In its history, the scholarly study of meditation has been the preserve of orientalists, historians and phenomenologists of religion, and, more recently, psychologists of consciousness. These investigators have, on the whole, been mindful of philological, textual, and descriptive matters. Little attention has been given to philosophical, theoretical, or sociological aspects of meditation. In particular, the many possible connections between characteristics of meditational practice and institutionalized theories of knowledge, brought to light in other areas by the sociology of knowledge, have been ignored.
    By way of innovation, I want to see how epistemological perspectives might illuminate the shape of Buddhist attitudes toward the gradual or sudden attainment of enlightenment. Using a modified and rather informal structuralism, I want to compare the structures of institutionalized theories of knowledge with the structures of meditational practices and beliefs to see whether one might understand the characteristics of these practices and beliefs in terms of their underlying epistemological structure. I want to argue that one can plot the salient characteristics of meditational practices—here, whether enlightenment occurs gradually or suddenly—as symptoms of the presupposed structure of their institutionalized theory of knowledge.
But before embarking on the critical study of meditational practices we ought to first clarify just what the Buddhists themselves thought about gradual and sudden enlightenment, and how they conceived the relation of these aspects of meditational practice to their beliefs about the acquisition of knowledge.

I. APPROACHES TO THE PROBLEM OF GRADUAL AND SUDDEN ENLIGHTENMENT: NATURALISM AND PHENOMENOLOGY

    It is commonplace to read that Theravāda Buddhism teaches that nirvāna is attained gradually and that Chan or Zen Buddhism teaches 'sudden' enlightenment. Little is said about the bases for studying what such meditational claims mean, and less is said about the logical grammar of words as peculiar as 'gradual'. Typically, it is facilely assumed that this problem is merely a factual matter about temporal duration. On this view, to say enlightenment is 'gradual' usually means that it takes a long time for this quasi-mental state to occur. Such a claim does not seem logically different than saying that it took a person a long time to get 'dizzy' or 'drunk,' and so on. Now, to put a factual stress on this matter should immediately strike anyone familiar with the pragmatic attitude of (early) Buddhism as odd. Surely, it must have been unedifying for an early Buddhist to be concerned with rather speculative matters of fact. Is this just an example of 'corruption' in early Buddhism, analogous to the storied medieval Christian scholastic problem of angels on the head of a pin? What could be the practical salvific value of talk of gradualism in various Buddhist contexts? What could have been the possible interest for an early Buddhist in saying that enlightenment was to be attained gradually?
    Despite such considerations, scholars of meditation have persisted in treating meditational discourses as mere descriptive matters of fact. This is true, even though these scholars disagree implicitly about what counts as a 'fact', or, perhaps more accurately, stress different views about what counts as a fact. Basically, two such emphases seem current. As applied to my earlier example of dizziness, one may take the fact of dizziness to be an experience in which case one might term such an approach 'phenomenological.' However, one might feel required to seek facts in some supposedly underlying neurophysiological process, in which case one might term such an approach 'naturalist.' Although both naturalist and phenomenologist would agree that temporal duration was crucial to the meaning of 'gradual', they would not agree about the nature of what endured in time.
    I am convinced both these approaches emphasize the wrong things about Buddhist gradualism—for whatever different reasons. Not only does the Pali Canon tell a more complete story, but another order of analysis of the texts is required. Basically, I believe those tempted by either of these two approaches mistake a norm for a matter of fact, and that where a fact may be indicated, it tends more often to be a spatial fact, rather than a temporal one. Although the temporal and the factual question may not be without interest, it does not seem to be the chief concern of the Pali Suttas. Here, the Buddha recommends a particular mode of life—an issue which reads far beyond any such unedifying factual matter of the speed of the attainment of nirvāna.
    Taking the temporal point first, it would seem important to note that the term 'gradual' is ordinarily used in two quite different ways: Insofar as 'gradual' is used factually, it may indeed mean something temporal, like 'slow.' But, it may also mean 'graded.' It may be a temporal word just as easily as it may be a spatial one. The same is true of the Pali term anupubba, as I shall show in the discussion of the Pali Canon's view of "gradual" enlightenment. Thus, "gradual' is like other words that play across the temporal and spatial conditions of experience. Does a 'dashing' man need to be fleet of foot? Does a 'snappy' dresser need to be quick with buttons and zippers? Although spatial and temporal uses of 'gradual' often coincide, they need not do so. Doing something gradually—by degrees, in stages—may take less time than trying to do the same task at one go. Gradual methods are, indeed, often devised to save time—say, in building a house, taming a horse, writing a book, or attaining nirvāna—especially when contrasted to available alternatives in achieving the same sophisticated result. Perhaps, part of the reason this spatial sense of 'gradual' escapes our attention may have something to do with the fact that the ordinary English contrast word, 'sudden,' does not seem to have a spatial sense at all. It only seems to have temporal uses, and thus by analogy, we think of 'gradual' in the same way. Attention to the contexts of the discourses on gradualism tells another story.
    As one might emphasize either temporal or spatial aspects of gradualism so also have scholars of meditation emphasized different senses in which meditational forces are facts. Through their reliance on neurological research, the psychologists of consciousness exemplify a naturalistic approach. The question of gradual attainment of enlightenment would become a question to be settled by measuring the duration of 'extent' of certain neurological processes. Now, the psychologists of consciousness have not, to my knowledge, dealt with our particular problem. Yet, it would seem important—at least in passing—to represent their increasingly popular work in this context— even if I am forced to extrapolate from their more general work on meditation. They seem to exemplify an extreme contrast to the kind of epistemological approach I advocate, since they seem to avoid the whole issue of the theory-ladenness of meditational "facts."
    A characteristic of this loosely related group of writers is their reliance on quantitative neurological investigation of meditation. Typical of this view is the work of Dan Goleman. Here, EEGs supposedly get the investigator behind "abstract concept"—'the realm of discourse' (the beliefs and reports of meditators) to the "raw data." [1] Conveniently, this move (if possible) liberates the investigator from the need to deal with troublesome institutions, beliefs, theories, and critics! Thanks to the EEG one reaches the promised land-of-value-free inquiry. Consistent with this supralinguistic approach, no arguments will be found supporting such claims that a conceptually neutral realm has been reached. In their stead one finds pronouncements and decrees—poor surrogates for solutions to our awkward epistemological position. But, instead of evading epistemological issues, I believe we ought to face them squarely: What presuppositions, theories, beliefs, and institutions condition mystical or meditative experience? What sense can one make of truth claims made under such conditions?
    It is to the phenomenologists of religion, like Winston King, however, that one must look for the most direct discussion of our problem. In a comparison of Theravāda and Zen meditation, King concludes that there is really no difference between sudden and gradual attainments of enlightenment. As one might expect from a phenomenologist, King believes that there is in fact no difference, because there is no experiential difference between sudden and gradual attainment of enlightenment.

