The Philosophy of Adi Sankaracharya

Introduction

This post will look at the classical Advaita Vedanta of Shankaracharya and how he dealt with some basic questions of epistemology and soteriology. The presentation will stay close to what Shankara actually said and avoid speculative interpretations of his thought, such as how Advaita Vedanta might be meaningfully adapted so as to suit the needs of modern Westerners. For the most part I will draw upon Shankara's commentaries on the Brahma Sutra and Brhadaranyaka Upanishad, perhaps his most important works, but I will also refer to his other writings. What follows will consist of translations of some of the more pertinent meta-theoretical discussions in Shankara's works followed by commentary upon selected passages. At times, I have modified and condensed Shankara's discussions so as to clarify their meaning. The translations are often not literal but I think I have faithfully encapsulated Shankara's sense. I invite readers to consult the standard translations of Thibaut and Madhavananda, both of which are fairly reliable.

I. Epistemology and Authority

A. Perception

Shankara begins his introduction to the Brhadaranyaka Upanishad by delineating the domains of revelation and worldly knowledge. With respect to the former, he says the Vedas have authority in two areas: with regard to the knowledge of the brahmanic ritual, which ultimately aims at attainment of the heaven-world (svarga), and with regard to soteriological knowledge, which aims at the highest end of man -- release (moksha). Here, Shankara acknowledges that the Vedas do not have authority in the worldly domain of practical affairs:

The Vedas are devoted to teaching the correct means to attain what is beyond the range of reason or perception. As for matters within the range of worldly experience, perception and reason alone are valid but not the Vedas.... Thus the Upanishads give instruction about the Self...

Because the Self transcends the worldly means of knowledge, it is only known by way of revealed scripture (shruti; agama), i.e., the Vedas (Brahma Sutra Bhasya 1.1.3; 2.1.3; 2.1.6; Brhad Up Bhashya 3.3.1; 3.9.26; 4.4.20; 4.4.22). Specifically the Self is known through those scriptures that teach about the nature of the Self, i.e., the Upanishads. The other means of knowledge, such as reasoning, can help in the imparting of such knowledge, but they are not valid sources of knowledge about the nature of the Self when they are not guided by scripture.

Early in Shankara's introduction to the Brhadaranyaka Upanishad, an objector asks if the self is not known from perception:

Interlocutor: Is not the existence of the self a matter of perception (pratyaksha)?

Answer: No, for we see a divergence of views (vadi-viprati) on the matter. The Buddhists and materialists, for example, dispute the existence of the self. So it cannot be a matter of perception for no one disputes the existence of a real object before oneself, like a jar held in one's hand.

Here, Shankara makes use of one of his favourite arguments against the worldly means of knowledge. When the worldly means of knowledge are extended beyond their legitimate application and delve into areas that are not their domain, they descend into conflict. Here, he points out that the nature of the self cannot be a matter of perception since we find so many different theories as to the nature of the self. If it were simply a matter of perception, we would not find so many different theories.

The question of the perception of the self is also raised in the opening sections of the Brahma Sutra commentary. There, Shankara says that though the Self, or brahman, is a reality (vastu), it is not an object (vishaya) of knowledge. At Brahma Sutra 1.1.2, an interlocutor suggests that if brahman is a reality, it ought to be an object of perception:

Interlocutor: Well then, if brahman be a real thing it should be amenable to the means of knowledge like perception (pratyaksha).

Answer: No, for brahman is not an object of sensory (indriya) perception.

The primary reason that Shankara gives as to why the Self cannot be seen is that it has no form (rupa). But generally, Shankara holds that the Self cannot be an object of knowledge because the Self is the pure subject (vishayin), and as such, it cannot become an object of knowledge. Here, Shankara basically follows the teaching of the Brhadaranyaka Upanishad, which says that the Self, as the Seer (drashtr), is never the seen (Brhad Up 3.7.23) as one cannot see that which is the Seer of sight (Brhad Up 3.4.2).

On the other hand, Shankara likens the knowledge that derives from scripture as akin to perceptual knowledge. On this point, he follows the Brahma Sutra itself, which, at 1.3.28, refers to scripture as "perception" (pratyaksha):

Only those who have quelled their conceit (shanta-darpa) and who follow the revealed scripture (shruti) are able to determine the meaning of scriptural passages concerning the nature of the gods and so on, as if they were the objects of perception (pratyaksha-vishaya). (Brhadaranyaka Upanishad Bhashya 1.4.6)

Likewise, Shankara says that scriptures concerned with knowledge of the Self teach by informing about the nature of the Self. In this regard, scriptural knowledge is akin to ostensive demonstration and perception:

But the teachings concerning brahman instruct by merely indicating, in a manner analogous to indicating some object of sight (aksha). (Brahma Sutra Bhashya 1.1.1)

Shankara also says that knowledge of brahman is like direct perception in that the cognition of brahman, like perception, is dependent upon a real thing, and not on some human construct:

Knowledge (vidya) of brahman is... dependent upon reality (vastu-tantra), like the other valid means of knowledge such as perception (pratyaksha). (Brahma Sutra Bhashya 1.1.4)

B. Apperception

But is the Self not known from apperception? Shankara acknowledges that reflexive awareness or apperception (aham-pratyaya, literally, the "I-cognition") can give knowledge of the existence of the self. But apperception cannot give specific knowledge about the nature of the Self. Shankara states that we do indeed know that the self exists from the fact of apperception. But he adds that though apperception demonstrates that the self exists, it does not tell us about the specific nature of the Self. Again, to back this claim, he points to the conflict of opinion as to the nature of the self:

Interlocutor: Is this brahman known to exist or not? If it is not known to exist, then how can we enter into enquiry about something that we know absolutely nothing about?

Answer: It is known to exist, for brahman is the self of all, and no one says, "I do not exist" (na na aham asmi iti).

Interlocutor: Well, then, there is no need for further enquiry, since the self is known (from apperception).

Answer: No, for there is a conflict of opinion (vipratipatti) as to the specific (vishesha) nature of the self. The materialists think it is the body; some think it is the senses endowed with the quality of sentience; others say that it is merely the stream of cognitive moments; others again say it is empty...and so on. (Brahma Sutra Bhashya 1.1.1)

Here, Shankara raises the issue as to what, specifically, the Self is. This is important because the answer to this question will give content to the teaching of the Upanishads and at the same time allow it to be distinguished from other teachings.

At Brahma Sutra 1.1.4, the question as to whether or not the Self is known from apperception is raised once again. Here, the objector wishes to do away with the necessity of scripture. In his answer, Shankara argues from the transcendent nature of the Self:

Interlocutor: It is not necessary to say that the Self is known only from the Upanishads because it is the object of apperception (aham-pratyaya).

Answer: No, because the Self is the transcendent witness of apperception. The Self, which is the witness of apperception, cannot be apprehended by any of the other means of knowing such as reasoning.

Here, Shankara makes it clear that the witness is not some kind of reflexive "state of consciousness" or "introspection." As the witness, the Self is the transcendental condition of such states; this is what Shankara means when he says that the Self "sees" or witnesses the I-cognition (aham-pratyaya) and when he speaks of the "Seer of sight." Since the Self is the condition for the possibility of such states, it cannot be known by way of them, any more than a tumbler can stand on his own shoulders.

While Shankara admits apperception, he does not accept the doctrine of apperception (svasamvedana) that the Vijnanavadins epsouse. The Vijnanavadins hold that cognition (vijnana) illumines both its object and itself. Though in a similar manner Shankara refers to the Self as self-luminous (svayam-jyotir), this does not mean transcendental apperception (svasamvedana) for him; it merely means that the Self needs no other of light than itself. At several points in his commentaries, Shankara rejects the possibility of transcendental apperception on the grounds noted above: the Self does not directly intuit the Self because the Self cannot become an object of knowledge, anymore than a eye can see itself, a knife, cut itself, fire, burn itself, or a tumbler stand on his own shoulders.

On the question of whether or not the inner self (pratyag-atman) is known by way of apperception, Shankara is less clear and his statements are somewhat paradoxical. Following Kena Upanishad 1.4, Shankara says that the Self is neither known nor entirely unknown (Upadeshasahashri 1.15.48-49; Brahma Sutra Bhashya 1.1.4).

In his comments on Brhadaranyaka Upanishad 3.4.2, which states, "we cannot see the Seer of sight," Shankara says it is not possible to see the inner self, which is the "Seer of sight" (pratyagatmanam drsterdrastaram na pashyeh). Elsewhere, in the comments at Brhad Up 1.4.10, an interlocutor asks if it is not contradictory to speak of Self-knowledge when, as the scripture says, we cannot see the Seer of sight. In his response, Shankara says that there is no contradiction. The Self is simply known as the Seer of sight. And when this is understood, the desire to see the Self falls away as an impossiblity (asambhava). Self-knowledge does not mean that the Self is an object of knowledge (vishayi-karana). The same objection is posed in the comments at Brhad Up 4.4.20. There, the Upanishad itself says that the Self is to be understood as eternal and one. It then says that the Self is unknowable (apramaya):

Interlocutor: But it is not contradictory to say that the self is known (jnayata) and then say it is unknowable (aprameya)?

Answer: There is no fault here. When the scripture says that the Self is not an object of knowledge (aprameya) this means that it is not known by any means of knowledge (pramana) other than scripture (agama). Identity with the self that is immediate (sakshat-atma-bhava) is not something that needs to be achieved (kartavya) because it is already existing (vidyamanatvat). For everyone is always already (nitya) identical with the Self (atmabhava).

And yet, at the same time, early in his introduction to the Brahma Sutra, Shankara admits that the inner self (pratyag-atman) does, in a way, present itself:

Interlocutor: How it is that the mind and body can be superimposed upon the Self when the Self is not an object; superimposition only occurs with respect to objects.

Answer: The Self is not absolutely (atyanta) a non-object, since, in a way, it appears as the "object" (vishaya) of the I-cognition, and because the inner self presents itself with a kind of immediacy (aparokshatva).

This last turn of phrase is a reference to Brhadaranyaka Upanishad 3.4.1, which refers to the self as immediately present (sakshat). In his comments at 3.4.1, Shankara says that this means that the inner self is well known or common knowledge (prasiddha).

In his comments on Gita 2.18, Shankara brings these two conceptions together. Gita 2.18 says that the transcendent reality, or supreme Self, is unknowable (aprameya). Shankara comments as follows:

The Self is unknowable means it is not an object of knowledge; that is, it is not definitely determinable (parichedya) by the regular means of knowledge (pramana) like perception (pratyaksha), etc.

Interlocutor: The self is determinable by scripture and by perception prior to scripture.

Answer: This is not entirely true, for the Self is self-established (svatahsiddha). Only when the self, as the knower (pramatr), is established (siddha) can the search for knowledge begin. For objects of knowledge are not determinable when the self, as the "I am," is not known. And it is not the case that the self not well known (aprasiddha) to anyone. Scripture, which is authoritative, teaches by merely removing what has been falsely superimposed upon the Self, not by indicating something (entirely) unknown.

