Wynne - The ātman and its negation (1984)

The ātman and its negation - A conceptual and chronological analysis of early Buddhist thought

Journal of the International Association of Buddhist Studies

Volume 33 Number 1-2 2010 (2011)

Alexander Wynne

JIABS Vol 33 pp. 103-171

The denial that a human being possesses a “self” or “soul” is probably the most famous Buddhist teaching. It is certainly its most distinct, as has been pointed out by G. P. Malalasekera: “In its denial of any real permanent Soul or Self, Buddhism stands alone.” ¹ A similar modern Sinhalese perspective has been expressed by Walpola Rahula: “Buddhism stands unique in the history of human thought in denying the existence of such a Soul, Self or Ātman.”² The “No Self” or “no soul” doctrine (Sanskrit: anātman; Pāli: anattan) is particularly notable for its widespread acceptance and historical endurance. It was a standard belief of virtually all the ancient schools of Indian Buddhism (the notable exception being the Pudgalavādins), ³ and has persisted without change into the modern era. Thus the classical Theravādin view of Buddhaghosa that “there is only suffering, but nobody who suffers”4 is identical to the view of Śāntideva, the famous Indian Mahāyānist, that “the person who experiences suffering does not exist,“5 and both views are mirrored by the modern Theravādin perspective of Mahasi Sayadaw that “there is no person or soul” and the modern Mahāyāna view of the fourteenth Dalai Lama that ” [t] he Buddha taught that … our belief in an independent self is the root cause of all suffering.” ⁴

This traditional understanding has been accepted by Buddhist scholars past and present. According to La Vallée Poussin, the Buddha did not accept “the existence of a Self (ātman), a permanent individual; he teaches that the so-called Self is a compound of material and spiritual data called skandhas.” ⁵ In a similar vein Norman has stated that “the Buddha denied the existence of the permanent individual self,” Collins has spoken of the Buddhist “denial of self,” ⁶ and De Jong has noted that “in early Buddhism impermanence and suffering imply the non-existence of the self as a permanent entity.” ⁷ There is early canonical support for all these statements. In the Vajirā Sutta of the Saṃyutta-Nikāya, where the bhikkhunī Vajirā reports the doctrine to Māra as follows:

Why do you believe in a living being?
Is not this your view, Māra?
This is nothing but a heap of formations:
No being is found here. (553)
When there is a collection of parts
the word ‘chariot’ is used;
In the same way, when the aggregates exist (khandhesu santesu)
the conventional term ‘being’ (satto) [is applied to them]. (554)
Only suffering (dukkham eva) comes into existence, and only suffering endures.
Nothing apart from suffering comes into existence, and nothing apart from suffering ceases to exist. (555) ⁸

The statement that “only suffering comes into existence, and only suffering endures” is akin to Buddhaghosa’s statement that “there is only suffering, but nobody who suffers,” and Śāntideva’s statement that “the person who experiences suffering does not exist:” all assume that there is no “ghost” in the machine. Such an understanding has been summed up by T. W. Rhys Davids as follows: “Man is never the same for two consecutive moments, and there is within him no abiding principle whatever.” ⁹ According to this definition, and depending on one’s perspective, the doctrine could be taken to mean either that a person has no “soul” (in the sense of a “spiritual part of man in contrast to the purely physical”), or that a person lacks an inherent identity, that which could be termed “self” (in its simple philosophical sense of “that which a person is really and intrisically he (in contradistinction to what is adventitious).” For these definitions see the OED). Given the close correspondence between the Vajirā Sutta, Buddhaghosa, Śāntideva and the more recent Buddhist authorities cited above, this understanding would seem to have been the norm in Buddhist circles for over two thousand years. Indeed the attestation of this idea in a canonical text means that it can most probably be taken back to the pre-Aśokan period, i.e. within roughly 150 years of the Buddha’s death. ¹⁰

1. The historical problem

Despite its importance and historical endurance, it is odd that the No Self doctrine is hardly attested in the early Buddhist literature, the Vajirā Sutta being perhaps the only Pāli discourse to state the idea explicitly. ¹¹ Indeed it is very easy to read a substantial amount of the early literature - such as virtually all of the important DīghaNikāya - without encountering anything remotely like it. There are only two plausible explanations for this historical peculiarity. The first is that the doctrine is implicit in the early texts, but for some reason was only explicated in the Vajirā Sutta (and perhaps a few other discourses). The second is that the doctrine was generally unknown to the composers of the canonical texts - either because it emerged at a later date or because it was initially a fringe idea - who therefore failed to record it. Both explanations imply that Buddhist thought changed over time: from implicit to explicit formulations of the doctrine, or so that a later development or minority concern eventually came to dominate the philosophical mainstream of the early saingha. In other words, there must have been either a terminological or philosophical change in early Buddhist thinking about the human being.

To establish the more likely eventuality, the early Buddhist teachings on personal identity must be reconsidered. This will involve going over much old ground, but since a general consensus has not been reached this is unavoidable. ¹² Such a study is further necessitated by the fact that a number of important text-critical problems have been ignored: the historical problem noted above has not been properly recognised, little thought has been given to the form of the important Not-Self teaching, the basic vocabulary of the teachings on personal identity has been misunderstood, nonBuddhist parallels to important teachings have been missed, and little attempt has been made to relate the teachings on personal identity to the wider doctrinal concerns of the early texts. In short, there is much scope for a more detailed exploration of this aspect of early Buddhist thought.

