Word.” In Pointing at the Moon: Buddhism, Logic, Analytic Philosophy, edited by Mario D’Amato, Jay L. Garfield and Tom J.F. Tillemans. Oxford: New York. D’Amato, Mario. 2012. Distinguishing the Middle from the Extremes: A Study and Annotated Translation of the Madhyāntavibhāga, Along with Its Commentary, the Madhyāntavibhaga-Bhāṣya. New York: American Institute of Buddhist Studies. Derge T. (199–?). Bstan ’Gyur Sde Dge’i Par Ma: Commentaries on the Buddha’s Word by Indian Masters (Electronic Edn, 11 Cds). [Reproduced from Editions of the Individual Sections Published in Delhi at the Delhi Karmapae Chodhey, Gyalwae Sungrab Partun Khang, 1986, Itself a Reproduction from Prints from the 18th Century Sde-Dge Blocks]. (213 vols.) New York: Tibetan Buddhist Resource Center. Dhammajoti, KL Bhikkhu. 2007. Sarvastivāda Abhidharma. 3rd rev. ed. Hong Kong: Centre for Buddhist Studies, University of Hong Kong. Dhammajoti, KL Bhikkhu. 2019. “Adhimukti, Meditative Experience andVijñaptimātratā.” in Investigating Principles: International Aspects of Buddhist Culture – Essays in Honour of Professor Charles Willemen, edited by Lalji “Shravak” and Supriya Rai. Hong Kong: The Buddha-Dharma Centre of Hong Kong. 135–172. Gold, Jonathan C. 2014. Paving the Great Way: Vasubandhu’s Unifying Buddhist Philosophy. New York: Columbia University Press. Kapstein, M. (2001). Reason’s Traces: identity and interpretation in Indian & Tibetan Buddhist thought (Studies in Indian and Tibetan Buddhism;). Boston: Wisdom Publications.

112

Tzohar

Lamotte, Etienne. 1973. La Somme Du Grand Véhicule D’Asaṅga (Mahāyānasaṃgraha). 2 vols, Publications De L’institut Orientaliste De Louvain. Louvain-la-Neuve: Université de Louvain, Institut orientaliste. La Vallée Poussin, Louis de. 1971. L’abhidharmakosa De Vasubandhu. Nouv. éd. anastatique présentée par Étienne Lamotte. ed, Mélanges Chinois Et Bouddhiques,; v. 16;. Bruxelles: Institut belge des hautes études chinoises. Matsuda, Kazunobu. 1996. “Nirvikalpapraveśadhāraṇī: Sanskrit Text and Japanese Translation.” In Bulletin of the Research Institute of Bukkyo University, 89–113. Kyoto: The Research Institute of Bukkyo University. Pradhan, P. (1950). Abhidharma samuccaya (Visva-Bharati studies, 12); Santiniketan: Visva-Bharati. Schmithausen, Lambert. 1987. Ālayavijñāna: On the Origin and the Early Development of a Central Concept of Yogācāra Philosophy, Studia Philologica Buddhica. Monograph Series. Tokyo: International Institute for Buddhist Studies. Tzohar, Roy. 2016 “Does Early Yogācāra Have a Theory of Meaning? Sthiramati’s Arguments on Metaphor in the Triṃśikā-Bhāṣya ”Journal of Indian Philosophy, 45 (1): 99–120. Tzohar, Roy. 2017. “Imagine Being a Preta: Early Indian Yogācāra Approaches to Intersubjectivity.” Sophia international journal of philosophy and traditions, 56 (2):337–354. Tzohar, Roy. 2018. A Buddhist Yogācāra Theory of Metaphor. New York: Oxford University Press. Tzohar, Roy. 2020. “Contemporary Non-Conceptualism, Conceptual Inclusivism, and the Yogācāra View of Language Use as Skillful Action.” Philosophy East and West, 70 (3):638–660. Waldron, William S. 2003. The Buddhist Unconscious: The Ālaya-Vijñaana in the Context of Indian Buddhist Thought. RoutledgeCurzon Critical Studies in Buddhism. London/New York: RoutledgeCurzon. Walpola, Rahula. 2000. Abhidharmasamuccaya – the Compendium of the Higher Teaching. Trans. Sara Boin-Webb. Fremont, CA: Asian Humanities Press. Yamaguchi, S., & Lévi, S. (1934). Madhyantavibhāgaṭīkā : exposition systématique du Yogācāravijñaptivāda. Nagoya: Librairie Hajinkaku.

Part 2 Meta-cognition

∵

Intoduction to Part 2 The papers in this section take up the question of meta-cognition, but what exactly is that question? Humans, like other animals, are sometimes conscious of objects and events, whether external, such as colors and sounds, or internal, such as throbbings and yearnings. They are also (perhaps unlike other animals) sometimes conscious of their states of consciousness. Not only are they able to act on the basis of their awareness of colors and such; they are also, at least sometimes, able to report the occurrence of these states. We attribute conscious mental states to ourselves, as well as to others. The question of metacognition is the question of how this is possible: if we report only that of which we are cognizant, how is the cognizing itself cognized? It is important to distinguish this question from other questions about the possibility of self-knowledge. Just as I can be wrong about my height or weight, so I can be mistaken about what my yearning is a yearning for – or even, in cases of thought insertion, about who is thinking the thought that I report. Knowledge of the content of my own mental states might come more readily than knowledge of what is in the minds of others, but there are still cases where others are better positioned to know what I am thinking or feeling than I am. Knowledge of the contents of our own minds is not likely to provide the sorts of secure epistemological foundations that have sometimes been sought there. But this is in any event quite distinct from the question of meta-cognition that was debated in classical Indian philosophy. No participant in that debate thought that an answer to the question might furnish a ground for epistemic warrant. The question concerns not our knowledge of the contents of mental states, but our knowledge that we (and others) have mental states at all. How, in a word, do we know that we are not zombies? Classical Indian philosophers distinguished between two basic sorts of answers one might give to this question: one might hold that cognitions are selfilluminating, or that they are other-illuminating. Their classification uses the common metaphor of consciousness as illumination: to be aware of object O is to shine a light of focal awareness on O, to make O visible by bringing it out of the darkness, to bring O to one’s attention so that one can act on it. To say that cognitions are self-illuminating is to say that a given cognition makes not just its object but also itself visible or manifest. To say instead that cognitions are only other-illuminating is to say that meta-cognition is achieved by means of a cognition distinct from the target cognition. Two sorts of other-illumination theory were propounded. The first sees metacognition as resulting from the operation of an inner sense (called manas) that works much like the external

© koninklijke brill nv, leiden, ���1 | doi:10.1163/9789004440913_007

Introduction to Part 2

115

sense faculties. The second claims that we cognize cognitions by deploying a theory of mind that posits conscious states as the hidden causes of bodily action and speech. Self-illumination theorists hold that every cognition is itself cognized, but other-illumination theorists deny this; they claim that since it is a cognition distinct from the cognition of O that brings it about that the cognition of O is cognized, one may be aware of O without, either at that moment or subsequently, being aware of being aware of O. Meta-cognition is optional, not mandatory, according to other-illumination theorists. The thesis that every cognition is aware of itself – that it is characterized by reflexive self-awareness (svasamvedana) – was most famously promulgated by Dignāga, and three of the papers in this section have something to say about his position and its defense. Coseru sees Dignāga’s view as the natural outcome of a project that is proto-phenomenological in character. His (controversial) reading of Dignāga is partly grounded in the fact that Dignāga fashioned his epistemology so as to be neutral between the representationalist realism of Sautrāntika and the idealism of Yogācāra. Coseru takes Dignāga’s metaphysical neutrality to signal the same insistence that we eschew metaphysics, and turn to the experience itself, that is behind the Husserlian project. And what Coseru takes Dignāga’s reflexivity thesis to amount to is that when we prescind from strictly metaphysical considerations, what we find is a first-personal for-­ me-ness character; this is the self-revelation component of every experiential state. It will be recalled from the Introduction that reflexivity theorists commonly use an infinite regress argument to reject higher-order accounts of meta-cognition. Coseru claims that higher-order accounts are prone to this difficulty because they fail to acknowledge the first-personal subjective character of conscious mental states as precisely what makes them conscious. This failure generates an infinite regress insofar as any higher-order cognition invoked to account for the conscious character of the target cognition will itself fail to be conscious in the absence of a yet higher-order meta-meta-cognition. The paper by Arnold takes up a debate among Tibetan scholastics concerning the status of Dignāga’s thesis in the wake of Madhyamaka critiques of reflexive self-awareness. While there is some scholarly disagreement about just what this debate amounted to, Arnold follows the line of interpretation according to which it concerns whether the later account of reflexive self-­ awareness given by Śāntarakṣita is sufficiently different from Dignāga’s that it can be used to answer Madhyamaka critiques of Dignāga’s account. Arnold claims that it can, since it avoids modeling cognition on the act-object structure that presumably left Dignāga’s account open to the charge that it violates irreflexivity. On Arnold’s telling, Śāntarakṣita means no more by the claim that cognitions are reflexively self-aware than the anodyne claim that they are

116

Introduction to Part 2

c­ onscious. But this, it turns out, consists in their being first-personally presented to the subject. Arnold is well aware that this invocation of subjectivity will not sit well with many Buddhists, so he hastens to invoke Śāntarakṣita’s status as a Mādhyamika: since Madhyamaka denies that there are any ultimate reals, this characterization of cognition must be taken as merely conventionally true, i.e., as simply reflecting folk psychology. This would, however, disqualify Śāntarakṣita’s account as an explanation of the possibility of meta-cognition. People do regularly attribute conscious states to themselves. Dignāga thought that his formulation of the reflexivity thesis would tell us how this is done. But if, as Arnold’s Śāntarakṣita has it, all the thesis amounts to is that each of us has access to our own experiences in a peculiarly first-personal way, it is not clear that this brings us further along in our quest to understand meta-cognition. The paper by Siderits raises difficulties for the reflexive awareness approach, and discusses sources of support for an alternative Buddhist higher-order thought approach. The principal problems with reflexivity are two: that reflexive self-cognizing by an ultimately real cognition would violate irreflexivity; and that the reflexivity thesis leads to solipsism. (The second difficulty echoes a theme in Kripke’s interpretation of Wittgenstein.) The alternative approach, which is based on some discussions in Madhyamaka, claims that first-person attribution of mental states proceeds in the same way as third-person attribution, by way of an abductive inference from behavior and speech. Siderits puts this in terms of an opacity thesis: our own minds are no less opaque to us than are the minds of others. This account calls into question an assumption common to most Buddhist discussions of cognition, that consciousness is an ultimately real constituent of the psychophysical complex. The inference to the occurrence of cognition being abductive in nature, consciousness acquires the status of a theoretical posit, perhaps no more than a useful fiction. This is acceptable to Mādhyamikas, insofar as their school denies the coherence of the idea of ultimately real entities. The evidence that Siderits discusses in support of this account, however, derives from recent work in cognitive science. The physicalist orientation of that work results in some tension between the evidence and the Madhyamaka Buddhist account: Madhyamaka, like all other schools of Indian Buddhist philosophy, rejects physicalism. A fourth paper in this section discusses both the Indian origins of the reflexivity view and its reception in East Asian Buddhism. Funayama’s thesis is that the fourfold division analysis of cognition attributed to the Indian commentator Dharmapāla as a refinement of Dignāga’s threefold division analysis is actually a Chinese innovation. To appreciate the significance of Funayama’s thesis, some explanation of the terminology is necessary. In Indian epistemology, the chief task is to develop a theory of the number and nature of the pramāṇas

Introduction to Part 2

117

or epistemic instruments – the processes that invariably cause veridical cognitions. Any occurrence of such a process always involves something that stands as the object about which veridical cognition is sought, the prameya or object of cognition. The operation of the pramāṇa on the prameya brings about a cognition that stands as the result or pramāṇaphala (literally the fruit of the epistemic instrument). So for a naïve realist, properly functioning vision can count as a pramāṇā, a mango might serve as the prameya, and the cognitive state of the perception of a mango would be the pramāṇaphala. But in order to accommodate his Yogācāra brand of idealism, Dignāga must collapse these three components into one. It will be recalled that Dignāga claims each cognition has two forms or aspects, the object-form or noematic aspect (arthākāra) and the noetic aspect, the form of a cognition as such (svākāra). In a key passage in his principal text Pramāṇasamuccaya, Dignāga identifies the objectform with the prameya and the form of the cognition with the pramāṇa. To this he then adds that the cognition’s cognizing itself (i.e., the self-presentation of the object-form) may be identified as the result. The point he wants to stress in this passage is that these three aspects are actually just the three different ways in which a single unanalyzable entity may be described given our interests and cognitive limitations. This is the price he must pay in order to keep consciousness in his ultimate ontology while also respecting the tenets of nonself and mereological nihilism. The fourfold division analysis developed (according to Funayama) by Chinese Buddhist epistemologists (who then attributed it to Dharmapāla) adds cognition of self-cognition to Dignāga’s three aspects. This is remarkable because it effectively undermines the point Dignāga sought to develop: that only by identifying his three aspects can we avoid an infinite regress in our account of meta-cognition. If we say, as Dignāga does, that every cognition is itself cognized, and we also hold that reflexive awareness is a cognition, then the possibility of our being aware of reflexive awareness will not demand yet another cognition only if what we call reflexive awareness is simultaneously noema or object, noesis or subject, and noetic activity of the cognizing of object by subject. For only this identity prevents the occurrence of the resultant activity from requiring that it be the object of a further cognition; only thus can the self-disclosing of the object be its own disclosure. The fourfold analysis model only prevents an infinite regress by allowing for the possibility that the cognition of self-cognition might itself occur in object position for the self-cognition that is its object. The model thus reverts to an earlier Indian model of the Buddha’s omniscience, one according to which there are two distinct cognitions, either simultaneous or successive, involved. But this opens the door to a ­higher-order account of meta-cognition.

118

Introduction to Part 2

Finally, the paper by Moriyama discusses not first-person meta-cognition but third-person meta-cognition, the ability to attribute mental states to others. There would seem to be nothing mysterious about this ability: we interpret the minds of others from their behavior. But the realist opponent thinks it may not be so easy for the idealist Yogācāra to explain it, given their denial of external objects and so of bodies whose behavior we can observe. Yogācāra responds, predictably, by invoking causal relations across distinct mental streams to explain the observations that we take as evidence of mental states in others. But this leads to a new puzzle. Buddhas and highly advanced meditational adepts are said to cognize the mental states of others not by inference but directly. The question is what this could be like. Moriyama discusses some answers developed in the Chinese tradition built around the work of Dharmapāla. (This is the tradition whose views on mental consciousness are discussed in Part 3 of this volume.) Perhaps not surprisingly, it seems that because a buddha’s cognition of cognitions (whether their own or those of others) is free of the superimposition of dichotomous concepts like subject and object, it is inexpressible. (This was Vasubandhu’s answer to the “what-is-it-like” question in Viṃśikā 21.) In this respect, buddhas can be said to know our minds better than we ourselves do. But this has the interesting consequence that our cognition of our own cognitions is at least to this extent always mediated by concepts. Our ability to report the occurrence of our own cognitions depends on our bring able to represent to ourselves our own representations. And this, Dharmapāla seems to want to say, depends on our conceptual capacities. One must wonder just how transparent the mind is to itself. This should raise questions concerning the reflexivity thesis. As for how advanced yogic practitioners’ cognition of the minds of others works, the answer given by Dharmapāla’s commentator Kuiji is the sensible if somewhat disappointing one that they are able to form an image that resembles the content of the other’s mind. It is not, after all, direct in the sense of being right there in the stadium while the match is played out. It is only direct in the sense of there being a recreation on the home screen through a “live and direct” cable feed. This is more direct than ordinary third-person attributions, which are comparable to reconstructing how it must have looked on the field after reading a play-by-play account the next day. This difference is said to be a matter of relative vividness, which once again suggests that conceptualization plays a greater role in ordinary third-person attributions. But we are still left to wonder how buddhas can have something even the yogic adept cannot have: cognition of the cognitions of others that is not structured by the

Introduction to Part 2

119

subject-­object dichotomy yet still mediated by representations. If the only difference between my awareness of my suffering and a buddha’s awareness of my suffering is that the latter is a representation of the former, then why is it not the case that it is for the buddha just as it is for me? Why, in a word, is this not the same as the buddha’s being in a bad way?

Chapter 4

Whose Consciousness? Reflexivity and the Problem of Self-Knowledge Christian Coseru Abstract If I am aware that p, say, that it is raining, is it the case that I must be aware that I am aware that p? Does introspective or object-awareness entail the apprehension of mental states as being of some kind or another: self-monitoring or intentional? That is, are cognitive events implicitly self-aware or is “self-awareness” just another term for the cognition that takes an immediately preceding instance of cognition for its object? Not surprisingly, intuitions on the matter vary widely, and classical Indian philosophy presents us with three distinct views on the problem of self-knowledge: (1) cognitive events are implicitly or “reflexively self-cognizant”: I am simultaneously aware of the rain and of my awareness of it; (2) awareness of a particular cognitive event occurs through an additional, but immediately following, cognition, most often identified as an “inward” perception (anuvyavasāya); (3) cognitive events become instances of self-knowledge only when taken up inferentially in “reflection”: I know that I was bitten by a spider upon apprehending the spider’s bite mark on my skin. Dignāga and Dharmakīrti defend versions of the first position, Nyāya and some Buddhist philosophers like Candrakīrti versions of the second and third positions. This paper proposes a novel solution to this classical debate by reframing the problem of self-knowledge in terms of the relation between phenomenal concepts and phenomenal knowledge. Concepts of consciousness such as “introspective awareness” (manovijñāna) and “reflexive self-awareness” (svasaṃvedana, svasaṃvitti) are grounded in phenomenal experiences rather than physical events and processes. As such they yield a different kind of self-knowledge than what can be gained by applying externalist conceptual schemas to understanding the mind. I argue that Dharmakīrti’s theory of content, which takes a bottom-up approach to apoha, can thus be seen as endorsing the efficacy of phenomenal experience as a vehicle for self-knowledge.

© koninklijke brill nv, leiden, ���1 | doi:10.1163/9789004440913_008

122

Coseru

Keywords consciousness – self-consciousness – self-knowledge – s­ubjectivity – higher order thought – phenomenal concepts – cognitive phenomenology

...

1 Introduction How do we come to know our sensations, beliefs, desires, and actions? On its face, this question can seem odd, even contrived. What more than, say, feeling the warm glow of the sun does it take to know that I am having a perception of some kind? How could my consciously entertaining the belief that Devadatta is a man not be epistemically salient with regard to the semantic content of the belief? Is there anything more obvious to my being in motion than the regular gait of walking as I lift and set down each foot? And yet it only takes one instance of mistaken self-apprehension to realize that we are no more immediately discerning about our own mental states than we are about external reality or the minds of others. This sort of philosophical awakening is typically short lived, for it is immediately followed by the recognition that what makes self-knowledge problematic is the assumption that knowledge of our own mental states has epistemic privilege over other kinds of knowledge (of nonmental reality and the minds of others). The contemporary debate on self-knowledge in Anglophone analytic philosophy has focused almost exclusively on those specific aspects that grant self-knowledge epistemic privilege, in the sense that some form of immediate acquaintance with one’s own mental states must be constitutive of human subjectivity. But the history of philosophy east and west presents us with a different conception of self-knowledge, one that looks to its role in the achievement of wisdom and, in the case of Buddhism, enlightenment. The Socratic ideal of self-knowledge articulates not simply a conception of who we are, but, and more importantly, of what we can become. As an ideal of self-knowledge’s ennobling and enlightening effects it is not unlike that championed by the historical Buddha, who famously urged his followers to seek the guidance of their own reasoned deliberations and disciplined practice. These two conceptions of self-knowledge may differ in terms of their scope – whether self-knowledge should be about its content or about the act of knowing itself – but they do share a common feature: their concern is primarily not with what it is to know a certain object (a self, subject, or aspect of experience); rather, their concern is with the kind of relation that human subjects have to their own mental states, one which takes self-reference to be a ­necessary

Whose Consciousness?

123

condition for the ability to entertain propositions of any sort. What these two conceptions of the scope of self-knowledge have in common is the notion that self-knowledge is essentially knowledge about oneself or of oneself: in short, it is de se rather than de re knowledge.1 Consider, for instance, your awareness of reading these words right now: to count as an instance of self-knowledge, it is not enough that you be aware of their place in the syntax or the meaning they seek to convey. Rather, you must know that the property of “reading these words right now” belongs to you, that it occurs in your own mental stream. On the problem of self-knowledge, Indian Buddhist philosophy presents us with three distinct views: (1) conscious cognition of an object or mental state is reflexively self-aware or self-intimating; (2) self-cognition is the result of a co-occurrent or immediately succeeding cognition of cognition, most often identified as a kind of “inward” perception akin to introspective awareness; (3) cognitive events become instances of self-knowledge when reflection bears upon a specific inferential sign: I know it must have been a nightmare when I wake up feeling anxious about the dream I had last night. Some philosophers (e.g., Dignāga (c. 480–c. 540), Dharmakīrti (c. 600–660), and their followers) defend versions of the first position, others (e.g., Saṃghabhadra (fl. c. 400) and Candrakīrti (fl. c. 600)) put forward versions of the second and third positions.2 Given the universal scope to the no-self doctrine, specifying what exactly this self-referential relation entails and how self-knowledge is achieved becomes paramount. It should be obvious that Buddhist accounts of substantive self-­ knowledge cannot rest on egological conceptions of self-consciousness, that is, on conceptions of consciousness as the property, function, or dimension of an enduring subject or self.3 Just what it means for there to be self-knowledge without subjectivity, thus, remains an open question. Indeed, if experience 1 I follow Renz’s (2017: 4–5; 8) discussion of classical and contemporary approaches to the problem of self-knowledge. As Renz makes clear, unlike the contemporary focus on the role and function of self-referential thought to self-knowledge (see Shoemaker 1968, Anscombe 1981, and Evans 1982), classical Western debates were circumscribed by a set of epistemic accomplishments that go all the way back to Socrates’ response to the pronouncements of the oracle at Delphi: (i) a firm grasp of the conceptual distinction between belief and knowledge; (ii) the realization that most people, when challenged to examine their own convictions, fail to grasp this important distinction; (iii) an acknowledgment of the limits of his own knowledge; and (iv) a recognition of the possibility of falling short of the demands of wisdom. 2 These positions, and their variants, are not confined to Buddhism. The Prābhākara Mīmāṃsaka shares with Dignāga and Dharmakīrti the view that knowledge episodes are intrinsically (svataḥ) ascertained, while the Naiyāyikas and the Bhāṭṭa Mīmāṃsaka followers of Kumārila come down on the side of the Mādhyamikas in holding a relational view of self-­ knowledge (see Matilal 1989, Ch. 5). 3 The Pudgalavādins (“Personalists”) are an exception, but their concerns are metaphysical (Do persons exist as conceived?) rather than epistemological (How are persons known?). Despite the dominance of the ultraminimalist Abhidharma accounts of agency there are

124

Coseru

lacks any self-specifying features, and if cognitions are merely transient events arising within a continuum of causally interconnected states, what explains the distinctively phenomenal character of first-person agency? How can the efficacy of our epistemic practices be explained without presupposing that conscious cognitive states are diachronically unified? On the upside, the rejection of a persistent owner and/or locus of experience does offer novel opportunities for exploring the structure of awareness and the problem of personal identity beyond the metaphysics of subjects. Investigations of the problem of self-knowledge in Buddhism so far suggest a profound concern with the possibility of achieving a greater degree of selfunderstanding than is ordinarily available through reflection, introspection, or intersubjective reports. As the story goes,4 the idea of cognition being selfintimating or self-reflexive arose among the Mahāsāṃghikas (“Great Assembly” thinkers), who sought to attribute the capacity for becoming an entrant on the path of cultivation (a srota-āpanna) to momentary flickers or flashes of self-luminosity. The Mahāsāṃghikas maintained that these instances of selfluminous awareness explain how it is possible for a mental state to become reflexively self-conscious in a single moment (given metaphysical commitments to the principle of momentariness). Against this view, the Vaibhāṣikas (viz., thinkers who relied on the Vibhāṣa or “Compendium”) – chiefly ­Vasumitra (fl. c. 100), Dharmatrāta (fl. c. 150), and Saṃghabhadra – put forth rival twomoment or two-state views, according to which cognitive awareness is the outcome of a successive or simultaneous co-occurrent cognition. The position of the Vaibhāṣikas was in turn critiqued by Harivarman (fl. c. 350), an early Sautrāntika thinker (so-called because of his association with the Sūtras or “Discourses” School), who took important steps to broaden the scope of inquiry into the problem of self-knowledge beyond merely soteriological concerns. The first proofs of self-knowledge as a one state (mental) event, which also appeal for the first time to the concept of simultaneity, the role of memory, and to arguments against infinite regress, are owed largely to Harivarman.5 Dignāga’s framing of the problem of self-knowledge in the context of debates about the nature and function of perception reflects this new epistemological orientation. In examining this new orientation, I will argue that Dignāga can be plausibly interpreted as making the case for why we can (and should) pursue the question of the subjectivity or subjective character of experience good and compelling reasons to give Buddhist Personalists credit for insisting that important features of personhood are ineliminable (see Carpenter 2015). 4 Yao (2009: 10) traces this preoccupation with identifying the nature of self-cognition to the Mahāvibhāṣā. See also La Vallée Poussin (1928–29: 129). 5 See, for example, Satyasiddhiśāstra, 278bc in Aiyaswami Sastri (1975).

Whose Consciousness?

125

i­ndependently of metaphysical concerns about what this subjective aspect is (and is like), and of what its fundamental attributes might be.6 In order to motivate this view of Dignāga as a proto-phenomenologist, I will propose several ways in which an analysis of the structure of consciousness can be pursued without assuming that such structure (with its subjective and intentional aspects) reflects an external relation of ownership between consciousness and some underlying substratum. The central concern here is whether reflexive awareness involves a direct self-referential relation. In § ii, I review Dignāga’s two-aspectual theory of mental states, and its implications for an account of self-knowledge that must negotiate the subjective and intentional dimensions of experience. In § iii I draw a parallel to contemporary debates in philosophy of mind between “onestate” and “two-state” theories of consciousness and consider what is at stake for accounts of consciousness that ignore its properly phenomenal features. In § iv, I outline Dharmakīrti’s so-called “egocentric predicament”7 (sahopalambha-niyama) argument, and provide an example of problematic issues that relational theories of self-cognition confront. In § v, i consider the conundrum of metaphysical interpretations, and appeal to the epistemic features of phenomenal concepts in order to articulate a conception of minimal self-reflexivity that – despite associations with the so-called “afflicted mind” (kliṣṭa-manas) – affords an understanding of consciousness in non-relational terms. I conclude in § vi with some suggestions for how we may move forward on the problem of self-knowledge without discarding the role of phenomenal consciousness in articulating its content. In defending an account of consciousness that takes self-reflexivity and intentionality to be its co-constitutive and self-specifying features, I also suggest a way in which no-ownership conceptions of reflexive self-consciousness can help us to get the structure of phenomenal consciousness right. 2

Subjectivity, Intentionality, and the Dual-Aspect Theory of Mind

Dignāga shares with Vasubandhu and his Abhidharma predecessors the view that a defining characteristic of conscious cognition is its co-arising together with an object or intentional content. But Abhidharma traditions also operate 6 The Ālambanaparīkṣā is sometimes taken to represent Dignāga’s idealist stance on the object of cognition. But it could also be read simply as a sophisticated phenomenological study of intentionality. 7 The American philosopher Ralph Barton Perry (1910) coined this term to capture the notion, first articulated by Berkeley, that a conception of how things are in themselves is ipso facto still a conception, and, as such, an event within consciousness.

126

Coseru

with an axiomatic distinction between mind (citta) and mental concomitants (caitta): the first stands for discerning awareness (vijñānati) in its many guises (perceptual, auditory, volitional, etc.), while the latter captures the distinctive character of conscious apprehension or its affective and dispositional saliences.8 Most important, the mind and its concomitants do not attend to different objects. Rather, the same object is consciously apprehended by a particular mode of apprehension such as, say, visual awareness, and, at the same time, disclosed as an event in the mental stream – for instance, as a sensed patch of color or a thought of some kind. At least in the context of Indian Buddhist Abhidharma, the general assumption is that all cognitions are inherently intentional: that is, they are necessarily about an object of their own. But this understanding of intentionality requires that there be particular types or modes of cognitive awareness that are uniquely constituted as such: that is, as always being directed to, or being about, something. Is intentionality9 – the apprehension of an object in its first-personal mode of givenness – a distinct mental act, to be added to a long list of mental concomitants? Or is it a structural feature of consciousness itself? In keeping with the Buddhist metaphysical commitment to the momentary nature of all phenomena, even these distinctly intentional modes of cognitive awareness are nothing but temporal instances in a stream of psychophysical events that correspond to phenomenally and physically real structural properties. But now we have a problem: how can presumably non-cognitive or subpersonal factors contribute to the arising of cognitive awareness, let alone sustain the self-referential relation necessary for self-knowledge? If intentionality is a relation between a conscious mental state and its content, how does its objectdirectedness come about? If all there is to the mental domain is a stream of momentary, object-directed mental events, what does it mean for an object to 8 The Abhidhammattha-Sangaha ii, 1 (in Nārada 1979) defines the “mental constituents” as follows: ekuppāda-nirodhā ca – ekālambanavatthukā, cetoyuttā dvipaññāsa – dhammā cetasikā matā (“The fifty-two states that are associated with consciousness, that arise and cease together with consciousness, that have the same object and basis as consciousness, are known as ‘mental states’”). Vasubandhu’s definition, as found in his Pañcaskandhaprakaraṇa (td 31, 848c3–9, in Dantine 1980: 46), is effectively a long enumeration of all elements “associated” (saṁprayukta) with the mind, among them universal (e.g., contact, attention, volition), particular (e.g., action, resolve, concentration), auspicious, and defiled mental states, and the so-called “others” (śeṣa), which include an eclectic list of affective and afflictive dispositions (e.g., hostility, anger, malice). 9 Various terms in the Buddhist philosophical lexicon, e.g., “apprehension of an object” (arthagrahaṇam, viṣayagrahaṇam), “obtaining cognition” (prāpakaṃ jñānaṃ), may be taken to capture roughly the notion of intentionality – that is, a mental state’s aboutness or directedness toward an object – as understood in Western philosophy since Brentano (1874: 88f). See Coseru (2012: Chapter 8) and Arnold (2005: 44f).

Whose Consciousness?

127

be known, and by whom? Whose consciousness is it that does the knowing, and what is it that gets known? The causal principle of dependent arising – an axiomatic principle of Buddhist metaphysics – demands that phenomena arise interdependently such that entities or events possess their properties only by virtue of their placement in the causal web. As the canonical formula for dependent arising goes, visual-consciousness arises in dependence upon a visual system and visual objects. This way of framing the causal relation confronts an explanatory problem: How can sensitivity to light and a surface give rise to a metacognitive awareness of blue: not the mere “seeing of blue,” but the awareness that blue is seen (nīlam iti ca vijānāti)? In short, how does such coming together of sense and object, of thought and its content, become an instance of self-knowledge that tells an individual that the cognition is hers, that it happens in her mental stream? If mental states are taken to be constitutively, rather than relationally, intentional, then self-consciousness is not a relation that consciousness has to itself when it attends to its own operations. Rather, it is a structural feature that certain (though not all) mental states have by virtue of being conscious. Furthermore, a conception of consciousness as constitutively self-intimating solves the problem of having to bridge the epistemic gap between the underlying factors of cognition (e.g., a living body) and its phenomenal character. One does not relate to pain as to an objective property of some sort (like being six feet tall or the husband of one’s spouse); rather one is in pain and the pain is phenomenally foregrounded as a sensation of some kind (of burning, stinging or throbbing). That there should be conscious mental events, such as bare thoughts and sensations that are unknown until they are attended to in a subsequent or co-occurrent moment of awareness is highly improbable.10 Indeed, 10

Armstrong (1984: 119f) – who rejects the Lockean self-intimation thesis that it is “impossible for anyone to perceive without perceiving that he does perceive” (Locke 1975: ii 27, 11) – uses the example of sleep-walking to build a case for the existence of unconscious visual and tactile sensations. While the evidence for preconscious human information processing is indisputable (Marcel 1983; Kouider and Dehaene 2007), whether the notion of “unconscious sensation” is coherent is highly controversial. Drawing on the literature on blindsight, Overgaard et al. (2008) give compelling evidence for why blindsight should be regarded not as a case of unconscious vision but rather of severely degraded conscious vision. Similarly, against Block’s (2012) appeal to two recent cases of neglect as evidence for unconscious seeing, Phillips (2016) argues that, while the various information extraction processes in these cases do provide support for sensory registration, they do not add up to perception, which requires phenomenal consciousness. Furthermore, neglect patients show various attentional and perception deficits, which biases their subjective reports in various ways. As Phillips notes apropos the paradigmatic case for unconscious perception, “the claim that blindsight involves unconscious perception is largely based on a dissociation between responding in a biased task and performance in an unbiased forced-choice task” (2016: 435).

128

Coseru

the notion of a pain that is unknown until it is reflected upon or attended to raises the question: how is such unknown pain ascertained? As a phenomenal concept (that is, as a concept that expresses knowledge of phenomenal qualities) “pain” is acquired experientially. On an account of phenomenal knowledge as mediated by phenomenal concepts, the idea of unknown or unconscious pain is, simply put, a category mistake.11 The problem of self-knowledge, then, is the problem of specifying how phenomenal salience – this distinctive quality that conscious mental states have – becomes a defining self-referential feature of conscious cognition. Is phenomenality, like intentionality, a structural or merely an add-on feature of consciousness? Do mental states acquire their character and perspectival stance as a result of self-grasping or is the mineness or for-me-ness12 of experience built in? Dignāga’s dual-aspect theory of mind can in large measure be understood as an attempt to bring the debate about the primacy of either “intentionality over phenomenality” or of “phenomenality over intentionality” to a resolution. As it stands, the theory rests on three distinct claims:13 (1) that we are directly aware of conscious events such as perceptions and thoughts; (2) that each mental event has a dual aspect: it has both a subjective (grāhakākāra) and objective (grāhyākāra) aspect; and (3) that each mental event is also reflexively ­self-­conscious (svasaṃvedana). The first claim goes against the view that conscious mental states are ultimately impersonal or anonymous (on the assumption that even nonconceptual content is first-personally given). To hold such a view would be akin to claiming that experiential states do not have a self-referential structure or acquire such structure only by virtue of the operations of what Asaṅga calls the “afflicted mind” (kliṣṭa-manas): an inauspicious 11 12 13

Of course, one can talk about tissue damage as the object (or cause) of a sensation of pain, but the physical state of the body prior to the occurrence of this sensation does not admit of phenomenal description. I take “mineness” or “for-me-ness” to be an irreducible structural feature of experience with a distinct phenomenology, not simply an occurrent feature of conscious mental states. See Kriegel and Zahavi (2015) for an articulation of this view. Funayama (Chapter 7) contrasts this threefold schema with the fourfold division of cognition popularized in East Asia by Dharmapāla, which he regards as a Chinese apocryphal theory, and as not particularly helpful in unpacking or making explicit the intent of Dignāga’s threefold division between the objective, subjective, and self-reflexive aspects of cognition. Indeed, Dharmapāla’s postulation of a forth aspect – the “cognition of selfcognition” or “awareness of self-awareness” (svasaṃvitti-saṃvitti) – to account for how self-cognition itself is known – is explanatorily redundant. On Dignāga’s understanding of the concept, svasaṃvitti, that third dimension of cognition, which guarantees that subjective and objective aspects are self-intimating, is itself epistemically salient. It is a metaphysical requirement of conscious cognitions, as opposed to subconscious or subpersonal mental factors (e.g., vāsanās), that they be self-intimating, that they exhibit a particular phenomenal character.

Whose Consciousness?

129

r­epository of  “I-making” and “self-grasping” tendencies.14 The second claim identifies ­subjectivity or the “subjective aspect” (svābhāsa) and intentionality or the “objective-­aspect” (viṣayābhāsa) as distinctive features rather than operations of the mental domain. The subjective aspect is constitutive of an implicit openness to what is given, while the objective aspect captures what the mental state is about: a content of some kind. Finally, the third claim is intended to capture the mode of presentation of all conscious cognitive states. Incidentally, the first and third claims can also be read as making the case that effortful self-­knowledge – of the sort gained through introspection or intersubjective ­reports – ­depends on tacit or non-propositional modes of acquaintance. This understanding of the structure of conscious experience is not unlike that put forward by some contemporary defenders of the view that our subjectivity is immersive rather than egological: as Zahavi observes, t­aking Heidegger’s lead on this issue, “I am acquainted with myself when I am captured and captivated by the world. Self-acquaintance is indeed only to be found in our immersion in the world, that is, self-acquaintance is always the self-acquaintance of a world-immersed self.”15 Likewise, as P.F. Strawson notes with regard to the mode of presentation of our mental states, “our desires and preferences are not, in general, something we just note in ourselves as alien presences. To a large extent they are we” (Strawson 1992:134). Even Wittgenstein comes close to articulating a tacit conception of self-knowledge when he declares: “It is correct to say ‘I know what you are thinking,’ and wrong to say ‘I

14

15

Mahāyānasaṃgraha i, 6–7 (in Lamotte 1938: 36): “How does one know that manas in the sense of ‘afflicted mind’ (kliṣṭa-manas) exists? Without it, there would be no uncompounded ignorance, that is, a basic ignorance not yet associated with all the faults (doṣa), but serving as their base (āśraya). Indeed, introspective awareness (mano-vijñāna) must also have a simultaneous basis, as do the five types of empirical consciousness whose support are their material organs. Such a simultaneous support can only be found in the ‘afflicted mind.’ The etymology of the word ‘manas,’ which means ‘mine,’ and has to do with the afflicted mind, could not be otherwise explained. Also, without it there would be no difference between the trance of non-identification (asaṃjñisamāpatti) and the cessation trance (nirodhasamāpatti), which would constitute a fault. Indeed, whereas the trance of non-identification is free of afflictions, the trance of cessation is not. For those who lack conceptuality, then, there would be neither self-grasping (ātmagrāha), not the conceit of ‘I’ (asmimāna); for as long as they dwell in a non-conceptual state, they would not be afflicted.” See also Kramer (2008) and Waldron (2003: 147) for further discussion of the concept of kliṣṭa-manas and its implication for a Buddhist account of subjectivity. See Heidegger GA 61: 95, and discussion in Zahavi (2005: 82). As Zahavi notes, echoing an argument that is central to Dharmakīrti’s defense of the self-presentational character of cognition, “self-acquaintance is not something that takes place or occurs in separation from our living in a world” (ibid.).

130

Coseru

know what I am thinking.’ (A whole cloud of philosophy condensed into a drop of grammar.)”16 To claim that access to our mental lives is always mediated in some fashion or another, perhaps by participation in a shared domain of language and reflection – is to ignore these essential features of consciousness. Against this view, Dignāga’s dual-aspect theory entails a fundamental difference in the two basic modes of existence of consciousness: as pre-reflective and as reflective. While the former is a kind of immersive non-objectifying acquaintance, the latter stands for a detached and objectifying self-acquaintance. It is to this latter type of conscious apprehension that we owe the distinction between observer and observed. Does Dignāga’s dual-aspect theory of cognition, then, reflect commitment to a subjective ontology, as his critics claim? Or do we have here the makings of a distinctly proto-phenomenological epistemology that brackets assumptions about a world beyond experience so as to lay bare the phenomenal character embodied in that experiential structure? I do not mean to suggest that the dual-aspect theory cannot be read as a statement about the metaphysics of mind and thus as open to the sort of criticism leveled against it by both Buddhist and non-Buddhist rivals.17 Rather, I want to claim that such a reading is less helpful if we take Dignāga’s project to be descriptive in scope, and concerned with making sense of experience itself rather than negotiating various theoretical assumptions about it. Specifically, the question is whether the first-personal givenness of experience entailed by the “subjective aspect” (svābhāsa) is itself an intentional or objectifying stance, and thus some kind of internal or inner perception. If it is, then we have to side with Yao (2005) against Hattori (1968), Nagatomi (1979), Franco (1993), and all those who take the self-intimating dimension of reflexive self-consciousness (svasaṃvedana) to be a distinct type of inner or mental perception rather

16

17

Wittgenstein (2001: 189). It is true that elsewhere in the Philosophical Investigations (e.g., Section 357), Wittgenstein insists that a meaningful articulation of our inner mental states requires the sort of accessibility that only public ways of knowing can provide. And yet, here he seems to concede that our phenomenal or subjective use of the first-person pronoun does have immunity to error through misidentification. See also Overgaard (2007: 135ff). Arnold (2012) makes the case that while Candrakīrti's and Kumārila’s critiques of the phenomenology of first-order experience in Dignāga’s memory argument for svasaṃvitti do raise legitimate concerns (for instance, about the absurd consequences of admitting that all cognitions, by virtue of being self-presenting, are epistemically salient), they fail to capture the intent of Dignāga’s phenomenological stance. What Dignāga is most likely after with his argument for svasaṃvitti is a criterion by means of which the occurrence of mental states, rather than their intentional content, can be indexed as such, that is, as states of a distinctly perceptual or conceptual kind.

Whose Consciousness?

131

than a structural feature of consciousness itself.18 On the other hand, if the “subjective aspect” is self-intimating without being constituted as the intentional content of a previous cognition, then the structure of self-knowledge is non-relational. What is at stake in this debate about the character of self-knowledge, then, is whether self-knowledge is best understood on a perceptual or acquaintance model. According to the perceptual model, mental states are logically (and perhaps ontologically) independent of the awareness that can attend to them. That a certain mental state can become the object of another mental state depends on contingent features of our cognitive architecture: I see color because I am sensitive to light (where sensitivity reflects an external relation of some sort between an organism and its surroundings). But on this model the relation between mental states and our acquaintance with them is causally indeterminate. A whole series of concurrent factors must be in place for an instance of self-awareness to occur: an object or mental state of sufficient salience, the lack of distracting factors, and a well-developed habit of attending to the present at hand. The perceptual model does not tell us how self-­ knowledge can be achieved, only that, given the right chain of events, it is so achieved. By contrast, on the acquaintance model, the link between a mental state and its conscious apprehension is constitutive rather than causal. If Dignāga’s account of reflexive self-consciousness (svasaṃvedana) follows the constitutive model, then having an experience of some kind, say of seeing blue, is akin to standing in an intimate epistemic relation to that experience.19 18

19

Franco’s position on “introspective awareness” (mānasa-pratyakṣa) is also a critique of Wayman’s (1991) view that Dignāga did indeed treat such awareness as a different type of perception. Yao rests his claim that Dignāga does take self-awareness as a type of perception on a close reading of the Chinese materials, specifically on translations of Nyāyamukha by Xuanzang (600–664) and Yijing (635–713), as well as on Kuiji’s (632–682) commentaries on Dignāga’s principal works. These texts seem to indicate, in no ambiguous terms, that Dignāga treats self-awareness as a distinct form of perception. For Yao mental perception (mānasa-pratyakṣa), then, and the mental faculty of cognitive awareness (mano-vijñāna) are to be clearly differentiated, the first being just an aspect of the latter. This perceptual model self-knowledge follows closely Shoemaker (1996: 224). See also Chalmers (1996: 195), for a discussion of constitutive versus causal accounts of epistemic access to our mental states, and Zahavi (2005: 22) for parallels to Sartre’s views on consciousness. The “acquaintance view” also resonates with Zahavi’s account of subjectivity as minimal mineness or for-me-ness, for which see Gallagher and Zahavi (2012) and Zahavi and Kriegel (2015). Ganeri (2012: 154) finds parallels to Vasubandhu’s account of an emergent subjectivity from the activity of kliṣṭa-manas in these contemporary accounts of minimal subjectivity or for-me-ness.

132

Coseru

It is important to remember that Dignāga develops his account of selfknowledge both in response to a received tradition of speculation about the so-called “luminosity” of the mind, and as a critique of alternative theories of perception, specifically those advanced by Nyāya, Bhāṭṭa Mīmāṃsā, and Buddhist realists (e.g., Sarvāstivādins).20 Thus, when Dignāga takes self-intimation to be the mode of presentation of all perceptual experience, he does not merely systematize into a system of epistemic warrants what had traditionally been known as empirical awareness and self-consciousness. Rather, he seeks to empirically ground his stance on cognition: perception does the job of apprehending particulars as uniquely characterized phenomena, but only if operating in a non-conceptual mode. If the goal is to account for the sense of mineness or for-me-ness that conscious cognitive events exhibit, naïve realist assumptions about how the mind interacts with the world are to no avail. Let us take a closer look at Dignāga’s claim, as articulated in his major work, the A Collection on the Sources of Knowledge (the Pramāṇasamuccaya and Vṛtti) that perception appears in its dual aspect as awareness of something coupled with self-awareness: Every cognition comes about with a double appearance, namely that of itself and that of the object. The awareness of itself as [possessing] this double appearance is the result [of the intentional act] – because the determination of the object [as cognized] conforms to it. When a cognition intending its object itself becomes an object of apprehension, then one apprehends it as either desirable or undesirable in conformity with self-awareness. When, on the other hand, the object to be apprehended is an external entity, then the source of knowledge is simply the cognition taking on the [intentional] aspect of the object. In this second instance, the source of knowledge refers simply to cognition as intending the object, thus ignoring the character of cognition as self-awareness, even though it is self-awareness that brings it forth. Why? Because the object as perceived [viz. the external object] is apprehended only by means of this [intentional aspect]. Thus, in whatever way the object may be apprehended, for instance as something white or non-white, it is an object in that form [viz., as intended] that is thus perceived.21 20 21

For a summary of the various arguments in this debate for and against using light as an analogy for cognitive awareness, see Watson (2014). ps i 9 and psv ad cit.: dvyābhāsaṃ hi jñānam utpadyate svābhāsaṃ viṣayābhāsaṃ ca.  tasyobhayābhāsasya yat svasaṃvedanaṃ tat phalam. kiṃ kāraṇam. tadrūpo hy arthaniścayaḥ̣ yadā hi saviṣayaṃ jñānam arthaḥ, tadā svasaṃvedanānurūpaṃ arthaṃ pratipadyata iṣṭam aniṣṭaṃ vā. yadā tu bāhya evārthaḥ prameyaḥ, tadā viṣayābhāsataivāsya

Whose Consciousness?

133

As I have argued elsewhere (Coseru 2015a: 223; Coseru 2015b), Dignāga’s central claim here concerns the character of cognition, specifically its self-intimating aspect. Regardless of what they are about (e.g., real or imagined objects), conscious cognitions are such as to disclose their occurrent for-me-ness: they happen to me, they occur in my mental stream. This notion that cognition is self-presenting is simply meant to address the problem of anonymity. But it also captures the modality-specific nature of conscious apprehension: perceiving differs from remembering, which, in turn, differs from conceiving. It is only insofar as cognitive awareness is self-intimating that it is possible to discern whether an object, say an earthen pot, is thematized as such or is merely ready to hand. The second claim is that a determination of the object, that is, how the object appears in cognition, for instance, as something desirable or undesirable, also depends upon various dispositional factors. This second claim addresses the phenomenal character of experience, the fact that conscious mental states have a distinct (perhaps even proprietary) phenomenology. Can I be mistaken about the content and character of a given mental state? For Dignāga, we can only be mistaken about the object or content of cognition, specifically about the nature of the object of cognition. We cannot be mistaken about cognition itself: “Even conceptual cognition is self-reflexive in its mode of presentation, though it differs [from perception] with regard to its object, because that [viz., the object, is what] it conceptualizes.”22 Given that what we are most intimately acquainted with are mental states, being mistaken about an occurrent mental state would be akin to being mistaken that one is conscious. That we are not so mistaken is because certain mental states exhibit immunity to error through misidentification:23 I can be mistaken about what I am experiencing, but not that I am having an experience. In shoring up his defense of the self-intimating character of cognition, Dignāga is also affirming the close link between consciousness and self-consciousness. However, his distinctly proto-phenomenological stance does not entail that conscious mental

22 23

pramāṇaṃ tadā hi jñānasvasaṃvedyam api svarūpam anapekṣyārthābhāsataivāsya pramāṇam. yasmāt so ’rthaḥ tena mīyate yathā yathā hy arthākāro jñāne pratibhāti śubhāśubhāditvena, tattadrūpaḥ sa viṣayaḥ pramīyate (Sanskrit in Steinkellner 2005). Translation, slightly adjusted for consistency, per Hattori (1968: 28). Cf. Kellner (2010: 207–210). ps 1:7ab: kalpanāpi svasaṃvittāv iṣṭā nārthe vikalpanāt. See also Yao (2004: 64) for a discussion of the same idea in the Nyāyamukha. In introducing the concept, Shoemaker (1968) also distinguishes between circumstantial immunity to error, as in instances where knowledge that one, say, faces a certain object is ordinarily assessed (by means of perception), and absolute immunity, as is the case with the identification of mental states as being “had” by a subject.

134

Coseru

states are completely knowable for the subject whose states they are. While the possibility of dwelling in a conscious mental state without any intentional content is presupposed by the thesis of omniscience, Dignāga’s immediate concern is pragmatic rather than soteriological: what relevant criteria must be in place for our mundane cognitions to succeed? As such, he addresses an issue that is at the very heart of phenomenological approaches to the problem of self-knowledge. As Zahavi observes: In my daily life, I am absorbed by and preoccupied with projects and objects in the world. Thus, pervasive pre-reflective self-consciousness is definitely not identical with total self-comprehension, but can rather be likened to a pre-comprehension that allows for a subsequent reflection and thematization. One should consequently distinguish between the claim that our consciousness is characterized by an immediate self-givenness and the claim that consciousness is characterized by total self-knowledge. One can easily accept the first and reject the latter, that is, one can argue in favor of the existence of a pervasive self-­ consciousness and still take self-comprehension to be an infinite task. (Zahavi 2005: 22f.) The reflexivity of awareness that svasaṃvedana is meant to capture is simply a condition for the possibility of self-knowledge. The mistake that critics of reflexivism make is to assume that this self-intimating or self-reflexive dimension of consciousness is the achievement of self-knowledge when in fact it is merely the condition for its possibility. In one of the most influential such critiques, Garfield writes: It may well be that the phenomenological project as prosecuted by Dignāga and Husserl, and as resurrected by Coseru and Zahavi, may be misguided for a simple reason: There may be nothing that it is like for me to see red, because I don’t. Instead of a single locus of consciousness contemplating a distinct world of objects – like a Wittgensteinian eye in the visual field or a Kantian transcendental ego – to be a person, from a Buddhist perspective, is to be a continuum of multiple, interacting sensory, motor and cognitive states and processes … My own access to them is mediated by my ideology, my narrative and a set of fallible introspectible mechanisms. (Garfield 2015: 209) As Garfield – echoing here Candrakīrti and fellow critics of svasaṃvedana, as well as contemporary champions of Higher Order Thought theories – would

Whose Consciousness?

135

have it, mental states lack any distinct phenomenal character of their own, so access to them is always mediated in some fashion or another. Once again, let me be clear that what is at stake in this debate is not the achievement of self-knowledge, which – and here I agree with Zahavi – may indeed turn out to be an “infinite task.” The issue rather is that self-knowledge comes bound up with a distinctively phenomenal character that tells the knower that the pain, the seeing of blue, the childhood memory are hers, that they occur in her mental stream.24 Even reflective and inferential modes of knowledge presuppose this tacit self-acquaintance. Otherwise, we confront the circularity implicit in the presupposition that specifying the content of reflective modes of self-­consciousness demands mastery of the concepts the reflective experience entails. Mastering such concepts, however, requires that there be some way of telling the specific mental state I happen to be in at the time: I can only ponder, judge or entertain that which is already present to me in thought.25 Dignāga’s attempt to frame the problem of self-knowledge in terms of the  distinctive character of cognition leaves open the question whether svasaṃvedana should be understood in intentional or representational term. The dual-aspect language is strongly suggestive of representationalism, which explains why debates arose in later Sautrāntika-Yogācāra circles about whether cognition (i) is aspectual or imagistic (sākāravāda), or (ii) lacks such representational features (nirākāravāda), or (iii) possesses such imagistic features only falsely (alīkākāravāda).26 Dharmakīrti’s elaboration of the argument for svasaṃvedana in terms of a constraint (niyama) that objects are known “only as apprehended” (sahopalambha), and Śāntarakṣita’s account of self-­awareness in terms of sentience come largely in response to the challenges that dual-­ aspect theories of consciousness face.

24

Sharf (Chapter 1), drawing on Saṃghabhadra’s discussion of perceptual objects in the Nyāyānusāra, frames the discussion in terms of the difference between conceptual and non-conceptual content: the “seeing of blue” is just a “raw feel” or a “quale” but one that lacks any discerning features, given that it has not yet come under the concept of “blue.” 25 The upshot is that we need nonconceptual content to make sense of conceptual accounts of self-knowledge, particularly as they find articulation in discussions about the relation between consciousness and content. I address this issue in detail elsewhere (Coseru 2020). See also Bermúdez (1998) for a compelling defense of the importance of nonconceptual content to solving the so-called paradox of self-consciousness. 26 See Dhammajoti (2007), Moriyama (2014), Kellner (2014), and McClintock (2014).

136 3

Coseru

Self-Awareness without Higher Order Thought

A central problem in contemporary discussions of consciousness is the problem of determining in precisely what a mental state’s being conscious consists. This problem raises a range of conceptual issues about the nature and structure of consciousness. Of particular importance here is the relation between consciousness and self-consciousness. The key question is: does consciousness imply self-consciousness or is self-consciousness the result of a higher-order cognition (a metacognition) co-occurrent with, and taking a previous instance of cognition as, its object? Solutions to this problem typically divide between those that take conscious cognition to be a “one-state” and those that regard it as a “two-state” process. One-state proponents argue that a mental state can be deemed conscious if and only if it possesses a specific character: it is reflexive or self-reflexive. This view, whose antecedents can be found in Descartes, Locke, and Kant, finds its clearest modern articulation in Brentano: “Every consciousness upon whatever object it is primarily directed is constantly directed upon itself” (Brentano 1982: 25). Following Brentano, the thesis of the unity of consciousness as reflexive awareness finds strong support in Husserl, Sartre, and many contemporary philosophers working in the phenomenological tradition. Critics of the reflexivity thesis in both India and the West have traditionally pointed to the conceptual problem of other minds and, more recently, to the findings of cognitive science to make the case about the fallibility of first-­ personal access to our own mental states. Proposing an alternative, two-state (or higher-order perception (hop)) conception of consciousness, they argue that a mental state is conscious in virtue of a distinct second-order state that is directed toward it. This latter group includes, among others, David Armstrong (1968, 1981/1997), David Rosenthal (1986, 2004), William Lycan (1996), and Rocco Gennaro (1996; 2006). Like the one-state model, the higher-order thought view too has antecedents in the Indian philosophical tradition. For instance, Mīmāṃsaka philosophers such as Kumārila, and notable Buddhist thinkers such as Candrakīrti and Śāntideva defend versions of the higher-order view. What Candrakīrti in particular takes issue with is the specifically imagistic position that the object of cognition is not extrinsic to cognition but is an aspect of cognition itself. Candrakīrti’s critique of reflexive awareness,27 then, targets this notion that there is a class of cognitive events that are essentially self-characterizing: they 27 See Prasannapadā and Madhyamakāvatāra, and discussions in Williams (1998), Garfield (2006) Arnold (2012), and Ganeri (2012).

Whose Consciousness?

137

reveal their own content without recourse to an additional instance of cognitive awareness, an object, or the positing of a subject of experience. More to the point, Candrakīrti rejects the notion that reflexive awareness has this unique property of giving access to the pure datum of experience.28 It is precisely with the intention of answering critics like Candrakīrti29 that Śāntarakṣita identifies the character of cognition as being contrary to insentient objects: “Cognitive awareness arises as something that is excluded from all insentient objects. This reflexive awareness of that cognition is none other than its non-insentience.”30 In effect, Śāntarakṣita simply follows Dharmakīrti’s critique of the physicalist (Cārvāka) claim that consciousness arises from the four material elements (Coseru 2017). Indeed, as Dharmakīrti maintains, if the four elements, or a special transformation thereof, are the ultimate basis of consciousness, then consciousness ought to arise whenever and wherever the elements occur, which is to say at any time and everywhere.31 Furthermore, even if consciousness were said to arise at a particular point in time and only in regions of space occupied by (or configured as) bodies, it cannot arise from something that is itself not sentient. The contemporary version of Candrakīrti’s view is that for a given mental state to be conscious a subject must have an appropriate higher-order belief, thought, or judgment that he or she is in that mental state. Take, for instance, Rosenthal’s influential defense of the higher-order view: “Conscious states are simply mental states we are conscious of being in. And, in general, our being conscious of something is just a matter of our having a thought of some sort about it” (Rosenthal 1986: 465). According to Rosenthal, for consciousness to be intelligible and analyzable, one must assume it to be an external, relational property of mental states, and to have something like an articulable structure. Here’s what I see as the main difficulty with this theory. Rosenthal argues that it is possible, in principle, that I be persuaded of my being angry through someone else’s testimony. Thus, I may realize that I am angry in the absence of 28

Arnold (2012) makes a compelling case that Candrakīrti’s uncompromising critique of the svasaṃvedana/svasaṃvitti thesis might be unwarranted if it is taken to show that selfawareness cannot be even conventionally, let alone ultimately, real. If svasaṃvitti is simply, with Śāntarakṣita, the mark of consciousness, rejecting it would in effect amount to denying “that we are conscious.” 29 Śāntarakṣita does not engage Candrakīrti directly, given that he does not become an influential thinker in Madhyamaka circles until the 11th century. His main Madhyamaka interlocutor is Bhāviveka. 30 ts 2000: vijñānaṃ jaḍarūpebhyo vyāvṛttaṃ upajāyate / iyam evā’tmasaṃvittir asya yājadarūpatā. 31 See pv ii, 35 and pvā loc. cit. Cf. Franco (1997: 171–172).

138

Coseru

any conscious feeling of anger, that is, without feeling angry much like a depressive might learn about her condition from a therapist without having any awareness of it. This example underscores the inferential conception of self-­ knowledge. On the higher-order theory, an individual must have a suitably unmediated higher-order thought about being in that state. But this higher-order state will not itself be conscious unless subject to another higher-order thought about it (thus leading to infinite regress). That means, a fairly large number of these higher-order thoughts are in fact unconscious. How exactly a series of unconscious cognitive events could generate conscious apprehension is not at all clear. As Dharmakīrti famously puts it in the Ascertainment of Knowledge (Pramāṇaviniścaya) (targeting the regress argument): Awaiting the end of a series of successive cognitions, a person does not comprehend any object, because cognition cannot be established as cognition when that cognition which is first-personally known (i.e. selfawareness) has not been first established. And since the arising of cognitions is without end, the whole world would be blind and deaf. For the series to come to an end, cognition’s own ascertainment and its apprehension of the aspect of the object must be concomitant.32 How, we might ask, can an unconscious mental state operate to confer consciousness on another unconscious mental state? In other words, if the hot theory claims that the thoughts required for consciousness can themselves be unconscious, we are owed an explanation of how the unconscious mental states can be a source for consciousness. Given these problematic issues, defenders of the higher-order view should not be allowed to gloss over the question of the phenomenal character of consciousness by assuming that consciousness owes its phenomenal character, indeed its very subjectivity, to an external or causal relation of some sort. Instead, the relational scenario ought to be unpacked in considerable detail so as to explain how it is possible for there to be such a thing as, for instance, pain that is unconscious or unknown until it becomes the object of a suitably cooccurrent cognition. On the view that I put forward here, “pain” is a phenomenal concept that can only be acquired experientially. Hence, talk of “unconscious pain,” as I noted above, amounts to committing a category mistake: specifically, using the language of subpersonal phenomena to refer to c­ onscious 32

pvin i: tan na tāvad ayaṃ puruṣaḥ kañcid arthaṃ pratyety upalambhaniṣṭhāṃ pratīkṣamāṇaḥ, ekāsiddhau sarvāsiddheḥ. na côpalambhānām utpattiniṣṭhety andhamūkaṃ jagat syāt. kvacin niṣṭhāyāṃ sa svayam ātmānaṃ viṣayākāraṃ ca yugapad upalabhata iti … (in Steinkellner 2007, pp. 41–42). Cf. Arnold (2012: 270).

Whose Consciousness?

139

mental states. Someone who has suffered no injury, discomfort, or distress or displays the rare condition of congenital analgesia cannot in principle grasp the concept of pain. Furthermore, even assuming that one can learn the concept of pain by definition does not entail that one grasps the property expressed by the concept (Nida-Rümelin 2007: 312). One need only point to conditions typically associated with various kinds of psychopathy and sociopathy to provide critical evidence for the relevance of phenomenal experience to phenomenal knowledge. These considerations should give pause and raise concerns that the higherorder theories – just like the relational stance at the heart of Madhyamaka dialectics, in view of their commitment to grounding consciousness in nonconscious mental (and even physical) states – are both more problematic and less equipped to handle analyses of phenomenal consciousness than one state theories. The representationalism that informs higher-order theories may be able to answer how consciousness comes to be about something in the absence of that something. But representationalism cannot tell us in a non-question begging way how consciousness comes to make its representational contents present and, in so doing, to be located in its occurrent for-me-ness. Specifically, these considerations invite us to go beyond traditional positions in metaphysics concerning the relation between mind and world, and corresponding debates in epistemology concerning externalist accounts of self-knowledge. 4

A Cognition Worthy of Its Name: Dharmakīrti’s Sahopalambhaniyama Argument

The tradition of epistemic inquiry that Dignāga helped to initiate rests on an important and somewhat radical claim regarding the relation between cognition and its result, between the cognitive event in question and its outcome: a reliable belief of some kind. As Dignāga notes, “a source of knowledge is something that performs an operation, but it is really effective only as a result.”33 Thus, the cognition that arises taking on the aspect of its object, while seemingly bound up with the intentional act itself, is nothing but the resulting apprehension: the act of knowing itself. In apprehending a given object, say an earthen pot, all that we can be certain about is that we are having a cognition of some kind: a visual or tactile experience. The aboutness of experience is a function of its intentional act: the earthen pot is only as seen, as handled, or as weighed. Similarly, the high-pitched sound is only as heard, and the judgment 33

ps i, 8: savyāpārapratītatvāt pramāṇaṃ phalam eva sat. Cf. Hattori (1968: 97).

140

Coseru

that numbers are real only as grasped under a Platonist metaphysics. Dignāga’s understanding of what counts as a reliable cognition is not unlike Husserl’s account of noematic content (e.g., the “perceived as such”). Access to the content as such, without recourse to ontological considerations about whether it is internal to cognition or external to it, is precisely what the phenomenological epoché is supposed to provide. Dharmakīrti’s argument for reflexive selfawareness (svasaṃvedana), then, is the closest we come to something like a proto-phenomenological account of noematic content. What is it about cognition that makes it worthy of its name? “Cognition” (jñāna), “discerning awareness” (vijñāna), and the warranted results (pramāṇaphala) of such epistemic practices (pramāṇa) depend on a complex vocabulary of phenomenal concepts (i.e., “concepts of experience”) to make them intelligible. How is it, then, that a reliable perception can be phenomenally distinguished from an instance of mistaken apprehension? That is, what are the criteria by means of which something can be veridically ascertained as “white” and “a lotus”? Whereas Dignāga thinks the only reliable basis for, say, perceptual knowledge is the presence of a nonconceptually contentful awareness (viṣayābhāsatā), Dharmakīrti instead argues that it is the seeming character of cognition itself that serves as a criterion for discerning what kind of cognition it is and of what. Bracketing considerations about the object at hand  – considerations inherited from his Sautrāntika and Yogācāra predecessors – Dharmakīrti’s sahopalambaniyama argument addresses the (otherwise quite trivial) concern that in seeking a basis for certainty, considerations about the epistemic character of cognition trump considerations about the ontological status of the object of cognition. There is no way to make sense of what we come to know without reference to how we know it: the lotus flower is only as seen, as drawn or as imagined; the childhood experience is only as remembered; the threat of losing a loved one is only as feared. Dharmakīrti’s egocentric predicament argument for reflexive self-awareness, then, does double duty: on the one hand, it makes explicit what the cognition’s relation to its object entails, and, on the other, it addresses concerns that, say, a Sautrāntika interlocutor might have about the consequences of removing external objects from considerations about the efficacy of our epistemic practices: Thus, “on account of there being no apprehension of [an object],” such as blue that lacks qualification (upādhi) – [specifically the] individuation (viśeṣana), which is “perception” or cognition – and “because there is apprehension of blue only when there is apprehension of that” – that is, because there is no awareness of blue and of its cognition [as blue] except

Whose Consciousness?

141

as co-occurrent (sahaiva) – it is established that “perception takes the form of blue” [stands for the view that it] just has the aspect (ākāra) of blue. On account of the requirement that the object be apprehended together with (sahopalambaniyama) the cognition, the so-called external (bāhya) object – for instance [a patch of] blue – does not differ from [its cognition], as in the case of apprehending the double moon [illusion, which no one takes to correspond to an external object].34 The central thesis at work in this extended version of the argument is that object apprehension is primarily the result of an individuation that occurs in the horizon of awareness itself. Phenomena such as “blue” appear as such only because of the operations of cognition in the relevant domain: in this case, that of color vision. In short, there is no such thing as “blue” apart from a blue quale. Rather, there are only experiential or self-presenting blue-continua. This way of framing the issue of the efficacy of our epistemic practices has consequences for the problem of self-knowledge. Specifically, it warns that postulating a basis for self-knowledge outside the structure of experience has the unfortunate outcome of assuming (mistakenly) that experience is an emergent property of something that is not itself experiential, but that nevertheless has the functional organization to support such experiential self-ascriptions. On the proposal outlined here, a solution to the question “What does selfknowledge contribute to our understanding of objects?” will demand that there be a class of primitive concepts, call them “phenomenal concepts,” that anchor our cognitive practices. Phenomenal concepts, then, are a special category of concepts uniquely suited to provide epistemic access to experience such that by virtue of possessing them a person can be said to have direct and infallible access to her mental states. Unlike the concepts of natural science, which mediate our knowledge of the external world, phenomenal concepts are the vehicle for phenomenal knowledge, that is, for knowledge of conscious experience. There is a long debate in contemporary analytic philosophy about whether the presence of phenomenal concepts shows that there is an epistemic gap between truths about the physical world and truths about phenomenal consciousness, and whether, given this epistemic gap, we can infer an ontological gap between mind and matter (or consciousness and the brain).35 34

35

pv 3. 335: tatra darśanena jñānenopādhinā viśeṣaṇena rahitasya nīlāder agrahāt tasya grahe ca nīlasya grahāt sahaiva nīladhiyor vedanād darśanaṃ nīlādinirbhāsaṃ nīlākaraṃ vyavasthitam. yad tāvan nīlādikaṃ bāhyam ity ucyate tad jñānena sahopalambhaniyamād tadabhinnasvabhāvaṃ dvicandrādivat. Cf. Ratié (2011: 347 n87). Proponents of the so-called “phenomenal concepts strategy” have argued that phenomenal concepts are in effect either recognitional (Loar 1990; Carruthers 2000; Tye 2000,

142

Coseru

Much of the debate, which reaches back to Frank Jackson’s (1982) now classic knowledge argument against physicalism, concerns whether or not phenomenal concepts can be deduced a priori from physical concepts. If they can, then their possession does not provide any additional knowledge beyond how the world seems to us (because, unlike physical, functional, and intentional concepts, they do not explain how the world must be in order for us to have the experiences that we do). If they cannot, then comprehensive knowledge of the physical domain cannot give us phenomenal knowledge: knowing all there is to know scientifically about color vision does not give us the knowledge of what it is like to see red. For the purpose of explaining Dharmakīrti’s sahopalambaniyama argument, I take phenomenal concepts to be concepts that apply to experience when we seek to explain why and how acquaintance with our own mental states differs from our grasp of external objects. Without the immediacy, fineness of grain, and epistemic stability of phenomenal concepts we cannot in principle report on the qualitative aspects of our experience. We may apply phenomenal concepts to experience directly, merely in virtue of having had some experience, say of seeing a white lotus, or indirectly, when recalling a past experience or making sense of someone else’s experience. But what sets phenomenal concepts apart is not how they relate to the experience in question (directly or indirectly), but rather that they convey the properties of the experience.36 Phenomenal concepts are what we need in order to provide description-level

36

2003; Carruthers and Vaille 2007) or indexical (Perry 2001 and O’Dea 2002) or informationtheoretic (Aydede and Güzeldere 2005) concepts of experience. As such they are said to provide causal, demonstrative, or informational access to experience in a manner that both bridges the explanatory gap and is compatible with physicalism. The problem with this strategy is the assumption that we need not invoke the nature of phenomenal consciousness itself in order to account for the special knowledge that phenomenal concepts provide (cf. Balog 2009: 301–302). But by not invoking the nature of phenomenal consciousness, proponents of this strategy – as Chalmers (2007) has convincingly argued – are confronted, inter alia, with a dilemma: either a scenario physically indistinguishable from ours that misses the physicalist’s account of phenomenal concepts is conceivable, in which case we cannot provide a physicalist explanation of how phenomenal concepts relate to phenomenal consciousness, or such a scenario is not conceivable, in which case the physicalist’s account of phenomenal concepts cannot explain our epistemic situation (cf. Balog 2009: 309). See also Alter and Walter (2007), and contributions therein. In this case, the property of having had an experience as if of a white lotus. In a defense of the reflexivity of awareness thesis, Jenzen (2008: 80f) offers an interesting argument “from spontaneous reportability,” which makes the case that in order for us to report on the objects of some experience, say the white lotus, we must be able to report on having had an experience since objects are known only as experienced.

Whose Consciousness?

143

a­ nalyses of experiential content,37 to tell what it is like to see red, smell roses, or entertain I-thoughts. As a phenomenal concept, “reflexive self-­consciousness” (svasaṃvedana, svasaṃvitti) designates that dimension of awareness that provides first-personal grounding for the apprehension of objects. Consider the function of phenomenal concepts when applied to perception. An object as perceived, thus, is such that it makes its causal efficacy present by occasioning different types of experience. For example, water causes the experience of wetness and fluidity, rocks the experience of resistance or hardness, and irregular surfaces the experience of roughness. Hence, our effective deployment of such concepts as “fluidity,” “resistance,” and “roughness” are instantiated by these phenomenal experiences. To invoke the special relation of phenomenal concepts to phenomenal consciousness, however, is not to deny the natural world. The object that is perceived, although unknowable apart from the perceptual occasion, does constrain the phenomenal concept’s referential and intentional properties. A wall limits movement, night restricts vision, and inattention renders us “blind” to salient features of our surroundings. In Dharmakīrti’s example above, the aspect that perception takes, which makes it seem “as if” of a blue object, is not a function of its representational content but a modality of its perceptual apprehension. We see a blue sky not because some determinate object gets represented as “sky” and as having the property of “blueness,” but rather because our visual system has adapted to seeing certain frequencies of light as blue. As a color quale “blue” can only be apprehended as such by a visual system directed by visual awareness. For this reason – argues Dharmakīrti in a more succinct rendition of the argument (this time in the Pramāṇaviniścaya) – “Blue and the cognition [by means of which something comes to be known] as such are not different because they are necessarily perceived together.”38 The consequence of this view, and a corollary to this thesis, is that, as Dharmakīrti explicitly states (concluding his verse): “The seeing of the object is not established with respect to someone unacquainted with [its] perception.”39 There is no apprehension of an object without an implicit awareness of its mode of apprehension. ­Reflexivity, 37 38

39

Chalmers (2011) applies this strategy in mitigating the distinction between questions of fact and questions of language, concluding that our verbal disagreements should not call into question facts about a certain domain, only their description. pvin i 54ab: sahopalambhaniyamād abhedo nīlataddhiyoḥ (in Steinkellner 2007: 40). A detailed genealogy of this widely cited verse is found in Watson (2010: 305 n26). For alternative translations and analyses, see Iwata (1991: 66–109), Dreyfus and Lindtner (1989: 47), Taber (2010: 292f), Kellner (2011: 419), and Arnold (2012: 267). pvin i, 54cd: apratyakṣopalambhasya nārthadṛṣṭiḥ prasiddhyati (in Steinkellner 2007: 40).

