One or None? Truth and Self-Transformation for ´Sa ˙nkara and Kamala´s¯ıla

: This article explores how two inﬂuential 8th-century Indian philosophers, ´Sa˙nkara and Kamala´s¯ıla, treat the threefold scheme of learning, reasoning, and meditation in their spiritual path philosophies. They have differing institutional and ontological commitments: the former, who helped establish Advaita Ved¯anta as the religious philosophy of an elite Hindu monastic tradition, afﬁrms an unchanging “self” (¯atman) identical to the “world-essence” (brahman); the latter, who played a signiﬁcant role in the development of Buddhist monasticism in Tibet, denies both self and essence. Yet, they share a concern with questions of truth and the means by which someone could gain access to it, such as what, if anything, meditation contributes to knowledge and its acquisition. By exploring their answers to this and related questions, including how discursive and conceptual practices like learning, reasoning, and meditation could generate nonconceptual knowledge or knowledge of the nonconceptual, this essay shows the difﬁculty of separating “philosophical” problems of truth from those related to self-transformation or “spirituality,” as Michel Foucault deﬁnes the terms. It also reassesses, as a framework for comparison, the well-known contrast between “gradual” and “sudden” approaches to the achievement of liberating knowledge and highlights them as tensions we still struggle to resolve today.


Introduction: Truth and Self-Transformation in Comparative Perspective
For much of the intellectual history of "the West," epistemological or "philosophical" matters, as Michel Foucault describes them-questions about "the conditions and limits of the subject's access to the truth"-were bound together with what Foucault calls "spirituality" (spiritualité): "the search, the practice, the experience by which the subject operates the transformations upon himself necessary to gain access to the truth." 1 In the opening of his 1981-1982 lectures, Hermeneutics of the Subject, Foucault distinguishes between "philosophy" and "spirituality" in order to decenter modern philosophy's central concern with knowledge (or self-knowledge, knowing oneself), that is, with epistemology. He argues instead for the centrality of "self-care" (souci de soi) or technologies of self-transformation for understanding the relationship between subjectivity and truth. Having made the distinction, however, Foucault immediately grants that "philosophical" and "spiritual" matters were rarely, if ever, separated in Greek and Roman antiquity. For Foucault, the turning point was what he calls "the Cartesian moment," but before then "philosophical thinking," as defined above, was subsumed within "spirituality" as one of its modes of practice. Foucault's primary example is Gnosticism, though he says the same is true of the Stoics, the Epicureans, the Platonists, the Neoplatonists, and in antiquity in general; his sole exception is Aristotle. 2 Scholarship has now begun to show even more that the situation in classical India is similar. One may find exceptions-this is still a topic of debate-but as a question of historical record there appear to be many more examples in classical India than exceptions, just as there are in Western antiquity. that the subitist's nondualist ontology and rejection of the path "only makes sense in the context of a community already committed" to the goal and the path. 7 This is a point we must continually bear in mind: Context matters, and the institutional context here for an ostensibly epistemological debate about the implications of nondualism and the means of knowing it-continuity or independence-gives voice to a central concern about spiritual path theory. Now, it may or may not be historically accurate to say that Kamalaśīla wrote three treatises, each called The Stages of Cultivation (Bhāvanākrama), while living in Tibet and engaged in an officially sanctioned debate or a series of written correspondences with representatives of a nondualist Chan Buddhist tradition from China. 8 It is certainly the case that Kamalaśīla was the product of a long tradition in India where related questions of epistemology, ontology, and spiritual path theory were still being debated. This becomes obvious when we trace the broader history and context for some of the concepts and arguments on which Kamalaśīla relies, such as the threefold scheme and nonconceptual meditation. 9 Adam (2006, pp. 82-83) points out that the threefold scheme used by Kamalaśīla throughout these treatises is not uniquely Buddhist; in fact, it is a broader Indian framework for conceptualizing practices upon a path of self-cultivation. 10 Despite its ubiquity in Indian and Tibetan Buddhism, it is perhaps better known throughout India from its presence in the mainstream Indian tradition of Vedānta (or Uttara-mīmām . sā), its scriptural basis therein being two nearly identical passages from the Great Forest Upanis . ad (Br . hadāran . yakaupanis . ad). 11 One wonders if the threefold scheme were not imported into the Buddhist tradition by scholar-monks (Asaṅga and Vasubandhu, circa 3rd to 4th century?) already familiar with the similar framework in the Veda. On the other hand, we will see below that making the link between the passages in the Br . hadāran . yaka-upanis . ad and the standard threefold scheme as it appears in later Buddhist and Vedānta systematic thought requires some interpretative work. Given the purported impact of Buddhist thought upon the early Vedānta interpretive systems (Gaud . apāda andŚaṅkara, if not Bhartr . prapañca and Man . d . anamiśra, approximately 6th to 8th century?), the direction of influence may well have gone both ways. 12 Setting aside these historical speculations, we have a situation in which comparable Indian philosophers discuss comparable techniques as comparable solutions to comparable problems.
It is also worth noting at the outset that comparable frameworks for conceptualizing a series of interlinking practices can be found well beyond India. In the opening of his 1980lectures, Subjectivity and Truth, Foucault (2017 identifies the typology of mathēsis (learning, teaching), meletē (reflection, meditation, thinking), and askēsis (exercises, ascetic practices), a framework he claims to be that of Epictetus but theorized more generally by the Stoics and implicit "in all the great arts of living" of the Hellenistic world. 13 Foucault's discussion of this framework provides the context for an introduction in those lectures to the concept of "technologies of self" (techniques de soi): "thought out, elaborated, systematized procedures taught to individuals in such a way that, through the management of their own life, through the control and transformation of self by self, they can attain a certain mode of being" (p. 35). His description of such practices fits well with the path philosophies of classical India that concern us here. The historical proximity betweeń Saṅkara (the most famous proponent of early Vedānta) and Kamalaśīla-the fact that both know and discuss a similar typology of practice-lends further credibility to the project of comparing them first as a basis for broader comparison and analysis. Since Kamalaśīla's use of the threefold scheme has been the subject of recent scholarly analysis, I will discuss him more briefly and mainly as a foil to highlight certain resonances and tensions with Saṅkara and the early Vedānta tradition, which will receive the more detailed treatment here. 14 As we will see, the comparison raises difficult questions that still resonate today.

Comparing Kamalaśīla withŚaṅkara and Early Vedānta
Typically, we rely on dualities for comparison, and the "sudden/gradual" dichotomy has been proposed as a comparative framework in this case, too. In his groundbreaking essay on the metaphor, THE MIND IS A MIRROR, Paul Demiéville describesŚaṅkara as someone who "leaned toward 'subitism'." Demiéville asserts that forŚaṅkara: [L]iberation . . . does not admit of any active effort on our part. It is a matter of knowledge, not of works . . . Liberation is inherent in ourātman . . . [I]t would be wrong to think that it could be revealed through purifying theātman by our own effort . . . In fact theātman can never be the object of any activity, for all activity directed toward an object implies a modification of that object-but theātman is eternally unalterable. 15 Demiéville depends on Deussen (1912), and therefore gives us a fairly reliable description of the main thrust ofŚaṅkara's nondualism and how it informs his passivism, 16 but Gómez (1987, p. 128) cautiously reminds us that "Śaṅkara . . . spawned three lines of disciples-only one of which took a 'leap-philosophy' position (to borrow K. H. Potter's terminology)." 17 Modern scholarship has shown that the sudden/gradual dichotomy is more "idealtypical" than actual. 18 It does not correspond in reality to a single duality, but gestures toward a number of different but related dichotomies that may concatenate in complex ways in particular cases. These dichotomies include the active and passive, the innate and acquired, the expressible and inexpressible, the conceptual (savikalpa) and nonconceptual (nirvikalpa), the dual and nondual. Gómez (1987, p. 132) concludes, "One would like to suggest that the various polarities associated with the sudden-gradual controversies respond somehow to one single duality, but that is not the case. Clusters of dualities, however, do overlap." "Sudden" and "gradual" cover a variety of tensions, which coexist or combine in complex ways, but often are reduced and then exploited for the purposes of polemical argumentation and identity formation. This recognition reduces or at least problematizes the usefulness of the dichotomy as a neutral typology for modern doxographical description, but it can still be helpful as a framing hypothesis because of the productive questions it raises for us.
Demiéville also describesŚaṅkara as a "systematizer," and Gómez has this to say about systems and their relation to the various dualities that comprise the sudden/gradual complex: [R]eligious systems often are dynamic attempts to solve all of these admittedly real and universal oppositions. The dichotomies are inherent to human thought, not constructed by philosophers or mystics, although religious effort and ideology often can be described as a resolution, or rather a balancing, of the tension between the two poles. (Gómez 1987, p. 132) The point here is that we misunderstand the nature of systematic thought and practice if we do not also recognize and respect the intractability of the tensions that such systems try to resolve. For instance, whether we label them subitist or gradualist, we can find a natural tension within spiritual path philosophies that attempt somehow to account for and describe a process of change, progression, or improvement that can seem fast at times and slow or stagnant at others. We do well to read systematic thought and polemical debates with such tensions in mind.
In order to put Kamalaśīla into conversation withŚaṅkara and early Vedānta, we can also begin by identifying some basic assumptions shared by them. Both address the same basic problem of human dissatisfaction, and both are concerned with the question of how one can achieve liberation from it. That may be obvious to anyone who knows anything about them, but the point underscores that both are working within the framework of spiritual path theory. Among the three characteristics of "spirituality" identified by Foucault are, first, that the truth is not simply given to the subject as such, and second, that accessing the truth requires a certain modification of the subject, a type of conversion, which can take different forms, but means that the subject must always be "torn away" from his or her present condition. It involves a transformation of the subject. Foucault calls it "a job" (un travail) and a "progressive transformation," a movement by which "the truth comes to him and illuminates him" (la vérité vient à lui et l'illumine). 19 Both Kamalaśīla andŚaṅkara begin from the assumption that there is something unstable or unfulfilling about our everyday lives-we find ourselves caught in patterns of behavior that we believe will bring happiness, but ultimately leave us dissatisfied-but it seems that each of us must begin the path by deciding for ourselves whether this is a fair description of our lives. For those unconvinced by such a characterization,Śaṅkara says, even the scriptures will not sway them. 20 For those who are ready to hear it, however, Saṅkara and Kamalaśīla will identify the same basic cause of the problem and the same basic solution. Ignorance causes our dissatisfaction, ignorance about the nature of the world and ourselves in relation to it, and thus liberation requires replacing ignorance with knowledge.
However, the two philosophers differ on the content of this liberating knowledge and the means of achieving it.Śaṅkara is committed to the claim that theātman, the self, is identical to brahman, the essence of reality, and that the diverse world of our everyday experience is therefore illusory; Kamalaśīla agrees that our everyday experience misleads us, but he holds the view that neitherātman nor brahman ultimately exists; everything is empty of essential reality, including human beings; we, like everything else, arise in dependence on other things, which for him undermines the apparent solidity or reality of our everyday world. Thus, the two philosophers agree on the illusory nature of everyday experience, but disagree on the ontological characterization of ultimate truth. We will see how they differ on the means of achieving such liberating knowledge when we look at how they treat the threefold scheme.