Such a study must begin with a very old and much discussed point: the Upaniṣadic background to early Buddhism. For as we will see, early Upaniṣadic speculation on the ätman was wellknown in early Buddhist circles, and determines the form and content of some important early Buddhist teachings.

Of the various senses in which the term ātman is used in the early Upaniṣads, the most important is the “spiritual self or the inmost core of a human being.” ¹³ According to the Yājñavalkyakānda of the Bṛhadāranyaka Upaniṣad, the ātman in this sense is a non-physical substance, ¹⁴ a principle of life on which a person’s cognitive functions (prāna) depend, ¹⁵ and also the inner subject of perception, i.e. an “unseen seer” ¹⁶ that consists of nothing but consciousness. ¹⁷ The ātman of Yājñavalkya, and henceforth of Vedāntic philosophy in general, is both a spiritual substance and an unchanging inner subject of phenomena. This understanding is not entirely different from the Cartesian “mind-soul,” which is also a spiritual substance as well as the true subject of experience. ¹⁸ Despite this similarity, however, the ātman of Yājñavalkya is also said to be a nondual consciousness identical to the underlying essence of the cosmos (brahman), the realisation of which is a state of pure bliss. ¹⁹ This equation of microcosm (ātman) and macrocosm (brahman) is of course philosophically problematic, since it involves the identification of the individual subject of perception with an impersonal essence. In the early Upanisads, however, this problem is resolved “mystically” rather than philosophically: as the Bṛhadāranyaka Upaniṣad states, by meditating on the ātman (when it is “seen, heard, contemplated and cognised …”), its true nature as a macrocosmic essence will be revealed (”… the whole world is known”). ²⁰ Such an understanding belongs more to the realm of religious experience than that of rational enquiry. The peculiar identity of microcosm and macrocosm was not initially problematic, therefore, but only became so for later generations of Vedāntic thinkers who puzzled it over to varying degrees of success.

It is against this conceptual background that the early Buddhist teachings on personal identity must be understood. As will become clear, these teachings refer to the Upaniṣadic ātman in both its microcosmic and macrocosmic apsects (as the inner perceiver and nondual essence respectively). Understanding exactly how this is the case will help resolve the problem of whether the “No Self” teaching of the Vajirā Sutta was implicit in other early teachings, or whether it was a philosophical development from an earlier period dominated by different concerns.

2. The Not-Self teaching

Perhaps the most important source for the early Buddhist critiques of the atman are the various sectarian accounts of the Buddha’s Second Sermon. The first teachings of this sermon states that the five aggregates (form, feeling, apperception, volitions and consciousness) are “not ātman/attan” (anātman/anattan) since they are beyond a person’s control. ²¹ The precise meaning of this teaching is unclear: quite what the word atman/attan means, and why the ability to control each of the five aggregates would mean that they constitute an ātman/attan, is difficult to make out. ²² The peculiar content of this teaching is matched by the fact that it is mentioned in only one Pāli discourse (the Cūlasaccaka Sutta) besides the two Pāli texts that record the Second Sermon. [^27] This peculiarity suggests that the teaching was of little importance in the early Buddhist period. Much more important is the second teaching of the Second Sermon. As Collins has pointed out, “a very high proportion of the discussions of Not-Self in the Suttas consist in various versions of this argument.” [^28] This teaching is in fact the most important early Buddhist negation of the ātman, and one that has distinct Upaniṣadic undertones:

What do you think, bhikkhus: is form permanent (nicca) or impermanent (anicca)?

‘Impermanent, master.’

Is that which is impermanent unsatisfactory (dukkha) or satisfactory (sukha)?

‘Unsatisfactory, master.’

And is it suitable to regard that which is impermanent, unsatisfactory and subject to change (viparināmadhamma) as “This is mine, I am this, this is my attan”?

‘No, master. ²³

[^27] Collins 1982: 97 has suggested that this teaching is directed against the Brahminic notion of the ātman as the “microcosmic reflection of the macrocosmic force of the universe (brahman).” But the teaching does not presuppose that the ātman should be an inner controller, and if so it would not appear to be a “microcosmic reflection” of a world-controlling force. The teaching instead states that if the five aggregates were ātman/attan, a person should be able to change them as he wants. The argument is thus directed against the notion that the five aggregates constitute the ātman/attan, and not against the notion that there is ātman/attan controlling them from within, as is also assumed by Siderits (2007: 46ff.). For a full study of the conceptual and historical implications of this teaching see Wynne 2009a; especially 86-88. [^28] Collins 1982: 98. Similar teachings begin book IV (Salāyatanavagga) of the Saṃyutta-Nikāya: it is stated that all sense faculties and their objects are impermanent (anicca), unsatisfactory (dukkha) and so Not-Self (anattan). Derivatives of this teaching, where the five aggregates are also stated to be anattan, can be found at SN III.20-21, 23-24 and 179.