144

Coseru

then, is simply a condition for the possibility of effective conscious cognition, not the self-knowledge that is gained while pursuing the various moral and existential goods that Buddhists ultimately seek. The metaphysical impulse that drives explanatory concerns about self-knowledge is rooted in an unjustified demand: that the problem of self-knowledge can only be addressed if the objects, situations, and events that occasion it are abstracted from livedexperience. But the reductionist framework of Buddhist metaphysics, with its anti-realist stance on “the self,” seems ill-equipped to handle the subjective features of lived-experience, thus creating seemingly insurmountable problems for externalist accounts of self-knowledge. 5 Self-Knowledge and the Conundrum of Metaphysical Interpretations There are substantive disagreements about how the problem of self-knowledge should be framed, the kind of evidence that is deemed reliable, and the lines of justification that are worth pursuing. It can be (and has been) framed in both metaphysical and epistemological terms, drawing on both dialectical and epistemic accounts of what it is that we thus come to know, and taking the form of both conceptual analysis and ostensive demonstration. Nāgārjuna’s dialectical stance, which takes the essencelessness (niḥsvabhāvatā) implicit in the relational view of emptiness to be the defining characteristic of all phenomena, targets specifically the metaphysical stance of Abhidharma Reductionism: namely, that for an entity to be real it must have svabhāva (regardless of whether svabhāva stands for the Sarvāstivāda conception of “own nature” or the Sautrāntika idea of a “defining characteristic”). With Asaṅga and Vasubandhu, this dialectical move is reworked into a critique of the relational view as reflecting an internal dialectic of cognitive distortions. Following Dignāga and Dharmakīrti, commitment to a conception of epistemically warranted subjectivity becomes the norm for Yogācāra conceptions of self-knowledge. This line of thought, in turn, faces a robust challenge from later Mādhyamikas such as Candrakīrti and Śāntideva, whose antirealist stance about phenomenal character blocks any conception of self-cognition as intimating or self-­illuminating (svaprakāśa).40 These debates, and their offshoots in Tibet and East Asia during 40

Arnold (Chapter 5) thinks that there is “no good reason for a Mādhyamika to resist the idea of self-awareness,” at least as Śāntarakṣita understands it, that is, as “nothing more than that we are conscious.” Simply put, consciousness ought to have a place in our conventionally reliable account of what there is even for the Mādhyamika. The Mādhyamika’s

Whose Consciousness?

145

the second millennium, inform a thriving analytic philosophy of mind that has endured in Buddhist intellectual circles to this day. Is there a tension between the no-self view and the notion that cognition of an object is possible only in virtue of cognition’s own self-specifying features? In other words, assuming a realist ontology of unique particulars that can be perceptually apprehended, what are the specific criteria for dissociating between self-knowledge and knowledge of external objects? I want to claim that such a tension arises only if self-knowledge is framed primarily as a problem about grounding, specifically about the nature of the objects, properties, and relations that are constitutive of self-knowledge. As a problem in the metaphysics of mind, self-knowledge concerns the subject-object relation or how consciousness in its various guises can relate to itself as an object. The assumption is that self-consciousness presupposes no prior acquaintance with the mental state one happens to be in. Rather, one comes to have self-­knowledge by recognizing oneself as the object of a particular mental state of some kind (e.g., of perception, memory, or introspective awareness). If this object (a self or subject) can be dissolved under analysis – given the Abhidharma ontological distinction between primary and secondary existents –, then it is not something that exists ultimately (dravyasat) (Cox 1995: 138f). Primary existents, unlike secondary existents, are irreducible and thus serve as the constitutive elements of what there is. On this ontological picture, then, ultimately only partless atoms and partless moments of consciousness exist; enduring selves and chariots derive their existence from these more basic elements and reflect our linguistic and conceptual practices (Williams 1981: 240). The grounding project thus embroils us in questions about exactly how these partless atoms and partless moments of consciousness ultimately exist, and about the relations that obtain both among and between primary and secondary existents. Combined with the principle of momentariness, which stipulates that these units of matter and experience are in effect very short concern, however, is with the kind of awareness svasaṃvitti claims to be, specifically a special type that makes itself present or known in the process of revealing its object or content. Since the Mādhyamika is blocked by the rejection of svabhāva from admitting that anything could possesses its own determination or intrinsic character, he cannot commit to the idea of a self-revealing or “autonomously intelligible” cognition. While I am in general agreement with Arnold that Mādhyamikas have no good reason to reject the notion of svasaṃvitti as entailing that consciousness is implicitly self-aware or self-revealing, the question whether this self-awareness also entails a privileged way of knowing, I think, unnecessarily complicates what in effect is a basic phenomenological stance: that experiences are what they are in virtue not merely of their content but also their character.

146

Coseru

lived, questions about the object of self-knowledge ultimately lead to considerations about causality and the nature of the dependency relation between partless atoms and partless moments of consciousness. Since the Abhidharma project has no place for the idea of a first cause, the only way to make sense of the arising and dissolution of phenomena is to say that they do so as manifestations of a beginningless and bottomless causal web. The problem of selfknowledge, then, becomes the problem of charting the order of causal events to explain how an awareness comes to have properties of some sort, say visual, because of its dependence upon a visual system and the reflectance properties of a given object. As a key taxonomical concept at the heart of Abhidharma, svabhāva stands for either “intrinsic nature” or for a “distinguishing feature” or “characteristic” of phenomena. Applied to cognition, the notion that mental states have distinguishing features or properties of their own yields a range of positions in epistemology, from naïve realism to representationalism. Take hearing: either there is something intrinsic about the auditory system that furnishes it with phenomenal content over and above the sound’s represented features or what we call “hearing” is just a set of features such as pitch, timber, and degree of acoustic acuity, that are represented as externally (perhaps spatio-temporally) mapped.41 One example is the difference between hearing a sentence in a language one does not understand, such as “gataṃ na gamyate”, and hearing it after one has learned the language in this case, Sanskrit. Clearly, what sets the two instances apart – that is, what marks their phenomenal contrast – is that although they involve the exact same auditory input, the second comes bound up with a rich phenomenal content. How far this phenomenal content reaches into the structure of cognition is precisely what the Yogācāras want to examine when they affirm the ultimate irreducibility of mental properties. That such self-specifying thresholds of experience as phenomenal contrast should exist and be discoverable is precisely what the Madhyamaka project calls into question when it argues that our tendency to see things in terms of their defining properties (that is, in terms of svabhāva) is the result of a fundamentally mistaken superimposition.

41

In drawing an analogous distinction, Bayne (2009) argues that the line between low-level and high-level perceptual content is rather blurry, and that it is doubtful that phenomenal content only becomes manifest at a certain threshold, or that we could talk about phenomenal content at all without considering such representational features as edge, corner, shade, acoustic acuity, etc. How far into perception phenomenal content reaches is highly debated, but various proposals for moving this investigation forward invoke, for instance, contrast arguments in order to make the case for something like a distinct cognitive phenomenology that combines these two kinds of content.

Whose Consciousness?

147

Thus, when approached from the standpoint of the two truths dialectic, it is not hard to see why the problem of self-knowledge becomes a problem of adjudicating between realist, idealist, and anti-realist positions in the metaphysics of mind. In this case, the solution is always articulated in terms of a “revisionary metaphysics” (as Claus Oetke (2003: 470) calls it, borrowing Peter Strawson’s distinction between “descriptive” and “revisionary” metaphysics) since a conception of consciousness as self-intimating cannot arise in relation to things that are impermanent or lack shareable features.42 For something to be defined as the locus of a specific property such as self-intimation–the argument goes– there needs to be some corroborative evidence beyond the conscious mental state itself. 6 Conclusion: Consciousness, Naturalism, and the Bounds of the Lebenswelt It should be obvious by now that framing self-knowledge as a problem in the metaphysics of mind creates more problems than it solves. First, questions about the content of consciousness become questions about ontological constitution: are the objects so apprehended real, are they merely representations, or are they entirely internal to the mind? Second, the possibility of treating such content as simply intentional content is blocked by the more pressing need to address the problem of reference, specifically the problem whether uses of the first-person pronoun refer to a subject or owner of experience. Third, the difference between action, object, and agent or cognition, cognized, and cognizer (pramātṛ, prameya, and pramiti) becomes a difference in modes of existence rather than a way of specifying the constitutive elements of a knowledge event, its modality, and its epistemic status. Just as pursuing questions in the metaphysics of mind by asking – with a view to closing the explanatory gap between mind and world – how consciousness and cognition can emerge from what are presumed to be unconscious and non-­cognitive ­elements and processes, so

42

This point is amply made by Āryadeva in cś 10.3 with his ingenious argument that, if something is real (say, the property that fires have to generate heat), it is so for all tokens of the type. The same cannot be said for selves, given that “what is your self is my nonself” and hence what one apprehends as my own self is not a universally shared property like “heat,” which is experienced by all those exposed to fire. Siderits (Chapter 6) appeals to Āryadeva’s argument to make the case that attribution of mentality must require criteria for attribution: if we employ criteria (of the inferential sort) to attribute states of consciousness to others, why should the establishment of self-consciousness be exempt from such criteria?

148

Coseru

also doing epistemology in a metaphysical key obscures in a fundamental way what is at stake in the debate on self-knowledge. What is at stake in these debates is not just the possibility of overcoming ignorance, but the very capacity to even entertain this possibility, let alone engage in the sorts of self-reflective and self-scrutinizing activities that are presupposed by substantive accounts of self-knowledge. According to one tradition of thought, the self-knowledge project is deeply embedded in the Abhidharma metaphysics: figuring out what phenomena reduce to, and what is ultimately real, serves as a proxy for self-knowledge. Whether Abhidharma Reductionism is eliminativist or not, then, remains an open question whose answers depend on whatever the no-self doctrine is taken to entail: dispelling the illusion of a permanent, substantive self, or dispensing with any talk of subjective experience altogether. But according to the view I defend here, reductionism must give way to phenomenological reduction for the project of self-knowledge to get off the ground. For without an understanding of the distinctively self-intimating character of experience, identifying, analyzing, and cultivating the range of presumably impersonal mental factors that Buddhists deem wholesome (and countering those they deem unwholesome), would have no way to proceed. Bibliography Aiyaswami Sastri, N. 1975. Satyasiddhiśāstra of Harivarman. Vol. 1. Baroda: Gaekwad’s Oriental Series. Alter, T. and Walter, S. 2007. Phenomenal Concepts and Phenomenal Knowledge: New Essays on Consciousness and Physicalism. New York: Oxford University Press. Anscombe, Elizabeth. 1981. “The First Person,” In Collected Papers, vol. i. Metaphysics and the Philosophy of Mind. Minneapolis: University of Minnesota Press. Armstrong, D. 1968. A Materialist Theory of Mind, London: Routledge. Armstrong, D. 1981/1997. “What is consciousness?” In Block, N., Flanagan, O. & Güzeldere, G. (eds.) The Nature of Consciousness: Philosophical Debates, pp. 721–728, Cambridge, MA: mit Press. Armstrong, D. and N. Malcolm. 1984. Consciousness and Causality, Oxford: Blackwell. Arnold, Dan. 2010. “Self-awareness (svasaṃvitti) and related doctrines of Buddhists following Dignāga: Philosophical characterizations of some of the main issues,” Journal of Indian Philosophy, 38: 323–378. Arnold, Dan. 2012. Brains, Buddhas, and Believing: The Problem of Intentionality in Classical Buddhist and Cognitive-Scientific Philosophy of Mind. New York: Columbia University Press.

Whose Consciousness?

149

Aydede, M., and Güzeldere, G. 2005. “Cognitive Architecture, Concepts, and Introspection: An Information-theoretic Solution to the Problem of Phenomenal Consciousness,” Noûs, 39: 197–255. Balog, Katalin 2009. “Phenomenal Concepts,” In Brian McLaughlin, Ansgar Beckermann & Sven Walter (eds.), Oxford Handbook in the Philosophy of Mind. Oxford University Press. pp. 292–312. Bayne, Tim. 2009. “Perceptual experience and the reach of phenomenal content,” Philosophical Quarterly, 59 (236): 385–404. Bermúdez, Jose Luis. 1998. The Paradox of Self-Consciousness. Cambridge, MA: mit Press. Block, N. 2012. “The grain of vision and the grain of attention,” Thought: A Journal of Philosophy 1 (3): 170–184. Brentano, F. 1874/1973. Psychology from an Empirical Standpoint, London: Routledge and Kegan Paul. Brentano, F. 1982. Descriptive Psychology, Muller, B. (ed. and trans.), London: Routledge. Carpenter, A. 2015. “Persons Keeping Their Karma Together The Reasons for the Pudgalavāda in Early Buddhism”. In K. Tanaka, Y Deguchi, J. L. Garfield, and G. Priest, eds. The Moon Points Back, 1–44. New York: Oxford University Press. Carruthers, Peter. 2000. Phenomenal Consciousness. Cambridge: Cambridge University Press. Carruthers, Peter and Veillet, Bénédicte. 2007. “The Phenomenal Concept Strategy,” Journal of Consciousness Studies, 14 (9–10): 212–236. Chalmers, David. 1996. The Conscious Mind: In Search of a Fundamental Theory. New York: Oxford University Press. Chalmers, David. 2007. “Phenomenal Concepts and the Explanatory Gap.” In Phenomenal Concepts and Phenomenal Knowledge: New Essays on Consciousness and Physicalism, edited by Torin Alter and Sven Walter, 167–194. New York: Oxford University Press. Coseru, Christian. 2012. Perceiving Reality: Consciousness, Intentionality, and Cognition in Buddhist Philosophy. New York: Oxford University Press. Coseru, Christian. 2015a. “Taking the Intentionality of Perception Seriously: Why Phenomenology is Inescapable.” Philosophy East and West 65 (3): 227–248. Coseru, Christian. 2015b. “Perception, Causally Efficacious Particulars, and the Range of Phenomenal Consciousness,” Journal of Consciousness Studies 22 (9–10): 55–82. Coseru, Christian. 2017. “Consciousness and Causal Emergence: Śāntarakṣita Against Physicalism,” in Oxford Handbook of Indian Philosophy, edited by Jonardon Ganeri, 360–378. Oxford: Oxford University Press. Coseru, Christian. 2020. “Reasons and Conscious Persons.” In Derek Parfit’s Reasons and Persons: An Introduction and Critical Inquiry, edited by Andrea Sauchelli, 160–186. London: Routledge.

150

Coseru

Cox, Collett. 1995. Disputed Dharmas: Early Buddhist Theories on Existence. Tokyo: International Institute of Buddhist Studies. Dantinne, Jean. 1980. Le traité des cinq agrégats: Pañcaskandhaprakaraṇa de Vasuban­ dhu. Bruxelles : Institut Belge des Hautes Études Bouddhiques. de la Vallée Poussin, Louis 1970. Madhyamakāvatāra par Candrakīrti, Bibliotheca Buddhica ix, Osnabrück: Biblio Verlag. Dhammajoti, K.L. 2007. “Ākāra and Direct Perception.” Pacific World Journal 3 (9): 245–272. Dreyfus, G. and C. Lindtner. 1989. “The Yogācāra Philosophy of Dignāga and Dharmakīrti.” Studies in Central and South East Asian Religions 2: 27–52. Evans, Gareth. 1981. The Varieties of Reference. Oxford: Oxford University Press. Franco, Eli. 1993. 1993. “Did Dignāga Accept Four Types of Perception?” Journal of Indian Philosophy 21: 295–299. Franco, Eli. 1997. Dharmakīrti on Compassion and Rebirth. Vienna: Wiener Studien zur Tibetologie und Buddhismuskunde. Gallagher, S. and Zahavi, D. 2012. The Phenomenological Mind. 2nd ed. London: Routledge. Ganeri, Jonardon, 2012. The Self: Naturalism, Consciousness, and the First-Person Stance, Oxford: Oxford University Press. Garfield, Jay L. 2015. Engaging Buddhism: Why It Matters to Philosophy, New York: Oxford University Press. Gennaro, R. 1996. Consciousness and Self-Consciousness: A Defense of the Higher-Order Thought Theory of Consciousness, Amsterdam: John Benjamins. Gennaro, Rocco. 2012. The Consciousness Paradox: Consciousness, Concepts, and Higher-Order Thoughts. Cambridge, MA: mit Press. Heidegger, Martin. 1994. Gesamtausgabe Band 61: Phänomenologische Interpretationen zu Aristoteles. Einführung in die phänomenologische Forschung. Frankfurt am Main: Vittorio Klostermann. Hattori, M. 1968. Dignāga, On Perception, Cambridge, MA: Harvard University Press. Husserl, E. 1983. Ideas Pertaining to a Pure Phenomenology and to a Phenomenological Philosophy, Vol. i, Kersten, F. (trans.), Dordrecht: Kluwer Academic Publishers. Iwata, T. 1991. “On the Classification of Three Kinds of Reason in Pramāṇaviniścaya iii – Reduction of Reasons to svabhāvahetu and kāryahetu.” In Studies in the Buddhist Epistemological Tradition: Proceedings of the Second International Dharmakīrti Conference, Vienna, June 11–16, 1989, edited by Ernst Steinkellner, 85–96. Vienna: Verlag der Österreichischen Akademie der Wissenschaften. Ichigō, Masamichi. 1985. Madhyamakālaṃkāra of Śāntarakṣita with His Own Commentary or Vṛtti and with the Subcommentary or Pañjikā of Kamalaśīla. Kyoto: Buneido. Janzen, G. 2008. The Reflexive Nature of Consciousness, Amsterdam and Philadelphia, PA: John Benjamins.

Whose Consciousness?

151

Keijzer, F. & Schouten, M. 2007. “Embedded cognition and mental causation: Setting empirical bounds on metaphysics.” Synthese 158 (1): 109–125. Kellner, Birgit. 2010. “Self-Awareness (Svasaṃvedana) in Dignāga’s Pramāṇasamuccaya and -Vṛtti: A Close Reading.” Journal of Indian Philosophy 38 (3): 203–231. Kellner, Birgit. 2011. “Self-Awareness (Svasaṃvedana) and Infinite Regresses: A Comparison of Arguments by Dignāga and Dharmakīrti.” Journal of Indian Philosophy 39 (4-5): 411–426. Kellner, Birgit. 2014. “Changing Frames in Buddhist Thought: The Concept of Ākāra in Abhidharma and in Buddhist Epistemological Analysis,” Journal of Indian Philosophy 42 (2): 275–295. Kobayashi, Hisayasu 2010. “Self-Awareness and Mental Perception.” Journal of Indian Philosophy 38 (3): 233–245. Kouider, S,, and Dehaene, S. 2007. “Levels of Processing during Non-Conscious Perception: A Critical Review of Visual Masking.” Philosophical Transactions of the Royal Society B: Biological Sciences, April: 857–875. Kramer, Jowita. 2008. “On Sthiramati’s Pañca-skandhaka-vibhāṣā: a Preliminary Survey.” Nagoya Studies in Indian Culture and Buddhism: Saṃbhāṣā 27: 149–171. Kriegel, Uriah. 2009. Subjective Consciousness: A Self-Representational Theory. New York: Oxford University Press. Lamotte, Étienne. 1938. La Somme du Grand Véhicule d’Asaṅga (Mahāyānasaṃgraha). Vols. i–ii. Louvain-la-Neuve: Université de Louvain. Loar, B. 1990. “Phenomenal states,” Philosophical Perspectives, 4, pp. 81–108. Locke, John. 1975. An Essay Concerning Human Understanding. Oxford: Clarendon Press. Lycan, W.G. 1996. Consciousness and Experience, Cambridge, MA: mit Press. Matilal, Bimal K. 1986. Perception: An Essay on Classical Indian Theories of Knowledge, Oxford: Oxford University Press. MacKenzie, M. 2007. “The illumination of consciousness: Approaches to self- a­ wareness in Indian and western traditions” Philosophy East and West, 57 (1): 40–62. Marcel, A J. 1983. “Conscious and Unconscious Perception: Experiments on Visual Masking and Word Recognition.” Cognitive Psychology 15 (2): 197–237. McClintock, Sara. 2014. “Kamalaśīla on the Nature of Phenomenal Content (ākāra) in Cognition: A Close Reading of tsp ad ts 3626 and Related Passages,” Journal of Indian Philosophy 42 (2): 327–337. Moriyama, Shinya. 2014. “Ratnākaraśānti’s Theory of Cognition with False Mental Images (*alīkākāravāda) and the Neither-One-Nor-Many Argument,” Journal of Indian Philosophy 42 (2): 339–351. Nagatomi, M. 1979. “Mānasa-pratyakṣa. A Conundrum in the Buddhist Pramāṇa System.” In Sanskrit and Indian Studies: Essays in Honour of Daniel H.H. Ingalls, edited by M. Nagatomi, B.K. Matilal, J.M. Masson, and E. Dimock, 243–260. Dordrecht: Reidel.

152

Coseru

Nārada, Maha Thera. 1979. A Manual of Abhidharma Being Abhidhammattha Saṅghata of Bhadanta Anuruddhācariya. London: Pali Text Society. Nida-Rümelin, M. 2007. “Grasping phenomenal properties”, in Alter, T. & Walter, S. (eds.) Phenomenal Concepts and Phenomenal Knowledge, pp. 307–336, New York: Oxford University Press. O’Dea, J. 2002. “The indexical nature of sensory concepts,” Philosophical Papers, 31 (2), pp. 169–181. Oetke, Claus. 2003. “Some Remarks on Theses and Philosophical Positions in Early Madhyamaka.” Journal of Indian Philosophy 31:449–478. Overgaard, S. 2007. Wittgenstein and Other Minds: Rethinking Subjectivity and Intersubjectivity with Wittgenstein, Levinas, and Husserl. London: Routledge. Overgaard, S. Fehl, K., Mouridsen, K., Bergholt, B., and Cleeremans, A. 2008. “Seeing without seeing? Degraded conscious vision in a blindsight patient” PLoS ONE 3(8): e3028. doi:10.1371/journal.pone.0003028. Perry, Ralph B. 1910. “The Egocentric Predicament,” Journal of Philosophy 7: 5–14. Perry, J. 2001. Knowledge, Possibility and Consciousness. Cambridge, Mass.: mit Press. Phillips, I. 2016. “Consciousness and Criterion: On Block’s Case for Unconscious Seeing,” Philosophy and Phenomenological Research 93 (2): 419–451. Pradhan, P. 1975. Abhidharmakośabhāṣyam, revised 2nd ed., Haldar, A. (ed.), Patna: Kashi Prasad Jayaswal Research Institute. Ratié, Isabelle. 2011. Le Soi et l’Autre: Identité, différence et alterité dans la philosophie de la Pratyabhijñā. Leiden: Brill. Renz, Ursula. 2017. Self-Knowledge: A History. New York: Oxford University Press. Ronkin, Noa. 2005. Early Buddhist Metaphysics: The Making of a Philosophical Tradition. London and New York: Routledge-Curzon. Rosenthal, David. 1986. “Two concepts of consciousness,” Philosophical Studies, 49: 329–359. Rosenthal, David. 2004. “Varieties of higher-order theory,” in Gennaro, R. (ed.) HigherOrder Theories of Consciousness: An Anthology, pp. 17–44, Amsterdam: John Benjamins. Shastri, Swami Dwarikadas. 1968. Tattvasaṃgraha of Achārya Shāntarakṣita with the Commentary “Pañjikā” of Shrī Kamalaśīla. Vols. 1–2. Vārāṇasī: Bauddha Bhāratī. Shoemaker, Sydney. 1968. “Self-reference and self-awareness.” Journal of Philosophy 65 (19): 555–567. Shoemaker, S. 1996. The First-Person Perspective and Other Essays. Cambridge: Cambridge University Press. Steinkellner, Ernst. 2005. Dignāga’s Pramāṇasamuccaya, Chapter 1: A Hypothetical Reconstruction of the Sanskrit Text with the Help of the Two Tibetan Translations on the Basis of the Hitherto Known Sanskrit Fragments and the Linguistic Materials Gained from Jinendrabuddhi’s ṭīkā. http://ikga.oeaw.ac.at/Mat/dignaga_PS_1.pdf (last accessed February 29, 2016).

Whose Consciousness?

153

Steinkellner, E. 2007. Dharmakīrti’s Pramāṇaviniścaya, Chapters 1 and 2. Beijing: China Tibetology Publishing House; Vienna: Austrian Academy of Sciences Press. Strawson, Peter F. 1992. “Freedom and Necessity,” In Analysis and Metaphysics. Oxford: Oxford University Press. Taber, John. 2010. “Kumārila’s Buddhist.” Journal of Indian Philosophy 38: 279–296. Tye, M. 2000. Color, Consciousness, and Content. Cambridge, MA: mit Press. Tye, M. 2003. “A theory of phenomenal concepts,” in Minds and Persons, ed. A. O’Hear Cambridge, MA: Cambridge University Press. Waldron, William. 2003. The Buddhist Unconscious: The Ālaya-vijñāna in the Context of Indian Buddhist Thought. London: Routledge. Watson, Alex. 2010. “Bhaṭṭa Rāmakaṇṭha’s Elaboration of Self-Awareness (Svasaṃvedana), and How It Differs From Dharmakīrti’s Exposition of the Concept.” Journal of Indian Philosophy 38 (3): 297–321. Watson, Alex 2014. “Light as an Analogy for Cognition in Buddhist Idealism (Vijñānavāda)” Journal of Indian Philosophy 42 (2-3): 401–421. Wayman, Alex. 1991. “Dharmakīrti and the Yogācāra theory of Bīja.” In Studies in the Buddhist Epistemological Tradition: Proceedings of the Second International Dharmakīrti Conference, Vienna, June 11–16, 1989, edited by Ernst Steinkellner, 419– 430. Vienna: Verlag der Österreichischen Akademie der Wissenschaften. Williams, Paul. 1981. “On the Abhidharma Ontology,” Journal of Indian Philosophy, 9: 227–257. Williams, Paul. 1998. The Reflexive Nature of Awareness: A Tibetan Madhyamaka Defense. London: Curzon Press. Wittgenstein, Ludwig. 2001. Philosophical Investigation. 3rd ed. Translated by G.E.M. Anscombe. Oxford: Blackwell. Wright, Crispin, Barry C. Smith, and Cynthia Macdonald, eds. 1998. Knowing Our Own Minds. Oxford: Clarendon Press. Yao, Zhihua. 2004. “Dignāga and the Four Types of Perception.” Journal of Indian Philosophy 32: 57–74. Yao, Zhihua. 2005. The Buddhist Theory of Self-Cognition. London and New York: Routledge. Zahavi, Dan. 2005. Subjectivity and Selfhood. Cambridge, MA: mit Press. Zahavi, D. and Kriegel, U. 2016. “For-me-ness: What it is and what it is not.” In Philosophy of Mind and Phenomenology: Conceptual and Empirical Approaches, edited by D.O. Dahlstrom, A. Elpidorou and W. Hopp, 36–55. London: Routledge.

Chapter 5

Should Mādhyamikas Refute Subjectivity? Thoughts on What Might Be at Stake in Debates on Self-Awareness Dan Arnold Abstract This paper will philosophically engage the question of why or whether Mādhyamika philosophers are right to refute svasaṃvitti, which arguably amounts to refuting consciousness. Are there good Mādhyamika reasons for thinking this problematic? Candrakīrti, we know, wrote a lengthy refutation of some version of the svasaṃvitti doctrine, whereas Śāntarakṣita’s more favorable inclination toward Yogācāra allowed him to embrace the idea; as some later Tibetan scholars understood, though, it’s not clear that Candrakīrti was refuting and Śāntarakṣita embracing the same idea, and whether there is good reason for a Mādhyamika to oppose svasaṃvitti depends of course on just how that is understood. I will suggest that at least on one understanding of the idea, Mādhyamikas ought not to think svasaṃvitti particularly problematic. I will also suggest, though, that even the non-problematic understanding of svasaṃvitti raises deep questions about the basic intelligibility of the Buddhist project.

Keywords Madhyamaka – Candrakīrti – Śāntarakṣita – svasaṃvitti – self-awareness – ­epistemology – phenomenology – givenness

...

1

Contours of the Debate

The Indian Buddhist philosopher Candrakīrti (fl. c.600 ce), a proponent of the Madhyamaka school of thought, is thought by many to have decisively refuted the doctrine of “self-awareness” (svasaṃvitti). That doctrine, which would become most influential as elaborated by Candrakīrti’s near-contemporary

© koninklijke brill nv, leiden, ���1 | doi:10.1163/9789004440913_009

Should Mādhyamikas Refute Subjectivity?

155

Dharmakīrti (c.600–660), was known to Candrakīrti from the work of Dignāga (c.480–540). Engaging Dignāga’s view, Candrakīrti concluded not only that self-awareness is not ultimately real, but that it is not even conventionally real. Given the significance of conventional truth for Candrakīrti, it is striking that he so uncompromisingly rejects self-awareness; this would have to be a problematic idea indeed if he can reasonably think it has no place even in our ordinary understanding of experience. It should, though, be emphasized that Dignāga’s elliptical introduction of the doctrine is elusive enough that it is not an entirely straightforward question whether Candrakīrti’s critique was on target. Matters become only more complicated if it is further considered how variously the idea of self-awareness was understood and deployed later in the Buddhist tradition. Much contemporary discussion of the issue is framed with reference to Paul Williams’s 1998 book The Reflexive Nature of Awareness, which takes an interesting angle on the question of whether there are versions of the idea that escape Candrakīrti’s critique. On Williams’s reading, the Buddhist philosopher Śāntarakṣita (c.725– 788) – himself a proponent of the Madhyamaka school of thought, although one clearly more sympathetic to the idealism of Yogācāra than was Candrakīrti (who was relentless in his critique of that school) – suggests that self-awareness be understood quite differently than it it was by many critics, at least, of Dignāga and Dharmakīrti. In developing that idea, Williams takes his bearings from the modern Tibetan philosopher Mipham (1846–1912), who cleverly leveraged Śāntarakṣita’s different emphasis to argue that if it is Śāntarakṣita’s idea that we have in view, something worth the name “self-awareness” does indeed withstand the critique by Candrakīrti – who cannot in any case have meant to target Śāntarakṣita, since the latter wrote about a century later. Mipham thus argued that the most reasonable view for Mādhyamikas to hold is that selfawareness, while not of course ultimately real, should figure in a conventionally true account of cognition, notwithstanding Candrakīrti’s contrary conclusion. To put Mipham’s contention in perspective, it should be acknowledged that thinkers in the epistemological tradition of Dignāga and Dharmakīrti can be (and were) read as deploying self-awareness toward ends rightly thought antithetical to Madhyamaka. If “self-awareness” denotes (as many have taken Dignāga and/or Dharmakīrti to hold) an epistemically privileged kind of awareness – a uniquely indubitable “inner perception” of what is immediately present to experience – proponents of Madhyamaka were right to resist the idea. (I here leave aside the question whether that really represents the best reading of Dignāga and/or Dharmakīrti; suffice it to say that is clearly how Candrakīrti read Dignāga.) Mipham recognizes, however, that Śāntarakṣita had other fish to fry when it comes to svasaṃvitti; on my reading, Śāntarakṣita’s

156

Arnold

is not an epistemological deployment of the concept, but rather something more like a transcendental view thereof.1 On Śāntarakṣita’s view, svasaṃvitti is perhaps best rendered not as “self-awareness,” which is among other things ambiguous as to whether it involves an objective genitive (awareness of some object as “self”), or a subjective one (the awareness had by oneself). I would instead venture that Śāntarakṣita’s idea might better be captured by the translation intrinsic awareness (also a possible reading of the Sanskrit); for he clearly has in mind not a particular kind of awareness, but something more like a defining characteristic of cognition. On his view, then, it is cognition’s svasaṃvitti – its intrinsic awareness (subjectivity, consciousness, self-awareness) – in virtue of which cognition is, as it were, cognitive. What Śāntarakṣita affirms when he introduces self-awareness, then, is just that cognitions are constitutively subjective – an idea so commonsensical as to be tantamount to his affirming simply that there are conscious experiences. Given that it is just how we are to understand svasaṃvitti that is at issue throughout, I will henceforth leave it untranslated, instead letting the different positions serve to explain.2 Nevertheless, it is apt to say that I will argue that there is no good reason for a Mādhyamika to resist the idea that subjectivity is among the conventionally real features of cognitions – and that, I take it, is just what is denied by would-be defenders of Candrakīrti’s critique. 1 It should at the outset be emphasized that the works of all these thinkers can be read as recommending various understandings of svasaṃvitti. Dignāga’s view, sketched in just a few verses of the ps (with commentary thereon), is particularly indeterminate, while Dharmakīrti’s treatment (in the pv and the pvin) is so expansive as to elude definitive characterization. On Dignāga’s arguments, see Kellner (2010); on the range of ways the doctrine can be taken, see Arnold (2010). In the present volume, see particularly the contributions by Coseru, Funayama, and Siderits. My views have particular affinities with those developed in Coseru (Chapter 4), though there are some differences worth noting. See, too, Dreyfus (2011), who suggests that Madhyamaka should be thought compatible with some of the Yogācāra tradition’s characteristic emphasis on subjectivity (though his focus is on the category of ālayavijñāna rather than on svasaṃvitti); and Thompson (2011), who agrees with me in thinking that Śāntarakṣita’s view helps us recognize the difference between subjectivity, and the kind of self that Buddhists refute. I would note in this regard that despite the extent of our shared sympathies, Coseru seems to me get things backwards in saying that “what it means for there to be self-knowledge without subjectivity” is the “open question” here; on my reading, Śāntarakṣita embraces the conventional reality of subjectivity, and the question thus is what that means in the absence of a self. 2 In a related vein, I beg the reader’s forgiveness for my imprecisely alternating among terms like “consciousness,” “subjectivity,” “cognition,” and the like. This is in part an artifact of the similar range of terms encountered in the literature on the Buddhist thinkers here in view. Generally when I say consciousness or subjectivity (e.g.), I mean to identify a fact about (or feature of) cognitions; reference to the latter, in contrast, is generally meant as denoting occurrent occasions of experiencing. (References to awareness, I reckon, are perhaps most inconsistent.)

Should Mādhyamikas Refute Subjectivity?

157

Śāntarakṣita is concerned, then, not with a distinct kind of cognition, but with a fact about cognitions, as such – the fact, in particular, that they are conscious, which surely has a credible claim to being integral to conventional understanding. I would allow, however, that Śāntarakṣita’s philosophical predilections are idealist enough that he accepts the reality of subjectivity as a phenomenologically basic datum – as, indeed, the best candidate for something uniquely possessed of “intrinsic identity” (svabhāva, the principal target of Mādhyamika critiques). Candrakīrti, who clearly wanted no truck with anything in the neighborhood of idealism, would surely worry that it concedes too much to the idealist to allow even that much; chief among my questions here is whether that is right. The contemporary interpreters who take Candrakīrti’s side here clearly agree with his thought that idealism shouldn’t be allowed even to get off the ground. Jay Garfield worries, in this regard, that insofar as subjectivity can be cashed out in terms of distinctively first-personal access to conscious experience, Śāntarakṣita’s account entails a whole raft of problems fashionably characterized as “Cartesian.” Mark Siderits (Chapter 6) similarly suggests that the problem with accepting subjectivity as criterial for cognition is that that just is to posit a privileged kind of awareness: first-personal awareness, in particular, of one’s “own” mental states, the supposedly unique immediacy of which is just what gives rise to the “problem of other minds.” Rejecting this asymmetry, Siderits takes the upshot of Candrakīrti’s critique to be that minds are, in fact, “just as opaque to themselves as they are to other minds”; indeed, he says, Madh­yamaka would have us conclude that “the first-person and third-person cases are exactly alike.” It’s not hard to imagine the sense it could make particularly for a Buddhist to say as much. The guiding thought of most Buddhist philosophy, after all, is anātmavāda, the doctrine that persons are “without selves” – and insofar as selves are posited as (among other things) the referents of first-person pronominal usage, it would seem that accepting the distinctive immediacy of firstpersonal access to experience does indeed open the way to conclusions that Buddhists are intent on foreclosing. Christian Coseru rightly says in this regard that “Buddhist accounts of substantive self-knowledge cannot rest on egological conceptions of self-consciousness, that is, on conceptions of consciousness as the property, function, or dimension of an enduring subject or self” (Chapter 4, p.123). And Siderits, again in this volume, worries that “even if one stops short of positing a single enduring entity as the subject of one’s mental states,” the admission of svasaṃvitti “will still motivate intuitions that make such a posit seem plausible” Chapter 6, p.191). I take it, however, that Śāntarakṣita’s view also makes sense as consistent with the Buddhist soteriological project. By way of showing as much, I aim here

158

Arnold

to develop a few points. After briefly considering the basics of Candrakīrti’s critique of svasaṃvitti, I will first sketch what I take to be Śāntarakṣita’s approach. Integral to understanding that, I think, is a recognition of the rhetorical significance of his willingness to entertain the idealist’s thought that cognition is uniquely basic. Particularly as developed in his concise Madhyamakālaṃkāra (“Embellishment of Madhyamaka”), the rhetorical force of Śāntarakṣita’s exposition of Madhyamaka lies in its showing that even cognition, the best candidate for the status of something intrinsically intelligible, turns out to be (as all Mādhyamikas are apt to say of all existents) empty of any intrinsic identity. In assessing Śāntarakṣita’s view of svasaṃvitti, then, it must be kept in mind that despite the lengths to which he goes in appropriating Yogācāra arguments, his point all along is finally to emphasize that cognition is no more ultimately real than anything else. Those who are sympathetic to Candrakīrti’s critique are nonetheless apt to resist his approach on the grounds that Śāntarakṣita seems to allow cognition a privileged status among conventionally real phenomena; on the view that Madhyamaka cannot thus endorse any sort of philosophical hierarchy among conventional existents, that alone compromises his Madhyamaka bona fides.3 If, though, we are really to judge whether svasaṃvitti represents an idea that Madhyamaka should reject, we must approach Śāntarakṣita’s contribution to the debate with an openness to the possibility that he is not addressing the question we had supposed he was; otherwise, it may not be Śāntarakṣita’s proposal that we are assessing. In the introduction to this volume (as also in Chapter 6), Mark Siderits says  in this regard that svasaṃvitti represents a Buddhist theory of “meta-­ cognition” – an answer, that is, to the question, “How is consciousness itself cognized?” I suggest, though, that it may be misleading to suppose from the outset that the orienting question here is that of how cognition itself is “cognized.” This stacks the deck for the idea that we must be concerned here with some additional cognitive act that is performed – for the claim that something further must be done to cognitions such that, as “cognized,” they can themselves appear together with whatever objects they disclose. It thus becomes reasonable to think we are owed an explanation of just what the agent and patient of this further “act” could be, making it hard to allow (or even to see) that svasaṃvitti, on the best understanding thereof, might be aptly characterized as a way of refusing the intelligibility of just such a demand. To ask, then, how cognition itself is cognized is arguably to fail to grasp something integral to the very concept of cognition. 3 See note 40, below.

Should Mādhyamikas Refute Subjectivity?

159

That is, at any rate, what I take Śāntarakṣita to argue, and my principal exegetical task is simply to show that, whatever the merits of Candrakīrti’s critique of svasaṃvitti, the modern Tibetan scholar Mipham was right that that critique affords no purchase with respect to Śāntarakṣita’s idea. While I think the case for that conclusion is fairly straightforward, I also want, more ambitiously, to motivate a philosophical case for Śāntarakṣita’s view, notwithstanding Can­ drakīrti’s thoughts on the matter, that a Mādhyamika can and should allow that subjectivity is real – as “real,” at least, as anything can be, where that is to say that it is relatively, conventionally real. Among the considerations that give pause is that the subjectivity of awareness seemingly entails that first-personal access to experience is privileged. Depending, perhaps, on what “privileged” means, though, that might not count against its nonetheless being just conventionally real. After all, the “problem of other minds” is not a peculiarly technical philosophical issue; rather, it arguably reflects eminently human feelings of aloneness, and thus makes sense as an aspect of conventional truth.4 It would seem, indeed, that it is not svasaṃvitti that sits uneasily with conventional truth, but rather the behaviorist idea that access to one’s own experience works exactly the same way as awareness of others’ experience. There is a steep price to be paid for denying that, particularly insofar as Candrakīrti rightly took the significance of conventional truth to be of central concern for Madhyamaka. The cost is evident in Siderits’s concession (in Chapter 6) that the “transparency thesis” he opposes is so widely experienced as intuitively obvious that it requires considerable work to counter its attraction. Significantly, the work that’s thus required takes, for Siderits, the form of a philosophically revisionary project; on his reading, it takes a physicalist recasting of Buddhist reductionism to “level the playing field between the two competing theses.” To the extent, however, that such philosophical heavy-lifting is necessary in order for Siderits thus to motivate the alternative to svasaṃvitti that he thinks Madhyamaka should embrace, it seems our ­ question  – was Candrakīrti right to think self-awareness not even conventionally real? – becomes only more acute. 4 Writing in Chapter 6, Siderits appears to consider it a problematic upshot of the svasaṃvitti idea that it is apt to exacerbate the problem of other minds: “if what we mean by ‘conscious’ is what we are aware of in our own case, then it is simply false that there are other conscious beings. Ratnakīrti was right and Dharmakīrti was wrong: solipsism follows from the YogācāraSautrāntika view” (p.196). As against, that, I am suggesting that the problem of other minds is always already lurking in conventional experience; if that’s right, the point that svasaṃvitti is problematic on account of its fostering that could, in fact, count in favor of svasaṃvitti’s having a place in conventional truth. On the generally human nature of the problem, see Valberg 2007: 168–200; Cavell 1979; and Moran 2011/12.

160

Arnold

I would suggest in this regard that the “transparency thesis” is hard to dislodge because, in fact, it is veritably constitutive of conventional truth. I take it that Madhyamaka’s guiding thought, in this regard, is that conventional truth cannot coherently be thought to be explained by anything ultimately real, and this because any putative explanation turns out itself to be intelligible only with reference to conventional truth. Śāntarakṣita is sufficiently idealist in his philosophical sensibilities that he can embrace svasaṃvitti for the very same reason: just as reductionist analyses of persons make sense only insofar as “persons” are in view, so, too, any investigation of reality makes sense only insofar as concern for such an investigation shows up for the experience of an inquiring subject. What’s more, that thought can be understood not only as consistent with the no-self doctrine, but as a basis for elaborating that. It matters in this regard that self-awareness need not be understood chiefly in epistemological terms; that first-personal access is distinctive entails no commitment to thinking such access infallible, or to the idea that we are transparent to ourselves. I will suggest in this regard that Candrakīrti’s would-be defenders are too quick to suppose that the phenomenological givenness that figures in Śāntarakṣita’s case is tantamount to (or entails) the kind of epistemic givenness influentially critiqued by Sellars. Candrakīrti is aptly characterized as having advanced good arguments against the latter, but the thought that the same considerations count against the arguably phenomenological concerns of Śāntarakṣita represents, I think, a category mistake. Let us begin, though, by getting some idea of just what Candrakīrti thought incoherent about self-awareness. 2

Candrakīrti’s Critique of Self-Awareness

Candrakīrti’s principal critique of Dignāga’s doctrine of svasaṃvitti comes in a sustained engagement, in Chapter 6 of the Madhyamakāvatāra, with the idealism characteristic of the Yogācāra school of thought. In that context, most salient for Candrakīrti is that self-awareness is supposed to explain how reality (understood by proponents of Yogācāra as finally consisting only in awareness itself) could be known even in the absence of the basic structure of “knowing” – in the absence, in particular, of an ultimately real distinction between what Dignāga refers to as the “apprehending” and the “apprehended aspects” of awareness (i.e., its subjective and objective poles).5 Candrakīrti thus launches 5 Dignāga’s invocation of self-awareness figures centrally in his case (made at ps 1.8–10) for thinking that both the content of cognition and the subjective apprehension thereof – the

Should Mādhyamikas Refute Subjectivity?

161

his critique of self-awareness with reference to (what is among the principal Yogācāra names for cognition as ultimately real) the “dependent” nature (paratantra-svabhāva): “If the dependent nature is devoid of these two – that is, without anything ‘apprehended’ and without an ‘apprehender’ – then by whom is its existence known? It doesn’t make sense that it exist without being apprehended!” For Candrakīrti, the point of Dignāga’s appeal to self-awareness is to meet this objection by affirming that it’s just in the nature of awareness for it to know itself.6 Notably, Candrakīrti’s argument against this idea typifies some characteristically Sanskritic modes of argument. Indeed, it might give a Mādhyamika pause that much the same line of argument – which exploits widely shared Sanskritic presuppositions about the orienting significance of Sanskrit grammatical categories for philosophical analysis – was also pressed by Kumārila, an influential proponent of the ideologically conservative Mīmāṃsā school of Brahmanical philosophy. It’s worth noting in this regard that despite their strong resistance to self-awareness – as generally advocating robustly realist views, Mīmāṃsakas were wary of giving any ground to idealism – Mīmāṃsakas strongly affirmed the reality of selves. That Kumārila and Candrakīrti could make common cause in refuting svasaṃvitti suggests, then, that the latter idea stands or falls quite independently of the self idea. In any case, Kumārila agrees with Candrakīrti that the very idea of self-awareness is just incoherent, both of them arguing against Dignāga that nothing can coherently be taken as agent and patient of the same act. Like Candrakīrti, Kumārila’s commentators in this regard favored the same examples typically adduced by Indian philosophers in adumbrating this “anti-reflexivity principle”: cognition can no more know itself than a finger-tip can touch itself, than a sword can cut itself, or that fire can illuminate itself.7 “aspect to be apprehended” (grāhyākāra) and the “aspect which is apprehender” (grāhakākāra) – are just intrinsic to cognition, which must therefore be reckoned as autonomously intelligible. 6 The quotation is from Madhyamakāvatāra 6.72, which I have translated from Li (2015: 13); cf. Huntington (1989: 166), La Vallée Poussin (1910: 349). For a helpfully concise sketch of the whole arc of argument launched by this verse, see MacDonald 2015 ii: 234–235, n459. 7 For Kumārila’s argument, see śv, Śūnyavāda section, verse 64; the finger-tip and fire examples are given by Kumārila’s commentator Pārthasārathimiśra, for whose comment see Shastri (1978: 205). Candrakīrti invokes the same examples in his auto-commentary on mav 6.72; see La Vallée Poussin (1970: 166; 1910: 349). See, too, Garfield, (2006: 202–203). Funayama (see Chapter 7), following the Chinese translation of a major Abhidharma work, gives the following examples: “A finger cannot touch itself; the blade of a sword cannot cut itself; the eye cannot see itself; and a strong man cannot carry himself on his shoulder.”

162

Arnold

Candrakīrti further argues that the cost of denying this principle is infinite regress. He considers, for example, Dignāga’s argument that memory makes sense only given cognition’s intrinsically having a reflexive aspect.8 The initial premise of Dignāga’s argument is that the phenomenologically distinctive character of memory – its being an experience not merely of some object, but of some object as past – consists in memory’s comprising not just the object(s) experienced on the remembered occasion, but also the fact of its having been previously experienced; when I recollect an occasion of seeing a colorful sunset, my recollection is of my having experienced that. Dignāga concluded from  this that the second-order aspect – awareness of having experienced something – could be available to memory only if it had in the first instance been part of the experience recalled; after all, one is not rightly said to “remember” something not previously experienced. As Evan Thompson puts the conclusion to this argument, it thus follows that “the original experience was not simply one of perceiving the object but also one of experiencing the seeing of the object” (2011: 162). With respect to this line of argument, Candrakīrti’s critique again has affinities with that of Kumārila, who challenges Dignāga’s basic premise regarding the phenomenology of memory. In particular, Kumārila argues that Dignāga’s account of memory would absurdly entail an endless “multiplication of representations” (ākārapracaya), with each further recollection of any occasion necessarily adding further aspects. It cannot be right, Kumārila thus argues, that the phenomenology of memory is as Dignāga presupposes.9 Similarly, Candrakīrti argues that if memory were not of objects but of experiences thereof, then a further moment of experience would be entailed as experiencing the memory, and a further one as experiencing that, etc.10 Insofar as the analysis of 8 9

10

On the general significance for Buddhists of the anti-reflexivity principle, see Siderits 2007: 46–50. On this argument, see Kellner (2010: 210, ff.), Ganeri (1999). Thompson (2011: 161–168) suggests a reconstruction of this argument with reference to Husserl; Siderits (Chapter 6) offers a different reading of the same line of argument. śv, Śūnyavāda 112–114 (Shastri 1978: 213–214). Kumārila’s argument is noted by Kellner (2010: 212) and Hattori (1968: 109, n.1.70). Interestingly, the 2015 film Inside Out cleverly depicts recent psychological theories according to which there is, in fact, something like a multiplication of “aspects” with each further recollection; according to such theories, each accessing of any memory necessarily taints it with the affective dimension involved in the recollecting, thus changing what is available to future moments of reflection. This argument is from Candrakīrti’s auto-commentary on mav 73a (La Vallée Poussin 1970: 167.11, ff.; translated in La Vallée Poussin 1970a: 350). Note that Huntington (1989: 244n101) mistakenly represents this as part of Dignāga’s argument for self-awareness, rather than as part of Candrakīrti’s critique thereof.

Should Mādhyamikas Refute Subjectivity?

163

the remembered experience must pertain also to any recollection thereof, infinite regress looms; for if it’s right that memory is not just of an object but of an experience of an object, it seems the recollection must in turn be an experience of an experience of an object, etc. Candrakīrti elsewhere critiques the doctrine from a different angle that he takes to reveal much the same problem. In the course of a sustained engagement with Dignāga in the first chapter of the Prasannapadā (pp), Candrakīrti closely considers the category of svalakṣaṇa as that figures in Dignāga’s epistemology. Exhibiting his predilections as an ordinary language philosopher, Candrakīrti asks whether Dignāga’s peculiarly technical use of this Sanskrit word can make sense of ordinary uses thereof. The problem is that on its ordinary usage, svalakṣaṇa denotes a “defining characteristic,” whereas Dignāga takes it as denoting something bare of any defining characteristics – specifically, the unique particulars that are the objects of perception (pratyakṣa). Insofar as Dignāga takes perception to be constitutively non-conceptual, it cannot apprehend objects as “characterized” in any way at all.11 While Candrakīrti takes it as integral to ordinary perceptual experience that we perceive things under a description, Dignāga’s revisionary epistemology instead holds that it is just unique particulars that are given to perception, and he appropriates the word svalakṣaṇa as denoting these. The problem with that, according to Candrakīrti, is not simply that it flouts conventional usage, but also that the conventional usage constitutively involves a relationship, as reflected in the Sanskrit grammatical tradition’s analysis of the word-form; a characteristic (lakṣaṇa) is intelligible only as the characteristic of what is characterized thereby (lakṣya). Dignāga’s idea instead appears to be that perceptible particulars are “characterized” just by themselves.12 In the the course of arguing thus, Candrakīrti anticipates that Dignāga might invoke svasaṃvitti to meet just that objection. Consider, here, one of the examples of conventional usage that Candrakīrti presses against Dignāga: the expression vijñāna-svalakṣaṇa. According to conventional usage, this expression denotes a defining characteristic of cognition – as, for example, when it is said that the defining characteristic of cognition is its being a “presentation of a particular object” (viṣayaṃ prati vijñaptiḥ). Given, however, Dignāga’s conception of the word svalakṣaṇa, the same expression would instead mean “the unique particular which is cognition.” For Dignāga, that would be to say the 11 12

Candrakīrti also resists Dignāga’s characterization of pratyakṣa (perception) on ordinarylinguistic grounds; see Arnold (2005: 175–183). For a detailed analysis of the arguments here, see Arnold (2005: 152–158); see, as well, Arnold 2003, 2005a).

164

Arnold

expression denotes cognition as an object of perception. Candrakīrti argues, however, that that makes no sense, since in that case “the defining characteristic of cognition could not be known, because it is not an object; after all, a defining characteristic can be known only insofar as it is the way an object appears.”13 Candrakīrti’s point is that cognition is not an object just by definition; it is, rather, the cognitive “instrument” by which objects are known – and as the grammarians tell us, the instrument of any act is necessarily distinct from the patient thereof.14 That, finally, is why svasaṃvitti is relevant here; Candrakīrti’s thought is that Dignāga might invoke self-awareness precisely as exemplifying a sort of awareness that does, in fact, take cognition itself as its object. In heading off this move, Candrakīrti refers the reader to the lengthier critique of selfawareness we have scouted following the Madhyamakāvatāra, but he here adds a brief comment on the upshot of that: It does not make sense that one svalakṣaṇa is characterized [or, “apprehended,” lakṣyate] by another svalakṣaṇa, and that one, in turn, by selfawareness. Moreover, that cognition – itself unestablished apart from its defining characteristic – does not exist in any way at all; there being nothing to be characterized, it is impossible, since there is no possibility of a characteristic’s pertaining without any basis.15 This passage resists translation insofar as the question of what svalakṣaṇa means is itself in play here, but the point is again that the idea of self-­awareness is simply incoherent. Here, the reasons have particularly to do with how ­cognition and its objects are, on a conventional understanding, to be defined. 13 14 15

pp 61.4–5; cf. Arnold (2005a: 432); MacDonald (2015, vol.ii, pp.232–233). Candrakīrti’s critique of Dignāga’s usage involves reference to the grammatical tradition’s theorization of the -ana affix (which figures in the word-forms pramāṇa and svalakṣaṇa), which typically denotes the instrument of an act; see Arnold 2005: 153–154. pp 61.11–62.3; here translated according to the variant readings given in MacDonald (vol.i, p.251); for MacDonald’s translation, see vol.ii, pp.234–236. Translation of this passage is complicated by Candrakīrti’s exploitation of the alternative senses of the terms at issue. For example, the verb in the first sentence (lakṣyate), on the most natural reading, means “characterized” or “defined”; in that case, the point would seem to be that one thing’s defining characteristic cannot itself characterize that of another. It would also make sense, however, for the same verb to be used with an epistemic sense like “apprehended” – in which case, the point is perhaps that a defining characteristic cannot itself “apprehend” anything. MacDonald (2015, ii: 236 n460) offers a slightly different gloss of the argument: “Dignāga maintains that the object (karman) in the case of self-cognition is jñāna, but since jñāna is nothing but its characteristic, i.e., awareness, in this case self-awareness, there will not be an object for this self-awareness; that is, jñāna-cum-svasaṃvitti cannot additionally be posited as a lakṣya.”

Should Mādhyamikas Refute Subjectivity?

165

It doesn’t make sense, in particular, to accept that the svalakṣaṇa (“defining characteristic”) of cognition is its being a “representation of a particular object,” and at the same time to say that that fact itself characterizes another such characteristic. Putting the same point in terms of cognition, per se, Candrakīrti adds that it makes no sense that cognition could be, in turn, the object of svasaṃvitti. Cognition, by definition, consists in a representation of a particular object, and it makes no sense that cognition both have that as its “defining characteristic” (svalakṣaṇa) – that cognition be such a “representation” – and at the same time that it be, itself, the object of just such a representation. Insofar as cognition thus makes no sense as the content of self-awareness, though, it follows that there is therefore nothing for “self-awareness” to be of – which means the selfawareness posited by Dignāga therefore fails to exhibit the defining characteristic of cognition. That Dignāga cannot reconcile his usage of the word svalakṣaṇa with ordinary usage, then, finally entails that svasaṃvitti fails even to count as an instance of awareness. What is most salient about Candrakīrti’s arguments is that they commonly presuppose that svasaṃvitti, on Dignāga’s view thereof, must exhibit the same kind of act-structure that admits of analysis in terms of the Sanskrit grammatical tradition’s theorization of the various ways in which nouns can syntactically relate to verbs. Insofar as the sort of act at issue is that of cognizing, the upshot is that self-awareness must exhibit the same intentional structure as any first-order awareness; it thus becomes hard to imagine it as anything other than a higher-order awareness of one’s first-order awareness – an additional act, that is, accompanying first-order acts of cognizing. To say that every awareness is somehow characterized by svasaṃvitti, then, could only be to say that every awareness is accompanied by a higher-order awareness, in virtue of which cognitions comprise themselves as part of their phenomenal content. If that’s right, infinite regress surely looms; for if it is only in virtue of a further act of cognition that we count as being aware of our own experience, the same analysis must pertain to the further act, too.16 As reflected in the argument’s debt to grammatical analyses, Candrakīrti’s critique thus presupposes that svasaṃvitti makes sense only as a particular kind of cognitive act. If it is right that svasaṃvitti must, in order to count as awareness, thus exhibit the

16 See Garfield (2006: 209) on the way in which all available options give rise to vicious regresses. Among the reasons to doubt whether Candrakīrti’s critique has purchase even against Dignāga’s view is that Dignāga thought reference to self-awareness necessary to avoid infinite regress; see Kellner (2010: 213–215).

166

Arnold

structure of cognitive acts, then it is surely relevant for a critique of the idea to show that nothing makes sense as the accusative of that act – and if there is nothing that such awareness could be of, it is indeed hard to see how it counts as awareness. This angle of critique effectively discloses the same sort of problem pressed by modern critics of self-awareness as that figures in post-Kantian German idealism – the “iteration problem,” as Robert Pippin thus explains: If consciousness and self-consciousness are treated as separate aspects of any consciousness, then the arguments that showed why consciousness of X must be accompanied by consciousness of consciousness of X would all apply to the latter too, since self-consciousness … would also be an instance of consciousness and so subject to its conditions. (1989: 46–47) Variations on this problem are unavoidable as long as it is supposed that (as Pippin puts it) “in “thinking a thought,” two mental events occur, or two twoplace relations, between my thinking and its thought, and between me and my thinking a thought.” (Ibid.) Insofar as that is just how Candrakīrti seems to have understood the idea first theorized by Dignāga, it is no wonder he thinks the idea is doomed.17 3 Self-Awareness According to Śāntarakṣita’s Madhyamaka Akeel Bilgrami, in developing his own account of self-awareness, introduces a distinction that can helpfully frame the Mādhyamika philosopher Śāntarakṣita’s version of svasaṃvitti as a genuine alternative to the idea Candrakīrti had in his sights: there are, Bilgrami says, perceptual conceptions of self-awareness such as Candrakīrti clearly targets, and constitutive conceptions such as Bilgrami himself elaborates. Self-awareness, on the latter sort of view, is constitutive of intentional cognitions in the sense that the latter are “conceptually dependent in a crucial sense on our self-knowledge of them.” (2006: 23) (To similar effect, Christian Coseru argues (Chapter 4) that insofar as mental states are “constitutively, rather than relationally, intentional,” the self-awareness thereof cannot be understood as “a relation that consciousness has to itself when it attends to its own operations”; the self-awareness of mental states is,  rather, a “structural feature” mental states have just “by virtue of being 17

Again, I am here setting aside the question of whether this line of critique actually has purchase against Dignāga’s view of self-awareness; our question is only whether it works against Śāntarakṣita’s view.

Should Mādhyamikas Refute Subjectivity?

167

c­onscious.”) The idea that self-awareness is thus constitutive of cognitions nicely captures, I think, Śāntarakṣita’s understanding of the doctrine, and the point of my invoking Bilgrami’s framing is to recommend particular caution in assimilating such a view to the kind of perceptual view Candrakīrti targets. Bilgrami’s own case for a constitutive view of self-awareness, then, is structured by what Bilgrami takes to be a “governing disjunction” between the two ­options – a disjunction such that “[o]nly one of the two models, perceptual or constitutive, can be right.” (Ibid.: 28) If Bilgrami is right, then insofar as it is apt to say (as I will) that Śāntarakṣita’s is clearly a constitutive view of svasaṃvitti, it follows that only one of the views at issue – Śāntarakṣita’s constitutive view, or Dignāga’s clearly perceptual view – can be right. To that extent, there is good reason to think that Candrakīrti’s critique of (what he took to be) Dignāga’s svasaṃvitti doctrine cannot, in fact, be relevant to Śāntarakṣita’s understanding of that. Whether or not Bilgrami is right, though, it is surely the case that Śānt­ arakṣita’s formulation is not, at least, obviously vulnerable to Candrakīrti’s critique. The locus classicus for Śāntarakṣita’s view of self-awareness is a pivotal few verses that occur early in his relatively concise Embellishment of Madhyamaka (Madhyamakālaṃkāra) – in particular, verses 16–18 of that text’s 97 verses.18 In verse 16, Śāntarakṣita first stipulates that svasaṃvitti is just criterial for a point that he not unreasonably takes to be uncontroversial: “Cognition is distinct from what consists of inanimate matter (jaḍarūpa); the fact that cognition is not inanimate is just its svasaṃvitti.”19 Śāntarakṣita thus affirms, in effect, merely that svasaṃvitti is the “defining characteristic” (svalakṣaṇa) in virtue of which awareness is to be distinguished from things like rocks and tables and stars. Among the salient facts about occurrences of cognition, Śāntarakṣita effectively proposes in this way, is that they are subjective or conscious; insofar as it denotes that fact about cognitions, we can reasonably say that svasaṃvitti, on this usage, refers to something like subjectivity, per se. We might also say that svasaṃvitti, for Śāntarakṣita, is transcendental rather than empirical, and thus 18

19

The same verses are repurposed in Śāntarakṣita’s Tattvasaṃgraha (ts), wherein they show up as verses 1999–2001ab. The significance of svasaṃvitti for Śāntarakṣita’s exposition of Madhyamaka is, however, more immediately clear in the context of the Madhyamakālaṃkāra, from which I here take my bearings. My own translation of the complete text of this is among the selections in Arnold (forthcoming). Translated from Ichigo (1985: 70); cf. Shastri (1997, vol.2, p.478), as well as Garfield (2006: 211), Williams (1998: pp.24–25). In the second half-verse and in the following verses, I have supplied the antecedents for all of the pronominal references to the first quarter-verse’s vijñānam.

168

Arnold

more akin to Kant’s “transcendental unity of apperception” than to anything like “inner perception.” That is so in the sense (perhaps among others) that we can in the first instance identify moments of cognition as tokens of the type “cognition” only if we have in view something subjective – otherwise it won’t be as cognitions that we are identifying them.20 Śāntarakṣita immediately emphasizes that this just is to say that svasaṃvitti cannot, in fact, be understood as exhibiting the same intentional structure as acts of knowing. Thus, he says in verses 17–18ab of the Madhyamakālaṃkāra that as denoting simply a fact about cognitions, svasaṃvitti is not vulnerable to critiques that exploit grammatical analyses of the structure of action: It is not, however, in terms of the relation of action and agent that cognition has svasaṃvitti, since it does not make sense that something unitary (which is without parts) be threefold; so, svasaṃvitti makes sense, in the first instance, just because of cognition’s consisting of consciousness (bodharūpa).21 Whether or not he is warranted in saying as much, it is abundantly clear that Śāntarakṣita thus means to clarify that critiques just such as Candrakīrti’s have no purchase with regard to the idea he (Śāntarakṣita) is stalking. However, Jay Garfield, a strong proponent of the view that Candrakīrti’s critique of svasaṃvitti is rightly taken as normative for Madhyamaka, says of Śāntarakṣita’s move here that it is “brutally simple, if not obviously coherent”; Śāntarakṣita simply “denies that there is any identity of agent, object, and act in reflexive awareness, since these three components are not present.”22 On Garfield’s reading, Śāntarakṣita thus offers only a question-begging assertion where an argument is needed. This is an uncharitable reading of Śāntarakṣita, to say the least; surely there is nothing “incoherent” in his clarifying his initial claim – viz., “the fact that cognition is not inanimate is just its svasaṃvitti” – by clarifying its scope. Garfield’s characterization of the move 20

21 22

The point is akin to John Searle’s often-engaged contention that subjectivity is ontologically subjective – that it is only as subjective, that is, that “subjectivity” can be in view (which means that attention to any of the objective factors that figure in producing consciousness is not, ipso facto, attention to subjectivity). See Searle 1997. Translated from Ichigo (1985: 70–74); cf. Shastri (1997, vol.2, p.478), Garfield (2006: 211). Garfield (2006: 212). In Engaging Buddhism, Garfield says to similar effect: “It might appear that there is something terribly flat-footed about any insistence such as … Śāntarakṣita’s that nothing relational, nothing not already subjective, could ever differentiate conscious from unconscious phenomena. But it is a position remarkably prescient of new mysterians…” (2015: 152).

Should Mādhyamikas Refute Subjectivity?

169

suggests that Śāntarakṣita is merely asserting, of what he must realize is an act, that this kind of act uniquely involves the identity of act, agent, and patient. It begs the question, though, to urge against Śāntarakṣita that he is absurdly conflating the distinct terms of an act; for his point, I take it, is to argue that just such analysis is irrelevant to the point he is developing. We must suppose, then, that it is not Śāntarakṣita’s idea that we have in view if we persist in thinking the analysis of actions is relevant. What’s more, there is good reason to think Śāntarakṣita is right about this; after all, a fact about cognitions is not itself an act of cognizing, and it is to that extent a category mistake to suppose that the subjectivity of cognitions must itself exhibit the same structure as cognitions themselves.23 Śāntarakṣita’s point is that svasaṃvitti, as J.J. Valberg says of self-consciousness in a relevantly similar context, “is not an act of first-person thinking. It is not an act, or event, of any kind, but a fact.”24 To that extent, concerns regarding the specific agent and patient of svasaṃvitti are basically misguided, and so Śāntarakṣita can reasonably hold that his is an uncontroversial point (a conventionally true point!) concerning the nature cognition. His point, he suggests in concluding the verses here scouted, is true just by definition: that cognition is characterized by svasaṃvitti makes sense “just because of cognition’s consisting of consciousness.” Insofar as this aptly characterizes the idea Śāntarakṣita is after, it seems to me there is good reason to think svasaṃvitti makes eminently good sense as conventionally true. Śāntarakṣita’s wager, on this reading, is that we could deny his claim only by denying that there is any difference between cognitions and inanimate objects – only by denying, in effect, that there are conscious experiences. That cognitions are distinguished from inanimate things by their svasaṃvitti, Śāntarakṣita thus holds, is just a phenomenologically basic datum. Mādhyamikas of Candrakīrti’s ilk, though, are apt to protest in this regard that the very idea of a “phenomenologically basic datum” makes no sense as being conventionally real; after all, that sounds a lot like the claim that subjectivity has an ontologically privileged status, which is just the sort of status Mādhyamikas are in the business of denying. If it is right, then, that Śāntarakṣita’s recuperation of the svasaṃvitti doctrine entails any conclusion to the effect that subjectivity is intrinsically intelligible – to the effect that it is, 23

24

To be sure, Garfield recognizes this: “Śāntarakṣita is hence denying that consciousness should be understood as an action, and so as subject to the tripartite analysis of actions into agent, action, and object.” (2006: 212) Nonetheless, Garfield’s arguments against Śāntarakṣita’s view, as I will emphasize later on, chiefly consists in his doubling down on Candrakīrti’s arguments, which are relevant only if the contention of Madhyamakālaṃkāra 17–18ab is ignored. Valberg 2007: 193; for Valberg’s further development of the point, see note 38, below.

170

Arnold

in Madhyamaka terms, possessed of “intrinsic identity” (svabhāva) – that would surely vindicate the suspicions of those sympathetic to Candrakīrti. To get a handle on this, we need some understanding of how Śāntarakṣita’s appeal to svasaṃvitti figures in the larger economy of his presentation of Madhyamaka. Among the ways we might get at that is by asking whether or how it could make sense for Śāntarakṣita to think anything in the neighborhood of idealism could count as conventionally true. In this regard, Śāntarakṣita is widely known for his distinctive synthesis of Yogācāra and Madhyamaka, and he is typically taken to have concluded that the idealism typical of Yogācāra represents the best elaboration of conventional truth; proponents of Yogācāra erred, he holds, only in supposing that idealist claims could be ultimately true. The latter conclusion finally places Śāntarakṣita in the Madhyamaka camp; while some Buddhist philosophers of Yogācāra stripe arguably held that cognition uniquely withstands critical analysis (and so is ultimately real), Śāntarakṣita’s whole point will be that nothing withstands ultimate scrutiny. Even cognition, he argues, finally fails to make sense as intrinsically intelligible, and his major insight was that the very arguments favored by Buddhist idealists could finally be turned against themselves to show as much. To that extent, Śāntarakṣita’s synthesis of Yogācāra and Madhyamaka does not involve his trying to show (as some East Asian projects in Buddhist philosophy have done) that the two schools of thought are finally in agreement. He sees Yogācāra as a necessary but finally provisional moment in a philosophical dialectic that culminates in Madhyamaka, which alone fully captures what he takes to be the ultimate truth: viz., that “all the existents posited by ­Buddhists and non-Buddhists are in reality without any essence,” as the opening verse of the Madhyamakālaṃkāra declares. Nevertheless, it’s right to emphasize that Śāntarakṣita embraces basically idealist intuitions that Candrakirti, for one, would scarcely entertain. Indeed, the rhetorical force of Śāntarakṣita’s exposition of Madhyamaka – the power, in particular, of its culminating disclosure of the conclusion that even cognition fails to exhibit intrinsic identity – arguably results from Śāntarakṣita’s initial willingness to allow that cognition does have a philosophically basic significance. The idea that cognition has a “philosophically basic significance” must, if this is to make sense as a Madhyamaka position, be understood so as to be consistent with cognition’s not being an ultimately real entity (paramārthasat bhāva). That is as Śāntarakṣita finally concludes, and I will say more in concluding about the sense this makes. More challenging, perhaps, is the not unrelated question whether the very idea of a “philosophically basic status” can be understood as consistent with Madhyamaka intuitions regarding conventional truth. Could anything make sense, that is, as being (say) a condition of

Should Mādhyamikas Refute Subjectivity?

171

the possibility of any inquiry, and as nonetheless being just conventionally existent? Śāntarakṣita clearly means to show that his close attention to cognition can take us as far as conventional truth allows; can he, though, coherently think so if he so much as entertains the idea that cognition nonetheless represents the best, most intuitively plausible candidate for something intrinsically intelligible?25 Candrakīrti’s intuition in this regard is that these ideas cannot, in fact, coherently be held together. Śāntarakṣita’s distinctive exposition of Madhyamaka, in contrast, can be understood as oriented precisely by the idea that the uniquely self-disclosing character of cognition – its svasaṃvitti or “intrinsic awareness” – makes it, prima facie, a plausible candidate for something uniquely possessed of intrinsic identity. I take it, that is, that Śāntarakṣita’s way of presenting Madhyamaka thus involves his allowing that the self-intimating character of cognition makes it plausible to suppose that cognition, uniquely, is just intrinsically intelligible; it is critical analysis of cognition that most preoccupies Śāntarakṣita, and surely that is because he recognizes the significant considerations that weigh in favor of its basicness. Among the considerations emphasized in this regard by Yogācāra thinkers are that anything at all that is known can be known only insofar as it shows up, in the first instance, to awareness – and while what is thus known can usually be doubted, that something appears thus is itself indubitable. (That is the gist of Dharmakīrti’s sahopalambhaniyama argument for self-awareness, which Christian Coseru ­(Chapter 4) aptly characterizes as an argument from the “egocentric predicament.”) Not only, though, is awareness always already evident along with anything else that becomes known; what’s more, cognition arguably contains any such thing, in that anything known shows up, ipso facto, in cognition. The Irish idealist George Berkeley held in this regard that “an idea can be like nothing but an idea”;26 that is much as Śāntarakṣita first concludes from the account of self-awareness here introduced following him. Immediately following the verses discussed above (Madhyamakālaṃkāra 16–18ab), then, Śāntarakṣita proceeds to ask: But how, in that case, could cognition be aware of anything distinct from itself? After all, cognition’s nature does not belong to anything else, such that it would be possible for something else to be cognized when ­cognition 25 26

In this regard, Śāntarakṣita epitomizes “Svātantrika” intuitions about how best to argue for Madhyamaka, whereas Candrakīrti’s resistance reflects his “Prāsaṅgika” intuitions; I say a bit more on this below. Principles of Human Knowledge §8 (Berkeley 1996: p.27).

172

Arnold

itself is experienced; for you have admitted that cognition and what is cognized thereby are distinct!27 Śāntarakṣita thus proposes his account of svasaṃvitti as motivating, first, the kind of epistemic idealism that arguably follows from empiricism – a view, that is, on which perceptual awareness is considered basic just insofar as it consists in immediate acquaintance with mental representations (ākāras or “sense data”). After showing how that option fails to withstand scrutiny, he will then turn to consider the full-blown idealism of Yogācāra. The ensuing development of Śāntarakṣita’s exposition is framed, then, precisely by the idea that cognition represents the most plausible candidate for something intrinsically intelligible – a thought explored first with reference to the epistemic idealism characteristic of empiricism (in traditional doxographic terms, “Sautrāntika,” considered and finally rejected in verses 20 through 34), then with reference to the idealism of Yogācāra (considered at verses 44–60). Though having motivated these positions by affirming cognition’s svasaṃvitti, Śāntarakṣita means to show them finally incoherent, and he probes them by pressing a basic question of the sort that begins many Buddhist arguments: are cognitions and the mental representations manifest therein identical to one another, or distinct? That framing of the challenge, which epitomizes Śāntarakṣita’s exposition of Madhyamaka, is introduced in the first verse of the Madhyamakālaṃkāra: “The existents posited by Buddhists and non-Buddhists are in reality without any intrinsic identity,” as we have already seen Śāntarakṣita say in the first half of verse 1; the reason for this, as the verse says in concluding, is that all existents are, “like reflections, neither intrinsically unitary nor intrinsically nonunitary.”28 According to Śāntarakṣita’s modus operandi, the best way to show that all existents are empty of intrinsic identity is thus to begin by recognizing that anything with an intrinsic identity would have to be specifiable as being, intrinsically, either unitary in nature, or non-unitary. Accordingly, the entirety of the ensuing text is dedicated to substantiating the claim that no existents make sense as either intrinsically unitary or intrinsically non-unitary. Śāntarakṣita is thus celebrated in the Indian Buddhist tradition for his epitomizing what would be christened the neither-one-nor-many argument – and his overwhelmingly predominant concern in the Madhyamakālaṃkāra is to drive 27 28

Verses 18cd and the first two thirds of verse 19 are again attested in Sanskrit (occurring as Tattvasaṃgraha 2001 and most of 2002). As in several of the preceding clauses, most of what I take to be references to cognition are again pronominal. Translated from Ichigo 1985: 22.