Saṅkara's and Kamalaśīla's basic difference regarding the content and means of liberating knowledge, however, belies another important similarity between them. In Vedānta and Buddhist thought, the goal may be called "liberation" (mukti, moks . a) or "cessation" (nirvān . a); in either case, however, this goal is comprised largely by knowledge of or insight into the true nature of reality. Common terms are yathābhūtajñāna ("knowledge of reality as it is"), samyagjñāna ("perfect knowledge"), or samyagdarśana ("perfect seeing or insight"); the Buddhist tradition also speaks of "seeing and knowing things as they really are" (Pāli: yathābhūtañānadassana). We see here the application in both systems of the widespread conceptual metaphor, KNOWING IS SEEING. 21 Moreover, such knowledge or insight has power; it is thought to be liberating. Such liberating knowledge is not merely a state of knowing certain information about the world. Again, this attitude toward the truth conforms to what Foucault says about spirituality. Its third characteristic for Foucault is that knowing the truth produces certain consequences for the knowing subject beyond the mere result of the steps taken to access it. He writes, "Truth is what illumines the subject; truth is what gives him 'supreme bliss' (béatitude); truth is what gives him tranquility of the soul." 22 Now, understanding how these Vedānta and Buddhist philosophers differ on the nature and means of liberation requires that we engage several interrelated problems of ontology, epistemology, and path theory. Consider the following questions aboutŚaṅkara in particular: How can he be both subitist and systematizer? How can he advocate for the spontaneity of liberation while at the same time upholding an epistemic continuity thesis with regard to learning, reasoning, and meditation? How can he be a subitist while discounting the importance of meditation, nonconceptual or otherwise? How does he manage the tension between the progressive method implied by the threefold scheme and his subitist tendencies? If he ultimately affirms that reality is inherently nondual, how can the various means of achieving knowledge of reality, such as the threefold scheme of learning and so on, which employ concepts rooted in difference, ever produce such knowledge? With respect to Kamalaśīla, even though he can be described as a gradualist, it is fruitful to consider how he uses the threefold scheme to argue for an epistemic continuity thesis while at the same time incorporating nonconceptual meditation into his path system. Such questions arise from applying the sudden/gradual dichotomy and the distinction between continuity and independence theses as interpretive tools in this comparative exercise.
These are complicated questions, and a more straightforward, interpretative one could help orient us at this point: how doŚaṅkara and Kamalaśīla treat the threefold scheme in their respective works? Here is a preliminary hypothesis: the threefold scheme helps Kamalaśīla articulate his gradualist system, whileŚaṅkara treats it with some ambivalence. Kamalaśīla uses it to explain key junctures in his system and as a descriptive framework for the gradual path, whereasŚaṅkara accepts the authenticity of the scheme, given its basis in his scriptural corpus, but deemphasizes its applicability or reinterprets its meaning in some contexts.
Why this difference between them? Here again, the answer becomes more complex, but one reason may be the differing importance each philosopher gives to meditative practice on the spiritual path and more specifically to "nonconceptual" meditation or knowledge (nirvikalpa-samādhi or jñāna) as a step on this path. Nonconceptual meditation plays a crucial role in Kamalaśīla's path philosophy, whereas it does not seem to be given much, if any, importance inŚaṅkara's writings, at least not as a stage on the path. Arguably, forŚaṅkara, the knowledge of the nonconceptual, that is, the highest, nondual brahman-which knowledge may itself be described as nonconceptual-may be defined as the goal of the path, but meditation is not the primary means to such knowledge. Saṅkara repeatedly emphasizes that scripture is the sole means of knowing the true nature of reality. This deemphasis on meditation and concurrent emphasis on scripture and hermeneutics may well demonstrateŚaṅkara's indebtedness to the orthodox Mīmām . sā tradition and his special contribution to the Vedānta tradition. 23 Both these reasons foŕ Saṅkara's ambivalence also concern his reputed subitism.
We could say thatŚaṅkara's subitism makes him distrust gradualism, and while this statement may be vague given Gómez's comments about the value of these terms as descriptions of actual philosophers' positions, it is still meaningful and suggestive of more specific claims, such as the following: His nondualist ontology, one in which the liberated state is innate within us, shapes his epistemological position, and his epistemological position, which emphasizes the revelatory nature of scripture, makes him distrust gradualism. His nondualist ontology also shapes his path theory, which treats knowledge as a mental state of passive receptivity, and his emphasis on passivity and nondualism makes him distrust causal schemes in general. At the same time, however, systematicity breeds gradualism and a teaching system requires it. More specifically, a reliable teaching strategy requires that there is regularity of causation, or at least a possible relation of cause and effect between actions and consequences. Otherwise, teaching and learning become theoretically impossible. So,Śaṅkara inherits the threefold scheme and cites it affirmatively in certain contexts, but he also argues against its use in other contexts. At times he defends the idea of immediate liberation, even though he also makes concessions to gradualism-just as Gómez tells us all subitist philosophers must do at one point or another because of the basic theoretical problems inherent in subitism.

The Threefold Scheme and the Epistemic Continuity Thesis
What, if anything, does meditation contribute to knowledge and its acquisition? What is the epistemic value of meditation practice? These questions can form the basis for comparing Kamalaśīla andŚaṅkara's use of the threefold scheme, while helping to confirm that both of them uphold an epistemic continuity thesis with regard to the relationship among learning, reasoning, and meditation. These questions raise others, such as what one means by knowledge, not to mention what is meant by meditation. Given that there are numerous types of meditation, from visual to verbal, kinetic to conceptual and possibly even nonconceptual meditation, is knowledge any easier to define? Can we say, for instance, that knowledge is "justified, true belief," as many 20th-century analytic philosophers have defined it, or is it something more or even less? 24 Perhaps knowledge requires no justification, no special cognitive process, but assuming we wish to retain the notion that knowledge requires some kind of "safety condition," here is another way of asking the question: Does meditation (of whatever kind) contribute anything uniquely necessary to justifying a belief as true? Independence theorists could say "yes," assuming they are willing to grant that justification is possible at all; continuity theorists might say "yes" or "no," depending on what they mean by justification.
Kamalaśīla's answer to the above questions has been a topic of recent scholarly discussion, and indeed the distinction between continuity and independence theses has arisen from these discussions. Thus, it makes sense to begin with Kamalaśīla and quickly rehearse the nature of the debate. Tillemans (2013Tillemans ( , 2016 makes the distinction while arguing that for Kamalaśīla, "yogic perception would not provide any new information from what had been given by philosophy. Kamalaśīla's yogic perception, in effect, appears to be . . . a type of amplification or integration of the contents of philosophical thought. Philosophy would be doing the significant epistemic work of discovering truths." 25 Here, we can understand yogic perception as another way of referring to the kind of understanding gained in meditation practice, which is said to align with the practice of reason or what Tillemans calls "philosophy." Adam (2016, p. 370) takes issue with Tillemans' characterization of Kamalaśīla's position and offers a different point of view: While Tillemans is certainly correct in stating that meditative understanding conforms to the conclusions reached through philosophy according to Kamalaśīla, this position is not inconsistent with one that holds meditative understanding to provide a kind of experiential verification of those conclusions.
Adam also suggests that for Kamalaśīla, "meditation plays an indispensable role in the quest for liberating knowledge, contributing insights that are unattainable by studying and thinking alone" (p. 354). Adam may well be right, but we can see multiple issues at play here. One is the transformative efficacy of meditation, its necessity for achieving liberating knowledge. Another is its epistemic value.
In responding to Adam, Kellner (2020, pp. 69-70) makes the distinction between epistemic value and transformative efficacy while arguing that Adam's claim that meditative cultivation provides "experiential corroboration" for conclusions reached through conceptual reasoning misrepresents Kamalaśīla's epistemology: . . . there is no indication that Kamalaśīla would consider the certainty obtained by reasoning to be insufficient evidence when compared to evidence provided by perceptual awareness. Conceptual certainty is not regarded as sufficiently strong to remove deep-seated misconceptions in the mind, but this is more like a psychological or phenomenological insufficiency, not one of strength of evidence.
Kellner thus makes a distinction between epistemic strength and transformative efficacy, a distinction that is perhaps less clearly made by Adam in his objections to Tillemans' interpretation of Kamalaśīla's position, but one wonders whether Kamalaśīla himself draws the distinction quite so clearly. The debate raises questions about the nature of truth and its relationship to the knowing subject in the context of spiritual path philosophy. Can conceptual knowledge provide sufficient evidence for certainty about the world, or does knowledge require a more intimate relationship between the knowing subject and the object known? What justification, if any, does knowledge require? And what should we make of the concept of a powerful, liberating knowledge? Is it the same knowledge we gain from ordinary perception or inference? One thing is clear: when we speak about liberating knowledge in the context of Buddhism, or Vedānta for that matter, we are not simply referring to information about the nature of the world but to a state of knowledge that results from and is a significant transformation of the way the knowing subject sees, thinks about, and acts in the world.
A key passage from the first Bhāvanākrama will suffice to demonstrate Kamalaśīla's use of the threefold scheme and his adherence to the continuity thesis. It also suggests how he seeks to integrate nonconceptual meditation into his path system: Regarding the three types of wisdom, that is, the wisdom arising from learning, reasoning, and cultivation, one should first of all generate the wisdom arising from learning, because with it one first determines the meaning of the scriptures. After that, by means of the wisdom arising from reasoning, one understands the meaning of the scriptures in terms of whether it is definitive or provisional. After that, once one has thereby generated certitude, one should cultivate the real object {true meaning} (bhūtam artham . ) and not a false {unreal} one (abhūtam). For otherwise perfect knowledge (samyagjñāna) would not arise, because one might cultivate false objects, too (viparītasyāpi bhāvanād), and doubt would not disappear. And then cultivation would be entirely meaningless {useless} (vyarthaiva), like that of the non-Buddhists. As the Blessed One There is no other cause that would bring peace. Therefore, after one has analyzed with the wisdom arising from reasoning, that is, though both logic and scripture, one should cultivate only the truly real nature of things (bhūtam eva vastusvarūpam). And the real nature of things has been certified by scripture and reasoning to be their absolute non-origination from the standpoint of ultimate reality {truth} (paramārthatah . ). 26 In this passage, Kamalaśīla makes a number of points that are relevant to the current discussion. For one thing, he outlines the threefold scheme as a logical or temporal sequence of practices. One begins by learning the scriptures, and then proceeds by reasoning to determine their correct meaning, before finally engaging in the practice of meditative cultivation on the true nature of reality, which for Kamalaśīla is the fact that all things are selfless: they arise in dependence on other things, and therefore from an ultimate standpoint do not arise at all. Such a process will give rise to "perfect knowledge" (samyagjñāna), the goal of the path. So, we have here something like an analysis of knowledge and its acquisition, but Kamalaśīla gives voice to the concern that one's meditation practice must at some point take an "unreal" object, that is, an imaginary or conceptual object, as its focus. For, as Kellner (2020, p. 53) says, "the yogi has to engage in reificatory practices in order to remove reification." This process will also involve the cultivation of nonconceptual knowledge (nirvikalpajñāna), which Kamalaśīla describes as "seeing the highest reality" (paramatattvadarśana) (Kellner 2020, p. 63).