Since the list of five aggregates denotes the different aspects of phenomenal being, ²⁴ this teaching therefore states that an ātman/attan cannot be found in conditioned experience. Although it is not immediately clear how the term ātman/attan is to be taken here, Norman has argued that it is to be understood as the blissful and unchangeable ātman of the early Upaniṣads. This is shown by the fact that the response to the Buddha’s final question can only be given by those who know, in advance, that the term attā is by definition nicca and sukha, and therefore anything which is anicca and dukkha cannot be attā. This gives us a clear indication of the type of attā which is being discussed. It is the Upaniṣadic idea of an ātman which is nitya and sukha … ²⁵

If this teaching negates the Upaniṣadic ātman in the sense of an unchanging, blissful essence, it would seem to be concerned with the ātman in its macrocosmic aspect (as brahman), for this is how the blissful ātman is considered in the early Upaniṣads. ²⁶ It does not make any sense, however, to read this teaching as a negation of a macrocosmic essence. While it might make sense to ask whether consciousness has the characteristics of the macrocosmic ātman (since a number of important early Upaniṣadic passages state that the ātman in its macrocosmic aspect is a nondual consciousness), ²⁷ it makes little sense to ask if form, sensation, apperception and volitions have the characteristics of an essence that transcends all phenomena.

That the teaching is not a straightforward denial of the macrocosmic ātman is also suggested by the use of the terms sukha and nicca. The primary sense of these terms seems not to be “permanent” and “blissful,” as would be the case in a straightforward negation of Upaniṣadic thought. The term sukha is rather an antonym of the term dukkha, which here has the weak sense of “unsatisfactory,” this being the only way in which the impermanence (anicca) of the five aggregates can be taken: the point is not that the general experiential condition of a person is “suffering,” but rather that it is ultimately unsatisfactory on the basis that something enduring cannot be found. The teaching therefore appears more like an enquiry into phenomenal identity rather than a direct negation of the Upaniṣadic ātman: it seeks to establish that form and so on are affected by causes and conditions (i.e. that they are adventitious: viparināmadhamma) and so cannot constitute what a person “really and intrinsically” is. If so, it would appear that the teaching is a philosophical enquiry into intrinsic identity or “self.” ²⁸

This impression is strengthened by the fact that the term ātman/attan must here be taken in the philosophical sense of “self.” For it makes no sense to ask whether “form” (i.e. the physical body), sensation, apperception and volitions constitute a “soul,” no matter how the latter is conceived. These are sensible questions in an enquiry into personal identity, however: since a person ordinarily identifies with form and so on as “oneself,” it makes sense to ask if they really can be considered as an intrinsic identity. On this point it is important to note that the Not-Self teaching is often placed directly after the statement that an ordinary person identifies with the five aggregates in the form “This is mine, I am this, this is my self” (ātman/attan). ²⁹ Such texts play on the flexibility in the term ātman/ attan, using it first in the sense derived from the reflexive pronoun (where it denotes a person’s phenomenal identity), and second in the more philosophical sense of intrinsic identity. In this way the teaching points out that although a person takes the five aggregates as his individual “self” (ātman/attan), this is unsatisfactory since no intrinsic identity (ātman/attan) can be found therein. It would seem, then, that the teaching addresses the problem of personal identity by questioning the identification with phenomenal being. To this end the Upaniṣadic notion of an ātman that is blissful and permanent/unchanging is certainly invoked, of course, but this would seem to be only for the sake of communicating a new idea in a particular intellectual context.

This is far from a statement of the No Self doctrine as described by the Vajirā Sutta, Buddhaghosa and Śāntideva: there is no denial of the self per se, but only a subtle argument that the concept of a “self” does not make sense of conditioned experience. Since the underlying metaphysic of the teaching is not made clear, it is possible that it presupposes a self beyond conditioned experience or the exact opposite. If so, the difference between this teaching and the “No Self” doctrine of the Vajirā Sutta remains to be determined. A better understanding of the problem requires an investigation of other early Buddhist teachings on personal identity.

3. Self-consciousness in the early Buddhist texts

The Not-Self teaching considers a person’s identification with the five aggregates in terms of the notion “This is mine, I am this, this is my self.” As such, it is closely connected to other early Buddhist teachings concerned with the notion “I” (ahan ti), the notion “I am” (asmī ti), the “conceit I am” (asmimāna), and the “underlying tendency towards conceit with regard to the notions ‘I’ and ‘mine’ (ahaṃkāramamaṃkāra-mānānusaya).” All of the teachings containing these formulations tackle the subject of personal identity by examining what we might call “self-consciousness,” i.e. a person’s awareness of his own “identity,” “acts” and “thoughts.” ³⁰ Rather than enquire into whether an intrinsic identity can be found in self-consciousness, as the Not-Self teaching does, such teachings explore the affective and cognitive aspects of reflexive awareness. Its affective nature is most apparent in these texts, this being indicated by the compounds the “underlying tendency towards conceit with regard to the notions ‘I’ and ‘mine’ (ahaṃkāra/mamaṃkāramānānusaya)” and “the conceit ‘I am’ (asmimāna).” Both formulations indicate that “for the unenlightened man, all experience and action must necessarily appear phenomenologically as happening to or originating from an ‘I’..‘37 This means, in other words, that self-consciousness is a basic sort of existential conceit, a grasping at individual existence and identity that underpins all conditioned experience.