Should Mādhyamikas Refute Subjectivity?

173

home the conclusion that cognition, in particular, fails to make sense as unitary. Integral to his driving this home is his sympathetically entertaining the best case for thinking cognition, uniquely, really is intrinsically intelligible; it seems he is idealist enough at least to entertain the possibility that cognition alone might exemplify just the way of being that Madhyamaka is concerned to show impossible. But his point in lengthily entertaining that idea is precisely to show that notwithstanding the plausibility of the idealist considerations that weigh in favor of cognition’s basicness, not even cognition can finally withstand Madhyamaka’s scrutiny. Śāntarakṣita’s guiding thought is thus that the strongest case for Madhyamaka involves allowing some Yogācāra impulses to play out as far as possible. Candrakīrti, I have noted, is apt to worry in this regard that even to entertain such impulses is already to grant too much. While it remains, however, for us to consider whether there are peculiarly Buddhist reasons for thinking Candrakīrti’s concerns warranted, it seems to me that Śāntarakṣita makes a good philosophical case for thinking his modus operandi does not, in fact, entail any commitment to the intrinsic intelligibility (the “ultimate reality”) of awareness. Śāntarakṣita’s angle of critique in this regard consists in recruiting some well-known Yogācāra arguments to show that that there can be no coherent account of the relation between awareness and the content thereof, because mental content, whether on a realist account or an idealist one, is just phenomenally complex – and as constitutively related to something complex, cognition thus makes no sense as unitary.29 Nor, however, can the unity of awareness be salvaged by allowing that awareness and its content are therefore identical, since that view cannot make sense of the phenomenologically intentional character of cognition. Indeed, such a view is tantamount to denying the reality of cognitive content, and non-existent content can no more explain the intentionality of awareness than can sky-flowers or square circles.30 The trend and cogency of Śāntarakṣita’s arguments in this regard is echoed in an astute comment C.S. Peirce would make centuries later. In an 1890 “Note on Kant’s Refutation of Idealism,” Peirce remarks that any sort of subjective idealism presupposes a tendentious idea of presence: “The idealist says that all that we know immediately, that is, otherwise than inferentially, is what is 29 30

For a good feel for his approach, see especially verses 46–51. The first horn of the dilemma is scouted, with respect to Yogācāra views, at verses 46–51, wherein Śāntarakṣita makes use of arguments familiar from Vasubandhu and Dharmakīrti; for a concise discussion of this family of arguments, see Eltschinger 2010: 29–32. The second horn is considered over verses 53–60, regarding which a helpful discussion by Moriyama (2014: 345–346) considers these verses with an eye towards the later Buddhist philosopher Ratnākāraśānti’s response to Śāntarakṣita.

174

Arnold

present in the mind.” Peirce particularly emphasizes the temporal dimensions of that idea: “The whole idealist position turns upon this conception of the present.” That means, however, that “the first move toward beating idealism at its own game” is just to note the phenomenological fact that “we apprehend our own ideas only as flowing in time.” It follows that “because neither the future nor the past, however near they may be, is present, there is as much difficulty in conceiving our perception of what passes within us as in conceiving external perception.”31 Śāntarakṣita’s critique of idealism is not explicitly expressed in terms of such temporal considerations,32 but his overall argument is to much the same effect. Śāntarakṣita’s point, like Peirce’s, is developed by analyzing the idea that cognition can be of only what is present therein – and he concludes, like Peirce, that the question of cognition’s relation to its content turns out to be no less problematic for an idealist than it supposedly is for naïve realists. To that extent, the animating impulse of idealism – the thought that the most basic philosophical question is whether we really know only what is “in” awareness, or whether instead there is a real relation between mind and a world “out there” – turns out to be a red herring. If cognition is to make sense (as the idealist holds) as just intrinsically intelligible, the question is really whether cognition makes sense just by itself, or whether instead it is constitutively intentional; questions about the spatial locations of the relevant terms are finally beside the point. 4

Buddhists on First-Person Perspectives and No-Self

Śāntarakṣita’s cogent critique of idealism shows, I take it, that notwithstanding Candrakīrti’s scruples, a willingness to entertain idealist intuitions entails no commitment to finally idealist conclusions – indeed, quite the opposite. Nevertheless, Mādhyamikas of Candrakīrti’s ilk are still apt to protest that the concession reflected in Śāntarakṣita’s modus operandi – i.e., the thought that there is particular force to a case for Madhyamaka that emphasizes that even cognition is finally empty – opens the door too widely, giving comfort to deeply 31 32

Peirce Edition Project 2010 (volume 8), p.80. See, however, Madhyamakālaṃkāra verse 49, where Śāntarakṣita rejects the idea that the phenomenal complexity of mental content is merely apparent, resulting from the rapid succession of unitary cognitions that are stitched together conceptually. The problem is this: “If, however, it is admitted that there are as many cognitions as there are sense data, it is in that case difficult to deny that this view will admit of the same analysis that applies to atoms.” (Translated from Ichigo 1985: 136).

Should Mādhyamikas Refute Subjectivity?

175

misguided thoughts that ought not to be allowed even to get off the ground. Śāntarakṣita’s approach allows, in particular, that cognition (subjectivity, consciousness…) is somehow distinctive – that there is, indeed, some reason to think cognition philosophically basic. The worry regarding this is that by effectively privileging anything like the first-person perspective, one is apt to bolster the attachment to self that Buddhists are chiefly concerned to eradicate. This is perhaps the most significant of the considerations that are apt to make some Mādhyamikas leery of Śāntarakṣita’s way of presenting the tradition. It is particularly this concern, perhaps, that moves Candrakīrti to argue that the idea of self-awareness ought not to be allowed any space even among conventionally real phenomena. His case for that conclusion, we saw, is to the effect that the very idea of svasaṃvitti is patently incoherent, and thus makes no sense even conventionally.33 It is not even conventionally true, for example, that a sword can cut itself or that a gymnast can stand on his own shoulders. While such examples are perhaps too crude to be helpful, we might with more nuance say that Candrakīrti can reasonably think it incoherent to affirm as conventionally true the deliverances of any kind of awareness that is thought to be uniquely indubitable. The contradiction inherent in Śāntarakṣita’s exposition, one might thus suppose, is that the very idea of cognition’s having a “philosophically basic status” is mutually exclusive of its being just conventionally real. More precisely, it could seem that to allow that cognition is distinctively subjective is, ipso facto, to grant the privileged character of first-personal access to mental events – and insofar as that is preeminent among the ideas conducive to self-grasping, surely the only option consistent with Buddhism’s tradition-orienting no-self doctrine is to foreclose that possibility, which Candrakīrti does by denying from the outset that distinctively first-personal access even makes sense. What’s more, I have said that Śāntarakṣita takes the distinctiveness of first-personal access as just given – that cognition is distinguished from inanimate things by its svasaṃvitti, he thinks, is just a basic phenomenological

33

While those with “Svātantrika” predilections in the matter of interpreting Madhyamaka often charge Candrakīrti with leaving us no way to distinguish between more or less “true” conventions, Candrakīrti actually has something to say about that. His standard response is exemplified in his comment on Madhyamakāvatāra 6.41, which clarifies that while nothing is intrinsically existent, there is nevertheless a difference between the kinds of things that typically come into view, and altogether nonexistent things: “For example, someone with diseased eyes sees things like non-existent hairs, but not such other things as a donkey’s horn or the son of a barren woman (though these are likewise non-existent).” (Translated from La Vallée Poussin 1970: 130.).

176

Arnold

datum, which must therefore be addressed by any philosophical account. Doesn’t that idea warrant Candrakīrti’s concern? It is clear from some contemporary interpretations of Candrakīrti’s approach that it is the supposed givenness of this phenomenological datum that sets off alarms. After all, many interpreters (the present writer included) have argued that some of Candrakīrti’s most promising arguments – particularly those directed at the revisionary epistemological project of Dignāga – are in much the same spirit as Wilfried Sellars’s cogent critique of the “myth of the given.”34 If it’s right that Śāntarakṣita accepts the subjectivity of awareness as a phenomenological datum (“given”), it may seem to follow that Candrakīrti – justly appreciated for his veritably Sellarsian critique of givenness – is right to suppose the idea problematic. That idea is prominently recurrent in Jay Garfield’s Engaging Buddhism (2015), which is much concerned to resist “any view according to which the mind is self-revealing, and immediately available to consciousness.” Garfield thus holds that the problem identified by Candrakīrti’s critique of svasaṃvitti is that if cognition “were self-revealing, if it existed in the way it appears, if introspection were inherently veridical,” the mind in that case “would be non-deceptive, and would exist ultimately.” (2015: 85) That would be a particularly problematic entailment from a Buddhist perspective; after all, the Buddhist tradition’s diagnosis of the human predicament is that suffering results from the extent to which we systematically misunderstand ourselves (taking ourselves, in particular, as selves). If it’s right, then, that to admit the distinctiveness of first-personal access to mental states is, ipso facto, to affirm that our acquaintance with our own mental states is indubitably infallible, that would surely warrant any Buddhist’s (and certainly a Mādhyamika’s) resistance to theorizing subjectivity. It is really this concern – to refute infallibly exhaustive self-awareness – that finally drives this reading of Candrakīrti’s critique. That is clear from the kinds of factors that some of Candrakīrti’s modern defenders have considered relevant in refuting any claims regarding first-­ personal immediacy. Garfield emphasizes, for example, that there is much to our mental lives that is phenomenologically inaccessible – this, he says with reference to the Tibetan scholar whose reading of Candrakīrti guides his own, is “the Freudian point, at the heart of Tsongkhapa’s concern.” In order for svasaṃvitti to be consistent with that, one has to have “a pretty good understanding of one’s own cognitive life, something that even many adults fail to have.” (2015: 150) In a different scientific idiom, Garfield emphasizes to similar effect that the sense of ego is “constructed, and constructed by means of innumerable unconscious cognitive processes. Those processes lie below the 34

For a good way into these scholarly conversations, see the essays in Garfield 2019.

Should Mādhyamikas Refute Subjectivity?

177

level of introspectibility, and so never by themselves have this kind of subjective character.” (2015: 210) Garfield thus argues that claims regarding the supposedly distinctive immediacy of first-personal awareness are refuted by the fact that the many enabling conditions of our mental lives vastly exceed introspection – by the facts, e.g., that we are not, as psychoanalysis shows, transparent to ourselves, and that introspection fails to yield any knowledge at all of the neurophysiological underpinnings of conscious experience. As Garfield elsewhere expresses the upshot of Candrakīrti’s critique, if self-awareness is admitted even as conventionally real, “meditation and cultivation would be pointless; in our own [conscious experience], cognitive science would be complete.” (2006: 223) Given that presupposition, Garfield finds it apt to characterize Candrakīrti’s critique of svasaṃvitti as being, along with his arguments against Dignāga’s account of perception and perceptibles, “another attack on the Myth of the Given much as developed by Sellars in the 20th century.” (2015: 85) And the upshot of Candrakīrti’s supposed extension of the Sellarsian critique is, for Garfield as for Siderits, that “our own experience … is as opaque to us and as deceptive to us as the objects we encounter”; again, “the mind, and even consciousness, are hidden, rather than manifest phenomena, known only by inference, and through imperfect processes.” (2015: 75, 170) Gilbert Ryle fueled misguided readings of his work as behaviorist in orientation when he wrote that “the sorts of things that I can find out about myself are the same as the sorts of things that I can find out about other people, and the methods of finding them out are much the same.”35 Garfield says more strongly that self-knowledge, “if it is to mean anything,” can only be “an understanding of our own cognitive lives and perceptual experience achieved by the same mechanisms by means of which we know external events.” (2015: 151; emphasis added) Throwing caution to the wind, Siderits similarly avers (Chapter 6) that Candrakīrti’s critique rightly shows that minds are “just as opaque to themselves as they are to other minds.” Siderits suggests, moreover, that what Mādhyamikas ought to have said by way of addressing the question to which self-awareness is supposedly the answer – viz., “how is consciousness itself cognized?” – is something along the lines of the Brahmanical Mīmāṃsā tradition’s proposal: it can only be from abductive inference that one becomes aware that it was to one’s own cognitions that anything experienced was present, and this because “the first-person and third-person cases are exactly alike” (emphasis added).36 35 36

Ryle 1949:155; emphasis added. On Ryle vis-à-vis behavio[u]rism, see Tanney 2015. According to the Mīmāṃsaka doctrine, the occurrence of our own cognitions is inferred from the fact that objects of awareness show up for us exhibiting the property of being known (jñātatā). For another development of the argument for thinking that is as

178

Arnold

Can it really be thought, however, that such claims fare better than Śāntarakṣita’s account in capturing what is conventionally true of cognition?37 Our framing question, recall, has to do not with the best philosophical analysis of awareness, but with whether Candrakīrti was right to deny that self-­awareness is even conventionally real. To the extent, however, that it is conventional truth that is at issue, it’s hard to see how the view that “the first-person and thirdperson cases are exactly alike” could be thought preferable to Śāntarakṣita’s more intuitively plausible contention that “cognition is distinct from all that consists of inanimate matter” (and that it is in virtue of cognition’s “intrinsic awareness” that that is so). It seems to me Śāntarakṣita rightly thinks he has identified a phenomenologically basic datum, and indeed that he has identified, to that extent, something veritably constitutive of conventional truth. The reasonableness of his contention finds support, I think, in Richard Moran’s characterization of the basic category mistake evident in the thought (attested, we have seen, by Garfield and Siderits) that the extent of our opacity to ourselves counts against the svasaṃvitti thesis: Whatever introspective access is, our understanding of it will have to make sense of its conceptual dependence on the level of commonsense psychological description. For the object of first-person awareness (on any account of it) is not all of psychological life, but primarily the states of mind identified under the categories of what is sometimes called “folk psychology” … (2001: 7) In Buddhist terms, that is of course to say (with Śāntarakṣita) that self-­ awareness does indeed pertain – indeed, constitutively pertains – to conventional truth. The attempts of Garfield and Siderits to bolster Candrakīrti’s critique of svasaṃvitti are, I suggest, misguided partly by the presupposition that the phenomenological givenness invoked by Śāntarakṣita entails all the problems that Candrakīrti rightly identified with respect to epistemic givenness. It is clear, certainly, that variations only on the latter problem are targeted by their arguments, which mostly consist in doubling down on the same lines of attack that Candrakīrti pursued. As I suggested at the outset, the deck is stacked in

37

Mādhyamikas ought also to have held, see Siderits 2011. I have myself entertained the idea that Mīmāṃsā’s account is tenable; see Arnold 2005: 208, 288n5. Siderits is not unaware of the claim’s counter-intuitiveness; immediately after saying that “the first-person and third-person cases are exactly alike,” he says: “It is in this claim of symmetry between first- and third-person attributions of consciousness that the opacity thesis meets its most vehement source of opposition.”

Should Mādhyamikas Refute Subjectivity?

179

Candrakīrti’s favor if it is presupposed that the problems addressed by svasaṃvitti are chiefly epistemological – if it is presupposed, as by Siderits, that svasaṃvitti must be understood as explaining what kind of act it is in virtue of which cognition is itself “cognized.” This presupposition, I said, occludes the possibility that svasaṃvitti might reasonably be understood as a way of rejecting views of cognition according to which anything further must be “done,” any additional intentional act performed, to explain how mental states are manifest to their subjects. That the tenability of such a view has not really been entertained is, in any case, clear from the extent to which would-be bolsterers of Candrakīrti’s case often beg the questions most at issue, doubling down on arguments that are relevant only if one does not take seriously Śāntarakṣita’s contention that svasaṃvitti, on his understanding thereof, is not to be understood “in terms of a relation of action and agent.” The arguments adduced as bolstering Candrakīrti’s refutation, then, almost invariably hinge on the presupposition that svasaṃvitti must itself exhibit the same intentional structure as particular acts of knowing. That idea is presupposed, for example, by arguments premised on the question of what the content or object of svasaṃvitti could be (which represents, I’ve said, a basic category mistake; that any particular cognition has an object does not mean that the subjectivity that defines it as a cognition must likewise have one). With respect, for example, to the familiar idea that “I know my own life from a first-person perspective,” Garfield says the “sleight of hand” that makes the idea intelligible becomes apparent if we ask “whether when an object is ‘given’ – that is, when I become perceptually aware of an object – my experience is also given – that is, made aware to me [sic] – as a second object of awareness.” (2015: 163) The idea of svasaṃvitti, he further says, makes sense only if experience is “more than the appearance of [an] object, if it is a special inner appearing. And there is simply no reason to think that when I am aware of a blue sky, I am also aware of a second thing, namely its appearance.” (2015: 165) Entertaining the idea that svasaṃvitti is best understood not in terms of “inner appearing” but in terms of Thomas Nagel’s idea that there is “something it is like” to be conscious, Garfield again expresses the point in terms of whether anything epistemically given could count as answering the question of what, exactly, it is “like”: “why not argue that what makes a perceptual state conscious is for there to be something that it is like to be in it, and then simply treat that something as other than the state? But what?” (2015: 207) On that reading, the title question of Nagel’s famous 1974 article – “What Is It Like To Be a Bat?” – must be understood so as to admit of a determinate answer, one presumably consisting in demonstrative reference to an additional “object” of inner appearing (as though one might answer, “why, it’s like this! ,”

180

Arnold

said while indicating the relevant presentation). But that, I think, no more takes Nagel’s idea seriously on its own terms than Candrakīrti’s critique has purchase with respect to Śāntarakṣita’s idea. Such arguments persistently presuppose that svasaṃvitti can only be an act, and that among the issues requiring explanation therefore is whether there is anything it makes sense to identify as the object of that act. Insofar as it is thus presumed that Śāntarakṣita could only have been concerned with epistemological questions – questions about whether and how anything really is “known” by this particular kind of awareness – it is difficult to see the possibility that Śāntarakṣita might have been stalking altogether different problems. I am suggesting in this regard that the epistemological preoccupations of Candrakīrti’s would-be defenders occlude the recognition that Śāntarakṣita’s is a basically conceptual point – that svasaṃvitti, on his understanding, is simply criterial for the identification of any token of the type cognition, and that to deny his contention is in effect to deny that we have the kind of experience we have. Such “idealist” intuitions as Śāntarakṣita retains are, to that extent, better characterized as transcendental intuitions. His thought, I thus take it, is that our having such experience as we do is in a sense absolute for philosophy – not, to be sure, “absolute” in anything like the ontological sense that Buddhists understand by ultimately existent (paramārthasat), but simply in the sense that there is no perspective from which any claims can be entertained other than the one we occupy, which fact necessarily figures, therefore, as the starting point for philosophical inquiry. Śāntarakṣita’s point, in other words, is thus that just insofar as cognitions occur, their defining characteristics (including svasaṃvitti or subjectivity) are incontrovertibly instantiated. As J.J. Valberg says to similar effect, “given that there is such a thing as my [cognitions]” – and that is something that cannot coherently be denied – “[their] existence, i.e., the existence of my [cognitions], cannot but already be manifest.” That is, it would not be cognitions that there is “such a thing as” if the items at issue were not such as to be manifest to their subjects. And, Valberg says, “this fact of manifestness – which is a fact about the fact of the existence of my [cognitions] – this metafact is self-­ consciousness.” This is, I suggest, much as Śāntarakṣita says in emphasizing that svasaṃvitti is not, in fact, such as to admit of the same analysis to be given of acts – and the upshot for Śāntarakṣita is, just as Valberg concludes, that this “self-­consciousness” therefore “does not generate a regress.”38

38

Valberg (2007: 193–194), here continuing the discussion quoted above (see note 24). I have taken the liberty of substituting the word cognitions where Valberg has “consciousness”;

Should Mādhyamikas Refute Subjectivity?

181

Even if that’s right, and if something worth the name “self-awareness” therefore escapes Candrakīrti’s supposedly decisive critique, there nonetheless remains something that a Mādhyamika, in particular, could be expected to think problematic in Śāntarakṣita’s account. Even, then, if the idea of distinctively first-personal access need not by itself entail any epistemological or metaphysical commitments (about, say, infallibility or “selves”), a Mādhyamika of Candrakīrti’s ilk is apt to think there is problem enough in merely distinguishing first-personal access. To allow that for any cognition there is just one subject with first-personal access to its manifestation is, it would seem, surely to allow that one subject has privileged access. What’s more, I have allowed that among the tenable idealist intuitions Śāntarakṣita’s Madhyamaka retains is the thought that just that fact represents a phenomenologically basic datum. Candrakīrti, for one, would surely resist the very idea of “phenomenologically basic data,” to say nothing of the first-personal access that Garfield and Siderits take him to have refuted. Oughtn’t a Mādhyamika to suppose, then, that just insofar as it is invoked as phenomenologically basic, the idea of svasaṃvitti, ipso facto, exemplifies the svabhāva criterion (that of having “intrinsic identity”) that is the principal target of Madhyamaka critique? Mark Siderits has in this regard said (in personal correspondence) that “as a Mādhyamika, Śāntarakṣita cannot help himself to the notion of a ‘best’ form of conventional truth.”39 This introduces considerations relevant to the bifurcation of the Madhyamaka tradition into “Svātantrika” and “Prāsaṅgika” lines of interpretation, regarding which it here suffices to say the division reflects a divergence in intuitions over whether, in effect, conventional truth will admit of philosophical systematization, or whether instead conventional truth is exemplified just by common-sense intuitions (by “folk psychology,” in the idiom Moran invoked). Candrakīrti, an ordinary language philosopher avant la lettre, held the latter, “Prāsaṅgika” view, and thus thought that any philosophically revisionary theory of the conventional inevitably renders that unintelligible. Śāntarakṣita, in contrast, exemplifies not only a distinctive synthesis of Madhyamaka and Yogācāra, but also (and just in virtue of that) typifies the “Svātantrika” approach to arguments for Madhyamaka – which just is to say that he, unlike Candrakīrti, is willing to countenance the idea that we can, by rightly using the tools of philosophy, more nearly approach reality even at the level of conventional truth.40

39 40

Valberg himself does not take the latter as a substantive term, and so it is not out of place to foreclose objections based on that thought. Personal correspondence, 20 January 2020; cited with permission. For a good introduction to this debate, see the essays in Dreyfus and McClintock 2003.

182

Arnold

That is surely an important point to be made about the evidently different views of conventional truth held by these Mādhyamikas, and it provides some context for understanding how Śāntarakṣita could reasonably take the best account of conventional truth to involve idealist conclusions, even as Candrakīrti could reasonably think something so counter-intuitive couldn’t possibly make sense as “conventionally” true. I cannot, of course, give a full account here of all the considerations that weigh for and against these alternative approaches, and it might thus have to suffice here to say, by way of response to Siderits’s scruples about a “best” form of conventional truth, that it begs questions centrally at issue between Svātantrikas and Prāsaṅgikas to take Candrakīrti’s thought as epitomizing a normatively “Madhyamaka” view. I can, though, say a bit more than that, as the question here at issue (that of the status of svasaṃvitti) represents, I think, an especially illuminating case of something embraced (by Śāntarakṣita as I read him) as philosophically basic – as a phenomenologically basic datum such as necessarily constrains any analysis – of which it is, at the same time, nonetheless denied that there pertains any “intrinsic identity” (svabhāva). Indeed, on my reading this issue paradigmatically exemplifies the case for thinking it tenable that there be something which is philosophically basic, and which is nonetheless just conventionally real. (I am suggesting, that is, that svasaṃvitti represents a particularly good test-case for the intuition that any conventionally real phenomena can without distortion be subject to philosophical refinement.) By way of appreciating the sense it could make thus to hold that subjectivity is at once philosophically basic and conventionally real, consider Siderits’ elaboration of the foregoing contention that Śāntarakṣita, as a Mādhyamika, is not entitled to the idea of any uniquely significant form of conventional truth: Some formulations of how things are turn out to be better – for certain purposes – than others. That’s all. And this is where the commensurability difficulty comes in, if we take Śāntarakṣita’s account of cognition to be intended as supporting the claim that this is what we must say about cognition in the end. He can say that his version of svasaṃvedana is a useful model for us to employ when we think about how cognition might work. But he would have to agree that for certain purposes other models might work better. This, I think, cuts to the heart of the matter; whether it’s right that Śāntarakṣita must agree that “other models might work better” is, I suggest, just what must be decided if we are to make sense of Śāntarakṣita’s prima facie counter-­ intuitive contention that Yogācāra idealism makes sense as the best account of conventional truth.

Should Mādhyamikas Refute Subjectivity?

183

My sense is that Śāntarakṣita would deny that svasaṃvitti is merely a “useful model” for thinking about cognition, and that he would argue that alternative accounts of the same fact cannot, in fact, coherently be imagined. Indeed, what I mean in saying that svasaṃvitti represents a phenomenologically basic datum for him just is that he understands this as a condition of the intelligibility of anything else we could know. The idea that we we could do without Śāntarakṣita’s idea of svasaṃvitti – that that is “conventionally” true merely in the sense that it has been (or could be) chosen from among a number of equally viable alternatives – is, on my reading, tantamount to a misguided idea thus epitomized by Daniel Dennett: “There seems to be phenomenology … but it does not follow from this undeniable, universally attested fact that there really is phenomenology.” (1991: 365–366) Galen Strawson’s apt rejoinder seems to me just as Śāntarakṣita would say: Dennett’s contention “fails immediately if it is taken as an objection to the claim … that we can be certain both that there is experience and that we can’t be radically in error about its nature. It fails for the simple reason that for there to seem to be rich phenomenology or experience just is for there to be such phenomenology or experience.”41 What is phenomenologically given is, to be sure, just conventionally real; subjectivity neither is nor discloses any kind of ultimately real existent. There is no kind of thing that subjectivity really is – only the fact (contingent but nonetheless incontrovertible) of there seeming to innumerable subjects that there is a world. But while there is nothing that subjectivity is, nonetheless it cannot coherently be claimed that this seeming is “really” something else; anything else it could be would itself be intelligible only given, in the first instance, that it seems so. Look: I get why a Mādhyamika might nonetheless resist the very idea that anything could thus be “philosophically basic”; it is not hard to hear that claim as tantamount to the svabhāva idea that Madhyamaka rightly denies. Insofar as that is the main objection, though, the concern can be met, I think, by emphasizing that what Madhyamaka chiefly aims to refute is an ontological criterion. That is why there is, on my view, no particular problem with a Mādhyamika’s embracing the formulation (long associated with Siderits himself) that the ultimate truth, for Madhyamaka, is that there is no ultimate truth – a formulation whose air of paradox I would dispel by unpacking it as saying it is really true (it is as a Buddha knows) that there can be no ultimately real existents. The upshot of this, I take it, is that there is no problem with a Mādhyamika’s allowing, as I’ve put it, that our having such experience as we do is “absolute” for philosophy; that is only to say that the perspective from which inquiry is undertaken is

41

Strawson 2008: 55n; cf. verses 53–56 of Śāntarakṣita’s Madhyamakālaṃkāra.

184

Arnold

itself basic for any inquiry, a thought that entails no commitments about what the perspective of our inquiry is, what kind of “thing” it must be.42 And that, finally, is why none of this need be thought at odds with the Buddhist no-self doctrine, either – which is, of course, as we must show if Śāntarakṣita’s view of self-awareness is to make sense as part of any Mādhyamika elaboration of Buddhist commitments. Here it is significant that notwithstanding their importantly different intuitions regarding conventional truth, Śāntarakṣita and Candrakīrti agree in thinking it significant that Madhyamaka recuperates the value of conventional truth; what’s more, they both think so as against the understanding of the two truths they find epitomized by the reductionist arguments of the Abhidharma schools of thought. According to the latter sort of view, an ultimately true account explains conventional truth, making reference only to ultimately real existents that have explanatory purchase with respect to conventional phenomena whose mode of being is, in contrast, deficient. That is why it makes sense that the principal target of Madhyamaka critique is an ontological criterion; regarding the ontological primitives posited as ultimately existent by Abhidharma schools, the concern is to show that no such things could make sense as what the phenomena they supposedly constitute “really” are. And the reasons for that epitomize Madhyamaka’s guiding thought that ultimate truth is so much as intelligible only relative to conventional truth. Typically Madhyamaka arguments often work, then, by showing that putatively explanatory categories depend for their intelligibility on the very phenomena they purportedly explain – as, for example, it makes sense to individuate ways of knowing (pramāṇa) only insofar as one already has the idea of things known thereby (prameya), or to individuate a locus of motion only insofar as one already has the idea of motion. Madhyamaka’s central insight, I 42

There are many contemporary non-perceptual accounts of self-awareness, I think, that might support this conclusion. With respect, for example, to the construal of first-person reference as comparable to indexical demonstratives, Sydney Shoemaker argues that “[t]he way out of this incoherence is to abandon completely, not just in part, the perceptual model of self-knowledge.” (1968: 564) In a grammatical idiom that Indian philosophers would surely recognize, Elizabeth Anscombe has argued to similar effect that it is “the (deeply rooted) grammatical illusion of a subject” that generates the many pitfalls here, and the dispute is bound to be “self-perpetuating, endless, irresoluble, so long as we adhere to the initial assumption, made so far by all the parties to it: that “I” is a referring expression.” (1981: 36, 32) What must be understood, Anscombe thus concludes, is that “‘I’ is neither a name nor another kind of expression whose logical role is to make a reference, at all.” (1981: 32) Consider, too, Moran: “It is modeling self-consciousness on the theoretical awareness of objects that obscures the specifically first-person character of the phenomenon, whether or not this theoretical perspective takes the specific form of the perceptual model of introspection.” (2001: 32).

Should Mādhyamikas Refute Subjectivity?

185

am thus suggesting, is that ultimate truth cannot coherently be conceived as altogether superseding conventional truth – and this because it is only at the level of conventional truth that claims regarding the ultimate truth can be so much as entertained.43 While it is true, then, that we are not selves – true, that is, that persons will indeed admit of the kind of reductive redescription pursued by the Abhidharma tradition – that cannot be taken to imply that there is something else that we really are. It is not the case, then, that the reason there are no selves is just that what really exists is causally continuous series of impersonal events; the more basic point, rather, is that there is nothing at all that being a person “really” consists in. “Self-grasping” most subtly consists, to that extent, precisely in attachment to the idea of oneself as being an enduring entity – in attachment, as we might also put it, to the idea that it is only as objectively existent (only as a substance) that subjectivity could be real. Nothing about Śāntarakṣita’s idea of svasaṃvitti, however, entails that idea; indeed, such a conclusion could be understood as just what his version of the doctrine denies. The concern for consistency with the no-self doctrine, then, doesn’t clinch a case for Candrakīrti’s critique, either. Accordingly, I take it that there is good reason to think Śāntarakṣita’s formulation of svasaṃvitti escapes Candrakīrti’s influential critique, and that it is misguided to deny that anything worth the name “self-awareness” could find a place in our conventionally valid practices. To the contrary, it is reasonable to admit the distinctiveness of firstpersonal access to experience, and yet to emphasize that that is not at all to say that we are therefore transparent to ourselves; Śāntarakṣita’s claim entails no commitment either to indubitably known sense data, or to any acquaintance with, e.g., the goings-on in one’s central nervous system. Indeed, insofar as a first-person perspective constitutively pertains (pace Moran) to the “states of mind identified under the categories of what is sometimes called ‘folk psychology,’” it stands to reason that svasaṃvitti makes sense, in fact, only as conventionally real. Notwithstanding the arguments of great thinkers like Candrakīrti and Tsong-kha-pa, then, it seems to me there is no good reason to think Mādhyamikas are committed to denying the common-sense truth that cognitions are conscious.

43

As I have argued at length in Arnold (2012), Mādhyamika arguments are to this extent comparable to some contemporary arguments against physicalism – arguments to the effect that (in Sellarsian terms) the “scientific image” of human experience is necessarily dependent on the “manifest image,” which therefore cannot coherently be thought to be explained away by the former.

186

Arnold

References Anscombe, G.E.M. 1981. “The First Person,” in Metaphysics and the Philosophy of Mind: The Collected Papers of G.E.M. Anscombe, vol. ii, pp. 21–36. Minneapolis: University of Minnesota Press. Arnold, Dan. 2003. “Candrakīrti on Dignāga on Svalakṣaṇas.” Journal of the International Association of Buddhist Studies 26 (1): pp. 139–174. Arnold, Dan. 2005. Buddhists, Brahmins, and Belief: Epistemology in South Asian Philosophy of Religion. New York: Columbia University Press. Arnold, Dan. 2005a. “Materials for a Mādhyamika Critique of Foundationalism: An Annotated Translation of Prasannapadā 55.11 to 75.13.” Journal of the International Association of Buddhist Studies 28 (2): pp. 411–466. Arnold, Dan. 2010. “Self-Awareness (svasaṃvitti) and Related Doctrines of Buddhists Following Dignāga: Philosophical Characterizations of Some of the Main Issues.” Journal of Indian Philosophy 38: pp. 323–378. Arnold, Dan. 2012. Brains, Buddhas, and Believing: The Problem of Intentionality in Classical Buddhist and Cognitive-scientific Philosophy of Mind. New York: Columbia University Press. Arnold, Dan. (Forthcoming.) A Madhyamaka Reader: Buddhist Thought in Classical Indian Philosophy. Columbia University Press (to be published in the series “Historical Sourcebooks in Classical Indian Thought”). Berkeley, George. 1996. Principles of Human Knowledge / Three Dialogues. Ed. Howard Robinson. New York: Oxford University Press. Bilgrami, Akeel. 2006. Self-Knowledge and Resentment. Cambridge: Harvard University Press. Cavell, Stanley. 1979. The Claim of Reason: Wittgenstein, Skepticism, Morality, and Tragedy. New York: Oxford University Press. Dennett, Daniel. 1991. Consciousness Explained. Boston: Little, Brown. Dreyfus, Georges. 2011. “Self and Subjectivity: A Middle Way Approach.” In Siderits, Thompson, and Zahavi 2011, pp. 114–156. Dreyfus, Georges, and Sara McClintock, eds. 2003. The Svātantrika-Prāsaṅgika Distinction: What Difference Does a Difference Make? Boston: Wisdom Publications. Ganeri, Jonardon. 1999. “Self-Intimation, Memory and Personal Identity.” Journal of Indian Philosophy 27: pp. 469–483. Garfield, Jay. 2006. “The Conventional Status of Reflexive Awareness: What’s at Stake in a Tibetan Debate,” Philosophy East and West 56 (2): pp. 201–228. Garfield, Jay. 2015. Engaging Buddhism: Why it Matters to Philosophy. New York: Oxford University Press. Garfield, Jay, ed. 2019. Wilfrid Sellars and Buddhist Philosophy. New York: Routledge.

Should Mādhyamikas Refute Subjectivity?

187

Hattori Masaaki. 1968. Dignāga, On Perception, Being the Pratyakṣapariccheda of Dignāga’s Pramāṇasamuccaya from the Sanskrit Fragments and the Tibetan Versions. Cambridge, MA: Harvard University Press. Huntington, C.W., with Geshe Namgyal Wangchen. 1989. The Emptiness of Emptiness: An Introduction to Early Indian Mādhyamika. Honolulu: University of Hawaii Press. Ichigō Masamichi, ed. 1985. Madhyamakālaṃkāra of Śāntarakṣita with his own commentary or Vṛtti and with the subcommentary or Pañjikā of Kamalaśīla. Kyoto: Buneido. Kellner, Birgit. 2010. “Self-Awareness (Svasaṃvedana) in Dignāga’s Pramāṇasamuccaya and -vṛtti – A Close Reading,” Journal of Indian Philosophy 38: pp. 203–231. La Vallée Poussin, Louis de, trans. 1910. “Madhyamakāvatāra: Introduction au Traité du milieu de l’Ācārya Candrakīrti, avec le commentaire d l’auteur, traduit d’après la version tibétaine” [6.1–6.80], Le Muséon ns. 11: pp. 271–358. La Vallée Poussin, Louis de, ed. 1970. Madhyamakāvatāra par Candrakīrti: Traduction tibétaine. Bibliotheca Buddha, vol. 9. Osnabrück: Biblio Verlag (reprint). La Vallée Poussin, Louis de, ed. 1970a. Mūlamadhyamakakārikās (Mādhyamikasūtras) de Nāgārjuna, avec la Prasannapadā Commentaire de Candrakīrti. Bibliotheca Buddhica, vol. 4. Osnabrück: Biblio Verlag (reprint). Li Xuezhu. 2015. “Madhyamakāvatāra-kārikā Chapter 6.” Journal of Indian Philosophy 43: 1–30. MacDonald, Anne. 2015. In Clear Words: The Prasannapadā, Chapter One. Vol. i: Introduction, Manuscript Description, Sanskrit Text; Vol. ii: Annotated Translation, Tibetan Text. Vienna: Verlag der Österreichischen Akademie der Wissenschaften. Moran, Richard. 2001. Authority and Estrangement: An Essay on Self-Knowledge. Princeton: Princeton University Press. Moran, Richard. 2011/12. “Cavell on Outsiders and Others,” Revue Internationale no. 256: 239–254. Moriyama Shinya. 2010. “On Self-Awareness in the Sautrāntika Epistemology.” Journal of Indian Philosophy 38: 261–277. Moriyama, Shinya. 2014. “Ratnākaraśānti’s Theory of Cognition with False Mental Images (*alīkākāravāda) and the Neither-One-Nor-Many Argument,” Journal of Indian Philosophy 42 (2): 339–351. Nagel, Thomas. 1974. “What Is It Like To Be a Bat?” The Philosophical Review 83/4: 435–450. Peirce Edition Project, 2010. Writings of Charles S. Peirce: A Chronological Edition, Volume 8: 1890–1892. Bloomington: Indiana University Press. Pippin, Robert B. 1989. Hegel’s Idealism: The Satisfactions of Self-Consciousness. Cambridge: Cambridge University Press. Ryle, Gilbert. 1949. The Concept of Mind. Chicago: University of Chicago Press.

188

Arnold

Searle, John. 1997. The Mystery of Consciousness. New York: New York Review of Books. Shastri, Swami Dwarikadas, ed. 1978. Ślokavārttika of Śrī Kumārila Bhaṭṭa, with the Commentary Nyāyaratnākara of Śrī Pārthasārathi Miśra. Varanasi: Tara Publications. Shastri, Swami Dwarikadas, ed. 1997. The Tattvasaṅgraha of Ācārya Śāntarakṣita, with the “Pañjikā” Commentary of Ācārya Śrī Kamalaśīla. 2 volumes, continuous pagination. Varanasi: Bauddha Bharati. Shoemaker, Sydney. 1968. “Self-Reference and Self-Awareness.” The Journal of Philosophy, vol. lxv, No.19: 555–567. Siderits, Mark. 2007. Buddhism as Philosophy. Indianapolis: Hackett. Siderits, Mark. 2011. “Buddhas as Zombies: A Buddhist Reduction of Subjectivity.” In Siderits, Thompson, and Zahavi 2011, pp. 308–331. Siderits, Mark, Evan Thompson, and Dan Zahavi, eds. 2011. Self, No Self? Perspectives from Analytical, Phenomenological, & Indian Traditions. New York: Oxford University Press. Strawson, Galen. 2008. Real Materialism and Other Essays. New York: Oxford University Press. Tanney, Julia. 2015. “Gilbert Ryle,” Stanford Encyclopedia of Philosophy (https://plato .stanford.edu/entries/ryle/). Thompson, Evan. 2011. “Self-No-Self? Memory and Reflexive Awareness.” In Siderits, Thompson, and Zahavi 2011, pp. 157–175. Valberg, J.J. 2007. Dream, Death, and the Self. Princeton: Princeton University Press. Williams, Paul. 1998. The Reflexive Nature of Awareness: A Tibetan Madhyamaka Defence. London: Curzon.

Chapter 6

Self-Knowledge and Non-Self Mark Siderits Abstract Carruthers (2011) claims that the transparency thesis – the thesis that the mind has transparent access to its own states – may be a “human universal,” something that humans at all times and in all cultures are strongly disposed to accept as true. One consequence of the transparency thesis is that there is a significant asymmetry between knowledge of one’s own mental states and knowledge of the mental states of others, the latter states being known only through inference from sensory cues. The assumption that we have privileged and possibly infallible access to our own mental states may nonetheless be questioned. One reason to question it derives from empirical investigation of the neurophysical basis of consciousness. A quite different reason stems from the denial of conscious agency that follows from the philosophical thesis of non-self. But in either case the asymmetry between self-knowledge and knowledge of other minds dissolves: knowledge of one’s own mental states is just as much the product of an inference-performing mind-reading faculty as knowledge of the minds of others. One’s own mind is just as opaque to oneself as are others’ minds. This paper investigates the role that the opacity thesis may play in a Buddhist account of consciousness. One claim is that Dignāga’s reflexivity thesis (the theory of svasamvedana) represents an unnecessary concession to the transparency thesis. But the chief question to be addressed is whether the opacity thesis has consequences that are incompatible with core Buddhist commitments. Recent espousals of the thesis grow out of attempts at naturalizing the mental. Would full-scale embrace of Kumārila’s opacity view (an embrace that is at least hinted at by Śāntideva) lead Buddhists in the direction of physicalism?

Keywords opacity – transparency – meta-cognition – svasamvedana – r­eflexivity – Dignāga – Śāntideva – HOT

…

© koninklijke brill nv, leiden, ���1 | doi:10.1163/9789004440913_010

190

Siderits

“How is consciousness itself cognized?” This question has played an important role in the development of Buddhist views about the nature of consciousness (vijñāna) and cognition (jñāna). Among the possible answers there is one – the opacity thesis – that I believe best coheres with what I take to be core Buddhist commitments. According to this thesis, while we may say that we are aware that we are conscious, we are never aware of consciousness as such. Cognitions are only cognized indirectly, through a kind of abductive inference. This means there is no essential asymmetry between first-person and third-person attribution of mental states. Minds are, in this respect, just as opaque to themselves as they are to other minds.1 We are accustomed to hearing that Buddhists subscribe to the thesis that cognitions possess the interesting property of being reflexively aware (svasamvedana). And of course this view is extensively developed and argued for by Dignāga and his successors in the Yogācāra-Sautrāntika line. (For further discussion see Chapters 4 and 5 by Coseru and Arnold in this volume.) But this view, which is a version of the transparency thesis, has its difficulties, so I should like to explore the less well-developed opacity thesis. In order to investigate how this might fit into a larger Buddhist project, I must begin by saying something about what I take such a project to be. I apologize for going over things that I realize are known to anyone knowledgeable about the Buddhist tradition. Sometimes, though, it can be helpful to say things that we think of as going without saying. Buddhism is like other Indian soteriological projects insofar as the aim of its practice is to overcome existential suffering. It also broadly agrees with other schools about the source of suffering, which is said to stem from our ignorance of our true identity. It is this ignorance that is thought to give the prospect of our mortality the power to alienate us from all our projects. What sets the Buddhist path apart from others lies in the specific form Buddhists claim our error takes: thinking there is an “I” that might have an identity. There is no such thing as the enduring substantial self posited by other philosophers. The concept of 1 Here it is important to be clear about what is not claimed by either thesis. The question at issue is whether we have direct, unmediated access to the occurrence of our own cognitions. This says nothing about whether we have privileged access to the contents of such states. As I shall be using it here, the transparency thesis is neutral with respect to the question whether we are ever infallible with respect to what it is that we are aware of. It is widely agreed that we are, e.g., sometimes wrong about our emotional states. A transparency theorist might agree that what I took as painful might really be indifferent. This does not directly contradict the transparency thesis as I understand it here. What such evidence may show, however, is that sometimes self-attribution of conscious mental states takes the form of confabulation. And sometimes confabulation involves the positing of a state that did not occur.

Self-Knowledge and Non-Self

191

an “I” is a mere shorthand device for keeping track of a causal series of impermanent psychophysical elements. Still one can understand all this and yet continue to experience suffering. The culprit here is the “I”-sense possessed by all but fully enlightened beings. What unites the wide variety of Buddhist practices is that all are thought to help extirpate this “I”-sense. Carruthers (2011) claims that the transparency thesis – the thesis that the mind has transparent access to its own states – may be a “human universal,” something that humans at all times and in all cultures are strongly disposed to accept as true.2 Like some Buddhists, Carruthers thinks the transparency thesis is false. But if he is right that it is a human universal, this would help explain why the “I”-sense is so widely shared and so difficult to expunge. Anyone who accepted the transparency thesis would be inclined to believe that each of us has privileged and perhaps infallible access to the occurrence of their own mental states. This in turn would fuel the notion that the mind is an inner subjective realm quite unlike the external world of publicly observable things. And even if one stops short of positing a single enduring entity as the subject of one’s mental states, its acceptance will still motivate intuitions that make such a posit seem plausible. Nowadays the term of art for the simple self posited by philosophers like Descartes, Vātsyāyana and Kumārila is “soul pellet.” As the disparaging tone of the term suggests, there are currently very few philosophers inclined to accept the view that there is such a thing. Buddhist philosophers likewise thought the disproof of such a self was relatively straightforward. They recognized, though, that its refutation left the “I”-sense unscathed.3 If the transparency thesis is a human universal, this would help explain why considerably more work needs to be done. For if one thinks that one has authoritative and direct access to one’s own mental states, one will likely retain the sense that these states must have some subject. 2 For a nice expression of the roots of the transparency thesis in common sense see Bogdan 2010: 70. It should be noted that “transparency” is being used differently here than in certain other discussions concerning knowledge of our own mental states. When, for instance, beliefs are said to be transparent, what is meant is that we come to know of our beliefs by looking at the world and not at some inner mental store. If we think of beliefs as inhabitants of the mind, to call them transparent in this sense is to say we see them by looking without and not within; they are said to be transparent in the sense that we see right through them. By contrast, what Carruthers calls the transparency thesis has it that our skulls are transparent to their owners – we see right through them to the mind within. 3 Thus Candrakīrti likens someone who thinks that refuting this “philosopher”s self’ will extirpate the “I”-sense to one who, knowing there are snakes living in the walls of their house, comforts themselves with the thought that there is no elephant inside (mav 6.141).

192

Siderits

The Buddhist reductionist strategy of reductively analyzing the person into a causal series of psychophysical elements is responsive to this difficulty. It aims to explain why the “I”-sense is so widely shared, given that it conveys the false belief that there is an enduring person. The explanation is that while strictly speaking there is nothing more than the causally linked psychophysical elements in serial succession, still given the vicissitudes of ordinary life it is generally useful to have a single concept that can be used to refer to these linked elements indiscriminately. Indeed the utility of this device is so great that we end up using it unthinkingly, forgetting that the “I” is a mere enumerative term like “pair,” “score” and “multitude.” The result is that this many, masquerading as a one, gets taken to be a single enduring thing. The result is reinforcement of the “I”-sense. This Buddhist strategy may go some way toward answering doubts about non-self fueled by the imperviousness of the “I”-sense to extirpation by refutation. But new problems emerge from the details of the proposed reduction of the person to psychophysical elements. Among the skandhas is consciousness (vijñāna). And it is a piece of Buddhist doctrinal orthodoxy that a fully enlightened being is directly aware of all ultimately real compounded things as having the three marks of impermanence, suffering and non-self.4 Now if consciousness is among these ultimately real compounded things, it would seem to follow that the consciousness whereby all dharmas are directly cognized is among the objects it is directly aware of. And this would be odd, since it violates the seemingly sound principle of irreflexivity, which has it that an ultimately real entity cannot act on itself. A fingertip cannot touch itself, fire does not burn itself, a knife-blade cannot cut itself, so how could it be that consciousness cognizes itself? This difficulty was the subject of a major Abhidharma controversy (see Yao 2005). Some Ābhidharmikas claimed that the fully enlightened being’s sweep of all dharmas occurs in two successive moments: at t1 a cognition surveys all dharmas save the consciousness (citta) and the associated mental concomitants (caitta) that perform the sweep; at t2 a new cognition takes the consciousness and mental concomitants operating at t1 as its object and ascertains their being characterized by the three marks. Others gave a similar account only 4 I would speculate that this doctrine was meant to explain why a mere philosophical demonstration that all dharmas have the three characteristics is unable to dislodge the “I”-sense. The knowledge gained from such a demonstration would be said to lack the force necessary to undermine this deeply ingrained habit because this knowledge is mediated by concepts; only the direct awareness of a fully enlightened being would be deemed sufficient. But the issue is also addressed in terms of the boundary between conceptual and non-conceptual cognition discussed in Chapters 2 and 3 by Spackman and Tzohar (this volume).

Self-Knowledge and Non-Self

193

claimed that all this happens in a single moment that consists of distinct phases. Yet others held that consciousness and the associated mental concomitants stand in a relation of mutual reciprocal causation, so that while a given consciousness can ascertain the relevant properties of everything else, including the simultaneously occurring mental concomitants, those mental concomitants can at the same time perform a similar function with respect to that consciousness. Finally there were those who bit the bullet and claimed that just as light illuminates itself while illuminating other things, so consciousness can cognize itself while cognizing its object. This was a strictly intramural debate among Buddhists. By the time that Dignāga formulates his version of the reflexivity thesis (svasamvedana), Brahmanical schools have joined the discussion. But there is a sense in which the Brahmanical views of a real subject (satkāyadṛṣṭi), as well as those of the heterodox Buddhist Personalists (pudgalavādins), can be seen lurking in the background of the debate from the outset. The Abhidharma dispute concerned whether a cognition can cognize itself. But the pan-Indian debate of which this was a chapter concerns the question how cognitions are themselves cognized. The latter question might be answered without having to worry about violating the principle of irreflexivity, by stating that it is the cognitive agent, in the form of a self or a person, that cognizes cognitions.5 Then one might think of cognition as a mode (guṇa) of the self (the view of Nyāya), or as something that bears no expressible relation to the person that is its subject (the view of the Personalists). Of course the Buddhist orthodoxy is that there are no substances and so there is no cognitive agent. The ultimate reals or dharmas are to be understood as just tropes. But the Sāṃkhya school can be seen as espousing an equally pure trope-theoretic ontology. And yet they offer a proof of the existence of a self that uses as evidence precisely the point that the psychophysical elements are all cognizables and characterized by suffering. The Sāṃkhya strategy was to show that, given the irreflexivity principle, our ability to survey all the skandhas entails the existence of a witnessing that transcends the empirical. One sees the same strategy at work in the claim one sometimes encounters that the Buddha only meant to deny the existence of an empirical

5 Another possible strategy for making cognition of cognition compatible with irreflexivity is to claim (as do Dignāga and Dharmakīrti) that strictly speaking, subject- and object-poles of a cognition are one. One difficulty with this is that cognitions then turn out to be ineffable – an odd result for those who (like Coseru, this volume) insist that there is something it is like to cognize. A second difficulty (one developed by Kumārila) is that it becomes inexplicable how something simple and ineffable should come to seem to possess the dual structure of having both noematic and noetic aspects.

194

Siderits

self.6 On this view what one reports on when one says that one is conscious of seeing blue is not the being conscious (understood by Sāṃkhya as a pure unchanging principle of illumination and not an agent of cognition) but the manifest effects of its association with the psychophysical complex. The evidence the Buddha offered in his stock arguments for non-self would then actually prove just the opposite. This is the threat that may lie behind the Abhi­ dharma debate over whether it is one cognition or two that cognizes the three marks in all dharmas. Now Dignāga had in hand arguments he took to show that existing things must be momentary. This would allow him to return the focus of the cognitionof-cognition debate to the question whether a cognition might cognize itself. Since selves or persons must endure, and nothing ultimately real endures, the difficulty involved in saying all cognitions are cognized is not to be resolved by positing a non-empirical cognitive subject. And we know how he explained the cognition of cognition, and the arguments he used to support that account (see Kellner 2010). It is worth pausing for a moment to reflect on the consequences of accepting his reflexivity thesis, however. On the one hand it accords with the transparency thesis, insofar as it explains why we should be immediately aware of our own mental states. If it is true that cognitions bring about awareness not only of their objects but also of themselves, it would seem to follow that every conscious mental state intimates itself to the person whose state it is. But should a Buddhist want to embrace the transparency thesis, given its tendency to reinforce the sense that for each of us there is an inner subjective realm to which we have privileged access? Granted, when momentariness is brought into the account, this sense of subjectivity cannot serve as grounds for positing an enduring self. But the Buddha spoke of two mistaken views concerning the “I”: eternalism and annihilationism. The latter is the view that the subject is impermanent, undergoing annihilation shortly after it comes into existence. It is a mistaken view precisely because it involves the supposition that there is an experiencing subject that serves as a locus of meaning and value. (The life lesson that the annihilationist takes away from impermanence is that we should live wholly in the present.) For a Buddhist, embracing the transparency thesis looks like a case of not only inviting Māra in the door but seating him at the head of the table.7 6 See Albahari 2006 for a recent instance of this claim. 7 This is one important reason to question Coseru’s (Chapter 4) interpretation of Dignāga as proto-phenomenologist. Husserlian methods may or may not solve the “crisis of [European] philosophy,” but no Buddhist would hold that they solve the existential crisis of sentient beings. A second reason is that it is difficult to deny that Dignāga engages in serious ontology in the argument for idealism he constructs in Ālamabanaparīkṣā. The case for a

Self-Knowledge and Non-Self

195

There is a second problem for the reflexivity thesis that is independent of Buddhist soteriological aims. It follows from this thesis that in the first-person case, attributions of mentality involve no criteria of application. Whenever one is in a conscious mental state, one needs no evidence in order to affirm that one is in some mental state. The third-person case is quite different, however. We can only attribute mental states to others inferentially; we interpret their bodily behavior. Of course we are often unaware that we are doing so. When we see someone walk across the room and reach for the doorknob, we see them as wanting to open the door. Attribution of the intention comes automatically. But reflection shows that in this case our judgment comes quickly and effortlessly only because the scenario is stereotypical. We employ criteria of application in attributing states of consciousness to others. This should make us wonder if it is the same concept we are applying in the first-person and the third-person cases. It follows from the reflexivity thesis that attribution of what we take to be the same concept necessarily requires evidence in one sort of case and necessarily involves no criteria of application in another sort of case. That would seem to be grounds for taking there to be two quite different concepts involved.8 The upshot is that if what we mean by “conscious” proto-­phenomenologist reading turns on seeing Dignāga as trying to bypass metaphysics altogether. 8 Candrakīrti’s comments on Āryadeva’s cś 10.3, although directly concerned with the question of the existence of a self, can also be read as making this point. The passage reads as follows: “And this self does not exist. For if the self existed by its own nature, then just as it is the object (ālambana) of one “I”-sense (ahaṅkāra), so it would be the object of every “I”-sense. For in the world the heat of fire is not the intrinsic nature of anything that is not hot. Accordingly, if the self existed by its own nature, it would be “the self of all,” and the object (viṣaya) of [everyone’s] “I”-sense. And this is not so, for: ‘3. What is self to you is non-self to me, by the stipulation [concerning meaning of “self”] it is not; The concept [of self] does not apply with respect to impermanent entities.’ (CS 10.3) What is your self, the object of your “I”-sense, the object of [your] self-love, that is non-self for me, for it is the object of neither my “I”-sense nor [my] self-love. Since this is so, it does not accord with the stipulation [that the intrinsic nature of an entity should be equally present in all instances]. A self that does not accord with the stipulation does not exist by its own nature; abandon this unreal object, the superimposition of a self. [Objection:] If the self does not exist, what about the “I”-sense and self-love, given that the concept (kalpanā) does not apply with respect to impermanent entities? As was [already] argued for extensively, because the self is everywhere absent save where the intrinsically established skandhas are, the concept (kalpanā) of the self is created as an unreal superimposing on the impermanent entities, rūpa, feeling, perception, dispositions and consciousness, [using such terms as] “self,” “being,” “living thing,” “creature.” Self is conceived in dependence on the skandhas just as fire is conceived in dependence on

196

Siderits

is what we are aware of in our own case, then it is simply false that there are other conscious beings. Ratnakīrti was right and Dharmakīrti was wrong: solipsism follows from the Yogācāra-Sautrāntika view.9 It is not, though, as if Dignāga and Dharmakīrti are merely trying to accommodate the widespread intuitions in favor of the transparency thesis. In the memory argument they have powerful support for their reflexivity thesis.10 It starts by considering an opponent’s higher-order account of how a cognition may be cognized. Suppose this were done not reflexively but reflectively: upon perceiving blue, one introspects or “looks within” and notes the visual cognition that apprehended the blue colour. Now since the perception of blue is carried out by the visual faculty, while the introspective awareness of that visual cognition is performed by the distinct faculty of the inner sense, reflective cognition of the visual cognition must occur after the visual cognition has taken place. (Indeed on this model some time may have elapsed between the two apprehensions.) But then given the momentariness of all ultimately real things, it follows that the reflective cognition reported as “I see blue” involves memory: one introspectively cognizes the visual cognition only by remembering its earlier occurrence. (So strictly speaking the report should be “I saw blue.”) For it is only by means of memory that one can cognize an object that no longer exists. But there is memory only of that which was experienced earlier. So the visual cognition that one is now aware of (by means of memory) must have been experienced earlier. Now either that earlier experiencing occurred at the time of the visual cognition (in which case the cognition cognized itself), or else it occurred at some time after the visual cognition but prior to the present introspective cognition. In the latter case, memory must once again be involved, and an infinite regress threatens. So the possibility of subsequent

9 10

fuel. And having been examined in the five ways with respect to whether it is identical with or distinct from the skandhas, it does not exist with intrinsic nature; it is imagined by means of a mere concept (prajñapti), it is established that the self is conceptually constructed on the basis of impermanent compounded things (saṃskāras).” cśv 47–8 The argument here is reminiscent of the one attributed to Wittgenstein in the Appendix of Kripke (1982). For why solipsism is particularly problematic for Buddhists see Moriyama (Chapter 8). Dignāga’s argument is at psv 11d, Dharmakīrti’s is at pvin 1.40,11–41,13 ad 1.54cd. My reconstruction differs from that of Coseru (Chapter 4). It differs as well from that of Kellner 2011, who renders the argument question-begging by having the regress go forward in time. (A forward-moving regress may feature in another of Dignāga’s arguments.) This is not how Mīmāṃsā critics understood Dignāga’s argument from memory, and Dharmakīrti’s formulation of the memory argument seems to allow the same reconstruction. (Kellner and I now agree on how to understand Dignāga’s formulation of the argument from memory.)

Self-Knowledge and Non-Self

197

r­ eflective awareness of a cognition requires that the cognition of which one is then aware, say the cognition of blue colour, involved reflexive awareness of itself as the cognizing of blue. While it may be true that not every cognition is actually made an object of reflection, it seems that all cognitions are possible objects of reflection. So it then follows that every cognition must cognize itself.11 This argument does not, however, address the alternative higher-order thought account of cognition of cognition. On this view cognition is never itself directly cognized; attribution of conscious mental states is as much a matter of interpretation in the first-person case as in the third-person case. Reflective awareness of one’s own cognition is strictly a matter of reconstruction, so no memory is involved. I shall here call this the opacity thesis. This was Kumārila’s view, and it is at least hinted at in Prajñākaramati’s comments on Śāntideva’s bca 9.24cd. This view facilitates explaining the possibility of cognizing cognition without violating the principle of irreflexivity.12 It also avoids the threat of solipsism facing the reflexivity thesis. Of course if cognitions are never themselves directly cognized and only inferred, one must wonder what could serve as the reason in the inference whereby they are cognized. Normally a reason is ascertained in cases of co-occurrence of the property to be proven (sādhya) and the proving property (hetu), and on this view the property to be proven is never given in experience. This is why Kumārila calls the cognition of cognition a case of arthāpatti or inference to the best explanation (abduction). The inferential reason is the object’s having the property of being available to the agent for action and speech; and the best explanation of the object’s having this property is that the object was cognized by the agent. The best explanation of someone’s reaching for the doorknob is that they want to open the door, they believe that grasping the doorknob is the best way to achieve this aim, and they see the doorknob. Likewise the best explanation of my being disposed to say “blue” when asked the colour of the room I am in is that I see blue. The firstperson and third-person cases are exactly alike. 11

12

Montague (2017) similarly distinguishes between reflection or introspection on the one hand and reflexivity (which she calls “awareness of awareness”), the former being indirect, retrospective and optional, the latter being simultaneous with presentation of the object and constitutive of every conscious mental state. Both Coseru and Arnold (Chapters 4 and 5) dismiss the objection to the reflexivity thesis based on irreflexivity. They hold that this objection applies only if reflexivity is understood in terms of an act-object structure, but not if reflexivity is understood as constitutive of consciousness. Kumārila’s objection to this latter formulation is that there is then an irreconcilable tension between the resulting thesis that cognitions have two forms or aspects, and Buddhist mereological nihilism. But I shall not discuss this controversy here.

198

Siderits

It is in this claim of symmetry between first- and third-person attributions of consciousness that the opacity thesis meets its most vehement source of opposition. Dharmakīrti gives a clear statement of the objection: ‘In that case there would be no difference in form between [cognition of] one’s own thought and determination of the thought of another’ (pv 3.179). Of course the defender of the opacity thesis can respond that the essential asymmetry between first- and third-person attributions introduced by the transparency thesis leads to the threat of solipsism. They can also point out that the evidence supporting self-attribution is typically more reliably available to the subject than the evidence supporting other-attribution; this might begin to explain away the intuitive support for the idea that we are directly aware of our own cognitions. But the result may still look like a stalemate, and the transparency thesis does seem just simply obvious to many. What a Buddhist defender of the opacity thesis needs is some way to neutralize that intuitive support. One option here would be to appeal to the sort of neuroscientific evidence used by Carruthers (2011) to level the playing field between the two competing theses. But doing this would raise another concern. Appealing to neuroscientific evidence in defense of a thesis concerning consciousness seems tantamount to conceding that consciousness is a physical phenomenon, and Indian Buddhists consistently reject physicalism. Indeed I shall shortly argue that it is difficult to see how fitting the opacity thesis to Buddhist reductionist metaphysics can avoid reduction of the mental skandhas to the physical. Now I think a Buddhist form of the opacity thesis is worth exploring, but this will be complicated. One will need to look at the sort of empirical evidence that can be offered in its support, and also at how it might fit into a recognizably Buddhist account of persons, given the other metaphysical theses in play. I begin, though, with the reasons Buddhists might have for rejecting physicalism. There is, first, Dharmakīrti’s argument against physicalism (pv 2.24–72), which seeks to show that consciousness could not supervene on the physical, on the grounds that there are changes in mental states where there are no changes in the body of their subject. The argument is widely thought to fail, and I shall not rehearse the reasons for this here.13 Needless to say, Dharmakīrti is not to be blamed for not knowing of the correlations between mental states and neurophysical states. Still his insistence that mental states do not supervene on bodily states has the air of being driven by an ideological agenda. Part of his agenda may be straightforwardly soteriological. As a Buddhist, Dharmakīrti feels compelled to support the claim that the Buddha is knowledgeable about 13

But see Tillemans 2015 for discussion and further references. See also Coseru (Chapter 4) for apparent endorsement of the argument.

Self-Knowledge and Non-Self

199

matters relevant to cessation of suffering, and the claim that there are five skandhas (and not just rūpa or matter) is central to the Buddha’s teachings. But this scruple does not stop Yogācāra from eliminating rūpa from their ultimate ontology. What seems a more likely explanation of this and other equally bad Buddhist arguments against physicalism is the conviction that if it were true then the goal of cessation of rebirth would be automatically obtained by all at death, and the Buddha’s path would be pointless. The pan-Buddhist opposition to physicalism seems driven by the concern that it would undermine the doctrines of karma and rebirth, which are seen as central to the Buddhist path. It might be thought that physicalism is also incompatible with the standard explanation of the efficacy of mindfulness meditation. Such practice is usually said to be instrumental in bringing about cessation of suffering because it allows one to directly observe mental contents and ascertain that they are nothing like a self. This explanation clearly invokes the idea that at least under the right observation conditions the mind can be transparent to itself. But here a proponent of the opacity thesis can say that such meditation methods are efficacious precisely because most people (mistakenly) believe the transparency thesis. It is not clear what other core Buddhist commitments besides the karma-rebirth complex would be threatened by a Buddhist physicalism. There are thus difficulties for a Buddhist who accepts the opacity thesis, insofar as doing so might have physicalist consequences that are deemed unacceptable for a Buddhist. One option here is to bring metaphysical quietism to bear on these alleged physicalist consequences. If, as some claim, the point of Madhyamaka dialectic is to rule out all projects of “serious ontology” (Tillemans 2015), then a Mādhyamika might embrace the opacity thesis and yet avoid the physicalism that is thought to follow. But I shall not discuss this option here, largely because I do not see how to reconcile its brand of quietism with the core Buddhist claim that the “I”-sense is fundamentally erroneous.14 I shall instead proceed to explore what a Buddhist formulation of the opacity thesis might look like, setting to one side for now the question whether physicalism represents a threat to core Buddhist soteriological projects. There is now widespread consensus in cognitive science concerning two points about consciousness: (1) not all mental states are conscious, and (2) global availability is the mark of consciousness.15 In the Buddhist tradition point (1) seems to have first been explicitly endorsed by Yogācāra with its posit of ālayavijñāna, though other attempts by Ābhidharmikas to solve the problem of mental continuity across periods of unconsciousness at least hint at the 14 15

For more on this difficulty see Siderits 2015: 187–206. There is, of course, no such consensus among philosophers.

200

Siderits

idea that there are unconscious mental states. The consensus in cognitive science results from empirical investigation of cognitive processes, and is at the heart of the distinction between modular and global processes. At one time, so-called dual-systems theorists made bold claims to the effect that all mental processes could be assigned to one of two systems, system 1 (the dorsal system) or system 2 (the ventral system). Current views are more nuanced, in part due to increasing evidence of the plasticity of the brain: when a brain structure that serves a particular function is damaged, other parts of the brain may take up that function. Still there are functions that are clearly modular: fast, informationally encapsulated and resistant to cognitive penetration. The ducking response, for instance, is much faster than conscious threat assessment (“that object is coming at my head”), and is not dampened by knowledge of illusoriness of trigger (we also duck in 3-D movies). At the other end of the scale are types of mental processing that clearly require informational input to be made globally available. When one sees rain falling outside and wishes to stay dry, perceptual content must be made available to long-term memory and semantic processing in order to arrive at the response of reaching for the umbrella in the entry. Of course the umbrella-taking response may seem just as automatic as ducking, but this is possible only when it has become habitual. We know that when this type of problem is first encountered the subject needs to formulate a prediction of the likely result of leaving the house unprotected, then initiate a search for ways to avoid the unwanted result. Since each step in this process requires attention, and the steps must occur sequentially, the process will be slow by comparison; they involve what are called von Neumann bottlenecks (see Dennett 2017). What the subject sees when they look out the window must be seen as rain: the visual system must present its informational content in such a way as to make that content available to other systems that might play a role in solving the problem.16 By contrast, we flinch or duck before we judge that we see an object coming toward our head. Because having heritable flinchresponse circuitry enhanced the reproductive fitness of our ancestors, we don’t need to be conscious of the flying rock as anything in order for the response of ducking to be triggered. The second sort of process thus exploits what is now widely considered the defining feature of conscious mental states, that they are “globally broadcast.” 16

For discussion of the modularity of perceptual processing see Mandelbaum 2018. The points made there are relevant to assessment of the Yogācāra-Sautrāntika doctrine of the nonconceptuality of perception, which is discussed by Keng, Sharf, Spackman and Yao (Chapters 1, 2, 9 and 10).

Self-Knowledge and Non-Self

201

That is, a conscious mental state is one that makes its representational or intentional content available to other cognitive systems, such as those responsible for action, speech and memory. This global broadcasting occurs in what is described as a common workspace where a given sensory input may be shared with other specialty sub-systems. At any one time there are, of course, many different sensory inputs that can be thought of as vying for the limited space on this mental blackboard. Which input does appear is largely determined by attention, generated either in the processing of the sensory input itself (seeing as rain), or by other systems (where was the umbrella just seen?). Indian discussions of consciousness make extensive use of the metaphor of illumination: cognitions are said to illuminate their object (as well as, according to reflexivists, illuminating themselves). The notion of global availability seems like a promising way of making this metaphor of illumination more precise given what we now know about our cognitive architecture. Suppose we accept the claim that all and only conscious mental states are globally available. Given the mereological nihilism that is accepted by almost all Buddhist schools, it will follow that consciousness is not ultimately real, and is reducible. Global availability is a dispositional property, and nothing can be said to bear a dispositional property in the unaccompanied state, so consciousness is not something that bears its intrinsic nature. To say of an informational state that it is globally available is to say that there are multiple pathways between it and a variety of consumer systems. The state could not be a conscious state in the absence of those many pathways; its being conscious is something dependent on other things, something borrowed and not intrinsic. Consciousness is, like the chariot, a many masquerading as a one, something not ultimately real. There remains the question whether it is to be reduced or eliminated: is consciousness conventionally real, or is our belief that there is such a thing merely the product of a discredited theory? Given the useful role that the notion plays in folk psychology as well as in Buddhist theories of meditational practice, a Buddhist would have to opt for the reductionist alternative. To what, though, is consciousness to be reduced? When persons were reduced to causal series of aggregations consisting of the five groups of elements (skandhas), the four groups of immaterial elements (nāma skandhas) were selected based on their seeming phenomenological simplicity. It is difficult to see how a particular painful feeling, a desire for food, or an awareness event could be reductively analyzed into entities that are both somehow more simple and also recognizably non-physical. Standing behind the property-dualism of Abhidharma is the intuition that mental tropes like pain are just the sorts of things that can be known only by acquaintance, along with the idea that intrinsic natures are precisely the properties that are knowable only by

202

Siderits

a­ cquaintance. Of course, one might say that the reduction of consciousness as global availability proceeds to entities that are both non-physical and yet somehow more basic, but whose nature we are simply unable to imagine. But this move has an air of desperation about it. Given all we now know about correlations between mental events and brain events, why suppose that the reduction base must turn out to be, in some completely unspecified sense, non-physical? Accepting an account of consciousness as global availability that is couched in terms compatible with physicalism rules out the reflexivity thesis, but it does not by itself rule out other formulations of the transparency thesis. Svasamvedana is ruled out because physicalist explanations are causal in form, and no account of event causation countenances reflexive causation. (A cognition’s cognizing itself would be a case of the same entity being both cause and effect of the same causal relation.17) But there are other accounts of how transparency might be achieved that are also compatible with physicalism, so adopting the global availability account of consciousness and its attendant bias toward physicalism does not yet establish the opacity thesis. Many theorists put forward an inner sense account, according to which transparency is made possible by our having, in addition to the five external senses, an inwardly directed sense faculty that serves as the organ of introspection. Indeed several classical Indian schools seem to have had just this in mind when they speak of manas.18 And an inner sense account of the cognizing of cognition is perfectly compatible with the claim that consciousness just is global availability. So while it is true that the inferential or interpretive account of cognition of cognition relies on the global availability account of consciousness, accepting the latter is not tantamount to establishing the opacity thesis. It might be, for all that has been established so far, that the mind is furnished with a metarepresenting detector that represents and globally broadcasts all first-order representations that appear in the global workspace. Bogdan (2010) provides considerable evidence in support of the opacity thesis drawn from developmental cognitive psychology. But here I shall confine my discussion to Carruthers (2011), who defends the opacity thesis by examining a wide variety of types of empirical evidence and comparing how the two 17

18

Vaibhāṣikas raise precisely this objection to svasamvedana, although of course their opposition is not grounded in physicalism. (See Dhammajoti 2007: 328). The convergence of modern physicalism and Abhidharma dualism on this point is at least partly a result of the fact that for both (albeit for different reasons), causation is to be understood as a relation between events. For discussion of the place of manas in Buddhist theories of consciousness, see Keng, Lin and Yao (Chapters 9, 11 and 10).