Kamalaśīla shows an awareness of the epistemological problem that rival, non-Buddhist philosophers disagree on the nature of reality even while they may also advocate for meditation practice as a means of attaining knowledge of it. Given the possibility, the necessity even, of reifying ultimately "false" objects, meditation can lead one astray. Kamalaśīla tries to allay concern by appealing to the epistemic authority of learning and reasoning, and here he expresses the continuity thesis most clearly. He says that learning and reasoning provide the certitude with which one can engage effectively in meditation practice. However, he seems to equivocate insofar as he also grants that meditation practice upon the "wrong" object will not destroy doubt, which seems to assume that meditation practice on the "right" object does play a role in dispelling it. It is as if the certitude provided by learning and reasoning still contains an element of uncertainty, but Kamalaśīla does not say this outright. Instead, he proceeds to emphasize the epistemic authority of scripture. He does so, one should note, by explicitly quoting it, but we should also note that the scriptural passage itself emphasizes the transformative efficacy of meditative cultivation on the true nature of reality. Kamalaśīla then restates his position that scripture and reasoning certify the true nature of reality, and that the true nature of reality is emptiness, the fact that things do not ultimately arise at all. Upon such a truth one should meditate, he says, and thereby begins to articulate an answer to the difficult question of how one can use concepts to eliminate conceptuality. 27 Saṅkara's answer to the question of meditation's contribution to knowledge and its acquisition has also been a subject of scholarly debate. Many modern interpreters, most seeming to follow later or competing Vedānta exegetical traditions, have argued that Saṅkara holds that meditative experience possesses special epistemic value over and above learning the scriptures and reasoning about them; meditative experience somehow confirms the truth of the scriptures and is their ultimate source. According to Rambachan (1991), however, these modern interpreters have mostly misunderstoodŚaṅkara's main epistemological position that "scripture" (śruti)-the Upanis . ads or Vedānta-vākyas, "the statements of Vedānta," asŚaṅkara calls them-is the sole valid means of knowing brahman, and that scripture is intrinsically valid (svatah . prāmān . ya) as a means of such knowledge. Rambachan would argue thatŚaṅkara's answer is therefore negative: meditation has no independent epistemic value apart from learning and reasoning. WithŚaṅkara's emphasis on scripture as the only valid means of knowing brahman, even the "spiritual" value of meditation, that is, its value as a method of bringing about liberation, is called into question.
We will explore these issues in more detail below, but here we will look at just a single passage fromŚaṅkara's commentary on the Br . hadāran . yaka-upanis . ad, the scriptural source for the threefold scheme in Vedānta. This comment occurs in a brief transitional passage that concludes the section in which the threefold scheme is first mentioned and introduces the next section. It shows thatŚaṅkara also asserts a "continuity" thesis with respect to the relationship among learning, reasoning, and meditation. Here is the comment: In all cases, however, since consideration through reasoning should be done only in accordance with what is ascertained by scripture, and since meditation should be done only in accordance with what is considered through reasoning, that is, through what has been certified by scripture and reasoning, a specific prescription for meditation is without purpose. 28 While the remark comes in response to a hermeneutical question, the point could not be clearer: scripture provides the epistemological basis for proper reasoning about the nature of reality, and scripture supported by reasoning provides the epistemological basis for meditation. Meditation by itself, and even reasoning if detached from scripture, serves no independent purpose and therefore requires no separate prescription. Meditation accords with scripture and reasoning. When compared to Kamalaśīla,Śaṅkara makes an even stronger claim because, while Kamalaśīla still argues for the soteriological (or transformative) necessity of meditation practice, hereŚaṅkara even seems ambivalent about its soteriological value. The question arises whether or how muchŚaṅkara even advocates for the threefold scheme as an accurate description of the path. Now, we will turn toŚaṅkara and the place of the threefold scheme in early Vedānta.

4.Śaṅkara and the Threefold Scheme in the Br . hadāran . yaka-Upanis . ad
In the preface to his translation ofŚaṅkara's Upadeśasāhasrī, Mayeda (1979, p. xiii) states, "In the Vedānta school there are three stages in the attainment of final release: (a) hearing (śravan . a), (b) thinking (manana), and (c) meditation (nididhyāsana)." This neutral statement masks a broader tension within Vedānta concerning the nature and purpose of meditation, including "nonconceptual meditation" (nirvikalpasamādhi), and the "spiritual" efficacy of the scriptures (śruti, "what is heard") and their ability to liberate the person directly. The tension can also be characterized as one between "gradual" and "immediate liberation" (krama-and akrama-mukti). Finally, the importance of the threefold scheme in Vedānta can be connected to the question of the epistemological status of scripture apart from its spiritual efficacy: is scripture inherently valid as a source of knowledge of reality, or must it be validated or confirmed by another means of knowledge, such as meditative experience or reasoning? Vedānta philosophers,Śaṅkara among them, debated these issues among themselves and with philosophers from rival Indian traditions.
According to Rambachan (1991, p. 10), "[Modern] writers who argue for anubhava [experience] as the true pramān . a [valid means of knowledge] of brahmajñāna [knowledge of brahman] generally treat the process of knowledge inŚaṅkara as progressing through three different phases." The first two phases are "viewed as preliminary and intellectual. It is only the experience which contemplation (nididhyāsana) affords that conclusively informs us about brahman." (Rambachan 1991, p. 14) Moreover, as Rambachan explains, such writers often equate experience with nonconceptual meditation. Rambachan himself argues that such a position fails to representŚaṅkara's own views, as expressed in his commentaries on the principal Upanis . ads, the Bhagavadgītā, and the Brahmasūtra. He instead argues that scripture is, forŚaṅkara, "the singular and exclusive means of knowledge about brahman," and concludes that "it is not possible to reconcileŚaṅkara's views with this seemingly well-ordered system," that is, with the threefold scheme. (Rambachan 1991, p. 97) For his part, Mayeda (1979, p. xv), following Hacker (1949, pp. 8-9), notes that the three stages, of hearing, thinking, and meditation, "appear to correspond to the first, second, and third chapters, respectively, of the Prose Part" of the Upadeśasāhasrī, which scholars generally maintain to be a legitimate work byŚaṅkara and perhaps his only non-commentarial work. 29 The correspondence is indeed striking, but there do not appear to be any direct references to the threefold list in the Upadeśasāhasrī itself. Furthermore, as  argues, chapter eighteen of the verse portion of the work can be read both as a defense of the claim that hearing the scriptures alone can liberate a person immediately, and an argument against the necessity of so-called "repetitive contemplation" (prasaṅkhyāna) as a means of securing and stabilizing liberating knowledge. Advocacy for prasaṅkhyāna meditation is traditionally associated with the Vedānta philosopher, Man . d . anamiśra, author of the Brahmasiddhi, who may have been a contemporary ofŚaṅkara and whose position on the threefold scheme will be assessed later in the essay. 30 Mayeda (1979, p. 254, note 1) thus raises the question: ifŚaṅkara rejects the necessity of prasaṅkhyāna meditation, what should we make of the fact that the third chapter of the prose portion of the Upadeśasāhasrī appears to prescribe an ostensibly similar style of meditation called "meditative rehearsal" (parisam . khyāna)? Clearly the place of meditation and of the threefold scheme in early Vedānta and inŚaṅkara's thinking in particular is a matter of some debate.
Most modern scholars who mention the threefold scheme in Vedānta point to one (or both) of two nearly identical passages from the Br . hadāran . yaka-upanis . ad, 2.4.5 and 4.5.6, which appear to provide a scriptural basis for the scheme. 31 To get a sense for howŚaṅkara treats the threefold scheme, we can begin with his comments on these passages. The context for both is a dialogue between the famous sage, Yājñavalkya, and one of his two wives, Maitreyī. Yājñavalkya is preparing to leave home and become an ascetic, and he wishes to settle his affairs with his wives. The Upanis . ad tells us that, while his other wife, Kātyayanī, concerned herself only with "women's wisdom" (strīprajñā), Maitreyī liked to discuss "religious topics" (brahmavādinī). She questions Yājñavalkya about the benefits of possessing "the whole world and its wealth" when it will not bring one immortality. She asks him to tell her what he knows. 32 Yājñavalkya agrees but instructs her to "concentrate" as he is speaking-the word he uses is nididhyāsasva, the imperative form of the same term typically translated as "meditation" in the threefold scheme. He then explains that "the self" (ātman) is the ultimate reason why people value things in the world, such as one's husband, wife, children, wealth, power, and so forth. The self is the ultimate ground of all things, he tells her. In the key passage that concerns us, Yājñavalkya states: You see, Maitreyī-it is the self (ātmā) that one should see (draśtavyah . ); it is the self that one should hear (śrotravyah . ); it is the self that one should consider (mantavyah . ); it is the self on which one should concentrate (nididhyāsitavyah . ). For by seeing the self, by hearing it, by considering it, and by knowing it (vijñānena), all this is known (idam . sarvam . viditam).
It is noteworthy that Yājñavalkya actually speaks of four actions, not three, which prompts an interpretive question about the notion of "seeing" (darśana) implied here, as well as how sight relates to the other three actions mentioned in the passage. If the passage speaks of four actions, why do commentators speak of a threefold process? A look atŚaṅkara's commentary can help us understand how, retrospectively, commentators saw three in four. Saṅkara begins his commentary on this passage as follows: Therefore [Yājñavalkya says], "you see, it is the self that should be seen;" [that is,] it is worthy of seeing, it should be made into an object of seeing. "It should be heard," [that is,] previously (pūrvam) from the spiritual teacher or scripture. Next (paścāt) "it should be considered," [that is,] with reasoning (tarka, logic). After that (tatah . ), "it should be concentrated on," [that is,] it should be meditated on (dhyātavyah . ) with conviction (niścayena). For in this way [the self] becomes something seen by completing the efficient causes/practices (sādhanas) of hearing, considering, and concentration. When these [efficient causes/practices] are obtained as a unity, then "perfect seeing" (samyagdarśana), 33 which has unity with brahman as its object, becomes clear (prasīdati), 34 and not otherwise, such as by hearing alone. 35 While the scriptural passage only offers a list and does not make clear that seeing is meant to be the outcome of the other three actions, the commentary interprets it to mean that learning, reasoning, and concentration are stages of a sequential process with "perfect seeing" or knowledge of brahman as its goal. The terms, "previously," "next," and "after that," introduce either temporal or logical stages of a sequence. The goal is apparent from the way the commentary interprets the concept of seeing. However, the commentary also speaks of the three practices being obtained as a unity, suggesting that they could be understood as elements of a singular practice with the same goal. The scriptural passage seems clear that the four actions of seeing and so forth have complete knowledge as their result, but it is worth emphasizing the description of the goal in the commentary as a "complete seeing." This is a key concept inŚaṅkara's path philosophy, equivalent to the liberating knowledge of brahman (brahmajñāna), the goal of the path: KNOWING IS SEEING.