Thus, though the Self is known only from scripture, scripture is not "proof" of the Self. The Self does not need of such "evidence" since it is self-established (svatah-siddha). And because it is self-established, it is also well known (prasiddha). As Shankara makes clear in the above, the Self is the condition of the possibility of knowledge; as such, it cannot itself become an object of knowledge. But as the condition of knowledge, it is, in a sense, "known" in all acts of knowledge (see Kena Upanishad Bhashya 2.4). It cannot be seen, and yet it shows itself through a kind of self-presentation whenever there is knowledge.

Nonetheless, though the Self is known directly (sakshat) in this manner, it is not seen for what it is in itself. As Gita 15.10 says, the deluded (vimudha) do not recognize the (anupashyati) the true nature of the Self. In Shankara's psychology, the individual (jiva) is a combination of the "I-sense" (ahamkara), mind (manas), intellect (buddhi), etc., on the one hand, and the inner self (pratyag-atman), the true and essential core of the individual, on the other. As noted above, Shankara calls the inner self (pratyag-atman) the "object" (vishaya) of the I-cognition (aham-pratyaya). What this means is that the I-cognition "denotes" the Self. How so? For Shankara, the Self is of the nature of consciousness, and because of its power to illuminate, is likened to light (prakasha; jyotir). This light illuminates the inner organ (antahkarana) and all objects of knowledge (prameya). The intellect (buddhi) catches some of this light and a "reflection" (chaya; pratichaya; abhasa) of the Self appears in the intellect. This reflection is the basis of the "I-sense." But due to ignorance (avidya), the functions (vyapara) of the inner organ (i.e., mind and intellect) are mixed up (mishra-bhuta) with the inner self (Brhad Up Bhashya 4.3.9). The true nature of the Self should be discernable by way of discrimination (viveka). But because of the conflation (samkirnatvat) of the Self with the mental self, it is not possible to determine (avadharitum) the true nature of the Self (Brhad Up Bashya 2.1.15). Thus, due to non-discrimination (aviveka), the Self is thought to be a knower (pramatr), doer (kartr), enjoyer (bhoktr), etc. when in truth it is none of these. In reality, the Self and its various limiting adjuncts (upadhi) -- the body, senses, vital airs, mind, intellect, I-sense -- are absolutely distinct (vivikta).

It is for this reason that a teaching (upadesha) based upon revelation (shruti) is required. Only in this way can the true nature of the Self be indicated. Shankara's general position is that knowledge of the nature of the Self needs the guidance of scripture. As he says in his comments on Brahma Sutra 4.1.2, the "Thou," in the scriptural formula "Thou are That," initially refers, for the student, to the inner self (pratyag-atman) understood as an agent and so on, but later it is finally ascertained as the nature of pure consciousness (chaitanya). Similarly, in his comments on Gita 8.3, Shankara says that the Self is first (pravrttam) presented as the inner self (pratyag-atman) and later, this presentation culminates (avasana) in ultimate reality (paramartha), that is, in the supreme Self (paramatman). In his comments on the Katha Upanishad, Shankara notes that this "continuum," from the inner self to the supreme self (paramatman), is known as the "adhya-atma." Thus, though the inner self presents itself with a kind of indeterminate immediacy, its true nature can only be indicated by means of scripture.

C. Experience

What about direct experience (anubhava)? Is the self not known through direct experience? While Shankara does admit that the self is known through a kind of direct experience, it is important to note that for him, this experience is carefully circumscribed by the Vedic revelation (shruti). One passage where Shankara speaks explicitly about experience (anubhava) occurs in the opening sections of the Brahma Sutra. He says:

But Vedic revelation (shruti) is not the only valid means of knowledge in the enquiry into brahman; both scripture and direct experience (anubhava) are, since brahma-jnana has its culmination (avasana) in direct experience (anubhava) and because it has an established reality (bhuta-vastu) as its object. (Brahma Sutra Bhashya 1.1.2)

Wilhelm Halbfass comments on the above: "This passage... is as significant as it is ambiguous and elusive" ("The Concept of Experience," India and Europe). One way to approach the question as to what Shankara means by "experience" here is to eliminate various possibilities.

For one, Shankara is not talking about a spontaneous mystical experience arising independently of the teaching (upadesha) of the Upanishads. Though he speaks of experience existing alongside scripture, he is also careful to say that this experience is the culmination of brahma-jnana. Since he refers to "culmination" here, the term "brahma-jnana" in the above passage refers to both the path of knowledge (jnana-marga), i.e., the inquiry into brahman (brahma-jijnasa), as well as to the final cognition of brahman. For Shankara, and the classical Vedanta in general, such inquiry always occurs in accordance with scripture since the Self can only be known from scripture. Thus, what he is saying here is that a particular cognition, fully comparable to direct experience, is the culmination of hearing (shravava), thinking (manana), and contemplating (nididhyasana) upon the meaning (artha) of the words (vakya) of the Upanishads.

Shankara also does not speak of this "experience" as some sort of "pure consciousness event" like the asamprajnata samadhi of the yogins. He is aware of the existence of such states, and though he associates samadhi with the state of deep sleep, wherein the jivatman temporarily "merges" with brahman, he does not associate brahma-jnana-anubhava with samadhi. We will deal with samadhi and meditation below.

Nor does Shankara refer to his own mystical experience or the personal attainment of extraordinary states of consciousness when he speaks of "anubhava." We may wonder why it is that Shankara does not refer to his own "experience." Similar questions have been asked of Meister Eckhart, Nagarjuna and others. This may strike us as odd, until we realize that the importance of "personal experience" has a relatively recent history. Wilhelm Halbfass comments astutely on this point:

The historically and philosophically significant question is not whether or how Shankara privately valued "personal experience," but why and how he tried to anchor it in a text, the Veda, and how he experienced this text itself as an objective revelation or epiphany that guides and anticipates all legitimate "personal experiences..." "The Concept of Experience," in India and Europe, p. 391.

The allusion to brahma-jnana as a kind of "experience" implies is that brahma-jnana, when fully developed as the final realization (samyag-darshana) of brahman, is akin to the direct perception of a real object, which is precisely what Shankara says in the passage above that speaks of brahman as an established reality (bhuta-vastu). Though this knowledge is direct, like perception, it is not contentless or indeterminate. It has as its content the identity of the inner self (pratyag-atman) with the supreme Self (paramatman). And it is determinate in that it indicates the specific (vishesha) nature (svabhava) of the supreme Self. These two facets are coordinated by Shankara in the first prose portion of his Upadeshasahashri. There, he says that once the qualifications for inquiry are met, the student should be taught the oneness of the Self (2.1.6); and then after this, the specific nature of brahman should be taught (2.1.7). Likewise, in his comments on Brahma Sutra 4.1.2, Shankara says that the experience of the Self (atma-anubhava) consists of the knowledge, "I am pure consciousness, one, and free from all suffering." (sarvaduhkhanirmukta-eka-chaitanya-atmako 'hamityesha atma-anubhavah). Shankara's favorite description of the nature (svabhava) of the Self is the compound "eternal, pure, awakened, and free" (nitya-shuddha-buddha-mukta). This specific nature is important for Shankara because it distinguishes the Advaita Vedantin's conception of self from that of the Buddhists and others.

Though this knowledge is determinate, it is, for Shankara, also immediate or direct, and it is for this reason that he refers to it as an "experience" (anubhava). Shankara likens this immediacy to the recognition or realization experienced by the "tenth boy," who, after counting his party several times to see if they have all safely crossed a river, neglects to count himself and thereby fails to see that he is the tenth -- until it is pointed out to him, "You are the tenth!" The example shows how it is possible for there to be a cognitive realization that is immediate and direct, and yet at the same time determinate, meaningful and with content.

In his comments on Chandogya 7.1.3, Shankara attempts to clarify the nature of this determinacy. The context is one in which there is a question as to how it is that the word "atma" denotes the Self:

Interlocutor: Is the Self not denoted by the term "atma?"

Answer: No. The Self is beyond description; as the Taittiriya Up says (2.4.1), "that from which words are turned away..."

Interlocutor: Then how do words denote the Self?

Answer: Though strictly speaking the Self cannot be denoted, the inner self is "denoted" by virtue of it being the remainder (pratishishta) once the adventitious conditions falsely associated with the Self are negated, just as there is the determinate specification (vishesha) of a king once his subjects are ruled out, even if we do not actually see the king.

But, it may be said, transcendent reality is beyond distinction and differentiation (nirvishesha). How can the Self be so determined? This is one of the central paradoxes in Shankara's thought. In a similar way, Shankara says that the Self is said to be discriminated (viveka) from its limiting adjuncts (upadhi), and that brahman is distinct (vyatireka) from name and form (nama-rupa). And yet brahman and the Self are also said to be beyond difference (abheda). In his comments on Brahma Sutra 3.2.34, Shankara briefly discusses this problem. There he says that difference (bheda) does not actually belong to brahman but is metaphorically said to apply to the relationship (sambandha) between brahman and its limiting adjuncts (upadhi). In reality, however, there is no real conjunction (samyoga) or contact (samparka) between brahman and its limiting adjuncts, for all relation (sambandha) is but a projection of ignorance (avidya). This same idea underlies the asparsha-yoga described in the Gaudapada Karika: In reality, the Self remains untouched (asparsha) by the effects of ignorance as its true nature is without a second (advaita) and beyond all relation (sambandha).

D. Independent Reasoning and the Conflict of Reason

In the opening sections of the second adhyaya of the Brahma Sutra, Shankara deals with the Samkhya and Yoga schools. It is this context that he treats reason as an independent means of knowledge. Here, an objection is raised that reasoning is closer to direct experience than scripture:

Interlocutor: Reason (yukti) allows us to determine something unseen (adrshta) on the basis of its accordance (samya) with what is already seen (drshta); in this sense it is closer to, and in accord with (samnikrshyate), direct experience (anubhava). Scripture (shruti), though, is less in accord and more remote (viprakrshyate) since it transmits (abhidhana) its meaning indirectly by oral tradition (aitihya). And, since inquiry (which makes use of reasoning) culminates in direct experience, its result is something seen (drshta), and so reasoning is applicable. Moreover, scripture itself says that the self is "to be enquired into" (mantavya) and thus it enjoins reflection (manana), showing that reasoning (tarka) is applicable. (Brahma Sutra Bhashya 2.1.4)

Shankara gives his answer to this objection in his comments on Brahma Sutra 2.1.6:

Answer: Although brahman is an established (parinishpanna) reality, because it is without form (rupa), it is not within the range (gocara) of perception (pratyaksha); and because it has no inferable marks (linga), it, as such, is not subject to inference (anumana). It can only be known through revealed scripture (agama). As the Katha Upanishad says, "This one cannot be attained through reasoning (tarka)" (Katha Up 1.2.9). And as the Gita says, "Neither the gods nor the maharshis know of my origin" (Gita 10.2). The passage from the Brhadaranyaka Upanishad, "the self is to be enquired into," should not be taken as referring to the application of autonomous (shushka; literally, arid or fruitless) reasoning (tarka), but to reasoning in accord with scripture (shruti-anugrhita) and as an auxiliary to direct experience (anubhava).