Other texts affirm this affective understanding of self-consciousness, e.g. the Tanhā Sutta, where eighteen “thoughts caused by thirst” (tanhāvicarita) are listed with reference to both oneself (ajjhatikassa upādāya) and that which is external to oneself (bāhirassa upādāya) respectively. ³¹ Thus the notion “I am” (asmī ti) is the most basic “thought caused by thirst” that paves the way for seventeen further forms of self-consciousness with reference to oneself (e.g. evasmī ti: “I am thus”); and the notion “I am in respect of this” (imināsmī ti) is the most basic “thought caused by thirst” that paves the way for seventeen further forms of self-consciousness with reference to that which is external to oneself (e.g. iminā evasmī ti: “I am thus in respect of this”). ³²

In contrast to this analysis of the affective aspect of self-consciousness, other texts suggest that self-consciousness is a sort of ignorance. A good example is the Sammāditthi Sutta, which equates the destruction of the “underlying tendency towards conceit with regard to the view ‘I am’ (asmī ti ditthimānānusayaṃ samūhanitvā)” with the abandonment of ignorance (avijjaṃ pahāya) and the attainment of knowledge (vijjaṃ uppādetvā). ³³ Other texts combine the affective and cognitive aspects of self-consciousness. Thus the Dutiyanānātitthiya Sutta concludes with the following “inspired utterance” (udāna) of the Buddha:

This generation is obsessed with the notion ‘I’ (ahaṃkāra)
and attached to the notion ‘another’ (paraṃkāra).
They have not understood this matter,
and have not seen that it is a barb.
For the person with vision who has removed this barb,
the notions ‘I am acting’ and ‘Another is acting’ do not occur.
This generation is mired, bound and trapped by conceit (māna);
it exerts itself over views, and so does not escape transmigration. ³⁴

The notions of “I” (ahaṃkāra) and “another” (paraṃkāra) are here related to conceit (māna) and the holding of speculative views (ditthi). Self-consciousness, therefore, is deeply involved in the affective and cognitive causes of a person’s suffering. Given that it is said to be an “underlying tendency” (anusaya), it would seem to be an ever-present factor of conditioned experience, and its transcendence would no doubt effect a substantial personal transformation. Although the early texts generally have little to say about the state achieved through eradicating self-consciousness, there are a couple of illuminating exceptions. A series of Suttas in the SamyuttaNikāya describe how Sāriputta attained all nine “gradual abidings” (anupubbavihāra) - from the first jhāna to the “cessation of apperception and feeling” (saññāvedayitanirodha) - despite lacking self-consciousness. These texts begin with a question from Ānanda about the reason for the unusually calm countenance of Sāriputta. The latter explains that this is due to the meditative states he attains without any prior intention, a state of affairs that Ānanda attributes to his lack of self-consciousness:

‘Herein, venerable sir, I pass my time having attained the first jhāna, that state of joy and bliss born of seclusion which is devoid of desire and bad thoughts, but which includes deliberation (vitakka) and reflection (vicāra). It does not occur to me, venerable sir, that I am attaining the first jhāna, or have attained the first jhāna, or have emerged from the first jhāna.’

‘It is so for the venerable Sāriputta because the underlying tendency towards conceit in the notions ‘I’ and ‘mine’ has for a long time been destroyed. Therefore it does not occur to the venerable Sāriputta that he is attaining the first jhāna, or has attained the first jhāna, or has emerged from the first jhāna. ³⁵

Apart from another text in which Ānanda states that Sāriputta’s “underlying tendency towards conceit in the notions ‘I’ and ‘mine’ has for a long time been destroyed,³⁶ the only other discourse on the unusual psychology of a person devoid of self-consciousness is the Upasena-āsīvisa Sutta. In this peculiar text, the venerable Upasena is said to have been bitten by a poisonous snake while both he and Sāriputta dwelt in the Sappasondika mountain cave. ³⁷ When he subsequently asks to be taken outside on a couch, before his body “falls apart right here, just like a fistful of chaff,”³⁸ Sāriputta exclaims that he sees no change in his body or decline in his faculties. ³⁹ To this Upasena states that his unusual countenance is due to the fact that he lacks self-consciousness with regard to his sense faculties:

Venerable Sāriputta, the person who might think that he is the eye or possesses it … that he is the tongue or possesses it … that he is the mind or possesses it, for him there might be a change in his body or a decline in his faculties. But it is does not occur to me, venerable Sāriputta, that I am the eye or possess it … that I am the tongue or possess it … that I am the mind or possess it. So how could there be a change in my body or decline in my faculties? ⁴⁰

Sāriputta thus concludes that “the venerable Upasena’s underlying tendency to feel conceit in the notions ‘I’ and ‘mine’ has for a long time been destroyed, [^348] and the story concludes with the account of how Upasena’s body fell apart “like a fistful of chaff” after he had been taken outside on a couch. ⁴¹ This text thus claims that Upasena achieved a completely impersonal state, one in which the automatic tendency to identify with conditioned experience had ceased to function.