"suddenness" or "gradualness" of enlightenment ... appears to depend primarily upon emphasis and/or point of specification. One may choose to emphasize the prior preparation .. . and call it "gradual"; or one may stress the experiential breakthrough and call it "sudden." But in both Theravada and Zen, there are development and pinpointed breakthrough.[2]

    For King, this virtually closes the case. If, however, one takes seriously the theory-ladenness of meditational experiences, the hard questions just begin. Why, indeed, the differences in "stress," as King himself is compelled to ask? Why the canonical, commentarial, and modern norm among Theravādins that nirvāna comes gradually? King's reply to his own question is couched in terms of "the Indian penchant for classification and analysis" versus the "Sino-Japanese impatience with metaphysical speculation and a fundamental reliance upon intuitional apprehension of existential truth."[3] One wonders what the Buddha would say to the implication that he was not impatient with metaphysical speculation. Or what the Hua-Yen philosophers would say to the implication that they were not among the most supreme speculative metaphysicians of all time. But, like many cultural generalizations, King's also contains an unexpected germ of truth. Surprisingly, King drops the matter at this point. Yet, one should not be altogether puzzled, since King's approach will not let him push beyond the reports of experiences to levels of structuring which may give rise to these experiences. I want to suggest that an appreciation of fundamental attitudes toward knowledge may help stimulate understanding of these divergent views of what may or may not be identical processes or experiences- In part, I aim to reinforce Jayatilleke's views about early Buddhist empiricism by arguing that its underlying structural pervasiveness accounts for much of the character of early Buddhist belief in the gradual attainmentof nirvāna.[4] Contrary to what Buddhist empiricists themselves might believe, I believe that their empiricist epistemology is symptomatic of a deep yet compromised empiricist structure.

II. THE PALI CANON ON GRADUALISM

The classical and principal discussions of gradualism occur in four places in the Majjhima Nikāya (MN). In two condensed analogies, the Buddha teaches what has become known crudely as "gradual enlightenment." Both these analogies—taming a thoroughbred colt (MN 1.445-446; MN III.1-6) and mastering complex skills (calculating and archery: MN III.1-6)—indicate much of the character of gradualism, which I shall explain shortly. In MN 1 (480-481), the Buddha deals directly with gradual attainment of paññā. Contrary to popular misconception, this shows that the distinction between gradual and sudden enlightenment differs from the distinction between those who attain nirvāna by paññā. and the jhānas, respectively. As the Buddha implies in MN I(478ff), the paññāvimutta seems to achieve nirvana immediately (in both spatial and temporal senses), because he has previously achieved those stages of sanctity which others may only now be set to achieve.
    The compounds of the Pali anupubba (Skt., anu-pūrva) "gradual" are numerous, and occupy nearly three columns in Trenckner's Critical Pali Dictionary.[5] For the purposes of this article, I shall treat only the relevant compounds and deal with the pertinent aspects of their logical grammar. This pragmatic approach may leave the linguistic survey of these compounds incomplete, but I believe I have covered all pertinent issues from the philosophical point of view. The compounds of anupubba have both broad and narrow references: they may refer to the entire effort of attaining enlightenment as well as to the stages of meditational attainment and pedagogical practice. Thus, terms like anupubba-kārana, "gradual training," anupubba-kiriyā, "gradual working," anupubba-patipadā, "gradual progress," anupubba-samā-patti, "gradual attainment," and anupubba-sikkhā, "gradual training" refer broadly to the systematic or successive character of the whole Buddhist way of life, from first silas to final release. Considering the narrower context of the jhānas, one completes a gradual cessation of consciousness (anupubba-nirodha), or one is said to come to dwell in certain graded levels of meditational abodes (anupubba-vihāra). Finally, one may speak about pedagogical matters, in what seems a prescriptive epistemological way, about the Buddha's normative gradual method of instruction (anupubba-kathā) and its correlative, the student's gradual method of study or training (anupubba-sikkhā).[6]
    Some of these notions need explaining. The early Buddhists held definite beliefs about the details and reality of the mental landscape. The meditator was thought to ascend a graded trail of real, though impermanent, mental steps (jhānas), one after another, until the summit of nirvāna was won. It is true that nirvāna is not itself another jhāna and, that, strictly speaking, is not necessarily 'won' by meditation: it is not the causal product of the process of meditation. Yet, there is some relation between meditation and nirvāna, although the precise nature of it is often difficult to make out. More on this matter shortly. Moreover, the progress of the meditator through the jhānas was also thought to be open to precise location in terms of a psychological map of the real, though impermanent, mind. To follow the Buddha meant, in part, to accept his map of the mind—at least provisionally for the purpose of testing its accuracy and its utility for attaining release. In meditation, these directions were, in turn, tested for their truth—although, of course, the question of vicious circularity is conveniently passed over by the Buddhists. One might also add that as the route to nirvāna by meditation was graded, so was the goal itself, in some sense, graded. Early Buddhist notions of levels of accomplishment, like "Streamwinner," "Once-returner," and so on, seem to point in the same direction of gradual-graded-attainment.
    Apart from these descriptive uses of the grades of attainment, two aspects of the early Buddhist attitude to saving knowledge are also termed "gradual" although in a different sense than we have seen thus far. The context of this new sense of "gradual" is the classical Buddhist milieu of learning and teaching. Gradual teaching or instruction (anupubba-kathā) refers to the Buddha's normative analytical and graded pedagogy.[7] This method of instruction exemplifies the Buddha's use of skill-in-means (upāya-kosalla) [8] which, as Jayatilleke has argued, encompasses a kind of openness to falsification and corresponding obligation for verification.[9] Because of his compassionate care and sympathy for humanity and its physical and intellectual suffering, the Buddha prescribed teaching the dhamma in orderly and logical ways, tailored to the needs and capacities of his listeners, and open, in large measure, to dispute and verification.[10] Although, at times he speaks in the didactic mode, the Buddha eschewed an abrupt, paradoxical, or esoteric mode, typical of the thwacks and slaps of some Zen Buddhist pedagogy and the later Mahāyāna uses of upāya, respectively.
    From the perspective of the student, gradualism requires a correspondingly earnest methodical and analytic study of the dhamma. A student is responsible for testing and verifying the dhamma experientially. If one follows Jayatilleke here, epistemological gradualism—this attitude of experiential scrutiny—applies to all aspects of the dhamma—both to preliminary matters as well as to those which arise at rarified meditational levels.[11]
    One cannot then conclude that the gradual attainment of enlightenment primarily meant that nirvāna came slowly, or that it was the norm of the slow-witted. This, at any rate, is not the view of the Pali Canon. For the early Buddhists, gradualism was a complex notion, involving both the description of a graded model of the meditational and cognitive landscape, along with certain values or prescriptions about the proper epistemological attitude of scrutiny and experiential testing needed at all levels of the teaching and learning process of attaining release.
    In another discussion on Theravāda meditation, Winston King underscores this opposition of description and prescription by repeating it in terms of the contrast between jhānic and vipāssanic aspects of meditation.[12] Although these two aspects are "set in tension with each other,"[13] they also complement each other.[14] Vipāssana (insight) supplies "critical awareness"[15] of the jhānic attainments, a "reviewing of the path."[16] The jhānic route thus describes a journey through a series of gradually ascending stages, while vipāssana censors and scrutinizes the quality of those achievements.
    For King, the central question still remains why these two disciplines are combined at all. What is achieved by their combination in the trance of cessation (nirodha-sampātti), or in the Theravāda tradition as a whole? Once again King couches his explanation in experiential or phenomenological terms:

    The jhānic discipline contributes meditational expertise, which may strengthen the concentration of the vipassanic meditator ... and very importantly gives a quality of depth and lastingness of experiential attainment... . On the reverse sides, vipāssana keeps the whole jhānic progression within Buddhist bounds so that none of its utterly peaceful states will be construed as the final goal of meditation.[17]

    Now,I do not wish to quibble with these admirable conclusions. They strike me as sensitive and germane. Indeed. I should like to confirm them and also take them a step further beyond the phenomenological level which they occupy. I am urging the reader to consider that there are deeper reasons behind this felicitous conjunction of meditational modes, which I, first of all, identify as epistemological in nature. My 'hunch' is that the connection between the jhānic description of cognitive growth with the vipassanic epistemological scrutiny suggests a fundamental connection with a comprised empiricist syndrome recently spelled out by Ernest Gellner.[18] There are parallels to the specific conjunction of the jhānic and vipāssanic modes of meditation in similar conjunctions in the general development of empiricist approaches to the growth of knowledge. Jhānic and vipāssanic modes of meditation are joined for the same reason similar aspects of the general empiricist theory of knowledge are joined.

III. THE "GHOST" MEETS THE "MACHINE"

    One can speak of an 'empiricist syndrome' today largely because it has been the subject of intense debate by modern epistemologists. This is perhaps especially true of north Atlantic analytic philosophy, although the ferment on the continent in Marxist and structuralist circles seems to focus on similar issues from the opposite philosophical shore. Among philosophers of science, Ernest Gellner has been particularly active in recent years in this area. Gellner believes one ought to distinguish two moments in the life of empiricism as it has developed in certain favored contexts: empiricism is both a description of how knowledge works and a prescription about what ought to count as knowledge. As a description, empiricism offers a mere "toddler's toy" model,[19] far too crude and simple to reflect the complexity of cognition; but, as a prescription, it provides a useful "touchstone,"[20] admirably stating a clear normative attitude toward the limits of cognition. In this latter sense, empiricism actsas a "censor" or "selector,"[21] laying downtwo imperatives: "Be sensitive to whether or not assertions are testable (in the specified approved manner)! Spurn those which are not". [22] Gellner realizes that an empiricist would not typically recognize that empiricism itself rests of prescriptions. Indeed, part of what being an empiricist has meant in the past, has been bound up with the conviction that our cognitive situation is grounded in unbiased observations. But, for Gellner and any one of the numerous critics of empiricism today, this is just not so.
    As for the empiricist "toddler's toy" model, it can be summarized along the lines of an acquisitive enterprise. Beginning with an active external world and a passive internal one, the inner world of 'concepts' or 'knowledge' is built up by accumulating sense-data. But, since all one 'knows' consists of sense-data, the existence of a world behind sense-data becomes theoretically problematic, and unless something intervenes, one is led down the primrose path to phenomenalism and nominalism. In this condition one can still 'generalize,' by assembling sense-data into complex 'beliefs' or 'ideas' by 'induction.' The truth of these beliefs is tested or verified by 'correspondence' with the facts of sense experience. This comparison of simple and complex is achieved through the process of 'analyzing' complex 'beliefs' into their constituent sense-data. Normative statements analyzed in this fashion reveal no world of 'good' or 'bad,' but mere pleasures, pains, or emotions. Science, especially in its reductionist and impersonalist moods, represents the kind of model explanation of the world of experience to which all other cognitive enterprises should aspire.
    For Gellner, empiricism tends toward solipsism and eventually idealism—as long as it remains pure. After all, experience is just my experience. My experience is composed of private sense-data, and the existence of the external world is necessarily left in doubt. Yet, historically and, in Gellner's view, happily, empiricism did not in every case actually retain its purity and develop into idealism and solipsism; Bishop Berkeley was not the sole heir to the empiricist tradition. The Utilitarians, Locke, Russell, and others, claim this birthright as well. Their thought embodies a salutary convergence of empiricism and materialism—the "ghost" and the "machine," in Gellner's words. These thinkers sought a "stable, recognisable structure that could somehow be reached through the qualitative sense-data available to the ghost."[23] Because of their confidence in knowing the world, they also believed that the world was improvable, and that analysis and scrutiny were both worthwhile and appropriate activities for human beings.