Shankara often points out in his commentaries that independent reasoning gives rise to conflicting theories. He derives this idea from the Gaudapada Karika (3.17), which is, no doubt, influenced by Chandrakirti and Buddhapalita on this point. In his commentary on the Brhadaranyaka Upanishad he writes:

Those logicians who reject the authority of revealed scripture (agama) give conflicting (viruddha) statements about the nature of the self -- that it is a doer, that it is not a doer, that it exists, that it doesn't exist, and so on -- and confound (akulikrta) the meaning of the shastras, and thereby make its purport difficult to grasp. (Brhadaranyaka Upanishad Bhasya 1.4.6)

Shankara's most developed statement of this idea of the conflict of reason occurs in the second adhyaya of the Brahma Sutra:

With respect to matters that are only to be known from revealed scripture (agama), independent reasoning (kevala-tarka) is not to be relied upon for the following reason: reasoning (tarka) that is without the guidance of revelation (agama) and instead attached only to human imagination/speculation (purusha-utpreksha) is without basis (apratishtata), because speculation (utpreksha) is without restraint (nirankusha; literally, "without a crook;" the image is of a lost sheep wandering about). For we see that the metaphysical arguments of clever men are shown by more intelligent men to be fallacious, and how these in their turn are refuted by still others; and so, there is no possibility of a foundation for reason (tarka) bereft of the guidance of revelation, because of the diversity of human views (purusha-mati-vairupya) on such matters. (Brahma Sutra Bhashya 2.1.11)

Shankara's language in this passage is much like that used by Bhartrhari in his Vakyapadiya (1.34). As an interesting aside, Bhartrhari also points out that much the same can be said with respect to the various interpretations of scripture. There is also the matter, first pointed out by the materialists of ancient India, that the Vedas themselves make all sorts of contradictory (vyaghata) statements. With respect to the problem of conflicting statements occurring within the scriptures, Shankara's approach toward them is to order them hierarchically in accordance with the principle of "harmonization" (samanvaya). On the problem of the conflict of interpretations of scripture, he is less clear. Although he follows the Mimamsakas with respect to the worldly means of knowledge and with regard to the nature of brahmanic ritual, with respect to ultimate soteriological concerns, he treats the Mimamsakas with the same regard he extends toward the other logicians (tarkika), i.e., with contempt. As for rival interpretations of the Vedanta, his approach is simply to reject the idea that there is any valid interpretation of scripture other than the non-dualist interpretation. He suggests that other interpretations are not "attuned" to the ultimate meaning and spirit of the scriptures, though he often defends his case by means of citation and logical argumentation. In any case, though interesting, such problems are beyond the scope of the present study.

E. The Authority of the Siddhas

As noted already, the second adhyaya of the Brahma Sutra deals with rival schools and sampradayas; in particular, it is concerned with the Samkhya. Its approach is first to question the authority and integrity of the source-texts of these other schools and then to refute their arguments. We have already looked at Shankara's general attitude toward independent reasoning. As for the secondary sources (Smrti), since they are attributed to various sages, the Brahma Sutra's tack here is to question the authority of the founders of these schools:

Interlocutor: Your account does not leave open the possibility of the authority of the Smrti-texts, such as the Yoga Sutras and the Samkhya source-texts, or the authority of rishis like Kapila. The Samkhya is also not concerned with things that are "to be done" but only with true knowledge, which is the means to release. But there is no room in your account for the texts of the Samkhya and so they thereby become meaningless. Since many people cannot understand the meaning of the shruti-texts, they rely on the Smrti-texts, which are composed by recognized authorities (prakhyata-pranatr). And the knowledge (jnana) of such men, like Kapila, is said to be unobstructed (aprahita) like that of the rishi (arsha).

Answer: If we admit your doctrine, then it, in turn, will render other Smrti-doctrines useless (like the "Vedantic" portions of the Gita, e.g.). And it is not possible for someone to perceive (upalabhate) super-sensory (ati-indriya) objects (artha) without the aid of revelation (shrutim-antarena), because there are no means (nimitta) to do so.

Interlocutor: It is possible in the case of siddhas like Kapila because they have unobstructed (aprahita) knowledge (jnana).

Answer: No, because powers (siddhi) such as super-sensory perception are dependent upon certain practices (anushthana) and such practices are characterized by things that are "to be done" (codana). (Brahma Sutra Bhashya 2.1.1)

Here, we can see that generally, Shankara does not accept the authority of sages who themselves do not recognize the Vedas as ultimately authoritative. This section of the Brahma Sutra is concerned with refuting rival schools on their own terms, but we can see here that where the authority of source-texts is concerned, there is very little that can be said when two parties simply disagree as to which source is authoritative. Shankara's argument here would appear to be that it is inconsistent to say that the Samkhya is not concerned with things that are "to be done" when its own authority is dependent upon its founding sages acquiring various siddhis, which are dependent upon things "to be done." Interestingly, as the passage continues, Shankara accuses the Samkhya of inventing the idea that Kapila was a siddha and then retrospectively reading his authority back into the tradition.

In the next passage, Shankara resorts once again to his favorite prasanga-style argument as to why sources other than the Vedas are not authoritative. Having dispensed with the validity of reasoning independent of scripture, he turns to the suggestion that the authority of the sages might serve as a foundation:

Nor can we count on some recognized (prasiddha) sage (mahatmya) like Kapila, since even here there will be no foundation, because the teachings of these recognized sages (mahatmya), as well as the founders of the other schools (tirthakara, i.e., the Buddha, Mahavira, etc.), all mutually contradict one another (paraspara-vipratipatti). (Brahma Sutra Bhashya 2.1.11)

Elsewhere, Shankara develops this argument, expanding on its implications. He then goes on to contrast revelation (shruti) with the secondary sources (Smrti):

Besides, even assuming that we can trust in the authority of these siddhas, because they instruct by way of so many different doctrines (bahu-siddhanta), their teachings will all be in conflict (vipratipatti) with one another. And then, as people are multiform (vaishvarupa) in their opinion (mati), (if we accept these teachings) the undesirable consequence (prasanga) will follow that truth (tattva) will be unregulated and without basis (avyapasthana). The Vedic revelation, on the other hand, is an absolutely independent (nirapeksham) and self-constituting authority (svarthe pramanyam). But human dicta (purusha-vacasam) are dependent upon an external basis and mediated (vyavahita) by memory (smrti) and discourse (vaktr; literally: 'talkers'). (Brahma Sutra Bhashya 2.1.1)

Here, Shankara relates the term for the secondary sources, Smrti, to its more basic sense: memory (smrti). What he is saying is that memory and verbal transmission form the basis of Smrti. (This is precisely what the interlocutor above had said about scripture, that it requires transmission (aithya).) In his commentary on Shankara's Bhashya, Vacaspati Mishra relates the issue back to the authority of the siddhas and brings "experience" (anubhava) into the equation. His argument appears to be that even though these siddhas may have various transcendent experiences, they are still required to remember those experiences, translate what they mean into a teaching, and then transmit that teaching to students. Here, being human as they are, this process may be fallible. He concludes that experience and memory (anubhvava-smrti) are less direct than scripture. Here, interestingly enough, we do have a explicit reference to the idea that the teachings of certain siddhas may be based upon their "experience." But the context is clearly one in which the authority of the siddhas is in question, which is to say that the need for authority is the actual source of the idea. And it is explicitly denied by the Vedantins that scripture has such a basis.

II. Soteriology

A. Knowledge vs. Action

Shankara's commentary upon the first four sutras of the Brahma Sutra constitutes his most basic statement of his interpretation of Vedanta. There, the over-riding concern is the relation between action (karma) and knowledge (jnana), duty (dharma) and release (moksha). The next two passages deal with the general difference between action and knowledge. Here, Shankara rejects the idea that the Upanishads teach an injunction to know the Self:

The fruits (phala) of religious duty (dharma) are transitory (anitya) since they are dependent (apeksha) upon the performance of certain practices (anushthana). But the fruit of the knowledge (jnana) of brahman, which is release (moksha), is permanent (nitya) since it is not dependent upon such actions. Religious practices involve that which is to be brought into being (bhavya), and they depend upon human effort and activity (purusha-vyapara). But the object of enquiry here, brahman, is something that is already existent (bhuta), for it is always-already (nitya). Scriptures dealing with religious and spiritual practices instruct people by enjoining (niyujyan) them to act. But the teachings concerning brahman instruct by merely indicating, in a manner analogous to indicating some object of sight (aksha).... Now, the Upanishads teach that the highest end of man is realized by the knowledge of brahman, which destroys ignorance and ends samsara. (Brahma Sutra Bhashya, 1.1.1)

Things that are "to be done" (kartavya) are dependent upon man (purusha-adhina). But there can be no option (vikalpana) with respect to what is really existing (vastu). Choosing whether to do something or not is entirely dependent (apeksha) upon the human intellect and will (purusha-buddhi-tantra). But the knowledge (jnana) of a real thing as it is in itself (vastu-yatha-atmya) is not dependent upon the mind of man; it is dependent upon the reality of the thing (vastu-tantra).... Just as validity with respect to really existing things depends upon the things themselves, so it is with brahma-jnana; it is dependent upon reality alone, because it has as its object an established reality. (Brahma Sutra Bhashya 1.1.2)

In the next passage, Shankara deals with a rival interpretation of the Vedanta, an interpretation that holds that knowledge in conjunction with action (jnana-karma-samuchaya) is the means to release. Almost all of the other Vedantins of Shankara's period held this teaching. Most of them -- such as Bhartrprapanca, Shankara's principle rival, and later, Bhaskara -- were bheda-abheda-vadins. But Mandana Mishra, the great Advaitin and contemporary of a Shankara, also held a version of this teaching. Here, Shankara is most likely addressing the bheda-abheda-vadins, who held a doctrine of gradual release (krama-mukti) as opposed to the doctrine of release in this life (jivan-mukti):

Interlocutor: The Vedas instruct about brahman but only in so far as that instruction is connected to injunctions (vidhi) to practice (karya).... Moksha arises due to ritually prescribed devotion and meditation (upasana). If the Vedic revelation only concerns declarations about what is, there will be nothing to be avoided and nothing to strive for (hana-upadana).... Moreover, we see that sometimes those who have only heard (shruta) about brahman continue to be affected by samsara. Moreover the scripture says that the self is not only to be heard, but to be enquired into (mantavya) and contemplated (nididhyasitavya). Therefore, enquiry and contemplation are enjoined by injunctions.