Various texts describe the means of attaining this state, e.g. following the path that leads through the four jhānas and culminates in the three knowledges, ⁴² or concentrating on the thought “this is calm, this is supreme, namely the calming of all mental formations, the relinquishment of all attachment (upadhi), the destruction of thirst, dispassion, cessation, Nirvana,” which is said to lead to the attainment of the “release of mind, a release through understanding. ⁴³ For the purpose of the present enquiry, however, more important are other texts which relate the transcendence of self-consciousness to the understanding of the five aggregates and the Not-Self teaching. The Dutiyasaññā Sutta of the AñguttaraNikāya, for example, states that the problem of self-consciousness (ahaṃkāra/mamaṃkāra) is resolved by regarding that which is unsatisfactory (dukkha) as Not-Self (anattan). ⁴⁴ A similar contemplation is outlined in the Mahāpunna Sutta, which states that seeing the five aggregates as Not-Self (attan) leads to the cessation of selfconsciousness as follows:

O bhikkhus, one should regard whatever form is past, present or future, internal or external, gross or subtle, inferior or superior, near or far - all form - as ‘This is not mine, I am not this, this is not my self (attan)’ … For the person who knows and sees it thus, bhikkhus, the tendency towards conceit in the notion ‘I’ with regards to the body and its consciousness, and towards conceit in the notion ‘mine’ with regards to external objects, does not arise. ⁴⁵

This passage does not state how contemplating the insubstantiality of conditioned experience (the five aggrgegates) aids the transcendence of self-consciousness, but the point is investigated in more detail in the Khemaka Sutta, where the bhikkhu Khemaka addresses the elders of Kosambī as follows:

Venerable sirs, the Blessed one has spoken of five aggregates of attachment, ⁴⁶ namely: the aggregate of attachment that is form … feeling … apperception … volitions … [and] consciousness. I have no view that any sort of self (attan) or its property (attaniya) is found in these five aggregates of attachment, venerable sirs, and yet I am not an arahant devoid of corruptions. For I still have the notion ‘I am’ (asmī ti) with regard to these five aggregates of attachment, venerable sirs, despite the fact that I do not have the view ‘I am this’ (ayam asmī ti na са samanираззамиі). ⁴⁷

The logic of this statement is relatively simple. Khemaka knows that he should be detached from the conditioned experience of the five aggregates, this being inherently unsatisfactory since it lacks intrinsic identity (attan). But he is unable to do so because of his automatic tendency to identify with conditioned experience in the form of the notion “I am.” Although Khemaka knows what he should know, according to Buddhist doctrine, and so does not intentionally identify with the the five aggregates, his identification with them runs deeper in the form of a sense of subjectivity (asmī ti) that takes them as its locus. What is required to achieve detachment from the five aggregates, according to Khemaka, is the following contemplation:

Although a noble disciple might have abandoned the five lower fetters, it might occur to him that the conceit (mäno), intention (chando) and underlying tendency (anusayo) ‘I am’ (asmī ti) with regard to the five aggregates of attachment has not been destroyed. At another time, he immerses himself in observing the rise and fall of the five aggregates of attachment: ‘form … feeling … apperception … volitions … consciousness is thus, its arising is thus, its fading away is thus.’ In doing this the conceit, intention and underlying tendency ‘I am’ with regard to the five aggregates of attachment that had not been destroyed is destroyed. ⁴⁸

The practice mentioned here - the contemplation of conditioned experience (the five aggregates) as a process - seems to be an attempt to see the truth of the Not-Self teaching at a deeper level: the bhikkhu does not simply think about the insubstantiality of the five aggregates, but attempts to see this truth experientially. According to the Khemaka Sutta, such a contemplation is a more powerful means of overcoming the subtle sense of identification with conditioned experience, but the aim of both contemplations is, however, the same, i.e. detachment from the five aggregates leading to the cessation of identification with them.

All of the above passages on self-consciousness elaborate the typically Buddhist understanding that desire and ignorance cause suffering. That self-consciousness is a problem of an affective nature is easy to understand, of course, for self-consciousness implies self-centredness which in turn implies psychological states - selfishness, desire etc. - which are, according to the Buddhist analysis, ethically and spiritually harmful. But if the fundamental problem of desire - the sole cause of suffering according to the Second Noble Truth - can be controlled and suppressed via various religious practices, should it matter that the person who suppresses it is still self-conscious? According to the Khemaka Sutta it does, for the locus of the notion “I am” is the five aggregates. This means that the self-conscious person is inevitably attached to conditioned experience, and if so it is not enough to conquer one’s desires and abide in a state of altruism, for self-consciousness is by its very nature a subtle form of clinging to conditioned experience: as we have seen, self-consciousness is a subtle, underlying form of grasping after individual experience and identity, a basic existential ‘conceit’ (p. 114). Being self-conscious, in other words, means to be both ignorant and attached, and so subject to suffering.