The passage prompts further consideration not only about the relationship between seeing and knowing, but also between "meditation" (or "concentration," nididhyāsana) and knowledge. 36 One may note the subtle shift in terminology in the second sentence of the scriptural passage whereby "knowledge" (vijñāna) replaces "concentration" (nididhyāsana). 37 Again and again in the Brahmasūtrabhās . ya,Śaṅkara distinguishes knowledge from action. He argues further that meditation is a type of action, thus distinguishing it sharply from knowledge-as we will see when we look at some key passages. What should we make of the claim made above that "hearing alone" does not produce "perfect seeing?" The claim would seem to support the gradualist idea that the three practices are all necessary for the attainment of perfect seeing, but this statement seems at variance with Saṅkara's claims in the Brahmasūtrabhās . ya, Upadeśasāhasrī, and elsewhere that scripture is the only valid means of knowing brahman, and that hearing scripture can immediately produce such knowledge. 38 It can do so,Śaṅkara tells us elsewhere, because knowledge of brahman, being our own self-awareness, is already directly available to us. Scripture tells us something we already know, but we know it imperfectly or incorrectly. Therefore, scripture removes our mistaken understanding, our ignorance, which allows the truth to shine forth, like the sun on a cloudless day.Śaṅkara says in the Upadeśasāhasrī: "As the sun has light as its nature, it has neither day nor night. In like manner I have neither knowledge nor ignorance since I have Pure Consciousness as my nature, without distinctions." 39 While metaphors like this one express the kind of nondualist sentiment that scholars associate with "subitism," when we try to understand such statements within the context inŚaṅkara's whole path philosophy, we can see the tension therein between "sudden" and "gradual" approaches to liberation. Saṅkara concludes his comments on Br . hadāran . yaka-upanis . ad 2.4.5 with a remark on the final sentence of the scriptural passage, which speaks to relationship between the goal of the path and the means of achieving it: When the scripture says, "For when the self is seen, heard, considered, and known, oh, Maitreyī, then all this is known (viditam . )," that is, it is truly known (vijñāta), the intent (artha, object/meaning) is the destruction of that, namely, the mental object of the belief caused by ignorance (avidyāpratyayavis . aya), which consists in the effects of action upon the doer; and is the object of the belief that is a superimposition of ignorance upon the self; and has as its characteristic the castes and estates of life (varn .āś rama) and so forth; and is the cause of karma, namely, of priestly power, royal power, and so forth; and is like the belief that a rope is a snake. 40 Recall that Maitreyī had asked about the benefits of possessing "the whole world and its wealth." This passage asserts that "all this" (idam . sarvam . ), which according toŚaṅkara refers to all dualistic appearances, is in fact only a mental object constructed by ignorance. In fact, "all this" is identical to our nondual essence. One could interpret him to say that the threefold process, with perfect seeing as its result, removes ignorance and enables one to know the self. When one knows the self, one knows everything as it truly is. For there is nothing other than the self.
Saṅkara restates several of the same points in the short commentary on Br . hadāran . yakaupanis . ad 4.5.6, a scriptural passage nearly identical to 2.4.5. Here is his entire comment on 4.5.6: When Yājñavalkya says, "For, oh Maitreyī, when the self is seen . . . ," he says the following in response to the question of how the self is seen: "When it is heard," [that is,] previously (pūrvam . ) from a teacher and scripture; once again (punas) "when it is considered," [that is,] when it is investigated (vicārite) by reasoning (tarka), [that is,] by argumentation (upapatti)-but hearing is by scripture alonewhen it is considered by argumentation, afterwards (paścād) "when it is known" (vijñāte), [that is,] when it is ascertained (nirdhārite) that it is so and not otherwise. What happens (kim . bhavati)? Yājñavalkya says, "This becomes known." "All this" (idam . sarvam) means what is other than self, because nothing exists that is different from the self. 41 Again, the threefold scheme refers to a sequential process with seeing or knowing as its outcome. We are told here that the threefold scheme explains how the self may be seen. The comment, "but hearing is by scripture alone" (śravan . am . tuāgamamātren . a), which seems inserted into the middle of the passage, suggests thatŚaṅkara views the practice of hearing or learning as a specific one distinct from reasoning, an interpretation that seems confirmed by looking at more evidence fromŚaṅkara's body of work. The commentary also highlights what happens when the self is known, and repeats the idea that knowing the self brings knowledge of "all this," that is, knowledge of the nondual nature of reality.
Saṅkara's commentary on the scriptural passages from the Br . hadāran . yaka-upanis . ad suggests that he accepts a gradualist path philosophy, and that the threefold scheme provides one way of conceptualizing its structure, but it remains to be seen whether this hypothesis is consistent with a broader survey ofŚaṅkara's works. The evidence from the Upadeśasāhasrī seems inconclusive given that the text does not explicitly refer to the threefold scheme and includes both "gradualist" and "subitist" elements. What further textual evidence can shed light on the place of the threefold scheme inŚaṅkara's path philosophy? The Brahmasūtrabhās . ya holds a preeminent place amongŚaṅkara's works. Mayeda (1979, p. 6) calls it "the yardstick against which to measure the authenticity of other works ascribed to him." When we look at the Brahmasūtrabhās . ya, however,Śaṅkara's attitude toward gradualism and the threefold scheme appears rather more ambivalent. 42

The Threefold Scheme and Knowledge as Experience in the Brahmasūtrabhās . yá
Saṅkara's commentary on the Brahmasūtra contains approximately ten passages that cite one of the two key passages from the Br . hadāran . yaka-upanis . ad. These references are given either by the opponent (pūrvapaks . in) or the authoritative voice (siddhāntin). When Br . hadāran . yaka-upanis . ad 2.4.5 is cited by the opponent, as it is on four occasions in the Brahmasūtrabhās . ya (1.1.4, 2.1.3, 2.3.39, and 3.4.51), it is given in support of the argument that scripture enjoins or requires certain practices, such as meditation, in order to know brahman, and in support of the general causal relationship between the process and the goal. This is true of the citation in the commentary on 1.1.4. It is also true of the citation in the commentary on 2.1.3, which discusses whether or not the Upanis . ads enjoin the practice of yoga as a means of obtaining "perfect seeing" (samyagdarśana). It is also true of the citation in the commentary on 2.3.39, where it occurs in the midst of an argument that theātman is an agent and the action of meditation can bring realization of the self. The same is true of the passage at 3.4.51, where the opponent makes the axiomatic statement: "Knowledge has at its antecedent cause learning and so on" (śravan .ā dipūrvikā hi vidyā). The opponent draws the parallel between ritual action and its relation to the fruits of ritual and the three practices of learning, reasoning, and meditation as the means of producing knowledge.
In these instances,Śaṅkara challenges the idea that scripture enjoins or requires anything at all. He argues for scripture as the sole means of knowing brahman and against the idea that meditation is necessary for achieving the goal of liberating knowledge. More generally, he questions whether a causal relationship of any kind can be said to pertain between the process and the goal of liberation. Liberating knowledge, he repeatedly says, is not an action. In his response to the opponent's argument at 3.4.51,Śaṅkara's view of the relationship between the means of knowledge and knowledge as its product is that the means of knowledge remove the obstacles to knowledge; they do not produce knowledge. This is consistent with the instances in the commentarial passages on 1.1.4 and 3.2.21 in which the authoritative voice (siddhāntin) offers an alternative interpretation of the intent behind Br . hadāran . yaka-upanis . ad 2.4.5. The basic idea is that the passage is meant to draw the practitioner's attention away from objects of the world and towards the self. It is not meant to require any actions or draw a direct relation of cause and effect between particular actions and liberating knowledge.
When Br . hadāran . yaka-upanis . ad 2.4.5 or 4.5.6 is cited by the authoritative voice, on three occasions it is given in the context of clarifying that the object under discussion is the highest brahman or the highest self. This is true of the commentarial passages on 1.4.19, 2.1.22, and 2.3.6. The other passages in whichŚaṅkara cites Br . hadāran . yaka-upanis . ad 2.4.5 in the authoritative voice are 1.1.2 and 4.1.1-2. The first instance, at 1.1.2, refers to the passage in order to support the idea that reasoning depends on the scriptures. The other citation occurs at 4.1.1-2, at the beginning of a short section on the means of liberating knowledge. This is an important section given whatŚaṅkara repeatedly says about the relationship among the means of knowledge and their relationship to knowledge of brahman, and I will conclude the discussion of the Brahmasūtrabhās . ya by analyzing it, but we will begin with first passage.
The first reference in the Brahmasūtrabhās . ya to the key passage from Br . hadāran . yakaupanis . ad 2.4.5 is found in the commentary on Brahmasūtra 1.1.2. It is cited in support of the complementarity of scripture and reasoning when the latter is based on the former: "Understanding brahman is accomplished by determination, through analysis, of the meaning of the statements in the Vedānta [that is, the Upanis . ads], and not by any other valid means of knowledge, such as inference and so on." Then, the point is made: "Moreover, scripture even allows for reasoning as a support, because such passages as '[The self] should be heard; it should be considered . . . ' make evident that the human intellect is an assistant for [knowing] the self." 43 . No mention is made of meditation at all.Śaṅkara puts the epistemological and soteriological weight firmly on scripture, with reasoning providing support for the process of inquiry into brahman (brahmajijñāsa). The above remarks are part of a larger argument at the beginning of the Brahmasūtrabhās . ya justifying theśāstra, that is, the Brahmasūtra itself, as a model for and enactment of the inquiry into brahman.
Just afterwards comes a key passage the meaning of which has been a source of debate. Rambachan tells us that it has frequently been cited as the "proof text" thatŚaṅkara claims experience (anubhava) to be a valid means of knowing brahman. The passage reads as follows: Unlike in the case of the inquiry into religious duty (dharmajijñāsa), scripture and so on are not the only valid means in the case of the inquiry into brahman. Rather scripture and so on as well as experience (anubhava) and so on are in this case, in accordance with their nature, valid means of knowledge (pramān . a), because knowledge of brahman culminates in experience (anubhavāvasānatva) and because it takes as its object a truly existent thing (bhūtavastuvis . ayatva). 44 Rambachan maintains that it would not be appropriate to regard this isolated passage as justification for the strong claim that experience must validate scripture, or that scripture is merely an indirect means of knowing brahman. The passage is framed by comments that scripture is, in fact, the sole means of knowing brahman. When seen in context, the abovecited passage suggests that perception, inference, and other valid means of knowledge can play a role in supporting the inquiry into brahman alongside scripture and reasoning based on it.
Comans (2000, p. 308) notes, however, that the passage also suggests that knowledge of brahman is itself an experience or culminates in one. We can add that it also says that knowledge of brahman takes as its object a "truly existent thing" (bhūtavastu). This second point reflects the argumentŚaṅkara makes elsewhere to the effect that knowledge of brahman relies for its truth on the object itself, and not on any human agency, because knowledge is not an action. This is an argument we will discuss when we look at the second passage in the Brahmasūtrabhās . ya that mentions the threefold scheme from Br . hadāran . yakaupanis . ad 2.4.5. However, Comans's first point raises a separate question about the status of knowledge of brahman. If it culminates in an experience, is this experience "conceptual" or "nonconceptual"?