Answer: No. Those who practice the ritual meditations in conjunction with esoteric knowledge only attain as high as Brahma-loka, which is a temporary state.... But moksha is eternal. If moksha were dependent upon such practices and activities, it would be impermanent. And the scriptures say that release follows immediately (anantara) from the knowledge of brahman.... Knowledge of brahman is not dependent upon human activity (purusha-vyapara); it is dependent upon reality, like the other valid means of knowledge such as perception (pratyaksha).... For this reason, moksha is not something attained (prapta), like an effect (karya) brought about (utpadya) in some way; and it does not involve some kind of transformation (vikara) of the self; nor is it a gradual union or mental identification (sampad) with brahman; nor does it involve some kind of gradual purification (samskara) of the self, like the polishing of a mirror. Moksha is nothing but the identity of the self with brahman. (Brahma Sutra Bhashya, 1.1.4)

Here, Shankara rejects four different soteriological theories concerning the "cause" of release. Though moksha is described as a "fruit" (phala) of knowledge, Shankara does not accept that it is an effect (karya). For the same reason he rejects the idea that it is a "transformation" of the self; nor does it involve some form of purification, though he accepts purification as a secondary, indirect means. He also regards the mental identification of the self with brahman as an artificial mental product. Release is nothing but the realization that the inner self is none other than the supreme Self.

In the next passage, an objection is raised concerning the nature of knowledge. Here, as in his introduction to the Chandogya Upanishad, he admits that knowledge involves mental action; but, he insists, knowledge is essentially different from such action:

Interlocutor: But knowledge too is a kind a kind of mental action (manasi kriya).

Answer: No. The two are different in nature. For such action is not dependent upon the nature of some real thing (vastu); it is dependent (adhina) upon the operation (vyapara) of the human mind (purusha-citta). Meditation (dhyana) and consideration (cintana) are mental (manasam), but because they are dependent upon man, they can either be performed or not. But knowledge has as its object a real thing (bhuta-vastu); it is not something man-made, but relates to reality only. It is not grounded in injunctions (codana), nor is it dependent upon the merely human. Thus, although knowledge involves the mental (manasatva) it is completely different from it. The mental state involved in the meditation "man and woman are the fire" described in the ritual portions of the Vedas is a human activity since it is dependent upon injunctions (codana). But the cognition (buddhi) of fire itself is not dependent upon the injunctions of the Vedas nor upon anything man-made; it is dependent upon a real thing, which becomes an object of perception. It is thus a form of knowledge (jnana) which is not like an action (kriya). The self is not something to be strived after, nor does it involve avoiding anything; and the knowledge concerning it is not something that needs to be performed or effected. (Brahma Sutra Bhashya 1.1.4)

In the above, Shankara provides a rather stark and absolute disinction between knowledge and action. At yet in his comments on Brahma Sutra 3.4.26-27; 33; and 35, he admits a role for action. Some have suggested that Shankara contradicts himself here. But what he is saying is that action is merely an indirect means. Basically, Shankara is constrainted by the Brahma Sutra itself which itself probably held to something like jnana-karma-samuchaya.

Shankara's commentary on the Gita is important because it is here that he explains the relationship between those means that he does not consider direct and those means he does consider direct. In his introduction to the Gita, he states this relationship clearly:

When the Lord created the world he first made Prajapati and others and had them to practice the dharma of the path of action (pravrtti). He then created others and had them adopt the dharma associated with the cessation of action (nivrtti) characterized by knowledge (jnana) and renunciation (vairagya). This twofold Vedic dharma sustains the cosmos and leads to both prosperity and happiness on the one hand, and the highest end of man, moksha, on the other... The aim of the Gita is the highest end of man and the cessation of samsara. This comes about by devotion to knowledge (jnana-nishta) and the renunciation of karmic action. The path of action leads to prosperity in this world and rebirth in the deva-lokas. But when it is practised with complete devotion to the Lord and without any expectations (abhisamdhi) concerning its fruits (phala), the path of action is conducive to the purification of the heart, mind and intellect (sattva-shuddhi). The one whose inner organ has become clear and pure (shuddi-sattva) qualifies for the path of knowledge (jnana-nishta), which leads to the arising of knowledge (jnana). Thus, the path of action, too, is also a kind a means (hetu) to the highest end of man.

Here, Shankara admits a role for those yogas other than jnana-yoga or "devotion to knowledge" (jnana-nishta). Ultimately, they prepare the aspirant for the path of knowledge. In his Gita commentary, Shankara implies a distinction between nivrtti-marga and moksha-marga. He allows room for the various yogas of the Gita by assigning them a place in the moksha-marga. But these "means" are, nonetheless, ultimately distinguished from jnana-marga, which, properly speaking, is the only direct means to release.

II. The Threefold Means

In the Brhadaranyaka Upanishad, we read that the Self is to be known through hearing, thinking, and contemplating. In his commentary upon the Taittiriya Upanishad Shankara states that these three are to be considered direct means to release:

While austerity (tapas), celibacy (brahmacharya), and so on, are aids to purification, they are not direct means to release, while hearing, consideration, and contemplation of the meaning of the shastras are direct means to moksha. (Taittiriya Upanishad Bhashya 1.11.2-4)

Perhaps the first point to be noticed here, is that consideration and contemplation are to be practiced strictly in accordance with what has been "heard" (shruta), that is, in accordance with scripture (shruti). Meditation is not, for Shankara, some kind of "experimental" method. It is only used to confirm the truth of scripture:

Realization is not possible through independent reasoning (anumana), nor any of the other means of knowledge (pramana), though reasoning is applicable so long as it does not contradict (virodha) the Upanishads (vedanta-vakya). (Brahma Sutra Bhashya 1.1.2)

Just as enquiry (manana) through reasoning (tarka) must be accord with scripture (agama), so too contemplation (nididhaysana) must be in accord with scripture and with what has been determined through enquiry. The idea of contemplation being something independent and separate (prthak) is meaningless (anarthaka). (Brhadaranyaka Upanishad Bhashya 5.5.1)

We have disussed the role of reasoning in the inquiry into brahman (brahma-jijnasa) above and will discuss contemplation, meditation and yoga in greater detail below. Before continuing it should be pointed out that "discussion" is one of the senses of the term "vichara." Thus, when Shankara speaks of "vichara" we can also take it as referring to discussion:

The realization (avagati) of brahman follows from ascertaining (adhyavasana) the meaning (artha) of pertinent passages (vakya) from the Upanishads after their consideration and discussion (vicharana). (Brahma Sutra Bhashya 1.1.2)

Shankara often talks of the role of the teacher (acharya) in instruction and the compounds "acharya-agama" and "shastra-acharya" occur throughout his works. In his commentary upon the Brhadaranyaka Upanishad he refers to the value of discussion:

Another traditional means of acquiring knowledge (vidya-prapti-upaya)... is association with those in possession of knowledge (vidvat). Association with these sages and discussing (vada-karana) with them increases one's understanding (prajna-vriddha). (Brhadaranyaka Upanishad Bhashya 3.1.1)

Some matters... are difficult to understand, even for a group of panditas let alone someone by himself. Where the determination of subtle matters (dharma-sukshma) is concerned, it may be desirable to seek counsel (parishad), depending upon the abilities of those involved. (Brhadaranyaka Upanishad Bhashya 4.3.2)

The next question to consider is whether inquiry and contemplation are necessary. Frankly, here, Shankara is not entirely consistent. At one point in the Brhadaranyaka Upanishad, he says that hearing is not enough:

The self is first to be heard (shrotavya), from scripture (agama) through a teacher (acharya), then considered (mantavya) through reasoning (tarka), and then contemplated (nididhyasitavya), that is, meditated upon (dhyatavya) with determination (nishchayena). It is seen (drshta) by the accomplishment of these means (sadhana), hearing, consideration, and contemplation. When there is the coincidence of these three, then the vision of truth (samyag-darshana), the oneness of brahman, can occur, but not otherwise, that is, with mere hearing (shravana-matra). (Brhadaranyaka Upanishad Bhashya 2.4.5)

In his comments above, Shankara may simply be constrained by the content of the passage he is commenting upon. The overwhelming evidence, however, is in favour of the interpretation that hearing, in some cases, is sufficient:

Interlocutor: But hearing (shravana) about brahman needs to be followed by consideration (manana) and contemplation (nididhysana).

Answer: Not necessarily. Reflection and contemplation only serve the end (artha) of realization (avagati), just as is hearing does. (Brahma Sutra Bhashya 1.1.4)

This point is reiterated by Shankara when he deals with the issue of repetition. In the Upadeshasahashri, the value of repetition is raised:

The means to moksha is knowledge. It should be repeatedly imparted until it is apprehended by the student. (Upadeshasahashri 2.1.2)

A passage from the Brahma Sutra commentary deals extensively with the issue of repetition. At the end of the passage, Shankara admits that for some, "hearing" is sufficient. Basically, the interlocutor argues that if brahman is like an object of perception, like a jar sitting on a table, of what use is repeated instruction? If you point out a jar to someone and they do not understand what you mean, what good will pointing it out again do?

Interlocutor: Of what use is repetition when the object of knowledge, the supreme brahman, is an established reality.... If the passage "Thou are That" does not impart knowledge the first time, what good will repetition do?

Answer: For the one who is unable to experience the true nature of brahman at first, repetition is useful. In the Chandogya Upanishad, for example, Svetaketu asks to be instructed several times.... For we see that some people only gradually come to a true understanding of the meaning (artha) of what they have heard by the removal of false understanding.... People wrongly superimpose various objects onto the self -- the body (deha), the senses (indriya), the mind (manas), and the intellect (buddhi). Thus by one act of inquiry, one of these parts is removed and by another act, another part is removed, and so on; and thus a kind of gradual cognition takes place, though it is prior to the actual cognition of the Self. But for those with an acute intelligence (nipuna-mati), the meaning of such phrases is not clouded over with ignorance, doubt and wrong knowledge, and they are able to intuit (anubhavitum) the meaning the first time they hear it; for them repetition is not needed. For once knowledge of the Self arises, ignorance is dispelled and in that case, gradual understanding is not necessary. (Brahma Sutra Bhashya 4.1.2)

From this passage, it can be seen that Shankara admits a kind of subitism wherein enlightenment (atma-bodha) can happen "all at once."

III. Meditation and Contemplation

In his commentary on Brhadaranayaka Upanishad 1.4.7, Shankara considers various soteriological means other than knowledge of the Self. There, various forms and aspects of meditation, contemplation and yoga are discussed. One form of meditation that he rejects as not conducive to release is "upasana," a term used collectively to refer to the meditations, devotions and rituals of the jnana-karma-samuchaya Vedantins. We have seen how Shankara generally deals with the jnana-karma-samucaya-vada in the opening portions of the Brahma Sutra commentary. In his commentary on the Brhadaranayaka Upanishad he deals repeatedly with this interpretation of Vedanta, as well as with the brahmanic ritualists in general.