IV. EARLY BUDDHIST EMPIRICISM AND MEDITATION

    Gellner's myth about this compromised empiricism fits remarkably well with K.. N. Jayatilleke's account of early Buddhist theory of knowledge—especially in the way it resists idealism (as later Buddhist thought does not) and allies itself with materialism.[24] Early Buddhism populates the vacuum between experience and the otherwise noumenal world with real, though impermanent and causally conditioned, causally agent, material sense-data. These sense-data, in turn, activate the causally passive (initially, at any rate) and material mind, producing 'knowledge' of the world. For both Gellner and early Buddhism this convergence of "ghost" and "machine" reinforces the characteristic empiricist epistemological attitude of analysis.[25] This analytic spirit—like perhaps the "spirit" of Protestantism or capitalism in Weber—fits with the spirit of the development of traditional empiricism and early Buddhism. Both take the world seriously, because it is not illusory; both exhibit a "salutary censoriousness" which "seems only to come when cognitive hope and confidence have already been raised high."[26] This is why both Gellner's compromised empiricism and early Buddhism (surprisingly and in different ways to be sure) lead to "puritanical orderly world-reform and cognitive exploration," rather than to Schopenhaurian pessimism, aestheticism, and mysticism, on the one hand, or to indulgent hippie grooviness, on the other.[27]
    To those who imagine Buddhism to be Schopenhaurian, pessimistic, mystical, and so on, this claim will come as a shock. And, it is true that much of the Buddhist tradition has been all these things. Yet, Jayatilleke's research, for one, has done much to rectify this image of Buddhism—at least as it seems to have taken shape in the Pali Canon. The Utilitarians, for example, "took the world seriously." But, this meant attention to political reform, technological development, and cognitive exploration in the natural sciences. With the early Buddhists, this earnest spirit took the form of individual ethical and psychological reform, the establishment of an alternative model society—the Buddhist monastic community, the Sangha—cognitive exploration and therapy aimed at seeking the psychological roots of suffering morein the style of Freud and the psychotherapists.[28]
    This is not to deny the differences between Gellner's compromised empiricism and early Buddhist empiricism; it is only to show that they are not differences of "spirit." Moreover, in some ways, early Buddhism is evenmoreoptimistic than its counterparts in European empiricism. It stands for the possibility of the radical development of human cognitive potentials: Men can know the real nature of the world and nirvāna. This enlarges the range of experiential knowledge, taking in meditational states, kinds of ESP, and states transcendental to the ordinary man.[29]
    The cognitive optimism of early Buddhism rests, in turn, on the presuppositions underlying the theory of meditation outlined earlier. The Buddhists thought they knew how the mind worked and what techniques would best serve to enable it to work for human happiness. Insofar as early Buddhist meditation methods are concerned, they are specific to the compromised empiricist theory of knowledge spelled out by Gellner. One can, in fact, generate the model of early Buddhist meditation merely by reversing the order of the empiricist model of critical accumulation of sense-data. As one had gradually accumulated sense-data and passed them before the inner censor, the "ghost," before risking knowledge-claims, so also in meditation one gradually surpasses classes of sense-data experience and knowledge. Urged on by vipāssana criticism, the meditator presses along the jhānic route to higher meditational levels, completely stripped of sense-data information.
    Thus, reliable ordinary knowledge as well as nirvāna require gradual, diligent, and critical attention—analytic care in sifting our perceptions and beliefs. In meditation, this becomes even more severe as the meditator empties the mind of these data, noting their content and form as they are transcended until nirvāna itself is attained. One is not typically encouraged to leap to conclusions (or nirvāna) in early Buddhism. One is invited to analyze and verify the dhamma experientially and ultimately in meditation. The meditator initiates a relentless and deliberate selection process, which seeks to liberate the perceiver from the bondage of the inward flowof causally agent sensations. In meditation, a Buddhist tries to understand sense-data, and therefore knowledge, in their own terms, and declare them for what they are.
    All this makes for a measured and certain optimism about man's potential for salvation unaided by occult power or cosmic fate. In the context of this analytic, trial-and-error cognitive quest, one is advised not to expect rapid results, although these could, of course, occur. The early Buddhists encouraged persistence. Effort brought results. The point was to keep at it, to form the habits of mind and action which would surely (but not automatically) bring results.