Self-Knowledge and Non-Self

203

competing accounts of self-knowledge – interpretive and inner-sense – fare in explaining the phenomena. I shall discuss just one sort of case he builds, based on the phenomenon of confabulation. To explain what this is and why it bears on the dispute, we must first say more about the two competing accounts. We know that humans, as well as some non-human animals, possess what is called a mind-reading faculty. This is a system designed to use sensory input concerning the behavior of con-specifics (in the case of humans, for instance, such things as facial gesture and tone of voice) to produce attributions of mental states. It is the presence of such a system that explains things like the differential attention paid by infants to highly stylized drawings of faces. The mindreading faculty is clearly a modular system. Its output can be globally broadcast and made available to other systems, but that output results from a process that is swift and automatic. When the infant sees a stylized drawing of a face with upturned mouth, the infant does not reason that this indicates a smile, and that reciprocating is likely to yield increased opportunities to receive care. The infant sees a smiling face, and smiles back. Of course, over time and with experience this repertoire is enriched, becoming what we can call a subject’s theory of mind: a set of concepts and beliefs that are used to interpret the behavior of others. But the judgments that are its output are produced equally rapidly, without awareness of intermediary interpretive processes. We see someone’s hand movement as a reaching for the doorknob. (In Dharmakīrti’s system such cognitions would count as cases of perceptual judgment: feeling like unmediated perception but actually inferential in structure.) Of course sometimes the mind-reading system yields no such judgment, and we must turn to overt theorizing in order to interpret the behavior of the target. But quite commonly the mind-reading system works smoothly and effortlessly, so that we are inclined to report our judgments concerning the mental states of others as a matter of “just seeing.” An inner-sense account of self-knowledge must agree that it is the mindreading system that is typically at work in our attributions of mental states to others. Proponents of such an account claim, though, that there is a distinct system at work in the case of self-attribution – the inner sense. Opacity theorists deny this. They claim instead that cognition of one’s own cognitions results from turning the mind-reading system on one’s own case. That is, sensory input concerning the behavior of the subject system is interpreted as warranting attribution of cognitive states to the subject making the attribution. Here is where Kumārila’s abductive inference is performed. Under certain circumstances the best explanation of someone’s saying “blue” is that they see blue. When one oneself sees blue, global broadcast of the perceptual cognition can give rise to an inner-speech occurrence of tokening “blue.” The best explanation

204

Siderits

of this occurrence is that one has oneself seen blue. So when one sees blue one is able to immediately and effortlessly report that one sees blue. But according to the opacity theorist the self-report is not different in kind from the attribution of cognitive states to others. In both cases the attribution can proceed rapidly and smoothly, so that it feels like a direct report. Crucially, in both cases we can also mis-attribute cognitive states that are not there. Confabulation is the phenomenon of confidently attributing to oneself mental states that one does not in fact have. The problem of confabulation is what doomed 19th century attempts to develop a scientific psychology based on introspection: there seemed to be no way to tell whether subject reports were veridical or were rather prompted by features of the experimental setup, so that the all-important element of public verification was beyond reach. There are, though, special conditions in which confabulation can be known to have occurred. In such cases it looks for all the world as though the mis-­ attribution results from an attempt by the subject to make sense of their own behavior. And these pose a challenge for the transparency thesis. It is open to the inner-sense theorist to claim that under the sorts of abnormal circumstances leading to confabulation, the inner sense fails to operate and the mindreading faculty takes over. The mind-reading faculty’s being modular means that it must rely on algorithms. But the transparency thesis claims that one is directly and authoritatively aware of one’s own mental states. So if this is how access is achieved through the inner sense, then the inner-sense theorist owes us an account of how to draw a line between the two sorts of case. Carruthers claims that this burden is not met. I shall not review all the evidence he utilizes, limiting discussion to one sort of case Carruthers does discuss, that of commissurotomy patients, plus another he does not discuss, that of blindsight. Commissurotomy is the severing of the corpus callosum, the brain structure that serves as the principal neural pathway between the two hemispheres of the brain. One method of inducing confabulation in commissurotomy patients takes advantage of the fact that contents of the right side of the visual field are processed by the left hemisphere (which in most persons is responsible for speech production and comprehension) while contents of the left side are processed by the right hemisphere (which typically has only very limited speech comprehension). Under the right conditions it is possible to flash a word on a screen in such a way as to make it available only to the right hemisphere of a commissurotomy patient. When the word in question is “Walk! ,” one patient got up from their chair and left the room. Asked why, the patient responded that they were thirsty and were going to get something to drink. In another setting, a picture of a saw was shown on the left side of the patient’s visual field while a picture of a hammer

Self-Knowledge and Non-Self

205

was simultaneously flashed on the right side. Asked to use the left hand (controlled by the right hemisphere) to draw what they had seen, the patient drew a saw. Asked what they just drew, they answered “A saw.” But then when asked what it is that they saw, they responded “A hammer.” Pressed to explain the discrepancy, they confessed that they did not know. In the first case the confabulation is in the patient’s explanation of their behavior. Because that part of the brain responsible for speech production has no access to the visual presentation of the “Walk!” prompt, it produces an interpretation of the subject’s action that seems plausible under the circumstances. That the subject produces this interpretation rapidly, effortlessly, and in full confidence of its veracity despite their knowledge of the procedure performed on their brain, strongly suggests that it is the output of the mind-­reading system. In the second case there is strictly speaking no confabulation, just conflicting belief-attributions concerning what the subject saw. Presumably the right answer in this case is that the subject saw both a saw and a hammer.19 If the inner sense is a reliable source of information about occurrences of one’s mental states, the transparency theorist owes us an explanation as to why it has failed in this case. Of course they may claim that commissurotomy renders the patient’s inner sense unable to process input from both ­hemispheres simultaneously. But the affected pathways of the commisurotomy patient are quite different from the regions of the brain that are affected in cases of blindsight, yet similar instances of confabulation are to be found in the literature on blindsight patients. Blindsight is a condition in which a patient with intact visual organs reports utter absence of visual consciousness (they are not aware of seeing anything), yet for instance performs significantly better than chance when asked to identify letters flashed on a screen before them. In one particularly striking case a patient walks down a hallway strewn with obstacles, safely navigating around them while insisting that they see nothing in their way. Asked how they walked, they immediately and confidently answer that they went straight down the hall. This report is of course what the mind-reading system would predict about the experience of a blind person who walks down a hallway. An inner-sense theorist cannot explain this apparent failure of the inner sense in the same way that they might want to explain its failure in the commisurotomy cases. For blindsight does not involve damage to the corpus callosum, 19

This might be denied on the grounds that the right hemisphere is not conscious. But then we must ask how the patient was able to draw a saw. It would be odd to suppose that the ability to draw artifacts is encapsulated within visual processing mechanisms in the right hemisphere.

206

Siderits

it involves damage to one of the brain’s visual systems. There are two such systems, an encapsulated system that uses visual input to directly control motor performance without making its content globally available; and a second system that performs basic categorization and globally broadcasts its product. It is due to lesions in the area of the brain responsible for the latter function that the patient reports lack of visual consciousness. So the confabulation in this case cannot be blamed on the same neural disruption that led to confabulation in the commisurotomy cases. The opacity thesis has a single explanation for all these cases: the mind-reading system is working with diminished input and so works to fill in the gaps with a plausible interpretation. Since defenders of the transparency thesis must give different explanations for the different cases of confabulation, their view comes to seem to that extent less defensible. But confabulation raises a further difficulty for the transparency thesis that goes beyond merely comparing the scores of competing theories. A major source of the intuitive plausibility of the transparency thesis is that self-­attributions of cognitive states often feel like direct reports. The early Sautrāntika Śrīlāta puts this in terms of the notion of vividness (spaṣṭatva), which in Abhidharma is generally taken as the hallmark of perception. Cognition of fire by means of vision or touch is said to be more vivid than cognition of fire by means of an inference from seeing smoke. This is what will lead Dignāga and Dharmakīrti to call the fire cognized in the first way a svalakṣaṇa (a “self-­ characterized” object) and the fire cognized in the second a sāmānyalakṣaṇa (a “generally characterized” object). When Śrīlāta claims that the reflective report “I see blue” is accompanied by vividness, he is asserting that our a­ wareness of our own c­ ognitive states has the status of perception or pratyakṣa, that it is of something “before the eyes” (Dhammajoti 2007: 329). What cases of confabulation demonstrate is that the mind-reading faculty is capable of generating this sense of direct awareness (perhaps because it utilizes fast-acting modular processes). But the cognition of cognitive states produced by the mind-reading system is not direct, it is mediated by inferential links. So Śrīlāta is wrong to take the feeling of vividness to support the thesis that mental states are transparent to their subjects. There are other considerations that might be brought to bear on the issue. For instance, there are reasons to think that adding an inner sense to the cognitive architecture of early hominids would not have conferred a reproductive advantage in the environment of evolutionary adaptation, especially given that they are likely to have already been equipped with a mind-reading faculty (something humans share with other primates). But I shall not go into any of that here. Suffice it to say that the results of empirical investigation give at least

Self-Knowledge and Non-Self

207

some reason to resist the intuitive appeal of the transparency thesis. Insofar as this widely held thesis is a major source of support for belief in the reality of the “I”-sense, something all Buddhists agree must be extirpated in order to overcome suffering, they ought to welcome this evidence against the transparency thesis. The question is whether fear of embracing physicalism will stand in their way. The intuitive appeal of seeing meditation as a matter of directly observing one’s own mental processes is clear. But Buddhist philosophers are in the business of scrutinizing our intuitions; they are revisionists at heart. (Think of the intuitive appeal of external-world realism, and Yogācāra’s revisionist response.) The real obstacle here seems to be the threat that physicalism poses to the karma-rebirth ideology. I shall not try to answer the question whether this ideology is a core component of Buddhism, that a system without it could not qualify as Buddhist. But I do think it is worthwhile to raise the question. Abbrevations cśv  = Āryadeva’s Catuḥśatakam with Candrakīrti’s Vṛtti, ed. Bhagchandra Jain Bhaskar. Nagpur: Alok Prakashan, 1971. MAV  = Madhyamakāvatāra of Candrakīrti. Li Xuezhu (2015), “Madhyamakāvatāra-kārikā Chapter 6,” Journal of Indian Philosophy 43: 1–30 psv  = Dignāga’s Pramāṇasamuccaya, Chapter 1: A hypothetical reconstruction by Ernst Steinkellner, published online at: http://www.ikga.oeaw .ac.at/Mat/dignaga_PS_1.pdf. pv    = The Pramāṇavārttikam of Dharmakīrti. Edited by Raniero Gnoli. Rome: Istituto Italiano per il Medio od Estremo Oriente, 1960. PVin = Dharmakīrti’s Pramāṇaviniścaya. Chapters 1 and 2. E. Steinkellner (ed), Beijing-Vienna: China Tibetology Publishing House, 2007. Bibliography Albahari, Miri (2006), Analytical Buddhism: The Two-Tiered Illusion of Self. Basingstoke and New York: Palgrave Macmillan. Bogdan, Radu. 2010. Our Own Minds: Sociocultural Grounds for Self-Consciousness. Cambridge, MA: mit. Carruthers, Peter (2011), The Opacity of Mind. Oxford: Oxford University Press. Dennett, D.C. 2017. From Bacteria to Bach and Back: the evolution of minds. New York : W.W. Norton & Company.

208

Siderits

Dhammajoti, Bhikkhu KL (2007), Sarvāstivāda Abhidharma. Hong Kong: Center of Buddhist Studies, University of Hong Kong. Kellner, B. (2010), “Self-Awareness (svasaṃvedana) in Dignāga’s Pramāṇasamuccaya and -vṛtti: A Close Reading.” Journal of Indian Philosophy 38: 203–231. Kellner, B. (2011), “Self-awareness (svasaṃvedana) and infinite regresses: a comparison of arguments by Dignāga and Dharmakīrti,” Journal of Indian Philosophy 39: 411–426. Kripke, Saul (1982). Wittgenstein on Rules and Private Language. Cambridge, MA: Harvard University Press. Mandelbaum, Eric. 2018. “Seeing and Conceptualizing: Modularity and the Shallow Contents of Vision,” Philosophy and Phenomenological Research 97 (2,1): 267–283. Montague, Michelle (2017), “What Kind of Awareness is Awareness of Awareness?,” Grazer Philosophische Studien 94: 359–380. Siderits, Mark (2015). Personal Identity and Buddhist Philosophy: Empty persons, 2nd ed. Farnham, Surrey: Ashgate. Tillemans, T. (2015), “On minds, Dharmakīrti and Madhyamaka,” in Koji Tanaka, Yasuo Deguchi, Jay Garfield and Graham Priest (eds), The Moon Points Back. NY: Oxford University Press. Pp. 45–66. Yao, Zhihua (2005). The Buddhist Theory of Self-Cognition. London: Routledge.

Chapter 7

The Genesis of *Svasaṃvitti-saṃvitti Reconsidered Toru Funayama Abstract The fourfold division of cognition is well known in the East Asian tradition of Yogācāra Buddhism as originated by Dharmapāla (mid 6th c.). However, no Indian and Tibetan sources confirm the Indian origin of the key term “cognition of self-cognition” (zheng zizheng fen, *svasaṃvitti-saṃvitti-bhāga) except for the two notable Chinese “translations” by Xuanzang, the Cheng weishi lun “Establishment of Consciousness-only” and the Fodi jing lun “Interpretation of the Buddha Land.” By re-examining the reason(s) for the postulation of this term in comparison with Dignāga’s threefold division of cognition (grāhyākāra, grāhakākāra, svasaṃvedana), I will endeavor to argue for a fundamental difference between Dignāga’s use of svasaṃvedana as the result of cognition (pramāṇa-phala) and that of Dharmapāla; consequently the so-called Dharmapāla theory of the fourfold division is rather difficult to explain in the Indian context. Further, I would like to suggest that Chengguan’s 澄觀 (738–839) illustration of the fourfold division of cognition by a clear mirror ((1) the image on the mirror 鏡像, (2) the light on the mirror 鏡光, (3) the mirror’s surface 鏡面, and (4) the backside of the mirror 鏡背) is possibly based on the actual form of Chinese bronze mirrors.

Keywords Dharmapāla– svasaṃvitti-saṃvitti – Yogacara Buddhism – self-cognition – self-awareness – Cheng weishi lun – Dignāga – the fourfold division of the mind – the threefold division of the mind – Pramāṇasamuccaya

…

During the second and fourth centuries ce, Indian Mahāyāna (“Great Vehicle”) Buddhism formed two major schools: the Middle Way school (Madhyamaka) was founded by Nāgārjuna (ca. 150–250) first, and then the Yogācāra school was founded by Maitreya, Asaṅga and Vasubandhu. Yogācāra Buddhism was disseminated in China no later than the beginning of the fifth century, and

© koninklijke brill nv, leiden, ���1 | doi:10.1163/9789004440913_011

210

FUNAYAMA

formed a notable school called Faxiang school (法相宗 or 法相學; School of the Intrinsic Nature of Real Entities). It was the influence of Xuanzang 玄奘 (d. 664) that contributed to the spread of Yogācāra in China. Soon thereafter this school prevailed throughout East Asia, including Japan. One of the many noteworthy theories of this school is its theory of the fourfold division (四分) of the mind (Skt. vijñāna /識, a synonym of Skt. citta /心) as a fundamental part of the school’s epistemology. This theory has considerable significance, but still even today scholars are unclear about some key points; one such unresolved question is the origin of the fourfold division. In this paper I will consider the question of how this theory came to be formed. The theory of the fourfold division of cognition is popular in East Asian Yogācāra. It is acclaimed as the original idea of Dharmapāla (Hufa 護法, fl. ca. mid-6th-century).1 It holds that a cognition has these four aspects: [1] “the object-aspect of cognition” (grāhya), [2] “the subject-aspect of cognition” or “cognitive agent” (grāhaka), [3] “self-cognition” or “self-awareness” (svasaṃvitti, svasaṃvedana or ātmasaṃvedana), [4] “cognition of self-cognition” or “awareness of self-awareness” (*svasaṃvittisaṃvitti).2 After briefly introducing, in Section I, earlier references to self-awareness and its illustration(s) in Indian Buddhism, I will check well-known passages of the two Chinese works which are arguably Xuanzang’s (d. 664) “translations” of Sanskrit texts. One of them is Dharmapāla’s Cheng weishi lun 成唯識論 (The Establishment of Consciousness-only; hereafter cwsl). It was translated into Chinese in 659.3 The other is the Fo di jing lun 佛地經論 (The Treatise on the Buddha Land; hereafter abbrev. as fdjl) of Qinguang 親光 et al.4 It was

1 For problems concerning Dharmapāla’s dating as 530–561 ce, see Funayama (2000: 7–11). 2 The Sanskrit term svasaṃvitti-saṃvitti is merely a reconstruction from the Chinese translation. It is based on de La Vallée Poussin’s assumption (1928: 131 ”svasaṃvitti-saṃvitti-bhāga” for zheng zi zheng fen 證 自 證 分 ). Slightly different reconstructions such as *svasaṃvedanasaṃvedana would also be possible. However, I will use his expression in order to avoid unnecessary confusion. 3 The translation year is based on the description of the Kaiyuan shijiao lu 開 元 釋 教 録 (T55, 556c). The cwsl is not a translation in the normal sense but a notably special type of arrangement called “mixed/mingled translation” (rouyi 糅 譯 , yirou 譯 糅 or herou 合 糅 ), meaning a Chinese-made arrangement of ten different commentaries on Vasubandhu’s root text Thirty Stanzas (Triṃśikā), centering on Dharmapāla’s commentary thereon. It is worth mentioning that no detail of the arrangement process of the cwsl is recorded in extant Chinese sources. 4 The Sanskrit form of Qinguang is not certain, though *Bandhuprabha is often assumed.

The Genesis of *Svasaṃvitti-saṃvitti Reconsidered

211

translated in 649.5 Only these two Chinese “translations” support the Indian origin of Dharmapāla’s theory of the four aspects of cognition. Not a single Sanskrit or Tibetan text mentions this theory. Thus, paying special attention to the argument of the two texts’ postulation of *svasaṃvitti-saṃvitti in Section ii, i will question whether this theory truly had an Indian origin. What is called Dharmapāla’s fourfold division of cognition is based on Dignāga’s (ca. 480–540 ce) threefold division – grāhya, grāhaka and svasaṃvedana – which is fully attested in Sanskrit and Tibetan texts. The threefold division is partially based on the distinction between prameya (the object of valid cognition), pramāṇa (a means of valid cognition), pramātṛ (the subject/cognizer of valid cognition), and pramiti (=pramāṇaphala, the result/fruit of valid cognition) of orthodox Brahmanism as found in the Nyāyasūtra 2.1.16 and Vātsyāyana’s Bhāṣya thereon. However, on the other hand, no Indian and Tibetan sources confirm the Indian origin of Dharmapāla’s key term “cognition of self-cognition” except for the abovementioned two notable Chinese “translations” by Xuanzang. In Section iii of this paper, I will attempt to cast doubt on the Indian origin of the fourfold division theory. Finally in Section iv I will speculate about whether there are any new perspectives or theories in the works of Chinese Yogācāra masters after Xuanzang. 1

Introductory Remarks on the Origin of Self-Cognition

In two monographs, Yamaguchi Susumu introduces an interesting passage from Candrakīrti’s Madhyamakāvatāra on the relationship between Dignāga’s idea of self-cognition and its school-affiliation. Paying special attention to Madhyamakāvatāra vi vv. 72–75, Yamaguchi points out the possibility that Dignāga’s svasaṃvedana-theory originally comes from a Sautrāntika doctrine that was later adopted by the Yogācāra school.6 However, it would be misleading to take his view in the historical sense; historically, either Sautrāntika or Yogācāra might have been the origin. It is indeed plausible that Dignāga was the first author to refine the theory of self-awareness at the philosophical level, 5 The year of translation year is based on the description of the Kaiyuan shijiao lu (T55, 556b). The fdjl was translated ten years earlier than the cwsl. However, on the other hand, the theory of *svasaṃvitti-saṃvitti in the fdjl is evidently based on the same theory in the cwsl. It is weird that no detailed record of translation as well as the author Qinguang et al. is found in extant Chinese sources. The name Qinquang is often retranslated into Skt. as *Bandhuprabha in previous studies but the Sanskrit name is no more than a conjecture. 6 Yamaguchi (1941: 281–295); ibid. (1951: 21).

212

FUNAYAMA

but it does not necessarily mean that nobody had mentioned the significance of self-cognition before Dignāga. An example of candle light which illuminates both the candle itself and other things is taken up by Nāgārjuna (ca. 150–250), too.7 By and large, the epistemic possibility of self-cognition or self-awareness is illustrated by a light (dīpa or pradīpa) which illuminates (pra√kāś) both itself and other things. On the other hand, the impossibility of self-cognition is mentioned by referring to several examples such as “a fingertip cannot touch itself,” “the blade of a knife cannot cut itself,” “the eye cannot see itself,” and “even a strong man cannot stand on his own shoulders.” For example, Xuanzang’s translation of the (Abhidharma-)Mahāvibhāṣā says, “A finger cannot touch itself; the blade of a sword cannot cut itself; the eye cannot see itself; and a strong man cannot carry himself on his shoulder.” (T27, 43a). These illustrations explain that cognizing oneself is self-contradictory. The fdjl refers to the first example in the context of the validity/invalidity of self-cognition (T26, 303a). What is more important is a reference to the same illustration in the (Abhi­ dharma-) Mahāvibhāṣā. It says: The Mahāsāṅghikas state as follows: A cognition cognizes that cognition itself, as in the case of a light, [the essence of which] is illuminating nature, which illuminates itself and illuminates others. The same is true of cognition. Having the nature of knowing, [the cognition] cognizes itself and cognizes others.8 The Mahāsāṅghika school is mentioned here as the proponent of the theory. It is quite possible that the theory introduced in the Mahāvibhāṣā predates Dignāga’s time, the sixth century. 7 E.g., Vigrahavyāvartanī (Johnston/Kunst 64,16–21), Mūlamadhyamakakārikā vii 8, Vaida­ lyaprakaraṇa (Kajiyama 1965: 135). Though these passages appear as an opponent’s view (pūrvapakṣa) in Nāgārjuna’s works, it is evident that he was aware of the significance of selfcognition as illustrated by a light. See also the Nyāyasūtra 2.1.16–19 and Nyāyabhāṣya thereon as a relevant topic. 8 浮 陀 跋 摩 、 道 泰 等 譯 《 阿 毘 曇 毘 婆 沙 論 》 如 摩 訶 僧 祇 部 作 如 是 説 。 自 體 能知自體，如燈是照性能自照，亦能照他。彼智亦爾。是智性能自知， 亦 能 知 他 。 (T28, 31c). Cf. 玄 奘 譯 《 阿 毘 達 磨 大 毘 婆 沙 論 》 謂 或 有 執 。 心 心 所 法 能 了 自 性 ， 如 大 衆 部 彼 作 是 説 :智 等 能 了 ， 爲 自 性 故 ， 能 了 自 他 ， 如 燈 能 照 ， 爲 自 性 故 ， 能 照 自 他 。 (T27, 42c). The former translation was made in 427 ce (see Sengyou’s 僧 祐 [445–518] Chu sanzang ji ji 出 三 藏 記 集 T55, 74a). It is older than the latter. The contents are nearly the same. See also Katsura (1976) as an earlier reference to the philosophical significance of this passage.

The Genesis of *Svasaṃvitti-saṃvitti Reconsidered

2

213

“Cognition of Self-Cognition” in the cwsl and the fdjl

Between the two “translations,” the passage in the cwsl is more popular than that in the fdjl because the former is ascribed to Dharmapāla, who was traditionally believed to be the originator of the fourfold division of cognition. However, the cwsl has only one section in which the “cognition of self-­ cognition” is mentioned.9 The passage in question begins as follows: When mind (*citta) and mental associates (*caitta/caitasika) are produced, a rational analysis will show that each [cognition] has these three parts, because it is divided into [1] the object of cognition (prameya; ≈ grāhya), [2] the subject of cognition (pramāṇa lit. “a means of valid cognition”; ≈ grāhaka), and [3] the result of cognition (pramāṇaphala; ≈ svasaṃvitti). The object and the subject must have a basis (/epistemic substratum, *āśraya); therefore, as a stanza of the Corpus of Valid Cognition (Pramāṇasamuccaya, Jiliang lun 集量論) states, “What looks like the form of an object is the object of cognition; the form of the subject and self-cognition are [respectively] the subject of cognition and the result of cognition. These three [aspects of cognition] are inseparable from one another.”10 If we make an even finer distinction concerning mind and its activities, we can say that there are four parts. In addition to the three parts as explained above, the fourth [part] exists as the part of cognition of selfcognition (*svasaṃvitti-saṃvitti-bhāga) 證自證分. If this fourth part did not exist, who would directly cognize the third part? Because, inasmuch as all parts belong to the mind, all must be directly cognized [i.e., even the third part must be directly cognized]. Moreover, [without the fourth part,] an undesirable conclusion would follow, namely that the self-cognition part would have no result; [this cannot be true] because all subject cognitions necessarily have their result.11 – cwsl (T31, 10b) 9

The term zheng zi zheng fen 證 自 證 分 is used twice (T31,10b18 and 10b25) in the section. Here I introduce only the first occurrence. 10 This stanza corresponds to Pramāṇasamuccaya I 10: yadābhāsaṃ prameyaṃ tat pramāṇaphalate punaḥ / grāhakākārasaṃvittyos trayaṃ nātaḥ pṛthak kṛtam // (Steinkellner 2005: 4, 17–18; Hattori 1968: 29 and 107 n.1.67). 11 玄 奘 譯 《 成 唯 識 論 》 : 然 心 心 所 一 一 生 時 ， 以 理 推 徴 ， 各 有 三 分 。 所 量、能量、量果別故。相見必有所依體故，如《集量論》伽他中説，“ 似 境 相 所 量 ， 能 取 相 自 證 ， 即 能 量 及 果 ， 此 三 體 無 別 ”。 又 心 心 所 若 細 分 別 ， 應 有 四 分 。 三 分 如 前 ， 復 有 第 四 證 自 證 分 。 此若無者，誰證第 三。 心 分 既 同 ， 應 皆 證 故 。 又 自 證 分 應 無 有 果 。 諸能量者，必有果 故。 (T31, 10b) Cf. Cook (1999: 62).

214

FUNAYAMA

Here especially Passage confirms that the third aspect of cognition can work also as the object of the fourth part of cognition. It is remarkable that such an idea cannot be found in Dignāga’s Pramāṇasamuccaya or Nyāyamukha. Dharmakīrti, the Indian successor of Dignāga’s epistemology, does not admit this view, either. In fact, the above idea is verified only in two Chinese texts: cwsl and fdjl. The passage of the latter reads as follows: The Corpus of Valid Cognition (Pramāṇasamuccaya, Jiliang lun 集量論) clarifies that any mind and mental associates invariably have three parts: one, the part of grasped object (grāhya); two, the part of grasping subject (grāhaka); and three, self-cognition (svasaṃvedana). These three parts are neither identical nor different. The first [part] is the object of cognition (prameya). The second [part] is the subject of cognition (pramāṇa). And the third [part] is the result of cognition (pramāṇaphala). If one analyzes [the mind] in more detail, there must be four parts to complete the theory. In addition to the first three as explained above, the fourth is [necessary] as cognition of self-cognition 證自證分. (…) The third [aspect], i.e., self-cognition, directly cognizes both the second (i.e., the subject of cognition) and the fourth (i.e., cognition of selfcognition). The fourth [aspect,], as self-cognition, directly cognizes the third [which in that case works as the object of cognition]. [Thus] both the third and fourth aspects are included in direct perception.12 – fdjl (T26, 303b) Passage of the fdjl asserts that the third aspect works as the cognitive agent to grasp both the second and fourth aspects. Passage agrees that the fourth aspect can work as the agent of cognizing the third. This means that according to the doctrine of the fdjl, both the third and the fourth parts of cognition can exchange roles as each other’s subject and object. This line of explanation has much in common with Passages and of the cwsl. Both texts claim that the third aspect is not merely the result of cognition as explained by Dignāga,13 but also functions as the cognitive agent, just as in the case of the second aspect. This kind of interchangeability of cognitive functions – in other words plural functions of a single aspect of cognition – deserves attention when 12

13

玄奘譯《佛地經論》: 《集量論》中辯心心法皆有三分。一所取分，二 能取分，三自證分。如是三分，不一不異。第一所量。第二能量。第 三量果。若細分別，要有四分，其義方成。三分如前。更有第四證自 證 分 。 … 第三自證能證第二及證第四。 第四自證能證第三。 第 三 、 第 四 皆 現 量 攝 。 (T26, 303b). Cf. Keenan (2002: 87–88). See Section 3 for Dignāga’s threefold division of cognition.

The Genesis of *Svasaṃvitti-saṃvitti Reconsidered

215

compared with Dignāga’s theory of threefold division of cognition in the Pramāṇasamuccaya. According to Fukihara Shōshin’s summary of the East Asian interpretation of Dharmapāla’s fourfold theory, the following four interpretations are possible:14

Interpretation One Interpretation Two Interpretation Three Interpretation Four

1. grāhya 相分

2. grāhaka 3. svasaṃvitti 見分 自證分

4. svasaṃvitti-saṃvitti 證自證分

prameya

pramāṇa

pramāṇaphala

n/a

n/a

prameya

pramāṇa

pramāṇaphala

n/a

n/a

pramāṇa

n/a

n/a

prameya pramāṇaphala pramāṇa

prameya pramāṇaphala

This diagram can be paraphrased as follows: [Phase One] – The object-aspect of cognition (grāhya) is the object of valid cognition (prameya). The subject-aspect of cognition (grāhaka) is a means of valid cognition (pramāṇa). And self-cognition (svasaṃvitti) is the result of valid cognition (pramāṇaphala). [Phase Two] – The subject-aspect of cognition (grāhaka) is the object of valid cognition (prameya). Self-cognition (svasaṃvitti) is a means of valid cognition (pramāṇa). And the cognition of self-cognition (svasaṃvitti-saṃvitti) is the result of valid cognition (pramāṇaphala). [Phase Three] – Self-cognition (svasaṃvitti) is the object of valid cognition (prameya). The cognition of self-cognition (svasaṃvitti-saṃvitti) is a means of valid cognition (pramāṇa). And self-cognition (svasaṃvitti) is the result of valid cognition (pramāṇaphala). [Phase Four] – The cognition of self-cognition (svasaṃvitti-saṃvitti) is the object of valid cognition (prameya). Self-cognition (svasaṃvitti) is a means of valid cognition (pramāṇa). And the cognition of self-cognition (svasaṃvitti-saṃvitti) is the result of valid cognition (pramāṇaphala).

14

Fukihara (1935: 270–273). The four phases are explained by Chinese commentators such as Kuiji 窺 基 (see Section 4 below) and the Japanese scholar-monk Zenju 善 珠 (723– 797) in his Yuishiki bunryō ketsu 唯 識 分 量 決 (T71, 443b).

216 3

FUNAYAMA

Can the Third Aspect Work as Cognitive Agent?

Next I want to consider a fundamental difference between Dignāga and Dharmapāla. What is called Dharmapāla’s view on the fourfold division of the mind may be acceptable, insofar as we read it in Chinese texts without careful comparison with Sanskrit texts such as the Pramāṇasamuccaya. An important point is that Dignāga always regards the third aspect, svasaṃvedana, as the result of cognition (pramāṇaphala). In other words, he never maintains that the third aspect can work as the subject of cognition (grāhaka or “a means of valid cognition [pramāṇa]”). In the first chapter, “Direct Perception,” of the Pramāṇasamuccaya, Dignāga explicitly states more than once that the third aspect, svasaṃvedana, is the result of cognition, to the effect that svasaṃvedana defines the inseparable relationship between the object-aspect and the subject-aspect of a single cognition. In stanza 10, he certainly means that svasaṃvedana is the result (pramāṇaphala); but this is not his only statement on the matter. He also states as follows: Or [one should maintain that] the self-cognition is here the result [of the act of cognizing] (v. 9a) Any cognition has a twofold appearance; viz., the appearance of [the cognition] itself and the appearance as the object. That both appearances have [the nature of] self-cognition is [the meaning of] “result.” … (Hattori 1968: 28) svasaṃvittiḥ phalaṃ vātra (i 9a) dvyābhāsaṃ hi jñānam utpadyate svābhāsaṃ viṣayābhāsaṃ ca. tasyobhayābhāsasya yat svasaṃvedanaṃ tat phalam…. (Steinkellner 2005: 43–45) The same idea is corroborated by the commentator Jinendrabuddhi (ca. 710– 770) and Dharmakīrti (7th century). For example, Jinendrabuddhi states, “On the other hand, the cognition itself occurs as self-cognized; therefore the very self-cognition is the result [of cognition].”15 Dharmakīrti explains, “In that case self-cognition is maintained to be the result [of cognition] because [the cognition] does not cognize anything else.”16 That is, Dignāga’s Indian successors Dharmakīrti and Jinendrabuddhi use the term svasaṃvedana only in the sense of pramāṇaphala, which is strictly distinguished from grāhaka and grāhya. 15 16

psṭ 68, 12: vijñānam eva tu svasaṃviditam utpadyata iti svasaṃvittir eva phalam. Pramāṇavārttika iii 332 cd: tadānyasaṃvido ’bhāvāt svasaṃvit phalam iṣyate //.

The Genesis of *Svasaṃvitti-saṃvitti Reconsidered

217

Dharmakīrti and Jinendrabuddhi take svasaṃvedana as the term signifying the ontological identity between grāhyākāra and grāhakākāra. When we compare such usage of the term with the above-mentioned two Chinese “translations,” the reference to svasaṃvedana as implied in Passages through looks quite odd as a theory of Indian Buddhism. The cwsl and the fdjl accept that the third aspect, svasaṃvitti, can work as the agent or subject to cognize the second aspect (grāhaka), but Dignāga never accepts this view; according to Dignāga svasaṃvitti always signifies the result side of a valid cognition (pramāṇaphala), in other words, the identity of the object and the subject as a single cognition. In the above-mentioned cwsl, Dharmapāla raises a question, “Who would directly cognize the third part? (誰證第三?)” in order to explain the necessity of postulating *svasaṃvitti-saṃvitti.17 As long as we read the passage in Chinese, this question sounds natural and convincing enough. However, when it is compared with Dignāga’s epistemology, it turns out that the question is impossible, since the third aspect is claimed to be the result in order to explain the epistemic identity of subject and object in Dignāga’s thought; the third aspect cannot be the object of anything else. This consideration suggests by extension the high probability of the apocryphal character of the cwsl and the fdjl. We may be able say that what has been called Dharmapāla’s theory has a totally different background from Dignāga’s system of epistemology. It seems to suggest that the two Chinese works, cwsl and fdjl, are not translations but Chinese arrangements of

1. grāhya object

2. grāhaka subject

3. svasaṃvedana self-cognition

Diagram 1

17

See n. 11 above.

218

FUNAYAMA

I­ndian doctrine, at least partially by Xuanzang himself or his disciples. As a working hypothesis I would like to suggest that the fourfold division of cognition is a Chinese apocryphal doctrine.18 To sum up my argument for the topic in question in Indian context, here are two diagrams of Dignāga’s theory of the threefold division of the mind. First, some modern studies use Diagram 1 on the preceding page, which I believe is quite misleading. I do not think that the above diagram correctly describes Dignāga’s epistemology. Dignāga certainly mentions a threefold division, but he never states that the third element, svasaṃvedana, is a third “portion” of the mind; he says that the third explains the inseparable relation of grāhya and grāhaka. In other words, he considers the third element not as a distinct, third part of the mind but as expressing “relation” or “status” in the sense that that there is no ontological difference between grasped (grāhya) and grasper (grāhaka). And second, consider the following diagram of the three-fold division, which I believe describes Dignāga’s intention far better than the first diagram. When we accept the second diagram as depicting Dignāga’s theory of svasaṃvedana, the above-stated question “Who would directly cognize the third part? (誰證第三?)” in Dharmapāla’s fourfold division of the mind

1. grāhya object

2. grāhaka subject

3. svasaṃvedana the inseparable relation between 1 and 2. 1 and 2 are the same mind.

Diagram 2

18

For a recent overview of the characteristics and kinds of Chinese Buddhist Apocrypha, see Funayama (2015).

The Genesis of *Svasaṃvitti-saṃvitti Reconsidered

219

c­ annot arise because the svasaṃvedana is not a distinct, third “portion,” at least according to Dignāga, the founder of the svasaṃvedana-theory. This question in the cwsl is most probably based on the thinking represented by diagram 1. 4

Chinese Acceptance of Dharmapāla’s Theory

In the previous section, I have raised doubts about the authenticity of the two Chinese translations. If my suspicion has merit, it means that the fourfold division of cognition itself is a Chinese apocryphal theory, but it would require careful consideration to reach a final conclusion. In what follows, I want to suggest some other phases of Chinese transformation of the theory. It might be worthwhile to check whether Chinese scholar-monks tried to add any new points in their own words to the theory of the cwsl and the fdjl. In his commentary on the twenty stanzas (Viṃśatikā/Viṃśikā) of Vasuban­ dhu, Kuiji 窺基 (632–682) summarizes his view of svasaṃvitti-saṃvitti in the following way: “Measurement” (liang 量) means “mensuration” (liangdu 量度);19 for example, when one measures silk, brocade and so on by a scale, the scale is the measuring instrument; silk and so on are the object of measuring. Understanding the quantity and number is the result of measuring. When mind and mental associates take things as cognitive support, we state that the fourfold division [of cognition is effective]. [Namely, first,] the seeing aspect is the subject of cognition; the appearance aspect (=object aspect) is the object of cognition; and self-­ cognition is the result of cognition. [Second,] when such self-cognition takes the seeing aspect as cognitive support, the seeing aspect is the object of cognition; self-cognition is the subject of cognition; and cognition of self-cognition is the result of cognition. [Third,] when cognition of self-cognition has self-cognition as cognitive support, self-cognition is the object of cognition; cognition of self-cognition works as the subject of cognition; and the self-cognition is the result of cognition….

19

For the term “mensuration” (liangdu) in the epistemological context, see Funayama (2014: 49 and 55).

220

FUNAYAMA

[And last,] when one takes, by the third [aspect], the fourth [aspect] as cognitive support, the fourth is the object of cognition; the third is the agent of cognition; and the fourth itself is the result of cognition. (T43, 998c–999a) This passage reveals that Kuiji’s idea is basically the same as Zenju’s.20 As summarized in the table of Section 2, what is called “the fourfold division” is actually not the division of the mind into four parts understood as working simultaneously, but either the threefold division [Phases One and Two] or the twofold division [Phases Three and Four]. It is worth noticing that both Kuiji and Zenju consider that a single aspect of cognition can function both as prameya and pramāṇaphala at the same time in Phases Three and Four. This is a remarkable difference between Dignāga and Dharmapāla: the former accepts only Phase One as his own theory. Kuiji admits the interchangeability of the four aspects. That is, he simply accepts the doctrine of the cwsl and the fdjl as his own. Huizhao 慧沼 (650–714), another notable Chinese scholar-monk of Yogācāra, also follows the basic tradition of Xuanzang’s two texts. The following passage of the Dasheng fayuan yilin zhang buque 大乘法苑義林章補闕 clearly reveals Huizhao’s deviation from Dignāgean epistemology: After [Dignāga], Bodhisattva Dharmapāla and so forth established the fourfold division, by adding “cognition of self-cognition.” The following is implied: If the appearance aspect works as the object of cognition, seeing as the subject of cognition, and self-cognition as the result, then, selfcognition has the seeing aspect as the object of cognition, self-cognition works as the subject of cognition. (Zokuzō kyō 續藏經 2.3.1,32d–33a = R98, 64b–65a.) Finally I want to briefly introduce two standard explanations which demonstrate Dharmapāla’s influence in China: one is a commentary on the Buddhāvataṃsaka-sūtra (Huayan jing 華嚴經) by Chengguan 澄觀 (738–839) (T36, 252a-b) and the other is a compendium entitled Mirror of Fundamental Teaching (Zongjing lu 宗鏡録) by Yanshou 延壽 (904–975) (T48, 761a-b). Their

20

As briefly mentioned in n. 14 above, Kuiji’s interpretation was succeeded by the Japanese commentator Zenju.

The Genesis of *Svasaṃvitti-saṃvitti Reconsidered

221

contents are nearly identical. Yanshou quotes the passage by referring to its source as “Qingliang Record” 清涼記. However, no text bearing such a name is found in the Buddhist canon. Judging from the identity of the passage and the fact that Chengguan’s place of activity was Mt. Qingliang 清涼山 (another name of Wutai shan 五臺山),21 it seems almost certain that Yanshou quotes Chengguan’s commentary on the Buddhāvataṃsaka-sūtra. Furthermore, we can also ascertain that both texts have references to the cwsl and Kuiji’s commentary thereon.22 What is more interesting is the use of the mirror as a metaphor by Chengguan and Yanshou. I will quote the passage from Chengguan’s text: [The fourfold division of cognition] can be illustrated by a clear mirror, too. [Namely,] the image on the mirror illustrates the object[-aspect of cognition]; the light of the mirror illustrates the subject[-aspect of cognition]; the surface of the mirror illustrates self-cognition; and the backside of the mirror illustrates the cognition of self-cognition. The surface depends on [the existence of] the backside, and likewise the backside depends on the surface. Thus both sides are mutually dependent. [In the case of a copper mirror] it is also possible to take copper as the cognition of self-cognition; mirror depends on copper and copper depends on mirror. (T36, 252b)23 To the best of my knowledge, the metaphor of the mirror in the context of selfcognition is not attested in Indic texts. In particular, it seems quite possible to me that the references to “the backside of the mirror” and “copper” are based on the actual shape of Chinese mirrors. Regarding the backside of the mirror, moreover, Yanshou’s Zongjing lu uses wording that illustrates “cognition of self-cognition” by the central part (ba 弝) of the backside.24 This also suggests the physical form of a Chinese mirror.

21

Chengguan was a monk of the Great Huanyan Monanstery of Mt. Qingliang 唐 清涼山大 華 嚴 寺 沙 門 . This expression is found in numerous locations (e.g., T36,1a10a, 17a). He is also called Qingliang guo shi 清 涼 國 師 in later texts. 22 Concretely cwsl (T31, 10b) and cwsl-shuji 述 記 (T43, 319b-c) are mentioned by Chengguan and Yanshou. 23 Almost the same explanation is found in Yanshou’s Zongjing lu, too (T48, 761b). See also the next note. 24 《 宗 鏡 録 》 鏡 如 自 證 分 。 鏡 明 如 見 分 。 鏡 像 如 相 分 。 鏡 後 弝如 證 自 證 分 。 (T48, 759c) The Chinese character ba 弝 literally means “the part of a bow grasped

222 5

FUNAYAMA

Final Thoughts

In this paper, I have endeavored to examine whether there is a significant difference between Dignāga’s and Dharmapāla’s epistemology. I would like to suggest that we should doubt the authenticity of Dharmapāla’s fourfold division of consciousness in Xuanzang’s “translations” of the cwsl and the fdjl. If my hypothesis is valid, it entails the “apocryphal” character of both texts, if not entirely then at least partially. No doubt there are more questions to be discussed on this issue. For example, I did not mention how Chinese Buddhist commentators explain the view that even a conceptual cognition (vikalpa) could be regarded as direct perception (pratyakṣa) insofar as it is treated as self-cognition. This idea is found in Dignāga’s Pramāṇasamuccaya i 7ab.25 In this paper I could not examine whether such a view is also found in Chinese Yogācāra texts because self-awareness of conceptual construction is not clearly discussed in Xuanzang’s translation of the Nyāyamukha. In any case, closer examination of *svasaṃvitti-saṃvitti in the Indian context awaits future research.

Abbreviations and References

PSṬ  Jinendrabuddhi’s Viśālāmalavatī Pramāṇasamuccayaṭīkā. Chapter 1. Part I: Critical Edition, edited by Ernst Steinkellner, Helmut Krasser, and Horst Lasic. China Tibetology Publishing House/Austrian Academy of Sciences Press, 2005. T Taishō shinshū daizō kyō 大正新脩大藏經 (Taisho Canon).

25

when shooting.” It is not easy to demonstrate the Chinese origin of expressions like this, but I suspect that it does not fit the ordinary description of an Indian mirror. In Sanskrit texts “mirror” is expressed by ādarśa, darpaṇa and the like, and the notions of the original image and the reflected image on the mirror are expressed by such Skt. words as bimba and pratibimba. However, I have not seen a Skt. text referring to copper and the backside of the mirror, not to mention the part ba 弝 of a mirror. ps i 7ab: kalpanāpi svasaṃvittāv iṣṭā nārthe vikalpanāt / (Steinkellner 2005 3, 13); “Even conceptual construction, when it is brought to internal awareness, is admitted [as a type of perception]. However, with regard to the [external] object, [the conceptual construction is] not [admitted as perception], because it conceptualizes [the object.]” (Hattori 1968: 27). This view is also accepted by Dharmakīrti in the Pramāṇavārttika iii 287: śabdārthagrāhi yad yatra taj jñānaṃ tatra kalpanā / svarūpaṃ ca na śabdārthas tatrādhyakṣam ato ’khilam //.

The Genesis of *Svasaṃvitti-saṃvitti Reconsidered

223

Cook, Francis H. 1999. Three Texts on Consciousness Only: Demonstration of Consciousness Only. Berkeley: Numata Center for Buddhist Translation and Research. Fukihara Shōshin. 1935. 富 貴 原 章 信 , “Shibun to ryōchi ni tsuite — yuishiki shibun setsu no ichibu to shite 四 分 と 量 智 に 就 て — 唯 識 四 分 説 の 一 部 と し て ,” Ōtani gakuhō 大 谷 学 報 16–2: 248–276. Funayama Toru. 2000. “Two Notes on Dharmapāla and Dharmakīrti,” Zinbun 35: 1–11. http://repository.kulib.kyoto-u.ac.jp/dspace/handle/2433/48788. Funayama Toru. 2014. “Chinese Translations of pratyakṣa.” In Chen-kuo Lin and Michael Radich (eds.), A Distant Mirror: Articulating Indic Ideas in Sixth and Seventh Century Chinese Buddhism. Hamburg: Hamburg University Press, 2014, pp. 33–61. http://hup.sub.uni-hamburg.de/volltexte/2014/146/chapter/HamburgUP_HBS03_ Funayama_LinRadich_Mirror.pdf. Funayama Toru. 2015. “Chinese Buddhist Apocrypha.” In Jonathan A. Silk et al. (eds.), Brill’s Encyclopedia of Buddhism, Vol. 1: Literature and Languages, Leiden: Brill, 2015, pp. 283–291. Hattori Masaaki. 1968. Dignāga, On Perception: Being the Pratyakṣapariccheda of Dignāga’s Pramāṇasamuccaya from the Sanskrit Fragments and the Tibetan Versions, Cambridge MA: Harvard University Press, 1968. Johnston, E. H. and Arnold Kunst, eds.1986. The Dialectical Method of Nāgārjuna: Vigrahavyāvartanī. Second Edition (Revised and Enlarged). Translated from the original Sanskrit with Introduction and Notes by Kamaleswar Bhattacharya, Text critically edited by E. H. Johnston and Arnold Kunst. Delhi: Motilal Banarsidass, First Edition 1978, Second Edition 1986. Kajiyama Yuichi. 1966. “An Introduction to Buddhist Philosophy – An Annotated Translation of the Tarkabhāṣā of Mokṣākaragupta,” Kyōto daigaku Bungakubu kenkyū kiyō 京 都 大 學 文 學 部 研 究 紀 要 (Memoirs of the Faculty of Letters, Kyoto University) 10, 1966, pp. 1–173. http://repository.kulib.kyoto-u.ac.jp/dspace/ handle/2433/72848. Katsura Shōryū. 1976. 桂 紹 隆 , “sarvālambanajñāna ni tsuite,” Indogaku bukkyōgaku kenkyīū 24–2: 160–161. https://www.jstage.jst.go.jp/article/ibk1952/24/2/24_2_678/_ article. Keenan, John P. 2002. The Interpretation of the Buddha Land. Berkeley: Numata Center for Buddhist Translation and Research. de La Vallée Poussin, Louis. 1928. Vijñaptimātratāsiddhi. La siddhi de Hiuan-tsang, Tome 1, Paris: Librairie orientaliste Paul Geuthner. Steinkellner, Ernst. 2005. Dignāga’s Pramāṇaviniścaya, Chapter 1: A Hypothetical Reconstruction of the Sanskrit Text with the Help of the Two Tibetan Translations on the Basis of the Hitherto Known Sanskrit Fragments and the Linguistic Materials

224

FUNAYAMA

Gained from Jinendrabuddhi’s Ṭīkā. www.oeaw.ac.at/ias/Mat/dignaga_PS_1.pdf, April 2005. Yamaguchi Susumu. 1941. 仏 教 に お け る 無 と 有 と の 対 論 , (Bukkyō ni okeru mu to u to no tairon). Tokyo: Sankibo busshorin, 1941, revised in 1975. Yamaguchi Susumu. 1951. Seshin no Jōgō ron 世 親 の 成 業 論 , Kyoto: Hōzōkan.

Chapter 8

Dharmapāla on the Cognition of Other Minds (paracittajñāna) Shinya Moriyama Abstract The problem of other minds has been discussed in the Buddhist epistemological tradition, three arguments by Vasubandhu, Dharmakīrti, and Ratnakīrti having received the most attention from modern scholars. However, there is another significant resource for considering the problem, viz Dharmapāla’s distinct analyses of a bodhisattva’s and of a buddha’s cognitions of other minds. Through examining his commentary on the Viṃśikā called Cheng wei shi bao sheng lun 成唯識寶生論 and the Cheng wei shi lun 成唯識論 together with the Wei shi ershi lun shuji 唯識二十論述記 by Kuiji (窺基, 632–682), I argue first for the specificity of Dharmapāla’s basic tenet that accepts other minds and negates external objects, pointing out how it differs from Dharmakīrti’s and Ratnakīrti’s arguments on the same topic. Based on this analysis, especi,ally focusing on Dharmapāla’s argument concerning a buddha’s cognition of another’s mind, I also propose to read Dharmapāla’s argument as a variant of the socalled transparency thesis, which claims that one’s mental states are transparent not only to oneself but also to the Buddha, or more correctly, one’s mind is more transparent to the Buddha than to oneself.

Keywords Dharmapāla – Kuiji – problem of other minds – transparency thesis – reflexive awareness (svasaṃvedana) – yogic perception – cognition of other minds (paracittajñāna)

…

1 Introduction The philosophical problem of other minds is often argued from the point of asymmetry between first- and third-person attribution of consciousness. © koninklijke brill nv, leiden, ���1 | doi:10.1163/9789004440913_012

226

Moriyama

While one’s own mental states seem immediate or transparent to oneself, another’s mental states are not accessible in the same manner. To this transparency thesis, there is another position called the opacity thesis that every cognition is cognized only indirectly by inference, regardless of whether it is first- or third-person attribution of consciousness.1 In this volume (Chapter 6), Mark Siderits explores the possibility of interpreting Buddhist philosophy as a version of the opacity thesis. On the other hand, what I will explore in the following is a different approach to the problem of other minds with respect to Yogācāra idealism, which will lead to two conclusions: there is (1) symmetry between cognition of one’s own mind and the cognition of another’s mind, and (2) asymmetry between the cognition of another’s mind and the cognition of external objects. The two conclusions are derived from the examination of some arguments by Dharmapāla, a 6th century Buddhist philosopher of the Yogācāra school. His ideas on the problem of other minds are found in his commentary on the Viṃśikā called Cheng wei shi bao sheng lun 成唯識寶生論 (cwsbsl)2 and the Cheng wei shi lun 成唯識論 (cwsl)3 together with the Wei shi ershi lun shuji 唯 識二十論述記(wsel-Shuji) by Kuiji (窺基, 632–682),4 where Dharmapāla’s ideas are often adopted. Previous studies concerning Buddhist approaches to the problem of other minds have examined three important treatises, namely, the final section of Vasubandhu’s Viṃśikā, Dharmakīrti’s Santānāntarasiddhi, and Ratnakīrti’s Santānāntaradūṣaṇa. Of these, Vasubandhu expressed the first point, that the cognition of another’s mind and one’s cognition of one’s own mind are equally false in comparison to the Buddha’s supreme knowledge of the ineffable nature of the object.5 Partly following Vasubandhu’s position, Dharmakīrti provided us with a clear account of the structure of yogic perception of other minds. According to his account, what the yogi immediately perceives is an 1 Carruthers 2011. 2 Ui (1963: 607–620) assumes the original title of the text is Vijñaptimātratāsiddhiratnasaṃbhava. As Ui has noted, the commentary does not contain Viṃśikā v. 1, vv. 11–15, and v. 22. Ui complains about Yi-jing’s translation, especially its different translations of the same Sanskrit term. 3 On the problem of other minds discussed in Xuanzang’s cwsl, Brewster (forthcoming) presents a detailed and systematic study based on Yogācāra textual sources from Indian and Chinese Buddhism. I thank Mr. Ernest Billings Brewster for sharing this valuable paper. 4 I follow Kuiji’s traditional name. See Schmithausen (2015: 17, fn. 21). 5 Viṃś 10.19–28. For the title of the text, see Kano (2008). For recent important studies on this text, see Kellner & Taber (2014); Silk (2016). For a detailed analysis of Vasubandhu’s and the cwsl’s arguments on paracittajñāna, see Yamabe (1998). For a recent Japanese study on the text, see Hyodo (2006).

Dharmapāla on the Cognition of Other Minds (paracittajñāna)

227

image of another’s mental state, which appears vividly to yogic cognition. Since the reliability of the mental image is justified by the subject’s action toward the object, like in external matters, this account presupposes the external existence of the other’s mental continuum.6 While Dharmakīrti argues on yogic perception of other minds within the context of conventional reality, his later successor Ratnakīrti asserts that other minds do not exist in the ultimate sense.7 That is to say, whereas Dharmakīrti accepts the existence of other minds as well as external objects in the conventional realm, Ratnakīrti denies both other minds and external objects in the ultimate sense. Compared to those Buddhist philosophers, Dharmapāla follows Vasubandhu’s claim, but takes a different position from both Dharmakīrti and Ratnakīrti. Dharmapāla maintains that other minds exist but external objects do not. In the following, we shall look closely at his discussion of these two points.8 2

Dharmapāla’s Commentary on Viṃśikā 21

Yogācāra idealism is established by the acceptance of the existence of consciousness and its mental states and the denial of any external objects. What exists is only what appears in one’s cognition. This doctrine is threatened by another Buddhist doctrine of paracittajñāna, that is, the Buddha’s and advanced yogis’ or bodhisattvas’ supernatural intuition of other persons’ mental states, which presupposes the existence of others’ minds outside the subject’s own mental continuum.9 If the latter is correct, it would establish the existence of something external from one’s own consciousness. Then the proponent of the Yogācāra doctrine would be required to explain in a plausible manner why only other minds exist and not other external entities. If Yogācāra

6 sas §88–94. For the text, see Katsura (1983). 7 On Ratnakīrti’s refutation of other’s mental continuum, see Inami (2001). 8 For Dharmapāla’s epistemology, see Tillemans (1990: 54–66). For Dharmapāla’s date, see Funayama (2000), esp. fn. 52. 9 The Buddhist doctrine of the cognition of other minds (paracittajñāna) has been traditionally acknowledged as one of the six kinds of supernatural intuitions (abhijñā) acquired through meditation by buddhas, yogis, or bodhisattvas. For the six kinds of supernatural intuition, see AK(Bh)vii 42–56. For paracittajñāna as one of the ten kinds of cognition (daśajñāna), see ak vii 11; AKBh 396.3–398.21, where Vasubandhu divides paracittajñāna into the non-contaminated (anāsrava) and the contaminated (āsrava), and discusses the image (ākāra) and the objective support (ālambana) of paracittajñāna. For some examples of the Buddha’s and bodhisattvas’ paracittajñāna, see YBh T. 30. 494c2–22.

228

Moriyama

idealism is correct, on the other hand, Ratnakīrti’s conclusion, that neither others’ minds nor external objects exist, would be more reasonable. On this dilemma, Dharmapāla presents his position in his cwsbsl on Viṃśikā 21 and its Vṛtti. But before examining that portion of the text, let us first look at its openingmirrorsion where a bodhisattva’s, compass,ion toward other sentient beings is explained in relation to the cognition of other minds. To the question, “How is compassion possible in Yogācāra idealism that denies all external entities?” Dharmapāla replies as follows: Provisionally having the other’s [mental] continuum as a determinant condition (zengshang yuan 増上縁, adhipatipratyaya), the image (xiang 相) of a sentient being appears in [a bodhisattva’s] own cognition, and by taking it (i.e., the image) as the object, [the bodhisattva] brings about great compassion and widely provides sentient beings with benefits.10 Here Dharmapāla uses the term “determinant cause” (adhipatipratyaya), which signifies an external cause that influences the effect differently from its material cause.11 That is, the other’s mental continuum is understood to exist externally to the subject’s mental stream. By relying on such a determinant cause, a bodhisattva’s compassion arises. In compassion, the image of a sentient being is formed in the, bodhisattva’s own consciousness. The same structure is also found in Dharmapāla’s explanation of para­ cittajñāna. In the commentary on Viṃśikā 21, he defines it as follows: It is generally said that [a bodhisattva] apprehends another person’s mind, when [the bodhisattva] apprehends his/her own mind’s image (xiang 相) by taking another’s mind as the condition (yuan 縁, i.e., the determinant condition). This is provisionally called the “cognition of another’s [mind].”12 Although another’s mind itself differs from its image that appears in paracittajñāna, it is provisionally referred to as the “cognition of another’s mind” due to the similarity between the image and the object. Defining paracittajñāna in this manner, Dharmapāla further discusses the same idea as applied to one’s cognition of one’s own mental states. He comments

10 11 12

cwsbsl 78a5–18: 借 外 相 続 為 増 上 縁 ， 於 自 識 中 現 有 情 相 ， 縁 此 成 境 起 大 悲心，於有情処而作弘益. For the translation of adhipatipratyaya, I follow Schmithausen (2015). cwsbsl 96b24–25: 凡 云 了 他 心 ， 以 他 心 為 縁 , 了 自 心 相 ， 仮 説 知 他 .

Dharmapāla on the Cognition of Other Minds (paracittajñāna)

229

on Viṃśikā’s phrase “like one’s cognition of one’s own mental state” (yathā svacittajñānam) as follows: Since two cognitions do not arise simultaneously, [the cognition to be cognized] is not a present one, but either a past or future one. A cognition can become the object to be cognized, but its body does not exist [when it is cognized]. However, one experiences one’s own cognition only in the manner that one’s own cognition forms a vivid image (zhang xian xiang 彰顯相) with respect to past and present mental states (xin ju fa 心聚法), retrospectively.13 Like a bodhisattva’s cognition of other minds, which is analyzed as his cognition of an image t“hatresembles the other’s mind, one’s “cognition of one’s own cognition is understood in the same manner: since thereis a gap between a cognition to be cognized and another cognition that cognizes it, the cognition’s cognition is possible only in an indirect manner, through the medium of an image of the cognition to be cognized. This explanation presupposes a Buddhist epistemological position called sākārajñānavāda, namely, a kind of representationalist thesis that cognition occurs endowed with a Representationn image (ākāra) of the object. We do not perceive external objects directly asthey are, but only indirectly through an internal representation of an external objet.Likewise, since one’s previous cognition and the other’s mind are temporally or spatially separated from one’s present cognition, there is no difference between yogic perception of the other’s mind and one’s cognition of one’s previous cognition, As summarized byense that both ognitions involve the medium images of objects. Notice, however, that the so-called reflexive awareness (svasaṃvedana) does not play any role in Dharmapāla’s argument in the cwsbsl. If reflexive awareness is taken into consideration, another picture of the discontestable will emerge, which we will examine later in relation to Kuiji’s description of Dharmapāla’s argument. On the nature of a bodhisattva’s perception of his own cognition and other minds, Dharmapāla also explains it in comparison with the Buddha’s wisdom concerning the ineffable nature of all objects. In the explanation, the term zi zheng 自証 is used in the sense that the Buddha’s enlightenment is accessible only to himself; others cannot imagine the world that the Buddha sees. Compared to such ultimate cognition, a bodhisattva’s cognitions of his own

13

cwsbsl 96b13–16: 二 心 同 時 不 共 聚 故 ， 固 非 現 在 ， 決 定 応 許 已 滅 未 生 。 但 可得一而為其境，体復是無。但唯自識還縁過現諸心聚法為彰顯相， 領納自心.

230

Moriyama

cognition and another’s mind are said to be mistaken. Dharmapāla states at the very end of this treatise: [Question:] What does [the Buddha] cognize rightly with the wisdom that is [obtained] after the annihilation of sleep (i.e., defilements)? [Reply:] Through this wisdom he cognizes the ineffable nature [of consciousness]. Since [that nature] exceeds expressions, it is known only to him. Therefore, it cannot be expressed. However, it is [known] that all natures [of entities] exist in [his] consciousness (i.e., his consciousness is omniscient). This is not known by other conscious beings. [The Buddha’s realm] is unknown and ineffable, because [other cognitions and words] have universal natures (zong xiang 総相, sāmānyalakṣaṇa) as their objects. On the other hand, this (bodhisattva’s cognition) has merely the nature of the imagined (wang gouhua xing 妄構画性, parikalpitasvabhāva), that is, the nature that is imagined with respect to what [the Buddha] knows by himself, because [the object that is cognized by a bodhisattva’s cognition] is far removed from the true nature of consciousness. [The bodhisattva’s cognition] does not cognize the true nature merely with respect to consciousness. This [characteristic] is established in both cases (i.e., one’s cognition of one’s own cognition and the cognition of other minds). [The two kinds of cognition] do not correspond to reality. Why? This is because [the bodhisattva] conceives an unreal entity as a real entity and determines [it as the object], due to the appearance of a mistaken image in his consciousness.14 According to this account, the Buddha perceives all natures of everything including other minds, but this realm is inaccessible to anybody else. This probably means that he perceives the true nature of consciousness, namely, the perfected nature (pariniṣpannasvabhāva). By contrast, bodhisattvas, who are inferior to the Buddha, cognize objects that are of the nature of the imagined (parikalpitasvabhāva). The same is true when a bodhisattva perceives his own cognition or another’s mind. In both cases, he perceives not the object’s true nature but an imagined nature that is superimposed by ignorance. In this 14

cwsbsl 96c6–14: 復 云 :何 通 睡 尽 之 智 能 正 了 斯 。 由 此 覚 知 無 言 境 性 。 超 過 語路、但自証知。是故不能以言詮及。然於此識所有自性、非是余識 之所能知。既非所知、非言能及。彼但総相為其境故。然斯唯有妄構 画性。即此構画於自証性。識之実相極遼遠故。唯於識処了不実相。 此二皆成。不称実境。所以者何。於非実事作実事解而為決断、由於 彼識現虚妄相故.

Dharmapāla on the Cognition of Other Minds (paracittajñāna)

231

sense, we can acknowledge symmetry between first- and third-person attributions of mental states, which are equally superimposed in a bodhisattva’s cognition. This position is stated first by Vasubandhu and reinforced by Dharmapāla with his theory of sākārajñānavāda or Buddhist representationalism. 3

A Passage Ascribed to Dharmapāla in the Cheng wei shi lun

Let us move to our second point, namely, the asymmetry between the cognition of other minds and the cognition of external objects. In the Cheng wei shi lun, Xuanzang (玄奘, 602–664) compiles ten authoritative commentaries on Vasubandhu’s Triṃśikā, including Dharmapāla’s. An important argument on cognition of other minds or paracittajñāna that is ascribed to Dharmapāla is contained in this text. Commenting on Triṃśikā 17, Xuangzang provides an additional argument consisting of ten questions and answers about Yogācāra idealism. The eighth question and answer are as follows: [Question:] If no external matter (se 色, rūpa) exists in reality, [such a nonexistent entity] should not be the object of one’s internal cognition. [However,] since the other’s mind exists in reality, why is [the other’s mind] not the objective support (suoyuan 所縁, ālambana) of one’s own [cognition]? [The other’s mind should be the immediate objective support for cognition.] [Reply:] Who says that the other’s mind is not the object of one’s own cognition? [Like you, we also claim that the other’s mind is the objective support for cognition. Yet,] we do not claim that the [other’s mind] is the immediate objective support (qin suoyuan 親所縁) [for the cognition]. When cognition arises, it has no real activity [toward an external other’s mind], unlike hands that directly grasp external things and unlike sunshine that directly illuminates external objects. [The cognition] is like a mirror on which a similar [image] of an external object appears. [Such a mirror-like cognition concerning the other’s mind] is [here] called the “cognition of another’s mind.” [The cognition] does not immediately cognize [an external other’s mind]. What is immediately apprehended [by this cognition] is [an image] into which [the cognition] transforms by itself. Therefore, a Buddhist text (Sandhinirmocanasūtra) states: “There is no [mental] entity that [immediately] grasps another entity. It is just that when a cognition arises and [an image] that is similar to the [object] appears [in the cognition], it is [provisionally] called ‘The [cognition]

232

Moriyama

grasps that thing.’”15 Just as one has the other’s mind as the objective support [as above], so matter, etc., are likewise.16 Here, too, paracittajñāna is interpreted as cognition that arises together with an image of another’s mind, like in the exposition of the commentary on Viṃśikā 21. However, unlike the previous commentary, the cwsl’s argument makes the characteristic of sākārajñānavāda clear by distinguishing between image-cum-cognition from cognition that acts directly toward the external object itself. If we use the term “mediate objective support” (shu suoyuan疎所縁) as the antonym of “immediate objective support” (qin suoyuan親所縁), the other’s mind is the mediate objective support, and its mental image is the immediate objective support. However, if another’s mind is accepted as existing mind-independently, the same would be applied to matter (rūpa), which also appears as an image in one’s consciousness. That is, Yogācāra idealism does not necessarily deny the existence of external matters. In fact, the last phrase ru yuan ta xin se deng yi er 如縁他心，色等亦爾 might be understood in this manner, as Dan Lusthaus points out: This is tantamount to admitting that rūpa exists independently, though not separate from my mind. It is known indirectly, it is a remote ālambana. How could such a crucial passage have gone unnoticed for so long? (Lusthaus 2002: 491) To this, however, Nobuyoshi Yamabe (1998: 31) has already noted that “[t]he structure appears actually to be closer to a realistic model, though we should note that ‘matter’ in this case is reduced to the ālayavijñāna of other sentient beings,” and Lambert Schmithausen has also leveled a severe criticism on the ground of thorough research on related evidence in the cwsl and early Yogācāra texts. He concludes that 15 As Schmithausen (2015: 14, fn. 13) notes, the passage is from sns viii, 7 (91, 8–11), whose Sanskrit text is found in jna 478.3–4: na hi maitreya tatra kaścid dharmaṃ pratyavekṣate, api tu tathāsamutpannaṃ tac cittaṃ yat tathā khyāti. 16 cwsl 39c9–16: 外 色 実 無 可 非 内 識 境 。 他 心 実 有 寧 非 自 所 縁 。 誰 説 他 心 非 自識境。但不説彼是親所縁。謂識生時無実作用，非如手等親執外 物，日等舒光親照外境。但如鏡等似外境現，名了他心。非親能了。 親 所 了 者 謂 自 所 変 。 故 契 経 言 :“無 有 少 法 能 取 余 法 。 但 識 生 時 似 彼 相 現 名 取 彼 物 。 ”如 縁 他 心 ， 色 等 亦 爾 . For translation and background, see Yamabe (1998: 31); Cook (1999: 239); Lusthaus (2002: 490–491); Schmithausen (2015: 13–46). Cf. wsesl Shu ji, 1007b17–24.

Dharmapāla on the Cognition of Other Minds (paracittajñāna)

233

[…] it would seem to me extremely unlikely that the words “visible matter, etc.” refer to any kind of matter existing independently. […] [T]here is sufficient internal evidence in the cwsl, supported by the earliest commentaries and partly even by independent Indian sources, indicating that wherever consciousness has an external “original” or “remote objective support” this [is] understood as consisting either in other consciousness or mental factors or, in the case of matter,mental image of matter contained in other consciousness, either in one’s own mind continuum or in that of other sentient beings. (Schmithausen 2015: 45) The “other consciousness” here refers to ālayavijñāna, the eighth consciousness, in one’s own mind or another’s mind. Material things, including bodies, are images that occur in ālayavijñāna, and the image becomes the mediate objective support for one’s ordinary cognition of matter. Thus, when one says that one cognizes matter as a bodhisattva cognizes another’s mind, it does not mean that matter exists just like other minds. Whereas matter is just an illusion, other minds really exist, and thus they become the mediate objective supports for paracittajñāna. 4

Kuiji’s Analysis of Dharmapāla’s Interpretation of paracittajñāna

So far we have seen how Dharmapāla discussed the equation of one’s prior cognitions with another’s mind, both being mental images that appear in a bodhisattva’s immediate cognition, as well as the difference between another’s mind and the external object, both of which are commonly presumed to exist outside the bodhisattva’s consciousness. As a Yogācāra philosopher, Dharmapāla does not accept any external objects, but he approves the ­independent existence of other minds because bodhisattvas’ altruistic activities out of compassion would be unintelligible otherwise. As we have seen before, however, an important element of Yogācāra idealism is missing here. That is the theory of reflexive awareness (svasaṃvedana). If the above argument is reexamined by bringing that theorySchwindto consideration, how would the argument change? We find an answer in the Weishi ershi lun shuji, a Chinese commentary on Viṃśikā and Vṛtti, by Kuiji.17 17

In this commentary, Kuiji explains the notion of paracittajñāna and summarizes the difference in its interpretations, between Dharmapāla and other masters like Sthiramati (470–550). For the difference of opinions on the Yogācāra doctrine between Sthiramati and Dharmapāla, see Katsura (1969: 15–16).

234

Moriyama

He explains the structure of paracittajñāna in the framework of Dharmapāla’s so-called “fourfold division theory,” based on the assumption that cognition is divisible into four parts:18 (1) the part of the [object’s] image (xiang fen 相 分, *nimittabhāga), (2) the part of seeing (jian fen 見分, *darśanabhāga),19 (3) the part of reflexive awareness (zi zheng fen 自証分, *svasaṃvittibhāga), and (4) the part of cognition of reflexive awareness (zheng zi zheng fen 証自 証分, *svasaṃvittisaṃvittibhāga). Of course, as Funayama argues in detail in this volume, the attribution of this theory to Dharmapāla is not without doubt given the extant textual evidences.20 Yet, at least for Kuiji, who believes the theory comes from Dharmapāla, the relationship between the notion of paracittajñāna and the fourfold division theory is an important issue to be discussed. The point is that a bodhisattva’s cognition of another’s mind clearly differs from another’s reflexive awareness of his/her own cognition. That is, in the bodhisattva’s paracittajñāna, another’s mind and mental concomitants take the role of object-image component (xiang fen) relative to the seeing component. However, this does not imply that the cognition has another’s mind itself as the immediate objective support. Otherwise, a bodhisattva would be able to grasp another’s mind transparently, but it is hardly intelligible that even a bodhisattva would have access to what it is like for the other person. Kuiji explains: Therefore, although paracittajñāna [is said] to cognize just the other person’s mind and the associated mental concomitants, it cognizes neither its (i.e., the other’s cognition’s) object nor the component of reflexive awareness that cognizes it, because such a cognition is not called the “cognition of other minds.” This explains [paracittajñāna] in the state of cause (i.e., the state of the bodhisattva) but not [paracittajñāna] as the Buddha’s cognition, etc.21

18 Cf. wsesl Shu ji, 1006b28–c4. 19 According to Yao (2005: 146), the two are probably derived from Asaṅga’s terms, *nimittavijñapti and *darśanavijñapti, respectively. See also Schmithausen (2015: 13–14, fn. 12) in reference to snsṬ’s passage: lta ba dang rgyu mtshan gyi rnam par yong su gyur pa. 20 See Funayama, Chapter 7. For the fourfold division theory, see also Frauwallner (1958: 403f.); Funayama (2000: 6, fn. 31), Yao (2005: 145–147), and Moriyama (2019). The theory is comparable with Dignāga’s theory of reflexive awareness in ps i, 10. 21 wsesl Shuji 1006c4–7: 故 他 心 智 但 知 他 身 心 心 所 法 ， 不 縁 彼 境 及 彼 能 縁 自 証 分 *， 不 名 他 心 智 故 。 此 説 因 位 。 非 仏 等 心 。 (*Read 彼 能 縁 自 証 分 for T: 彼 能 縁 自 自 証 分 ).

Dharmapāla on the Cognition of Other Minds (paracittajñāna)

235

Kuiji carefully avoids confounding the cognition of another’s mind and mental concomitants as a whole with the cognition of the constituents of another’s cognition’s such as the object-image component and the reflexive awareness component. Since the reflexive awareness and the cognition of reflexive awareness are related to one’s first-person privileged access to what it is like for one’sconsciousness “at one moment, that realm is inaccessible to any cognition other than the very cognition itself. In this sense (as Siderits also points out), the theory of reflexive awareness supports a version of the transparency thesis. The same point is repeated in Kuiji’s explanation of the phrase “like one’s cognition of one’s own mental state” (yathā svacittajñānam, 如自心智): [A bodhisattva’s] cognition of another’s mind is called so because it takes another’s mind as the original (zhi 質, i.e., mediate objective support) and the [bodhisattva’s] mind changes into the condition (yuan 縁, i.e., immediate objective support), but not because it immediately cognizes another’s mind, etc. It is like the fact that various cognitions that cognize [their own minds] do not immediately grasp [them]. Since [one’s own cognition] just changes into the condition (i.e., the image of one’s mind), which is different from the original (i.e., one’s mind itself), [the cognition of one’s own mind] is said to not correspond to reality (bu rushi不如実). The “cognition of one’s own mind” (zi xin zhi自心智 svacittajñāna) indicates the seeing component (jian fen 見分), because it is admitted that it (i.e., the seeing component) cognizes the image into which one’s own prior or subsequent cognition changes. However, the reflexive awareness component (zi zheng fen 自証分) and the other (i.e., the component of cognition of reflexive awareness) are not called the “cognition of one’s own mind,” because there is no objection to the [view] that they (i.e., reflexive awareness and the cognition of reflexive awareness) correspond to the reality (rushi zhi 如実知).22 The cwsl’s two concepts, mediate and immediate objective supports, are here discussed in relation to the distinction between “original” (zhi 質) and “condition” (yuan 縁). In both paracittajñāna and svacittajñāna, one cognizes the object mediated by its condition, namely, its image that appears in the cognition,

22

wsesl Shuji 1007a2–5: 以 他 心 為 質 而 心 変 縁 , 名 他 心 智 。 非 能 親 取 他 心 等 故 名他心智。如縁自心諸所有智，亦不親取。但変而縁，与本質異，名 不如実。此「自心智」説見分者。前後許自変相縁故。非自証分等名 為自心智。彼如実知無異解故.