In the next passage, the jnana-karma-samucayin suggests that the Upanishads provide an injunction to practice a particular form of meditation (upasana) and that this meditation creates a special knowledge through which the Self is known:

Interlocutor: Ritually prescribed meditation (upasana) generates another knowledge, a special state of consciousness (vishishtam vijnana-antaram), and it is through this, and not merely through hearing scripture, that the Self is known.

Answer: This is wrong. The Vedic teaching, "the self is to be meditated upon" is not an injunction, and it does not enjoin this form of meditation, for nothing is to be done either inwardly or outwardly with regard to the Self. (Brhadaranyaka Upanishad Bhashya 1.4.7)

Again, it is apparent that Shankara admits that in some cases, hearing is sufficient. In the above, Shankara reiterates that knowledge is not something "to be practiced" and that nothing needs "to be done" with respect to it.

This line of interogation continues in the following tract. Shankara begins by rejecting the idea that yoga is necessary to first calm the mind:

There is no other way to silence the mind than knowledge of the self and its continuous remembrance (smrti)... And no effort in involved in this...

Interlocutor: Is not the continuous (samtana) remembrance (smrti) of cognition (vijnana) of the self something different from knowledge arising from hearing and hence something enjoined?

Answer: No; and the remembrance of the self arises spontaneously. (Brhadaranyaka Upanishad Bhashya 1.4.7)

Here, we encounter the idea that no effort is involved in knowledge of the Self. This idea is related to Shankara's suggestion that knowledge of the Self is like a form of perception. The idea is that no effort is required to see a jar that is before one, as long as the line of sight between oneself and the jar is not occluded. A more developed presentation of this idea is given in the Gita Bhashya:

The self is not something unknown to us at any time. It is not something to be acquired.... In this sense, for those who qualify, the devotion to knowledge (jnana-nishta) is easy... It is not for the knowledge of brahman that any effort (yatna) is required as something "to be done" (kartavya); it is only required for the cessation (nivrtti) of the false cognition of the Self, that is, of what the Self is not. (Gita Bhashya 18.55)

At this point, a consideration of what Shankara means by "contemplation" (nididhyasana) might be helpful. Generally, in English, the terms "meditation" and "contemplation" are used synonymously. And yet, the senses of the Latin terms "meditatio" and "contemplatio" are very different. Basically, the distinction between the two is the distinction between thinking and knowing, ratio and intellectus, or as the Greek has it, logos and nous, dianoia and episteme. In this sense, the terms "meditation" and "contemplation" make for useful translations of the Sanskrit terms "manana" and "nididhaysana." This distinction parallels Shankara's comments at Gita 3.42. There he distinguishes the functions of mind (manas) and intellect (buddhi). Mind (manas), he says, in involved in thinking (samkalpa/vikalpa), while intellect (buddhi) is concerned with ascertaining (nischaya). As noted above, Shankara asscociates contemplation (nididhyasana) with determination (nischaya).

That "nididhyasana" is actually a kind of knowing is also apparent in Shankara's own comments. The reference above to the "continuous (samtana) remembrance (smrti) of cognition (vijnana) of the self" (Brhad Up Bh) would appear to be a reference to contemplation (nididhyasana). This turn of phrase can be compared with the following two excerpts from his commentary on the Gita:

Meditation (dhyana) consists in a continuous (samtana) uninterrupted (avicchina) cognition (pratyaya), like a stream of flowing oil. (Gita Bhashya 13.24)

Devotion to knowledge (jnana-nistha) is intent application toward effecting the continuous (samtana) cognition (pratyaya) of the inner self (pratyag-atman). (Gita Bhashya 18.55)

Indeed, at times, Shankara omits the third "means" altogether and simply refers to the cognition of the Self in its place:

Through knowledge of brahman we become brahman, that is, through having heard (shrutva) from scripture (agama) and through a teacher (acharya), having considered (mantva) it through reason (tarka), and having cognized (vijnaya) it directly (sakshat). (Brhadaranyaka Upanishad Bhashya 2.5.15)

To close this subsection, I would like to make a general comment about the nature of inquiry and meditation in the soteriology of Shankara. In modern Western appropriations of Advaita Vedanta, there is, among some, a tendency to regard "spiritual inquiry" and "meditation" as involving a kind of special, "supra-mental" process that somehow transcends "mere thought" and "intellection." I would suggest that this attitude arises from the fact that in modern Western appropriations of Advaita Vedanta, inquiry and meditation are lifted from their original, classical Indian context, a context that anchored such practices in scriptural revelation. For Shankara, there is nothing special or "transcendental" about inquiry or meditation as such. What sets his form of inquiry and meditation apart is that they are guided by the Vedic revelation. As far as he is concerned, inquiry and meditation become "special," if you wish, when they are in accord with scripture. But in certain forms of modern spirituality, reliance upon scriptural revelation is seen as "dogmatic." In that case, however, something is will be missing, something that distinguishes "inquiry" and "meditation," something that sets it apart from mere worldly "ratiocination." It is for this reason, I would suggest, that "inquiry" and "meditation" have come to be seen as some kind "esoteric" cognitive process. My point here is simply that this kind of mentality is missing in the soteriology of Shankara. Thinking and contemplation are what they are. When they are in accord with revealed scripture, they are valid means to knowledge of the Self; and when they are not in accord, they are not valid.

V. The Role of Yoga

Generally, Shankara rejects the idea that the classical yoga of Patanjali plays a direct soteriological role in the final end of man:

Interlocutor: Several Upanishads have enjoined (vihita) yoga as a means to the realization of truth (samyag-darshana). The Brhadaranyaka Upanishad itself implies meditation when it says, "the self is to be heard, considered and meditated upon." And the Yoga-shastra says that yoga is the correct view (samyag-darshana).

Answer: No. The Yoga tradition is only partly true. And yoga as such is not a direct means to the highest state. Only the knowledge (vijnana) of the oneness of the self (aikatmatva) as revealed by the Vedas gives moksha. The Samkhya and Yoga, which are dualist (dvaitin), do not reveal the oneness of the self (atmaikatva). (Brahma Sutra Bhashya 2.1.3)

In his commentary on Brhadaranayaka Upanishad 5.5.1, Shankara speaks briefly of Patanjali's yoga:

Interlocutor: What about controlling the fluctuations of the mind-stuff (citta-vrtti-nirodha). Is it not enjoined?

Answer: No, and it is not a means to moksha. There are no other means to the attainment of the highest end than brahma-atma-jnana. (Brhadaranyaka Upanishad Bhashya 5.5.1)

On the other hand, with respect to Shankara's attitude toward yoga, there are two passages in the Brahma Sutra Bhashya that stand out as anomalies. In his comments on Brahma Sutra 3.2.24, Shankara says that in perfect concentration (pranidhana), certain yogins see (pashyanti) the Self, free from all plurality (prapancha) and they do so by means of absorption (dhyana) and devotion (bhakti). He then goes on to refer to those passages from the Katha Upanishad, Mundaka Upanishad and Mahabharata that speak of "seeing" the self while in meditation or through the purification of the mind. His commentary here parallels comments made at Brahma Sutra 3.2.5. There, Shankara says that occasionally, the supreme Lord (parameshvara) dispels the ignorance of those who meditate devotedly (abhidhyayate) on the Him and through his grace (prasada) these yogins are given extraordinary powers of "sight."

How are we to understand such passages? First, I think it is important to note that Shankara is, once again, constrained by the content of what he is commenting upon. This is to say that he is required to follow what the Brahma Sutra says here. But I think we can also understand what he says in light of his more attitude toward yoga as found in his commentary on the Gita.

In his comments on Gita 2.10, Shankara distinguishes karma-yoga and jnana-yoga. This parallels the distinction made by the Gita itself between Yoga and Samkhya. Here, "karma yoga" is used in a general sense to refer to any yoga of action. But in his comments at Gita 2.39. Shankara divides this yoga into karma-yoga proper and samadhi-yoga. In a similar manner, at Gita 6.2, Shankara distinguishes karma-yoga from dhyana-yoga. Now by "karma-yoga," we not mean the free giving of one's time to peel potatoes for the communal ashram, or charitably volunteering one's services at Mother Theresa's orphanage. Here, the term "karma" refers to the prescribed rites of brahmanism and "karma-yoga" means performing those rites while remaining detached (asanga) from their fruits (phala).

In his commentary on chapter 12 of the Gita, Shankara provides further distinctions. At 12.10-11 he distinguishes mere karma-yoga from karma-yoga practiced in conjunction with bhakti. And in his comments running from 12.6-9, he distinguishes mere dhyana-yoga from dhyana-yoga practiced in conjunction with bhakti. Thus, in his commentary on the Gita, Shankara provides a kind of hierarchy of yogas: karma-yoga; bhakti-karma-yoga; samadhi-yoga; and bhakti-samadhi-yoga.

This being the case, I take the comments at Brahma Sutra 3.2.24 and 3.2.5 as referring, in a conciliatory manner, to the practice of bhakti-dhyana yoga. Shankara acknowledges that in the case of some of those who practice this form of yoga, the Lord grants special powers of insight. But this yoga is still, properly speaking, only propaedeutic to jnana-yoga for Shankara. As we read on in his commentary, he makes it clear that such practices are still within the domain of duality. As he says in his comments at 3.2.6, the self is in fact not distinct from the Lord. And as he says at 3.2.25, there is, in truth, no one meditating and no one being meditated upon, as the Upanishads only really teach non-difference. In the end, then, though Shankara acknowledges the practice of yoga in this manner, it remains subordinate to the knowledge of the oneness of the self and brahman, which for Shankara, is the only true means to release.

For further discussion on the role of samadhi in Shankara's thought, see Micheal Comans' article "The Question of the Importance of Samadhi in Modern and Classical Advaita Vedanta"

This concludes the series of posts introductory to the forthcoming series on Neo-Vedanta.

Knowledge and Metaphysics in Ancient India: Some Themes

Introduction

This post will examine selected aspects of the problem of knowledge as it presented itself in ancient Indian thought, in particular, in the Upanishads and early Buddhism. It will begin by looking at various ways that the Upanishads arrived at their doctrines and will then move to the problem of how the self was known in the early and middle Upanishads. The problem of Vedic authority will then be taken up through an examination of the nature of the Buddhist critique of the Vedas as well as the brahmanic response to that critique. The post will then close with a critique of K.N. Jayatilleke's contention that early Buddhism constituted a form of "empiricism."

For the most part here, I follow Jayatilleke's excellent monograph, Early Buddhist Theory of Knowledge. Indeed, the post as a whole can be read as a reflection upon, and review of, Jayatilleke's book.