V. GRADUAL AND SUDDEN ENLIGHTENMENT: HISTORIC DEBATES 1: CHINA

    Thus far, I have tried to illuminate the nature of early Buddhist meditation and the belief in gradual enlightenment by appealing to the notion of early Buddhist empiricism. In a nutshell, I have argued that early Buddhist meditation theory is imbedded in a compromised empiricist epistemology and, as such, will  reflect salient characteristics of this epistemological syndrome. Even though ordinary knowledge requires accumulating sense-data, both processes occur by 'gradual' means—in both the descriptive and prescriptive senses of that term. As a structuralist, I have shown that Gellner's compromised version of empiricism is homologous to early Buddhist empiricism in both descriptive and prescriptive dimensions. Meditation in early Buddhism constitutes a counterpoint variant of this common theme, seeming for the most part a structural inversion of the empiricist statement about ordinary acquisition of knowledge.
    The critical reader will want some test of this thesis. And, if structuralism is not to become just another occasion for clever dialectical shenanigans, structuralists must offer some check on their own method. The perfect test of this thesis would be a debate between a proponent of early Buddhist empiricism who held the gradualist position, and another kind of Buddhist who held the sudden position—typically a Rinzai Zen Buddhist. The nature of the test would be to see if one could correlate opposed beliefs about the attainment of enlightenment with opposed epistemological beliefs—understanding all the while that both kinds of epistemologies may operate in these contexts in compromised forms.
    In the history of Buddhism, the issue of gradual and sudden enlightenment has arisen on two conspicuous occasions: the eighth-century controversies between the Northern and Southern schools of Ch'an Buddhism in China, and between the Indian and Chinese parties at the Council of bSamYas (792-794) (the so-called Council of Lhasa) in Tibet.[30] Of the two, the Chinese controversy gives fullest treatment to the sudden position. Indeed, the focus classicus of the sudden view remains the Platform Sutra of the Sixth Patriarch, attributed to the 'victor' of the debate and founder of the Southern school of Ch'an, Hui-Neng (638-713). Thanks to Yampolsky's recent research of this text and its historical context,[31] much has become clear. For one, Yampolsky argues that one should attribute the authorship of the sūtra to Shen-hui, one of Hui-Neng's disciples, rather than to the sixth patriarch himself. Together with Dumoulin's work in this area, one can be reasonably certain in correlating Hui-Neng's sudden theory of enlightenment with a certain epistemological-cum-ontological position opposed to that of early Buddhism.
    We know that Hui-Neng (Shen-hui)[32] taught the "sudden" attainment of enlightenment against the celebrated Ch'an teacher, and sixth patriarch according to the Northern school, Shen-hsiu. But, what did he mean? Dumoulin claims that Hui-Neng even makes it the sole criterion for orthodoxy![33] What can be contained in this cryptic claim to merit such importance? And what can explain the fierce attacks Hui-Neng aimed at Shen-hsiu? Well, Hui-Neng most certainly did not mean enlightenment was "easily obtainable" or even quickly won," [34] although these were not ruled out. Like the early Buddhists, Hui-Neng had higher purposes in mind. Both were concerned to make certain points about human psychology and knowledge, using the idioms of temporal duration and spatial levels when these suited their purposes. Both seem to insist, quite often without apparent purpose, that enlightenment occurred in a way harmonious with their practices and basic views.
    Dumoulin[35] and Yampolsky agree that the belief in sudden enlightenment has two sides. Negatively, it denies that the goal, prajñā, can be produced by a "step-by-step process of meditation,"[36] dhyāna—odd, one would have thought, for the Dhyāna school (Ch'an) to assert. Positively, it was a way of asserting the truth of a priori nonduality—that prajñā is "something possessed from the outset by everyone."[37] The point is to realize the imminent a priori nature of enlightenment and to let it shine through. Meditation cannot effect enlightenment because, strictly speaking, meditation and the passions it seeks to purge are ontologically empty and illusory.
    Thus, at bottom, the doctrine of sudden enlightenment is a way of denying the jhanās and of asserting the a priori nature of enlightenment in the idiom of meditational practice. As the early Buddhists set out to operationalize early Buddhist empiricism with the descriptive and prescriptive senses of gradualism, so also does Hui-Neng seem intent on operationalizing his own philosophical position in the sudden view of the attainment of enlightenment: there are no real—even impermanent—grades of enlightenment; thus there is no need to test for a priori enlightenment, since all beings are enlightened by nature.
    One will recall that the Pali Canon would certainly tell another kind of story. Although meditational progress through the jhanās does not causally produce nirvāna on the early Buddhist view, in conjunction with vipāssana, it is one important way to attain it. However impermanent they may be, one seeks to transcend the constraints of real (though impermanent) mind and world. Impermanence itself signals that progress can be made. But, early Buddhist soteriological optimism could not lead to the seemingly exaggerated optimism embodied in the belief in a priori enlightenment—that the battle was already won, or that virtually no battle needed to be fought. Nirvāna, on the other hand, transcends experience without being prior to experience. It is not, strictly speaking, posterior to experience either, since it is not, as it were, an inductive, empirical generalization, or caused by meditation. If I may be permitted a neologism, the word 'transposterior' (to experience) may capture the flavor of the relationship of nirvāna to ordinary experience. By this I mean that nirvāna is not a priori, and only can be said to be a posteriori if one stipulates that it is held to transcend experience.[38] Historically, this position may have arisen from conflict with brahminical rationalists, if we follow Jayatilleke's[39] suggestion. What remains important is the early Buddhist aversion to apriorism—even if it meant constructing an empiricism which finally may have (to put it charitably) transcended itself in the special case of the nature of nirvana. Hui-Neng and the Southern school of Ch'an Buddhism felt no such aversion for the a priori. In fact, they celebrated it, and consequently thought that it merely had to be seen beneath the surface of an already illusory world. Enlightenment was 'sudden' because it was a priori and without even ontological competition from an impermanent world.