These teachings on personal identity seem to have the same pragmatic point as the Not-Self teaching, i.e. the comprehension and abandonment of the underlying affective and cognitive cause of suffering: self-consciousness. And just like the Not-Self teaching, these teachings seem to have no obvious metaphysical significance: it is not clear if detachment from the five aggregates and transcendence of self-consciousness means that a person transcends intrinsic identity per se or whether he has some other sort of transcendent identity. This is true even of the texts that describe how Sāriputta attained certain meditative states without being aware of it, apparently because he lacked self-consciousness: the focus of these texts is the psychology of Sāriputta rather than more abstract concerns, such as the ontological nature of the state attained by him.

Attempts to read a particular metaphysic into such texts are far from persuasive. Pérez-Rémon, for example, has argued that two kinds of personal identity can be indentified in the series of texts on Sāriputta’s meditative attainments. On the one hand there is the socalled “asmimanic self” which is “contained in expressions such as ‘I am attaining the first jhāna, I have attained the first jhāna, I have emerged from the first jhāna’ …“57 On the other hand, PérezRémon believes that forms of the first personal pronoun in statements such as “I dwell having attained the first state of meditation” (idhāhaṃ … pathamaṃ jhānaṃ upasampajja viharāmi) and “it [did not occur] to me thus, venerable sir” (tassa mayhaṃ, āvuso, na evam hoti) is “incompatible with the ‘asmimanic I’ contained in the expressions that follow.” He therefore argues that

Sāriputta is able to say ‘I dwell having attained the cessation of awareness and feeling.’ If the ‘I’ of this sentence cannot stand either for the the asmimanic self or for the genuine empirical moral agent, what kind of self does it stand for? Nothing is left but to say that it stands for the true self who, in that condition attains to a complete aloofness from the empirical factors in the isolation that is his very being. ⁴⁹

This argument is not very convincing. Pérez-Rémon supposes that there is a metaphysical difference between the subject of the verb in expressions such as paṭhamajjhānaṃ upasampajja viharāmi, which apparently indicates the “true self,” and expression such as ahaṃ paṭhamajjhānaṃ samāpajjāmī ti, which apparently indicates the “asmimanic” or phenomenal self. Such a distinction is entirely arbritary, and does not convince that an intrinsic identity is presupposed by the early Buddhist teachings. Indeed key phrases, such as tassa mayhaṃ āvuso na evaṃ hoti: ahaṃ paṭhamajjhānaṃ samāpajjāmī ti, seem to cancel out the positive language that they follow, such as paṭhamajjhānaṃ upasampajja viharāmi. This would seem to incline towards the position that there is no true subject of experience. A better argument, then, would in fact be that since Sāriputta is unaware of what he experiences, these passages imply the No Self doctrine.

Although the Saṃyutta-Nikāya texts on Sāriputta’s meditative attainments perhaps incline towards the No Self doctrine, another argument for intrinsic identity in early Buddhism can be made. This is that since the Not-Self teaching considers only the experiential aspects of the human being (the five aggregates), it might leave room for a non-phenomenal or transcendental self. Indeed the five aggregates are presented as objects of identification or appropriation for the perceiving subject, who is denoted by the term “I” in the expression “I am not this.” It is possible that such an “I” could stand for a transcendent subject of experience. This being the case, the arguments for or against a self would seem to be well-balanced. There are further passages to consider, however, and these are more useful in determining the metaphysical presuppositions of the early Buddhist texts on personal identity.

4. Dependent origination and self-consciousness

We have seen that the Khemaka Sutta advises a contemplation of the rise and fall of the five aggregates as a means of achieving the cessation of the notion “I am” with regard to them. In this context, the “I” that ceases would seem to be a person’s notion of an individual identity comprised of form, sensation, apperception and so on, for it is this that is undermined by seeing that the five aggregates are impermanent. If so, the text would not seem to deny that subjectivity remains intact in the liberating experience, and the possibility remains that it presupposes a subjectivity abstracted from the five aggregates as a person’s real self, i.e. a sort of transcendent “I” that experiences detachment from the five aggregates. The same could be said for the Not-Self teaching: the statement ‘I am not this’ might presuppose a transcendent “I” beyond the five aggregates. In another respect, however, this point is not so clear. For the Khemaka Sutta goes on to compare the notion “I am” to the subtle persistence of a flower’s scent, a simile in which the “I” seems to indicate simple subjectivity:

‘It is just like the scent of a blue lotus, a red lotus or a white lotus. Would a person be describing it correctly if he were to say that the scent belongs to its leaves, its colour or filaments?’

‘It is not so, venerable sir.’

‘How, then, would one describe if correctly? If one were to describe it correctly, one would say that the scent belongs to the flower. In just the same way, venerable sirs, I do not declare ‘I am’ with regard to or apart from form … feeling … apperception and consciousness. And yet I still have the notion of ‘I am’ (asmī ti) with regard to these five aggregates of attachment, venerable sirs, despite the fact that I do not think ‘I am this’.⁵⁰

Khemaka’s primary point is that the locus of the notion “I am” is the five aggregates as a whole, in a manner comparable to how a scent lingers around the whole flower. But the simile also points out that just as scent emerges from the whole flower, so too does the notion “I am” arise from conditioned experience as a whole. This might indicate the understanding that the notion “I am” is an emergent state of consciousness, one in which the “I” is felt to stand apart from its objective locus as a quasi-independent subject of experience. The text is therefore ambiguous: it is not entirely clear if the term “I” refers to a person’s sense of being a composite entity made up of different phenomenal aspects, or whether it refers to a person’s sense of being a quasi-independent subject of consciousness that observes the different aspects of conditioned experience. Forming a correct understanding of this ambiguity is vitally important. For if the early Buddhist texts understand the “I” in expressions such as asmimāna, ahamkāra, asmīti and so on in the sense of a quasi-independent subject of consciousness, they would imply that liberation involves the cessation of subjectivity per se, in which case there would be little possibility that a person has a self.