I. Metaphysics and Ways of Knowing in the Upanishads

The Upanishads are often associated with the earliest developments of Indian speculative thought. They are also associated with some of the earliest expressions of Indian mysticism. A general claim is sometimes made, by Neo-Vedantins and others, that the contents of the Vedas are based upon the mystical and illuminative "experiences" of the rishis who composed them. A version of this theory is the contention that the various metaphysical conceptions found in the Upanishads are based upon the "yogic experiences" of ancient rishis.

While there may indeed be a relationship between certain experiences and various conceptions found in the Upanishads, it would be difficult to maintain that the entire contents of the Upanishads are based upon the yogic experiences of the ancient rishis. This is not to say that this is a general claim of the Neo-Vedantins; Aurobindo, for one, made a point of distinguishing those parts of the Upanishads that are "experientially based" from those that are not. Nonetheless, in light of the above, a critical account of how specific conceptions in the Upanishads find their basis might be in order.

Before beginning, a review of the idea that metaphysical concepts, in general, are derivative of mystical experiences might be useful. The idea that the metaphysical content of a spiritual discipline is derived from the mystical experiences following from its practice is an idea most closely associated with what is sometimes called "mystical empiricism." Empiricism generally holds that any legitimate metaphysical conception must be empirically verifiable. It also contends that all our concepts ultimately derive from experience. Mutatis mutandis, mystical empiricism holds that the truth of a particular spiritual teaching can be "verified" by way of direct, transpersonal or mystical experience. Its contention is that various metaphysical truths -- as well as various metaphysical "structures," such as "higher planes of being," and so on -- were "discovered" by the founders of the different traditions of spirituality, and that these truths, and structures, can be personally "verified" by experimentally replicating the experiences of the founders via the spiritual techniques first used to discover them. The implication of this contention is that the metaphysics underpinning a spiritual discipline is based upon, and drawn from, the experiences of its founders. Recently, a similar notion has beed put forward by certain scholars of Yogachara and Vijnanavada thought, who have suggested that the contents of the philosophy of that school is, in part at least, derivative of meditative experience.

While such ideas are suggestive, they are also theoretically problematic. They are problematic as it is difficult to say which of the two, the metaphysic or the experience, is logically prior. Did an experience determine a particular metaphysic, or did the teaching, which implies a metaphysic, determine the experience? The two seem inextricably intertwined. In what follows will argue from the direction of the latter point of view and offer an alternate interpretation of what is actually achieved by the practice of a spiritual tradition.

Though the Upanishads relate various states of consciousness, such as dreaming, to certain metaphysical conceptions, there is no unanimity in the Upanishads as to how such states are to be conceived. We find, for example, more than one theory associated with the state of dreaming. According to one account, the self wanders around outside the body during dreaming. It is in the context of this idea that Brhad Up 4.3.14 says, "do not wake him too suddenly," lest his dream-self not find its way back into its body. But according to another idea the self remains inside the body while dreaming (Brhad Up 2.1.8). This is the interpretation favored by Shankara. Here, it is apparent that we have two very different conceptions of dreaming. Now if these conceptions of dreaming were derived entirely from the experience of dreaming itself, we would not find different interpretations of what happens when one dreams. This means that account of dreaming found in the Upanishads also derives from speculation and that it is not based solely on the "raw" experience of dreaming.

A parallel example can be found in the later yogic tradition. In the kundalini yoga of the Naths, the yogic process, and the experience associated with it, is said to occur "within" the body. But in the shabda yoga of the northern Sants, the yogic process, and the experience associated with it, is thought to occur "outside" the body. Here, it is apparent that how a teaching initially conceives of yogic experience is determinative of how that experience is later interpreted. This is to say that the metaphysical conceptions underpinning a particular teaching will, to a significant extent, pre-determine the experiences that follow from that teaching.

Given this point -- that a pre-conceived metaphysics is already at work interpreting "yogic" experience even while it is occurring -- there is certain circularity to the "verification" theory of the mystical empiricists. According to mystical empiricism, spiritual truths, and various metaphysical structures, are "re-discovered" or "verified" by practitioners of a spiritual tradition through its practice. But accepting a particular practice will mean adopting the specific form of that practice; it will mean accepting the teaching that informs it, along with the metaphysical edifice underpinning that teaching. It is no accident, then, that these "metaphysical structures" are "re-discovered" by practitioners. For what is practice (abhyasa; bhavava) but the inculcation of a teaching to the point where one comes to understand one's experience in terms of it, where one begins to see reality in accord with that metaphysic? We will return to this point below.

One experience that appears to have been, more or less, directly related to certain metaphysical conceptions is the experience of ecstasy (i.e., the so-called "out of body experience"). Since the time of the Rg Veda, the experience of ecstasy has been related to the idea of the separability of the personality or "soul" (anu; manas; atma) from the body. The Keshi-sukta section of the Rg Veda (10.136), for example, describes the ecstatic experiences of the keshins, long-haired munis, who appear to drink some sort of psychotropic compound (visha) and fly through the air with the winds, "looking down on those below." More significantly, Rg Veda 10.59 describes how the soul (manas) of someone thought to be dead has "wandered" some distance in the nether worlds before returning to its body. Later, the Katha Upanishad speaks of the practice of separating the inner self (antar-atma) from the body just as the core from a stalk of grass is pulled from its sheath (2.3.17). Even the Buddhists speculate about the ability to separate the "mental body" from the physical body (mano-maya-rddhi).

While there appears to be a relationship between the experience of "bodily transcendence" and certain metaphysical conceptions, such as the separability of the soul from the body, a one-way relationship between the two cannot be established, for the same reason that the experience of dreaming does not necessarily imply the idea that the soul wanders around outside the body during sleep. There are, in other words, other ways of intepreting the experience of ecstasy. Again, this is because there is a degree of speculation involved in understanding of the experience of ecstasy. Thus, descriptions of the ecstatic experience cannot be taken as accounts of some "raw" experience.

In his comments on Rg Veda 10.59, A. MacDonnell suggests a relationship between the experience of ecstasy and ideas concerning the continuation of the soul after the death of the body. Eric Frauwallner and K.N. Jayatilleke have also both suggested that meditative experience (jhana) may have contributed to speculations concerning the existence of a "mental-body" (mano-maya-kaya). The Vedic tradition itself appears to have attempted to use the experience of ecstasy as grounds for positing the existence of the disincarnate soul. That they had attempted to do so is evident from the remarks of the materialists. Interestingly, the materialists take the very argument, turn it around, and use it against the Vedic tradition. The materialists argue that the disincarnate soul does not exist precisely because it cannot be separated from the body the way we can separate a stalk of grass from its sheath. In other words, we do not actually see a soul rising from the bodies of people who have recently died.

While it would appear, then, that the experience of ecstasy was related to conceptions concerning the existence of the soul, I think it is worth pointing out that speculation about the afterlife, the immortality of the soul, the existence of heaven, and so on, had as much, if not more, to do with wonder about, and desire for, life after death than it did with trying to simply understand the experience of ecstasy. In other words, the attempt to base the existence of the soul on the experience of ecstasy would appear to be an ad hoc attempt to rationalize and justify the desire for life after death. The Buddhists, who are critical of just such desire, appear to have recognized this; the Potthapada Sutta points out that we should not be lead astray by the language of the Buddha when he speaks of the mental body (mano-maya) as some sort of "constructed self" (atta-patilabha). In other words, the mental body is not to be taken as an actual self; this is indicated by the fact that it "comes and goes." Much the same can be said for later conceptions concerning the existence of the "subtle body" (sukshma-sharira; linga-sharira). The postulation of a "subtle body" had more to do with the explanatory need for a vehicle for the karmic "seeds" (bija) that remain after the death of the physical body than it did with the need to explain or understand a particular "meditative experience."

Another domain where mystical experiences would seem to be determinative is soteriology. While we cannot rule out the possibility that mystical experience was involved in conceptions of liberating insight or enlightenment (avabodha), it is important to note that the eschatological and soteriological conceptions in the Upanishads are also essentially related to certain patterns of thought. While the early Upanishads do contain content relating to the older brahmanic ritualism, they generally offer a significant critique of the personal eschatology associated with the sacrificial cult and its ritualism. At several points in the Upanishads (Brhad Up 1.4.15; 3.8.10; 4.4.6; Chan Up 5.10.5; 8.1.6; Mundaka Up 1.2.9) it is stated that when the karmic results of the rite are "used up," the beneficiary of the rite leaves the heaven-realms (svarga) and returns back to the world of men. In other words, there is a growing recognition in the Upanishads that the fruits of ritual action (karma) are only temporary. The logic behind this kind of thinking is as follows: whatever comes into being, goes out of being; the effects (karya) of ritual action (karma) come into being; therefore, they will, in time, go out of being. This same logic underpins the soteriology of Advaita Vedanta, with its notion that moksha is, in reality, "always already" established (nitya; siddha), and it is this same reasoning that Shankara uses when he rejects the idea that liberation is the causal effect of some means (sadhana). Moksha must be permanent; therefore, it cannot be the result of an action. Here, rather than an experience, it is a certain kind of inferential thinking that is determinative of how moksha is to be conceived.

Other metaphysical conceptions in the Upanishads can be shown to be related to basic observation and inference. For example, discernable in the Upanishads are certain clusters of ideas forming rudimentary "theories" about the mysterious processes of life and death. These theories are each based upon one of the elements of life: "fire" (tejas); breath (prana); and water (apah). That these theories derive in large part from simple observation and inference is not difficult to see. The monsoons bring new life to plants; vegetation dies during periods of drought; living bodies are warm and respiring; corpses are cold and breathless. This act of observation is related to the term "darshana" in the oldest Upanishads, and "observation" is one of the earliest senses of this term. We also find the term used in the later classical tradition with a similar sense. In philosophical writing, darshana and its cognates mean, "it can be seen that..." or "we see that...." This "seeing" acts as a substitute for other senses, like the sense of touch. We might say that the term "darshana" refers to perception in general. Accordingly, Chandogya 3.13.8 says, "This is the perception (drshti) of that, when one cognizes (vijanati) the warmth (ushniman) in the body (sharira) by touch (samsparsha)."

In the Brhararanyaka Upanishad, knowledge of the self is related to this "seeing." There we read that the self is "to be seen" (drashtavyam), as well as to be heard, reasoned about, and contemplated (Brhad 2.4.5). The later tradition takes this four-fold group as referring to a sequential process through which truth is gradually revealed. But there is no reason to assume that the four were originally meant to reflect a progressive unfolding of knowledge and understanding. What these four probably originally referred to were four different, though inter-related, manners of knowing.

The early Upanishads did, however, distinguish knowledge of the self from other forms of knowledge. Chandogya Upanishad 7.1.2-3 reads: "O sir, I have studied the Rg Veda, the Yajur Veda, the Sama Veda and the Atharva Veda, the Epics and Puranas, grammar, ritual, mathematics, meteorology, mineralogy, debate, etiquette, etymology, ritual science, the elements, archery, astronomy, herpetology, and the fine arts. O sir, I know all these. I am a knower of these mantras and sciences (mantra-vid). But I am not a knower of the self (atma-vid)."