236

Moriyama

and never grasps its original immediately.23 In arguing so, Kuiji assumes the gap between the original and the condition, due to which not only the cognition of another’s mind but also the cognition of one’s own mind is judged to be ultimately false, that is (provinces), not corresponding to reality. While the original of another’s mind never appears immediately to the cognition of another’s mind, in the case of the cognition of one’s own mind, too, it never cognizes one’s own mind immediately. Interestingly, in the above argument, Kuiji introduces some technical terms of Dharmapāla’s fourfold division of consciousness and makes the contrast between the correctness of reflexive awareness and the incorrectness of other types of cognitions, or between the indubitability of first-person givenness and the dubitability of the third-person approach to another’s mind. Even though a bodhisattva’s paracittajñāna is admitted as a perception, which differs from any inferential approach to another’s mind, there is still an asymmetry between object-oriented cognition and reflexive awareness. Thus, if we follow Kuiji’s line of argument, the transparency thesis will be affirmed. In fact, this is exactly the direction that Dharmakīrti once suggested in his pv iii 455–459, where he distinguishes between yogic perception of another’s feeling and one’s reflexive awareness of one’s feeling.24 However, the transparency thesis is nothing to do with Dharmapāla’s argument on the cognition of another’s mind. The fact that he mentions neither reflexive awareness nor the fourfold division theory of consciousness in the context of a bodhisattva’s cognition of another’s mind in his commentary on Viṃśikā 21 shows his indifference concerning the transparency thesis. What is instead important for him is the structural similarity between paracittajñāna and svacittajñāna, both of which are mediated by the object’s image, whereby the cognition’s immediate connectedness to the object itself is denied. In this sense, unlike Dharmakīrti, Dharmapāla’s position has affinity with the opacity thesis. Lastly, turning to the Buddha’s cognition of other minds, which is usually explained in connection with the notion of the so-called “subsequent cognition”

23 Cf. wsesl Shuji 1007b26–28: “That is, one’s own cognitions, etc., have another person’s real mind as a determinant condition (adhipatipratyaya), namely, the original to be grasped. One’s own mind transforms itself and forms the image, which is similar to the real [other’s mind]. Since it is said that the part of seeing cognizes [the image of] the other’s mind, it is called the “cognition of another’s mind.” (即 自 心 等 以 他 実 心 為 増 上縁，所取本質。自心別変作相分心，似他本物。説此見分為了他 心 ， 名 他 心 智 ). 24 See Moriyama 2010.

Dharmapāla on the Cognition of Other Minds (paracittajñāna)

237

(pṛṣṭhalabdhajñāna), we can see the issue from yet another viewpoint.25 According to Kuiji’s detailed analysis, paracittajñāṇa is divided into the contaminated (sāsrava) paracittajñāna and the non-contaminated paracittajñāna. Of the two, the latter is the Buddha’s and the bodhisattvas’ so-called “subsequent cognition,” which is further divided into two types according to whether clinging to entities (fa zhi 法執, *dharmagrāha)26 is completely removed or not. The Buddha’s subsequent cognition that cognizes other minds is the one that lacks clinging to entities. On this Buddha’s non-contaminated paracittajñāna, Kuiji introduces three different interpretations: it contains (1) neither the object-image component nor the seeing component, (2) only the seeing component, or (3) both the object-image and the seeing components.27 Of these, the third is attributed to Dharmapāla,28 though no textual evidence is found in the cwsbsl. According to Kuiji’s explanation, Dharmapāla claims that the Buddha’s paracittajñāna has both the object-image component and the seeing component, like all other cognitions. If the cognition is mediated by the object’s image, it would be a false cognition that does not correspond to reality. However, this is not the case for the Buddha’s subsequent cognition. As Kuiji states: The Buddha’s paracittajñāna, which changes into the object (i.e., the image of another’s mind), is said to be [a correct cognition] that corresponds to reality (rushi 如実), because it strongly resembles (qin si 親似) another’s mind. Due to the lack of clinging [to entities], it cognizes that the [true] nature [of consciousness] is ineffable (xing liyan 性離言). The other cognitions of another’s mind, which weakly resemble (shu si 疎似) another’s mind due to clinging [to entities], are said to be [false cognitions] that do not correspond to reality (bu rushi 不如実). Due to the clinging [to entities], it does not cognize that the nature of all entities is ineffable. Therefore, it is said that a divergence [between cognition and 25 26

27 28

For the translation “subsequent cognition,” I follow Schmithausen (2015). The term “clinging to entities” (法 執 ) is a concept paired with “clinging to Self” (我 執 ), and divided into one deemed innate (俱 生 ) and one considered conceptual (法 執 ). Cf. cwsl 6c26–7a17. For the detailed investigation of the cwsl’s usage of the term, see Schmithausen (2015: 25–46). wsesl Shuji 1008a2–3. Cf. cwsl 50b17–21. wsesl Shuji 1008a2–3: “The [Buddha’s] subsequent cognition possesses two parts (i.e., the component of the object’s image and the component of seeing). [Dharmapāla] claims that this thought (siwei 思 惟 i.e. subsequent cognition) resembles the characteristic of suchness (真 如 相 , tathatā-lakṣaṇa), because it does not cognize true suchness.”; 後 得 智二分倶有。説此思惟似真如相，不見真実真如性故。.

238

Moriyama

its object’s true nature] exists. Here, in general, all paracittajñānas except for the Buddha’s are said to be [false cognitions] that do not correspond to reality. But it is not the case that the Buddha’s cognition is said to be [a correct cognition], one that corresponds to reality because it immediately illuminates another’s mind.29 The contrast between the Buddha’s paracittajñāna and other paracittajñāna is clearly shown here. The point of the final sentence is that a buddha’s paracittajñāna does not fully capture what the cognition itself is like because, although it is non-dual, it is indirect, being mediated by the image of another’s mind. Although mediation by the object’s image is common to both cognitions, only the former strongly resembles the object, and therefore is called a correct cognition that corresponds to reality. Certainly, since Kuiji makes no reference to reflexive awareness in this context, it is unclear whether the Buddha’s paracittajñāna gives transparent access to another’s mind as does one’s reflexive awareness. Yet, as far as the section of Kuiji’s commentary on Viṃśikā 21 is concerned, the term rushi 如実 (yathārtha) is applied only to the reflexive awareness component and this paracittajñāna of a buddha. This may suggest a version of the transparency thesis; yet, unlike the ordinary transparency thesis, this version indicates a symmetry between the Buddha’s cognition of another’s mind and one’s reflexive awareness of one’s mind. Alternatively, if we focus on the Buddha’s paracittajñāna as the cognition of the ineffable, true nature of consciousness, which is inaccessible for anybody else, it can be also understood as another version of transparency thesis: since the true nature or ­suchness of one’s mind is transparent only to the Buddha but not to the subject who introspects one’s own mind, we see here an asymmetry between the Buddha’s third-personal and one’s first-personal approaches to one’s mind. Note that in this case, contrary to the ordinary transparency thesis, the third-person approach is rather veridical in the sense that it makes known the true nature of one’s mind. The Buddha knows more about myself than I do. Although those two versions of transparency thesis are highly speculative, it is at least clear that the above Buddhist account of paracittajñāna has the potential of giving hints for considering the problem of other minds and inter-subjectivity through fresh eyes. 29

wsesl Shuji 1008a18–23: 仏 他 心 智 雖 変 為 境 親 似 他 心 , 名 為 「 如 実 」 。 以 無執故, 知性離言。余他心智亦変為境，未断執故，疎似他心, 名「不 如実」。以有執故，不知諸法体性離言。故説有異。此中通説除仏以 外諸他心智，説「不如実」。非仏此智親能照了他心等故名爲「如 実」.

Dharmapāla on the Cognition of Other Minds (paracittajñāna)

239

5 Conclusion As Dharmakīrti argued in his Santānāntarasiddhi, the argument from analogy is a common method for ordinary people to infer others’ mental states from their behavior. In this case, asymmetry between first- and third-person attribution of mental states is well established. However, once yogic perception of other minds is introduced into the discussion, the problem becomes more complicated. As we have seen, Dharmapāla maintains both symmetry between a bodhisattva’s cognition of the other’s mind and one’s cognition of one’s own mind, and asymmetry between the cognition of other minds and the cognition of external objects. However, the two conclusions, which are derived from the examination of a bodhisattva’s paracittajñāna, are inapplicable to the case of the Buddha. According to Kuiji’s account based on the view attributed to Dharmapāla, the Buddha’s subsequent cognition grasps the other’s mind through an image that strongly resembles the other’s mind. This suggests that one’s mental states are transparent not only to oneself but also to the Buddha. Furthermore, since the Buddha’s cognition grasps the suchness behind the mind, which is inaccessible to oneself, it can also be said that one’s mind is more transparent to the Buddha than to oneself. In this manner, the Yogācāra doctrine that is ascribed to Dharmapāla shows us alternatives to the ordinary transparency thesis. Through knowing the diversity of interpretations of the cognition of other minds in Buddhist epistemology, we can reconsider our tacit presumption of first-person privileged access to one’s own mind, which will lead us to a new horizon beyond the distinctions between self and other, namely, the ultimate statN/A of non-duality. But this issue clearly requires examination from different perspectives – something that must await another occasion. Primary Literature AK(Bh): Vasubandhu’s Abhidharmakośabhāṣya. P. Pradhan (ed.), Abhidharma­ kośabhāṣya of Vasubandhu. Patna: K.P. Jayaswal Research Institute, 1967. cwsl: Cheng wei shi lun 成 唯 識 論 of Dharmapāla et al., trans. Xuanzang, T. 1585. cwsbsl: Cheng wei shi bao sheng lung 成 唯 識 寶 生 論 of Dharmapāla, trans. Yijing, T. 1591. jna: Jñānaśrīmitranibandhāvali. A. Thakur (ed.), Jñānaśrīmitranibandhāvali. Patna: Kashi Prasad Jayaswal Research Institute, 1987. ps i: Dignāga’s Pramāṇasamuccaya, i (pratyakṣa). E. Steinkellner (ed.), Dignāga’s Pramāṇasamuccaya, Chapter 1. Online: http://www.ikga.oeaw.ac.at/Mat/dignaga_

240

Moriyama

PS_1.pdf (2005); A revision of the Sanskrit text, http://www.ikga.oeaw.ac.at/mediawiki/images/f/f3/Dignaga_PS_1_revision.pdf. pv iii: Dharmakīrti’s Pramāṇavarttika iii (pratyakṣa): H. Tosaki, Bukkyō ninshiki ron no kenkyū. Vol. ii. Tokyo: Daitō shuppansha, 1985. PVin ii: Dharmakīrti’s Pramāṇaviniścaya ii (svārthānumāna): E. Steinkellner (ed.), Dharmakīrti’s Pramāṇaviniścaya: Chapters 1 and 2. Beijing-Vienna: China Tibeto­logy Publishing House & Austrian Academy of Science Press, 2007. sas: Dharmakīrti’s Santānāntarasiddhi: Th. Stscherbatsky (ed.), Saṃtānāntarasiddhi Dharmakīrti. Osnabrück: Biblio Verlag, 1970 (Petrograd, 1916). sns: Saṃdhinirmocanasūtra. É. Lamotte (ed.), Saṃdhinirmocanasūtra. Louvain: Université de Louvain, 1935. Viṃś: Vasubandhu’s Viṃśikā. S. Lévi (ed.), Vijñaptimātratāsiddhi: Deux Traités de Vasubandhu: Viṃśikā et Triṃśikā. Paris: Libraire Ancienne Honoré Champion, 1925. wsesl Shuji: Wei shi er shi lun shu ji 唯 識 二 十 論 述 記 of Kuiji 窺 基 . T. 1834.

Secondary Literature Brewster, Ernest Billings. Forthcoming. “The Problem of Other Minds in Cheng weishi lun《 成 唯 識 論 》 .” Journal of Indian Philosophy. Carruthers, Peter 2011. The Opacity of Mind. Oxford: Oxford University Press. Cook, Francis H. 1999. Three Texts on Consciousness Only. Berkeley: Numata Center for Buddhist Translation and Research. Frauwallner, Erich. 1958. Die Philosophie des Buddhismus. Berlin: Akademie-Velag. Funayama, Toru. 2000. “Two Notes on Dharmapāla and Dharmakīrti.” zinbun 35: pp. 1–11. Hyodo, Kazuo. 2006. Yuishiki to iu koto: Yuishiki niju ron o yomu. Tokyo: Shunjusya. Inami, Masahiro. 2001. “The Problem of Other Minds in the Buddhist Epistemological Tradition.” Journal of Indian Philosophy 29: pp. 465–483. Kano, Kazuo. 2008.“Two Short Glosses on Yogācāra Texts by Vairocanarakṣita: Viṃśikāṭīkāvṛtti and *Dharmadharmatāvibhāgavṛtti.” In Sanskrit Texts from Giuseppe Tucci’s Collection Part i, edited by F. Sferra, pp. 343–380. Roma: Istituto italiano per l’Africa e l’Oriente. Katsura, Shoryu. 1969. “Dharmakīrti ni okeru ‘jiko-ninshiki’ no riron.” Nanto Bukkyo 23: pp. 1–44. Katsura, Shoryu. 1983. “Dharmakīrti ‘Ta-sozoku no sonzai ronsho’: Wayaku to synopsis.” Bulletin of the Faculty of Letters, Hiroshima University 43: pp. 102–120. Kellner, Birgit, and John Taber. 2014. “Studies in Yogācāra-Vijñānavāda idealism i: The interpretation of Vasubandhu’s Viṃśikā.” Asiatische Studien 68: pp. 709–756. Lusthaus, Dan. 2002. Buddhist Phenomenology. London/New York: Routledge Curzon.

Dharmapāla on the Cognition of Other Minds (paracittajñāna)

241

Moriyama, Shinya. 2010. “On Self-Awareness in the Sautrāntika Epistemology.” Journal of Indian Philosophy 38: pp. 261–277. Moriyama, Shinya. 2019. “Dharmapāla.” In Brill’s Encyclopedia of Buddhism. Volume ii: Lives, edited by J. Silk et al., pp. 168–172. Leiden-Boston: Brill. Schmithausen, Lambert. 2015. On the Problem of the External World in the Ch’eng wei shih lun. Electronic Edition with Corrections and Additions. Tokyo: The International Institute for Buddhist Studies of the icpbs. url: https://icabs.repo.nii.ac .jp/?action=pages_view_main&active_action=repository_view_main_item_ detail&item_id=341&item_no=1&page_id=15&block_id=18. Silk, Jonathan. 2016. Materials Towards the Study of Vasubandhu’s Viṃśikā (i). Cambridge MA: Department of South Asian Studies, Harvard University. Tillemans, Tom. 1990. Materials for the Study of Āryadeva, Dharmapāla and Candrakīrti. Vol. i. Wien: Arbeitskreis für Tibetische und Buddhistische Studien Universität Wien. Ui, Hakuju. 1963. “Jō yuishiki hōshōron kenkyū.” In: Daijō butten no kenkyū. Tokyo: Iwanami Shoten. Yamabe, Nobuyuki. 1998. “Self and Other in the Yogācāra Tradition.” In Essays on Japanese Buddhsit Culture: In Honor of Dr. Tensei Kitabatake On the Occasion of His Seventieth Birthday, pp. 15–41. Kyoto: Nagata Bundhodo. Yao, Zhihua. 2005. The Buddhist Theory of Self-Cognition. London & New York: Routledge.

Part 3 Mental Consciousness in East Asian Buddhism: msf

∵

Introduction to Part 3 Three papers in this volume touch upon the notion of mental consciousness simultaneous with the five sensory consciousnesses (henceforth abbreviated as msf). It refers to the function of the mental consciousness (mano-vijñāna; the 6th type of consciousness) that arises simultaneously with a sensory consciousness. The origin of this notion can be traced back at least to the Saṃdhinirmocana-sūtra, which claims that manas consciousness must arise simultaneously with one or more types of sensory consciousness. To the best of our knowledge, this notion never appears in any extant Indian and Tibetan sources. Only some Chinese texts refer to it. What is interesting about this notion is that it squarely contradicts previous conceptions in Buddhist philosophy about how the mental consciousness functions. Both Abhidharma and early Yogācāra claim that mental consciousness can only arise at the next moment after a sensory consciousness has arisen. The main reason for this is that mental consciousness does not have direct access to external objects but can only conceptualize the information that is gleaned from sensory consciousness. Thus mental consciousness is also held to be responsible for our false views about what really exists. For example, it is the mental consciousness that falsely judges “there is a cup” when we merely cognize a variety of tropes through sensory consciousnesses. The mental consciousness is also responsible for “cognizing” a circle where there is merely a whirling torch. In contrast, the notion of msf demands that, in addition to functioning in the above manner, mental consciousness can also function simultaneously with a sensory consciousness. It remains to be explored what are the precise cognitive functions carried out by msf. Moreover, as shown in Keng’s paper, the Yogācāra tradition in East Asia identifies the function of msf with an even more mysterious notion in Dignāga: the notorious notion of mental perception (mānasa-pratyakṣa), i.e., a perceptual cognition that belongs to the mental consciousness. According to preDignāga Buddhist philosophy, it is generally agreed (though Sharf (Chapter 1) claims otherwise) that perception belongs exclusively to sensory consciousness, and so does inference to mental consciousness. But Dignāga’s notion of mental perception suggests a third cognitive function: mental consciousness that functions as perception. Again, it’s not clear at all why Dignāga stipulated such a notion. Almost all post-Dharmakīrti Buddhist philosophers could not make sense of it, and they ended up concluding that the only reason Dignāga did this was out of respect for scriptural authority.

© koninklijke brill nv, leiden, ���1 | doi:10.1163/9789004440913_013

INTRODUCTION TO PART 3

245

As mentioned in the Introduction, Dignāga’s distinction between perception and inference follows from the fact that there are just two sorts of cognizables: pure particulars and generalized objects. Only the former has causal efficacy and hence is really existent (dravyasat). In contrast, the latter has no causal power and is only nominally existent (prajñaptisat). Under this distinction, the notion of mental perception suggests a third category of existents that are particulars but have no causal power. Keng’s paper concerns a specific site of perception-thought interface: perception of shape. While some Ābhidharmikas held that the object of visual consciousness is a single color-and-shape dharma (on the grounds that one never perceives color without shape), others disagreed. (They might have had in mind what one sees when a clear blue sky fills the field of vision.) They claimed instead that shape is conceptually constructed and only color particulars are ultimately real. Dignāga agreed with the latter, but felt compelled to make room for the intuition that shape is perceived and not merely inferred. The result, according to the Chinese commentator Kuiji, is the doctrine of “mental perception” (perception by manas), as something both direct and immediate like visual consciousness, but also involving conceptual mediation. And since subjects typically report that they are aware of color and shape simultaneously, this mental perception is said to be simultaneous with external sense-consciousness – hence msf, and more work for a modularity account of cognition. This puts some strain on Dignāga’s claim that perception and inference have distinct objects. The difficulty is that something not ultimately real is said to be an object of perception, when perceptual consciousness is understood to be caused by real entities. Yao’s essay also takes up the question of mental consciousness, raising the general question how manas, as a conceptual faculty, can be said to perceive external objects directly. His focus is not only Dignāga but also earlier Yogācāra. He starts with the point that it should not be possible for a mental consciousness that employs concepts to arise simultaneously with the perceptual content that it takes as input. He then investigates a variety of approaches to be found in early Yogācāra, Dignāga, and Chinese commentarial literature. One such approach reverts to the older Abhidharma idea that the basic categorization performed by perceptual identification is not conceptual in an ontologically problematic way. (This approach would not work for a radical nominalist like Dignāga.) Here again, msf and the parallel processing it implies make an appearance. Yao concludes, though, that no consensus was ever reached on the issue. Yao also suggests that the idea that msf has direct access to external objects explains why some Yogācāra texts do not subscribe to Sautrāntika’s

246

INTRODUCTION TO PART 3

i­ ndirect or representationalist realism, which insists that perceptual cognition functions by containing an image (ākāra) of what is cognized. Lin’s essay discusses the account of memory deployed by Vasubandhu in response to the sorts of objections to non-self posed by self-theorists. Because memory, being of an object that no longer exists, is not counted as valid cognition, mental consciousness is also held to be involved in memory. Here again the Chinese commentator Kuiji and his msf theory make an appearance. On Lin’s reading, the puzzle of manas operating as both conceptual and non-­ conceptual is resolved by distinguishing between msf, in which manas functions simultaneously with an external sense faculty, and diachronic manas, in which the faculty operates on past objects of cognition. In the latter function, manas is conceptual in nature, but in the former it merely produces conceptual output in the succeeding moment (just as the smoke detector proceeds from specific stimulus to stereotyped response). The claim would then be that the immediacy of perceptual consciousness is not threatened by the participation of manas insofar as it operates on the object simultaneously with the external sense faculty. But as Yao points out, this sort of solution requires an idealist metaphysics: only the Buddhist idealist holds that the object of perception exists simultaneously with the consciousness that takes it as intentional object. The above discussion may suggest that manas functions on a track parallel to that of the external sense faculties, processing at the level of conventionally real objects at the same time that vision and the rest process at the level of the ultimate reals. If so, this would represent a solution to the puzzle of bordercrossing – the puzzle of how transitions between the non-conceptual and the conceptual are effected – quite different from Dharmakīrti’s, and one that might align with the levels approach of Dharmottara that was discussed in the Introduction. Partly due to its absence from the Indian and and Tibetan sources, and partly due to its own density, msf remains understudied. There is much more to explore. For example, the role played by msf in meta-cognition remains to be examined. We may wonder: Is it the visual consciousness itself that is selfaware of its cognition, or is it the msf that makes the content of a visual consciousness globally available? Outside of the Buddhist context, contemporary cognitive science suggests that it would be impossible to cognize color and shape at the same time. The two types of cognition need to be computed successively, which takes time. This would pose serious challenges to the notion of msf.

Chapter 9

Mānasa-pratyakṣa as the Perception of Conventionally Real (prajñaptisat) Properties – Interpreting Dignāga’s mānasa-pratyakṣa Based on Clues from Kuiji Ching Keng Abstract In his writing, Kuiji, (632–682), the foremost disciple of Xuanzang (602–664) interprets the notion of mental perception (mānasa-pratyakṣa) of Dignāga (ca. 480–540) as a function of the mental consciousness that arises simultaneously with a sensory consciousness. For this reason, it is named as “the mental consciousness arising simultaneously together with the five [sensory consciousnesses] (MSF).” To define this as a kind of perception would raise a few issues: Why do we need such a notion? Would it lead to the absurd consequence that even a person born blind can still see? If not, then what is the relation between such consciousness and a simultaneous sensory consciousness? If this qualifies as a kind of perception, then it must be without conceptualization. But what kind of role does it play in cognition as a non-conceptualizing mental perception? Why is it necessary? What is its conscious content? How does the content of a sensory consciousness get transmitted or shared with a MSF? And how do its conscious content and function differ from sensory consciousnesses on the one hand and from conceptualizing mental consciousness on the other? Does it function like a sensus communis so as to accommodate all kinds of sensory content such as visual, auditory and olfactory? This paper seeks to answer the above questions and depict a picture of such cognition based on the works of Kuiji. With this as a clue, we may have a better understanding of Dignāga’s notion of “mental perception.”

Keywords Dignāga – mental perception (mānasa-pratyakṣa) – mental consciousness ­simultaneous with the five sensory consciousnesses – Kuiji – ­ultimately real (dravyasat), conventionally real (prajñaptisat) – shape perception

…

© koninklijke brill nv, leiden, ���1 | doi:10.1163/9789004440913_014

248 1

Keng

Introduction: Dignāga’s Notion of Mental Perception

How does the mind work to cognize an object? How do the sense faculties and the intellect work together to achieve successful cognition? Is it true that all cognition is conceptual, or is some initial moment of cognition non-­ conceptual? These are the sorts of questions that draw our interest to the study of consciousness or cognitive neuroscience, which arguably has become one of the most fascinating research topics of the 21st century. When studying the intriguing phenomena connected with consciousness, using the most advanced equipment, under ever-evolving hypotheses, it might be useful to compare our current understanding against a distant and different tradition, in order to shed light on possible hidden presuppositions we entertain. Among the ancient traditions, Buddhism is widely regarded as having offered a quite detailed map of how the mind works. It is common to both Western and Eastern traditions to divide cognition into a relatively passive and receiving aspect (sense perception) and a relatively active and processing aspect (intellect). Buddhism accordingly divides cognition into two aspects: the five types of sensory consciousness (visual, auditory, olfactory, gustatory, tactile); and mental consciousness, which is responsible for complex functions such as conceptualization, memory and imagination. Under this division, the Buddhist tradition famously holds a two-step theory of sensory cognition: First, there occurs a step of sensory consciousness, which perceives via direct perception; then there occurs a step of mental consciousness, which processes the information gathered, for example, by considering “What is this?” via comparison and conceptualization, or by making a verbal judgment such as “This is a cup.” The first step belongs to “pure perception,” which is free from conceptualization, but in the second step, mental consciousness conceptualizes objects and gives names. A major figure promoting such a two-step model is Dignāga (ca. 480–540 ce), a Buddhist philosopher who had a huge impact on Buddhist logic and epistemology. In his monumental masterpiece of epistemology entitled A Compendium of Valid Sources of Knowledge (Pramāṇasamuccaya, henceforth ps), Dignāga distinguishes two valid sources of knowledge (pramāṇa), viz., direct perception (pratyakṣa) and inference (anumāna) (ps verse i.2). According to Dignāga, the major difference between these two is that the former is “devoid of conceptualization” (kalpanāpoḍha) (ps verse i.3) but the latter is conceptualized. Under this model, our everyday experience such as “I am seeing a blue cup” is regarded as false because it is conceptualized. There is no cup, because real things are unique particulars, distinct from each other, despite their looking similar and playing a similar role in our experience as “containing liquid.”

Mānasa-pratyakṣa as Perception

249

There is no blue, because the blueness in this particular thing is definitely not identical with the blueness in that particular thing. Finally, there is no “I” either. In cognition, there is merely the subjective aspect that is constantly changing. The idea that there is a constant and sustaining self that remains the same throughout all experience is also the result of conceptualization. So the falsehood of our perception lies in that we apply universals (sāmānyalakṣaṇa) such as the concepts of “cup,” “blue,” “I” onto particulars (svalakṣaṇa), which alone are ultimately real (dravya-sat). Ordinary experience is false, or better, is merely conventionally true, because the universals that are superimposed upon particulars are merely conventionally real (prajñapti-sat). It is the mental consciousness (manovijñāna) that conceptualizes or superimposes universals onto the particulars that are perceived merely by sensory consciousnesses (indriya-vijñāna). Under Dignāga’s philosophical scheme, there are dichotomies in both epistemology and ontology. In epistemology, there is the dichotomy between direct perception and inference: the former produces cognitions that are ultimately true but the latter yields mere conventional truths. In ontology, there is the dichotomy between particulars and universals: the former are ultimately real but the latter are merely conventionally real. We may ask several questions about Dignāga’s scheme such as: from the perspective of ontology, is it true that our world consists only of particulars and every universal is unreal? From the perspective of epistemology, is it true that what we perceive through sense perception is always real, and hence sense perception always has epistemological primacy or superiority over inference or reasoning? The dualistic distinction is probably not so clear-cut.1 Actually, the dualistic framework does not seem to be the whole story in Dignāga, either, as he further divides direct perception into four types:2 (1) sensory perception (indriya-pratyakṣa); (2) mental perception (mānasapratyakṣa); (3) self-awareness (svasaṃvitti or svasaṃvedana) and (4) yogic perception (yogi-pratyakṣa). Among these, the first and the last are accepted without much dispute: The first refers to the perception carried out via sense organs by an ordinary sentient being and the last refers to the perception in the meditative state of a yogi, an established meditator. In contrast, there is

1 Sharf, Chapter 1, charges that the boundary between these two aspects can never be satisfactorily drawn. 2 As detailed below, there are debates among scholars about whether there are three or four types of direct perception (pratyakṣa) according to ps. This paper disagrees with the reading by Eli Franco, who claims that there are only three types of perception. The reasons are stated below.

250

Keng

much dispute regarding (2) and (3).3 In particular, the notion of mental perception is one of the most mysterious parts of the ps. As Vincent Eltschinger (2010, 410) observes: “Mental perception’s nature and function in the Buddhist epistemological school remain obscure, but it is quite often presented (at least from Dharmottara on) as an acknowledgement of scriptural statements. Mādhyamikas argue, however, that mental perception, by perceiving a subsequent phase of a real entity on the basis of the immediately preceding sensory perception, bridges the gap between perception and conception.”4 This paper aims to provide help for understanding Dignāga’s notion of mental perception by drawing clues from Kuiji 窺基 (632–682 ce), aka Master Ji 基法師, the foremost disciple of the renowned Chinese pilgrim Xuanzang 玄 奘 (602–664). Faced with the seemingly irresolvable difficulties surrounding the notion of mental perception, it is a pity that most scholars do not pay more attention to Kuiji’s report in his Dacheng fayuan yilin zhang 大乘法苑義林章 (T1861), where he cites an opinion that the perceptual function of the “mental consciousness simultaneous with the five sensory consciousnesses” (wu ju yishi 五俱意識; henceforth abbreviated as msf) is considered identical with Dignāga’s notion of mental perception. (T45:1861.258a8–11) This shows that at least according to Kuiji and the opinion he cites, the mental perception stipulated by Dignāga is the function of msf discussed in the Cheng weishi lun 成唯識論 (T1585, henceforth cwsl). As shown below, this attribution of mental perception to msf would provide very useful information about mental perception. The key provided by Kuiji for resolving the mysteries surrounding mental perception is the distinction between ultimately real properties such as colors (varṇa) and conventionally real properties such as shapes (saṃsthāna). By “shape” here I do not mean, for example, the subject-dependent looking-round of a round plate, but the subject-independent being-round of a round plate. The issue here is whether shapes are ultimately real or merely conventionally

3 Regarding the disputes surrounding (3), see Funayama (Chapter 7), Coseru (Chapter 4), and Arnold (Chapter 5). 4 For two recent interpretations, see Park 1999 and Arnold 2012, 165ff. For a brief overview of the interpretations provided by the Mādhyamikas in India and Tibet, see Tillemans 1989. Tillemans’ article is not useful for my purpose because the Prāsaṅgika positions he reviewed, for example the view provided by Jam dbyangs bzhad pa (1648–1722), consider mental perception (mānasapratyakṣa) to be conceptual (Tillemans 1989, 72ff.). These Prāsaṅgika positions, in my opinion, clearly deviate from Dignāga’s view that mental perception is by definition non-conceptual, and hence cannot be regarded as faithful interpretations of Dignāga’s own view.

Mānasa-pratyakṣa as Perception

251

real.5 This is a debate that can be traced at least to the Abhidharmakośa (Verse 4.3b), where in the auto-commentary Vasubandhu reports that shapes are considered ultimately real for Sarvāstivāda but merely conventionally real for Sautrāntika.6 Dignāga’s position in the Ālambanaparīkṣā (henceforth āp) agrees with the Sautrāntika. In his āp Verse 5, while challenging his third opponent, Dignāga claims that shapes (such as the form (ākāra) of the neck of a pot, a cup, etc.) are not ultimately real because once we analyze a cognitive object of a certain shape all the way to the level of atoms that alone are ultimately real, the perception of that shape ceases. Hence shapes must be merely conventionally real properties, unlike colors that are regarded as essential parts of atoms.7 But then the problem here is: While shapes are not ultimately real, shapes are particulars. Unlike the situation where one applies the same concept “plate” to two collection of atoms at different places, one cannot apply the same “being round” to those two different collections. Each collection of atoms is “being round” in its unique way and its unique “being roundness” cannot be reproduced in a different collection at a different place. Thus, shapes are to be cognized through perception, which alone cognizes particulars. When we say “this is round” and “that is round,” we are actually applying the concept “roundness” to those two different collections of atoms at different places. Thus, shape poses a challenge to the dichotomy mentioned above between being particular and ultimately real on the one hand, and being universal and conventionally real on the other. According to Dignāga, the former is cognized through sense perception; the latter through conceptualized mental consciousness. But shapes belong to a third category: They are particulars but are merely conventionally real. And this is precisely why Dignāga needs to stipulate the category of mental perception carried out by msf to account for our ordinary experience of simultaneously perceiving the color and the shape of an object. In this paper, I investigate the main features of msf as depicted in the cwsl and in Kuiji’s works in order to shed new light on Dignāga’s notion of mental 5 I believe that it might be easier to defend Dignāga’s idea that shapes are not ultimately real if we interpret shapes as referring to the subject-dependent looking-round. Another way to go is to interpret shapes as involving edges, i.e., the boundaries between the object and its background. In this case, edges become parabhāva (existing dependent upon others), and hence not ultimately real. Many thanks to Mark Siderits for suggesting this alternative. But here I choose not to pursue these two lines, mainly because Dignāga himself never takes these paths. 6 Pradhan 1967, 194.14–196.2. For Xuanzang’s Chinese translation, see T29: 1558.68b1–c25. 7 See Tola and Dragonetti 2004, 35–36; for the Tibetan translation, see Tola and Dragonetti 2004, 30–31; for Xuanzang’s Chinese translation, see T31:1624.888c09–14.

252

Keng

perception. Before taking a closer look at those texts, let’s briefly review why mental perception has so far proven inextricably difficult to clarify. 2

Why Does Mental Perception Remain Difficult?

Before I proceed, a clarification of the term “mental perception” is in order. The idea of mental perception appears in Dignāga’s ps i.6, which reads: mānasaṃ cārtharāgādisvasaṃvittir akalpikā || ps i.6ab|| Regarding this verse, there have been debates among scholars about whether Dignāga endorses three or four types of direct perception.8 According to the three-types reading, an English translation of the verse would read something like: “And there is mental [perception] that is without vikalpa, namely, the selfawareness of objects (artha) and greed, etc. (rāgādi)” According to the fourtypes reading, an English translation of the verse would read something like: “And there is mental [perception] that is without vikalpa, namely, the cognition (saṃvitti) of objects (artha) and the self-awareness (svasaṃvitti) of greed, etc. (rāgādi)” I think “three versus four” is somewhat misleading, depending as it does on whether one treats self-awareness (svasaṃvitti) as a sub-category or as something parallel to mental perception. No matter which way one chooses, the real issue still remains: Is mental perception simply a form of internal or introspective cognition, which concerns only internal mental states (internal mental representations, desires, etc.), or is mental perception also about external objects? This concerns the interpretation of the term artha in the compound artharāgādisvasaṃvittir. According to the three-types reading, artha refers to internal objects (such as the mental image of an object); but according to the four-types reading, artha refers to external objects. For the reasons stated below, I agree with the four-types reading and hence agree with Hattori’s English translation of ps i.6, which reads: “there is also mental [perception, which is of two kinds:] awareness of an [external] object and self-awareness of [such subordinate mental activities as] desire and the like, [both of which are] free from conceptual construction.” (Hattori 1968, 27) For this reason, it helps to distinguish between a broad sense of mental perception versus one of its narrower senses. The broad sense refers to the mental perception of both external and internal objects, but one narrow sense refers only to that of external objects. The broad sense hence includes as its sub-­ categories both self-awareness (svasaṃvitti) and mental perception in its 8 Including Franco 1986, Wayman 1991, Franco 1993, Yao 2004.

Mānasa-pratyakṣa as Perception

253

narrower sense. In this chapter, unless otherwise noted, the term “mental perception” is used in its narrower sense to avoid confusion. The obscurity about mental perception remains mainly because there are some irresolvable difficulties regarding its essence. As early as the 1960s, Hattori raised two challenges to Dignāga’s mental perception as follows: Two problems respecting mental perception of objects are: (1) If the mind perceives the same object that had been perceived by the immediately preceding sense, this mental perception could not be recognized as pramāṇa, because pramāṇa is defined as anadhigatārtha-gantṛ; see above, n. 1.20. (2) If, on the other hand, the object of the mental perception were absolutely different from that of the sense perception, then even blind and deaf persons would be able to perceive color and sound, for their minds are not defective like their senses. (Hattori 1968, 93) Here the Sanskrit term anadhigatārtha-gantṛ literally means “arriving at an unknown object” and hence Hattori translates it as “make known an unknown object” (Hattori 1968, 74, footnote 1.3) According to Hattori, faced with these difficulties in Dignāga, Dharmakīrti (600–660 ce) suggests the following solution: Dharmakīrti solves these difficulties in the following way: (1) What is perceived by means of mental perception is the object in the moment that immediately follows the moment of sense perception. Therefore mental perception is held to be anadhigatārtha-gantṛ. (2) Mental perception is conditioned by the immediately preceding sense perception as its samanantara-pratyaya. Accordingly, blind and deaf persons who have no sense perception are unable to have mental perception. (Hattori 1968, 93) (my emphasis)9 Dharmakīrti’s solution under Hattori’s interpretation can be put into the following diagram (where obj.1 means external object at moment t1) t1 t2 t3 obj.1 obj.2 obj.3 Sense perception (of obj.1)  Mental perception (of obj.2)

9 Also Cf. Tosaki 1979, 343ff. For Dharmakīrti’s interpretation in his Nyāyabindu, which I think is the same as what is meant here, see Stcherbatsky 1962, 26ff.

254

Keng

In my view, Dharmakīrti’s solution at least under Hattori’s interpretation hardly makes sense because the above (1) and (2) are incompatible. Given that every object is momentary, if, following (2), mental perception has no content that has not already been perceived by the sense perception at the previous moment, then given that the content of sense perception at t2 is about obj.1 (the object at t1), then the mental perception at t3 cannot cognize obj.2 (the object at t2), which has not been perceived by sense perception at t2. So the content of mental perception at t3, if anything at all, must be about obj.1, either being exactly the same as the content of sense perception or a modification of it. The same criticism can be voiced against the interpretation of Jinendrabuddhi (circa 710–70), which is the same as Dharmakīrti’s view.10 A few years after Hattori gave his interpretation, Masatoshi Nagatomi (1979) tried to resolve the conundrum differently, by the reinterpretation that the two types of mental perception (in the broad sense of the term) mentioned in ps i.6ab do not refer to two kinds of perception but to two aspects of the same mental perception.11 According to Nagatomi, the mental perception at t3 has two aspects: The first is the object-cognizing aspect, having the appearance of an object, and being free from conceptual construction; the second is the self-cognizing aspect, having the appearance of passion etc., and being free from conceptual construction.12 I am not at all convinced by Nagatomi for two reasons: (i) Basically, Nagatomi adopts the Yogācāra model of multiple aspects (ākāra or bhāga) in a cognition to account for the relation among sense perception, mental perception and self-awareness (svasaṃvitti). Under his ­construal, the grasped-aspect (grāhya-ākāra) in sense perception and the ākāra of selfawareness become the grasped-grasping (grāhya-grāhaka) aspects in a given mental perception. But I think to take the ākāra of self-awareness (passion, etc.) as the grāhaka-ākāra is quite misleading because, according to Dignāga’s distinction in ps i.8cd-9 among the object to be cognized (prameya), the means of cognition (pramāṇa) and the resulting cognition (pramāṇa-phala), self-awareness corresponds to pramāṇa-phala whereas the grāhaka-bhāga corresponds to pramāṇa. The sense perception at t2 and the mental perception at t3 should each be accompanied by individual episodes of self-awareness. (ii) Moreover, since the grāhya-ākāra in the mental cognition is exactly the same as the grāhya-ākāra in the sense perception at t2 under Nagatomi’s interpretation, he still does not explain how this mental perception, while providing no new information, could be considered as a means of cognition (pramāṇa).

10 See Hattori 1968, 93, footnote 1.46. Also cf. Yoshida 2012, 101ff., Shi 2015. 11 Nagatomi 1979, 254ff. in particular. 12 Ibid., 255.

Mānasa-pratyakṣa as Perception

255

So despite the efforts by Hattori and Nagatomi, we are still left with a conundrum. Between the above two difficulties identified by Hattori, (2) is easier to solve as long as we admit that mental perception must have sense perception as a pre-condition. In this way, we can avoid the unwanted consequence that even a person born blind can have knowledge about external objects through mental perception. But (1) is much more difficult. Namely, if mental perception has content that is different from that of sense perception, then it should be modified, i.e., conceptualized, and hence would not qualify as direct perception; but if mental perception has content that is the same as that of sense perception, then it should not be regarded as a means of cognition at all since it does not provide any new information. Thus, the difficulty is how we are to understand mental perception in such a way that its content is not the same as that of sense perception but is still non-conceptualized. With this difficulty in mind, let’s move to Kuiji’s discussion. 3 Kuiji: Mental Perception Is Carried Out by msf To begin, I cite passages where Kuiji explicitly links the function of msf to mental perception in the ps. In his Commentary on the Yogācārabhūmi (T1829), Kuiji asserts that “As claimed in the ps and in the Nyāyamukha, ‘[The cognition of] the five sensory consciousnesses merely belongs to direct perception; the same thing holds for the mental consciousness simultaneous [with the five sensory consciousnesses.]’” (T43:1829.198c6–7) This passage claims that msf can function in the manner of direct perception. Such direct perception, I argue below, should be regarded as the same thing as mental perception in the ps. In a more extended passage from his Dacheng fayuan yilin zhang (T1861), Kuiji cites two opinions about whether the five stages of minds (wuxin 五心)13 belonging to msf14 operate in the manner of direct perception alone (but not 13

14

The five stages of mind refer to the five steps of mental activities. Taken together, these five stages complete the process of a cognition. These are: the sudden (aupanipātika) mind, the searching (paryeṣaka) mind, the settled (niścita) mind, the defilement or purification (saṃkleśo vyavadānaṃ ca) mind, and the uniform-flowing (naiṣyandika) mind. See Bhattacharya 1957, 10, lines 2–5. The idea is that the mind suddenly becomes aware of an object, then it searches and analyzes it until it makes a decision about it. Due to this decision, defilement or purification of the mind then arises. Finally, the mind continues bearing defilement or purification. A main issue with these five stages of mind is which among these five belong to the sense consciousnesses and which belong to the mental consciousness. For more details, see Kramer (2016), Keng (2018) and the next footnote. Note that, according to Kuiji, each of the six consciousnesses can be analyzed into five stages of mind. (T45:1861.256b9–10) This is quite different from the Yogācārabhūmi, which

256

Keng

to inference (anumāna) nor to invalid means of cognition (apramāṇa)). The first opinion, by interpreting Dignāga’s ps as claiming that msf operates in the manner of direct perception, claims that all the five stages of mind belonging to msf operate in the manner of direct perception. In contrast, the second opinion disagrees by saying that Dignāga does not assert that msf operates merely in the manner of direct perception. According to this opinion, only the first and the fifth stages of the five minds belonging to msf must operate in the manner of direct perception. The rest, namely the second, the third and the fourth can also operate in the manner of inference (biliang 比量; anumāna) and invalid means of cognition (feiliang非量; apramāṇa). (T45:1861.258a8–17) Despite their difference, both opinions agree that ps claims that msf can operate in the manner of direct perception. Given that mental perception by definition belongs to the mental consciousness, based on Kuiji’s discussion here, we may infer that the kind of direct perception carried out by msf is precisely what Dignāga meant by mental perception. Further, an essential characteristic of msf is of course that it arises simultaneously with the five sensory consciousnesses. In Fascicle Four of the cwsl, extended discussion about msf appears under the discussion of simultaneous basis (juyou yi 俱有依; sahabhū āśraya15). There three kinds of bases are identified, with the second one being the basis qua contributory condition (zengshang yuan yi 增上緣依; *adhipatipratyaya-āśraya), also named as the simultaneous basis. Regarding this simultaneous basis, there are three positions held respectively by Nānda (6th century ce), by Sthiramati (7th century ce) and by Dharmapāla (sixth century ce). Among these three, the particularly relevant ones are the second and the third positions held by Sthiramati and Dharmapāla. Sthiramati holds that the five sensory consciousnesses have two simultaneous bases: sense organs and a simultaneous mental consciousness. The mental consciousness always has the seventh consciousness as its simultaneous basis, but if the mental consciousness arises simultaneously with the five sensory consciousnesses, then that mental consciousness also has the five sensory consciousnesses as its simultaneous basis. (T31:1585.20b3–16) In contrast, Dharmapāla disagrees, claiming that although the mental consciousness can arise simultaneously with up to five sensory consciousnesses and hence should have the five as its simultaneous basis, this does not always happen. Namely, there are cases where mental consciousnesses do not arise simultaneously with the five. Hence, the mental

15

claims that among the five stages of mind, the first and the fifth belong to sense consciousness, the second, third and fourth belong to mental consciousness. The Sanskrit terms “sahabhūr āśrayaḥ” are attested in Bhattacarya 1957, 4, line 6.

Mānasa-pratyakṣa as Perception

257

consciousness merely has the seventh and the eighth consciousnesses as its simultaneous basis. (T31:1585.20c12–19) Despite the differences between Sthiramati and Dharmapāla, it is striking that both endorse the existence of msf, namely, a mental consciousness simultaneous with the five sensory consciousnesses. The idea that msf has five sensory consciousnesses as its simultaneous basis helps answer the above-mentioned challenge of “a person born blind can perceive external objects.” Namely, the arising of a sensory consciousness of external objects is a pre-condition for msf to have any access to external objects. For a person born blind, since that person’s eye-consciousness never arises, the mental consciousness of that person cannot have access to external objects. 4

Main Features of msf

Having established that mental perception is the direct perception carried out by msf, we now need to investigate how exactly msf operates. The main section in the cwsl that discusses the msf appears in Fascicle Seven, which reads as follows: [Text 1] If one does not allow that mental consciousness (i.e., msf) arises simultaneously with [one or more of the] five [sensory consciousnesses], then the grasping (qu 取; Skt. √grah) [by msf] of that [same]16 object (suoyuan 所緣) of those [sensory consciousnesses] should not be vivid (mingliao 明了), because it would be like an unconcentrated mental consciousness grasping an object that has long ceased.17 [Question] How is it that simultaneous with [one or more of] the five [sensory consciousnesses] there is only one mental consciousness? Regarding cognitive objects (jing 境; viṣaya), matter, etc. does it (msf) cognize one or many? [Answer] This is like the situation where [sensory] consciousnesses, eye-consciousness, etc. cognize each of its own (i.e., proper) objects, no matter one or many. What would be the fault with this [claim of mine]?

16 17

Kuiji explains that this means that msf has the same cognitive object (ālambanapratyaya) as the five sensory consciousnesses. See cwslsj, T43:1830.485c6–8. That is, memory of an object in the past.

258

Keng

This is because both the image[-part] (xiang 相; nimitta-[bhāga]) and the seeing[-part] (jian 見; dṛṣṭi-[bhāga]) have many kinds of forms. [Question] Why is it that [one of the] sensory consciousnesses is not simultaneous with its own kind? [Answer] Regarding its own proper objects, if [that one sensory consciousness] can know the object, then just one [of that kind of consciousness] can know, the rest of [the same kind of consciousnesses that are simultaneous] would be useless. [Question] If so, then [one or more of the] five consciousnesses has already known each of their individual objects, why do we still need the msf? [Answer] The msf assists the five to arise, and hence [its function] is not merely limited to knowing the objects of the five consciousnesses. Moreover, [the msf] can vividly grasp (mingliao qu 明了取) the objects of those [five]. This function is different from that of [sensory] consciousnesses, eye-consciousness, etc., and hence [msf] is not useless. From this the scriptures claim that that msf is named as “with vikalpa (savikalpa),”18 which is not the case with five [sensory] consciousnesses.19 Based on this passage, we learn that there are three features of msf: (A) having the same cognitive objects as the five consciousnesses; (B) assisting the arising of the five consciousnesses; (C) grasping vividly the cognitive objects of 18 “With vikalpa” (sa-vikalpa) here should be understood as referring to the words of the Saṃdhinirmocanasūtra translated by Xuanzang (T16:676.692b18–28). Interestingly, the Chinese translation of the same passage by Bodhiruci (?–527 ce) describes the sixth consciousness simultaneous with the five sensory consciousnesses as “without vikalpa” (nirvikalpa). The Tibetan translation, in contrast, merely has “the sixth consciousness” (yid kyi rnam par shes pa) but says nothing about whether it is with or without vikalpa (Lamotte, 1935, 56, Section v.4). As detailed below, I argue that the function of mental perception is to grasp the shape of an object qua a particular (svalakṣaṇa) via direct perception (pratyakṣa). But since shapes are not ultimately real, a cognition of shapes necessarily involves vikalpa. Note that vikalpa here does not mean “conceptualization” (in the sense of forming a concept out of the perception of particulars) but refers to whatever cognitive function of the mental consciousness that superimposes upon its cognitive object anything that is not ultimately real. In this volume, vikalpa is in some contexts translated as “conceptualization” (such as Spackman (Chapter 2) and Tzohar (Chapter 3)) but in other contexts translated as “discrimination” (as in Sharf (Chapter 1)). Given the reason mentioned above, I think vikalpa should not be understood as conceptualization in this context. For a more detailed discussion of the Sanskrit term vikalpa, see Keng (2019). 19 cwsl, Fascicle 7: 又 若 不 許 意 與 五 俱 ， 取 彼 所 緣 應 不 明 了 ， 如 散 意 識 緣 久 滅 故 。 [問 ]如 何 五 俱 唯 一 意 識 ， 於 色 等 境 取 一 或 多 ?[答 ]如 眼 等 識 各 於 自 境 取 一 或 多 ， 此 亦 何 失 ?相 、 見 俱 有 種 種 相 故 。 [問 ]何 故 諸 識 同 類 不 俱 ?[答 ]於 自 所 緣 若 可 了 者 ， 一 已 能 了 ， 餘 無 用 故 。 [問 ]若 爾 ， 五 識 已 了 自 境 ， 何 用 俱 起 意 識 了 為 ?[答 ]五 俱 意 識 助 五 令 起 ， 非 專 為 了 五 識 所 緣。又於彼所緣能明了取，異於眼等識，故非無用。由此，聖教說彼 意 識 名 有 分 別 ， 五 識 不 爾 。 (T31:1585.38b21–c2).

Mānasa-pratyakṣa as Perception

259

the five consciousnesses. In the sections that follow, I discuss each of these three features. Feature (A): Having the Same Objects as the Five Sensory Consciousnesses In addition to the above passage in Fascicle Seven, there are passages in Fascicle Five of the cwsl that also claim that msf has the same objects as the five sensory consciousnesses. The first passage discusses the nature of consciousnesses, whether being wholesome (kuśala; 善), unwholesome (akuśala; 不善) or indeterminate (avyākṛta; 無記). There it is asserted that “Several passages simply claim that msf also has five kinds of objects as its objects, but does not claim that the nature [of msf and five consciousnesses] is the same.” (T31:1585.26b24–c11) This passage affirms that msf has the same objects as the five consciousnesses. From the same fascicle, another passage discusses whether the six consciousnesses can be accompanied with three kinds of feeling (pain (duḥkha), pleasure (sukha) and neither-pain-nor-pleasure (aduḥkhāsukha)); at the same time, it is affirmed that “since msf has the same objects as the five consciousnesses, if the five can have three feelings at the same time, then it would also be the case with msf. But this would be illogical and hence all six consciousnesses cannot be accompanied by three feelings at the same time.” (T31:1585.27c17–21) The above two passages clearly affirm that msf has the same objects as the five consciousnesses. But then of course the obvious question is: Why do we need msf if the five consciousnesses have already cognized the objects? To answer this question, cwsl refers us to the other two features of msf. 4.1

4.2 Feature (B:) Assisting the Arising of Five Sensory Consciousnesses Regarding the idea that the msf assists the arising of five sensory consciousnesses, it is noteworthy that this must not be confused with the basis qua the proximate condition (dengwujianyuan yi 等無間緣依; samanantara-pratyayaāśraya) or the basis that makes way and leads to rise [of the succeeding cognition] (kaidao yi 開導依; *avakāśadānāśraya).20 The msf is a kind of mental consciousness that arises together with – i.e., at the same moment with – five sensory consciousnesses and has the same cognitive object with those sensory consciousnesses. The latter, however, refers to the kind of consciousness, whether sensory or mental, that serves as a causal basis for the arising of another consciousness at a subsequent moment.

20

For the Sanskrit reconstruction, see Chu 2014, 272ff. Chu translates kaido as “making way and leading to arise.”

260

Keng

This function of msf in the respect of assisting the arising of the five sensory consciousnesses is somewhat mysterious. So far the only clue I have found is the idea in one passage of the cwsl that msf serves as one of the conditions for the arising of sensory consciousnesses. This passage appears when the cwsl comments on Triṃśikā verse 15b. The whole verse 15 reads: In the fundamental Consciousness, the five consciousnesses are produced according to their condition, Either together or not, like the waves on the water.21 The cwsl comments by saying that the five sensory consciousnesses appear by succeeding conditions. By “conditions” is meant: intentionality (zuoyi 作意; manaskāra), organs (gen 根; indriya) and objects. On this sentence, Kuiji’s cwslsj comments: [Text 2] Eye-consciousness arises based on nine conditions, eye-organ, etc. Namely, [the first five conditions are:] space, light, sense organ, sense objects, and intentionality (zuoyi 作意; manaskāra).22 These five [conditions] are also admitted by the Hīnayāna. If [according to the Mahāyāna, we] add the eighth consciousness, i.e., the fundamental consciousness, the seventh consciousness that is either defiled or pure, the simultaneous (ju 俱) sixth consciousness that is with vikalpa (sa-vikalpa),23 and the seeds that can produce [effects].24 [Eye-consciousness] arises with these nine as the bases. In the case of divine eyes, [out of nine] are excluded light and space. Ear-consciousness [arises] based on eight [conditions] excluding light. The three consciousnesses, nose-consciousness, tongueconsciousness, etc. [arise] based on seven, further excluding space,25 because they grasp [objects] only after having reached those objects.26 21

22 23 24 25 26

English translation by Leslie Kawamura. See Kawamura 1964, p. 93. The Sanskrit text reads: “pañcānāṃ mūlavijñāne yathāpratyayam udbhavaḥ | vijñānānāṃ saha na vā taraṅgāṇāṃ yathā jale.” See Buescher 2007, *33. Xuanzang’s Chinese translation reads: 依 止 根 本 識 五 識 隨 緣 現 或 俱 或 不 俱 如 濤 波 依 水 (T31:1586.60c5–6). This must be the intentionality as an all-pervading (sarvatraga) dharma belonging to the consciousness at the previous moment, which serves as a similar and immediately antecedent condition (samanantara-pratyaya) for the currently arising eye-consciousness. For the meaning of “with vikalpa” here, see above footnote 18. This should refer to the seeds in the fundamental consciousness (ālayavijñāna). That is, in addition to light. cwslsj, Fascicle 7: 眼 識 依 肉 眼 ， 具 九 緣 生 。 謂 空 、 明 、 根 、 境 、 作 意 ， 五同小乘。若加根本第八、染淨第七、分別俱六、能生種子九依而

Mānasa-pratyakṣa as Perception

261

This means that the eighth consciousness, the seventh consciousness, the msf and the seeds also serve as the conditions for the arising of five sensory consciousnesses. Judging from Kuiji’s words, all these nine conditions should be taken as necessary conditions. To conclude this section, the feature of the msf as “assisting the arising of the five consciousnesses” means that, together with other conditions, it serves as a necessary condition for the arising of five sensory consciousnesses. So far we have reviewed two features of msf. Combining these two, however, we still cannot adequately explain why the function of msf must belong to direct perception. Feature (A) merely suggests that the content of msf is the same as that of sensory consciousnesses. Feature (B) has little to do with the issue of whether msf belongs to direct perception. This leads us to the third, the most important feature of msf. Feature (C): Grasping Vividly the Objects of the Five Sensory Consciousnesses The last and the most important feature of msf is: vividly grasping the objects of the five sensory consciousnesses. The Chinese term translated as “vividly grasping” is “mingliao qu 明了取,” which also appears in Xuanzang’s translation of the Abhidharmasamuccaya. Based on that, the Sanskrit term behind “vividly grasping” could be “spaṣṭa-grahaṇa.”27 In his cwslsj, Kuiji discusses extensively the notion of “vividly grasping.” This refers to the function of msf that cognizes conventionally real properties of an ultimately real object of sensory consciousnesses, such as grasping the conventionally real properties of length while grasping the ultimately real properties of “blueness or yellowness” in an ultimately real material object. As Kuiji says 4.3

27

生。若天眼唯除明、空。耳識依八除明。鼻、舌等三依七。復除空。 以 至 境 方 取 故 。 (T43:1830.476a3–7). For Xuanzang’s translation of the Abhidharmasamuccaya, see T31:1606.703a01–5. For the Sanskrit text, see Hayashima 2003, 109. In the Viniścayasaṃgrahaṇī portion of the Yogācārabhūmi where it gives three reasons for why there must be a mind with vikalpa (rnam par rtog pa can) that cognizes the present object together with a mind with no vikalpa (rnam par rtog pa med pa), Xuanzang again uses ji mingliao 極 明 了 . The corresponding Tibetan is gsal ba (Derge 4038(1).58a4), which can be reconstructed as spaṣṭa in Sanskrit. I agree with Yao’s reading that here the mind with vikalpa should be taken as msf. See Yao’s chapter in this volume, section entitled “The Simultaneous Mental Consciousness.”

262

Keng

[Text 3] Question:28 Regarding such conventionally real properties as shapes and manifest [karma] (vijñapti-karma) on an object of matter, etc., do the five consciousnesses cognize the conventionally real properties or the ultimately real properties when cognizing [that object]? If [you claim that] they cognize length, etc., then your claim would be the same as the Sautrāntikas, but in fact there should not be any sense of “cognizing.”29 But if they do not cognize, then how could one say that [my cognition is in conformity with] the largeness or smallness [of that object]? Further, how could you claim that conventionally real material properties (jiase 假色) such as length can be included under the sphere of matter (sechu 色處; rūpa-āyatana)? Answer: Because of this issue, there are two interpretations from the western [masters in response.] One [master] claims: the five [sensory consciousnesses] merely cognize ultimately real [objects and properties] because the five consciousnesses merely [perform] direct perception (xianliang 現量; pratyakṣa) and vividly (mingliao 明了) cognize the particulars (zixiang 自相; svalakṣaṇa). For example, in the material sphere the eye-consciousness merely cognizes the ultimately real properties, blueness, etc. It is because the msf vividly grasps the length, etc. that [we claim] conventionally real material properties (jiase 假色), length, etc. are included under the sphere of matter. If a different organ encounters a different object,30 then the length is included under the sphere of dharma (fachu 法處; dharmaāyatana) because it is cognized by the mental consciousness alone. Here the phrase “in conformity with largeness or smallness” means largeness or smallness are suddenly transformed (dunbian 頓變) in conformity with the manifest matter (xianse 顯色; varṇa-rūpa). When the eye-­ consciousness cognizes it, it does not reach the understanding of “it is large or small.” Now the reason we can say “it is large or small” is that the 28 Read 間 as 問 . 29 I think the argument here against the Sautrāntikas goes back to the point made in Dignāga’s āp, namely, a conventionally real (prajñapti-sat) thing cannot produce cognition of the five sensory consciousnesses because such a thing cannot have causal efficacy. See āp Verse 2. 30 This is somewhat mysterious. My guess is that this refers to the situation where the sense organ ear hears the sound from a piano, whereas the accompanying msf is not only paying attention to the (e.g., duration of the) sound but is also measuring the piano’s length, which is not a proper sense field for ear. In this case, the length is included under the sphere of dharma.

Mānasa-pratyakṣa as Perception

263

mental consciousness cognizes it and [understands] it in terms of “being large or small” but not because the five consciousnesses can cognize it and understand it as being large or small. Hence [largeness or smallness] belongs to shape that is conventionally real. This being the case with respect to the sphere of matter, it can be extended all the way to the sphere of touch, namely, [the body organ] also does not cognize conventionally real [properties]. The [sensory consciousnesses] merely cognize each of their proper (ben 本)31 and ultimately real (shi 實) [spheres], having the four great elements (mahābhūta) as their objects. This position of mine is different from the Sautrāntikas. The second master claims: the five [sensory] consciousnesses also cognize conventionally real [properties] because they can vividly illuminate their own particulars (zixiang 自相; svalakṣaṇa), which in this context refers to the particulars at the level of sphere (chu zixiang 處自相; āyatana-svalakṣaṇa32) but not to the particulars at the level of an entity (shi zixiang 事自相; dravya-svalakṣaṇa) nor the particulars at the level of a particular (zixiang zixiang 自相自相; svalakṣaṇa-svalakṣaṇa). “Sphere” means the twelve spheres. “Entity” means this or that individual entity such as a blue [entity] and a yellow [entity]. “Particulars of particulars” means that in one blue [entity] there are many atoms (wei 微; paramāṇu), among which each atom is differentiated from every other, or each chunk [of a group of atoms] is differentiated from every other chunk [of a group of atoms]. Based on this reasoning, [the cognition of the five consciousnesses] is named as “direct perception.” I do not claim that all that are cognized through direct perception are ultimately real properties because an untainted (anāsrava) [direct perception] can also cognize conventionally real things. Yet, there are two kinds of conventionally real [things]: first, conventionally real [things] that have no substance (wuti; 無體); second, conventionally real [things] that are dependent upon others. The former include anger, etc.; the latter include regret, etc. because it is said to be an attribute (lakṣaṇa) of ignorance (chi 癡; moha). “Length, etc.” are included only under the conventionally real [things] that are dependent upon others, unlike blueness, etc. Conventionally real [things] that are dependent upon others are still ultimately real. They are named as conventionally real but their substance is something that [really] exists. [For this reason] 31

I think “original” here means “proper.” Namely, matter is the proper object for eye-­ consciousness, etc. 32 Cf. Hattori 1968, 26, Section Dab. Also see Abhidharmakośabhāṣya ad Abhidharmakośa i.10, see Pradhan 1967, 7, lines 20–21.

264

Keng

I do not commit the fault of the Sautrāntikas that [consciousnesses] cognize conventionally real [things]. Length, etc. are ultimately real because they are dependent upon other [substances]. Length, etc. are merely [different] status (fenwei 分位) of blueness, etc. We say “cognizing length” but in fact it is the five consciousnesses that reach multiple [segments] of blueness. There is no separate object [called length]. The claim “the mental consciousness alone reaches it, and hence is named as ‘cognizing conventionally real [properties]’” is made because the five consciousnesses also cognize [that] status of blueness. Hence Fascicle Three of the Yogācārabhūmi says, “when the consciousness (i.e., the ālayavijñāna) develops into matter, the development takes place in accordance with [the size of] being small, large or middle [of that matter].”33 [Kuiji’s comments:] Based on this [second position], length, etc. are also developed by the base consciousness (i.e., the ālayavijñāna). This is difficult to understand, and hence the former interpretation is better, as [I have discussed] in my commentary on Fascicle One of the Abhidharmasamuccayabhāṣya.34,35

33 34 35

This seems to refer to the passage in Fascicle Three of Xuanzang’s translation of the Yogācārabhumi: T30:1579.290a17–b5. 對 法 here refers to the text T1606 attributed to Sthiramati. Tib.: chos mngon pa kun las btus pa’i rnam par bshad pa; also reconstructed as *Abhidharmasamuccaya-vyākhyā. For Kuiji’s comments, see Fascicle Two of his Zaji lun shuji 雜 集 論 述 記 : X48:796.33a2-b9. CWSLSJ, Fascicle Two: 間 :如 色 等 法 形 、 表 等 假 ， 五 識 緣 時 為 緣 假 、 實 ?若 緣 長 等 ， 即 同 經 部 ， 應 無 緣 義 ； 若 不 緣 者 ， 如 何 此 中 言 隨 大 小 ?又 如 何 說 長 等 假 色 色 處 所 收 ?答 :由 此 義 故 ， 西 方 二 釋 。 一 云 :五 唯 緣 實 ， 五 識唯現量，明了緣自相故。如色處中唯青等實，眼識緣之。五識同時 意 識 明 了 取 得 長 短 等 故 ， 長 等 假 色 色 處 所 攝 。 若 以 別 根 .境 相 對 ， 長 等 法處收，唯意緣故。此中所言「隨大小」者，隨其顯色大小頓變。眼 識緣之無大小解。今談之為大小等也，意識緣之作大小相，非五識能 緣作大小解，即是假形。色處既爾，乃至觸處亦不緣假，唯緣本實四 大 為 境 ， 不 同 經 部 。 第 二 師 云 :五 亦 緣 假 ， 以 能 明 了 照 其 自 相 。 是 處 自相，非事自相，亦非自相自相。「處」者十二處。「事」者謂青、 黃等各各別事。「自相自相」者，於一青中復有多微一一各別，或多 分段各各有別。由如是理故名現量，非言現量皆是實法，無漏亦緣諸 假 法 故 。 然 假 有 二 :一 無 體 假 ； 二 相 待 假 。 前 如 忿 等 ； 後 如 悔 等 ， 以 癡 相說。長等但是相待假收，非如青等。相待仍實，名之為假，體是有 法，無如經部緣假之失。長等有體，依他法故。長等但是青等分位， 其實五識得多青等，名緣長等，無別緣也。「唯意得之，名『緣假』 」 者 ， 五 識 亦 緣 青 分 位 故 。 故 《 瑜 伽 論 》 第 三 卷 說 :識 變 色 時 隨 小 大 中。由此長等本識亦變，此甚難解。前解為勝，如《對法》第一疏。 (T43:1830.272a27–b26).

Mānasa-pratyakṣa as Perception

265

The core of the debate between the two masters is this: Are conventionally real properties such as length, etc. cognized by the five consciousnesses or by the mental consciousness? The former master claims that it is the mental consciousness that cognizes conventionally real properties; the latter master claims that it is the five sensory consciousnesses that cognize conventionally real properties. Kuiji endorses the former view, his reason being that conventionally real properties – length, etc. – are not developed36 by the base consciousness (ālayavijñāna). His presupposition seems to be that only ultimately real properties – color, etc. – are developed from the base consciousness.37 Kuiji’s discussion is crucial for us to understand the function of msf. Namely, it is the msf that, simultaneous with the five sensory consciousnesses, cognizes conventionally real properties – length, etc. – of an object. In contrast, the five sensory consciousnesses cognize the ultimately real properties – ­blueness, yellowness, etc. – of that object. There are two underlying presuppositions here. The first underlying presupposition is that in our ordinary experience of cognition of an object, there exist both conventionally real as well as ultimately real properties, and we simultaneously cognize both of them.38 This sounds very similar to the distinction made by philosopher John Locke (1632–1704 ce) between primary and secondary qualities. The difference is that, for Locke, color is secondary but extension (and so shape) is primary, whereas for Kuiji and some Buddhists color is ultimately real but shape is conventionally real. The second underlying presupposition is that sensory consciousnesses can only access ultimately real properties. Hence regarding such conventionally real properties as length, msf must be introduced to account for our simultaneous perception of them together with our perception of ultimately real properties such as colors. The idea that sensory consciousnesses only cognize ultimately real things can also find support in Dignāga’s āp. The whole purpose of the āp is to refute the realist position that real (i.e., consciousness-independent) entities serve as 36 37

38

This means that, when the seeds in the ālayavijñāna become mature due to suitable conditions, there is development (pariṇāma) into mental images of long or short. Kuiji also comments that the eighth consciousness does not cognize conventionally real properties such as shapes and shades. See Fascicle Three, cwslsj, T43:1830.321c2–14. Thus, it appears that there is a close connection between the “development of the consciousness” (vijñāna-pariṇāma) and the Abhidharmic notion of being “ultimately real.” This is quite noteworthy and deserves further research. For the reason why Dignāga regarded shapes are merely conventionally real, see the Introduction above.

266

Keng

the cognitive objects (ālambana-pratyaya) for cognition. Given the Abhidharmic framework of understanding the relation between cognitive object and cognition in terms of cause and effect, the realist position amounts to the idea that real entities serve as the cause for cognition. Now the problem is: What kind of cognition? Do real entities cause sensory perception (i.e. of the five sensory consciousnesses) alone or do they also cause cognition of the sixth consciousness? From āp verse 1, it is clear that Dignāga only means the five sensory consciousnesses. As it reads: Even if the atoms are the cause of the cognition through the senses (Tibetan: dbang po rnam par rig pa; Sanskrit reconstruction: indriyavijñapti), since (the cognition) does not appear under the form of those (atoms), the atoms are not the object of that (cognition), in the same way as the sense-organs (Tibetan: dbang po) (are not). (Tola and Dragonetti 2004, 33) In this verse, Dignāga is concerned only with whether atoms can be cognitive objects for cognition through the senses (indriya-vijñapti). This refers to cognition through the five sense organs but not cognition through the mental consciousness. Dignāga himself does not explicitly explain the exclusion of mental consciousness from his discussion, but Dharmapāla in his commentary entitled Guansuoyuan lun shi (“A Commentary on the āp”; T1625) gives the reason as follows: for the opponents the five sensory consciousnesses always cognize ultimately real things but the mental consciousness can cognize both ultimately real and conventionally real things because the opponents allow that mental consciousness cognizes such things as a chariot that are considered real at the conventional level. (T31:1625.889a23–25)39 Since Dignāga’s intention is to refute the existence of what are regarded by the opponents as ultimately real things, he only needs to debate with the opponents within the range of the five sensory consciousnesses. Dignāga’s idealist position in the āp would be justified as long as he successfully shows that no ultimately real entities can serve as the cause of the perception of the five sensory consciousness. Hence in order to set a mutual consensus (ubhayasiddha) for debate’s sake, Dignāga needs to assign the function of cognizing conventionally real properties such

39

Of course, the opponents also agree that mental consciousness cognizes conventionally real things such as the name or the concept “car.”

Mānasa-pratyakṣa as Perception

267

as shapes not to sense perception but to mental perception.40 For the same reason, Dharmapāla also claims that “whatever has a shape arranged in directions and corners is not a sense object for sense consciousnesses.” (T31:1625.891b6–7) Besides Dignāga and Dharmapāla, Kuiji also emphasizes this point in his cwslsj when discussing the example of “a double moon” of Dignāga’s āp. According to Kuiji, even the Sautrāntikas do not claim that it is sensory consciousnesses that cognize a double moon. If we admit that sensory consciousnesses cognize conventionally real properties such as “length, etc.,” then we would commit the “fault of inconclusiveness” (buding guo 不定過; anaikāntikadoṣa). (T43:1830.270b22–c3)41 Given the above two underlying presuppositions, msf must be postulated in order to account for our ordinary experience of perception, where we perceive both color and shape of an object at the same time. If there is no msf that perceives shape at the same time as sensory consciousnesses perceive color, then we should face the difficulty of cognizing the color of an object without cognizing its shape! We can also be sure that msf perceives shapes via direct perception because, despite the fact that conventionally real properties such as shapes or length cannot be cognized by sensory consciousnesses, they are particulars instead of universals, as indicated explicitly by the second master cited in TEXT 3 above.42 The shape of a patch of blue is always a particular. Despite the fact that it can look roundish, the shape of this patch of blue must be ever so slightly different from that of another patch of blue, and hence both must be particulars. Suppose we say: This is roundish; that is roundish, and hence both are alike in being “roundish,” we are already making up a concept of “being round” and subsuming our cognition of those two particulars under that

40

41

42

It is difficult to tell what Dignāga’s own position is in the ps, especially if we take into account his idealist position in the āp. But if we adopt the ascending scale of analysis, then at least we can say that at the conventional level – under which the ps was composed – there are ultimately real things, and it is the sensory consciousnesses that cognize them. Same as ps, Dharmapāla’s and Kuiji’s account of what ultimately real things are also operate at the conventional level. Wŏnch’ŭk (613–696 ce) also cites debates about whether it is the eyes or the mental consciousness that sees a double moon. Interestingly, he ascribes the position that it is the eyes that see a double moon to Sthiramati and the position that the eyes merely cognize ultimately real things to Dharmapāla. See Jieshenmi jing shu, Fascicle 4: X21:369.255a19–b4. The second master claims that even conventionally real properties belong to the particulars of sense sphere (chu zixiang 處 自 相 ; āyatana-svalakṣaṇa).