The general assumption among modern writers seems to be thatŚaṅkara thinks of knowledge of brahman as a "nondual" experience akin to "nonconceptual meditation" (nirvikalpasamādhi). Consider the recent statement in Duerlinger et al. (2020, p. 38): "Knowledge of the Self takes the form of an immediate intuition (anubhava) that transcends the subject-object duality." Rambachan (1991, p. 11ff) offers many more examples of this interpretive stance, and against such an interpretation he says: It is exceedingly important to note thatŚaṅkara all along sees brahmajñāna [knowledge of brahman] as a mental process occurring in the mind and not transcending it. Brahmajñāna is of the nature of an antah . karan . a vr . tti [a mental state] coinciding with the nature of brahman and produced by its authoritative pramān . a [valid means of knowledge], theśruti [the scriptures]. (Rambachan 1991, p. 109) Rambachan emphasizes here thatŚaṅkara sees knowledge of brahman as a state of mind, but one may ask, ifŚaṅkara says that knowledge conforms to its object, and he also says that brahman is nondual, undifferentiated, and free from conceptualization (nirvikalpa), as he does on several occasions in the Brahmasūtrabhās . ya, 45 then would not knowledge of brahman as a mental state also be nonconceptual or at least give rise to a nonconceptual experience? 46 If so, and since the scriptures, which have words as their nature, are conceptual, then it becomes possible to ask the question: How, forŚaṅkara, do concepts give rise to a nonconceptual experience of knowledge? Rambachan (1991, p. 110 Rambachan readsŚaṅkara as saying that knowledge of brahman is a conventional, conceptual state of mind. Given this interpretation, Rambachan (1991, p. 68ff) explains hoẃ Saṅkara can maintain that scripture, limited as it is by its linguistic nature, can still reveal the limitless. If we grant further that the absolutely real is itself beyond conceptuality, then Rambachan's explanation may begin to suggestŚaṅkara's possible answer to the question of how conceptual knowledge practices can give rise to knowledge of the nonconceptual. 47 Comans points out a couple more passages in the Brahmasūtrabhās . ya in whichŚaṅkara speaks of the knowledge of brahman producing an effect that is "present to experience" (anubhāvārūd . ha) and "made perceptible" (pratyaks .ī kr . ta), 48 such as the following remarks on 3.3.32: Because knowledge [of brahman] has a perception as its effect, it is not reasonable to have doubt about the absence of its effect. In the case of the effects of action (karma, i.e., ritual action), however, such as heaven and so on, which are not present to experience, there could be a doubt as to whether the result will occur or not, but the result of knowledge [of brahman] is present to experience, as it says in the scriptures, "That brahman, which is direct and immediate . . . " and "You are that (tat tvam asi)," which instructs one that [the effect of knowledge of brahman] is already accomplished. The statement, "You are that," cannot be interpreted to mean, "You will be that after you have died." Furthermore, another scriptural passage indicates that one will perceive the fruit of perfect seeing (samyagdarśana), namely [the understanding] that everything is the self, precisely at the moment of perfect seeing: "Seeing that this [self] is that [brahman], the seer, Vāmadeva, understood, 'I was Manu and the sun.'" 49 Saṅkara makes several important claims here. One we will discuss more below: knowledge is distinct from action. This claim recurs when he discusses the difference between knowledge and meditation. Meditation, he will claim, is an action like ritual action. He also says that knowledge of brahman is immediately available to experience because it is something we already possess. One is reminded here of many passages inŚaṅkara's writings, including the analogy of the sun cited above from the Upadeśasāhasrī, which suggest that brahman is awareness itself. Insofar as we are already aware that we are aware, we already possess knowledge of brahman. While the scriptures may be said to remove the apparent ignorance that makes us identify ourselves with our mind/body complex, they cannot remove or grant us our awareness of ourselves. Rather, scripture simply enables our awareness to know itself. Apparently, it must do more than this, however, because it also informs us of the identity, the nonduality, of self and world. Whether this knowledge is also directly available to experience would seem to depend on whether such an experience of knowledge is conceptual or nonconceptual.

6.Śaṅkara on Meditation and Its Difference from Knowledge
The second reference in the Brahmasūtrabhās . ya to Br . hadāran . yaka-upanis . ad 2.4.5 is made by the opponent in the commentary on Brahmasūtra 1.1.4.Śaṅkara's opponent argues that the scriptures enjoin certain actions like hearing, reasoning, and meditation as the means by which one achieves the goal of knowing brahman. So, for the opponent, scripture would be a valid means of knowing brahman only insofar as it prescribes certain actions that result in such knowledge. 50Ś aṅkara rejects this argument, arguing instead that scripture is the only valid source of knowing brahman, and furthermore that scripture only makes brahman known. It does not prescribe any action, because knowledge of brahman cannot be an action. If liberating knowledge were based on an action,Śaṅkara argues, then its object, brahman, would be subject to change, but brahman is not an object, properly speaking, and thus liberating knowledge cannot be an action. The opponent even uses the gradualist metaphor of wiping clean the mirror to describe the spiritual path, and this metaphor is rejected byŚaṅkara, because, again, the self, which is identical to brahman, cannot be the object of any action. 51 Saṅkara distinguishes knowledge of brahman from reasoning and meditation. The latter he classifies as types of mental action while knowledge, by contrast, depends on the object known, and not on any human effort: Meditation (dhyāna) and thinking (cintana) are, indeed, mental in the sense that a person has the ability to do them, or not to do them, or to do them differently, because they are dependent on the person. Knowledge (jñāna), however, is born from a valid means of knowledge (pramān . ajanya), and a valid means of knowledge takes as its object something that is truly existent. Knowledge is entirely dependent on the object alone. It is not dependent on a rule. It is not even dependent on a person. Therefore, even though it is mental, knowledge differs greatly [from meditation and the like]. 52 So, forŚaṅkara, the intent of the passage at Br . hadāran . yaka-upanis . ad 2.4.5, which appears to the opponent to prescribe certain actions, is simply to divert one's attention away from the sense objects of the world and towards the self. 53 It does not require any action from us. It simply directs the mind toward inquiry into brahman. It certainly does not provide anything like a complete map of the spiritual path.Śaṅkara offers the same interpretation again in the commentary on Brahmasūtra 3.2.21: "Even those scriptural passages such as '[the self] should be seen' and so on, which speak of the duty to know the highest brahman, have as their principal purpose to bring one face to face with reality (tattvābhimukhīkaran . apradhānā) and not to require that one realize reality." 54 Such passages can only direct one's attention to reality. They are prompts, not rules. They literally bring one "face to face" with the knowledge of brahman, but they cannot cause or enjoin it because such knowledge is not something one can cause or effect. In fact, we already possess it.
There is obviously a tension here betweenŚaṅkara's desire, on the one hand, to emphasize the effortlessness or passivity (the lack of agency) with which the practitioner knows the liberating knowledge of reality, and his need on the other hand to provide a set of instructions or some kind of description of a regular process by means of which one realizes that one already possesses such knowledge. This tension is present whenever the subitist attempts to provide instructions while remaining true to the implications of spontaneously achieving a goal that is thought to be innate.

Brahmasūtra-Bhās . ya
We have seen thatŚaṅkara draws a parallel between seeing and knowing, and that he distinguishes knowledge from meditation. He does so again and again. Another instance is found in the first explicit reference in the Brahmasūtrabhās . ya to the important concept of "perfect seeing" (samyagdarśana). ForŚaṅkara it is tantamount to the liberating knowledge of brahman, as we have already seen in his commentary on the Br . hadāran . yaka-upanis . ad. The passage is found in the commentary on Brahmasūtra 1.3.13, the context of which is the assertion that the object of a particular meditation is, in fact, the highest brahman. To distinguish knowledge from meditation,Śaṅkara notes that meditation can take an unreal or imaginary thing for its object, whereas knowing, like seeing, requires a real object: In this case, an act of meditation can also take an unreal thing as its object, as, for instance, the merely imaginary object of a wish. But an act of seeing can only take a real thing as its object, as we know from experience in the world; we therefore conclude that in the passage last quoted only the highest self, which is the real object of perfect seeing (samyagdarśana), is indicated as the object of sight. 55 Saṅkara, like Kamalaśīla, is sensitive to the concern that one can meditate upon unreal objects and that meditation can thereby become ineffective or misleading. In responding to this concern,Śaṅkara draws upon the conceptual metaphor that KNOWING IS SEEING in order to distinguish knowledge from meditation. He wants to ensure the reality of the object of knowledge. However, the conclusion to the commentary on this verse suggests a kind of gradualism that seems to accept that meditation can play a preparatory role in the attainment of liberating knowledge: With reference to the objection that a fruit confined to a certain place is not an appropriate reward for him who meditates on the highest self, we finally remark that the objection is removed if we understand the passage to refer to liberation by stages (kramamukti). He who meditates on the highest self by means of the syllable Aum, consisting of three mātrās, obtains for his reward the world of Brahma and afterwards gradually "complete seeing" (samyagdarśana). 56 Saṅkara makes a noteworthy concession here. Elsewhere in the Brahmasūtra-bhās . ya, he is adamant that liberation itself can have no degrees and neither can knowledge of brahman: "For all the scriptures assert that the state of liberation has only one form. For the state of liberation is nothing but brahman, and brahman is not connected with different forms because it has the character of being one." 57 However,Śaṅkara admits that knowledge of brahman may arise after a longer or shorter period of time, depending on the strength of the means employed and the qualifications of the individual practitioner. So, here again we see a manifestation of the tension between immediacy and progression, as well as between spontaneity and regularity of causation.
These tensions come to the foreground in a short sequence of twelve verses at the beginning of the fourth and final chapter of the Brahmasūtrabhās . ya. 58 In the commentary on 4.1.1,Śaṅkara explains that these twelve verses constitute an "extra discussion" (vicāraśes . a) of the means of knowledge (sādhana). He then immediately quotes from the Br . hadāran . yaka-upanis . ad 2.4.5: "Verily, it is the self that should be seen; it is the self that should be heard; it is the self that should be considered; it is the self on which one should concentrate." The question at issue is whether one needs to do these mental practices "repeatedly" (āvr . tti), or do them "suddenly, immediately, all at once" (sakr . t). 59 Interestingly, Saṅkara has the opponent argue here that one should do them only once and the authoritative voice then counters that they should be done repeatedly. He further argues that knowledge and meditation are often used synonymously in the scriptures, and that this implies repeated practice. The commentary on 4.1.7 even gives a definition of meditation as "the activity of extending a single train of thought." 60 Throughout these verses, a variety of different meditations are accepted byŚaṅkara as valid forms of practice for the practitioner who seeks to know brahman.
Saṅkara does grant the possibility of a person experiencing liberating knowledge of brahman from hearing the scriptural passage, "You are that" (tattvamasi), only once. For those who are "slow-witted" (mandamati), however, and do not have such spontaneous experience, repetition is useful.Śaṅkara thus introduces another kind of gradualism here by appealing to the distinction-one might call it a duality-between those who are "slowwitted" and those who are "quick-witted" (nipun . amati). This distinction echoes comments at the beginning of the Brahmasūtrabhās . ya on the four sādhanas or qualifications for inquiry into brahman. One must possess (1) the ability to discern (viveka) the difference between the permanent (i.e., brahman) and the impermanent; (2) dispassion (vairāgya) towards the fruits of one's actions; (3) mental quietude (śama), restraint (dama), inwardness (uparati), patient acceptance (titiks .ā ), single-mindedness (samādhanā), and faith (śraddhā); and (4) a desire for liberation (mumuks . utva). Several of these same qualifications are also found at the beginning of the prose portion of the Upadeśasāhasrī. 61 Thus, even ifŚaṅkara asserts, as he repeatedly does, that scripture is the only valid means of knowing reality, and that knowledge is not an action, and that the highest brahman admits no distinctions, we must nevertheless understand these broader epistemological and ontological claims within the context ofŚaṅkara's overall path philosophy. The passages cited above make clear that the knowledge of reality arising from inquiry into brahman is dependent on the transformation of the person, and such a transformation implies both path and goal. The practitioner must be properly prepared for the kind of inquiryŚaṅkara envisions and enacts in the Brahmasūtrabhās . ya itself, and such inquiry occurs within a specific social context. In the commentary on 4.1.12, the final verse of the short section on the means (sādhana), he acknowledges once again that some meditations have as their purpose the attainment of perfect seeing: The first section of the present chapter has established that one should practice repetition of all meditations. With respect to those meditations that have "perfect seeing" (samyagdarśana) as their goal, however, a distinction is made that such meditations terminate when their effect is accomplished, just as one stops beating rice when the husks are separated from the grains. For as soon as the effect, that is, perfect seeing, is attained, no further effort can be enjoined, since scriptural injunctions do not apply to one who knows brahman, which is not the object of any injunction, as his own self. 62 Saṅkara states here that some meditation practices have liberating knowledge of brahman as their goal. He even employs a metaphor of repeated effort to describe the process. However, he does not say that such practices will actually produce liberating knowledge. It seems the best we can do is improve the conditions for liberation by removing the obstacles to knowledge in our minds. We do this primarily through the prerequisites to knowledge, preparing ourselves for study, thought, and meditation upon the scriptures under the guidance of the authoritative teacher; but we still cannot thereby bring about our liberation, since in truth we are already liberated.