In the Taittiriya Upanishad, which is later than the Brhad Up and Chandogya Up, we read that study (svadhyaya) and discussion (pravacana) are to be practised (Tait Up 1.9.1). This can be taken as referring to the practices of 'hearing' (shravana) and 'thinking' (manana) referred to above. It would appear, then, that at this point in the tradition, listening, discussion, recitation, and debate were important aspects of the teaching of the Upanishads. Accordingly, some etymologists take the term "upanishad" to mean, "come, sit near (and listen to this teaching)."

In the Katha and Mundaka Upanishads, which are later still, we find the beginnings of a more critical gnoseology. Katha Up 1.2.7 says, "The self cannot be attained through mere hearing (shravana)." The next verse, 1.2.8, reads, "This self is not known when spoken of (prokta) by people... (and) it is beyond reasoning (tarka)." Apparently referring to the Taittiriya Up, Katha Up 1.2.23 says, "The self cannot be grasped through discussion (pravacana), through the intellect (medhyaya), or by repeated hearings (bahuna shrutena)." And Katha Up 2.3.12 says, "The self cannot be obtained through speech (vacas), through the mind (manas), nor by the eyes (cakshus)." Such passages can be read as a critique of the practices of "seeing," "hearing," and "reasoning" referred to above.

However, according to Katha 1.3.12, the self can be seen (drshyate) with the subtle intellect (sukshmaya buddhya) by those seers (darshi) who are able to perceive the subtle (sukshma). In the Mundaka Upanishad this idea is clarified. At 3.1.8-9 we read that the self is seen (pashyate) while meditating (dhyanamanah) through the clarification of knowledge (jnana-prasadena), or by the mind (citta) that has been purified (vishuddha). The Gita contains similar expressions. Gita 13.24 also says that one can see (pashyati) the self by way of meditation (dhyana).

It would appear then that by the time of the Katha Upanishad, yogic meditation entered the picture as a new means of knowledge in the Indian tradition. It would also appear that this tradition of mystical vision begins to understand itself in contradistinction to the older tradition of inculcation through recitation, listening, discussion and debate. Significantly, we find mention of two kinds of knowledge in the Mundaka Upanishad: a "lower knowledge" (apara-vidya) and a "higher knowledge (para-vidya). Here, the four Vedas and various sciences mentioned above in Chandogya Up 7.1.2-3 are explicitly referred to as the "lower knowledge" (Mundaka Up 1.1.5) and the "higher knowledge" is related to that which is "unseen" (adrshya), "ungraspable" (agrahya), "subtle" (sushkshma), and "without eyes or ears" (Mundaka Up 1.1.6). It is this "higher knowledge" that is said to be the means to attaining the "indestructible" (akshara).

II. Authority and Knowledge in the Brahmanic Tradition and Philosophical Darshanas

When the Vedic tradition was the only tradition practiced in India, there was little questioning of its authority, except perhaps among the materialists. But with the rise of heterodox traditions such as Buddhism and Jainism, its authority became a matter of great interest. According to the early Buddhist accounts, the brahminic tradition claimed that the Vedas were "composed" by the rishis. We do find occasional references to the Vedas being the "utterances (vakya) of the rishis" that support this contention. But it is difficult to determine whether or not the brahmins of the Buddha's time actually made this claim or whether this is merely a polemical construction on the part of the Buddhists.

In the Rg Veda, we find the idea that the Vedas are produced by the sacrifice of the cosmic person, Purusha (Rg Veda 10.90). In the later Brahmanas, the Vedas are said to arise from the divine Word, Vak (Taittiriya Brahmana 2.8.8.5). Elsewhere in the Brahmanas, Prajapati, the creator god, is said to be the source of the Vedas (Taittiriya Brahmana 3.3.2.1; Shatapata Brahmana 6.1.1.8). At times, Prajapati is identified with Brahma, and Brahma is related to the three Vedas. In the Upanishads, Brahma is said to have taught the Vedic tradition to Prajapati (Chandogya Up 8.15.1) and Prajapati is said to have taught the gods (Brhadaranyaka Up 5.2.1) and Manu (Chandogya Up 8.15.1). If Brahma is understood to be the creator and source of the Vedas, this can be taken as a reference to a personal (pauruseya) if superhuman origin for the Vedas.

In the later brahmanic commentarial tradition we find a more novel theory as to the origin of the Vedas. According to Yaska (5th century BCE), the rishis directly intuit (sakshat-krta) the Vedic dharma (Nirukta 1.20). In a similar way, the Vedartha-prakasha, a commentary on the Taittiriya Samhita, states that the Vedas are "seen" (drshta) by the rishis through "super-sensory perception" (ati-indriya pratyaksha). Here, it would appear that the brahmanic tradition is beginning to be concerned with establishing the authority of the Vedas by relating their origin with the most direct means of knowledge: perception.

In the Indian philosophical systems (darshana), we find a further development of the idea that the rishis directly intuit the Vedas through super-sensory perception. In his commentary on the Nyaya Sutras, Vatsyayana states that the rishis are "reliable witnesses" (apta) and legitimate sources of knowledge (1.1.8). The reference to the idea of a "reliable witness," or authoritative expert, is a reference to the epistemic theory that in some cases "word of mouth" (shabda), that is, testimony from a reliable source, is a legitimate source of knowledge (pramanya). The case of the rishis, here, is interesting. According to Vatsyayana, the rishis are reliable because they have directly intuited (sakshat-krta) the Vedic dharma (Nyaya Sutra Bhashya 2.1.68), even though the Vedas are said to belong to the domain of non-empirical facts (adrshtha). This is clearly an adaptation of Yaska's idea recast in terms of the epistemological categories of classical philosophy.

The Samkhya tradition also makes use of the idea of the seer's super-sensory ability, but applies it in another manner. It claims that the insights of its founder Kapila, whom it regards as a primordial knower (adi-vidvan) and great sage (maha-muni), are authoritative. The implication appears to be that great sages like Kapila have the same kind of authority as the rishis of old. This theory is referred to, and rejected, by Shankara in his account of the Samkhya's claim to authority.

Another later Naiyayika philosopher, Jayanta Bhatta, makes some rather interesting comments about the nature of immediate, pre-predicative perception (nirvikalpaka-pratyaksha). Some of the Mahayana Buddhists and Advaita Vedantins of his day had made claims about nirvikalpaka perception as a means of apprehending absolute reality or "things as such." Jayanta notices that even though the Buddhists and Advaitins both lay claim to this unmediated perception, they differ as to what it is that "direct experience" apprehends. Jayanta asks, "Is it pure being as such (sat), as Mandana Mishra claims? Or is it the thing in itself (svalakshana) as the Mahayanists claim?" What he's getting at here is that though both the Advaitins and the Buddhists had a doctrine of nirvikalpaka pratyaksha, they held diametrically opposed views as to what it is that pre-predicative perception apprehends. For Mandana, it is pure being as such where being is the supreme, all-pervasive universal (samanya). For Dharmakirti, however, it is the thing in itself, which for the later Buddhists is the concrete particular shorn of all qualities. Jayanta remarks that, at first glance, it may seem that "direct perception" offers a means of resolving the problems of metaphysics. But, he asks, who is it that determines which awareness is most direct and truly immediate? He concludes his remarks on the topic by commenting that even the "greatness (mahatmya) of perception (pratyaksha)" has its limits.

The Mimamsakas, who inherit the brahmanic tradition, reject in its entirety the notion that the Vedas are "composed" by any being, human or divine. For the orthodox Mimamsakas, the Veda is "authorless" (apaurushya), and by this they mean that it has no reference to any agency, human or super-human or divine. They also reject the contention that the Vedic dharma is "intuited" by the rishis through some kind of super-sensory perception. For the Mimamsakas, the Veda exists eternally as the body of timeless knowledge and transcendental truth. Kumarila, perhaps the most sophisticated of the Mimamsa exegetes and philosophers, rejects the contention that "super-sensory" (ati-indriya) knowledge, or "yogic perception" (yogi-pratyaksha) as he calls it, is authoritative. Being the traditionalist that he is, Shankara follows the Mimamsa acharya Kumarila almost entirely on these points. We will look at Shankara's position on such matters in another post.

Kumarila also subjects the dharmashastra theory, concerning the various means of deriving the dharma, to critical scrutiny. The dharmashastra literature gives four sources from which the dharma can be derived. The dharma can be learned from revealed scripure (shruti); it can be learned from transmitted scriptures and authoritative secondary literature (smrti); it can be learned from the conduct (achara) of "virtuous individuals" (sadhu); and it can be obtained via "personal conviction" (atma-tushti). But for Kumarila, the conduct of "virtuous individuals" can only refer to the conduct of those who are already learned (shishta) and "well-cultured" (samskrta) in the dharma. For Kumarila, the authority of the "sadhu" cannot in any way compete with the Veda; the conduct (achara) of the good (sadhu) can only mean conduct that conforms with the Vedic norms. The same goes for the so-called fourth source of the dharma, "personal conviction" (atma-tushti), also referred to as the "inner voice of conscience" (hrdaya-koshana). In Kumarila's view, "personal conviction" (atma-tushti) can only mean the "inner voice" of those who have thoroughly internalized the Vedic dharma, to those who "know the Veda" (veda-vid). Here, the Veda itself is the ultimate source of dharma (veda-mulatva), and the other means are but derivative from it.

Later commentators on the Nirukta place similar limitations on Yaska's theory that the Vedas are "intuited" by the rishis. In his commentary on the Nirukta, Durga insists that the term for "seeing" used by Yaska (sakshat-krta) must be taken as fiugurative; the dharma cannot actually be "seen" in this manner. Likewise, in the later dharmashastra literature, limitations are imposed on the powers of "extraordinary individuals" -- the seers, sages, and siddhas. In his writings, Medhatithi comments that deriving the dharma from the authority "super-human" (purusha-atishaya) individuals is a view belonging to the non-Vedic "outsiders" (bahya), such as the Buddhists and Tantrikas. Besides, he adds, seers and sages come and go, but the Veda is eternal.

III. Authority and Knowledge in Early Buddhism

As is well known, the early Buddhist tradition was critical of the authority of both the Vedas and the Vedic tradition. The term the early Buddhists use for Vedic traditionalism is "anussavika," which appears to mean "listening repeatedly." This may be a reference to what the Katha Upanishad means when it refers to those who attempt to apprehend the self "by repeated hearings" (bahuna shrutena).