268

Keng

c­ oncept. But in fact, the shape of neither can be adequately described as “being round.” Given that Dignāga claims that “perception has only the particular for its object” (Hattori 1968, 24), the function of msf in its cognition of shapes must belong to direct perception rather than to inference. An important implication here is that if this interpretation is right, then Dignāga and Dharmakīrti actually have very different ideas about what particulars are. For Dignāga, since conventionally real properties such as shapes are objects of mental perception, they are considered as particulars because the objects of all kinds of perception are particulars. But in the case of Dharmakīrti, no matter whether he has three43 or four44 criteria for particulars, “having causal efficacy” is one of them. So conventionally real properties such as shapes do not qualify as particulars according to Dharmakīrti because conventionally real properties do not have causal efficacy. Dharmakīrti’s denial of conventionally real properties as particulars explains why he did not construe mental perception as the kind of cognitive function that cognizes conventionally real properties such as shapes. Another implication here is that a cognition of conventionally real properties such as “length, etc.” of an object is still recognized as knowledge. It is msf that provides knowledge of conventionally real properties by means of direct perception. Given that both mental perception and inference are treated as valid means of knowledge, we know that Dignāga here is talking about means of knowledge in a broad sense, including knowledge at both the conventional (saṃvṛti) level and at the ultimate (paramārtha) level. We can have knowledge about conventionally real things such as “what a bridge is” (“bridge” as a universal (sāmānya-lakṣaṇa) through anumāna by mental consciousness) and “that is long” (“being long” as a particular through direct perception by msf). This can be summarized by the following diagram:

43 See Thakchoe 2011. These three are: (a) having determinate spatial locations (deśaniyata) of their own; (b) temporally determinate (kālaniyata); (c) ontologically determinate (ākāraniyata) as they are causally conditioned. Conventionally real properties do not meet the last criterion because only ultimately real dharmas are causally conditioned. 44 See Hattori 1968, 80. These four are: (a) has a power to produce effects (artha-kriyā-śakti); (b) is specific (asadṛśa); (c) is not denotable by a word (śabdasyāvisayaḥ); (d) is apprehensible without depending upon other factors such as verbal conventions. Conventionally real properties do not meet the first criterion because only ultimately real dharmas have the power to produce effects. For Dharmakīrti’s definition of svalakṣaṇa, also see Yoshimizu 2004.

269

Mānasa-pratyakṣa as Perception

Ultimately real properties

Conventionally real properties

Means of knowledge (Pramāṇa)

Sense perception Mental perception

Yes

No

No

Mental con­sciousness= conceptualization

No

Yes (“length, etc.”) Yes (“a bridge, etc.”)

Perception (Pratyakṣa) Perception (Pratyakṣa) Inference (Anumāna)

Now we can answer the conundrum posed at the beginning of this paper. The existence of msf does not lead to the unwanted consequence that a person born blind can see external objects. This is because the arising of msf has the arising of one of the sensory consciousnesses as its pre-condition (Feature (B)). msf has a content that is qualitatively different from that of sensory consciousnesses: msf cognizes conventionally real properties where sensory consciousnesses cognize ultimately real properties. (Feature (C)). Thus, mental perception performed by msf provides new information (anadhigatārthagantṛ) that has not been cognized by sensory consciousnesses. Finally, msf cognizes conventionally real properties as particulars (svalakṣaṇa) (Feature (C)). Hence msf provides new information by means of direct perception (pratyakṣa). With all these taken together, the cognition of msf can be properly named “mental perception.” 5

Does Kuiji’s Interpretation Fit the ps?

A remaining question is: Should we trust Kuiji’s interpretation as highlighted above? I argue that Kuiji’s interpretation has two merits: (1) It helps us better understand the transition from ps i.5 to i.6; (2) It agrees with both Dharmakīrti and Jinendrabuddhi to the extent that all three scholars construe mental perception as directed toward external objects. Regarding (1), in ps i.5, Dignāga says, “a thing possessing many (aneka) properties cannot be cognized in all its properties by the sense. The object of

270

Keng

the sense is the form which is to be cognized [simply] as it is and which is inexpressible.” (Hattori 1968, 27)45 What does Dignāga mean by “many properties” here, and why did he emphasize that some properties escape the senses? Now with clues from Kuiji, it is likely that “many properties” here refers to both ultimately real and conventionally real properties. And this explains why immediately afterwards Dignāga introduces mental perception that cognizes conventionally real properties.46 Further, in the svavṛtti ad i.6 Hattori’s translation reads, “The mental [perception] which, taking a thing of color, etc., for its object, occurs in the form of immediate experience (anubhava) is also free from conceptual construction.” (Hattori 1968, 27)47 Now with Kuiji’s interpretation, Hattori’s translation of “rūpādiviṣayālambanam” as “taking a thing of color, etc., for its object” may not be correct because color is an ultimately real property. A better translation would be “taking a thing of visible form, etc., for its object.” The term “immediate experience” (anubhava) is used with the intention of emphasizing that the cognition of conventionally real properties such as “length, etc.” still belongs to direct perception. The second merit of Kuiji’s interpretation is that he joins Dharmakīrti and Jinendrabuddhi in construing mental perception as related to external objects rather than to internal feelings. In the case of Dharmakīrti, mental perception is about the entity at the following moment (i.e., momentary object2 at t2) of the same thing (i.e., momentary object1 at t1) that was cognized by sense perception (arising at t2). In the case of Jinendrabuddhi, he unpacks the mysterious compound in ps i.6ab “artharāgādisvasaṃvitti” as “artha-saṃvitti rāgādi-svasaṃvitti ca”48 instead of the seemingly more natural reading as 45

The Sanskrit reconsctruction reads: “dharmiṇo ’nekarūpasya nendriyāt sarvathā gatiḥ | svasaṃvedyam hy anirdeśyaṃ rūpam indriyagocaraḥ || 5 ||” (Steinkellner 2005, 3). Xuanzang’s Chinese translation of this verse from the Nyāyamukha reads: 有 法 非 一 相 ， 根 非 一 切 行 ， 唯 內 證 離 言 ， 是 色 根 境 界 (T32:1628.3b18–19). 46 I admit that my interpretation here suffers from the problem that Dignāga himself does not provide any clue concerning the specific objects of mental perception, which I take to be shapes, length, etc. In his discussion of a thing possessing many properties, he simply says, “The object of the sense is the form which is to be cognized [simply] as it is and which is inexpressible.” (Hattori 1968, 27) But the difficulty for my interpretation is that a shape, or a length, etc. is also something “to be cognized [simply] as it is and which is inexpressible.” 47 The Sanskrit text reads: “mānasam api rūpādiviṣayālambanam avikalpakam anubhavākārapravṛttaṃ|” (Steinkellner 2005, 3). Xuanzang’s Chinese translation of this verse from the Nyāyamukha reads: 意 地 亦 有 離 諸 分 別 ， 唯 證 行 轉 (T32: 1628.3b20). 48 See Hattori 1968, 92, footnote 1.45. Also see Yoshida 2012, 100.

Mānasa-pratyakṣa as Perception

271

“artha-svasaṃvitti rāgādi-svasaṃvitti ca.” The clear intention of Jinendrabuddhi is that mental perception is about external objects (artha here) rather than about internal objects (rāga here). 6

Conclusion and Larger Implications

To conclude, my main points are summarized as follows. 1. According to Kuiji, together with the arising of the sensory consciousnesses, there also arises a mental consciousness simultaneous with the five sensory consciousnesses (msf). The cognitive function of msf is precisely mental perception (mānasa-pratyakṣa) stipulated by Dignāga in his ps i.6. 2. The msf has three features: (a) having the same object as the five sensory consciousnesses; (b) assisting the arising of the five sensory consciousnesses; (c) grasping vividly the objects of the five consciousnesses. 3. The five sensory consciousnesses cognize the ultimately real properties (color, etc.) of an object whereas the simultaneous mental consciousness cognizes the conventionally real properties (shape, length, etc.) of an object. Both operate in the form of direct perception (pratyakṣa) and are without conceptualization (kalpanāpoḍha). But as indicated above, since msf cognizes shapes, etc. that are not ultimately real, msf is regarded as with vikalpa (sa-vikalpa). Based on the above discussion, I draw a few further implications below. 1. Based on Kuiji’s interpretation, mental perception has a very different role in cognition from that of self-awareness (svasaṃvitti). The objects for the former are external (matter, etc.) but the objects for the latter are internal (desire, etc.). From this I agree with Yao that we should “understand mental perception solely as the experience of external objects.” (Yao 2005, 135) Thus, when Franco insists that Dignāga “subsumes the self-apprehension of desire etc., under mental perception” (Franco 1986, 82), I think this simply means that both mental perception (in its narrow sense) and self-awareness are carried out by mental consciousness. We can still distinguish the function of mental perception (in its narrow sense) from self-awareness and construe “artha-rāgādi” in the compound “artha-rāgādi-svasaṃvitti” as referring to two different types of objects: external (artha) and internal (rāgādi). 2. A lesson from this paper is that Dignāga philosophizes largely under the Abhidharmic framework, which was then the mainstream. This is most obvious in Dignāga’s āp where the whole force of argument relies on the

272

3.

Keng

distinction between ultimately real and conventionally real dharmas. As argued in this paper, the same distinction also underlies Dignāga’s distinction between sensory perception and mental perception. Despite Hattori’s excellent study of the ps, he mainly interpreted Dignāga from a post-Dignāga perspective. I think we should balance this approach with reading Dignāga from a pre-Dignāga perspective, namely, to contextualize Dignāga in the Abhidharma and the Yogācāra frameworks prevalent during the sixth century ce. I am aware that Kuiji’s interpretation deviates dramatically from most if not all post-Dharmakīrti interpretations of Dignāga. This, however, does not necessarily prove that Kuiji is wrong but speaks to the huge discrepancy between the Chinese and the Indo-Tibetan traditions. Franco (1993, 296) once remarked that “we have to read Dignāga’s text independently of his so-called ‘Great Commentator.’ And unfortunately we do not have any other commentarial tradition except that of Dharmakīriti and his followers.” I cannot agree more with the first sentence but disagree with his second. The interpretative tradition initiated by Xuanzang in China may have preserved a quite different lens through which to see Dignāga from what was passed on by Dharmakīrti.49 I suggest that we consult both traditions to achieve a better understanding of Dignāga. Bibliography

Arnold, Dan, 2012. Brains, Buddhas, and Believing: The Problem of Intentionality in Classical Buddhist and Cognitive-Scientific Philosophy of Mind. New York: Columbia University Press. Bhattacharya, Vidhushekhara. 1957. The Yogācārabhūmi of Ācārya Asanga: The Sanskrit Text Compared with the Tibetan Version. [Calcutta]: University of Calcutta. Buescher, Harmut. 2007. Sthiramati’s Trimśikāvijñaptibhāṣya: Critical Editions of the Sanskrit Text and Its Tibetan Translation. Wien: Verlag der Österreichischen Akademie der Wissenschaften. Chu, Junjie. 2014. “On the Notion of Kaidaoyi (*Avakāśadānāśraya) as Discussed in Xuanzang’s Cheng weishi lun.” In A Distant Mirror: Articulating Indic Ideas in Sixth and Seventh Century Chinese Buddhism, edited by Chen-kuo Lin and Michael Radich, pp. 271–311. Hamburg: Hamburg University Press.

49

For example, as discussed above, Dignāga might have a very different notion about particulars from Dharmakīrti.

Mānasa-pratyakṣa as Perception

273

Dhammajoti, KL. 2007. Abhidharma Doctrines and Controversies on Perception. Hong Kong: Center of Buddhist Studies, The University of Hong Kong. Eltschinger, Vincent. 2010. “Dharmakīrti.” Revue Internationale de Philosophie 64 (3): pp. 397–440. Franco, Eli. 1986. “Once again on Dharmakīrti’s deviation from Dignāga on pratyakṣābhāsa.” Journal of Indian Philosophy 14 (1): pp. 79–97. Franco, Eli. 1993. “Did Dignāga accept four types of perception?” Journal of Indian Philosophy 21(3): pp. 295–299. Hattori, Masaaki. 1968. Dignāga, On Perception. Cambridge, Mass. U.S.A.: Harvard University Press. Hayashima, Osamu. 2003. 《 梵 ・ 蔵 ・ 漢 対 校 Electric-Text、 大 乗 阿 毘 達 磨 集 論 ・ 大 乗 阿 毘 達 磨 雑 集 論 》 (“The Electric-Text of a Critical Edition of the Parallels among the Sanskrit, Tibetan and Chinese Texts of the Abhidharmasamuccaya and the Abhidharmasamuccayabhāṣya”). url = http://www.shiga-med.ac.jp/public/ yugagyo/AS.html (Accessed August 30, 2016). Kawamura, Leslie. 1964. A Study of the Triṃśikā-vijñapti-bhāṣya. Unpublished M.A. Thesis, Kyoto University. Keng, Ching. 2018. “How Do We Understand the Meaning of a Sentence Under the Yogācāra Model of the Mind? On Disputes Among East Asian Yogācāra Thinkers of the Seventh Century.” Journal of Indian Philosophy 46 (3): pp. 475–504. Keng, Ching. 2019. “What Is Svabhāva-vikalpa and with Which Consciousness(es) Is It Associated?” Journal of Indian Philosophy. 47 (1): pp. 73–93. Kramer, J. 2016. “Some Remarks on the Proofs of the “Store Mind” (Ālayavijñāna) and the Development of the Concept of Manas.” In B. Dessein and W. Teng (Eds.), Text, Philosophy, and History: Abhidharma across Buddhist Scholastic Traditions. Leiden: Brill, pp. 146–166. Lamotte, Étienne. 1935. Saṃdhinirmocana Sūtra: L’explication des Mystères. Louvain: Bureaux du recueil Bibliothèque de l’Universite. Nagatomi, Masatoshi. 1979. “Mānasa-pratyakṣa: A Conundrum in the Buddhist Pramāṇa System.” In Sanskrit and Indian Studies, edited by M. Nagatomi, B. K. Matilal, J. M. Masson and E. Dimock, pp. 243–260. Dordrecht, Holland/ Boston, U.S.A. / London, England: D. Reidel Publishing Company. Park, Kiyeol (朴 基 烈 ). 1999. “デ ィ グ ナ ー ガ の 意 知 覚 (Mānasapratyakṣa) 説 ” (“Dignāga’s Theory of Mental Perception (Mānasapratyakṣa)”). Indogaku bukkyōgaku kenkyū 印 度 学 仏 教 学 研 究 94 (47–2): pp. 105–109(L). Pradhan, P. 1967. Abhidharmakoshabhāṣya of Vasubandhu. Patna: K.P. Jayaswal Research Institute. Shi, Renyou (釋 仁 宥 ). 2015. Chenna xianliang lilun ji qi hanchuan quanshi 陳 那 現 量 理 論 及 其 漢 傳 詮 釋 (“On the Theory of Perception of Dignāga and Its Interpretations in China”). Taipei: Dharma Drum Publishing Corp.

274

Keng

Stcherbatsky, Theodore. 1962 (1930). Buddhist Logic, Vol. 2. New York: Dover Publications Inc. Steinkellner, Ernst. 2005. “Dignāga’s Pramāṇasamuccaya, Chapter 1: A hypothetical reconstruction of the Sanskrit text with the help of the two Tibetan translations on the basis of the hitherto known Sanskrit fragments and the linguistic materials gained from Jinendrabuddhi’s Ṭīkā.” url= http://www.ikga.oeaw.ac.at/Mat/digna ga_PS_1.pdf (accessed August 30, 2016). Thakchoe, Sonam. 2011. “The Theory of Two Truths in India,” The Stanford Encyclopedia of Philosophy (Summer 2011 Edition), Edward N. Zalta (ed.), url = http://plato.stan ford.edu/archives/sum2011/entries/twotruths-india/ (accessed August 30, 2016). Tillemans, Tom T.F. 1989. “Indian and Tibetan Mādhyamikas on Mānasapratyakṣa,” The Tibetan Juornal, 14(1), Special Issue: Tibetan Contributions to the Madhyamaka: pp. 70–85. Tola, Fernando and Carmen Dragonetti. 2004. Being as Consciousness: Yogācāra Philosophy of Buddhism. New Delhi: Motilal Banardsidass. Tosaki, Hiromasa (戶 崎 宏 正 ). 1979. Bukkyō ninshikiron no kenkyū 佛 教 認 識 論 の 研 究 (“A Study of Buddhist Epistemology”). Tokyo: Daito shuppan. Wayman, Alex. 1991. “Dharmakirti and the Yogacara Theory of bīja.” In Studies in the Buddhist Epistemological Tradition: Proceedings of the Second International Dharmakīrti Conference, Vienna, June 11–16, 1989, edited by Ernst Steinkellner, pp. 419–430. Wien: Verlag der Österreichischen Akademie der Wissenschaften. Yao, Zhihua (姚 治 華 ), 2004. “Dignāga and Four Types of Perception.” Journal of Indian Philosophy, 32(1): pp. 57–79. Yoshida, Akira (吉 田 哲 ), 2012. “ジ ネ ー ン ド ラ ブ ッ デ ィ に よ る 意 知 覚 解 釈 ” (“Jinendrabuddhi’s Interpretation of Mental Perception”). Bukkyō gaku kenkyū 佛 教 學 研 究 68: pp. 95–114. Yoshimizu Chizuko (吉 水 千 鶴 子 ). 2004. “Defining and Redefining Svalakṣaṇa: Dharmakirti’s Concept and its Tibetan Modification.” In Three mountains and seven rivers: Prof. Musashi Tachikawa’s felicitation volume, edited by Shoun Hino and Toshihiro Wada, pp. 117 – 133. Delhi: Motilal Banarsidass Publishers.

Chapter 10

Mental Consciousness and Its Objects Zhihua Yao Abstract This chapter examines the following question: whether the Yogācāras would be in agreement with the Sarvāstivādins or the Sautrāntikas in their debate on the issue of what constitutes the cognitive objects of mental consciousness. The short answer is that they would lean more towards the Sarvāstivādins in their conviction that sensory objects are included in the cognitive objects of mental consciousness and that mental consciousness does have direct access to these physical objects; nevertheless, there are some complications related to the different stages of Yogācāra Buddhism.

Keywords mental consciousness – Yogācāra – physical objects – s­ imultaneous consciousness – five states of mind

…

1 Introduction In Buddhism and other Indian philosophical schools, consciousness (vijñāna) is generally classified into six different types, i.e., visual, auditory, olfactory, gustatory, tactile, and mental consciousnesses. The first five consciousnesses are characterized as sensory, and in virtue of this characterization they perceive their respective sensory objects, i.e., visible matter, sound, smell, taste, and the tangible, all of which are exclusively physical objects. Now what are the objects of the sixth mental consciousness? Do they include these physical or sensory objects? In the history of Buddhist philosophy, the Dārṣṭāntikas explicitly held that the objects of mental consciousness do not include sensory objects. This view is reported in the Mahāvibhāṣā, a Sarvāstivāda Abhidharma text: “The Dārṣṭāntikas thus speak: Each of six consciousnesses (such as visual

© koninklijke brill nv, leiden, ���1 | doi:10.1163/9789004440913_015

276

Yao

­consciousness) has its own cognitive objects (ālambana). They say that mental consciousness does not take as objects the objects of five [sense-] consciousnesses (such as visual consciousness) as it has other cognitive objects [than these ones].”1 Although the Dārṣṭāntikas held back from specifying what exactly are the objects of mental consciousness, they nevertheless made it abundantly clear that any such codification would not include the objects perceived by the five sense-consciousnesses. Later Sautrāntikas further clarified this Dārṣṭāntika view by holding that only sensory faculties (indriya) have unmediated contact with their respective sensory or physical objects; for sense-consciousness and mental consciousness are only able to cognize these objects through the mediation of mental images (ākāra). This view is usually defined as “indirect realism”; it is generally perceived as an innovative idea of the Sautrāntikas and their predecessor, the Dārṣṭāntikas. The Sarvāstivādins, in contrast, refuted such a view and held that mental consciousness takes as objects the sensory objects and directly apprehends them without the mediation of mental images; in fact, there is ample written evidence for the debate on this issue between the two parties.2 But my question is: How do the Yogācāras, being, as they were, closely associated with both groups, treat this issue? For instance, do they accept or refute the notion that mental consciousness cannot take as objects the sensory objects? Do they think that mental consciousness can have direct access to physical objects without the mediation of mental images? Moreover, which side would they take in the debate: the Sarvāstivāda or Sautrāntika position? This issue has attracted some attention among contemporary scholars. For instance, after noticing a discrepancy in both Xuanzang’s (玄奘 602– 664) and Paramārtha’s (499–569) translations of Dignāga’s (ca. 480–540) Ālambanaparīkṣāvṛtti, Lin (2008) puts forward the argument that the two Yogācāra translators must have held different views on whether it is possible for mental consciousness to have direct access to sensory objects. Lusthaus (2013), too, makes a rather interesting point on this very issue in his discussion on the opening section of the Yogācārabhūmi (hereafter YBh), a Yogācāra text of voluminous proportions. In the current chapter, I will begin with Lusthaus’s discussion on the YBh passage, which lists the cognitive objects of mental consciousness. I will argue, contra Lusthaus, that this listing includes the five sensory objects as perceived 1 Mahāvibhāṣā T1545.449a16–18: 謂 譬 喻 者 作 如 是 說 ， 眼 等 六 識 身 所 緣 境 各 別 。 彼 說意識別有所緣，不緣眼等五識所緣. 2 Dhammajoti (2007a: 136–170) provides some general background on the SarvāstivādaSautrāntika debate on this issue.

Mental Consciousness and Its Objects

277

respectively by five sense-consciousnesses. In other words, in this passage from the opening section of YBh, it follows the Sarvāstivāda view that the cognitive objects of mental consciousness consist in all dharmas, which certainly include the five types of sensory objects. If sensory objects are the cognitive objects of mental consciousness, then how does mental consciousness directly apprehend these physical objects? To reply to this question, I further examine the Yogācāra theory of cognitive process. When a physical object is present to cognition, how do sensory and mental consciousnesses cognize such an object? The Yogācāras held that they have to arise simultaneously to cognize this object. If mental consciousness did not arise simultaneously with sense-consciousness, then it would not be able to have direct access to the physical object. As we know, mental consciousness, in contrast to sense-consciousness, is defined as a conceptual cognition. If mental consciousness functions simultaneously with sense-consciousness to cognize the physical object, then is it conceptual or non-conceptual? Early Yogācāra texts often suggest that it is still a “conceptual mental consciousness.” But the difficulty is: how can the same cognitive process be non-conceptual and conceptual at the same time? Later Yogācāra commentators were at pains to overcome this difficulty. One way is to acknowledge the existence of a special state of non-conceptual mental consciousness that functions simultaneously with sense-consciousness in the initial stage of cognition. The other way is to interpret the conceptual construction in terms of “association with inquiry (vitarka) and investigation (vicāra),” which is a minimal sense of conceptual construction and can be loosely considered as “non-conceptual.” Either way, it is justified to conclude that, for the Yogācāras, mental consciousness may have direct access to physical objects by functioning simultaneously with sense-consciousness. In the next two sections, I will test this conclusion with two challenging cases: one being the Yogācāra theory of the five states of mind, and the other the meditative state of mind. The early Yogācāra theory of the five states of mind provides a primarily successive model of cognition. In this model, the first state of the spontaneous (aupanipātika) mind is identical to sense-­ consciousness and arises first to cognize its sensory objects, then followed by the searching (paryeṣaka) and discerning (niścita) minds, both of which are associated with mental consciousness. For later Yogācāra commentators, it became a thorny issue to ease the tension between this successive model of cognition with the dominant simultaneous model. Divergent opinions were advanced to reinterpret the theory of the five states of mind. Some interpret this theory as a thoroughgoing successive model of cognition, which would support the Sautrāntika view that mental consciousness cannot have direct access

278

Yao

to physical objects. But the majority view tends to impose the simultaneous model of cognition to this theory to the extent that the immediacy (samanan­ tara) of successive mental states is interpreted as simultaneous immediacy, but not successive immediacy. So even in the context of the five states of mind, it is a favorable position that mental consciousness can apprehend physical objects directly and simultaneously with sense-consciousness. At this juncture, it should be emphasized that mental consciousness cannot and should not always function simultaneously with sense-consciousnesses. For instance, the conceptual mental consciousness has to arise after senseconsciousnesses and to further process conceptually the “sense data” as acquired through senses and sense-consciousnesses. There are also cases of the so-called “independent” mental consciousness, which functions independently without the aids of sense-consciousnesses. Typical examples are dreaming and meditation. Can this independent mental consciousness also have direct access to physical objects? The Yogācāras discussed an interesting case that a meditator is roused into wakefulness upon hearing a sound. Again, divergent opinions were advanced to interpret this case, but the majority view held that the meditative state of mind is one of the exceptional cases when mental ­consciousness, being functioning independently, cannot have direct access to physical objects. 2

The Cognitive Objects of Mental Consciousness

Let me begin by quoting Lusthaus’s (2013: 587) interesting comments: Quite strikingly, unlike the previous bhūmi, here in the mental bhūmi what is not being taken as an ālambana is the aggregate of physical matter (rūpaskandha). …But in this bhūmi the mental faculties of manas and manovijñāna do not have direct access to rūpa, to physical objects. That is the job of the senses, not the mental faculty. It views sense-­ objects only indirectly, as sense-spheres (āyatana). Xuanzang adds the word “inner,” i.e., the six inner sense-spheres (六內處 liu nei chu) to emphasize that the mental sphere operates at a remove from the physical world. The senses feed it objects, which it then processes in its own way, in its own sphere. It is not simply that manas and manovijñāna acquire whatever knowledge of physical things they obtain only as mediated through the sense – that, after all, is not only standard Abhidharma fare but common sense as well. Rather, it is that rūpa and the rūpaskandha ceases to provide cognitive supports (ālambana), so that attention and

Mental Consciousness and Its Objects

279

analysis turn exclusively to the mental sphere, even when pondering how physical events and causes produce cognitive repercussions. This is the critical Yogācāra move that has been mistaken for idealism ever since. These comments, as in many writings of Lusthaus, are full of philosophical insights. To wit: his musings are of a kind that attracts the attention of any philosophically minded scholar. Nevertheless, and apropos of the above quotation, I find his claim baseless. If mental consciousness does not have “direct access” to physical objects, it may or may not suggest a “critical” shift in the direction of “idealism.” I am not going to argue about this point here, rather I will focus on the issue itself: is it possible for mental consciousness to have direct access to physical objects? Or, to put it another way, do all the cognitive objects (ālambana) that fall into the purview of mental consciousness include physical objects? First of all, it is worth pointing out that Lusthaus’s claim that it is not possible for mental consciousness to take as objects “the aggregate of physical matter” is partially the fault of poor translation. His claim was based on a passage from the Manobhūmi section of YBh, which outlines the objects of the mental faculty (manas) or mental consciousness (manovijñāna): ālambanaṃ katamat | sarvadharma ālambanaṃ | kevalaṃ tu vedanā­ skandhaḥ saṃjñāskandhaḥ saṃskāraskandho ’saṃskṛtaṃ cānidarśanam apratighaṃ ca rūpaṃ ṣaḍāyatanaṃ sarvabījāni ca ||3 Lusthaus (2013: 586) translates this passage as follows: What is its ālambana? It takes all phenomena (sarvadharma) as its ālambana. Its ālambana is not shared [publicly, i.e., it is only accessible subjectively]. It takes [as its ālambana] only the aggregates (skandha) of hedonic tone (vedanā), associative-thinking (saṃjnā), and embodiedconditioning (saṃskāra); the unconditioned; invisible and non-resistant physical things (rūpa); the six [inner] sense spheres (ṣaḍ-āyatana); as well as everything derived from mental seeds. Note that he translates the term kevala (’ba’ zhig, bu gong zhe 不共者) as “to be not shared publicly,” and goes on to argue that the cognitive objects (ālambana) of mental consciousness are “to be only accessible subjectively.” This, in my 3 YBh (S) 11,12–14. See YBh (C) T1579.280b11–13 and YBh (T) D4035.5b6–7.

280

Yao

opinion, is contextually at odds with the whole sentence. It is better, therefore, to translate the passage in the following way: What are its cognitive objects (ālambana)? It takes all dharmas as its cognitive objects. [The cognitive objects] that are exclusively of (kevala) [the mental faculty (manas) or mental consciousness (manovijñāna) include] the aggregates of feeling (vedanā), ideation (saṃjnā), and volition (saṃskāra); the unconditioned (asaṃskṛta); invisible and non-resistant matter (rūpa); the six [inner] abodes (ṣaḍāyatana); as well as all mental seeds. The key difference between my translation and his is that it pushes the term kevala back within the context of the whole sentence; and, by doing so, the translation corresponds with the predicate “to be exclusively of” or, alternatively is consonant with the Chinese translation “to be not shared with (five sense-consciousnesses).” Furthermore, my translation is also supported by the commentaries on YBh provided by both Kuiji (窺基 632–682) and Dunryun (遁倫 active during the seventh century): “The exclusive objects [of the mental faculty or mental consciousness] are not shared with five [sense-] consciousnesses.”4 A familiarization with the discussions surrounding mental objects in the Sarvāstivāda Abhidharma, will yield the truism that there is so much highly contested historical debate behind this simple listing. The mental objects that are not shared with the five sense-consciousnesses can be further classified into three groups: (1) all mental seeds; (2) the six inner abodes; (3) the other five items. Among them, the mental seed is characteristically Yogācāra in its conceptual definition, and was not mentioned in the Abhidharma listing. Kuiji and Dunryun further clarify things by remarking that these mental seeds include only the defiled seeds, but not the undefiled ones.5 As for the second group, the term ṣaḍāyatana (six abodes) in Sanskrit is not clear, but Xuanzang’s translation adds the word “inner” (nei 內) and clarifies that this term refers to the six inner abodes, which contain the five senses and the mental faculty (manas). Crucially, these six inner abodes are distinguished from the six “outer” abodes, i.e., the five sensory objects and dharma. Lusthaus seems to misread the term by attributing it to a line of sensory objects, seeming to take Xuanzang’s qualification as meaning that “the mental sphere operates 4 yjsdllz T1829.7a29: 不 共 境 者 ， 不 共 五 識 故 。 yjlj T1828.319a1–2: 不 共 境 者 ， 不 共五識故. 5 See yjsdllz T1829.7a20; yjlj T1828.318c24–25.

Mental Consciousness and Its Objects

281

at a remove from the physical world” (op. cit.). If the six inner abodes are taken to be the objects of mental consciousness, then what can be said about the six outer abodes? Are they also mental objects? Since they are not mentioned explicitly in this listing, there is room for Lusthaus to speculate that they are not objects of mental consciousness. But if we examine this passage carefully, we will find that the six outer abodes are implicitly mentioned under “all dhar­ mas,” which is the very first part of the definition for mental objects. In the Abhidharma literature, there are several ways to explain the inferred multiplicity of the term “all dharmas,” depending on the different schemes of classifying reality. One such scheme classifies reality into twelve abodes. In the Vijñānakāya, a Sarvāstivāda Abhidharma work, these twelve abodes, and how they function as the cognitive objects of respective consciousness, are outlined as follows: There are twelve abodes (āyatana): the eye-abode, the visible-abode, the ear-abode, the sound-abode, the nose-abode, the odor-abode, the tongueabode, the taste-abode, the body-abode, the tangible-abode, the mentalabode, and the dharma-abode. Question: By how many consciousnesses is the eye-abode cognized? … By how many consciousnesses is the dharma-abode cognized? Answer: The visible-abode is cognized by two consciousnesses: visual and mental consciousnesses. The sound-abode is cognized by two consciousnesses: auditory and mental consciousnesses. The odor-abode is cognized by two consciousnesses: olfactory and mental consciousnesses. The taste-abode is cognized by two consciousnesses: gustatory and mental consciousnesses. The tangible-abode is cognized by two consciousnesses: tactile and mental consciousnesses. The remaining seven abodes are cognized by mental consciousness alone.6 If reality is broken down into twelve abodes, then all these elements or dhar­ mas combined are constitutive of the objects of mental consciousness. For instance, among them, the five sensory objects pertain to the shared objects of their respective sense-consciousness and mental consciousness; however, the other seven abodes, i.e., the six inner abodes and the dharma-abode, are 6 Vijñānakāya T1539.546c18–24: 有 十 二 處 ， 謂 眼 處 、 色 處 、 耳 處 、 聲 處 、 鼻 處 、 香 處 、 舌 處 、 味 處 、 身 處 、 觸 處 、 意 處 、 法 處 。 問 :眼 處 幾 識 所 識 ， 乃 至 法 處 幾 識 所 識 ?答 :色 處 二 識 所 識 ， 謂 眼 識 及 意 識 。 聲 處 二 識 所 識 ， 謂 耳 識 及意識。香處二識所識，謂鼻識及意識。味處二識所識，謂舌識及意 識。觸處二識所識，謂身識及意識。餘七處唯意識所識.

282

Yao

exclusively the objects of mental consciousness. This explains why the six inner abodes are listed as mental objects which are “not shared with the five sense-consciousnesses” in YBh. By the same logic, the five sensory objects are included in “all dharmas,” and perceived as the cognitive objects of mental consciousness. Another popular scheme classifies reality into eighteen realms (dhātu), consisting of the twelve abodes and six consciousnesses. Accordingly, the cognitive objects of mental consciousness would include all these eighteen elements; again, the five sensory objects must be perceived as the shared objects of their respective sense-consciousness and mental consciousness. In his Abhidharmakośabhāṣya (hereafter AKBh), Vasubandhu (ca. 400–480) makes this point explicit: “The visible, sound, odor, taste and the tangible are perceived (anubhūta) respectively by the visual, auditory, olfactory, gustatory and tactile consciousnesses. All of them are cognized by mental consciousness. They are respectively cognized by two consciousnesses. It is therefore known that the [other] thirteen realms are cognized by a single mental consciousness, because they are not objects of the five groups of [sense-]consciousness.”7 His critic Saṃghabhadra (active during the fifth century) echoes this point of view;8 such concordance indicates that they are in full agreement with each other, and that Vasubandhu at this point had not yet deviated from the orthodox Vaibhāṣika position. In the above discussions, one may notice that the key concept of dharma is ambiguous. On the one hand, being one of the twelve abodes or eighteen realms, it is designated as the object of the mental faculty (manas) or mental consciousness (manovijñāna). On the other hand, it seems to cover all the twelve abodes or the eighteen realms. The corollary for this orientation seems to be: when mental consciousness cognizes all these abodes or realms, it is able to apprehend all dharmas as objects. In the Sarvāstivāda Abhidharma, we see two ways of dealing with this incongruity. One is to consider the dharmaabode or dharma-realm as metonyms for all dharmas. If this is agreed, then all the twelve abodes or the eighteen realms can be subsumed under the nomenclature for a single abode or realm – dharma. Such an opinion is reported in the Mahāvibhāṣā: “Some other masters say that the dharma-realm covers completely all dharmas”; “some say that the dharma-abode covers all 7 AKBh 57,16–20 ad ak i.48a: rūpaśabdagandharasaspraṣṭavyadhātavo yathāsaṃkhyaṃ cak­ ṣuḥśrotraghrāṇajihvākāyavijñānair anubhūtā manovijñānena vijñāyante | evam ete pratyekaṃ dvābhyāṃ vijñānābhyāṃ vijñeyā bhavanti | śeṣās trayodaśa dhātavaḥ pañcānāṃ vijñā­nakāyānām aviṣayatvād ekena manovijñānena vijñeyā ity ākhyātaṃ bhavati | 8 See Nyāyānusāra T1562.377a3–6.

Mental Consciousness and Its Objects

283

dharmas.”9 In a typical Sarvāstivāda ontology, the dharma-realm should also include dharmas of the past, present and future. This position is explained thusly: Question: What is the dharma-realm? Answer: Dharmas that have been, are being, and will be cognized by the mental faculty are called the dharma-realm. Those that have been cognized by the mental faculty refer to the dharma-realms that have been cognized by the past mental faculty. Those that are being cognized by the mental faculty refer to the dharma-realms that are being cognized by the present mental faculty. Those that will be cognized by the mental faculty refer to the dharma-realms that will be cognized by the future mental faculty.10 Another way of resolving this problem is to limit the content of the dhar­ ma-abode or dharma-realm to “seven dharmas” only, namely, to the non-­ informative matter (avijñaptirūpa), the three aggregates, i.e., feeling, ideation, and volition, and the three unconditioned, i.e., space, cessation through understanding (pratisaṃkhyānirodha), and cessation without understanding (apratisaṃkhyānirodha). In his AKBh, Vasubandhu states: “The aggregates of feeling, ideation and volition should also be established as abodes and realms. That is, these three [aggregates], together with the non-informative [matter] and the three unconditioned – these seven entities can be called dharma-abode and dharma-realm.”11 Again, similar statements are found in Saṃghabhadra’s Nyāyānusāra (T1562.342a2–4) and also in the Mahāvibhāṣā (T1545.65a29–b1, T1545.985.b15). According to Dhammajoti (2007b: 38–39), this development that tapers down the dharma-abode or dharma-realm to the specifics of the seven dharmas was originated from the Jñānaprasthāna. By doing so, it reformulates the older classification schemes of aggregates (skandha), abodes (āyatana) and realms (dhātu) into a new five-grouped taxonomy in which the non-informative matter and the three unconditioned are integrated.

9

Mahāvibhāṣā T1545.370c19–20: 有 餘 師 說 ， 法 界 總 攝 一 切 法 盡 。 Mahāvibhāṣā T1545.985b8: 或 說 ， 法 處 攝 一 切 法 . 10 Mahāvibhāṣā T1545.370c3–7: 問 :法 界 云 何 ?答 :諸 法 為 意 已 、 正 、 當 了 是 名 法 界。已為意了者，謂諸法界已為過去意界所了。正為意了者，謂諸法 界正為現在意界所了。當為意了者，謂諸法界當為未來意界所了. 11 AKBh 17,2–5 ad ak i.15cd: ete punas trayaḥ | vedanāsaṃjñāsaṃskāraskandhā āyatanadhātuvyavasthāyāṃ dharmāyatanadhātvākhyāḥ sahāvijñaptyasaṃskṛtaiḥ || ity etāni sapta dravyāṇi dharmāyatanaṃ dharmadhātuś cety ākhyāyante ||

284

Yao

As we see, these seven dharmas correspond to the group three category in the Yogācāra listing, which consists of the aggregates of feeling (vedanā), ideation (saṃjnā), and volition (saṃskāra), the unconditioned (asaṃskṛta), invisible and non-resistant matter (rūpa). The invisible and non-resistant matter is apparently a Yogācāra adoption of the Sarvāstivāda concept of non-informative matter (avijñaptirūpa). The Yogācāras here do not specify the number of unconditioned dharmas; nevertheless if recourse is taken to the standard Yogācāra list for what is unconditioned, we will find that it includes six elements: (1) space, (2) cessation through understanding, (3) cessation without understanding, (4) motionless cessation (āniñjya), (5) cessation of ideation and feeling (saṃjñāvedayitanirodha), and (6) thusness (tathatā). As a whole, this tabulation by Yogācāra of mental objects seems to combine the two incongruous Abhidharma traditions (“all dharmas” versus “seven dharmas”) through an augmentation of its own elements (“mental seeds,” etc.). Even with this complication, it is patently clear that the five sensory objects are not only included, but are mutually coexisting objects prevailing within their respective sensory and mental consciousnesses. Just a few lines above the YBh passage under discussion, there is further evidence to support these findings. For instance, this is what it has to say about the objects of visual consciousness: All the visible, extensive and informative [matter] are the operative field (gocara) of eyes, objects (viṣaya) of eyes, the operative field of visual consciousness, objects of visual consciousness, cognitive objects (ālambana) of visual consciousness, the operative field of mental consciousness, objects of mental consciousness, and cognitive objects of mental consciousness.12 According to this passage, physical matter, as far as it is visible, extensive and informative (vijñapti), can be the operative field (gocara) or object (viṣaya) of 12

YBh (S) 5,8–10: sarvāsāṃ varṇṇasaṃsthānavijñaptīnāṃ cakṣurgocara[ś cakṣurviṣaya]ś cakṣurvijñānagocara[ś cakṣurvijñānaviṣaya]ś cakṣurvijñānālambanaṃ manovijñāna­ gocaro manovijñānaviṣayo manovijñānālambanam iti paryāyāḥ || See YBh (C) T1579. 279b15–17. All these three Sanskrit words, gocara, viṣaya and ālambana, are denotative of the term “objects.” To distinguish them, I have translated them respectively as “the operative field,” “objects” and “cognitive objects.” Note that eyes, as well as visual consciousness and mental consciousness, can have gocara or viṣaya; but only visual consciousness and mental consciousness can have ālambana. So one way to distinguish ālambana from its related “objects” is to see that the term is usually associated with consciousness (vijñāna) and thus acts as its intentional or cognitive object. In contrast, gocara as the operative field or viṣaya as objects, can be more objective and independent.

Mental Consciousness and Its Objects

285

eyes, visual consciousness and mental consciousness. At the same time it can be the cognitive object (ālambana) of visual and mental consciousnesses. However, similar to the case of the Sarvāstivāda concept of non-informative matter (avijñaptirūpa), when the physical matter is invisible or non-resistant, it can only be cognized by mental consciousness, and so serves as its cognitive objects. The same applies to the other four types of sensory objects: sound, odor, taste, and the tangible. They are the objects shared by the mental consciousness with their respective sense and sense-consciousness.13 So, according to this opening section in YBh, the cognitive objects of mental consciousness consist in “all dharmas,” which include the sensory objects such as visible matter, sound, odor, taste, and the tangible. This implies that mental consciousness can have direct access to physical objects as sense-­ consciousnesses do. Moreover, I would further assert that this passage does not in any sense suggest any movement toward “idealism.” 3

The Simultaneous Mental Consciousness14

If sensory objects are the cognitive objects of mental consciousness, then how does mental consciousness directly apprehend these physical objects? To reply to this question, we have to examine how the Yogācāras analyze the actual cognitive process. Further on in the Viniścayasaṃgrahaṇī Chapter in YBh, specifically its comments on the saying “[mental consciousness] cognizing and ­taking as objects its own objects and others’ objects” (svaparaviṣayālambanavi­ j­­ñaptiḥ),15 it makes the cogent point that: “Mental consciousness can take as objects others’ objects and its own objects. ‘Taking others’ objects’ means that it apprehends the objects of the five groups of [sense-]consciousness, either simultaneously or otherwise. ‘Taking its own objects’ means that it arises to apprehend the dharmas which are not the objects of the five groups of [sense-] consciousness.”16 Remarking on this passage, Dunryun, citing Shentai (神泰 13 14 15

16

See YBh (S) 6,19–7,2; 7,16–18; 8,8–10; 9,5–7 and YBh (C) T1579.279c12–14, 279c22–24, 280a3– 5, 280a15–17. Keng (Chapter 9) abbreviates it as msf, whereas Lin (Chapter 11) calls it “synchronic mode of mental consciousness” (sm). This is found in the Manobhūmi section of YBh (S) 12,1: svaparaviṣayālambanavijñaptiḥ. Lusthaus (2013: 587 n29) notes that Xuanzang’s translation (YBh (C) T1579.280b22) omits the word “others,” and he rightly points out that it refers to “the five [sense-]consciousnesses whose cognitive objects can be taken up by the mental faculty,” but without noticing that this would contradict his earlier remarks that we cited in the beginning of the chapter. YBh (T) D4038(1).5b5–6: yid kyi rnam par shes pa de ni gzhan gyi yul la dmigs pa dang | rang gi yul la dmigs pa yin te | de la gzhan gyi yul la dmigs pa ni rnam par shes pa’i tshogs

286

Yao

active in the seventh century), relates the issue of mental objects to a more complicated problematic, namely, the simultaneity of sensory and mental consciousnesses: When there arise five [sense-]consciousnesses, why must there be mental consciousness [accompanying them]? My interpretation: This question means that when five [sense-]consciousnesses arise there must be mental consciousness [accompanying them], because mental consciousness can also apprehend the objects of the other five [sense-]consciousnesses.17 This reasoning of Dunryun and Shentai is not fully justified since it is impossible to infer that sensory and mental consciousnesses can arise simultaneously merely from the fact that they each have access to sensory objects; not only that: it is possible for mental consciousness to arise in a later moment and access these objects as past objects. Indeed this is one way to justify the Sarvāstivāda epistemology. However, the Yogācāras take a stronger stance on the simultaneity of mental and sensory consciousnesses, for they believe that mental and sensory consciousnesses can cognize simultaneously the present sensory object. This issue is discussed elsewhere in YBh: Question: With regard to the conceptual mind (sems, *citta) and the nonconceptual mind, should we say that they take as objects the same ­present-arising object or not? Answer: We should say that they take the same [present] object. Question: Why? Answer: There are three reasons: (1) clarity; (2) paying attention to it; and (3) nurtured by two bases (āśraya).18

17 18

lnga po dag gi yul la cig car ram | cig car ma yin par dmigs pa gang yin pa’o || rang gi yul la dmigs pa ni rnam par shes pa’i tshogs lnga po dag gi dmigs pa med par chos la dmigs nas ’byung ba gang yin pa’o || The latter part “which are not the objects of five groups of [sense-]consciousness” is omitted in YBh (C) T1579.580c11–13: 又 復 意 識 能 縁 他 境 及 縁自境。縁他境者，謂縁五識身所縁境界，或頓不頓。縁自境者，謂 縁法境. yjlj T1828.600c16–18: 若 有 五 識 ， 何 故 定 有 意 識 ?解 :此 問 意 由 意 識 亦 得 解 他 五識境故，五識起時，定有意識. YBh (T) D4038(1).58a3–4: de la ci rnam par rtog pa can dang rnam par rtog pa med pa’i sems da ltar byung ba’i yul nyid kyis yul mtshungs pa ’am | yul mi mtshungs par brjod par bya zhe na | smras pa yul mtshungs par brjod par bya’o || rgyu gang gis she na | smras pa | rgyu gsum gyis gsal ba’i phyir dang | de la rtsol ba’i phyir dang | gnas gnyis kyis brtas par byas pa’i phyir ro || YBh (C) T1579.601b15–19: 問 :有 分 別 心 、 無 分 別 心 ， 當 言 同 緣 現 在 境 耶 ?為 不 同 耶 ?答：當言同緣*現 在 *境 界 。 何 以 故 ?由 三 因 故 :謂 極

Mental Consciousness and Its Objects

287

In his commentaries, Dunryun first points out that the delineation of conceptual and non-conceptual mind refers respectively to mental and sensory consciousnesses, that, moreover, the main point of the passage is “to clarify that the conceptual mental consciousness and the non-conceptual five [sense-] consciousnesses simultaneously cognize the object present to them.”19 In the Saṃdhinirmocana Sūtra (hereafter sns), a text which was incorporated in the Viniścayasaṃgrahaṇī Chapter of YBh in its entirety, the Yogācāras first explicitly advocated the idea that mental and sensory consciousnesses arise simultaneously and share the same object. By first contextualizing the idea that the six consciousnesses arise on the basis that they abide in the appropriating consciousness (ādānavijñāna), the text goes on to argue for the simultaneity of mental and sensory consciousnesses: “Simultaneously accompanying (sahānucara) visual consciousness, at the same time and with the same object, the conceptual mental consciousness arises.”20 The same applies to the other four sense-consciousnesses. This idea is further emphasized: no matter how many sense-consciousnesses function simultaneously, there is only a single mental consciousness arising simultaneously with them and sharing their objects. Expatiating on this, the authors attest that: If a single visual consciousness arises simultaneously (sahabhūta) [with its objects], only a single conceptual mental consciousness (vikalpakaṃ manovijñānam), having the same object as the visual consciousness, arises simultaneously. Even if two, three, or four groups of [sense-]consciousness arise simultaneously, or five [groups of sense-consciousness] arise simultaneously, in these cases still only a single conceptual mental ­consciousness, having the same objects of the five groups of [sense-]­ consciousness, arises simultaneously.21

19 20 21

明 了 故 ， 於 彼 作 意 故 ， 二 依 資 養 故 。 *This term “present” is omitted in the Tibetan translation. yjlj T1828.639a27–28: 分 別 意 與 無 分 別 五 識 同 緣 現 境 . sns v.4, 56,6–8: [mig gi rnam par shes pa de dang lhan cig rjes su ’jug pa | dus mtshungs pa | spyod yul mtshungs pa | rnam par rtog pa’i yid kyi rnam par shes pa ’ang ’byung ngo ||] See sns (C) T676.692b21–22 and YBh (C) T1579.718b1–2. sns v.4, 56,14–20: gal te mig gi rnam par shes pa gcig lhan cig ’byung na ni mig gi rnam par shes pa dang spyod yul mtshungs pa | rnam par rtog pa’i yid kyi rnam par shes pa gcig kho na lhan cig ’byung ngo || gal te rnam par shes pa’i tshogs gnyis sam | gsum mam | bzhi lhan cig gam | lnga char lhan cig ’byung na ’ang der rnam par shes pa’i tshogs lnga po dag dang spyod yul mtshungs pa | rnam par rtog pa’i yid kyi rnam par shes pa ’ang gcig kho na lhan cig ’byung ngo || See sns (C) T676.692b25–28 and YBh (C) T1579.718b4–8, both of which omit the first lhan cig (simultaneously).

288

Yao

Within such a context of defending a theory of “multiple-layered cognitive series,” the Yogācāras sound unusual in their emphasis on the singularity of mental consciousness. It is prudent to make sense of it in a wider context and therefore go beyond the single versus multiple layered cognitive series contrast as advocated by Schmithausen (1967). In his commentaries, Dunryun discusses four different views on this matter. The Sautrāntikas, in their advocacy of the so-called “single-layered cognitive series,” believe that only one consciousness (vijñāna) or mind (citta) functions in a given moment, and that there need not be any mental activity (caitta) accompanying it. For the Sautrāntikas, it is possible for a certain mental activity (caitta), e.g., feeling, to function at a given moment, but without the need to presuppose the functioning of mind (citta). Indeed, feeling shares equal status with the mind or consciousness.22 The Sarvāstivādins, although insisting upon the singularity of mind or consciousness, allow for the simultaneity of mind with at least ten kinds of mental activities, otherwise known as the ten mahābhūmikas. These include feeling (vedanā), ideation (samjñā), thinking (cetanā), contact (sparśa), desire (chan­ da), discrimination (prajñā), memory (smṛti), resolution (adhimokṣa), attention (manaskāra), and meditation (samādhi). Both schools would be loath to admit the simultaneity of mental and sensory consciousnesses. The other two schools, namely, Yogācāra and Mahāsāṃghika, do admit, however, their simultaneity, but with some subtle differences. Dunryun says: “Now in the Mahāyāna (=Yogācāra), although it is admitted that eight consciousnesses arise at the same time, two [consciousnesses of the same kind] cannot arise simultaneously within a single consciousness [series]. This repudiates the view held by the Mahāsāṃghika masters: that there are two mental consciousnesses even when one is in the womb.”23 In other words, the Mahāsāṃghikas adhere to the theory of “multiple-layered cognitive series” in its proper sense, a paradigm analogous to “the society of mind” theory in contemporary cognitive science.24 The Yogācāras, apparently influenced by the Mahāsāṃghikas, are not willing to go so far in conceding that two consciousnesses of the same kind can function simultaneously. Hence they emphasize the singularity of mental consciousness while defending the simultaneity of mental and sensory consciousnesses. For them, it is this single and simultaneous mental consciousness that apprehends the objects of sense-consciousnesses. 22 23 24

See Yao 2005: 99–101 for more discussions on this early Sautrāntika view. On the view of the late Sautrāntikas after Dharmakīrti, see Funayama 2000 for further discussions. yjlj T1828.668a10–12: 今 大 乘 中 ， 雖 許 八 識 同 一 時 生 ， 於 一 識 中 無 二 并 起 ， 則破大眾部師，在胎有二意識. See Yao 2005: 14.

Mental Consciousness and Its Objects

289

When apprehending sensory objects, is the simultaneous mental consciousness conceptual or non-conceptual? The phrase “conceptual mental consciousness” (vikalpakaṃ manovijñānam) in this sns passage suggests that it is conceptual. But the difficulty is: how can the same cognitive process be nonconceptual (in virtue of sense-consciousness) and conceptual (in virtue of mental consciousness) at the same time? To explain this puzzling point, Dunryun cites Kuiji, who in turn invokes no less an authority than Dignāga, to distinguish mental and sensory consciousnesses. Kuiji says: As elucidated in [Dignāga’s] Pramāṇasamuccaya and Nyāyamukha, the five [sense-]consciousnesses are only perception, so too the simultaneous mental consciousness. Now here the five [sense-]consciousnesses are apparently the perception that apprehends [their objects] with clarity. They do not lend themselves to conceptual construction: this is because they are not associated with inquiry (vitarka) and investigation (vicāra). The sixth [mental consciousness], although able to cognize the present25 [object] simultaneously with the five [sense-consciousnesses], is designated as “conceptual construction,” only because it is associated with inquiry and investigation. If [mental consciousness] cognizes the present [object] simultaneously, it has the ability to apprehend [the object] with clarity. If it is the mental consciousness that [arises] after the five [sense-] consciousnesses, then it cannot apprehend [its object] with clarity as it cognizes a past [object]. Just as one cannot apprehend with clarity things of the past hundreds or thousands of cosmic ages (kalpa).26 In Dignāga’s system, perception (pratyakṣa) is defined as “devoid of conceptual construction.”27 It includes sensory perception, mental perception (manopratyakṣa), self-cognition (svasaṃvedana), and yogic perception. Sensory perception corresponds to five sense-consciousnesses, hence Kuiji’s asseveration that they are “only perception.” Mental perception is what he called “the simultaneous mental consciousness” which arises simultaneously with five sense-consciousnesses and takes as objects their present objects. This simultaneous mental consciousness, by virtue of being a type of perception, should 25 Read xianzai 現 在 for xianliang 現 量 . 26 yjlj T1828.639b6–11: 如 《 集 量 理 門 論 》 云 :五 識 唯 現 量 ， 同 時 意 識 亦 爾 。 今 此 中 五 識 顯 (reads 雖 )明 了 現 量 ， 不 名 有 分 別 ， 不 與 尋 伺 等 相 應 故 。 第 六 雖 與 五 同 緣 現 量 ， 名 為 分 別 ， 以 與 尋 伺 相 應 故 。 名 (reads 若 )同 緣 現 在，可得明了。若彼五識後意識，即不明了，緣過去故，如緣百千劫 事 ， 即 不 明 了 。 Corrected after Kuiji’s yjsdllz T1829.198c6–11. 27 PS 3c: pratyakṣaṃ kalpanāpoḍhaṃ.

290

Yao

be considered as “non-conceptual.” However, here Dunryun and Kuiji, in following the wording of YBh, still designate it as “conceptual construction,” but only in the sense of “being associated with inquiry and investigation,” which, however, does not denote “conceptual construction” in the Dignāgian system. They further distinguish the simultaneous mental consciousness from mental consciousness proper or “the mental consciousness that arises after five senseconsciousnesses,” which, by definition, is capable of conceptual construction and lacks clarity in apprehending its mental objects. In contrast, the simultaneous mental consciousness can apprehend the present sensory or physical objects with great clarity. Primarily under the influence of Dignāga, the Chinese Yogācāras made exhaustive efforts to single out “the simultaneous mental consciousness” as a non-conceptual perception. Probably owing to this reason, Bodhiruci’s (active during the sixth century) translation of the passage from the sns cited earlier amends the phrase “conceptual mental consciousness” (fenbie yishi 分別意識) to “non-conceptual mental consciousness” (wu fenbie yishi 無分別意識).28 But the crucial issue is: how do we define “conceptual construction”? In what sense is sense-consciousness non-conceptual and mental consciousness conceptual? Can a conceptual consciousness apprehend sensory objects? As we know, Dignāga defines “conceptual construction” (kalpanā) in terms of the “association with name, genus, and so forth,”29 whereas Kuiji here understands it in terms of “association with inquiry (vitarka) and investigation (vicāra).” Both definitions reveal a long and complicated history within the Buddhist theories of “conceptual construction.” The background for Kuiji’s definition is the Sarvāstivāda theory, which distinguishes three types of conceptual construction, namely, (1) intrinsic conceptual construction (svabhāva-vikalpa), (2) conceptual construction in terms of reasoning (abhyūhanā-vikalpa) or determination (nirūpaṇā), and (3) conceptual construction in terms of recollection (anusmaraṇa-vikalpa).30 Among them, intrinsic conceptual construction is defined as “inquiry and investigation”31 or “inquiry.”32 Inquiry and investigation refer respectively to gross and subtle states of mind. In his AKBh, Vasubandhu proposes that all five sense-consciousnesses are perennially associated with inquiry and 28 See sns (Cb) T675.669a28. This discrepancy was reported in Wǒnch’ǔk’s (圓 測 613–696) commentaries on sns, see jsmjs X369.248c20–22. 29 ps 3d: nāmajātyādiyojanā || 30 See Mahāvibhāṣā T1545.219b7–11 and ak i.33ab. 31 Mahāvibhāṣā T1545.219b7–8. 32 AKBh 35,9 ad ak i.33ab. Saṃghabhadra also agrees with Vasubandhu, see his Nyāyānusāra T1562.350b11.

Mental Consciousness and Its Objects

291

investigation, arguing that “these consciousnesses are gross (audārika) in nature in that they are directed toward external things.”33 Having said this, he then challenges his own assumption with the following question: “If five groups of [sense-]consciousness have inquiry and investigation, how can they be said to be free from conceptual construction?”34 Vasubandhu answers his own question accordingly: “They are free from conceptual construction insofar as the conceptual construction is in terms of determination (nirūpaṇā) and recollection (anusmaraṇa).”35 He explains that these sense-consciousnesses possess only the first type of conceptual construction, namely, intrinsic conceptual construction, but are lacking in the other two, namely, conceptual construction in terms of determination and recollection. “For this reason, they are said to be free from conceptual construction, just as a horse with one leg is said to be legless.”36 As we can see, although Kuiji adopts the Sarvāstivāda definition of “conceptual construction” in terms of inquiry and investigation, he finds himself in disagreement with Vasubandhu’s assertion that the five sense-consciousnesses possess the intrinsic conceptual construction; in fact, Kuiji explicitly states that “they are not associated with inquiry and investigation.” On this point, Kuiji represents a later but dominant view among the Yogācāras, which was probably influenced by Dignāga. In the Cheng wei shi lun (*Vijñaptimātratāsiddhi) translated and edited by Xuanzang, two opposite views on the non-conceptual nature of sense-­ consciousness are introduced. One notable view is that the five sense-­ consciousnesses are associated with inquiry and investigation, most typically represented by the aforementioned Vasubandhu in AKBh and in Sthiramati’s (ca. 475–555) Abhidharmasamuccayabhāṣya.37 Sthiramati further relates the issue to a more sophisticated theory of conceptual construction as developed by the Yogācāras in YBh. This theory classifies conceptual construction into 33

AKBh 34,6–7 ad ak i.32ab: tathā hy eta audārikā bahirmukhatvāt | The sentence is missing in the Sanskrit edition by Pradhan (p. 22, l. 4 ad ak i.32ab) and Paramārtha’s Chinese translation (AKBh (Cp) T1559.168a13–14), but it is restored provisionally in Ejima’s critical edition on the basis of the Tibetan translation (AKBh (T) D4090.42a5) and Xuanzang’s Chinese translation (AKBh (Cx) T1558.8a12–13). It is reported and refuted in Saṃghabhadra’s Nyāyānusāra (T1562.350a7–9), and also commented on by Sthiramati and Pūrṇavardhana in their commentaries on AKBh. 34 AKBh 35,3–4 ad ak i.33ab: yadi pañca vijñānakāyāḥ savitarkāḥ savicārāḥ katham avikalpakā ity ucyante | 35 ak i.33ab: nirūpaṇānusmaraṇavikalpenāvikalpakāḥ | 36 AKBh 35,7–8 ad ak i.33ab: tasmād avikalpakā ity ucyante | yathaikapādako ’śvo ’pādaka iti || 37 See Cheng weishi lun T1585.36a24ff and lvp 1928–9: 389.

292

Yao

seven different types: they are conceptual constructions that are (1) with image (naimittika), (2) without image (anaimittika), (3) spontaneously forthcoming (svarasavāhī), (4) searching (paryeṣaka), (5) examining (pratyavekṣaka), (6) defiled (kliṣṭa), and (7) undefiled (akliṣṭa).38 As we can see, the conceptual constructions of searching (paryeṣaka) and examining (pratyavekṣaka) closely resemble inquiry (vitarka) and investigation (vicāra). He further holds that the five sense-consciousnesses possess the spontaneously forthcoming conceptual construction: “because it comes forth absolutely spontaneously in respect to its objects as it is, without making it diverse.”39 His explanation for this type of conceptual construction follows closely the definition given in YBh: “What is the spontaneously forthcoming [conceptual construction]? It is that which comes forth spontaneously in respect to the present object, exclusively on the strength of the object.”40 The second view proposes that inquiry and investigation are only associated with mental consciousness, but not the five sense-consciousnesses. One of the reasons given to justify this view is “because inquiry and investigation regard names, etc., as objects, but the five groups of [sense-]consciousness do not take them as objects.”41 If inquiry and investigation are involved with names, then they would be conceptual construction even in its Dignāgian sense, and hence they cannot be associated with the five sense-consciousnesses. This argument is strongly influenced by Dignāga, who more strictly defines conceptual construction as pertaining to “association with name, genus and so forth.” It may have been proposed by his follower Dharmapāla (ca. 530–561). In response to Sthiramati’s view that the five sense-consciousnesses possess the spontaneously forthcoming conceptual construction, Dharmapāla insists that in the YBh itself all seven types of conceptual construction are introduced via the mental level (manobhūmi) and thus belong exclusively to this level. When one attests that the five sense-consciousnesses possess the spontaneously forthcoming conceptual construction, it really means that “this [spontaneously forthcoming conceptual construction] is the inquiry and investigation that are associated with the conceptual mental consciousness, which accompanies the five [sense-]consciousnesses.”42 In other words, the five sense-consciousnesses 38 39 40 41 42

See YBh (S) 12,9–11; ybh (C) T1579.280c2–4. ASBh 16,13–14: acitrayitvālaṃbanaṃ yathāsvaṃ viṣayeṣu svarasenaiva vahanāt…. YBh (S) 12,15: svarasavāhī katamaḥ | pratyupasthite viṣaye svarasena yo viṣayabalād eva vartate || See YBh (C) T1579.280c7–8. Cheng weishi lun T1585.36b4–6: 尋 伺 以 名 身 等 義 為 所 緣 ， 非 五 識 身 以 名 身 等 義 為 境 故 。 See lvp 1928–9: 389. Cheng weishi lun T1585.36b9–10: (瑜 伽 說 ):此 是 五 識 俱 分 別 意 識 相 應 尋 伺 。 See lvp 1928–9: 390.

Mental Consciousness and Its Objects

293

themselves are never by themselves conceptualizing, but are always accompanied simultaneously by mental consciousness, which is conceptualizing in its very nature. As we see, Dharmapāla-Kuiji and Vasubandhu-Sthiramati have different views on the non-conceptual nature of sense-consciousnesses. The former holds that they are non-conceptual in virtue of being disassociated with inquiry (vitarka) and investigation (vicāra), while the latter believes that they are still associated with inquiry and investigation, which entails the so-called “intrinsic” conceptual construction (svabhāva-vikalpa) – a minimal conceptual construction that can be considered as “non-conceptual.” In the same vein, we can understand the conceptual simultaneous mental consciousness in YBh, sns and their commentaries in terms of this minimal sense of conceptual construction, which can be loosely called “non-conceptual.” This interpretation would be consonant with the Dignāgian view that the simultaneous mental consciousness (i.e., mental perception43) is non-conceptual in the sense of disassociation with name, genus, and so forth, which is a stronger requirement for conceptual construction. Despite their different views on the non-conceptual nature of sense-consciousnesses and the simultaneous mental consciousness, both Vasubandhu-Sthiramati and Dharmapāla-Kuiji would agree that senseconsciousnesses and their accompanying mental consciousness are capable of cognizing simultaneously present sensory objects. So in the current case, it is not just any kind of mental consciousness that is being addressed here; on the contrary, it is only the simultaneous mental consciousness that is capable of having direct access to sensory or physical objects. Mental consciousness proper or “the mental consciousness arising after sense-consciousnesses,” in contrast, does not have such a capacity. 4

The Five States of Mind

The issue whether it is possible for the mental and sensory consciousnesses to cognize the same present object is further discussed in YBh, but in this instance within the context of the five states of mind. This can be a testing case for our earlier conclusion that mental consciousness can have direct access to physical objects. In the section that opens YBh, believed to be the earliest layer of this voluminous text, the Yogācāras develop an influential theory that delineates the 43

See Keng (Chapter 9) for further discussions on the relationship between the simultaneous mental consciousness and mental perception.

294

Yao

cognitive process into five states. The following passage is regarded as the locus classicus for this theory: In this case, when visual consciousness arises, three [states of] mind (cit­ ta) are attained. Sequentially, [they are] (1) spontaneous (aupanipātika) [mind], (2) searching (paryeṣaka) [mind], and (3) discerning (niścita) [mind]. Of these, the first is none other than visual consciousness; the other two are mental consciousness. Then, subsequent to discerning mind, (4) [mind that is] defilement or purification is to be seen. On account of that, visual consciousness also arises as (5) [mind] uniformly flowing from that (tannaiṣyandika), being wholesome or unwholesome, but not owing to its own capacity of conceptualization. So long as this mental faculty (manas) is not distracted elsewhere, mental consciousness and visual consciousness are wholesome or defiled.44 Just as visual consciousness arises, [the other four consciousnesses] up to tactile consciousness should be thought of in the same way.45 This early Yogācāra theory emphasizes the successive sequence of these various different states of mind, which is consonant with the Sautrāntika theory of mind. It explicitly maintains that visual consciousness is identical to the first state of the spontaneous mind; this is followed by the searching and discerning minds, both of which are associated with mental consciousness. Thus, ostensibly, these two types of consciousness, being as they are respectively sensory and mental, arise consecutively. Schmithausen (1967: 124–125) suggests that this Yogācāra theory was influenced by the Mahīśāsakas and so treats it as supporting evidence for the Sautrāntika “presupposition” (Voraussetzung) in Vasubandhu’s thought. In this sense, a certain tension arises between this successive model of cognition and the idea of simultaneity as expounded in both the sns and the Viniścayasaṃgrahaṇī Chapter of YBh discussed in the last section. 44 45

Here Xuanzang’s Chinese translation adds the phrase “coming forth continuously” (xiangxu er zhuan 相 續 而 轉 , YBh (C) T1579.280a26–27). YBh (S) 10,2–7: tatra cakṣurvijñāna utpanne trīṇi cittāny upalabhyante yathākramam aupanipātikaṃ paryeṣakaṃ niścitaṃ ca | tatra ca ādyaṃ cakṣurvijñānam eva | dve manovijñāne | tatra niścitāc cittāt paraṃ saṃkleśo vyavadānaṃ ca draṣṭavyaṃ | tatas tannaiṣyandikaṃ cakṣurvijñānam api kuśalākuśalaṃ pravarttate | na tu svavikalpavaśena | tāvac ca dvayor manovijñānacakṣurvijñānayoḥ kuśalatvaṃ vā kliṣṭatvaṃ yāvat tan mano na anyatra vikṣipyate || yathā cakṣurvijñāna utpanna evaṃ yāvat kāyavijñānaṃ veditavyaṃ || See YBh (C) T1579.280a22–27. Also see Lusthaus (2013: 585) for his translation, but he did not further discuss this passage, nor did he use it to support his remarks that we cited in the beginning of the chapter.

Mental Consciousness and Its Objects

295

This tension became a theoretical burden for the later Yogācāras who tried to reconcile the two models of cognition. In the Cheng wei shi lun, for instance, after citing the aforesaid contentious passage, the issue is addressed full-on: “The purpose of this [passage] is to demonstrate that so long as [this mental faculty (manas) is engaged with its objects], both mental consciousness and visual consciousness usher forth continuously. Since there is never a case where there is visual consciousness but no mental consciousness, it is impossible that these two kinds of consciousness could surge forth one by one in consecutive fashion.”46 Apparently, this is an attempt to impose a simultaneous model of cognition upon a successive one. I am not going to deal with the tension between these two models of cognition;47 instead I will focus on the central issue of the current chapter, that is, whether mental consciousness has direct access to sensory objects. As we know, the Sautrāntikas, by insisting that mental and sensory consciousnesses have no direct access to sensory objects, vehemently argue for a successive model of cognition. So it would be interesting to see how the Yogācāras would treat this issue in their sequential theory of the five states of mind. The following passage discusses two possible scenarios: In this case, when mental consciousness is disengaged (anābhoga), distracted, or has unfamiliar cognitive objects, desire and so forth do not arise. This mental consciousness is to be explained as being spontaneous [mind], taking only the past cognitive object (ālambana). The searching or discerning mental [consciousness] (manas) that arises immediately subsequent to (samanantara) the five groups of [sense-]consciousness is explained as having only the present object (viṣaya). It is so when it takes exclusively the objects (viṣaya) of those [sense-consciousnesses] as its cognitive objects (ālambana).48 Here in the first instance, mental consciousness is also perceived as a spontaneous mind when it is disengaged, distracted, or countenancing unfamiliar cognitive objects. In this case, mental consciousness is not accompanied by 46

Cheng weishi lun T1585.21a17–19: 彼 意 定 顯 ， 經 爾 所 時 ， 眼 意 二 識 ， 俱 相 續 轉。既眼識時，非無意識，故非二識互相續生. 47 For detailed discussions on this issue, see Chu 2014 and Chu (forthcoming). 48 YBh (S) 59,12–15: tatra manovijñāne ’nābhogavikṣipte ’saṃstutālambane na asti chandādināṃ pravṛttiḥ | tac ca manovijñānam aupanipātikaṃ vaktavyam atītālambanam eva | pañcānāṃ vijñānakāyānāṃ samanantarotpannaṃ manaḥ paryeṣakaṃ niścitaṃ vā vartamānaviṣayam eva vaktavyaṃ | tac cet tadviṣayālambanam eva tad bhavati || See YBh (C) T1579.291b17–20.

296

Yao

any sense-consciousness, and hence it is an “independent” mental consciousness. Furthermore, this mental consciousness only takes as its cognitive object the past object, which, presumably, is the sensory object of the previous sense-consciousness. In the second instance, when it is engaged with its objects, mental consciousness arises in the form of the searching mind or the discerning mind immediately subsequent to the five sense-consciousnesses. Its only object is the present object, which is also the object perceived by sense-consciousness. According to Dunryun, however, this understanding is only one of the three interpretations circulating at that time in India. Besides this earlier and more standard interpretation, he also reports two alternative understandings that he learned from Xuanzang’s oral transmission.49 The first is attributed to Jinaputra (active during the sixth century),50 who uses the supernatural mental state of the Buddha and Bodhisattva as an example of spontaneous mental consciousness. This mental state, however, is not restricted to cognizing exclusively the past object: because the Buddha and Bodhisattva are supposed to be able to cognize objects at any time – even beyond the temporal sphere – and at their will. To reconcile this view with the text that is YBh, Jinaputra proposes a reading of the phrase “taking only the past cognitive object” (atītālambanam eva) as part of the subsequent sentence, and that henceforward we should retranslate the relevant portion of the text as: “This mental consciousness is to be explained as being spontaneous [mind]. Taking only the past cognitive object, the searching or discerning mental [consciousness] arises immediately subsequent to the five groups of [sense-]consciousness.” This interpretation is radically different from the earlier standard interpretation. According to this interpretation: in the first instance, mental consciousness, being the independent spontaneous mind, can cognize objects at any time; in the second instance, mental consciousness, being the searching or discerning mind, apprehends only the past object. He justifies the latter case by claiming that: “[The searching and discerning minds] arise by taking the past object exclusively. This is because they can recollect the objects of the five [sense-]consciousnesses.”51 In all probability, Jinaputra was influenced by the Sautrāntikas on this point as it sounds extremely similar to their view that mental consciousness can only access past sensory objects. Seen in this way, we find a correlation between the 49

Kuiji’s commentaries on YBh only introduce two opinions without indicating their proponents. See yjsdllz T1829.21b17–c22. 50 See yjlj T1828.333c18–27, but the relevant passage is not found in the extant abridged commentaries on YBh by Jinaputra (see yjsdls T1580). 51 yjlj T1828.333c25–26: 由 追 緣 五 識 所 緣 境 故 ， 唯 緣 過 去 境 生 .