Saṅkara's subitist impulses, arising perhaps from his nondualist ontology, also sometimes bring ethical implications, as we see in the above passage when he states that nothing can be required of the liberated person who knows brahman. Such a claim, of course, must be understood within the context of a path philosophy in which access to liberating knowledge has certain prerequisites. While one can see the apparent contradiction here, it is important to bear in mind that misunderstandings often arise when we juxtapose ostensibly contradictory claims without respecting the place of each claim within the overall system. Here, we are dealing with the difference between what Kellner (2020), following Siderits, calls "entry" and "exit" rules. ForŚaṅkara, however, there is a basic consistency between them because knowledge of brahman, being innate within us, cannot be caused, nor can it require anything of us.

An Alternative Voice from Early Vedānta: Man . d . anamiśra on the Threefold Scheme and Using Conceptuality to Attain Knowledge of the Nonconceptuaĺ
Saṅkara probably had in mind Bhartr . prapañca or an interpretive tradition he represented when he made his criticisms of meditation and the threefold scheme in the Brahmasūtrabhās . ya. It seems that Man . d . anamiśra may have hadŚaṅkara's criticisms in mind when he defended Bhartr . prapañca's views in his Brahmasiddhi. Recall that representatives of this alternative tradition upheld the practice of prasaṅkhyāna meditation, which Saṅkara appears to criticize in the Brahmasūtrabhās . ya and Upadeśasāhasrī, and they used the scriptural passages on the threefold scheme as a justification for the claim that meditation practice can help one acquire knowledge of brahman. Thus, Man . d . ana's Brahmasiddhi contains an unambiguous statement that the threefold scheme describes the means by which one destroys ignorance: Again, by what means is ignorance ended? It is by the repeated practice of learning, thinking, and meditation (dhyāna), and by the religious life (brahmacarya) and so forth, which are the different means (sādhanabheda) stated in the treatises (śāstras). How? The person who previously practices learning and thinking and then repeatedly practices meditation on the entire diffusion of differences that is denied with respect to the self-"It is not this, not this"-such a person, opposed to seeing difference, clearly causes it to end. 63 Man . d . ana goes on to explain how the practices of learning, thinking, and meditation, though conceptual, can eliminate both ignorance and conceptuality at the same time, leaving only the "pure, transparent nature" of self-awareness. Toward the end of the same passage, he raises the question explicitly: "How can one destroy difference with difference itself (katham . bhedenaiva bhedah . pratisam . hriyate)?" Man . d . ana responds, "because of its being the antidote of difference, like powder eliminates powder" (yathā rajasā rajah . ). He explains the analogy here by referring to a method by which a powdered substance is introduced into water dirtied by another substance. The introduction of the powder causes a chemical reaction whereby both substances are precipitated, leaving the water pure. He then offers two more analogies: The repeated practices of learning, thinking, and meditation on brahman, which is beyond difference, are clearly the antidote to seeing difference, even though they are connected with ignorance, just like stomach fluid digests fluid and is itself digested or poison destroys another poison and is itself destroyed. 64 Man . d . ana thus addresses the question of how a practice that depends on conceptuality and duality could give rise to a nondual and nonconceptual experience.
The Brahmasiddhi also contains a citation from Br . hadāran . yaka-upanis . ad 2.4.5, "It is the self that one should consider; it is the self on which one should concentrate (mantavyo nididhyāsitavyah . )," which is given in support of the view that repeated practice of meditation, ritual action, and so forth is necessary even after a correct vision of reality has arisen, because it helps to eliminate any regression toward seeing difference and makes a correct vision of reality continuous and stable. 65 The repetition of meditation and other practices is necessary, according to Man . d . ana, because: False appearances (mithyāvabhāsa) persist even for a person who has understood the true nature of the self from the scriptures, about which there is no doubt, because of the strength of latent dispositions that have power due to the accumulation of repeated false perceptions from beginningless time . . . Therefore, even if a vision of reality has arisen from a valid source of knowledge, it is thought that repetition of the vision of reality serves to overpower or destroy the latent dispositions made firm by having arisen through the repetition of false views from beginningless time. Thus it is said, "It is the self that one should consider; it is the self on which one should concentrate" (mantavyo nididhyāsitavyah . ), and there are the regulations (vidhāna) on the qualifications (sādhana) of mental quietude (śama), restraint (dama), living the religious life (brahmacarya), ritual action (yajña), and so forth. Otherwise, what is the purpose of teaching them? 66 We can see how Man . d . ana understands and interprets the key passage from the Br . hadāran . yakaupanis . ad. He, likeŚaṅkara, walks a fine line between enjoining a series of practices and emphasizing the singular power of the scriptures to reveal the true nature of reality. Interestingly, the comments above immediately lead to another question, namely, whether the scriptures can give rise to knowledge of reality only with the support of such methods as living the religious life. Man . d . ana quickly responds, just asŚaṅkara does, that hearing the scriptures can by itself give rise to knowledge of reality; the supporting qualifications are not necessary for acquiring such a vision of the truth. Man . d . ana argues that hearing the scriptures brings "certitude" (niścaya), but he also admits that one can fall back into error. Thus, repetition is said to be useful for the purpose of stabilizing the correct vision of reality and destroying false views. Kellner (2020, p. 58) writes of Kamalaśīla: On the path to liberation, a perceptual awareness of reality was regarded as necessary for the particular reason that only this kind of awareness, when preceded by a gradual process of acquiring insight, has the power of removing the afflictive and epistemic obscurations. Ascertainment by inference alone is simply not powerful enough to effect this kind of fundamental transformation of consciousness.
The problem that the repetition of meditation and other such practices, including the performance of Vedic ritual, is meant to solve for Man . d . ana is the same problem that, on Kellner's reading, motivates Kamalaśīla's arguments for the necessity of cultivating nonconceptual meditation on the gradual path. We also see here that Man . d . ana makes a distinction between epistemic certainty and spiritual efficacy similar to the one Kellner finds in Kamalaśīla. More broadly, this example reminds us of the importance of looking carefully at immediate context: both Kamalaśīla andŚaṅkara were involved in debates within their respective traditions as much as, if not more than, between them. 67 Yet, it also suggests the value of broader comparisons for what Foucault (1998) calls "the history of thought," which he describes as distinct from both "the history of ideas" and "the history of mentalities," because the history of thought focuses on "problems" or "problematizations. 68

Different Types of Meditation and the Addition of Nonconceptual Meditation to the Threefold Scheme in Post-Śaṅkara Vedānta
The nature and purpose of meditation and its relationship to various discursive practices, such as philosophical reasoning, have been among the central problems raised in this essay, but of course meditation is a rather vague term. It has various specific names, types, and definitions, some of which we have already encountered: bhāvanā, nididhyāsana, dhyāna, upāsana, prasaṅkhyāna, parisaṅkhyāna, and samādhi, not to mention vidyā and jñāna. Now, it would seem that by the time ofŚaṅkara and Man . d . anamiśra (circa 8th century) a gradual threefold process of learning, reasoning, and "meditation" (nididhyāsana, upāsana, dhyāna), culminating in the state of "perfect seeing" or liberating knowledge, either had become or was in the process of becoming an established doctrine in Vedānta. At some point, a fourth stage, samādhi, "meditative absorption" or "concentration," was added to these three practices. For example, Mayeda (1979, p. xv) draws our attention to the much later example of the Vedāntasāra of Sadānanda (circa 15th or 16th century?), which includes samādhi after the threefold scheme. 69 An earlier example of the threefold scheme being augmented by a fourth stage involving nonconceptual meditation (nirvikalpasamādhi) is the Vivekacūd .ā man . i, but how much earlier is unclear: the verse text's attribution toŚaṅkara has been the subject of ongoing debate. 70 In the midst of a striking series of verses (361-365ff) in which we also find the "gradualist" metaphor of purifying gold, 71 we find the following statement: "The virtue of thinking (manana) should be considered a hundred times greater than learning (śruti), and meditation (nididhyāsa) has a virtue one hundred thousand times greater even than thinking, while [the virtue of] nonconceptual meditation (nirvikalpaka) is limitless." 72 The next verse then clearly states that nonconceptual meditation is the means by which one comes to know the reality of brahman (brahmatattva).
The addition of this fourth stage called samādhi suggests several things. For one thing, it points to the general popularity in Indian thought of the idea that meditation defined somehow is necessary for liberation. Insofar as samādhi here stands primarily for nirvikalpasamādhi, nonconceptual meditation, the addition of a fourth stage may be seen as further evidence of the growing popularity of the notion of nonconceptual meditation among the path systems of classical India at the time. Additionally, and quite significantly, the addition of a fourth stage called samādhi suggests how the earlier Vedānta tradition, includinǵ Saṅkara and Man . d . anamiśra, understood the concept of nididhyāsana or dhyāna itself. The tradition could have simply reinterpreted the concept of nididhyāsana to include samādhi and even nirvikalpasamādhi. The fact that it did not seems to suggest that nididhyāsana was understood primarily as a conceptual meditative practice, for instance, a practice involving repeatedly fixing one's mind upon the "great sayings" (mahāvākyas) of the Upanis . ads, such as "You are that" (tat tvam asi), which are, of course, discursive and conceptual.
Whether the experience of liberating knowledge to which the conceptual practices of learning, reasoning, and meditation pave the way was thought to be nonconceptual, and if so, then how such conceptual practices as these ones could lead to nonconceptual experience: these remain, for me, questions worthy of discussion. In this respect, it would be interesting to engage in a more detailed comparison of the meditation practices described byŚaṅkara and Man . d . anamiśra, particularly the practices of "repetitive contemplation" (prasaṅkhyāna) and "contemplative rehearsal" (parisaṅkhyāna). In fact, after composing the musings and questions above in the process of completing the penultimate draft of this essay, I became aware of , which engages in precisely such an investigation. He concludes, in short, that the difference between prasaṅkhyāna and parisaṅkhyāna meditation is akin to the older distinction between yoga and sāṁkhya, respectively. 73 His analysis makes for compelling reading and could form a solid basis, alongside works like Kellner (2020) and Adam (2016), for pursuing a broader comparison of these meditation methods in early Advaita Vedānta with Buddhist conceptions of meditation, such as Kamalaśīla's description of the "contemplation of the real" (bhūtapratyaveks .ā ) as a method of "analytical" (vipaśyanā) meditation, and how all these meditation practices relate to and emerge from practices of reason in their respective traditions. 74 Whether or not they give nonconceptual meditation or knowledge a prominent place in their path philosophies, it seems that all three philosophers mentioned above place an emphasis on the conceptual and discursive nature of meditation practice.

Conclusions
Although focused on a common scheme or model of acquiring knowledge, this essay has highlighted some significant epistemological questions arising from basic differences in the final ontology of the two philosophical traditions under discussion. Advaita Vedānta claims that each person possesses an essential nature, the self orātman, which is both numerically and qualitatively identical to the essence of all things, brahman. Kamalaśīla denies that self and essence exist. Instead, according to him, everything arises in dependence on other things; that is, everything is selfless, empty of essential nature. While the philosophers under analysis agree that our ordinary experiences mislead us, they hold seemingly contradictory positions about the ultimate nature of things. If we apply the traditional formula, we get four possibilities: (1)Śaṅkara is right and Kamalaśīla is wrong.