VI. GRADUAL AND SUDDEN ENLIGHTENMENT: HISTORIC DEBATES II: TIBET

    The second classic locus of this debate is the late eighth-century controversy which occasioned the Council of bSam Yas (so-called Council of Lhasa). Here, the Indian Mādhyamika logician, Kamalaśīla (742-804) argued a gradualist position against a Chinese Ch'an teacher, Hva San, and his Tibetan allies, the rDzogs-chen. Far more importance is attached to this debate than may seem warranted. Yet, the issue was clearly thought to have been central to the subsequent development of Buddhism in Tibet. Our accounts of the debate records the point of view of the victor, in this case Kamalaśīla in his own Bhāvanākrama.We learn little of the views of Hva-San and his company from this text and are thus led to speculate about their fuller form and the possible relationship with the earlier teachings of Hui-Neng and his school. Although the connections between these two are not certain, many similarities of points of view can be established, which in themselves may point in the direction of relationship.
    The interesting thing about Kamalaśīla is that he seems to argue a gradualism similar to what we have discovered in early Buddhist meditation but, at first sight, without sufficient theoretical basis to do so. His philosophical position, as best one can make out from the often conflicting accounts of it, is exceedingly rich and complex. He seems at once a Svatāntrika Mādhyamakin,[40] Śūnyavādin, as well as logician and pragmatist in the tradition of Dharmakīrti.[41] Historians of Indian philosophy have also identified him as a critic of the Yogācārins.[42] Kamalaśīla, himself,[43] seems to recognize that these philosophical positions produce in him a certain amount of intellectual and practical tension. This is so especially in connection with his desire both to acknowledge the transcendent primacy of Śūnyavāda monism, along with the rather mundane, though nonetheless wholehearted, devotion to the bodhisattva ideal of compassion and meditation. What makes Kamalaśīla interesting then, is his conviction that enlightenment comes gradually and that one should press on with dhyāna and karunā, despite the awkward higher truth of the Śūnyatā.[44]
    This cannot have been a concern original to Kamalaśīla. Other Mādhyamikas must have shared it. But, it must have been especially acute in th like Hui-Neng, taught sudden enlightenment in the sense that meditation in the progressive manner was unnecessary.[45] Perhaps reflecting the supposed Yogācāra background of Ch'an, Hva-San teaches the idealist view that thought is at the root of all se face of Hva-San's idealist monistic teachings which reflected no such qualms about pressing on with the worldly exercises of dhyāna and karūna. One need merely stop thinking to stop suffering. And thinking could be stopped suddenly—without progress through the jhānas or bodhisattvabhūmis.[46] In this way a priori enlightenment simply shines through. Hva-San sounds very much the Yogacārin or close relative of Hui-Neng's Ch'an Buddhism in this passage:

[We] ourselves [are] coessential with the Buddha, and all representations which constitute the world being illusory or a magic play of the Absolute.... What we need is only to jump ... from the plane of representations into that Buddhahood, our true nature, by sudden elimination of those mental representations. We must arrest the play of their emanation, stop our mind, and see into our own nature.[47]

    But, what is it about having the Buddha-nature within us that requires a sudden interpretation of the attainment of enlightenment, along with the rejection of the jhānas, analysis, and the compassion of the bodhisattva? Except for the doctrine of a priori enlightenment, grounded in the possession of the Buddha-nature, Kamalaśīla and Hva-San would seem to share at least the transcendental monism, characteristic of both Śūnyavāda and Yogācāra, respectively. One will, of course, want to make appropriate qualifications for differences in these characteristics of the Absolute. Yet, in spite of that, one wonders how and why Kamalaśīla can commit himself so thoroughly to the worldly practices of dhyāna and karūna, knowing full well that these are onto-logically insubstantial? Has not Hva-San really drawn the natural consequences of transcendental monism? Is not Kamalaśīla quixotically supporting some venerable, but outmoded, tradition of the sūtras, which by some kind of intellectual inertia, now soldiers on without adequate theoretical basis?
    Tucci is one of the few scholars to have appreciated the awkwardness and poignancy of Kamalaśīla position.[48] But, his rather oblique solution to Kamalaśīla's quandry only precipitates a puzzle of his own. Speaking first of Hva-San, Tucci claims the sudden enlightenment doctrine follows from the simultaneous granting of ontological status to both abhūtapariikalpita ("power of subjective representation") and Śūnyatā. By contrast, Kamalaśīla then would be said to hold gradualism because he maintains loyalty to Śūnyavāda monism by refusing to grant ontological status to anything but the Absolute.
    Yet, it seems incoherent of Tucci to say that it is Hva-San's simultaneous admission of ontological status to both these principles which breaks "the monism of Mahāyāna,"[49] causing meditation to recede into the background and dictating a subitist view of the attainment of enlightenment. If anything, the opposite should occur; if one breaks the monism of the Mahāyāna into such a dualism, how then can either of these realities pass away suddenly? If the abhūtapariikalpita is empowered to project the world of representation, how does it also pass away in the face of sūnyatā which Tucci implies is ontologically distinct? It seems that either the "monism of Mahāyāna" is not really broken, in which case Tucci's solution does not even get started, or that it is broken, in which case one is not yet enlightened, because one has not yet penetrated into  sūnyatā. Either way, Tucci does not seem to have succeeded in his aim. Moreover, in the fact of his own supposed monism Kamalaśīla's gradualism becomes all the more mysterious, and not less so.
    I would merely point out that the text of the Bhāvanākrama gives no indication that Hva-San is any kind critic of monism. And, if he were, he would probably prefer gradual enlightenment over the sudden view. Kamalaśīla, on the other hand, does give indications of having watered down the transcendental monism one might expect him to have observed.
    This stems from Kamalaśīla's philosophical indebtedness to Bhāvaviveka and Dharmakīrti through his teacher, Sāntaraksita. Potter[50] and Warder[51] argue independently that Kamalaśīla's  thought represents a partial synthesis of the epistemological traditions of the Pramānavarttikam of Dharmakīrti and the Svatāntrika Mādhayamaka of Bhāvaviveka. Taken together, these influences seem to confirm Kamalaśīla's belief in the worth of logic and analysis,[52] against what Potter believes to have been the Yogācārin attempt to downgrade them.[53] Bhāvaviveka is said to have made this kind of point by advancing the unique view of graded levels of truth within sūnyatā—as well as within the empirical realm.[54] If this be monism, it is certainly highly modified. To admit grades of being is virtually to admit kinds of being, which is really to break the purer forms of the monism of Mahāyāna. For Dharmakīrti, the ontological basis of his positive attitude toward reason seems to be a certain materialist or physicalist—tending convictions. Against the Yogācārins, Dharmakīrti  argued the "relative independent reality of objects,"[55] and that reality has "arthakryātra, the character of doing something ... of making a difference."[56] Empirical perception (pratyakna) is therefore a pramāna (a means of knowledge), and '"effect of reality"' and not an illusion.[57] In this way Dharmakīrti undercuts any attempt to empower thought alone to make real changes in the status of a person seeking enlightenment.
    The views of Bhāvaviveka and Dharmakīrti, then, seem remarkably similar to Kamalaśīla's conviction, throughout the Bhāvanākrama, that the world and ordinary knowledge could not merely be thought away, but had to be undermined by serious meditative and analytic praxis. Dharmakīrti even explicitly holds this view. In Potter's words, "one obtains yogic insight ... by sharpening one's understanding or insight by meditation and dialectic,"[58]
    For both Dharmakīrti and Kamalaśīla this seemed also to mean that testing and a spirit of censoriousness (Gellner) become important. In classic empiricist style, Dharmakīrti believed a theory of knowledge ought to stand the test of experience" and "practice."[59] Quite probably reflecting this influence while quoting the sutras in his Nyāyabindupūrvapaksasanksipiti, Kamalaśīla reports the Buddha saying: " 0 Brethern! . .. never do accept my words from sheer reverential feelings! Let learned scholars test them... ."[60]
In the Bhāvanākrama, Kamalaśīla himself brings meditation into play with experiential testing: " Having thus ascertained reality by means of gnosis consisting in investigation, in order to make this evident, one should have recourse to the gnosis consisting in contemplation. .. ."[61]
    Kamalaśīla even seems to share the view of King about the complementary roles of the jhānas and vipāssana in Theravāda Buddhism. Here speaking of the jhānas in terms of samādhi, Kamalaśīla seems to repeat the division of labor between these two branches of meditation which I also linked with Gellner's claims about the descriptive and prescriptive aspects of empiricism: " ... when his mind has been taken hold of by the hand, as it were, of samādhi, the yogin, by using the sharper weapon of gnosis should root out the seeds of false imagination. .. ."[62]
    In these ways, Kamalaśīla seems to conform to much of the empiricist-cum-materiahst spirit of early Buddhism through the influence of Bhāvaviveka and Dharmakīrti. To the extent that these empiricist and materialist tendencies inform Kamalaśīla's thinking about meditation one would explain Kamalaśīla's teaching of the doctrine of gradual attainment of enlightenment on the same grounds as I have tried to do with the early Buddhists.