Mental Consciousness and Its Objects

297

above and Lusthaus’s earlier remarks that mental consciousness has no direct access to physical objects. However, given the fact that the remaining section of the text still closely adheres to the view that a searching mind and a discerning mind posit objects in the form of the present object, Jinaputra has to concede that “sometimes” they can also take the present object, thus not merely positing the past object as he had insisted earlier. Therefore, I would say that this new formulation by Jinaputra offers an interesting, but inconsistent interpretation, one that, however, is not justified by the text of YBh itself. The second alternative interpretation is attributed to an anonymous master, who is apparently influenced by the simultaneous model of cognition. Pertaining to the first instance, it is reaffirmed that the independent spontaneous mental consciousness can only apprehend the past object, and that the reason for this can be attributed to a lack of clarity. With regard to the second instance, it emphasizes that mental consciousness, being that of a searching or discerning mind, arises simultaneously with sense-consciousnesses and cognize the present object that is perceived by sense-consciousnesses. Moreover, this interpretation applies the simultaneous model of cognition thoroughly, and does so through a radical reinterpretation of the term “immediacy” (sa­ manantara) – specifically, “It is the simultaneous immediacy, but not the suc­ cessive immediacy.”52 In addition to the three interpretations from India, Dunryun, this time citing his contemporary Huizhao (慧沼 648–714), provides an interpretation that is suffused with idealism. According to this interpretation, mental consciousness, by virtue of being the spontaneous mind or the searching and discerning minds that arise after the five sense-consciousnesses, can apprehend both past and present objects. “If it is in the sense of cognizing the original objects that [are produced] by the store consciousness (ālayavijñāna), [mental consciousness] can be considered as apprehending the past object. If it is in the sense of cognizing the objective aspect of [mental consciousness] itself, then [mental consciousness] only apprehends the present object.”53 We can make sense of Huizhao’s claim under two frameworks of idealism. External objects in the realistic sense are always already there when mental consciousness arises; therefore, they can be considered as the past object for mental consciousness. But these objects themselves were produced by the store consciousness in the remote past. We can call this type of reasoning metaphysical idealism. On the other hand, when mental consciousness emerges, its objective aspect is also 52 53

yjlj T1828.334a4: 此 是 同 時 無 間 ， 非 前 後 無 間 也 . yjlj T1828.334a5–7: 若 緣 賴 耶 本 境 義 邊 ， 名 緣 過 境 。 若 緣 自 境 相 分 義 邊 ， 唯緣現境.

298

Yao

produced simultaneously; therefore, it is capable of cognizing its objective aspect as the present object. Such reasoning, based as it is on the simultaneity of cognition and its objects, can be identified as epistemological idealism. Either way the controversial issue concerning the cognitive objects of mental consciousness is resolved within an idealistic framework. As we see, the early Yogācāra theory of the five states of mind provides a primarily successive model of cognition, which is consonant with the Sautrāntika theory of cognition, and hence would support the view that mental consciousness cannot have direct access to physical objects. For later Yogācāra commentators, it is a thorny issue to ease the tension between this successive model of cognition with the dominant simultaneous model as discussed in earlier sections. Divergent opinions were advanced to reinterpret the theory of the five states of mind. Some interpret this theory as a thoroughgoing successive model of cognition, while some others impose the simultaneous model to this theory to the extent that the immediacy (samanantara) of successive mental states is interpreted as simultaneous immediacy, but not successive immediacy. The latter interpretation gradually became dominant with the development of Yogācāra from India to China. As a result, even in the context of the five states of mind, it became a favorable position that mental consciousness can apprehend physical objects directly and simultaneously with sense-consciousness. 5

The Meditative State of Mind

Another problematical case related to the cognitive objects of mental consciousness has to do with the mind in a meditative state. In the Viniścayasaṃgrahaṇī Chapter in YBh, a possible scenario is discussed. Suppose someone in a meditative state (mnyam par bzhag pa, *samāhita), upon hearing a sound, is roused into wakefulness. Now is it only her auditory consciousness that hears this? Or does her mental consciousness also apprehend the sound? Moreover, does this auditory experience occur in the meditative state or not? Concerning this last question, YBh puts forth the argument that this auditory experience occurs in the meditative state. “Therefore, it is not the case that one immediately wakes up [from her meditative state] upon hearing a sound. Having experienced the sound, if there is a desire [to seek out the sound], then she would have woken up [from meditation] at some later point”54 Dunryun adds that there are cases of deep meditation in which a 54

YBh (T) D4038(1).181a1–2: de’i phyir sgra ’dzin pa tsam gyis ldang ba ma yin gyi | sgra so sor rig nas ’dod na | de’i ’og tu ldang bar ’gyur ro || YBh (C) T1579.650c23–24: 非 取 聲 時 ， 即 便出定。領受聲已，若有悕望，後時方出.

Mental Consciousness and Its Objects

299

sound is heard, but only for a moment, yet the subject continues to stay in the meditative state. In such occurrences, he thinks that there is only the spontaneous, but no searching or discerning, mind at work. In any case, the above implies that consciousness is responsive to external stimuli during the meditative state. But the question is: Which consciousness exactly is responsive? Mental or sensory consciousness? For instance, YBh contends that “a certain other auditory consciousness, accompanying the mental consciousness associated with meditation, arises [to hear the sound]; because it is not the mental consciousness associated with meditation that apprehends this very sound.”55 Dunryun comments that it is certainly the auditory consciousness that hears the sound because mental consciousness is still in the meditative state, and that, further, “if [mental consciousness] takes its meditative objects, it cannot apprehend this sound.”56 To support his argument he quotes extensively from Shentai and Kuiji, who both hold that in such a case mental consciousness would not cognize the same object with sensory consciousness. If this is so, does it not contradict our earlier conclusion that mental consciousness can access sensory objects directly and simultaneously with sensory consciousness? Dunryun anticipates this question by explaining that there is a difference between a discursive mind and a meditative one. For instance, the discursive mind does not have a strong enough mental consciousness. When sensory consciousness arises, there must be simultaneous mental consciousness accompanying it. But when the auditory consciousness hears a sound while the mind is in a meditative state, the mental consciousness is still mindful of its own meditative objects to such a degree that it does not apprehend this sound. Toward the end of his commentaries, Dunryun also records two alternative views from India. One view agrees with the Chinese commentators who believed that in the meditative state the auditory consciousness does not have a simultaneous mental consciousness accompanying it and apprehending its objects. In other words, the mental consciousness in a meditative state is still mindful of its own meditative objects. The other view, however, holds that “at that time when mental consciousness is mindful of meditative objects, it also apprehends the sound simultaneously.”57 Dunryun concludes that “now we say 55

56 57

YBh (T) D4038(1).180b7–181a1: *mnyam par bzhag pa’i* yid kyi rnam par shes pa dang | lhan cig byung ba’i rna ba’i rnam par shes pa gzhan kho na skye bar ’gyur te | mnyam par bzhag pa’i yid kyi rnam par shes pa de nyid sgra ’dzin pa ma yin pa’i phyir ro || (*The sDe dge edition reads mnyam par ma bzhag pa’i, corrected after the Peking edition P5539.188a7.) YBh (C) T1579.650c20–22: 與 定 相 應 意 識 俱 轉 餘 耳 識 生 ， 非 即 彼 定 相 應 意 識 ， 能 取此聲. yjlj T1828.703b21: 若 緣 定 境 ， 不 取 此 聲 . yjlj T1828.704a21: 爾 時 意 識 正 緣 定 境 ， 兼 亦 緣 聲 .

300

Yao

that [in this case] the five [sense-]consciousnesses also cognize the same object as the mental consciousness.”58 The latter view, not so dissimilar to the view discussed in the last section, takes the simultaneous model of cognition to extremes, and even applies it to the case of meditative state. But the majority of Yogācāra commentators did not seem to accept this view. In the Yogācāra system that was endorsed by commentators such as Kuiji, Dunryun, and Huizhao, mental consciousness is distinguished into two basic types: (1) mental consciousness that accompanies sensory consciousness and (2) mental consciousness that does not accompany sensory consciousness. The first type is also called simultaneous mental consciousness: it takes the same object as sensory consciousness in most cases, but there are exceptional cases when they do not perceive the same object. The second type is further distinguished into (1) mental consciousness subsequent to sensory consciousness and (2) independent mental consciousness such as the mental states of dreaming, meditation, and so on. Neither of these two would share the same object with sensory consciousness. Therefore, in this section the meditative state is singled out as an example of the independent mental consciousness, which is an exceptional case of mental consciousness not taking the same object as sensory consciousness. At this juncture, we can also respond to the challenge of the so-called blind argument.59 Those who agree with the Sautrāntikas would insist that if mental consciousness could access physical objects directly, then the blind or deaf would not exist because their mental consciousness is still intact, and hence they are supposed to be able to apprehend physical objects as normal people do. The Yogācāras would argue that the mental consciousness of the blind or deaf is such an exceptional case of the independent mental consciousness. Without the aids of sense-consciousnesses, the mental consciousness of the blind or deaf cannot apprehend any physical objects. 6 Conclusion By examining these textual sources, we can tentatively answer the question raised at the beginning of this chapter: whether the Yogācāras would be in agreement with the Sarvāstivādins or the Sautrāntikas in their debate on the issue of what constitutes the cognitive objects of mental consciousness. The 58 59

yjlj T1828.704a21–22: 今 謂 五 識 亦 與 意 識 同 緣 一 境 . Lin (2008: 129) discussed such a blind argument as found in Dharmapāla’s commentary on Dignāga’s Ālambanaparīkṣā, see T1625.889b6–8.

Mental Consciousness and Its Objects

301

short answer is that they would lean more towards the Sarvāstivādins in their conviction that sensory objects are included in the cognitive objects of mental consciousness and that mental consciousness does have direct access to these physical objects; nevertheless, there are some complications related to the different stages of Yogācāra Buddhism. In the initial chapters of YBh, there is a strong mark of the Sarvāstivāda influence. This is indicated in the mental objects tabulation provided by the Yogācāras, which entails a combination of two incongruous traditions (“all dharmas” versus “seven dharmas”) of the Sarvāstivāda Abhidharma imbued with a number of its own elements (“mental seeds,” etc.). Furthermore, the Yogācāra theory of the five states of mind suggests a successive model of cognition, one that is dominant in the Sarvāstivāda and Sautrāntika epistemology. In either case, however, the Yogācāras do not subscribe to the Sautrāntika view that mental consciousness cannot have direct access to sensory objects. In the later chapters of YBh and sns, the Yogācāras adopt the simultaneous model of cognition, and thus explicitly state that the mental and sensory consciousnesses are capable of perceiving the same present sensory object simultaneously. To reconcile its opposition to the earlier successive model of cognition, the later Yogācāra commentators in India and China held divergent opinions. Some, such as Jinaputra, stuck to the successive model by arguing that mental consciousness in the capacity of the searching and discerning minds can only access the past object; as a matter of fact, this is closer to the Sautrāntika view. But the majority of Yogācāra commentators maintain that these minds still apprehend present sensory objects. They attempt to interpret the theory of the five states of mind as a simultaneous process of cognition. Some of them even apply the simultaneous model of cognition to the meditative state of mind, which, in fact, is a case of the independent mental consciousness that does not apprehend physical objects simultaneously with sense-consciousnesses. As is evident in the aforementioned divergent opinions, the question whether mental consciousness can apprehend physical objects directly without the mediation of mental images is a complicated and controversial issue throughout the history of Buddhist philosophy. The early debates between the Sarvāstivādins and Sautrāntikas were carried on by those who hold cognition without images (nirākāravāda) and those who hold cognition with images (sākāravāda) until the final stage of Indian Buddhism. In this chapter, I focus on YBh and its commentaries which reflect an early Yogācāra view that tends to agree with the Sarvāstivādins and those who hold cognition without images. For them, mental consciousness can have access to physical objects directly

302

Yao

and simultaneously with sense-consciousnesses. But this view is far from conclusive. Later Yogācāras as represented by Dignāga and Dharmakīrti lean more towards the Sautrāntika position of cognition with images and argue against their opponents by developing their sophisticated theory of mental images (ākāra). Acknowledgment This work was supported by the Academy of Korean Studies (ksps) Grant funded by the Korean Government (moe) (AKS-2012-AAZ-2102). My thanks to Achim Bayer, Ching Keng, Mark Siderits, and John Spackman for their stimulating comments. Bibliography and Abbreviations AK(Bh) Yasunori Ejima, Abhidharmakośabhāṣya of Vasubandhu. Chapter i: Dhātunirdeśa. Bibliotheca Indologica et Buddhologica 1. Tokyo: Sankibo Press 1989. AKBh (Cp) Abhidharmakośabhāṣya of Vasubandhu, Chinese translation. Apidamo jushe shilun 阿 毘 達 磨 俱 舍 釋 論 , trans. Paramārtha, T29: 1559. AKBh (Cx) Abhidharmakośabhāṣya of Vasubandhu, Chinese translation. Apidamo jushe lun 阿 毘 達 磨 俱 舍 論 , trans. Xuanzang, T29: 1558. AKBh (T) Abhidharmakośabhāṣya of Vasubandhu, Tibetan translation. Chos mngon pa’i mdzod kyi bshad pa, D4090: Ku,26a1-khu,95a,7. ASBh Abhidharmasamuccaya-Bhāṣya. Deciphered and edited by N. Tatia. Patna 1976. Cheng weishi lun *Vijñaptimātratāsiddhi of Dharmapāla et al. Cheng weishi lun 成 唯 識 論 , trans. Xuanzang, T31: 1585. Chu 2014 Junjie Chu, “On the Notion of Kaidaoyi (*Avakāśadānāśraya) as Discussed in Xuanzang’s Cheng weishi lun.” In A Distant Mirror: Articulating Indic Ideas in Sixth and Seventh Century Chinese Buddhism, ed. by Chen-kuo Lin & Michael Radich (Hamburg: Hamburg University Press 2014), 271–311. Chu forthcoming Junjie Chu, “Sthiramati and the Thesis of Mental Awareness Accompanying Sensory Awareness.” Forthcoming. D sDe dge Edition of Tibetan Tripiṭaka in www.asianclassics.org. Dhammajoti 2007a K.L. Dhammajoti, Abhidharma Doctrines and Controversies on Per­ ception. 3rd ed. Hong Kong: The University of Hong Kong 2007. Dhammajoti 2007b K.L. Dhammajoti, Sarvāstivāda Abhidharma. 3rd ed. Hong Kong: The University of Hong Kong 2007.

Mental Consciousness and Its Objects

303

Funayama 2000 Toru Funayama 船 山 徹 , “Dharmakīrti no rokushiki-kuki-setsu” ダ ル マ キ ー ル テ ィ の 六 識 倶 起 説 . In Indo no bunka to ronri: Tosaki Hiromasa Hakushi koki kinen ronbunshū イ ン ド の 文 化 と 論 理 :戸 崎 宏 正 博 士 古 稀 記 念 論 文 集 (Fukuokashi: Kyushu Daigaku Shuppankai 2000), 319–345. jsmjs Jie shenmi jing shu 解 深 密 經 疏 of Wǒnch’ǔk, X21: 369. Lin 2008 Chen-Kuo Lin, “Object of cognition in Dignāga’s Ālambanaparīkṣāvṛtti: On the controversial passages in Paramārtha’s and Xuanzang’s translations.” Journal of the International Association of Buddhist Studies 30–1 (2008): 117–138. Lusthaus 2013 Dan Lusthaus, “A Note on Medicine and Psychosomatic Relations in the First Two Bhūmis of the Yogācārabhūmi.” In The Foundation for Yoga Practitioners: The Buddhist Yogācābhūmi Treatise and Its Adaptation in India, East Asia, and Tibet, edited by Ulrich Timme Kragh (Harvard University Press 2013), 578–595. lvp 1928–29 L. de la Vallée Poussin, Vijñaptimātratāsiddhi. La Siddhi de Hiuan-Tsang. 2 vols. Paris: Librairie Orientaliste Paul Geuthner 1928–29. Mahāvibhāṣā ascribed to 500 arhats. Apidamo da piposha lun 阿 毘 達 磨 大 毘 婆 沙 論 , trans. Xuanzang, T27: 1545. Nyāyānusāra of Saṃghabhadra. Apidamo shun zhengli lun 阿 毘 達 磨 順 正 理 論 , trans. Xuanzang, T29: 1562. ps Ernst Steinkellner, Dignāga’s Pramāṇasamuccaya, Chapter 1. A hypothetical recon­ struction of the Sanskrit text with the help of the two Tibetan translations on the basis of the hitherto known Sanskrit fragments and the linguistic materials gained from Jinendrabuddhi’s Ṭīkā. Available online http://www.ikga.oeaw.ac.at/Mat/dignaga_ PS_1.pdf, 2005 (last accessed 15 September 2015). Schmithausen 1967 Lambert Schmithausen, “Sautrāntika-Voraussetzungen in Viṃśatikā und Triṃśikā.” Wiener Zeitschrift für die Kunde Süd- und Ostasiens 11 (1967): 109–136. sns Saṃdhinirmocana Sūtra, Tibetan translation. dGongs pa nges par ’grel pa, in É. Lamotte, Saṃdhinirmocanan Sūtra, L’Explication des Mystères, Texte Tibétain. Édité et traduit. Louvain 1935. sns (C) Saṃdhinirmocana Sūtra, Chinese translation. Jie shenmi jing 解 深 密 經 , trans. Xuanzang, T16: 676. sns (Cb) Saṃdhinirmocana Sūtra, Chinese translation. Shenmi jietuo jing 深 密 解 脫 經 , trans. Bodhiruci, T16: 675. T Taishō Edition of Chinese Tripiṭaka in tripitaka.cbeta.org. Vijñānakāya of Devaśarman. Apidamo shi shen zu lun 阿 毘 達 磨 識 身 足 論 , trans. Xuanzang, T26: 1539. Yao 2005 Zhihua Yao, The Buddhist Theory of Self-Cognition. London and New York: Routledge 2005. YBh (C) Yogācārabhūmi, Chinese translation. Yujiashi di lun 瑜 伽 師 地 論 , trans. Xuanzang, T30: 1579.

304

Yao

YBh (S) Yogācārabhūmi, Sanskrit edition. The Yogācārabhūmi of Ācārya Asaṅga, ed. by V. Bhattacharya. University of Calcutta 1957. YBh (T) Yogācārabhūmi, Tibetan translation. rNal ’byor spyod pa’i sa, D4035–4042 (Tshi 1a-, Dzi, Wi, Zhi, Zi, ’i 68b7). yjlj Yujia lun ji瑜 伽 論 記 of Dunryun, T43: 1828. yjsdllz Yujiashi di lun lue zuan 瑜 伽 師 地 論 略 纂 of Kuiji, T43: 1829. yjsdls Yujiashi di lun shi 瑜 伽 師 地 論 釋 of Jinaputra, trans. Xuanzang, T30: 1580.

Chapter 11

Vasubandhu’s Theory of Memory: A Reading Based on the Chinese Commentaries Chen-kuo Lin Abstract In this chapter I take an exegetical approach to the philosophical issue of memory treated in Vasubandhu’s Ātmavādapratiṣedha (Refutation of the Theory of a Self), an appended treatise in the Abhidharmakośa (Treasury of Knowledge), and Viṃśikā (Twenty Verses on Consciousness-Only). These two texts will be examined within the continuity of two texts in Vasubandhu’s philosophical enterprise. Unlike Janet Gyatso’s edited volume, In The Mirror of Memory (1992), this chapter deliberately reads these texts from the perspective of Chinese commentaries composed by Puguang (普光, died in 664) and Kuiji (窺基, 632–682), two eminent disciples of Xuanzang (玄奘, 602– 664). The main thrust of this chapter is to show how Vasubandhu argues against realist theories of memory, which affirm either the existence of the self as the owner of memory or the existence of external objects as the support (ālambana) of recollection. Vasubandhu concludes that memory can be explained without presupposing the existence of self (ātman) and external world. This is equivalent to saying that one should be more cautious when memory is employed by the realists, such as the Jains and the Vaiśesikas, as a means of proving the existence of ātman and external world. To recap the arguments in the Ātmavādapratiṣedha and the Viṃśikā as a whole, Vasubandhu’s theory of memory is presented in his critique of the ­realist’s employing memory to prove the existence of self and external world, arguing that memory occurs in the continuum of consciousness without necessarily presupposing either an ātman (permanent self) as the agent or an external world as object of recollection. In the continuum of consciousness, recollection occurs with the aid of attentive mindfulness (smṛti): that the “seeds” first deposited in the same continuum of consciousness are re-activated later to appear as that which has been formerly experienced in sense perception. For the realist, memory simply means retrieving the previous perception of an external object. Without the external object, perception and recollection cannot be caused to arise. In reply, though Vasubandhu agrees that memory should rely on the previous perception, he does not accept the claim that perception implies the e­ xistence

© koninklijke brill nv, leiden, ���1 | doi:10.1163/9789004440913_016

306

Lin

of external objects. The same argument also applies to the critique of the existence of self as the agent of memory. In the same vein, Kuiji further developed Vasubandhu’s theory by arguing that memory is possible because mental consciousness (manovijñāna) or mental faculty (manas) functions simultaneously with sensory perceptions to give memory meaningful content. Kuiji’s innovative interpretation set the ground for the later development of the theory of memory in East Asian Buddhism.

Keywords Vasubandhu – Kuiji – memory – smṛti – Ātmavādapratiṣedha – Viṃśikā

…

1 Introduction It is obvious that memory constitutes the core of our experience. Without memory we, no matter whether as individuals or as a society, are unable to talk about our past. Some philosophers also argue that personal identity and history could not be explained if memory is not properly accounted for. The main questions they ask are: Is a certain form of self theoretically required for the occurrence of memory? Can memory be counted as a valid means of cognition? Is it a kind of perception, or merely a form of imagination? Regarding these questions, some modern neuroscientists believe that the phenomenon of memory can be reduced to the function of the brain only, while phenomenologists attempt to demonstrate that memory as a mental phenomenon cannot be explained without taking the first-person stance (Bernecker 2008; Sutton, 2016). When we come to Indian philosophy, we also find that these questions are subject to heated debates. The Naiyāyikas argue for the existence of ātman (self) by appealing to memory as cognition of what has been cognized by the same self. However, how far does memory serve as a valid source of knowledge is a disputed issue among various Indian philosophical schools. For instance, though the Naiyāyikas used a memory argument for the existence of ātman, they did not accept memory as valid knowledge (pramā) for the reason that it does not present an object that has not been previously cognized. Accordingly, memory should be reduced to a more fundamental form of cognition. Still they recognize that there is such a thing as veridical memory, and this needs explaining (Chatterjee 1950: 23). It is in this philosophical milieu that the

Vasubandhu’s Theory of Memory

307

Buddhists joined the debate on the epistemological role of memory, which will be examined in this chapter. As we will see below, in the Refutation of the Theory of a Self (Ātmavādapratiṣedha) and Twenty Verses on Consciousness-Only (Viṃśikā) Vasubandhu also deploys an account of memory to defend his idealist metaphysics of nonexistence of self and external world. Vasubandhu as the author of Ātmavādapratiṣedha took a realist stance, which means that while denying the existence of the self (ātman), he did not deny the existence of an external world. The same author in the Viṃśikā changed his stance to idealism and denied the existence of both self and external world. Unlike Janet Gyatso’s edited volume, In the Mirror of Memory in 1992, which is indeed a milestone for its studies on the theory of memory (smṛti) in the tradition of Indo-Tibetan Buddhism, this chapter deliberately reads these texts from the perspective of Chinese commentaries composed by Puguang (普光, died in 664) and Kuiji (窺基, 632–682), two eminent disciples of Xuanzang (玄奘, 602–664). The main aim of this chapter is to show how Vasubandhu argues against realist theories of memory, which affirm either the existence of the self as the owner of memory or the existence of external objects as the support (ālambana) of recollection. Vasubandhu concludes that memory can be explained without presupposing the existence of self (ātman) and external world. This is equivalent to saying that one should be more cautious when memory is employed by the realists, such as the Jains and the Vaiśesikas, as a means of proving the existence of ātman and external world. To recap the arguments in the Ātmavādapratiṣedha and the Viṃśikā as a whole, Vasubandhu’s theory of memory is presented in his critique of the realist’s employing memory to prove the existence of self and external world, arguing that memory occurs in the continuum of consciousness without necessarily presupposing either an ātman (permanent self) as the agent or an external world as object of recollection. In the continuum of consciousness, recollection occurs with the aid of attentive mindfulness (smṛti), so that the “seeds” deposited in the same continuum of consciousness are re-activated to appear as that which has been formerly experienced in sense perceptions. For the realist, memory simply means to retrieve the previous perception of an external object. Without the external object, perception and recollection cannot be caused to arise. Though Vasubandhu agrees that memory should rely on previous perception, he does not agree with the claim that perception implies the existence of external objects. Vasubandhu concludes that the realist fails to prove the existence of external objects through the example of memory. The same argument also applies to the critique of the existence of self as the agent of memory. In the same vein, Kuiji further developed Vasubandhu’s theory by arguing that

308

Lin

memory is possible because mental consciousness (mano-vijñāna) or the mental faculty (manas) functions simultaneously with sensory perceptions to make memory meaningful. Kuiji’s innovative interpretation set the ground for the later development of the theory of memory in East Asia Buddhism. 2

Memory Without the Self: Vasubandhu’s Argument in the Ātmavādapratiṣedha

In the Ātmavādapratiṣedha, in addition to arguments against the Personalists (Pudgalavādins), Vasubandhu criticizes the non-Buddhist theory of ātman held by the Nyāya-Vaiśeṣikas (Duerlinger 2003: 238). According to the Naiyāyikas, memory as knowledge of the past refers to the revival of the latent impressions (saṃskāras) left by past experiences and retained in the self (ātman). Memory arises when mind (manas) is in contact with the latent impressions stored in the self (Chatterjee 1950: 22–23). As we will see, in reply to the Nyāya-Vaiśeṣika theory of memory Vasubandhu did not reject the idea that memory is retained in some form of storage. Instead he replaces the idea of ātman with the idea of the continuum of consciousness, and the idea of impressions (saṃskāras) with the idea of seeds (bījas). This change crucially differentiates two major Indian philosophical schools, for the Buddhist the continuum of consciousness is considered impermanent (being conditioned) while for the Naiyāyikas ātman is permanent (being unconditioned). Structurally speaking, the difference is not as radical as it looks, because both sides attempt to employ the evidence of memory either to prove or to refute the existence of the self. In the end, the main issue turns on the question of the ownership of memory. The epistemological approach to the issue of memory is also seen in the Viṃśikā, a text composed by Vasubandhu after the Abhidharmakośabhāṣya. In the Viṃśikā, Vasubandhu challenges the realist argument for the existence of an external world. The realist who affirms the existence of an external world claims that the experience of sense perception (pratyakṣa) can be employed to prove the existence of external objects, because if the external object does not exist as what is perceived, sense perception would not be possible. For example, I see a jar on the table. The perception as such is possible only if there exist that which sees, i.e., eyes, and that which is seen, i.e., visual form. According to the Buddhist theory of causation, perception results from the contact of the subject of cognition and the object of cognition. Since the cause and the effect cannot reside in the same entity and at the same moment, that which is seen as the cause of perception should not exist in perception as an effect, just as fuel does not exist in fire. Hence the realist proves the externality of the object

Vasubandhu’s Theory of Memory

309

of cognition. That is, a jar must exist as the cause external to the perception of that jar. In the same vein, the realist also claims that memory is possible only if the object of memory is what one has experienced in the past. For instance, I remember seeing the river Seine because I was in Paris last summer. For the realist, though I might recollect the past cognition of the Seine, there must have been the real object the Seine to cause the arising of the cognition in the past. Therefore, no matter whether what I recollect is the past cognition or the object of past cognition, some form of object external to mind must have existed as the cause of the memory. Before coming to Vasubandhu’s reply to the above realist challenge, some technical terms need to be clarified. First, when Vasubandhu talks about memory, he uses “smṛti” in the sense of “no loss (no dropping) of the support of cognition” (smṛti ālambanāsampramoṣaḥ).1 To be noted is that this definition connotes the meaning of “mindfulness” more than “memory,” which can be discerned in the Chinese rendering, nian (念).2 According to the Shuowen 說文, one of the oldest Chinese lexicological texts, composed around the second century, CE., “nian” 念 is defined as “constantly keep in mind,” which is very close to the meaning of “smṛti” as mindfulness.3 The same interpretation is also seen in Puguang’s commentary on AKBh: “The function of smṛti is to clearly sustain [the image of] the object of cognition (ālambana), which can serve as the cause of being not forgotten in the later time. It does not merely mean remembering the past object.”4 It is clear that smṛti does not mean memory in the literal sense; it rather refers to a casual factor associated with a particular attentiveness that allows memory of that experience to arise. According to this theory, people forget most of their trivial experiences, except for a small portion that are remembered. Why so? It is because in the beginning they are not mindful of most objects of experience. One forgets those experiences because smṛti (mindfulness) as the fundamental quality of mind does not properly function. On this issue, whether smṛti is an invariable mental factor or an occasional f­actor is 1 AKBh-X: 念 謂 於 緣 明 記 不 忘 (T29.19a20–21); AKBh-Z: 念 謂 不 忘 所 緣 境 (T29.178b 14–15). 2 I agree with Collett Cox’s rendering “smṛti” by “mindfulness,” which “actually encompasses the psychological functions of memory as they were understood within Indian Buddhism” (Cox 1992: 68). 3 As an editor suggested, one might better render smṛti by “attentiveness” instead of by “mindfulness,” since the meaning of “mindfulness” has been recently overloaded in other contexts. However, I have chosen to retain the term “mindfulness” because it is a widely accepted standard translation. 4 Puguang, jslj: 念 之 作 用 於 所 緣 境 分 別 記 持 ， 能 為 後 時 不 妄 失 因 ， 非 謂 但 念 過 去 境 也 (T41. 74.b. 21–22, emphasis added). Also see Bauddha Kośa, http://www.l.u-tokyo .ac.jp/~b_kosha/html/b_kosha_2014/html/index_75dharma.html#.

310

Lin

subject to dispute in the Abhidharma literature. However, this does not change the character of smṛti as the cause of memory (Jaini 1992: 47). In the Ātmavādapratiṣedha, Vasubandhu divides remembrance into two forms: (1) smaraṇa (yi憶/yinian憶念): the act of remembering or calling to mind, remembrance, reminiscence, recollection; (2) pratyabhijñāna (zhi知/zizhi記知/gengzhi更知): recognition, which means knowing a thing as the same as what was known before. Satischandra Chatterjee explains this notion correctly: “[T]he Buddhists conclude that pratyabhijñā is a dual cognition including both perception and memory which refer respectively to the two aspects of an object as ‘this’ and ‘that,’ or as present and past” (Chatterjee 1950: 206). For example, when someone says, “This is the jar I saw yesterday,” he means “I remember this jar as that jar I saw before.” The meaning of this statement consists of two parts: (1) the perception of the jar, and (2) perceiving this jar as that jar. Owing to the second meaning, the Naiyāyikas consider recognition to be a special mode of conceptual (savikalpaka) perception (Chatterjee 1950: 207).5 This interpretation is also testified to in the Chinese renderings. Xuanzang renders pratyabhijñāna by jizhi 記知, literally meaning “noting what is known or perceived” or simply “recognition.” In comparison, Paramārtha’s rendering by gengzhi 更知 (re-cognition) catches the nuance even more precisely. 2.1 Vasubandhu’s Reply to the Realist Theory of Memory In a section on the issue of memory in Vasubandhu’s Ātmavādapratiṣedha, the ground for the Self Theory of Memory (stm) is stated as follows: The same self that has cognized the past object recognizes the same object at present. If the self, a real enduring subject in time, does not exist, it will be impossible to explain why one remembers and recognizes the same object that was cognized some time ago. For the Non-Self Theory of Memory (nstm), which is held by the mainstream Buddhists, including Vasubandhu as the author of Ātmavādapratiṣedha, cognition and recognition can be adequately explained without assuming the existence of the self as the subject of cognition. They can be explained by referring to the internal mechanism of consciousness. For 5 Chatterjee goes on to explain the meaning of savikalpaka in Nyāya: “All savikalpaka or determinate perceptions of objects consist of certain given or presentative elements and certain representative factors like ideas and images of similar objects experienced in the past. In an ordinary savikalpaka perception the representative factors do not remain distinct but are associated with the presentative elements to make up one percept. In pratyabhijñāna or recognition, the representative factor has the form of a definite recollection of some past experience of an object and modifies the present perception of it. Still it is perception, since it is brought about by sense-object contact” (Chatterjee 1950: 207). Also see Wayman 1992: 142–143.

Vasubandhu’s Theory of Memory

311

nstm, nothing exists permanently as the owner of consciousness. On the contrary, stm challenges: If consciousness exists momentarily, how could C2 at t2 recollect the same object of C1 at t1 when C1 has ceased to exist? Since C1, C2 … and Cn+1 exist momentarily, they cannot be taken as the ground that accounts for the continuity of the experience of the same subject. If it is not the same enduring self who owns the series of mental episodes, an absurd consequence will be entailed: that the C1 of John can be recollected in the C2 of Paul.6 Hence for stm the self must be taken as the best candidate for owner of the memory.7 2.2 Vasubandhu’s Causal Theory of Memory Replying to the Naiyāyikas, Vasubandhu offers his own account of memory: Remembrance and recognition as such arise from a special [power of] mind (cittaviśeṣa) which is associated with ideation (saṃjñānvaya) of the object that was mindfully intended (smṛtiviṣaya).8 In Xuanzang’s Chinese translation this dense sentence is glossed to make it a bit more transparent, yet as shown in Puguang’s commentary, the main compound smṛtiviṣayasaṃjñānvaya is open to various interpretations. My own interpretation of this sentence basically follows Xuangzang’s and Paramārtha’s translations and Puguang’s commentary: When one is mindfully aware of an object, the perception will be turned into an idea or image. The mental act of forming an idea or image is called “ideation” (saṃjñā, xiang 想), which is defined in the AKBh as “grasping the marks of the object” (saṃjñānaṃ 6 The example raised by the realist is rather similar to that which Derek Parfit employs in the fission argument. 7 āvp: yadi tarhi sarvathāpi nāsty ātmā kathaṃ kṣaṇikeṣu citteṣu cirānubhūtasyārthasya smaraṇaṃ bhavati pratyabhijñānaṃ vā. p. 122; AKBh-X: 若 一 切 類 我 體 都 無 ， 剎 那 滅 心 於 曾 所 受 久 相 似 境 何 能 憶 知 。 (T29.156c26–27). 8 smṛtiviṣayasaṃjñānvayāc cittaviśeṣāt. Xuanzan’s Chinese translation: [如 是 憶 知 從 相 續 內 ]念 境 想 類 心 差 別 生 . My translation is based on Xuanzang’s translation and Puguang’s commentary, which is different from Duerlinger’s and Poussin/Gelong Kodrö Sangpo’s translations. Cf., Duerlinger’s translation: “[A]n object is remembered because immediately before the memory [of it] occurs a special kind of mind arises that is [causally] connected to a [prior] discrimination of the object to be remembered” (Duerlinger 96); Poussin/Gelong Kodrö Sangpo’s translation: “Memory (or reminiscence) [smṛti] (and recognition [pratyabhijñāna]) arise, in a stream, immediately after a special kind of thought which is [causally] connected with the ideation of formerly perceived object to be remembered” (Poussin/Gelong Kodrö Sangpo iv, 2558); Takeda’s translation: “念 (smṛti)の 境 に つ い て の 想 の 流 類 (smṛtiviṣaya-saṃjñānvaya)を 有 す る 特 殊 な 心 (cittaviśeṣa)か ら [想 起 や 再 認 知 が 生 じ る ]” (Takeda 2007: 2007). According to Takeda’s translation, “saṃjñānvaya” is understood as the type of ideation that is causally derived from the previous moment of ideation.

312

Lin

viṣayanimittodgrahaḥ).9 Here Puguang provides a lively example for us to make sense of saṃjñā: When one sees someone on the street, he knows this one is male, not anyone who is not male, through perceiving certain marks, like body shape, beard, dress, and so forth. By the same token, one is able to recognize a woman as woman through perceiving certain other marks. The distinction of objects is thus made by the act of saṃjñā, the mental act of forming an image in the cognitive process from sensation to ideation.10 The same type of ideation (saṃjñānvaya, xianglei 想類) is deposited in the causal continuum of consciousness, which is also called “bīja” (seed) in the Abhidharmakośa, though Vasubandhu did not explicitly mention this term in this context.11 It is in this mental process that later recollection as a mental phenomenon develops from the ideas which are stored in the mental continuum in the form of seeds (bīja). The ideas are produced through ideation of those mindfully intended objects. If the perception of the object does not occur attentively with mindfulness in the first place, it would not leave the idea in the consciousness. Attempting to unpack the dense meaning of the compound “smṛtivi­ ṣayasaṃjñānvaya,” Puguang recorded three options for interpretation. The first interpretation is said to represent Vasubandhu’s view, according to which recollection and recognition arise from the seeds of ideation only.12 The second 9

saṃjñā saṃjñānaṃ viṣayanimittodgrahaḥ. Pradhan, 1967, 54, 20–21, Chapter ii v.24a; AKBh-X: 想 謂 於 境 取 差 別 相 (T29.19a18). 10 Puguang, jslj: 想 謂 於 境 取 差 別 相 者 。 想 謂 於 境 執 取 男 女 等 種 種 差 別 相 ， 能 於 境 中 封 疆 畫 界 。 此 是 男 等 ， 非 非 男 等 ， 故 名 男 等 (T.41.74a24). 11 anvaya, m. ( √ i » anv- √i) , means “race,” “lineage,” and “family” among other usages (cited from the Monier-Williams Online Sanskrit Dictionary). In this text it is rendered by “lei” ( 類 ) in both Paramārtha’s and Xuanzang’s translations. I take it to mean the type of ideas or images as the result of ideation. When a certain type of idea is stored in the mental continuum, it is called “seed.” 12 Puguang, jslj: 如 是 憶 念 ， 如 是 記 知 ， 從 自 相 續 內 ， 有 念 境 想 熏 成 種 子 ， 名念境想類。此種在心功能差別，名心差別。後之憶知從此念境想類 種子心中差別功能而生。經部念知無別有體，故想種生. T41.447a6–10. “Recollection (smaraṇa) as such and recognition (pratyabhijñāna) as such [can be explained as follows:] The ideation (saṃjñā) of the object (viṣaya) of mindfulness (smṛti) is perfumed to become a seed (bīja) in one’s own [psycho-physical] continuum, which is called “the type of ideation of the object of mindfulness” (smṛtiviṣayasaṃjñānvaya). The seed in the mind has a distinct power, which is called “the distinct power of mind” (cittaviśeṣa). At a subsequent time, recollection and recognition arise from the special power in the mind which contains the seeds of ideation of the object of mindfulness. According to the Sautrāntika, recollection and recognition are not different entities. Hence [they hold the view that both recollection and recognition] arise from the seed of ideation (saṃjñā).”

Vasubandhu’s Theory of Memory

313

view holds that recollection and recognition arise from the seeds of mindfulness and from the seeds of ideation respectively.13 The second reading might be based on Vasubandhu’s classification of dharmas in which ideation (saṃjñā) and mindfulness (smṛti) are separate mental factors. According to the Abhidharma theory of dharmas, ideation and mindfulness should not be confused with each other. For this interpretation, there are two kinds of seed, the seed of ideation and the seed of mindfulness, which should also be separated. Accordingly, the same object of recollection and recognition consists of two aspects, one aspect corresponding to the mindful perception and another aspect corresponding to ideation. In other words, the same object of recollection and recognition has dual aspects: one is the particular, which is intended with the mindful perception, while the other is its property, which can be differentiated through ideation. This explains why, as the result, the object of recollection can be re-cognized. Unlike the second interpretation, which holds that recollection and recognition correspond to the dual aspects of the same object, namely, the phenomenal aspect of the object developed out of the seeds of mindfulness and another phenomenal aspect of the object developed out of the seeds of ideation, the third theory seems to hold that both recollection and recognition arise from the seed of mindful ideation. There is no distinction of the seed of mindfulness and the seed of ideation. Hence it is named “seed of mindfulnessideation.”14

13 Puguang, jslj: 又 解 :如 是 憶 知 從 相 續 身 內 念 境 類 種 、 想 境 類 種 ， 境 通 兩 處。此文但應言念類種，而言想者，想強別標。所以不言知境類者， 知 由 念 引 ， 故 不 別 言 。 故 下 論 云 :由 此 憶 念 力 有 後 記 知 生 。 T41.44a10–14. “Regarding the second interpretation, recollection and recognition as such [arise] from the seed qua the type of the object of mindfulness (smṛtiviṣayānvaya) and the seed qua the type of the object of ideation (saṃjñāviṣayānvaya) within one’s body as a continuum. The object [of recollection and recognition] corresponds to two locations [of seeds]. This text should mention the seed qua the type of mindfulness [only because smṛti is the topic for discussion]. It also mentions [the seed qua the type of] ideation, because ideation is strong enough to assign a special role for it. The reason why the type of the object of recognition is not mentioned is that recognition is induced from mindfulness. Therefore it is not mentioned separately. Hence in the following part of the treatise it says: ‘Recognition thereafter arises due to the force of recollection.’” 14 Puguang, jslj: 又 解 :憶 念 從 念 境 類 生 ， 記 知 從 想 境 類 生 。 以 經 部 知 無 別 體 故，此念想種熏在心中差別功能，名心差別。現行憶念及與記知從彼 種 生 (T41.44a15–18). “There is a further interpretation. Recollection arises from the type of the object of mindfulness, while recognition arises from the type of the object of ideation. Since the Sautrāntika holds that [recognition] is not an entity separate [from recollection], mindfulness and ideation are perfumed and become the seed in the mind, while

314

Lin

The three interpretations recorded in Puguang’s commentary are varied with regard to the nature of the seeds stored in the continuum of consciousness. According to the third interpretation, a seed as the impression left in the consciousness is the result of the ideation of the object that has been mindfully intended. That means the same seed preserves both perceptual and conceptual content of the past cognition. What one remembers and recognizes cannot be only bare perception without distinctive content. For example, when one remembers a man and recognizes him as John, memory as such must involve ideation to make possible knowing someone or something as X. This interpretation fits ordinary experience well.15 As for the issue of memory in the everyday sense, Vasubandhu explains that recollection and recognition develop out of the seeds stored in the continuum of consciousness, while seeds are formed by the ideation of something that has been the intended object of mindfulness.16 Vasubandhu goes on to explain the conditions for memory to arise. We have to note that underlying Vasubandhu’s interpretation are the principle of causation and the principle of momentariness, which should be either agreed upon by the two parties of the debate or subject to further justification. As we will see later, the stm theorist tries to argue for the existence of the self through causal reasoning from action (effect) to agent (cause). Vasubandhu, as seen above, has pinned down the special power of mind (cittaviśeṣa) as the main cause of memory. In addition to cittaviśeṣa as the main cause, Vasubandhu lists four other conditions for memory to arise:17 (1) The intending of the past object to be

15 16

17

the distinct power [of the seed] is called ‘distinctive power of mind.’ On the level of mental phenomena, recollection and recognition arise from this seed.” However, if one has pure perception only, such as the Buddha’s perception, can his cognition be “perfumed” to become a seed? This question has been open to debate for a long time in the Chinese Yogācāca literature. After reviewing the three interpretations, I favor the third. I do not wish to dive into the doctrinal disputes carried on by Kuiji and his disciples in their commentaries on the Cheng Weishi lun. Briefly speaking, the whole controversy is centered on the relation between mental phenomena and seed. Since memory is divided into recollection and recognition, logically speaking, seed should be also divided into two types. However, the question is pressed further: In what sense is recollection said to be distinct from recognition? Are they two cognitive events, or are they merely two aspects of the same cognitive event? Likewise, should the seed of memory also be divided into two kinds, the seed of mindfulness and the seed of ideation? Or does the nature of memory seed contain the function of mindfulness and ideation? These questions force the Chinese Buddhist scholastics to debate on the issue of memory in the context of Yogācāra idealism, which needs to be treated separately. My reading is based on Xuanzang’s Chinese translation, Puguang’s commentary, and Gelong Lodrö Sangpo’s English translation (2012: 2558–2559).

Vasubandhu’s Theory of Memory

315

remembered. This description indicates that Vasubandhu considers intentionality to be one of the phenomenological features of memory.18 (2) Similarity between the object of ideation (saṃjñā) and the object to be remembered.19 The object of ideation should be similar to the object of remembrance, and this factor is called “similarity.” This is because the object of both ideation and remembrance refers to the ākāra (representation) of a mental act, such that the ākāra at t1 is similar to the ākāra at t2. (3) Association (saṃbandha) of ideation with the object to be remembered. (4) Memory arises only when the mind is not affected by sickness, sorrow, mental distraction, etc. Regarding the phenomenological nature of the Buddhist theory of memory as shown in the above auxiliary conditions, Paul J. Griffiths once stated that “[t]he Yogācāra account of memory, and indeed Buddhist accounts in general, … have, whatever their conceptual advantages and disadvantages, little to offer those interested in the phenomenology of remembering” (Griffiths 1992: 123). The reason for this conclusion is that, for Griffiths, phenomenology of remembering makes no sense if there does not exist one who remembers. In his own words, “One reason for this lack of interest in the phenomenology of remembering is clear: a key element in any account of it must be the sense of identification that the subject of the memory feels with the subject of the event remembered” (122–123). According to the distinction of stm and nstm, he seems to argue that only stm (which of course includes the more specific theory that it is the soul that remembers) can offer the phenomenological first-person account of remembering. If the self is denied, for Griffiths, memory loses existential meaning for someone who remembers. I disagree with this interpretation. Though the self is denied in Buddhist doctrine, it does not mean that the first-person approach is also denied. On the contrary, in Buddhist philosophy, in Yogācāra in particular, the continuum of consciousness is considered most pivotal in the formation of saṃsāric experience. While the self is denied, the ego-consciousness in the same stream is rather (wrongly) viewed as the subject to which all experiences are fixed. The 18

19

By “phenomenology” I mean a methodology of explaining experience through bracketing metaphysical claims about the existence of self and external world. Experience in phenomenology is considered that which has always already been experienced in consciousness. Whether this type of methodology leads to the phenomenology of reflexive self-awareness (svasaṃvedana) is another question. Although early Vasubandhu did not develop the theory of reflexive self-awareness as seen in Dignāga’s philosophy, nevertheless he is consistent in employing a method of bracketing to explain the nature of experience. His account of memory is no exception. This is Gelong Lodrö Sangpo’s interpretation. Puguang provides a different interpretation: “‘Similarity’ is named for the likeness of the object of the past and the object of mindfulness. Because of the force of the similarity of objects as the condition, memory arises” (過 去 境 界 與 念 境 等 名 為 相 似 ， 由 彼 相 似 境 界 力 故 為 緣 生 念 ).

316

Lin

experience of memory is no exception. Furthermore, for Buddhism, memory should not be merely taken as the source of existential meaning. Conversely it always ties us to the karmic past from which we should be free. Memory is tied to a fabricated subject that is often mistakenly taken as a true self. 2.3 The Issues of Agency and Ownership of Memory For those who hold stm, the most crucial concepts are those of the agency and ownership of memory. If the self does not exist, while memory occurs momentarily only in a continuum of consciousness that lacks any basis for a proper account of personal identity, absurd consequences would result, such as Yajñadatta recollecting what Devattada has experienced. In reply, Vasubandhu restates his theory of personal identity, claiming that “personal identity” is a name for the psychophysical continuum (saṃtāna) in which moments of thoughts are causally connected without reliance on any metaphysically assumed substance, whether called “self” or “person.” Vasubandhu goes on to employ a causal theory of “transformation of the stream” (saṃtatipariṇāmaviśeṣa) to explain memory, a theory discussed above. He argues that the theory of mental causation in the transformation of the stream of mind-body is enough to explain the experience of memory.20 However, those who hold the view of stm will consider nstm counterintuitive. They ask: If a self does not exist, who remembers? If a self does not exist, whose is this memory? Those who hold stm cannot accept Vasubandhu’s account of memory unless they are satisfied that Vasubandhu’s theory properly explains the issues of agency and ownership of memory. For stm, there should be someone who remembers, just like there should be a doer who does. Hence stm insists that the self is the one who recollects. In reply, Vasubandhu argues that, though in everyday discourse people say, “Caitra remembers,” “Caitra” is in fact merely a conventional name for a mindbody continuum. What really cause memory to arise are elements in this continuum, not a mere name. For Vasubandhu, a real agent of memory is the special power of mind (cittaviśeṣa) that causes memory to arise. It is to be noted that Vasubandhu never denies the pragmatic use of the firstperson pronoun and proper names. What he warns us of is the metaphysical 20

Vasubandhu defines the meaning of “saṃtatipariṇāmaviśeṣa” on the theory of seeds (bīja): “Stream” or “continuum” (saṃtati) is the conditioning forces (saṃskāra) of the past, present and future that causally constitute an uninterrupted life-stream. “Transformation” (parināṃa) means the causal change of the stream between the prior and subsequent moments. “Viśeṣa” (difference) refers to the capability of producing the effect. The notion of this compound is central to Vasubandhu’s theory of mind in his early and late stages. Cf., Gelong Lodrö Sangpo’s English translation (2012: 544). For Xuanzang’s Chinese translation, see T29.22.c11–15.

Vasubandhu’s Theory of Memory

317

impulse underlying everyday discourse that often moves beyond the boundary of language. For Vasubandhu, it is a fundamental mistake to identify something not existing as the referent of the proper name. What truly matters is the existence of mind and body (five aggregates). Among them, the role of consciousness is the most crucial in the theoretic account of memory. Regarding the ownership of memory, stm argues that as the statement “Caitra is the owner of cow” or “Caitra owns his cow” shows, there must exist the owner of memory. In reply, Vasubandhu impressively analyzes the above statement step by step to conclude that “Caitra” as a name for a mind-body continuum denotes nothing but a collection of conditioned entities (saṃskārasamūha), and “cow” as another name for another stream denotes a collection of conditioned forces too. Under the law of mental causation, two mental streams can be causally connected. In this case, Caitra is not a self (ātman) who owns or controls his cow. What owns cows is nothing but a mind-body continuum. By implication, there is no self as the owner of memory. In reality, memory is “owned” by the cause, i.e., citta-viśeṣa (special power of mind). Vasubandhu concludes: “Here there is no single real entity called ‘Caitra’ or ‘cow.’ But the Caitra-stream has no relation of ‘owner’ or ‘master’ (svāmibhāva) and ‘what is owned’ other than that between a cause and its effect (hetubhāva)” (Gelong Lodrö Sangpo 2012: 2562). Summing up, in the Ātmavādapratiṣedha Vasubandhu denies the self (ātman) as the owner or the agent of memory. Memory can be satisfactorily explained just by the idea of the causal continuity of mental streams. At this stage, Vasubandhu still stood for representational realism (i.e., Sautrāntika) before he changed his metaphysical position to idealism as seen in the Triṃśikā and the Viṃśikā. 3

Memory Without the External World: Vasubandhu’s Argument in the Viṃśikā

In the Viṃśikā, the Buddhist and non-Buddhist realist opponents use a memory argument, in a context of epistemological analysis, to prove the existence of the external world:21 1) First, both realist and idealist share a common epistemological premise: “Whether an object exists or not is determined by the valid means of cognition. Among all means of valid cognition, perception is recognized as 21

For the following summary of Vasubanhdu’s arguments, see the verses 16 and 17ab of Viṃśikā as well as Vasubandhu’s self-commentaries. Also cf. Xuanzang’s Chinese translation (T31.76.b15–c4) and Stefan Anacker’s English translation (2005: 170–171).

318

Lin

the most important.”22 This universally accepted premise claims that the problem of existence should be answered only through accredited knowledge sources. This premise leads to another working principle: Existence can be either perceived or inferred through a valid means of cognition. 2) Since both parties agree that perception is most fundamental among all sources of knowledge, it should be analyzed first. For the realist, perception cannot occur without what is perceived or inferred as the cause of perception, which should exist outside of perception. Hence, the object of perception exists as being outside of perception. In refutation, Vasubandhu provides the following counter-arguments: 1) In dreams perception exists without the external object too. According to the rules of the Indian syllogism, this single example is sufficient for falsifying the realist claim about the existence of an object external to perception.23 2) In view of the theory of momentariness, which is also accepted by both parties, when perception arises, the known object and the knowing subject have already ceased to exist. So, again, no external object can be known in the present moment of perception. 3) The cognitions, “This is my perception” and “This is the external object” are the result of the conceptual judgment (pariccheda) through the mental consciousness (mano-vijñāna).24 In fact, there is perception only. The adjectives “my” and “external,” which result from the conceptual discrimination of mind, are simply superimposed on perception. Below we see how the realist employs the memory argument to prove the existence of an external world:

22

pramāṇavaśād astitvaṃ nāstitvaṃ vā nirdhāryate sarveṣāṃ ca pramāṇānāṃ pratyakṣaṃ pramāṇaṃ gariṣṭham. VS; 諸 法 由 量 刊 定 有 無 ， 一 切 量 中 現 量 為 勝 (T31.76. b15–16). 23 When the realist uses dream as a similar example to prove the existence of external object, Vasubandhu the idealist also uses dream as a similar example to prove the opposite thesis, i.e., non-existence of external object. The strategy of argument adopted by Vasubandhu here is a form of reductio ad absurdum. 24 Kuiji, wseslsj: “As said in the Cheng weishi lun (Vijñaptimātratāsiddhi), one does not take [the object] as external in direct perception. It is rather at the later moment that the thought of externality arises delusively through discrimination of the mental faculty (manas). For this reason, the seeing and the seen result from the transformation of consciousness, which can be viewed as existence too; that which is grasped by the mental consciousness (mano-vijñāna) is illusive construction, which is said to be non-existent” (成 唯識說，現量證時，不執為外。後意分別，妄生外想。故自相分，識 所 變 故 ， 亦 說 為 有 。 意 識 所 執 ， 妄 計 度 故 ， 說 之 為 無 ) (T43.983.a5–8).

Vasubandhu’s Theory of Memory

319

1)

Recollection (smaraṇa) is made possible through mental consciousness under the condition that the object must have been experienced first in perception. The reason is that, if there is nothing experienced out there, what will be the object of recollection? 2) Recollection is a fact. 3) Hence, the existence of an external object is proven through recollection. Vasubandhu’s refutation can be summarized as follows: 1) Recollection occurs when mental consciousness arises together with mindfulness (smṛti) to trigger the act of recollection. 2) It is not the case that what was perceived in the first place is the external object, because, as demonstrated elsewhere,25 what was perceived is the mere mental appearance of an object. 3) Therefore, what one recollects is the experience of the mere mental appearance of an object, not an external object. Vasubandhu notes that in Verse 9 of the Viṃśikā he has explained how the appearance of cognitive subject and cognitive object develops out of the seeds (bīja). In contrast to the representation theory in the Ātmavādapratiṣedha, Vasubandhu shifts his stance to idealism in Viṃśikā, claiming that the objects of cognition are developed out of the seeds that are deposited in the stream of consciousness.26 As a logical consequence, no existence of external objects can be inferred through recollection, nor can any external objects be taken to serve as the causal condition of memory. According to Kuiji’s commentary on the Viṃśikā, which did not differentiate its metaphysical standpoint from the Triṃśikā, Vasubandhu’s idealist theory of memory can also be reconstructed as follows: 1) Five types of sense perception arise from the contact of the sense faculties and the sense objects that are nothing but the mental appearance developed out of the seeds. For Yogācāra, all experiences occur within the internal loop of consciousness only. That is, no epistemic conditions external to consciousness are required to explain the experience of cognition, including memory. In this regard, Vasubandhu takes a radical idealist turn in his account of memory. 2) Recollection occurs in mental consciousness; the latter is further divided into two modes: (1) diachronic mode of mental consciousness (dm) and

25 26

Vasubandhu developed a full-blown theory of consciousness-only (vijñapti-mātra) in the Triṃśikā, To claim that Vasubandhu takes an idealist stance in the Viṃśikā is subject to debate. Here I simply follow the traditional interpretation as seen in East Asian Buddhism.

320

Lin

(2) synchronic mode of mental consciousness (sm).27 dm works at the next moment after sense perception. Its main function is to conceptualize the content of perception. As a result, the object of dm is a concept or conceptual image left by the previous perception. However, this cannot fully explain the experience of memory, because in memory the object of recollection is not a mere concept, but full of sensory content. For example, I can vividly remember the taste of coffee I enjoyed in Vienna last summer. I do not merely remember the idea of coffee taste. So, the problem is how mental consciousness at the later time produces memory of the perceptual contents of the earlier experience. According to Kuiji’s Commentary, Vasubandhu claims that in the past, sensory perception functioned simultaneously with the mental consciousness to produce a seed. The seed thus produced by both sensory perception and mental consciousness contains both the perceptual and the conceptual contents due to the synchronic function of both sensory perception and mental consciousness. While the seeds are developed to manifest as the object of recollection, it is the mental consciousness that is capable of recollecting the object in the previous experience. In conclusion, it is mental consciousness that remembers. In the Viṃśikā 17b, Vasubandhu explains how memory arises in the mental consciousness: “‘From that’ means ‘from the consciousness.’ A mental consciousness arises with the discrimination of a visible, etc. when that appearance is linked with memory” (tato hi vijñapteḥ smṛtisaṃprayuktā tatpratibhāsaiva rūpādivikalpikā manovijñaptir utpadyate).28 Let’s closely read Kuiji’s commentary on this passage: “This means, at the present succeeding stage of the previous five [sensory] perceptions which have cognized the appearance (ābhāsa) of the object, when associated with the mindfulness (smṛti) of particular objects, the discriminative mental consciousness (manovijñāna), which apprehends the past as its object, transforms [the past object] to appear as like the support of cognition (ālambana) of the previous five sensory perceptions. It is not the case that the five sense perceptions, such as visual perception, etc., experienced the object separate from mind first and 27 28

In Kuiji’s commentaries, SM is either called “tongshi yishi” (同 時 意 識 ) or “tongshi yi” (同 時 意 ). On other occasions it is either called “wuju yishi” (五 俱 意 識 ) or “wuju yi” (五 俱 意 ). For discussion on this concept, also see Ching Keng’s paper in this volume. Anacker 2005: 171, with slight modification; Xuanzang’s translation: 從 此 後 位 與 念 相 應 分 別 意 識 似 前 境 現 ， 即 說 此 為 憶 曾 所 受 (T.31.76c2–3).

Vasubandhu’s Theory of Memory

321

thereafter arises a mental consciousness, which being associated with the mindfulness (smṛti), apprehends the external object of the previous five sense perceptions.”29 According to Kuiji’s interpretation, “memory of past experience” means that, for instance, when one remembers a friend whom he met some time ago, he seems to see him as a real visual object in recollection. In reality, this friend as a visual object in recollection is nothing but the mental appearance that results from the function of the mental consciousness. Furthermore, the mental appearance in remembrance is developed out of the seeds that were “perfumed” by the mental faculty (manas) that had simultaneously accompanied the five sense perceptions, while the object of sensory perception in the past is the mental appearance too. Kuiji goes on to explain: In this [connection] the discriminative mental consciousness apprehends the object that has been present to, while not separate from, the consciousness. This is called “the consciousness of remembering what has been experienced.” What does it mean? Through the past five [sense] perceptions and the synchronic mental faculty (tongshi yi 同時意) one apprehended the object that was not separate from consciousness; [the cognition as such] was perfumed to become the seeds. [The seeds] continued to develop up to the present stage. At this stage the mental faculty (manas) is able to recollect the object of the past. This is called “recollection.” It is not the case that the cognition of the past experience is thus named because the object of the past five [sense] perceptions was necessarily separate from the mind whereas it is apprehended by the mental consciousness at the present.30 Kuiji’s interpretation is clearly based on the idealist premise, which holds that recollection of the past object will be impossible if the object exists outside of sensory perception. According to Kuiji, when S1 and O1 appear as the conditions of cognition, C1 arises while the mental consciousness also simultaneously 29 Kuiji, wseslsj: 謂 從 過 去 似 境 五 識 ， 今 此 後 位 與 別 境 念 相 應 之 時 ， 有 緣 過 去分別意識，變似前五識所緣境現，無曾現在受離心境，眼等五識， 從 此 今 時 ， 與 念 相 應 ， 有 一 意 識 ， 緣 前 五 識 離 心 之 境 (T43.1000.c22–26). 30 Kuiji, wseslsj: 於 此 分 別 意 識 緣 曾 現 在 不 離 識 境 ， 名 為 憶 持 曾 所 受 識 。 所 以 者 何 ?由 曾 五 識 及 同 時 意 ， 緣 即 識 境 ， 熏 成 種 子 。 今 時 相 續 ， 意 於 此 位，能憶前境，名為憶持。非曾五識境定離於心，今時猶有意識緣 之 ， 名 曾 受 識 (T43.1000c27–1001a3).

322

Lin

perceives the same object (O1).31 As mentioned above, mental consciousness (mano-vijñāna) can be divided into two types: diachronic mode of mental consciousness (dm) and synchronic mode of mental consciousness (sm). This distinction seems to appear only in the Chinese Yogācāra literature. In the cited passage of Kuiji’s commentary, he takes “synchronic mode of mental faculty” (tongshi yi) as the agent of recollection. In the same passage, manas appears twice to account for the process of recollection. Obviously his usage of “synchronic manas” is not a corrupted phrase of “synchronic mode of mental consciousness” (wuju yishi, 五俱意識/tongshi yishi, 同時意識). The crucial point is to find the reason for specifically identifying synchronic manas as the basis for memory to occur. In the Abhidharma literature, manas (mental faculty, mind) is referred to as that which functions as the supporting basis (āśraya) for the arising of the subsequent consciousness.32 However, this supporting basis cannot be understood as something that exists in the form of a substance. On the contrary, it may be suitable to interpret it as a space for allowing certain kinds of mental factors, such as smṛti (mindfulness), saṃjña (ideation) and prajñā (understanding), to come into play with the sense perceptions simultaneously. Precisely in this context the notion of manas is designated as a theoretic toolkit for explaining the capacity of discriminative thinking, including recollection. It is incorrect to take manas as some sort of Cartesian mental substance that serves as the faculty of thinking. On the contrary, it might be better to understand it as a special space in an internal loop of consciousness in which the seeds of the previous cognition can be retrieved from the storehouse consciousness to appear as the object of re-cognition, while some mental factors like understanding (prajñā) and ideation (saṃjñā) join to work simultaneously.33 If this interpretation makes sense, then memory should not be considered only as a sort of sensory representation. It is a karmically reconstructed mental perception whereas manas is taken as the baseless basis of memory.

31 32

33

Here, I adopt Hattori’s symbols: C1 = S1-O1; C = Cognition; S = Subject; O = Object. See Hattori 1968: 108. AKBh-X: “Manas means the incessant cessation of the group of six types of consciousness. That is, it is called the realm of manas because it can give rise to the subsequent consciousness when the group of six types of consciousness have been ceased.” (由 即 六 識 身 無 間 滅 爲 意 ， 即 六 識 身 無 間 滅 已 ， 能 生 後 識 故 名 意 界 ) (T29.4b3–6). Here I would not like to go too far in claiming that Vasubandhu in the Viṃśikā takes a Kantian position. I think in many respects his philosophy is still styled in naturalistic thinking. However, it is quite certain that the notion of manas in the vijñaptimātra system cannot be understood from the perspective of naturalism.

Vasubandhu’s Theory of Memory

323

Note that to interpret manas as a space does not necessarily preclude characterizing it as a faculty of thinking. By definition, “faculty” in English means “an inherent mental or physical power,” which exactly matches the usage of indriya in Sanskrit. For both Vasubandhu the realist and Vasubandhu the idealist, manas is also powered by the residual karmic seeds that are stored and evolve in the storehouse consciousness. For Buddhist philosophy, manas never functions as pure thinking. It is always already embodied in karmic historicity and linguisticality. Right within this space the powers of karmic seeds and epistemic seeds function. In the case of memory, recollection and recognition occur in the mental consciousness while the manas simultaneously works together with the sensory consciousness. This theory perfectly accounts for memory as a meaningful experience, for memory would not be meaningful without the karmic function of manas.34 4 Conclusion As we have seen, in Vasubandhu’s theory of memory developed from the Ātmavādapratiṣedha to the Viṃśikā, he specifies the role of mental consciousness (mano-vijñapti) in association with mindfulness (smṛti) in causing the arising of a mental appearance (pratibhāsa) that is copied from a prior perception. He did not explain further about the function of mental consciousness in details. Taking this as a clue, Kuiji moves further to claim that it is due to the synchronic manas, which is able to function together with sensory perceptions, that one cognizes with the various mental factors the mental objects that are not separate from cognition; the cognitions as such are further copied to become the seeds that continue to develop into appearances in the enduring process. According to this Yogācāra line of analysis, recollection is possible not only owing to the function of mental consciousness simultaneously in association with smṛti, but also owing to the fundamental nature of cognition, that is, 34

In Asaṅga’s Mahāyānasaṃgraha, manas is divided into two kinds: (1) the immediately ceased consciousness of the previous moment that serves as the diachronic basis for the succeeding consciousness to arise, and (2) defiled manas (kliṣṭaṃ manas) that serves as the synchronic basis of consciousness. In Vasubanhdu’s Triṃśikā, kliṣṭa-manas is also regarded as ego-consciousness that is always associated with four afflictions: view of self, confusion of self, pride of self and love of self. In this context, Kuiji’s notion of “synchronic manas” as the agent of memory makes more sense if it is understood as the manas that is the synchronic basis, namely as kliṣṭa-manas. This can be seen as Yogācāra’s option to account for the agent of memory. For a study of kliṣṭa-manas and mana-indriya, see Odani 2000.

324

Lin

memory is possible because one or more of the five sensory perceptions occurs simultaneously with mental consciousness; the latter is able to organize the discrete sensory experiences as a meaningful event. These experiences will be retained in the form of seeds in the continuum of consciousness, which is called storehouse consciousness, and reactivated later through the function of attentive smṛti in the mental consciousness. According to Kuji’s interpretation, the synchronic manas is the agent of remembrance. Why is manas able to remember the past event? It is because manas occurs simultaneously with sensory perceptions. This is the reason why we can remember in the mental consciousness the taste of coffee sipped in Vienna last summer.35 This also explains why the content of remembrance is always meaningful instead of being bare episodes of sensation. Acknowledgments This chapter is greatly indebted to the editors’ and anonymous reviewers’ invaluable comments. I especially thank Mark Siderits for his meticulous comments and patience that helped me substantively to finalize this work. Primary Sources and Abbreviations AKBh: Pradhan, Prahlad. 1967. Ed., Abhidharma koshabhāṣyaṃ of Vasubandhu. K.P. Jayaswal Resear. AKBh-X: Vasubandhu. Apidama jushelun (阿 毗 達 磨 俱 舍 論 ). Tr. by Xuanzang (玄 奘 ). T29. No. 1558. AKBh-Z: Vasubandhu. Apidamojusheshilun (阿 毗 達 磨 俱 舍 釋 論 ). Tr. by Zhendi (真 諦 ). T29. No. 1559. AKH: Kaidō (快道), Abidatsuma kusharon hōgi 阿毘達磨倶舍論法義. T64. No. 2251. āvp: Lee, Jong Cheol. 2005. Ed. Abhidharmakośabhāsya of Vasubandhu, Chapter ix, Ātmavādapratiśedha. Tokyo: Sankibo. jslj: Puguang (普 光 ), Jushelunji (俱 舍 論 記 ). T41. No. 1821. jsls: Fabao (法 寶 ), Jushelunshu (俱 舍 論 疏 ). T41. No. 1822. wsesl: Vasubandhu. Weishi ershi lun (唯 識 二 十 論 ). Tr. by Xuanzang. T31. No. 1590. 35

According to Kuiji, sm is pratyakṣa (perception). See Kuiji, yjsdllz: “As stated in the Pramāṇasamuccya and the Nyāyamukha, five kinds of sensory consciousness are perception only; the same for mental consciousness” (如 集 量 理 門 論 云 。 五 識 唯 現 量 。 同 時 意 識 亦 爾 ) (T43.198c6–7).

Vasubandhu’s Theory of Memory

325

wseslsj: Kuiji (窺 基 ). Weishi ershilun shuji (唯 識 二 十 論 述 記 ). T43. No. 1834. yjsdllz: Kuiji (窺 基 ). Yiujiashidilun luezuan (瑜 伽 師 地 論 略 纂 ). T43. No. 1829.

Secondary Sources Anacker, Stefan. 2005. Seven Works of Vasubandhu. Delhi: Motilal Banarsidass. Arnold, Dan. 2012. Brains, Buddhas, and Believing: The Problem of Intentionality in Classical Buddhist and Cognitive-Scientific Philosophy of Mind. New York: Columbia University Press. Bauddha Kośa. Bukkyō yōgo no yōreishū (Bauddakōsha) oyobi gendai kijun yakugoshū (仏 教 用 語 の 用 例 集 (バ ウ ッ ダ コ ー シ ャ )お よ び 現 代 基 準 譯 語 集 ). Accessed at http://www.l.u-tokyo.ac.jp/~b_kosha/html/b_kosha_2014/start_index.html. Bernecker, Sven. 2008. The Metaphysics of Memory. Dordrecht: Springer Science + Business Media B.V, 2008. Bernecker, Sven, 2010. Memory: A Philosophical Study. Oxford: Oxford University Press. Bhikkhu KL Dhammajoti. 2007. Abhidharma Doctrines and Controversies on Perception. Hong Kong: Centre of Buddhist Studies, The University of Hong Kong. Chadha, Monima. 2014. “A Buddhist Explanation of Episodic Memory: From Self to Mind,” Asian Philosophy, 24/1: 14–27. Chatterjee, Satischandra. 1950. The Nyāya Theory of Knowledge. Calcutta: University of Calcutta, 1950. Cox, Collett. 1992. “Mindfulness and Memory: The Scope of Smṛti from Early Buddhism to the Sarvāstivādin Abhidharma,” in Janet Gyatso, ed., In the Mirror of Memory: Reflections on Mindfulness and Remembrance in Indian and Tibetan Buddhism, Albany: State University of New York Press. Pp. 67–108. Duerlinger, James. 2003. Indian Buddhist Theories of Persons. London, New York: Routledge. Gelong Lodrö Sanpo. 2012. Annotated English Translation. Abhidharmakośa-Bhāṣya of Vasubandhu: The Treasury of the Abhidharma and its (Auto) commentary. Translated into French by Louis de La Vallee Poussin. Delhi: Motilal Banarsidass. Griffiths, Paul. 1992. “Memory in Classical Indian Yogācāra,” in Janet Gyatso, ed., In the Mirror of Memory: Reflections on Mindfulness and Remembrance in Indian and Tibetan Buddhism, Albany: State University of New York Press. Pp. 109–131. Gyatso, Janet. 1992. Ed. In the Mirror of Memory: Reflections on Mindfulness and Remembrance in Indian and Tibetan Buddhism, Albany: State University of New York Press. Hattori, Masaaki. 1968. Dignāga, On perception. Cambridge, MS: Harvard University Press. Jaini, Padmanabh S. 1992. “Smṛti in the Abhidharma Literature and the Development of Buddhist Accounts of Memory of the Past,” in Janet Gyatso, ed., In the Mirror of

326

Lin

Memory: Reflections on Mindfulness and Remembrance in Indian and Tibetan Buddhism, Albany: State University of New York Press. Pp. 47–59. Katou, Toshio (加 藤 利 生 ). 1994. “Nyāyavārttika ad ns, iv.2.31–35 に 於 け る 唯 識 說 批 判 ” (Nyāyavārttika ad ns, iv.2.31–35 ni okeru yuishiki setsu hihan) (A Study of Uddyotakara’s Criticism of Vijñānavāda). 仏 教 學 研 究 Bukkyogaku Kenkyū, 49: 26–50. Kellner, Birgit. 2010. “Self-Awareness (svasaṃvedana) in Dignāga’s Pramāṇasamuccaya and – vṛtti: A Close Reading.” Journal of Indian Philosophy, 38: 203–231. Kellner, Birgit and John Taber. 2014. “Studies in Yogācāra-Vijñānavāda Idealism i: The interpretation of Vasubandhu’s Viṃśikā,” Asiatische Studien – Études Asiatiques, 68 (3): 709–756. Matilal, Bimal Krishina. 1974. “A Critique of Buddhist Idealism,” in Cousins et al. (eds.), Buddhist Studies in Honour of I.B. Horner, Dordrecht: Springer. Pp. 139–169. Odani Nobuchiyo (小 谷 信 千 代 ). 2010. “Kliṣṭa-manas and mana-indriya: The Reason for the Deep-rootedness of Ego-attachment” (染 污 意 と 意 根 – 我 執 の 根 深 さ の 根 拠 ), Buddhist Seminar, 92: 1–13. Slotnick, Scott D. and Daniel L. Schacter. 2007. “The Cognitive Neuroscience of Memory and Consciousness,” in Philip David Zelazo, Morris Moscovitch, and Evan Thompson, eds., The Cambridge Handbook of Consciousness. Cambridge: Cambridge University Press. Pp. 809–827. Takeda, Hiromichi (武 田 宏 道 ). 2007. Muga no ronshō – “Kusharon” hagabon no kenkyū. Fukuronbun: “Kusharon” hagabon wayaku (無 我 の 論 證 – 『 俱 舍 論 』 破 我 品 の 研 究 ・ 副 論 文 :『 俱 舍 論 』 破 我 品 和 譯 ). PhD Dissertation, Ryukoku University. Vasubandhu: Viṃśatikā vijñaptimātrāsiddhi (= Vvs) http://www.sub.uni-goettingen. de/ebene_1/fiindolo/gretil/1_sanskr/6_sastra/3_phil/buddh/vasvvmsu.htm. Wayman, Alex. 1992. “Buddhist Terms for Recollection and Other Types of Memory.” In Janet Gyatso, ed., In the Mirror of Memory: Reflections on Mindfulness and Remembrance in Indian and Tibetan Buddhism, Albany: State University of New York Press. Pp. 133–147.