According to K.N. Jayatilleke, the Buddhist critique was based on the idea that both the brahmins and the Vedic poets had no personal knowledge of that which they talked about. The Buddhists contended that the brahmins, their teachers, and their teacher's teachers, all the way back to the Vedic seers, are not "reliable witnesses" (apta) and that their testimony cannot be trusted. In the Canki Sutta, for example, the Buddha says of some brahmins: "They do not know, and do not see what their scriptures (mantapadam) speak of. They do not say, 'I know this, I see (passami) this'." The Buddha also criticizes the dharma of the brahmins on the grounds that neither the brahmins, nor their teachers, nor the original Vedic seers, claim to have the super-sensory knowledge (abhinna) of their dharma. In the Tevijja Sutta, we also find the implied charge that none of the teachers of the Vedic tradition have direct knowledge of Brahma (Digha Nikaya 1.238): "they have not seen Brahma face to face (brahma sakhidittho)."

Given that this was the nature of their challenge to the authority of the Vedic tradition, we are now in a position to understand why the later brahmanic tradition posited the theory that the rishis intuited or "saw" the Veda through super-sensory vision or direct perception. It is clearly an ad hoc response to the Buddhist critique that is retrospectively read back into the tradition.

In contrast to the Vedic tradition, the Buddha claims that he does not teach his dharma on the basis of what he has heard from another ascetic or brahmin, but rather from the authority of his own knowledge: "It is from what I have seen myself, what I myself have realized, what I know personally, that I speak." (yad eva me samam natam samam dittham samam viditam tam evaham vadami) (Majjhima Nikaya 3.186) The operative term here is "samam" which is a reflexive term like "myself." Thus, the distinction being drawn here is between what one has "seen for oneself" and what one has "heard from tradition" (samam dittho va hoti anussavasuto) (Majjhima Nikaya 1.465).

The Buddha also occasionally exhorts others to follow him and his teaching (dharma): "Let an intelligent person come to me, sincere, honest, and straightforward; I will instruct him and teach the doctrine so that on my instructions he would conduct himself in such a way that before long he would himself know and see for himself." (Majjhima Nikaya 2.44) Thus, the contrast between the two traditions is such that the Buddhist truth is "personally known in this life (sacchikato sayam)" (Theragata 1.331), while the Vedic tradition is not "personally realized by oneself" (samam sayam abhinnatam attapaccakkhadhamman).

IV. Jayatilleke on Early Buddhist Theory of Knowledge

Jayatilleke compares this claim, that the Buddhist dharma is to be "personally known," to the principle of verifiablity put forward by empiricists and positivists. In Early Buddhist Theory of Knowledge, he writes:

The emphasis that 'knowing' (janam) must be based on 'seeing' (passam) or direct experience, makes Buddhism a form of Empiricism. We have, however, to modify the term to mean not only that all our knowledge is derived from sense-experience but from super-sensory experience as well.... Early Buddhism should therefore be regarded not as a system of metaphysics but as a verifiable hypothesis discovered by the Buddha in the course of his 'trial' and 'error' experimentation with different ways of life. (Early Buddhist Theory of Knowledge, pp. 463-464)

At the same time, Jayatilleke wishes to distinguish the "empiricism" of early Buddhism from mysticism. He begins by distinguishing the Upanishadic gnosis from Buddhist wisdom. The Upanishadic gnosis, he says, is due to the grace of God, while in Buddhism, mental concentration (samadhi) is the causal factor (upanisa) in the rise of knowledge (p. 420). Discussing the nature of abhinna, he continues:

Buddhism does not make the claim of the mystic that this knowledge was derived from a supernatural source... but that it is a product of the natural development of the mind.... It would be wrong to call this mystical or intuitive knowledge in the context of Buddhism.... We shall therefore refer to this kind of knowledge as 'extrasensory perception' in the Buddhist context. (Ibid., p. 426)

Apparently, when describing this early Buddhist "naturalism," it is not correct to characterize it as a form of mysticism though it is appropriate to speak of it in terms of ESP. But Jayatilleke is not consistent on this point. Given some of his other characterizations of early Buddhism, it seems difficult not to admit that what he is describing is indeed some form of mysticism. In the closing pages of Early Buddhist Theory of Knowledge, he writes:

The transempirical cannot be empirically described or understood but it can be realized and attained.... It is not that there was something that the Buddha did not know, but that what he 'knew' in the transcendent sense could not be conveyed in words because of the limitations of language and empiricism. (Ibid., p. 476)

If this is not a description of some form of mysticism, it is difficult to imagine what might count as mysticism. I am not suggesting that early Buddhism needs to be understood as a form of mysticism. My point here is simply that Jayatilleke's account of early Buddhism ends up sounding very much like a description of mysticism.

In order to bolster his claim that early Buddhism is a form of "empiricism" in which practitioners can personally "verify" the Buddha-dharma, Jayatilleke must put in check the element of faith (shraddha) that the Buddhist tradition speaks of. He locates a passage from a later philosophical work in support of his contention that faith is not essential to Buddhism:

Just as a learned one (pandita) tests gold by burning it and rubbing it with a touch-stone, so should my statements be accepted after examination (pariksha) and not out of respect for the guru (gaurutva). (Tattvasamgraha 3588)

This passage, from Shantirakshita's Tattvasamgraha, is not found in the Pali canon, and it is not clear if it is from some text that has been lost or if it is a construct of the later tradition. In any case, according to Jayatilleke, it "reflects the attitude of the Buddha" and he takes it to mean that the practice of Buddhism is not based upon faith in the Buddha. This is an odd claim given that, as Jayatilleke himself notes, the Buddhist critique of the Vedic tradition had hinged on the matter of who counts as a "reliable witness" (apta) and on whose teaching can be trusted.

Jayatilleke also interprets the Pali expression "safeguarding the truth" (saccanurakkhana) along positivistic lines. He suggests that it means that the Buddha-dharma should only be provisionally accepted by the practitioner until such time as its truth may be personally verified. The analogy here is with the scientific method of assuming a hypothesis until its veracity can either be confirmed or discounted. As for those passages from the early Buddhist writings that are more explicitly dogmatic, Jayatilleke says that they "emerged only in the latest stages of the Pali Canon" (p. 401).

In an analogous manner, Jayatilleke interprets the parable of the arrow as the expression of a kind of Buddhist Pragmatism:

...the gist of it is that a man struck with a poisoned arrow should be concerned with removing the arrow and getting well rather than be interested in purely theoretical questions (about the nature of the arrow, who shot it, etc.) which have no practical utility. The moral is that man should only be interested in truths which have a practical bearing in life. (Ibid., p. 357)

This interpretation jibes well with Jayatilleke's positivist reading of the Buddhist rejection of the various "views" (drshti): just as positivism regards metaphysical questions as "meaningless" insofar as they cannot be empirically verified, so too Buddhism dismisses the various metaphysical theories or "views" (drshti) as pragmatically irrelevant.

Admittedly, there are elements of Jayatilleke's reading of early Buddhism that ring true. For one, he is correct to suggest that early Buddhism did not advocate some form of agnosticism, as we often find the rejection of this position in the Buddhist writings. But other aspects of his interpretation are intermingled with problematic elements. For example, the various metaphysical theories (drshti) may indeed have been rejected by the early Buddhists as soteriologically irrelevant. But Buddhism also presents itself as the correct view (samyag-darshana), and that claim would seem to entail that it held to its own share of metaphysical presuppositions. So it is unlikely that the various views were rejected purely on the grounds that they are merely "theoretical" or "metaphysical."

It may be that the metaphysical theories of the rival teachings were understood as irrelevant simply because they were seen as incapable of bringing release (nirvana). As the Vacchhagotta Sutta says, the views are to be rejected because they are not efficacious in that they are not conducive to nirvana. This point is developed by Richard Robinson in his article, "The Unexplained Points." Robinson argues that early Buddhism understood its teaching as a kind of metaphysical gnosis (vidya) like other teachings of its time; the difference was that it understood its teaching as the only teaching conducive to nirvana. Thus, the other views are rejected not because they are "metaphysical," but because their gnosis does not lead to freedom from rebirth. As Robinson puts it, "It is not a question of metaphysics versus pragmatic wisdom, but rather one of which metaphysics is most efficacious in attaining an existential objective." Robinson continues:

Przyluski drew attention to the deep significance of the terms derived from upa-as: "Upas, upasana, upasaka bears witness to a fund of common ideas of an ancient teaching in which deliverance is the fruit of the act" (J. Przyluski and E. Lamotte, "Bouddhisme et Upanisad"). In primitive Buddhism, upasaka seems to have meant a devotee or adherent either in the householder life or in the homeless life. This is why Vacchagotta... can be said to have become an upasaka.... His becoming an upasaka involved commitment to a teaching he accepted. His goal... was arhatship, and eventually he attained it by a process of combined moral and mental cultivation, which is essentially the Upanisadic upasana, except that it concentrated on a different gnosis -- the fourfold truth and the twelvefold dependent coarising, rather than the identity of Brahman. ("The Unexplained Points.")

Robinson's article is as much a paper on early Buddhism as it is a methodological reflection on various approaches to early Buddhism. The value of his presentation lies in the fact that he attempts to understand early Buddhism on its own terms, or at the very least, in terms of how the various movements at that time in India understood themselves. Much of Jayatilleke's interpretation, on the other hand, comes off as a projection of modern concerns onto the blue-screen of ancient India. This makes his account contentious on several points. For one, it is highly unlikely that early Buddhism advocated the kind of systematic doubt and openness to alternative hypotheses that characterize modern scientific method. Like other soteriological teachings, Buddhism speaks of the insidious nature of doubt (samshaya) and lays stress on the importance of faith (sraddha) in the teaching.

Let us look again at a passage that Jayatilleke himself quotes:

Let an intelligent person come to me, sincere, honest, and straightforward; I will instruct him and teach the doctrine so that on my instructions he would conduct himself in such a way that before long he would himself know and see for himself. Majjhima Nikaya 2.44

As I read the passage, the Buddha is saying, "Let me teach this sincere, intelligent person the dharma; and let that person then conduct himself faithfully and earnestly in accordance with the dharma; and then let that person see if the dharma makes a difference in his life." While it may be tempting to interpret this point in accordance with James' Pragmatic Rule, I would suggest that it might more favourably be seen in terms of existential commitment.

In this case, "conducting himself in such a way..." does not refer to the implementation of some kind of experimental method; and "seeing for oneself" will not mean the empirical verification of some fact. "Conducting oneself in such a way that one may see the truth of the dharma" means inculcating the dharma into one's life, into one's very existence, to the point where one begins to see oneself and the world in a new, more meaningful and more liberating way. It is, in other words, to understand the world and our place in it in terms of the Buddha-dharma. Accepting the Buddha-dharma as only "provisionally true," as Jayatilleke suggests, will, quite simply, not achieve this end, because such an approach will not involve the kind of existential commitment that is required. Rather, understanding oneself and one's place in the world in terms of the Buddha-dharma will only be effected through faith in that very dharma.