Just as ambiguous as the Khemaka Sutta is the Ānanda Sutta, in which the bhikkhu Punṇa teaches that the notion “I am” occurs only in relation to (upādāya) the five aggregates, rather than independently (anupādāya). ⁵¹ This could indicate the understanding that the ” I ” is a quasi-independent observer of the five aggregates, rather than a person’s sense of being an “I” made up of the five aggregates. On the other hand, however, this text also includes the Not-Self teaching. ⁵² Since this teaching deconstructs a person’s identification with the five aggregates, and since it is here presented as a means of overcoming the notion “I am,” the term “I” would here seem to denote the phenomenal person as a whole, i.e. the individual “I” understood as an aggregate of five experiential parts. On balance, the presence of the Not-Self teaching in the Ānanda Sutta, probably indicates the latter possibility, i.e. that the term “I” refers to a person’s sense of individuality comprised of different phenomenal aspects. Furthermore, since the Khemaka Sutta’s contemplation of the five aggregates is simply an expansion of the NotSelf teaching, the same conclusion probably applies to it.

Other texts definitely do not share this understanding, however, but seem to veer more towards taking the term “I” as a person’s sense of being a quasi-independent subject of experience. They therefore suggest a relationship of dependence or emergence rather than identification between the “I” and the five aggregates. One such text is the Vinopama Sutta, which uses the simile of the lute (vīṇā) and its sound to describe the relationship between the subjective aspect of self-consciousness and the five aggregates:

This thing called a lute, venerable sir, consists of many different components so that when played it makes a sound by means of them: dependent on the parchment sounding board, the belly, the arm, the head, the strings, the plectrum and the appropriate effort of the musician, this lute, venerable sir, which consists of many various components, is played and makes a sound by means of them. ⁵³

In this simile the different parts of the lute denote the different aspects of a person’s phenomenal being, whereas the sound that emerges from them denotes the notion “I am,” i.e. a person’s sense of subjectivity. The simile of the lute thus suggests that the sense of being an inner perceiver emerges from the different aspects of conditioned experience functioning as a whole, so that an apparently independent subject of experience is constructed out of an insubstantial process. This implies that the subjective aspect of selfconsciousness does not exist apart from the five aggregates, despite the cognitive separation of functions (subject vs. object) that might give this impression. In the language of Buddhist philosophy, this means that the ” I ” is dependently originated and so not ultimately real. If so, a contemplation of the five aggregates could be used to emphasise the fact that the inner subject is dependent on an insubstantial process, and in this way lead to the cessation of a person’s sense of being a quasi-independent subject of experience. Such a contemplation is suggested in the Vīnopama Sutta, which states that the notions “I” (ahan ti), “mine” (maman ti) and “I am” (asmī ti) can be transcended by investigating the limits of conditioned experience as follows:

In just this way, bhikkhus, the bhikkhu investigates form to its full extent, he investigates feeling … apperception … volitions … [and] consciousness to its full extent. When he does this, the notions ‘I’, ‘mine’ and ‘I am’ are found in him no longer. ⁵⁴

This passage suggests that by contemplating the limitations of conditioned experience, the subjective aspect of self-consciousness the “I” that perceives the five aggregates - ceases to function. This contemplation is similar to the contemplation of the rise and fall of the five aggregates outlined in the Khemaka Sutta. But the purpose here seems to be that of emphasising the limitations of that on which the notion ” I ” is founded. This seems to show that the ” I ” is limited to the impermanent processes of conditioned experience, and so cannot be separated from them. Dependence on the five aggregates, then, would here seem to indicate that the subjective aspect of self-consciousness is not independent or ultimately real.

Suggestive as they are, the similes of the lute and flower do not explicitly state that the term ” I ” is to be understood as a quasi-independent subject of experience. There is still some room to doubt, then, that these texts deny the notion of an independent subject of experience per se. Other texts certainly do focus on this understanding, however. We have seen, for example, that the Sāriputta Sutta describes the destruction of the underlying tendency towards conceit in the notions “I” and “mine” (asmimāna/mamaṃkāraanusaya), so that Sāriputta attains each of the nine states of meditation (anupubbavihāra) without an awareness of the fact that he attains, abides in or emerges from them. ⁵⁵ Such a description seems to imply the complete cessation of the subjective aspect of selfconsciousness, as if Sāriputta is in a totally impersonal and “selfless” state in which there is no sense of being an inner perceiver or “I” that observes and comprehends what is happening. In such texts there seems to be little room for a subjective aspect of consciousness that could be taken as the true self in opposition to the phenomenal self (consisting of the five aggregates).