VII. CONCLUSION: BELIEF, PRACTICE, AND STRUCTURE

    To gain a unifying structural insight into Kamalaśīla's situation I want to conclude this discussion by pushing beyond the rather straightforward discussion of the history and content of Kamalaśīla's thought. Granted that Kamalaśīla was influenced by both Bhāvaviveka and Dharmakīrti, one might go on to ask what conditions of Kamalaśīla's practical situation reinforced his adherence to an empiricist and materialist-tending tradition? Here, I want to suggest that Kamalaśīla's practical discipline of analysis and compassion may have 'fit' better with the world-view he inherited from Bhāvaviveka and Dharmakīrti, and thus that, in consequence, it was favored. Kamalaśīla could not have been a pure Śūnyavādin Absolutist without suffering substantial disharmonies in his overall approach to the world. Kamalaśīla may have thought and taughtmore like an empiricist-cum-materialist early Buddhist, partly because he also acted like one. In taking the world of thought and being as at least provisionally real in meditation, analysis, and compassionate behavior, Kamalaśīla may very well have come to think about the appropriate means of release as gradual—much as did the early Buddhists.
    I am suggesting that Kamalaśīla's belief in gradual enlightenment may have been connected to his practice in somewhat the same way some beliefs might be said to be induced by certain practices. In the Buddhist tradition one thinks of the belief in the transcendental Buddha as having possibly been induced by the practice of buddhapūjā, which does not in itself require such a transcendental objectof worship. Although buddhapūjā is, strictly speaking, an act of remembrance, such practices tend, quite often, to induce a belief in the existence of their object. Gombrich suggests that in modern Theravādin countries one can observe this movement from mere recollection of the exemplary earthly life of the long-deceased historical Buddha to the belief in the transcendental existence of the Buddha, now thought to be available to human entreaties.[63]
    I do not believe these processes happen mechanically or through causal connections, as typically conceived. Human culture seems too intricate and human beings too subtle for the mechanistic process to be the strongest candidate explanation here. A likelier model might be one which takes its rise from Levi-Strauss[64]: to the degree one finds structural affinities between practices and beliefs, perhaps one should consider such affinities either a working out of certain deep common structures, or perhaps related by a sort of formal or structural causality. Men often do things for structural reasons—whether the structures lie behind the things in question or whether they operate on the same level. In Kamalaśīla's case, he may have advocated the belief in gradual release and the practice of meditation, analysis, and compassion because of some deep common structure, or because either belief or practice were causally prior, and same in form.
    This speculation suggests that a kind of formal causality may be at work in the passage from deeper levels of culture to other more accessible to common sense, or between things on the same level of culture. In the context of Buddhist meditation and theories of release, I offer that one does not practice analytic methods of meditation and painstaking human compassion for lengths of time without having something of those activities 'rub off' on other levels of life—in our case the gradualist theory of the attainment of release. (The same also goes for the effect of beliefs on practices.) Among the things which 'rub off', I want to identify the notion of form or structure. Critical analytic meditational methods and serious concern for ordinary human well-being, conform to the gradual kind of enlightenment, at once described as a graded route and prescribed as a critical, analytic censoriousness about claims to, knowledge.