Index abductive inference (inference to the best explanation) 116, 177, 203 See also arthāpatti ābhāsa (appearance) 320 Abhidhammattha-Sangaha 126n8 Abhidharma 4, 4n3, 5–6, 8–9, 9n6, 10, 12, 26–27, 31, 33, 38n14, 53, 63, 66n4, 100n15, 101n17, 123n3, 125–126, 144–146, 148, 184–185, 192–193, 202n17, 206, 244–245, 265n37, 266, 271–272, 275, 278, 280–282, 284, 301, 310, 322 *Abhidharmahṛdaya (or *Abhidharmasāra, Apitan xinlun 阿毘曇心論) 41, 41n20, 43, 45 Abhidharmakośa (Vasubandhu) 36n10, 37n12, 39, 41, 41n20, 42–44, 47–50, 50n30, 51, 51n32, 70n10, 251, 263n32, 305, 312 Abhidharmakośabhāṣya (Vasubandhu) 70n10, 92, 105, 105n26, 227n9, 263n32, 282, 282n7, 283, 283n11, 290, 290n32, 291, 291n26, 291n33, 291n34, 308–309, 309n1, 311, 311n7, 322n32 Abhidharmasamuccaya 261n27 Abhidharmasamuccayabhāṣya 105n25, 261, 263n32, 264, 264n34, 291 *Abhidharmāvatāra (Ru apidamo lun 入阿 毘達磨論) 41n21, 49–50, 50n29 abhinirūpaṇāvikalpa (discrimination through examination). See under vikalpa aboutness 126n9, 139 ādarśana 221n24 ādānavijñāna (appropriating consciousness) 287 adhimukti (aspiration, wish) 100n15, 104n23, 110 afflicted mind. See kliṣṭa-manas agency 84, 95–97, 100, 123n3, 124, 189, 316 aggregates. See skandhas Aiyaswami Sastri 124n5 ākāra (aspect, form, mental image, representation) 6, 8, 14, 18, 21, 27, 36, 38, 38n14, 48, 50, 70–72, 117, 141, 172, 198, 213, 227, 227n9, 229, 232–233, 246,

251–252, 254, 265n36, 270, 270n46, 276, 301–302, 315 ākāśa (space) 51, 78, 91n6, 95, 109, 137, 260, 283–284 akliṣṭa (unstained) 93, 292 ālambana (intentional object, objective support) 40, 70–71, 109, 195n8, 227n9, 231–232, 270, 270n47, 276, 278–280, 284, 284n12, 285, 292n39, 295, 295n48, 296, 305, 307, 309, 320 immediate vs. mediate objective support 231–235 ālambana-pratyaya (cognitive object) 257n16, 266 Ālambanaparīkṣā (Dignāga) 19–20, 125n6, 251, 262n29, 265–267, 267n40, 271, 281n2, 300n59 Ālambanaparīkṣāvṛtti (Dignāga) 276 ālayavijñāna (store-consciousness, storehouse consciousness, base consciousness, eighth consciousness) 11, 156n1, 199, 232–233, 257, 260, 260n24, 261, 264–265, 265n36, 265n37, 297 Albahari, Miri 194n6 Alter, T. 141n35 Anacker, S. 317n21, 320n28 Anālayo, Bhikkhu 36n11 anaikāntika-doṣa (fault of inconclusiveness) 267 anadhigatārtha-gantṛ (providing new information) 253, 269 Aṅguttara-nikāya 34n7 annihilationism 3, 194 Anscombe, Elizabeth 123n1, 184n42 anti-realism. See under realism anti-reflexivity principle 161, 161n7 See also irreflexivity anubhava (immediate experience) 270, 270n47 anumāna (inference) 248, 256, 268–269 anusmaraṇa 97–8, 291 See also recollection anusmaraṇavikalpa (discrimination through recollection). See under vikalpa

328 apoha (nominalism) 13, 14n9, 121 apramāṇa 256 argument argument from blindness 300, 300n59 argument “from spontaneous reportability” 142n36 fineness of grain argument for nonconceptualism 27, 27n1, 54–55, 66, 67n6, 75, 142 infinite regress argument for reflexive (same-order) theory of consciousness 17, 21, 115, 117, 138, 162–163, 165, 165n16, 196 memory argument (argument from memory) for the existence of ātman (Nyāya) 306 for the existence of external world 317–318 for reflexive self-awareness (svasaṃvitti) 17–18, 53n37, 124, 130n17, 162, 196, 196n10 neither one nor many argument for svasaṃvitti 172–173 nondualist argument for nonconceptualism 28, 79–85, 101n18 private language [argument] 102, 110 sahopalambhaniyama (“egocentric predicament”) argument 125, 135, 139–142, 171 Armstrong, D. 127n10, 136 Arnold, Dan V 34n8, 69–70, 92n7, 115–116, 126n9, 130n17, 136n27, 137n28, 138n32, 143n38, 144n40, 154, 156n1, 163n11, 163n12, 164n13, 164n14, 167n18, 177n36, 185n43, 190, 197n12, 250n3, 250n4 arthāpatti (abductive inference, inference to the best explanation) 197 Āryadeva (3rd c) 147n42, 195n8 asaṃjñisamāpatti (trance of nonidentification). See under samāpatti asaṃskṛta 279–280, 283, 283n11, 284, 308 Asaṅga (4th c) 2, 63, 69, 103, 103n19, 104–105, 105n25, 107–108, 128, 144, 209, 234n19, 323n34 āśaya (aspiration) 29, 96–98, 100, 100n15, 109, 109n34, 110 Ascertainment of Knowledge. See Pramāṇaviniścaya aspiration. See āśaya and adhimukti aspect. See ākāra

Index āśraya (basis) 98n14, 129n14, 213, 256, 256n15, 259, 286, 322 āśrayaparāvṛtti (transformation of [the] support) 98, 98n14 associated. See saṃprayukta association. See saṃbandha ātman (permanent self). 35, 51, 305–308, 317 See also self ātmavādapratiṣedha (Refutation of the Theory of a Self ) 305, 307–308, 310, 317, 319, 323 ātma-grāha (self-grasping) 129 attention. See manaskāra attribution [of mental states] 147n42, 203–206 first-person vs. third person 19, 116, 118, 178n37, 190, 195, 197–198, 225–226, 239 *avakāśadānāśraya (basis that makes way and leads to rise [of the succeeding cognition]) 259 avijñapati-rūpa (non-informative matter). See under rūpa āyatana (sense sphere) 262, 278–283 āyatana-svalakṣaṇa. See under svalakṣaṇa Aydede M. 141n35 Balog, Katalin 141n35 bare thought 127 base consciousness. See ālayavijñāna basis qua the proximate condition. See samanantara-pratyaya-āśraya basis that makes way and leads to rise [of the succeeding cognition]. See *avakāśadānāśraya Bayne, Tim 146n41 belief 29, 54–55, 67, 76, 82n21, 122, 123n1, 137, 139, 191n2, 192, 201, 203, 205, 207 Berkeley, George 11, 125n7, 171, 171n26 Bermudez, José Luis 55n40, 67n7 bhāga (part, portion, aspect)  218, 254 darśana-bhāga/dṛṣṭi-bhāga (part of seeing) 234 nimitta-bhāga (part of the [object’s] image) 234, 258 saṃkleśa-bhāga (defiled portion) vs. vyavadāna-bhāga (pure portion) 98n14 *svasaṃvittisaṃvittibhāga (part of cognition of reflexive awareness) 209, 210n2, 213, 234

Index *svasaṃvittibhāga (part of reflexive awareness) 234 bhājanaloka (receptacle world) 105, 108 Bhāṭṭa Mīmāṃsā 123n2, 132 Bhāvanākrama (Kamalaśīla) 72 bhavaṅga-citta (latent mind) 34n7 bhūmi (stage, level) 278 bīja (seed) 105, 108, 279, 308, 312, 312n12, 316n20, 319 bimba and pratibimba (original image and reflected image) 221n24 binding problem 36 blind argument from blindness. See under argument blindsight 127n10, 204–205 person born blind 247, 255, 257, 269 Block, Ned 32n4, 127n10 Bodhi, Bhikkhu 35n9, 36n11 Bodhiruci 菩提流支 (5th-6th c) 258n18, 290 bodhisattva 15, 19, 26–29, 64–65, 69, 70, 73, 73n12, 74, 77–80, 89, 93–97, 99–100, 100n15, 103–104, 108, 107n32, 109, 220, 225, 227, 227n9, 228–231, 233–237, 239, 296 Bogdan, Radu 191n2, 202 Brahma-nimantanika-sutta 34n7 Brentano, F. 126n9, 136 Brewer, Bill 55, 55n40 Bronkhorst, Johannes 41n19 Buddha historical Buddha 122 mindless Buddha vs. mindful Buddha 97, 97n12 buddhahood 96–97 Buddhavarman (Futuobamo 浮陀跋摩) (4th-5th c) 44 Buddhāvataṃsaka-sūtra (Huayan jing 華嚴 經) 220–221 Buescher, Hartmut 90n3, 91n4, 91n6, 260n21 Cahen, Arnon 55n40 caitta/caitasika (mental associate/ concomitant/factor) 5n5, 38, 47, 98n14, 126, 192, 213, 288 Candrakīrti (ca. 700–750) 2, 73, 121, 123, 130n17, 134, 136–137, 137n28, 137n29, 144, 154–161, 161n7, 162, 162n10, 163, 163n11, 164, 164n14, 164n15, 165, 165n16, 166–169,

329 169n23, 170–171, 171n25, 173–175, 175n33, 176–182, 184–185, 191n3, 195n8, 211 Carpenter, A. 123n3 Carruthers, Peter 32n2, 141n35, 189, 191, 191n2, 198, 202, 204, 226n1 Cartesian 157, 322 Cārvāka (Materialist school) 116, 137, 141n35, 147, 199, 202 causality/causation causal theory of memory 311–317 causal theories of reference 107n32 causation, theory of 146, 193, 202, 202n17, 308 knowledge of causality 90, 92, 94–95, 100, 102, 110 Cavell, Stanley 159n4 cessation See nirodha Chachakka-sutta 35n9 Chalmers, David 131n19, 141n35, 143n37 Chan [and non-conceptual states] 34, 63 Chatterjee, Satischandra 14n9, 306, 308, 310, 310n5 Cheng wei shi bao sheng lun 成唯識寶生論 (Dharmapāla) 225–226 Cheng weishi lun 成唯識論 (*Vijñaptimātratāsiddhi) 20, 209–210, 225–226, 231, 233, 250, 291, 291n37, 292n41, 292n42, 295, 295n46, 314n16, 318n24 Chengguan 澄觀 (738–839) 209–221, 221n21 Chödrön, Gelongma Karma Migme 92n9, 93, 95–98, 98n14, 100n15, 108 Chu sanzang ji ji 出三藏記集  212n8 Chu, Junjie 259n20, 295n47 Chuard, Philippe 68, 74n13, 76n15 ciṇṭāmaṇi (wish-fulfilling jewel)  96–97 citrīkāra (objectification) 69 citta (mind) 5n5, 34n7, 98n14, 126, 192, 210, 213, 232n15, 286, 288, 294 cittamātra (consciousness only) 11 cittaviśeṣa (special/distinctive [power of] mind) 311, 311n8, 312n12, 314, 314n14, 316–317 cognition cognition of one’s own mind. See svacittajñāna

330 cognition (cont.) cognition of other minds. See paracittajñāna cognition of self-cognition/awareness of self-awareness. See *svasaṃvitti-saṃvitti cognition with images. See Sākāravāda cognition without images. See Nirākāravāda invalid means of cognition. See apramāṇa means of valid cognition. See pramāṇa result of valid cognition. See pramiti subject/cognizer of valid cognition. See pramātṛ Collection on the Sources of Knowledge. See Pramāṇasamuccaya and Vṛtti continuum (mental continuum). See saṃtāna Corpus of Valid Cognition. See Pramāṇasamuccaya Coliva, Annalisa 55n40 Collins, Arthur W. 55n40 Collins, Steven 34n7 color (varṇa) 4–5, 7–8, 9n6, 14, 37n12, 39, 54–55, 114, 126, 131, 141–143, 245–246, 250–251, 253, 265, 267, 270–271 Commentary on the Yogācārabhūmi (Kuiji) 255 commissurotomy 204–205 concept 8–10, 13–15, 26–29, 32, 37–38, 40, 43, 49, 54, 54n39, 55, 62, 64–72, 75, 75n14, 76–82, 85, 93–94, 96–99, 100n15, 101n18, 118, 121, 124–125, 127–128, 128n13, 129, 129n14, 133n23, 135, 138–141, 141n35, 142–143, 146, 156, 158, 190, 192, 192n4, 195, 195n8, 203, 235, 237n26, 245, 249, 251, 258n18, 266n39, 267–268, 282, 284–285, 316, 320, 320n27 See also kalpanā concept possession 75–76 conceptual (savikalpa) perception 310, 310n5 conceptual analysis vs. ostensive demonstration 144 conceptualist vs. non-conceptualist accounts of cognition 12, 22 demonstrative concepts 55, 75–76 devoid of conceptualization. See kalpanāpoḍha mere concept. See prajñapti

Index conceptual 6, 8, 12–16, 18, 26, 27n2, 28–33, 38–41, 43, 47–48, 50, 53–56, 62–63, 64n3, 65–67, 67n7, 68–70, 72–75, 77, 79, 81–82, 85, 89–92, 94, 97, 99, 101, 101n16, 101n18, 102, 105n27, 106, 107n31, 109–111, 118, 121, 123n1, 130n17, 133, 135n24, 135n25, 136, 144–145, 178, 180, 192n4, 222, 222n25, 237n26, 245–246, 248, 250n4, 252, 254, 270, 277–278, 280, 286–287, 289–293, 315, 318, 320 confabulation 190n1, 203–206 consciousness See also vijñāna access consciousness 32 global availability account of consciousness 199, 201–202 phenomenal consciousness 32, 43, 43n25, 125, 127n10, 139, 141, 141n35, 143 non-relational accounts of consciousness 125, 131 contaminated. See sāsrava contextualist 33, 33n6, 102 continuum [of consciousness] (mental continuum) 305, 307–308, 312, 314–316, 324 See also saṃtāna convenient designator 6–7 conventional/conventionally (saṃvṛti) 6–8, 12, 26, 51, 66n4, 79, 81, 83, 93–94, 116, 137n28, 144n40, 155–156, 156n1, 157–159, 159n4, 160, 163–164, 169–171, 175, 177–178, 181–185, 201, 227, 246–266, 267n40, 268, 268n44, 269–272, 316 See also saṃvṛti conventional reality/conventionally real (pratjñaptisat). See under prajñapti conventional truth (saṃvṛti-satya). See under truth Coseru, Christian 34n8, 115, 126n9, 133, 134, 135n25, 156n1, 157, 166, 171, 190, 193n5, 194n7, 196n10, 197n12, 198n13, 250n3 Cousins, Lance 41, 41n18, 41n19 Cox, Collet 37n12, 52n32, 52n34, 52n35 Crowther, Timothy 55n40 D’Amato, Mario 69n8, 70, 73, 97, 97n12 Dantinne, Jean 126n8 Daotai 道泰 ( 5th c) 44 darpaṇa (mirror) 221n24

331

Index Dārṣṭāntika 38n13, 45, 52, 275–276 Dasheng fayuan yilin zhang 大乘法苑義林 章 (Kuiji) 250, 255 Dasheng fayuan yilin zhang buque 大乘法 苑義林章補闕 (Huizhao) 220 de se vs. de re 123 Deleanu, Florin 34 Demiéville, Paul 72, 72n11 demonstrative concept. See under concept Dennett, Daniel C. 183, 200 dependent arising (dependent origination). See pratītyasamutpāda dependent nature. See under svabhāva Descartes, René 136, 191 Dessein, Bart 38n13, 51n32 determinant condition (adhipatipratyaya). See under pratyaya *Devakṣema or *Devaśarman (d.u., possible name of the author of the Vijñānakāya-śāstra) 39n15 devoid of conceptualization. See kalpanāpoḍha Dhammajoti, Bhikku K.L. 37n12, 38n14, 41n21, 46n26, 50n29, 50n31, 51n32, 52n35, 53, 53n36, 92, 135n26, 202n17, 206, 276n2, 283 dharma 3, 6–8, 16, 18, 36, 39, 42n22, 45, 47, 49–51, 51n33, 70–71, 77–78, 98, 192, 192n4, 193–194, 245, 260n22, 262, 262n30, 268n43, 268n44, 272, 277, 280–285, 301, 313 dharmakāya (reality-body) 73 Dharmakīrti (7th c) 2, 13–14, 14n9, 15–22, 31, 106n29, 121, 123, 123n2, 125, 129n15, 135, 137–140, 142–144, 155, 156n1, 159n4, 171, 173n30, 193n5, 196, 196n10, 198, 203, 206, 214, 216–217, 222n25, 225–227, 236, 239, 244, 246, 253, 253n9, 254, 268, 268n44, 269–270, 272, 272n49, 288n22, 302 Dharmapāla (fl. mid-6th c) 2, 19–21, 116–118, 128n13, 209–210, 210n1, 210n3, 211, 213, 215–220, 222, 225–231, 233, 233n17, 234, 236–237, 237n28, 239, 256–257, 266–267, 267n40, 267n41, 292–293, 300n59 Dharmaskandhapāda-śāstra (Fayun lun 法蘊 論) 46n26

Dharmottara (8th c) 2, 246, 250 dhātu (element) 35, 282–283, 283n11 Dignāga (6th c) 2, 10, 12–13, 15–22, 31, 34, 43n24, 53n37, 101n17, 106n29, 115–117, 121, 123, 123n2, 124, 125n6, 125, 128, 128n13, 130, 130n17, 131, 131n18, 132–135, 139–140, 144, 155, 156n1, 160–167, 176–177, 189–190, 193, 193n5, 194, 194n7, 196, 196n10, 206, 209, 211–212, 214, 214n13, 215–220, 222, 234n20, 244–245, 247–254, 256, 262n29, 265–272, 272n49, 276, 289–293, 300n59, 302, 315n18 direct perception. See pratyakṣa discerning awareness 126, 140 discrimination, doctrine of three kinds of (trivikalpa; san fenbie三分別). See under vikalpa discriminative thinking 322 discursive thought 41, 43, 103–104, 104n21, 108 dispositional property, consciousness as a 201 Dokic, J. 74n13, 75 double moon 141, 267, 267n41 Dragonetti, Carmen 251n7, 266 Dretske, Fred 76n16 Dreyfus, Georges 34n8, 84, 143n38, 156n1, 181n40 dual-system theory 200 duality [of subject and object] 28, 70–71, 78, 80–82, 84, 91, 98n14 See also grāhya-grāhaka Duerlinger, James 308, 311n8 dukkha/duḥkha 35n9, 259 See also suffering Dunne, John 73 Dunryun 遁倫 (active during the 7th century) 280, 285–290, 296–300 Dzogchen 34, 63 egological 1, 7, 123, 129, 157 immersive vs. egological 129–130 non-egological 1, 7 eighth consciousness. See ālayavijñāna eliminativism 53, 148 Eltschinger, Vincent 173n30, 250 emergent property 141 enlightened mind 90, 108

332 entrant on the path of cultivation. See srota-āpanna epistemic gap vs. ontological gap 38, 127, 141 epistemic privilege 15, 122 epoché (phenomenological reduction) 140, 148 essenceless 29, 70, 89–91, 94, 144 essencelessness. See niḥsvabhāvatā Establishment of Consciousness-only, The. See Cheng weishi lun eternalism 3, 194 Evans, Gareth 27n1, 54, 54n39, 66, 67n6, 75, 123n1 evidence 15, 75, 100n15, 116, 118, 127n10, 139, 144, 147, 190n1, 193, 195, 198, 200, 202, 204, 207, 232–234, 237, 276, 284, 294, 308 existent/existence ultimately real existent 183–184 really existent (dravyasat) vs. conventionally existent (prajñaptisat) 245 See also under sat primary (dravyasat) and secondary existent 145 external object 11, 100n15, 132, 140–143, 225–229, 231–233, 239, 244–245, 252–253, 255, 257, 269–271, 297, 305–306, 308, 318, 318n22, 319, 321 external reality/world 11, 89–90, 103, 122, 141, 191, 305, 307–308, 315n18, 317–318 fallibility 136, 181 Farah, Martha 77n18 fault of inconclusiveness. See anaikāntika-doṣa Faxiang zong 法相宗 (School of the Intrinsic Nature of Real Entities) 20, 210 figurative theory of meaning 107n32 fineness of grain argument for nonconceptualism. See under argument first-person first-person [stance] 19, 115–116, 138, 184n42, 238, 306, 315 first-person vs. third-person 56, 118, 157, 178, 178n37, 190, 195, 197, 239

Index first-personal access [to consciousness] 124, 157, 159–160, 175–176, 181, 239 See also privileged access [to (one’s own) consciousness] first-personal awareness 157, 177 first-personal givenness of experience 84, 126, 128, 130, 143, 236 first-personal perspective 106, 174–175, 179, 185 five states of mind 275, 277–278, 293, 295, 298, 301 spontaneous (aupanipātika) mind 277, 292, 294–297 searching (paryeṣaka) mind 255n13, 277, 292, 294–297, 299, 301 discerning (niścita) mind 277, 294–297, 299, 301 [mind that is] defilement or purification (saṃkleśo vyavadānaṃ ca) 255n13, 294 [mind] uniformly flowing from that (tannaiṣyandika) 294 Fo di jing lun 佛地經論 (The Treatise on the Buddha Land) 209–210 folk psychology 1, 4, 116, 178, 181, 185, 201 Forman, Robert K.C. 33n5, 62 fourfold division of the mind (fourfold division theory) 116–117, 128n13, 209–211, 213, 215–222, 234, 234n20, 236 Franco, Eli 130, 131n18, 137n31, 249n2, 252n8, 271–272 Frauwallner, Erich 39n15, 234n20 Fukihara, Shōshin 215, 215n14 Fukuda, Takumi 38n13 Funayama, Tōru 21n11, 116–117, 128n13, 156n1, 210n1, 218n18, 219n19, 234, 234n20, 250n3, 288n22 fundamental consciousness. See ālayavijñāna Gallagher, Shaun 1n1, 84n22, 131n19 Ganeri, Jonardon 131n19, 136n27, 162n8 Garfield, Jay L. 34n8, 134, 136n27, 157, 161n7, 165n16, 167n19, 168, 168n21, 168n22, 169n23, 176, 176n34, 177–179, 181 Gelong Lodrö Sangpo 311n8, 315n19, 316n20, 317 generalized objects. See sāmānyalakṣaṇa Gennaro, Rocco 136

Index Ghoṣaka (d.u.) 51n32 given epistemological givenness 160, 178–179 myth of the given 176–177 phenomenological givenness 84, 100, 102, 109n34, 126, 128–130, 135, 143, 160, 175–176, 178–179, 183, 236, 254, 310n5 gocara (operative field) 270n45, 284, 284n12 Gombrich, Richard 34n7 grāhaka[-ākāra] (subject-aspect of cognition, cognitive agent) 91n4, 93, 128–129, 160n5, 209–211, 213, 213n10, 214–218, 254 grāhya[-ākāra] (object-aspect of cognition, grasped-aspect) 71, 91n4, 93, 128, 160n5, 209–211, 213–218, 254 grāhya-grāhaka (grasper-grasped) 71, 91n4, 93, 98n14, 210–211, 213–218, 254 Granoff, Phyllis 56 grasped-aspect. See grāhya-[ākāra] grasper-aspect. See grāhaka[-ākāra] grasper-grasped. See grāhya-grāhaka great elements. See mahābhūta Griffiths, Paul J. 65–66, 70–71, 315 Guansuoyuan lun shi 觀所緣論釋 (A Commentary on the Ālambanaparīkṣā) (Dharmapāla) 266 guṇa (mode of the self) 193 Gyatso, Janet 305, 307 Güzeldere, G 141n35 Hanna, Robert 54n38, 67n7 Harada, Wasō 38n13, 51n32, 51n33 Harivarman (ca. 4th c) 47n27, 124 Harvey, Peter 34n7 Hattori, Masaaki 34n8, 130, 132n21, 139n33, 162n9, 213n10, 222n25, 252–254, 254n10, 255, 263n32, 268, 268n44, 270, 270n46, 270n48, 272, 322n31 Hayashima, Osamu 261n27 Heck, Richard 29, 66, 67n7, 74n13 Heidegger, Martin 129, 129n15 hetu (proving property) 93, 197 Honjō, Yoshifumi 38n13 Hossō School (hossō shū 法相宗) 20 Huizhao 慧沼 (651–714) 220, 297, 300 Huntington, C.W.  161n6, 162n10

333 Husserl, Edmund 115, 134, 136, 140, 162n8, 194n7 Hīnayāna 260 I-making 129 idealism/idealist 4n4, 10–12, 18–20, 115, 117–118, 125n6, 147, 155, 157–158, 160–161, 170, 172–174, 180–182, 194n7, 226–228, 232–233, 246, 266, 267n40, 279, 285, 297, 307, 314n16, 317, 318n23, 319, 319n26, 321 epistemological idealism 172, 298 metaphysical idealism 11n7, 297 ideation. See saṃjñā ignorance (moha) 16, 107n31, 129n14, 148, 190, 230, 263 illumination, metaphor of 16, 18, 114, 193, 201, 212 image-part. See under bhāga immediate experience. See anubhava immersive vs. egological. See under egological impermanent/impermanence 4, 5, 9–10, 12, 16, 26, 90, 147, 192, 194, 195n8, 308 impressions 45, 105, 308, 314 Inami, Masahiro 221n7 incommensurability epistemic 92n7, 94, 110 linguistic 107, 107n32 indirect realism (representationalism). See under realism indriya ([sense] faculty, organ) 35, 38, 249, 260, 266, 270n45, 276, 323, 323n34 indriya-pratyakṣa (sense perception). See under pratyakṣa ineffable 70, 90, 99, 103, 193n5, 226, 229, 230, 237, 238 inference. See anumāna inner sense/inner-sense 3, 7, 20, 22, 114, 196 See also manas inner sense theories or hop 32, 202–206 insentient object 137 intentional object. See ālambana intentionality 53, 109–110, 125n6, 125–126, 126n9, 128–129, 173, 260, 260n22, 315 See also manaskāra

334 intersubjective intersubjective (experience) 104–106, 106n28, 108, 238 intersubjective agreement 106 intersubjective report 124, 129 intimating 144, 194 intrinsic nature. See svabhāva introspection (pratyātmavedya) 7, 98n14, 176, 124, 129, 134, 176–178, 184n42, 196, 197n11, 202, 204, 238, 252 introspective awareness 17, 121, 123, 129n14, 131n18, 145, 196 irreflexivity 10, 16, 18, 22, 115–116, 192–193, 193n5, 197, 197n12 “I”-sense 3–4, 9–10, 82n21, 191, 191n3, 192, 192n4, 195n8, 199, 207 Iwata, T. 143n38 Jackson, Frank 142 Jain 305, 307 Jaini, Padnamabh Shreevarma 41n18, 46n26, 51n32, 310 Jha, Amishi 64n3 Jieshenmijing shu 解深密經疏 (Wŏnch'ŭk) 267n41 Jīnaputra (d.u.) 20 Jinendrabuddhi (ca. 700) 216–217, 254, 269–271 jñāna (cognition) 5n5, 69, 91n6, 95, 126n9, 132n21, 140, 164n15, 190, 216, 222n25 Jñānaprasthāna 283 justification 144, 314 kalpanā (conceptual construction) 6, 8, 12, 53, 107n31, 113n22, 195n8, 222, 222n25, 277, 289–293 kalpanāpoḍha (devoid of conceptualization) 248, 252, 254, 270–271, 289n27, 291 Kamalaśīla (8th-9th c) 2, 72–73, 73n12, 82 Kano, Kazuo 226n5 Kant, Immanuel (1724–1804) 54n38, 134, 136, 166–168, 173, 322n33 Kapstein, Matthew 105n26 karma 11, 96, 105, 105n25, 107n31, 108, 164, 199, 207, 262, 316, 322–323 karmic seed 105, 323 Katusra, Shōryū 56, 212n8, 227n6, 233n17 Katz, Steven 33n6

Index Kawamura, Leslie 260n21 Kellner, Birgit 14n10, 34n8, 38n14, 51n32, 132n21, 135n26, 143n38, 156n1, 162n8, 162n9, 165n16, 194, 196n10, 226n5 Kelly, Sean 67n7, 74n13, 75 Keng, Ching 34, 43n24, 56, 63n2, 79n20, 101n17, 200n16, 202n18, 244–245, 255n13, 258n18, 285n14, 293n43, 302, 320n27 Kevaddha-sutta 34n7 kliṣṭa-manas (afflicted mind) 125, 128, 129n14, 131n19, 323n34 knowledge claims 97–98 Kouider, S. and Dehaene, S. 127n10 Kramer, Jowita 49n28, 56, 129n14, 255n13 Kriegel, Uriah 128n12, 131n19 Kripke, Saul 116, 196n9 Kritzer, Robert 38n13, 51n32, 51n33 Kuiji 窺基 (632–682) 2, 20–21, 118, 131n18, 215n14, 219–220, 220n20, 221, 225–226, 226n4, 229, 233, 233n17, 234–239, 245–247, 250–251, 255, 255n14, 256, 257n16, 260–261, 264, 264n34, 265, 265n37, 267, 267n40, 269–272, 280, 289, 289n26, 290–291, 293, 296n49, 299–300, 305–308, 314n16, 318n24, 319–323, 323n34, 324n35 Kumārila (ca. 700) 123n2, 130n17, 136, 161, 161n7, 162, 162n9, 189, 191, 193n5, 197, 197n12, 203 La Vallée Poussin, Louis de 124n4, 161n6, 161n7, 162n10, 175n33, 210n2, 311n8 Lamotte, Étienne 69, 69n9, 71, 77–78, 92n9, 93–98, 98n14, 100n15, 105, 108, 129n14, 258n18 Lau, Hakwan 32n2 Levine, Joseph 54–55, 55n40 Li, Xuezhu 161n6 light (dīpa or pradīpa) 212 Lin, Chen-kuo 56, 202n18, 246, 276, 285n14, 300n59 Lin, Qian 35n9, 47n27, 56 Lindtner, C. 143n38 lived-experience 144 Loar, B. 141n35 Locke, John (1632–1704) 127n10, 136, 265 Lopez, Don 56

Index Lusthaus, Dan 232, 232n16, 276, 278–281, 285n15, 294n45, 297 Lycan, William 136 MacDonald, Anne 161n6, 164n13, 164n15 Madhupindika-sutta 35n9 Madhyamaka/Mādhyamika 15–16, 18, 62–63, 72, 115–116, 137n29, 139, 146, 154–155, 156n1, 157–160, 166–167, 167n18, 168, 170–171, 171n25, 172–174, 175n33, 181–184, 199, 209 Madhyamakālaṃkāra (Śāntarakṣita) 158, 167, 167n18, 168, 169n23, 170–173, 174n32, 183n41 Madhyamakāvatāra (Candrakīrti) 73, 136n27, 160, 161n6, 164, 175n33, 211 mahābhūta (great elements) 263 mahābhūmika 45–46, 288 attention (manaskāra). See manaskāra contact (sparśa) 4–5, 35, 36n10, 39, 101, 101n18, 126n8, 276, 288, 308, 310n5, 319 desire (chanda) 8, 83, 122, 201, 252, 254, 271, 288, 295, 295n48, 298 discrimination/knowledge (prajñā) 27, 37, 45–48, 288 See also prajñā feeling (vedanā) 35, 49, 259, 280, 283–284, 288 See also vedanā ideation (samjñā) 37–38, 49, 50, 280, 283–284, 288, 311–315, 322 See also samjñā meditation (samādhi) 288 memory/recollection (smṛti) 37, 45–48, 288 See also smrti resolution (adhimokṣa) 288 thinking (cetanā) 288 Mahānidāna-sutta 36n10 Mahāsāṃghika (Great Assembly) 53, 124, 288 Mahāvedalla-sutta 36, 36n10, 36n11, 37 Mahāvibhāṣā (Abhidharma-Mahāvibhāṣā) (Apidamo dapiposha lun 阿毘達磨 大毘婆沙論) 37n12, 41, 41n20, 42n22, 43–46, 46n26, 47, 50, 124n4, 161n7, 212, 275, 276n1, 282–283, 283n9, 283n10, 290n30, 290n31

335 Mahāyāna 4n3, 63–65, 73, 95, 103, 209, 260, 288 Mahāyānasaṃgraha (Asaṅga) 29, 69, 71, 77–78, 80, 89–90, 108n33, 129n14, 323n34 Mahāyānasūtrālaṃkāra (Asaṅga) 69, 69n8, 70–71, 78, 97, 97n12 Mahīśāsaka 294 Maitreya 209, 232n15 manas (inner sense/inner-sense, mental faculty, mental sense) 3, 7–8, 22, 35, 38, 114, 202, 202n18, 278–280, 282, 294–295, 306, 308, 321, 323 manas consciousness (= mental consciousness). See mano-vijñāna mānasa-pratyakṣa (mental perception). See under pratyakṣa manaskāra (attention, intentionality) 9, 64n3, 104, 108, 114, 126n8, 127n10, 168n20, 200–201, 203, 260, 260n22, 278, 286, 288 Mandelbaum, Eric 200n16 manobhūmi (the mental level)  292 mano-vijñāna/manovijñāna (mental consciousness, mind-cognition) 13, 31, 35, 121, 129n14, 131n18, 244–246, 249, 278–280, 282, 282n7, 284n12, 287–289, 294n45, 295, 295n48, 306, 308, 318, 318n24, 320, 322–324 Marcel, A.J. 127n10 Master Ji. See Kuiji Matilal, Bimal Krishna 34n8, 123n2 Matsuda, Kazunobu 91n5 maturation (vipāka) 105 māyā (magic show) 93 McClintock, Sara 38n14, 135n26, 181n40 McDowell, John 13, 27n1, 27n2, 55, 55n40, 67n6, 72–75, 76n17 memory 3, 6, 11, 17–18, 48, 53, 75, 97–98, 98n13, 99, 124, 135, 162–163, 196, 196n10, 197, 200–201, 246, 248, 257n17, 288, 305–311, 311n8, 314–324 memory argument. See under argument ownership of memory 308, 316–317 mental associate/concomitant/factor. See caitta/caitasika

336 mental consciousness See also mano-vijñāna mental consciousness simultaneous with the five sensory consciousnesses (msf) 20, 244–245, 250, 257–258, 258n18, 265, 271, 288–290, 293, 293n43, 299–300 diachronic mode of mental consciousness 319, 322 synchronic mode of mental consciousness 285n14, 320, 322 mental image. See ākāra mental perception. See under pratyakṣa mental stream 8–9, 12, 18–19, 107n31, 118, 123, 126–127, 133, 135, 228, 317 See also saṃtāna mereological nihilism (Buddhist) 6–7, 18, 117, 197n12, 201 meta-cognition 9, 16–18, 21–22, 114–118, 121–129, 131–135, 138–139, 141, 144–148, 156n1, 157, 158, 166, 177, 184n42, 189, 203, 246 See also higher-order theories of consciousness, under theories of consciousness metaphysics, revisionary vs. descriptive 147 metaphysical quietism 199 middle path 3 Milindapañha 35n9 Millikan, Ruth 76n16 Mīmāṃsā 132, 161, 177, 177n36, 196n10 mind-reading faculty 189, 203–206 mindfulness. See smṛti mineness/for-me-ness 84, 115, 128, 128n12, 131n19, 132–133, 139 Mipham (1846–1912) 73, 155, 159 Mirror of Fundamental Teaching 宗鏡 錄 200 misidentification 130n16, 133 Miyashita, Seiki 51n33 modality-specific nature of conscious apprehension 133 mode. See guṇa momentariness/momentary 5–7, 11, 17, 124, 126, 145, 194, 196, 254, 270, 311, 314, 318 Moran, Richard 159n7, 178, 181, 184n42, 185 Moriyama, Shinya 34n8, 107, 107n30, 118, 135n26, 234n20, 236n24 Morrison, A.B. 64n3

Index msf. See mental consciousness simultaneous with the five sensory consciousness Müller, V.R. 67n7 multiple-layered cognitive series 288 mundane vs. transcendent 92, 92n7 mutual consensus. See ubhayasiddha Nagatomi, Masatoshi 130, 254, 254n11, 255 Nagel, Thomas 54, 179, 180 naïve realism. See under realism nāma-skandhas. See under skandhas Ñāṇamoli, Bhikkhu 36n11 Nārada, Mahā Thera 126n8 nian 念 (mindfulness) 309, 311n8 See also smṛti Nida-Rümelin, Martine 139 Nidānavagga-sutta 34, 35n9, 36n10 niḥsvabhāvatā (essencelessness) 144 nimitta (inherent characteristic) 49, 70, 71, 103, 258 See also under bhāga nirākāravāda 135, 301 nirmāṇakāya (emanation-body) 73 nirodha (cessation) apratisaṃkhyānirodha (cessation without understanding) 283 pratisaṃkhyānirodha (cessation through understanding) 283 saṃjñāvedayitanirodha (cessation of ideation and feeling) 284 nirodhasamāpatti (cessation trance). See under samāpatti nirvikalpa (nonconceptual) 65–66, 93–94, 98, 110 nirvikalpajñāna (nonconceptual cognition, nonconceptual discernment) 26, 31, 33, 62–85, 89, 92n7 mūlanirvikalpakajñāna (fundamental nonconceptual cognition) 69, 94 nirvikalpalokottarajñāna (transcendent cognition that is free of conceptual differentiation) 91, 91n6, 110 prāyogikanirvikalpakajñāna (preparatory nonconceptual cognition) 69, 94 pṛṣṭhalabdhajñāna/ pṛṣṭhalabdhanirvikalpajñāna (subsequent cognition/subsequent

Index nonconceptual cognition) 14, 28, 69, 77, 89, 91–94, 97, 237 niyama (restriction, delimitation) 105n26, 135 niḥsvabhāva. See essencelessness noematic content 140 non-conceptual 10, 12–16, 26–34, 39–40, 43n24, 45, 47–48, 50, 54–56, 62–74, 77–80, 82, 85, 89–90, 94–97, 98n14, 99–103, 109n34, 109, 129n14, 132, 135n24, 163, 246–247, 250n4, 255, 277, 287, 289–291, 293 See also nirvikalpa content nonconceptualism (cnc) 29– 30, 32, 67, 67n7, 68, 70–71, 73, 78 nonconceptual content 26, 28, 54, 55, 64–66, 69–71, 78, 97, 101, 101n18, 128, 135n24, 135n25, 140 nonconceptualism 12, 29–30, 54n38, 55, 55n40 nondualist argument for nonconceptualism. See under argument supervenience nonconceptualism (spnc) 30, 62, 68, 73–74, 77, 80, 85, 101n18 state nonconceptualism (snc) 29–30, 32, 66–67, 70, 73 nondual/nondualistic/nonduality 16, 28–29, 65, 70–71, 74, 78, 80–85, 94, 101n18, 238–239 nondual cognition 65, 70–71, 74, 78, 80–85 nondualist argument for nonconceptualism. See under argument non-reifying 29, 79, 101, 101n18, 102 non-self/no-self (anātman) 1, 3, 9, 26, 28, 123, 145, 148, 160, 175, 184–85, 189, 192, 194, 195n8, 246, 310 nonconceptual cognition. See nirvikalpajñāna noematic content 140 Nyāya 13–15, 21, 121, 123n2, 132, 193, 306, 308, 310, 310n5, 311 Nyāyabhāṣya 212n7 Nyāyamukha (Dignāga) 20, 131n18, 133n22, 214, 222, 255, 270n45, 270n47, 289, 325n35

337 Nyāyapraveśa (Jīnaputra) 20, 215n14 Nyāyasūtra 211, 212n7 *Nyāyānusāra (Apidamo shun zhengli lun 阿毘達磨順正理論) (Saṃghabhadra) 40, 41n21, 42n23, 48, 50, 50n30, 51n32, 52n35, 135n24, 282n8, 283, 290n32, 291n33 object to be cognized. See prameya object-aspect of cognition. See grāhya objective support. See ālambana Odani Nobuchiyo 323n34 omniscience 117, 134, 230 opacity opacity of self-consciousness 106, 178n37, 190, 197–199, 202–204, 206, 226, 236 opacity thesis vs. transparency thesis 189–190, 202, 206 operative field. See gocara organ. See indriya other minds 18–19, 107n30, 107n31, 157, 177, 189–190, 225–227, 230, 233, 237, 239 cognition of other minds. See paracittajñāna problem of other minds 106, 106n29, 107n30, 136, 157, 159, 159n4, 225–226, 226n3 ownership of memory. See under memory Overgaard, S. 127n10, 130n16 O’Dea, J. 141n35 pabhassara citta (luminous mind) 34, 34n7 Pacherie, E. 74n13, 75 pain 19, 127–128, 128n11, 135, 138–139, 201 Pañcaskandhaprakaraṇa (Vasubandhu) 126n8 Pañcavastuka-vibhāṣa (Wushi piposha lun 五 事毘婆沙論) 51n32 paracittajñāna (cognition of other minds) 225, 226n5, 227, 227n9, 228–233, 233n17, 234–239 paradoxical 92, 99 Paramārtha 真諦 (499–569) 19, 44, 276, 291n33, 310–311, 312n11 paramārthasatbhāva (ultimate real entity) 170 Park, Changhwan 38n13

338 Pārthasārathimiśra (11th c) 161n7 particular 249, 258n18, 262–263, 268–269 See also svalakṣaṇa Peacocke, Christopher 27n1, 55n40, 74n13, 76n17 Pierce, C.S. 173–4, 174n31 perception/perceptual (pratyakṣa, indriyapratyakṣa) 5–6, 10–17, 21, 26–27, 27n2, 31–33, 37n12, 39–43, 43n24, 45, 49–56, 63n2, 67, 77, 81, 84, 89–90, 94, 100, 100n15, 101n17, 103, 105, 117, 121–124, 127n10, 128, 130, 131n18, 132–133, 133n23, 140–141, 143, 146, 146n41, 155, 163, 163n11, 164, 167–168, 174, 177, 195n8, 196, 200n16, 203, 206, 216, 222, 222n25, 226–227, 229, 236, 244–258, 261–263, 265–270, 270n46, 271–272, 289–290, 293, 293n43, 305–314, 317–322, 324, 324n35 See also pratyakṣa doctrine of three kinds of perception 52, 52n35 mental perception (mānasa-pratyakṣa). See under pratyakṣa perception based on the sense faculties (yigen xianliang 依根現量, *indriyāśrita-pratyakṣa) 52n35 perception of [inner] feelings (lingna xianliang 領納現量, *anubhāva-pratyakṣa) 52n35 perception of awareness or comprehension (jueliao/juehui xianliang 覺了/覺慧現量, *buddhi-pratyakṣa) 52n35, 53 perceptual content 55, 67, 67n7, 68, 73–77, 146n41, 200, 245, 320 low-level and high-level perceptual content 146n41 perceptual identification. See saṃjñā perceptual judgment 14, 203 theories of perception 16, 27, 34, 101n17, 132 perennialism 33, 33n5 Perry, Ralph Barton 125n7 personal identity 106, 124, 306, 316 Personalism. See Pudgalavāda phenomenal concepts strategy 141n35 phenomenal salience 128 phenomenological reduction. See epoché. Phillips, I. 127n10

Index Philosophical Investigations (Ludwig Wittgenstein) 130n16 physicalism 4, 116, 142, 141n35, 185n43, 189, 198–199, 202, 202n17, 207 Prābhākara Mīmāṃsā 123n2 Pradhan, P. 70n10, 105, 105n25, 251n6, 263n32, 291n33, 312n9 pragmatic 134, 316 prajñā (knowledge, understanding, insight) 27, 37, 45–49, 51–52, 322 Prajñākaramati (fl.975) 197 Prajñapti-śāstra (Shishe lun 施設 論) 46n26 prajñaptisat (conventionally real). See under sat pramā (knowledge) 306 pramāṇa (epistemic instrument, means of valid cognition) 116–117, 133n21, 139n33, 140, 164n14, 184, 211, 213–216, 248, 253, 254, 268–269, 318n22 pramāṇaphala (warranted results) 117, 140, 209, 211, 213, 213n10, 214–217, 220, 254 Pramāṇasamuccaya and Vṛtti (A Collection on the Sources of Knowledge) (Dignāga) 117, 132, 213, 213n10, 214–216, 222, 248, 289, 324n35 Pramāṇavāda 31, 34 Pramāṇavārttika (Dharmakīrti) 216n16, 222n25 Pramāṇaviniścaya (Ascertainment of Knowledge) (Dharmakīrti) 138, 143 pramātṛ (subject/cognizer of valid cognition) 147, 211 prameya (object to be cognized) 117, 147, 184, 211, 213, 213n10, 214–215, 220, 254 pramiti (= pramāṇaphala, result of valid cognition) 147 Prāsaṅgika 171n25, 181–182, 250n4 Prasannapadā (Candrakīrti) 136n27, 163 pratibhāsa (mental appearance) 323 pratipakṣa (antidote) 98n14 pratītyasamutpāda (dependent origination) 3, 83, 126–127 pratyabhijñāna (recognition) 310, 310n5, 311n7, 311n8, 312n12 pratyakṣa (sense perception, direct perception) 52, 163, 163n11, 222, 248, 249n2, 258n18, 262, 269, 271, 308, 318n22, 324n35

339

Index See also perception indriya-pratyakṣa (sensory perception) 249 See also perception mānasa-pratyakṣa (mental perception) 21, 39, 43n24, 63n2, 101n17, 130, 131n18, 244–245, 247, 249–257, 258n18, 267–269, 270n46, 271, 289 yogi-pratyakṣa (yogic perception) 249 pratyaya (cause/condition) adhipatipratyaya (determinant cause/ condition) 228, 228n11, 236n23, 256 samanantarapratyaya (similar and immediately antecedent condition) 253, 259, 260n22 pre-reflective pre-reflective awareness 32 pre-reflective self-consciousness 134 pre-reflective vs. reflective 84, 130 primary (dravyasat) and secondary existent. See under existent primary and secondary qualities. See under qualities private language [argument]. See under argument privileged access [to (one’s own) consciousness] 106–107, 181, 190n1, 194, 235, 239 property to be proven. See sādhya property dualism 201 propositional 29, 54, 67, 97, 99 proto-phenomenological 115, 125, 130, 133, 140, 194n7 Proudfoot, W. 33n6 proving property. See hetu pṛṣṭhalabdhajñāna (subsequent cognition). See under nirvikalpajñāna Pudgalavāda (Personalism) 123n3, 193, 308 Puguang 普光 (?-664) 305, 307, 309, 309n4, 311, 311n8, 312, 312n10, 312n12, 313n13, 313n14, 314, 314n17, 315n19 Qingguang 親光 (d.u.) 210, 210n4 quale/qualia 32, 40, 89, 135n24, 141, 143 qualities

primary and secondary (John Locke) 265 Raftopoulos, A. 37n7 Rahula, Walpola 105n25 Ratié, Isabelle 141n34 Ratnakīrti (11th c) 19, 106n29, 159n4, 196, 225–227, 227n7, 228 Ratnākāraśānti (c. 1000 ce) 173n30 realism 5, 18, 118, 132, 145, 147, 161, 173, 207, 232, 265, 266, 297, 305–310, 311n6, 318, 318n23, 323 anti-realism 144, 147 indirect realism (representationalism) 6, 12, 115, 135, 139, 146, 231, 246, 276, 317 naïve realism 117, 132, 146, 174 realism about universals 13, 13n8 reality ultimate reality 6–9, 16, 18, 116, 137n28, 148, 155, 158, 160–161, 170, 173, 183–184, 192, 194, 196, 201, 245, 249–251, 258n18, 261–272 See also under sat conventional reality 83, 156n1, 227 See also under sat recognition. See pratyabhijñāna recollection (anusmaraṇa, smaraṇa, smṛti) 27, 37–38, 42, 44–48, 83, 162, 162n9, 163, 290–291, 305, 307, 310, 310n5, 312–314, 319–323 See also smṛti reductionism 53, 144, 160, 184, 192, 198, 201 reflection 13, 27n2, 29–30, 33, 42, 46, 48, 50, 56, 72, 81, 101n17, 121, 123–124, 130, 134, 162n9, 172, 195, 197, 197n11 reflexive infinite regress argument for reflexive (same-order) theory of consciousness 17 memory argument for reflexive (sameorder) theory of consciousness 17 reflexive awareness/reflexive selfawareness/reflexive selfconsciousness 21, 115–117, 121, 123, 125, 128, 128n13, 130, 130n17, 131, 132n21, 134–137, 137n28, 140, 143, 144n40, 168, 197, 229, 233–236, 238, 315n18 See also svasaṃvedana

340 reflexive (cont.) reflexive (same-order) theory of consciousness 9, 16–18, 21–22, 53, 99, 115–116, 118, 134, 136, 142n36, 143, 162, 161n7, 189–190, 192–197, 197n11, 197n12, 202 reflexivity thesis (svasaṃvedana) (Dignāga) 16, 18, 21, 115–116, 118, 136, 142n36, 189, 193–197, 197n12, 202 Renz, Ursula 123n1 representation/representational 6, 14, 18, 20, 32, 54–55, 62, 66n5, 99, 118–119, 135, 139, 146n41, 147, 172, 201–202, 229, 252, 315, 317, 319 See also ākāra representationalism 6, 11, 17, 115, 135, 139, 146, 231, 246 See also under realism Rosenthal, David M. 32n2, 136–137 rūpa (material, form) 3, 38, 72, 195n8, 231–232, 262, 270n45, 279–280, 282n7, 283–285 avijñpati-rūpa (non-informative matter) 283–285 rūpaskandha. See under skandhas. Ryle, Gilbert 177, 177n35 sādhārana (shared) 89, 104–105, 105n27, 106–110 sādhya (property to be proven) 197 sahabhū āśraya (simultaneous basis) 256, 256n15 sahopalambha-niyama (necessary concomitance of object and cognition) 125, 135, 139–142, 143n38, 171 See also under argument sākāravāda 135, 301 sākāravāda (cognition is aspectual or imagistic) vs. nirākāravāda (cognition lacks such representational features) vs. alīkākāravāda (cognition possesses such imagistic features only falsely) 135 sākārajñānavāda 229, 231–232 salvific (activity) 69, 94–95, 99 samāhita (meditative state [of mind]) 298 samanantara (immediacy) as simultaneous 297

Index as successive 295, 295n48, 297–298 samanantara-pratyaya-āśraya (basis qua the proximate condition) 259 sāmānyalakṣaṇa (universal, generalized object, generic characteristic) 12, 37, 37n12, 40, 206, 230, 245, 249, 268 samāpatti (trance) asaṃjñisamāpatti (trance of non-identification) 129n14 nirodhasamāpatti (cessation trance) 15, 129n14 saṃbandha (association) 315 Saṃdhinirmocana-sūtra 244, 258n18, 287 Saṃghabhadra (5th c) 39, 48, 52, 52n35, 53, 53n37, 55, 123–124, 135n24, 282–283, 290n32, 291n33 saṃjñā (ideation/perceptual identifications) 3, 37, 47, 47n27, 48–49, 49n28, 311–312, 312n9, 312n12, 313, 315, 322 saṃjñānvaya 想類 (type of ideation) 311, 311n8, 312, 312n12 Sāṃkhya 134, 139 saṃprayukta (associated) 8, 18, 35–37, 41–43, 45–50, 53, 126n8, 129n14, 139, 192–193, 234, 277, 284n12, 289–294, 299, 309, 311, 320–321, 323n34 saṃskāra (embodied-conditioning, mental forces) 3, 11, 279–280, 283n11, 284, 308, 316n20 saṃsāra 72, 80, 90, 9499, 103–104, 108, 315 saṃtāna (mind-stream, [mental] continuum) 105, 105n26, 107n31, 124, 228, 233, 305, 307–308, 312–316, 324 saṃtānapariṇāmaviśeṣa (transformation of the stream to produce the effect) 316, 316n20 saṃvṛti-satya. See under satya *Samyuktābhidharmahṛdaya (Za apitan xin lun 雜阿毘曇心論) 50, 50n30, 51n32 Saṃyuktāgama (Za ahan jing 雜阿含 經) 35n9 Sanskrit grammatical tradition 164, 164n14, 165, 168 Santānāntaradūṣaṇa (Ratnakīrti) 226 Santānāntarasiddhi (Dharmakīrti) 226, 239

Index Śāntarakṣita (725–788) 2, 115–116, 137, 137n28, 137n29, 144n40, 154–155, 157–160, 167–173, 175–176, 178, 180–182, 183n41, 184 Śāntarakṣita’s account of svasaṃvitti 116, 135, 155–158, 166, 166n17, 167–170, 172, 179–180, 183, 185 Śāntarakṣita’s critique of idealism 174–175 Śāntarakṣita’s neither one nor many argument for svasaṃvitti 172–173 Śāntideva 136, 144, 189–190 Sarvāstivāda (Vaibhāṣika) 4, 26–27, 31, 36–50, 51n32, 53–54, 124, 132, 144, 202n17, 251, 275–276, 276n2, 277, 280–286, 288–291, 300–301 sāsrava (contaminated) 237 Sastri, Shanti Bhikshu 51n32 sat (being, existence, reality) dravyasat (ultimately real) 6–7, 16, 116, 145, 173, 193, 245–246, 249 paramārthasat (ultimately existent) 180 prajñaptisat (nominally existent, conventionally real) 6, 8, 12, 83, 137n28, 155–156, 156n1, 158–159, 169, 171, 175, 177–178, 182–183, 201, 227, 245–246, 246n42, 249–251, 261–272 satkāyadṛṣṭi (views of a real subject) 193 satya (truth) paramārtha-satya (absolute truth/ ultimate truth) 6, 79, 92, 170, 183–185 saṃvṛti-satya (conventional truth) 6–7, 12, 66n4, 116, 155, 159, 159n4, 160, 169–171, 175, 175n33, 178, 181–185, 249 Satyasiddhiśāstra 124n5 Sautrāntika 4, 6, 12, 14–16, 19, 26, 31, 38, 38n13, 39, 42n22, 47, 47n27, 51, 51n33, 52–53, 55, 115, 124, 135, 140, 144, 159n4, 172, 190, 196, 200n16, 206, 211, 245, 251, 262, 262n29, 263–264, 267, 275–276, 276n2, 277, 288, 288n22, 294–296, 298, 300–302, 312n12, 313n14, 317 sa-vikalpa (with vikalpa) 258n18, 260, 271, 310, 310n5 Schear, Joseph K. 55n40 Schmithausen, Lambert 105n27, 226n4, 228n11, 232, 232n15, 232n16, 233, 234n19, 237n25, 237n26, 288, 294

341 School on the Intrinsic Nature of Real Entities. See Faxiang zong Searle, John 168n20 seed. See bīja seeing-part. See dṛṣṭi-bhāga or darśanabhāga, under bhāga self self theory of memory (stm) vs. non-self theory of memory 310 self-grasping. See ātma-grāha self-apprehension 122, 271 self-awareness/self-cognition/selfconsciousness 34, 53, 115, 117, 121, 123–124, 124n4, 125, 127–128, 128n13, 130–136, 140, 143–144, 144n40, 145, 147n42, 154–157, 160, 160n5, 161, 162n10, 164–167, 169, 171, 175–177, 180–181, 184, 184n42, 185, 209–216, 219–222, 249, 252, 254, 271, 289 See also svasaṃvedana/svasaṃvitti/ ātmasaṃvedana as conventionally real 159, 177–178 as ultimately real 161 as intrinsic awareness 156, 171, 178 constitutive account of 166–167 memory argument for. See under argument perceptual account of 166 sahopalambhaniyama argument for. See under argument transcendental view of 156 self-characterized object. See under svalakṣaṇa self-characterizing 136 self-cognition 53, 117, 123, 124n4, 125, 128n13, 144, 164n15, 209–216, 219–222, 289 self-consciousness 123–125, 127–128, 130–135, 135n25, 136, 143, 145, 147n42, 157, 166, 169, 180, 184n42 self-familiarity 145 self-intimating (svaprakāśa) 31, 53, 123, 127, 127n10, 128n13, 130–134, 144–145, 147–148, 171 self-knowledge. See meta-cognition self-luminosity 124 self-reference/self-referential relation 122, 123n1, 123, 125–126, 128 Sellars, Wilfrid 160, 176–177, 185n43 Sengyou 僧祐 (445–518) 212n8

342 sense data 48, 73, 101n17, 172, 174n32, 185, 278 sense-faculty. See indriya sensory perception. See pratyakṣa, indriya-pratyakṣa sense-cognitions, five 27, 36–37, 37n12, 38, 40–50 sense-consciousness/sensory consciousness 7–9, 20–21, 26, 31, 55, 244–245, 247–250, 255–263, 265–267, 267n40, 269, 271, 276–278, 280–282, 285–293, 295–302, 323, 324n35 shape (saṃsthāna) 4, 7–8, 9n6, 43n24, 75, 221, 245–246, 250–251, 251n5, 258n18, 262–263, 265, 265n37, 265n38, 267–268, 270n46, 271, 312 shared. See sādhārana Sharf, Elizabeth Horton 56 Sharf, Robert 3, 10, 26–27, 27n2, 33–34, 34n8, 41n18, 52n35, 63, 101n17, 102, 135n24, 200n16, 244, 249n1, 258n18 Shastri, Swami Dwarikada 162n9, 167n19, 168n21 Shentai 神泰 (7th c) 285–286, 299 Shoemaker, Sydney 123n1, 131n19, 133n23, 184n42 Shuowen Jiezi 說文解字 309 Siderits, Mark 1n1, 63, 79n20, 106, 106n29, 116, 147n42, 156n1, 157–159, 159n4, 161n7, 162n8, 177–179, 181–183, 199n14, 226, 235, 251n5, 302, 324 Silk, Jonathan 226n5 simultaneous basis. See sahabhū āśraya simultaneous mental consciousness. See mental consciousness simultaneous with the five sensory consciousnesses (msf). single layered cognitive series 288 six sense spheres/six abodes (ṣaḍ-āyatana) 278–282 skandhas (aggregates) 3–4, 4n4, 5, 5n5, 9–11, 37n12, 49, 83, 192–193, 195n8, 198–199, 201, 278–280, 283–284, 317 nāma-skandhas 201 rūpaskandha (aggregate of physical matter) 199, 278 Skandhila 塞建陀羅 (d.u.) 50n29 sliding (ascending) scale of analysis 147, 267 smaraṇa (recollection, act of remembering) 310, 311n7, 312n12, 319

Index smṛti (mindfulness, recollection) 27, 37, 45–48, 64, 64n3, 199, 305, 307, 309–314, 314n16, 319–324 society of mind, theory of 288 solipsism 19, 116, 159n4, 196–198 soteriological 1, 13–15, 26–27, 62–63, 69, 98n14, 102–103, 124, 134, 154, 190, 195, 198–199 space 51, 56, 78, 91n6, 95, 99, 110, 137, 260, 283–284, 322–323 See also ākāśa Spackman, John 28–29, 34, 67, 74, 101n18, 192n4, 200n16, 258n18, 302 spaṣṭa-grahaṇa (vividly grasping) 258, 261–262 spaṣṭatva (vividness) 9n6, 15, 53, 118, 206 śrāvaka (listener) 93, 100n15 Śrīlāta (4th c) 206 srota-āpanna (entrant on the path of cultivation) 124 śrutavāsanā (propensity of the teachings) 93 Stcherbatsky, Theodore 253n9 Steinkellner, Ernst 132n21, 138n32, 143n38, 143n39, 213n10, 216, 222n25, 270n45, 270n47 Sthiramati (7th c) 20–21, 29, 89–92, 100, 107n32, 110, 233n17, 256–257, 264n34, 267n41, 291, 291n33, 292–293 store-consciousness. See ālayavijñāna storehouse consciousness. See ālayavijñāna Strawson, Galen 32n2, 183, 183n41 Strawson, P.F. 129, 147 subject-aspect of cognition/cognitive agent. See grāhaka subjectivity 84, 84n22, 106, 116, 122–125, 129, 129n14, 131n19, 138, 156n1, 156n2, 157, 159, 167, 168n20, 169, 173, 176, 179–180, 182–183, 185, 194, 238 subsequent cognition. See under nirvikalpajñāna successive model (of cognition) 277, 294–295, 298, 301 suffering (dukkha/duḥkha) 3, 9, 15, 19, 35, 49, 84, 119, 139, 176, 190–193, 199, 207, 259 superimposition 18, 118, 146, 195n8, 230–231, 249, 258n18, 318

Index supervenience 8, 28, 30, 62, 68, 72, 74–77, 81, 85, 101n18, 198 svābhāsa (subjective aspect) vs. viṣayābhāsa (objective aspect) 129–130, 132n21 svabhāva (intrinsic nature, own-being [Sarvāstivāda], defining characteristic [Sautrāntika]) 16, 43n24, 44, 66–67, 69, 78, 91n6, 144, 144n40, 146, 157, 170, 181–183, 192n8, 195n8, 201 paratantra-svabhāva (dependent nature) 89–90, 90n3, 90n4, 90n6, 91, 91n6, 93, 98n14, 102, 110, 161 parikalpita-svabhāva (imagined nature, the nature of the imagined) 91, 98n14, 104, 110, 230 pariniṣpanna-svabhāva (perfected nature) 90–91, 91n6, 100, 110, 230 svabhāvavikalpa (inherent discrimination). See under vikalpa svacittajñāna (cognition of one’s own mind) 229, 235–236 svalakṣaṇa (particular, defining characteristic, inherent characteristic, selfcharacterized object) 37, 37n12, 40, 163–164, 164n14, 165, 167, 206, 249, 258n18, 262, 269 āyatana-svalakṣaṇa (particular at the level of sphere) 37n12, 263, 267n42 dravya-svalakṣaṇa (particular at the level of an entity) 37n12, 263 svalakṣaṇa, Dharmakīrti’s definition 268, 268n44 svalakṣaṇa-svalakṣaṇa (particular at the level of a particular) 263 svasaṃvedana/ svasaṃvitti/ ātmasaṃvedana (self-awareness/self-cognition/ self-consciousness) 34, 34n8, 53, 53n37, 121, 128, 128n13, 130–135, 137n28, 140, 143, 144n40, 154–156, 156n1, 157–159, 159n4, 160–161, 163–164, 164n15, 165–172, 175–183, 185, 209–211, 213–219, 222n25, 225, 229, 233–234, 249, 252, 254, 270–271, 289 svasaṃvitti-saṃvitti (cognition of selfcognition) 21n11, 128n13, 209–211, 213, 215, 217, 219, 222 Svātantrika 171n25, 175n33, 181–182 syllogism 318

343 Taber, John 143n38, 226n5 tacit or non-propositional modes of acquaintance 129 tacit self-acquaintance 135 Takatsukasa, Yūki 40n17 Takeda, Hiromich 311n8 Tanney, Julia 177n35 Tantra 34, 63 tatpṛṣṭhalabdhajñāna (cognition subsequent to it). See under nirvikalpajñāna Tattvasaṃgraha (Śāntarakṣita) 167n18, 172n27 *Tattvasiddhi-śāstra (Chengshilun 成實論) (Harivarman) 47n27 Thakchoe, Sonam 268n43 Thanissaro Bhikku 34n7 Theories of consciousness higher-order theory of consciousness See also meta-cognition higher-order account of consciousness 9, 17, 21, 32, 115, 117, 136–139, 195–196 higher-order belief 137 higher-order cognition 17, 21, 115, 136 higher-order perception (hop) theory of consciousness 9, 17, 21, 32, 136 higher-order thought (hot) theory of consciousness 9, 32, 32n2, 116, 121, 134, 136, 138, 197 same-order (first-order) theory of consciousness 32, 32n2 Theravāda 4–5, 34n7 Thompson, Evan 56, 156n1, 162, 162n8 threefold division of the mind (Dignāga) 214n13, 215, 218 Three natures, theory of dependent nature. See under svabhāva nature of the imagined/ the imagined nature. See under svabhāva perfected nature. See under svabhāva Thusness (tathatā) 284 Tillemans, Tom J.F. 198n13, 227n8, 250n4 Tīrthika 147 Tola, Fernando 251n7, 266 Tosaki, Hiromasa 253n9 transparency transparency thesis [about selfconsciousness] 159–160, 190n1, 191

344 transparency thesis vs. opacity thesis  189–190, 202, 206 Treatise on the Buddha Land. See Fo di jing lun Triṃśikā (Vasubandhu) 20, 29, 90, 90n3, 91, 210n3, 231, 260, 317, 319, 319n25, 323n34 Triṃśikābhāṣya (Sthiramati) 29, 89–90, 91n4, 91n6, 92, 107n32, 179n3 trope 6, 14, 193, 201, 244 truth (satya) absolute/ultimate truth (paramārtha-satya) 92 conventional truth (saṃvṛti-satya) 6–7, 12, 66n4, 116, 155, 159, 159n4, 160, 169–171, 175, 175n33, 178, 181–185, 249 two truths, doctrine of 6, 102, 147, 184 Tsongkhapa 176, 185 two-aspectual theory of mental states/dual aspect theory of mind 125 Tye, Michael 67n7, 75, 141n35 Tzohar, Roy 28–30, 34, 63n1, 63n2, 69, 78n19, 79, 90n1, 90n2, 101n16, 103n19, 106n28, 107, 192n4, 258n18 ubhayasiddha (mutual consensus) 266 Ui, Hakuju 226n2 ultimate reality (dravyasat). See under reality ultimate truth. See under truth unconditioned. See asaṃskṛta universal. See sāmānyalakṣaṇa upacāra (figurative) 103 Upanibandhana (Asvabhāva) 69, 71, 93 upādhi (lack of qualification) 140 Vaiśeṣika 305, 307–308 Valberg, J.J. 159n4, 169, 169n24, 180, 180n38 Vasubandhu (4th-5th c) 2, 13–14, 18–20, 29, 39, 44, 51–53, 70n10, 90–92, 92n7, 94, 105, 105n26, 118, 125, 126n8, 131n19, 144, 173n30, 209, 210n3, 219, 225–226, 226n5, 227, 227n9, 231, 246, 251, 282–283, 290, 290n32, 291, 293–294, 305–315, 315n18, 316, 316n20, 317, 317n21, 318, 318n23, 319, 319n25, 319n26, 320, 322n33, 323, 323n34 Vātsīputrīya School 52 Vātsyāyana (2rd-3rd c) 191, 211

Index vedanā (hedonic states, hedonic tone) 3, 279 pain (duḥkha), pleasure (sukha), neither-pain-nor-pleasure (aduḥkhāsukha) 3, 259 vicāra (fine discernment, investigation) 27, 40–41, 41n19, 42, 42n22, 43, 43n24, 44–46, 46n26, 47, 47n27, 48–50, 277, 289–290, 292–293 Vigrahavyāvartanī (Nāgārjuna) 212n7 vijñāna (cognition, consciousness) 3, 5n5, 11, 35, 43, 137n30, 140, 167n19, 190, 192, 210, 216n15, 265n36, 265n37, 275, 284n12, 288, 306 See also discerning awareness Vijñānakāya-śāstra (Apidamo shishen zulun 阿毘達磨識身足論) 39, 39n15, 281, 281n6 vijñapti-mātra (consciousness-only) 11, 319n25, 322n33 vikalpa (conceptualization, conceptual construction) 3, 7–9, 12, 15–16, 26–27, 42–44, 46, 47n27, 65–66, 66n4, 69, 71, 83, 90n1, 92, 94–96, 100, 100n15, 101n17, 102–103, 109, 111, 118, 222, 222n25, 247–249, 252, 258n18, 261n27, 269, 271, 289, 294 See also conceptualization (abhi-)nirūpaṇāvikalpa / abhyūhanāvikalpa (discrimination through examination/ conceptual construction in terms of reasoning) 42, 43n24, 46, 290–291 anusmaraṇavikalpa (discrimination through recollection/ conceptual construction in terms of recollection) 42, 43n24, 46, 290–291 svabhāvavikalpa (intrinsic conceptual construction, inherent discrimination) 27, 40, 42–43, 43n24, 44, 48, 290, 293 trivikalpa (three kinds of discrimination) 27, 31, 38, 40–43, 43n24, 44, 46–49 Viniścayasaṃgrahaṇī (Asaṅga) 103, 261n27, 285, 287, 294, 298 vipāka (maturation) 105 Visuddhimagga (Buddhaghosa) 33

Index Vitakkasaṇṭhāna-sutta 41n18 vitarka (coarse discernment, inquiry) 27, 40–41, 41n19, 42, 42n22, 43, 43n24, 44–47, 47n27, 48–50, 277, 289–290, 292–293 vividness. See spaṣṭatva Viṃśikā (Vasubandhu) 5n5, 20, 118, 219, 225–226, 226n2, 226n5, 228–229, 232–233, 236, 238, 305, 307–308, 317, 317n21, 319, 319n26, 320, 322n33, 323 viññanam anidassanam (featureless consciousness) 34n7 viṣaya (sense object, perceptual object) 35, 38 von Neumann bottleneck 200 von Rospatt, Alexander 56 Waldron, William 105n27, 129n14 Walshe, Maurice 36n10 Walter, S. 141n35 Watson, Alex 34n8 Wayman, Alex 131n18, 252n8, 310n5 Wei shi ershi lun shuji 唯識二十論述記 (Kuiji) 225–226, 233 Westerhoff, Jan 1n2, 11n7 wholesome (kuśala), unwholesome (akuśala), indeterminate (avyākṛta) 148, 259, 294 Williams, Paul 34n8, 72, 136n27, 145, 155, 167n19 wish-fulfilling jewel. See ciṇṭāmaṇi Wittgenstein, Ludwig (1889–1951) 116, 129, 130n16, 134, 196n9 Wŏnch'ŭk 圓測 (613–696) 20, 267n41 Xuanzang 玄奘 (602–664) 20, 22, 43–45, 50n29, 131n18, 161n7, 209–212, 218, 220, 222, 226n3, 231, 247, 250, 251n6, 251n7, 258n18, 260n21, 261, 261n27, 264n33, 270n45, 270n47, 272, 276, 278, 280, 285n15, 291, 291n33, 294n44, 296, 305, 307, 311, 311n8, 312n11, 314n17, 316n20, 317n21, 320n28 Yamabe, Nobuyoshi 38n13, 40n17, 56, 226n5, 232, 323n16 Yamaguchi, Susumu 92n8, 211, 211n6

345 Yanshou 延壽 (904–975) 220–221, 221n22, 221n23 Yao, Zhihua 34n8, 52n35, 53, 53n37, 124n4, 130, 131n18, 133n22, 192, 200n16, 202n18, 234n19, 234n20, 245–246, 252n8, 261n27, 271, 288n22, 288n24 yathārtha (rushi 如實) (accordant with reality) 238 Yijing 義淨 (635–713) 20, 131n18, 226n2 Yogācāra 4n3, 4n4, 5n5, 10–11, 11n7, 12, 18–22, 27–31, 34, 38n13, 63, 63n2, 66n4, 69, 89–90, 90n1, 92, 98n14, 99–100, 100n15, 101, 101n16, 101n18, 102–104, 104n21, 105–106, 106n28, 107, 107n30, 107n31, 107n32, 110, 115, 117–118, 135, 140, 144, 146–147, 154–155, 156n1, 158, 160–161, 170–173, 173n30, 181, 183, 199, 207, 209, 210–211, 220, 222, 226, 226n3, 227–228, 231–233, 233n17, 239, 244–245, 254, 272, 275–280, 284, 285–288, 290–291, 293–295, 298, 300–302, 314n15, 314n16, 315, 319, 322–323, 323n34 Yogācāra-Sautrāntika School 12, 14–16, 19, 27, 159n4, 190, 196, 200n16 Yogācārabhūmi/ Yogācārabhūmiśāstra 51n33, 103–104, 104n22, 227n8, 255, 255n14, 261n27, 264, 264n33, 276–277, 279, 279n3, 280, 282, 284, 284n12, 285, 285n13, 285n15, 285n16, 286, 286n18, 287, 287n20, 287n21, 290–292, 292n38, 292n40, 293–294, 294n44, 294n45, 295n48, 296, 296n49, 296n50, 297–298, 298n54, 299, 299n55, 301 Manobhūmi [of Yogācārabhūmi] 279, 285n15 yogi 72, 104, 107n30, 109, 109n34, 118, 226–227, 227n9, 249 yogic perception. See yogi-pratyakṣa/ yogipratyakṣa Yoshida, Akira 254n10, 270n48 Yoshimizu, Chizuko 268n44 Zahavi, Dan 84, 128n12, 129, 129n15, 131n19, 134–135 Zenju 善珠 (723–797) 215n14, 220, 220n20