(3) Both of them are right (in some way).
(4) Both of them are wrong (in some way).
The first and second options most closely parallel traditional polemics. Both Kamalaśīla andŚaṅkara are committed to the truth of their respective positions and the corollary that its truth entails the falsity of the opposing view. The third option could lead us down the road toward perennialism, the idea that all religious philosophies ultimately express the same truth, but we need not take it so far. Perhaps there is something in what both are saying that could point toward the truth. For example, they seem to agree that some kind of conscious awareness remains present even in ostensibly nonconceptual or nondual states of mind. I take this to be the import of Kamalaśīla's characteristic doctrine of "reflexive awareness" (svasam . vedanā), 75 while forŚaṅkara it is a basic implication of the conscious nature of brahman. The fourth option opens the door to entirely different ways of conceptualizing reality. Perhaps both final ontologies overdetermine the empirical evidence that experimentation can provide. In any case the basic epistemological question remains: how can one determine the truth of the situation?
Kamalaśīla andŚaṅkara appear to agree that meditation will not help one solve the problem. As "continuity" theorists, both argue that effective and epistemically valid meditation practice must follow from correct reasoning and correct reasoning follows from and issues in a correct understanding of scripture, one likely developed in close contact with a teacher or community of practitioners. Neither would go so far as to accept the possibility of a scenario in which an experienced practitioner of his tradition, who had been devoted wholeheartedly to its teachings and practices, might discover on the basis of advanced meditation experience that his own tradition is wrong and the rival's tradition is right. 76 Each philosopher is committed to a particular body of knowledge that he seeks to defend. Yet, the continuity thesis gives rise to a challenging epistemological question: if the truths accessible through meditation merely conform to those already available through learning and reasoning, how do we know that we are not simply fooling ourselves into believing that our meditation practice gives us access to the truth? In other words, how do we overcome the problem of "autosuggestion" or even "confirmation bias" in our meditation practice? 77 ForŚaṅkara, the answer to this question could be as simple as saying that scripture tells us the truth. It has intrinsic validity. Scripture removes ignorance, whereas knowledge is not an action. The concern with autosuggestion arises only from the perspective of a skeptical application of reason detached from belief in scripture. Such skepticism generates an infinite regress whereby no knowledge can ever be validated. According to Tillemans (2016, p. 191), Kamalaśīla's answer is to say that meditation must be validated by reasoning, but one wonders whether this answer entirely solves the problem and if Tillemans has adequately represented his position. Kamalaśīla deserves further attention on this question. Does not valid reasoning for a Buddhist philosopher like him also require some foundation in empirical investigation in order to grant access to the truth? On the other hand, when Kamalaśīla speaks about the nature of reality, he relies heavily on the authority of scripture, the Buddha's words. It is commonly said that for Buddhist philosophy the validity of the Buddhist scriptures derives from the Buddha's own personal experimentation, his own direct insight into the nature of reality after having rejected alternatives. That is fair. The traditional narratives of the Buddha's attainment of liberating knowledge do not characterize him as accepting one among existing alternatives. Instead, he rejects all of them, forges his own path to the truth, and subsequently establishes his own institution. The situation is quite different with Kamalaśīla andŚaṅkara, however. Both are defenders of established institutions as well as rationalist philosophers working from within centuries-old traditions. Most traditional Indian narratives of religious conversion likewise feature philosophical debate or the performance of wonder-working or both, contests waged between already well-established positions. 78 That said, it would be valuable to explore classical Indian narrative literature to see if there are any traditional stories of religious conversion or liberation that lend support for an independence theory, and if so, how.
Independence theorists face some equally challenging epistemological questions of their own. For example, if the independence theorist argues for the unique epistemic value of nonconceptual meditation, a mental state in which all discursivity or conceptual thought has stopped, then how could nonconceptual meditation provide any information at all? We have not dealt with actual independence theories in this essay, but it would be valuable to do so. 79 If we grant for the sake of argument that something like nonconceptual experience is possible, then it stands to reason that nonconceptual meditation must be available to conceptualization in something like the same way that, for Kamalaśīla and Saṅkara, conceptual or theoretical knowledge can give rise to and become available for nonconceptual experience. Beyond that, we might ask whether the independence theorist, as someone who appears committed to the claim that all conceptual constructions distort reality, would find any value in working out a theory of epistemic independence at all. I began this essay with the observation that epistemological or philosophical questions, as Foucault defines them, do not go away simply by decentering the concern with truth in order to focus on the transformative techniques of traditional path philosophies. We have seen thatŚaṅkara and Kamalaśīla were deeply concerned with questions of truth and falsehood. We have also seen how this concern motivates some of their uses and critiques of the threefold scheme. I want to conclude by emphasizing once again the importance of understanding their engagement with epistemological questions within the broader context of their spiritual path philosophies. We are not dealing here with detached speculations on the nature of reality but with committed articulations of particular strategies of self-transformation.Śaṅkara and Kamalaśīla were equally, if not more, concerned with techniques of "spiritual" efficacy. This does not mean they were "pragmatists" concerned only with what works, or that they upheld a pragmatic theory of truth, but that they saw epistemology, ontology, and path philosophy as mutually implicating areas of concern. It remains an open question, at least for me, how best to compare their situation with the one in which we find ourselves today. Although many people in modern times have run from truth, albeit in several different directions, the framework of the threefold scheme and the conceptual metaphor of liberating knowledge as a form of perception point to a concern some still share with understanding the nature of knowledge, the process whereby a person comes into relationship with knowledge, and the effect of knowledge upon the knowing individual. Such a concern brings us face to face with the aporias of truth and self-transformation: why must we transform ourselves to access a truth we possess innately? On the other hand, if we do not already possess a certain knowledge, how could we acquire it without ourselves being changed in some way? We cannot easily disentangle epistemological questions from problems of path philosophy.
Funding: This research received no external funding.

Informed Consent Statement: Not applicable.
Data Availability Statement: Data sharing is not applicable to this article. No new data were reported in this study.

Conflicts of Interest:
The author declares no conflict of interest.
Notes 1 Foucault (2001, p. 16). My translation. See also Collins (2020, pp. xxxi-ii). 2 Pierre Hadot had reached a somewhat similar conclusion, although he developed a reading of Aristotle that emphasized theory as a form of practice. See, for instance, Hadot (2002), especially chapter six, and the comments of Kapstein (2001, p. 9). 3 Foucault uses the term "spiritual technique" in the 1978 interview published as "The Theatre of Philosophy" (La scène de la philosophie) on which see Collins (2020, pp. 89-90). As is now well-known, Foucault developed his notion of "technologies of self" in dialogue with the work of Pierre Hadot and his notion of "spiritual exercises" (exercices spirituels). On the use of Hadot (and Foucault) in exercises of comparison with Buddhism and non-western traditions, see Collins (2020) as well as the essays collected in Fiordalis (2018a) and the references therein. 4 On Foucault's skepticism, see the passing remark of Collins (2020, p. xxxv). 5 Fiordalis (2018b) looks at the use of the threefold scheme in Asaṅga and Vasubandhu (circa 4th century) and argues that the "continuity" thesis also gives us an accurate description of Vasubandhu's position on the relationship between reasoning and meditation. 6 Gómez (2015Gómez ( , pp. 1125. The above is Gómez's translation. The concept of "all-knowledge" or "omniscience" (sarvajñā) indicates the goal of the path. For more on this concept in the works of Kamalaśīla, see McClintock (2010). 7 Gómez (1987, pp. 77-78), italics in the original. The translation of Huineng's poem is also from Gómez (1987, p. 73). 8 A great deal has been written on the so-called Samye (bSam yas) debate. In the introduction to his edition of the first Bhāvanākrama, Tucci ([1956Tucci ([ -1958Tucci ([ ] 1986) discusses some of the Tibetan sources that link the composition of the three Bhāvanākrama treatises to a formal debate. On the nature of the debate itself and whether it was actually a series of written correspondences, see the secondary sources by Demiéville and Ueyama cited in Kellner (2020, p. 41), footnote 5. See also Higgins (2016) for a bibliography of other relevant studies. 9 In this respect, we could also consider the materials selected and studied in Gómez (1983a). 10 For evidence, Adam cites Balagangadhara (2005, p. 105ff), which actually makes only a somewhat cursory reference to the threefold scheme, but Adam also cites an unpublished masters thesis by Christine Fillion from McGill University on the threefold scheme inŚaṅkara's Upadeśasāhasrī. On that basis he refers to the two passages in the Br . hadāran . yaka-upanis . ad that will concern us in this article, especially when we look at the threefold scheme inŚaṅkara and early Vedānta. Kapstein (2001, p. 25), footnote 29, also includes a brief reference to the Br . hadāran . yaka-upanis . ad and notes the similarity of its progression with the Buddhist scheme. By contrast, no mention is made of this broader non-Buddhist context in Eltschinger (2010), an excellent article focused on the development in Buddhist thought of the concept of wisdom arising from reasoning and a commonly cited source on the threefold scheme. Fiordalis (2018b, p. 255) makes only a passing reference to Adam (2006) and Balagangadhara (2005). 11 Rambachan (1991) gives many concrete examples in the introductory chapter of his work from which it becomes abundantly clear that the threefold scheme is well-known in Indian religious philosophy from its presence in the Vedānta tradition. A casual, if well-informed, search of the internet reveals the same thing. Ramana Maharshi would be an interesting modern figure to consider, but there is no space to do so in this article. 12 Scholars debate the dating of and historical relationships among these important early Vedānta figures, but we can be reasonably certain that Gaud . apāda and Bhartr . prapañca precededŚaṅkara and Man . d . anamiśra, who may have been contemporaries. On the vexed figure of Gaud . apāda, see Potter (1981, pp. 103-5). On Bhartr . prapañca, who wrote a now lost commentary on the Br . hadāran . yaka-upanis . ad, see Hiriyanna (1957, pp. 79-94) and more recently Andrijanić (2015) and . On Man . d . anamiśra and the question of his relationship toŚaṅkara, see Thrasher (1979) and Tola (1989). See Harimoto (2006) for a discussion of the dating ofŚaṅkara. On the apparent influence of Buddhism onŚaṅkara (and Gaud . apāda) see, for instance, Whaling (1979), which also discusses the view thatŚaṅkara introduced into the Hindu tradition a monastic organizational structure on par with and influenced by Buddhist monasticism. Regarding Bhartr . prapañca and Man . d . anamiśra, it seems clear that they were influenced by the Yoga tradition, and if so, then they were also influenced by Buddhism, albeit indirectly, as Buddhism probably had a strong influence on development of the Yoga tradition. 