NOTES
[1]. D. Goleman, "Perspectives on Psychology, Reality, and the Study of Consciousness." Journal of Transpersonal Psychology 4 (1974): 4.  
[2]. W. King, "A Comparison of Theravāda and Zen Meditation." History of Religions, (1969); 310.
[3]. Ibid., p. 311.  
[4]. K. N. Jayatilleke, Early Buddhist Theory of Knowledge (London: George Alien and Unwin, 1963). See also D. Kalupahana, "A Buddhist Tract on Empiric ism," Philosophy East and West 19, N6. 1 (1969). and Causality: The Central Philosophy of Buddhism (Honolulu: University of Hawaii, 1975) and Buddhist Philosophy (Honolulu: University of Hawaii, 1976).  
[5]. V, Trenckner, A Critical Pali Dictionary, Volume 1 (Copenhagen: Royal Danish Academy of Sciences and Letters. 1924-1948): 201-202. 
[6]. T. Rhys-Davids and W. Stede, eds., Pali Text Society's Pali English Dictionary (London: Luzac, 1966): 39. 
[7]. Ibid., pp. 39,101. Nyanatiloka, Buddhist Dictionary (Colombo, Sri Lanka: Frewin, 1972): 17. 
[8]. Jayatilleke, Early Buddhist Theory of Knowledge, p. 40.   
[9]. Ibid., chap. 8.   
[10]. Ibid.,p- 277.   
[11]. Ibid., p. 466.      
[12]. W, King, "The Structure and Function of the Trance of Cessation in Theravada Meditation," manuscript read at Annual Meeting of the American Academy of Religion, Chicago, November, 1975. 
[13]. Ibid., p. 4.      
[14]. Ibid., p. 11.   
[15]. Ibid., p. 7.     
[16]. Ibid., p. 8.     
[17]. Ibid,, pp. 14f.  
[18]. E. Gellner, Legitimation of Belief (Cambridge: Cambridge University Press, 1974). 
[19]. Ibid., p. 36.          
[20]. Ibid.                       
[21]. Ibid., pp. 32f.        
[22]. Ibid., p. 38.           
[23]. Ibid., p. 115.        
[24]. Ibid., chaps. 5, 6.       
[25]. Ibid., p. 124.                   
[26]. A. K.. Warder, "Early Buddhism and Other Contemporary Systems," Bulletin of the School of Oriental and African Studies, 17 (1956); 43-63.                   
[27]. Gellner, Legitimation, p. 120.       
[28]. P. De Silva, Buddhist and Freudian Psychology (Colombo: Lake House, 1974). 
[29]. R. Johanssen, The Psychology of Nirvana. (London: Alien and Unwin, 1969).   
[30]. P. Demieville, Le Concile du Lhasa (Paris: Imprimerie Nationale, 1952). E. Obermiller, "A Sanskrit Ms. from Tibet—Kamalasila's Bhavana-krama," Journal of the Greater India Society 2(1935): 1-11. G. Tucci, trans., Minor Buddhist Texts, Part 11. The First Bhavana-Krama of Kamalasila, Serie Orientale 9 (2) (Rome: Institute Italiano per il Medio ed Estremo Oriente, 1958).
[31]. P. Yampolsky, trans.. The Platform Sūtra of the Sixth Patriarch (New York: Columbia University Press, 1967).     
[32]. For "Hui-Neng" one may therefore read "Shen-hui," the historical proponent and/or source of the teaching attributed to Hui-Neng.       
[33]. H. Dumoulin, History of Zen Buddhism, P. Peachev, trans., (New York: Pantheon, 1963), 87.  
[34]. Yampolsky. Platform Sūtra. p. 116.   
[35]. Dumoulin, History of Zen, p. 95.  
[36]. Yampolsky. Platform Sutra, p. 116.  
[37]. Ibid..p. 115.
[38]. G. Dharmasiri, A Buddhist Critique of the Christian Concept of God (Colombo, Sri Lanka: Lake House. 1975), pp. 199ff.
[39]. Jayatilleke, Early Buddhist Theory of Knowledge, chap. 5.   
[40]. Obermiller, "A Sanskrit Ms.." p. 5.   
[41]. A. K. Warder, Indian Buddhism (Delhi: Motilal Banarsidass, 1970). pp. 467f.   
[42]. Ibid.,pp.477ff.     
[43]. Tucci. Minor Buddhist Texts, pp. 173-175.   
[44]. Ibid., pp. 175f.
[45]. Ibid- p. 60.
[46]. Ibid.. p, 105.      
[47]. Ibid., pp. 64. 103', 104-111.  
[48]. Ibid., pp. 104-111.   
[49]. Ibid.. p. 105. 
[50]. K, Potter, Presuppositions of India's Philosophies (Englewood Cliffs, New Jersey: Prentice-Hall. 1963),pp.239f.  
[51]. Warder, Indian Buddhism, pp. 476f.   
[52]. Tucci, Minor Buddhist Texts, p. 160.   
[53]. Potter, Presuppositions, p. 233.           
[54]. Ibid., p. 240.   
[55]. Ibid., p. 233.   
[56. Ibid., p. 141.    
[57]. Warder, Indian Buddhism, p. 468.   
[58]. Potter, Presuppositions, p. 194. Also Kalupahana, Buddhist Philosophy, pp. 138f., that Kamalaśīla held the principle of causality to be central to Buddhist conceptions of reality.   
[59]. Warder, Indian Buddhism, p. 468.   
[60]. T. Stcherbatsky, Buddhist Logic, 2 volumes (New York: Dover, 1962 of original 1930), 1: 76f.  
[61]. Tucci, Minor Buddhist Texts, p. 477.   
[62]. Ibid., p. 170.   
[63]. R. Gombrich. Precept and Practice. (Oxford: Oxford University Press, 1971), pp. 121f.   
[64]. C. Levi-Strauss, Structural Anthropology (New York: Doubleday, 1967), chaps. 1-4, 11,15, 16.