13 Foucault apparently misidentifies the relevant chapter from Epictetus' Dialogues, but the editor points the reader toward book 1, chapter 4, "Of progress or improvement." For a translation, see Long (1890, pp. 13-17). There is also a question about the Greek terminology. In his French translation of Plato's Republic, Leroux (2002, p. 585), note 128, comments on a passage from Book III, section 407b-d: "Plato distinguishes here between study (mathēsis), activities of reflection (ennoēseis), and concentration on oneself (melétas pròs heautòn) . . . this list puts us in the presence of the three registers of philosophical exercise . . . " My translation of the French. In their English translation of the relevant passage from the Republic, Grube and Reeve give " . . . learning, thought, or private meditation . . . " (Cooper 1997(Cooper , p. 1043. At issue is the distinction between the practices of "thinking" or "reflection" (ennoēseis) and "concentration" or "meditation" (meletē), and the conceptual question of whether meletē and indeed all these practices should be classified as types of askēsis or not. A note of thanks to Marc-Henri Deroche for drawing my attention to this noteworthy passage from the Republic. 14 I became aware of Aleksandar Uskokov's dissertation, , only at the final stage of revising this article for publication, thanks to a comment from one of the anonymous reviewers for the journal. Uskokov's work includes a thorough discussion of the threefold scheme inŚaṅkara and early Advaita Vedānta and explores a number of questions raised in this article. In fact, it is precisely the recent study I wish I had known earlier because it would have made my research easier and allowed me to do more with this article. Rather than revising my entire article in light of his research, however, I have chosen to keep it intact, making only a few references to his dissertation here and there, and I would encourage interested readers to consult it and compare our interpretations. 15 Demiéville (1987, p. 16), noting Deussen (1912, p. 403). 16 The terms "passivism" and "passivity" will be used here to refer primarily to the distinctionŚaṅkara draws between knowledge as a state of passive receptivity and a more "active" sense of acquiring knowledge or liberation. It is thus distinct from "quietism," at least as Tillemans (2016, p. 3) defines the term: "reasoned disengagement from all philosophical theses and hence debates (vivāda) about them." Gómez (1987, p. 128) notes that Zaehner consideredŚaṅkara a "quietist," but I leave open the question of whetherŚaṅkara's position has any similarities with the "quietism" of the Spanish mystic, Miguel de Molinos, and other Christian "quietists." We should bear in mind that "quietism" and "passivism" (as distinct from "pacifism") have been given various pejorative connotations which should not be imported unconsciously into this context but considered critically alongside the positions described here and below. 17 Gómez refers to the well-known classifications found in Potter (1963). 18 Gómez (1983aGómez ( , 1983bGómez ( , 1987 is exemplary in this respect. 19 Foucault (2001, p. 17). My translation. 20 See the clear statement byŚaṅkara in the commentary on the Br . hadāran . yaka-upanis . ad cited and translated by Rambachan (1991, pp. 66-67). Consider also Uskokov (2018, p. 441ff), and chapter seven, on the key role played by the desire for liberation in this path system. 21 This conceptual metaphor is identified and discussed in Lakoff and Johnson (1980, pp. 48, 103-5). See also Lakoff and Turner (1989, pp. 48, 158, 190-91, 206). Following them, I capitalize key metaphors for the sake of clarity and emphasis. KNOWING IS SEEING seems related to, but distinct from, the conceptual metaphor, SEEING IS BELIEVING, which is only half of a quote from the 17th-century English clergyman, Thomas Fuller: "Seeing's believing, but feeling's the truth." There is clearly a productive tension to be considered further between seeing as belief and seeing as knowledge, both in general and in the history of Indian thought. In this respect and others, it would be useful to consider Bouthillette (2020) more fully than I have been able to do in this article. 22 Foucault (2001, p. 18). My translation. 23 For an overview of the influential Mīmām . sā philosopher, Kumārila's position on the differing epistemic values of scripture and meditation (or yogic perception), see McCrea (2009). On the special contribution ofŚaṅkara to the Vedānta tradition, consider . 24 Few specialists today seem to feel these conditions are sufficient for describing knowledge, but most still agree that they are either necessary for or even largely constitutive of knowledge. For a lucid discussion, see Ichikawa and Steup (2018). 25 Tillemans (2016, p. 191). Italics in original. 26 Tucci (1958, p. 198): tatra prathamam . tāvatśrutamayī prajñotpādanīyā | tayā hi tāvadāgamārtham avadhārayati | tataś cintāmayyā prajñayā nītaneyārtham . {Ms: nītaneyārthatayā} nirvedhayati | tatas tayā niścitya bhūtam artham . bhāvayen nābhūtam | anyathā hi viparītasyāpi bhāvanād vicikitsāyāś cāvyapagamāt samyagjñānodayo na syāt | tataś ca vyarthaiva bhāvanā syāt | yathā tīrthikānām | uktam . ca bhagavatā samādhirāje | nairātmyadharmān yadi pratyaveks . ate | tān pratyaveks . ya yadi bhāvayeta | sa hetur nirvān . aphalasya prāptaye | yo anyahetu na sa bhotiśāntaye | iti | tasmāc cintāmayyā prajñayā yuktyāgamābhyām . pratyaveks . ya bhūtam eva vastusvarūpam bhāvanīyam | vastūnām . svarūpam . ca paramārthato 'nutpāda evāgamato yuktitaś ca niścitam. The translation above is my own, but the passage is also cited and mostly translated in Kellner (2020, pp. 55-56), and I have benefitted from her reading and translation as well as that of Adam (2002, p. 128). Another translation is found in Beyer (1974, pp. 104-5). 27 Kellner (2020) explores Kamalaśīla's answer to this question in more detail. 28Ś aṅkara (1964, p. 770): sarvathāpi tu yathāāgamenāvadhāritam . tarkatastathaiva mantavyam | yathā tarkato matam . tasya tarkāgamābhyām . niścitasya tathaiva nididhyāsanam . kriyata iti pr . thaṅnididhyāsanavidhiranarthaka eva. The editor's name is not provided in my copy of this edition. This edition and indeed the editions of all the Vedānta works cited in this article can be compared with the online editions of these texts available on the Göttingen Register of Electronic Texts in Indian Languages (GRETIL). All the translations from Sanskrit here and below are my own unless otherwise noted. 29 See Mayeda (1979, p. xv), note 8. In the introduction to his Sanskrit edition of the Upadeśasāhasrī, Mayeda (1973, pp. 66-67) makes the same suggestion while discussing the formation of the text: that the content of the three chapters of the prose part appears to illustrate "respectively the stage of hearing (śravan . a), the stage of thinking (manana), and the stage of meditation (nididhyāsana), which constitute the three Vedāntic stages to attainment of final release (moks . a)." Again, citing the same pages from Hacker (1949), Mayeda concludes that the prose part "constitutes the whole which is complete both in content and in form," and that "the three prakaran . as of the Prose Part were written at one time after the composition of at least the 15th prakaran . a of the Metrical Part," and "that the three prakaran . as originally constituted a work independent of the Metrical Part." Mayeda describes the prose text as "a handy 'Guide' for teachers while the Metrical Part is, at it were, a 'Text Book' for students." 30  further demonstrates that Bhartr . prapañca (circa mid 6th century) also supported prasaṅkhyāna meditation, and that the practice likely had its origins in the yoga tradition. See especially chapter five of his dissertation, which explores the topic in detail. Therein Uskokov (2018, p. 236) describes Bhartr . prapañca as being "along with Kumārila,Śaṅkara's main foil," which suggests the significant influence he had onŚaṅkara's thinking. 31 See Mayeda (1979, p. xv), note 7. Rambachan (1991, pp. 1-14), especially 10ff, offers a useful survey of views; in chapter five, Rambachan (1991, pp. 97-116) goes into a more detailed analysis of the threefold scheme and attempts to reconcile it with his previous arguments aboutŚaṅkara's epistemology and soteriology. See also  for a thorough treatment. 32 For the Sanskrit of the Br . hadāran . yaka-upanis . ad itself, I have mainly followed the edition in Olivelle (1998, p. 66ff). 33 This concept seems key inŚaṅkara's thought. We will look at its presence in the Brahmasūtrabhās . ya more below. The first occurrence of the term therein is found in the commentary on 1.3.13, a translation of which is found below, but for the whole passage one can also look at Thibaut (1890, p. 172) and also consider his footnote 2. 34 That is, the mind becomes settled down or calm and therefore clear, like pure, still water. This state of mental clarity would seem to be the precondition for experiencing oneness with brahman. The verb used here is related to the Buddhist concept of prasāda, "faith or trust." 35Ś aṅkara (1964, p.  Nididhyāsana must derive from ni-dhyai, which puts it in the same etymological stemma as dhyāna, which also derives from dhyai. Another key term for meditation in the writings ofŚaṅkara, that is, in the Brahmasūtrabhās . ya and his commentaries on the Upanis . ads, is upāsana. Samādhi is not commonly used therein. According to Comans (1993, pp. 23-24), it is found only three times in the Brahmasūtrabhās . ya and never prominently as a feature of the authoritative (siddhānta) viewpoint. For instance, in Brahmasūtra 2.3.39 and commentary it is found in close proximity to a reference to Br . hadāran . yaka-upanis . ad 2.4.5, but the commentary on Brahmasūtra 2.3.40 makes it clear that these verses are considered byŚaṅkara to be arguments of an opposing viewpoint. 37 Rambachan (1991, p. 157), note 62, sees this change as significant, perhaps as a basis for arguing that nididhyāsana can be seen as equivalent to knowledge and thus distinct from meditation. Rambachan also prefers "contemplation" to meditation as a translation for nididhyāsana. However, the work of Uskokov (2018, pp. 191, 204ff), especially chapter three, suggests to me a different possibility, as he argues for the general equivalency of knowledge (vidyā) and meditation (upāsana) in the doctrine of the Upanis . ads and in early Vedānta, that is, in Bādarāyan . a's Brahma-sūtra itself, and for the importance of meditation as the primary means to liberation therein. 38 For the first claim, see, for example, the comment on Brahmasūtra 2.1.3: "Knowledge of reality, however, arises only from the Vedānta texts alone" (tattvajñānam . tu vedāntavākyebhya eva bhavati). Rambachan (1991, pp. 46, 140), note 89, also cites this passage. On the second claim that hearing scripture can produce immediate knowledge of brahman, see particularly Madhavananda (1921, p.161); Grimes (2004, p. 209):śruteh .ś atagun . am . vidyānmananam . mananādapi | nididhyāsam . laks . agun . am anantam . nirvikalpakam. The verse is numbered 364 in the former and 365 in the latter. 73 See Uskokov (2018, p. 480) for a clear statement to this effect. He also demonstrates the strong influence of the Yoga tradition upon the development of prasaṅkhyāna, and shows how the roots of parisaṅkhyāna can be found not only in the Mīmāṁsā tradition but also in the Mahābhārata. The story he tells is of the centrality of meditation in the Upanis . ads, the Brahma-sūtra, and pre-Śaṅkara Advaita Vedānta, and for its recovery as a central practice in the work of Vācaspatimiśra (circa 9-10th century), who was responsible for synthesizing the philosophies ofŚaṅkara and Man . d . anamiśra. The window forŚaṅkara's rationalism appears narrow indeed. 74 It would also be interesting to consider further the place of the threefold scheme in the Yoga tradition, more broadly, and how Yoga path philosophers account for the shift from conceptual practices to nonconceptual meditation. Note, for instance, that the Yogasūtra, book 1, verses 48-50, mentions the concept of "truth-bearing wisdom" (r . tambharā prajñā), which seemingly arises from a type of samādhi, and contrasts it with wisdom that derives from learning (śruta) and inference (anumāna). 75 For a brief discussion of this concept, see Kellner (2020, pp. 65-67). For fuller treatments, see Williams (1998) and Arnold (2010). 76 See Adam (2016, p. 372), note 37, for his take on Kamalaśīla's position here. 77 This concern is raised in Tillemans (2013Tillemans ( , 2016. 78 For an overview of Buddhist debate narratives, see Cabezón (2008, pp. 71-92). 79 Tillemans (2016, pp. 193-94) identifies the 14th-century Tibetan Buddhist philosopher, Longchenpa (Klong chen Rab 'byams pa), as a good candidate, and indeed a comparison between him andŚaṅkara could be fruitful. In the process, it may be that the distinction between "continuity" and "independence" theses will need to be rethought along with, of course, the one between "subitism" and "gradualism."