A similar understanding is suggested in the Aggivacchagotta Sutta. Like the Khemaka Sutta it advocates the practice of contemplating the rise and fall of the five aggregates, although the end result is expressed slightly differently. The key passage occurs when the Buddha responds to Vacchagotta’s question as to whether he has any views (ditthigata):

‘Does the venerable Gotama have any view (ditthigata)?’

The very notion of ‘view’ has been dispelled by the Tathāgata, O Vaccha, for the Tathāgata has seen this: ‘Form is thus, its arising is thus, its fading away is thus; sensation is thus, its arising is thus, its fading away is thus; apperception is thus, its arising is thus, its fading away is thus; volitions are thus, their arising is thus, their fading away is thus; consciousness is thus, its arising is thus, its fading away is thus.’ Therefore I say that the Tathāgata is released through the destruction, fading away, cessation, abandonment and relinquishment of all thoughts (mañ̃ita), agitations (mathita), and every underlying tendency towards conceit (māna) with regard to the notions ‘I’ (ahimkāra) and ‘mine’ (mamiṃkāra). ⁵⁶

Understanding the five aggregates as a process is here not said to eradicate a person’s identification with conditioned experience (the notion “I am” with regard to the five aggregates), but rather to eradicate the notions “I” and “mine.” While the understanding of the term “I” is not made clear, this emphatic description of the Buddha’s liberated state seems to indicate a complete transcendence of phenomena: the cessation of all conceptuality (mañ̃ita) implies the cessation of the entire contents of consciousness, including the sense of being a quasi-independent subject of experience. If so, the text indicates that in the Buddha’s awakened state the sense of subjectivity per se has been transcended.

In support of this interpretation are a number of texts that focus on the dependent origination of the notion “I am.” One such text is the Vepacitti Sutta, which relates the myth of the defeat of the demons (asura) by the gods (deva), and the shackling of their leader Vepacitti. The doctrinal point of this myth is that the bonds of Vepacitti operate as a function of his thoughts:

O bhikkhus, when Vepacitti the leader of the demons thought ‘The gods are righteous, but the demons are not, and so I will go, right here and now, to the citadel of the gods,’ he saw that he was released from the five bonds wrapped round his neck, and being presented and endowed with the five sorts of heavenly sensual pleasure he enjoyed himself. But when, O bhikkhus, Vepacitti the leader of the demons thought ‘The demons are righteous, but the gods are not, and so I will go there right now, to the citadel of the demons,’ he saw that he was shackled by the five bonds wrapped round his neck, and so was deprived of the five sorts of heavenly sensual pleasure. That is how subtle the bonds of Vepacitti are, O bhikkhus, but even subtler is the bond of Māra. The person who thinks, O bhikkhus, is shackled by Māra - but by not thinking he is released from the Evil One. ⁵⁷

There is perhaps no clearer Buddhist text on the notion that thinking or conceptualisation (maññita) causes bondage. The text goes on to state that the foundation of this conceptual bondage is the subjective aspect of self-consciousness:

‘I am,’ O bhikkhus, is a thought (maññita); ‘I am this,’ O bhikkhus, is a thought; ‘I will be,’ O bhikkhus, is a thought; ‘I will not be,’ O bhikkhus, is a thought; ‘I will possess form,’ O bhikkhus, is a thought; ‘I will not possess form,’ O bhikkhus, is a thought; ‘I will be conscious,’ O bhikkhus, is a thought; ‘I will be unconscious,’ O bhikkhus, is a thought; ‘I will be neither conscious nor unconscious,’ O bhikkhus, is a thought. Thought, O bhikkhus, is an illness, a boil and a barb. Therefore, O bhikkhus, you should train yourselves with the thought ‘I will pass my time with a mind free from thinking. ⁵⁸

A similar analysis is found in the Dhātuvibhanga Sutta, which adds to this the point that thinking itself (maññita), especially in terms of the notion “I am” (asmī ti), is fundamentally problematic - “an illness, a boil and a barb” that ought to be transcended (samatikkama). ⁵⁹ The reason for this critique is made clear in the Vepacitti Sutta, which explains that each of the notions beginning with “I am” is “an impulse” (iñjita), “a palpitation” (phandita), a “conceptual proliferation” (papañcita) and “a conceit” (mānagata). ⁶⁰ All these terms indicate that the various manifestations of the subjective aspects of self-consciousness - the “I” as a quasi-independent observer of phenomena - arise in dependence on the conceptual activity of the mind. This is especially true of the term papañcitaṃ. In contrast to the Brahminic notion of prapañca as the manifoldness or diversity of the external world, ⁶¹ the term in early Buddhist texts refers to the tendency of the mind towards conceptual diffuseness or proliferation. ⁶² If so, it would seem that the subjective aspect of self-consciousness is conceptually constructed in the processes of the dependent origination of consciousness, and thus has no independent reality. The clearest explanation of the dependent origination of conceptual proliferation (papañca) is found in the Madhupindaka Sutta:

Visual consciousness arises dependent on the eye and forms, the coming together of the three is contact, sensation arises from contact, one apperceives (sañjānāti) what one senses, thinks over (vitakketi) what one apperceives, and conceptually proliferates (papañceti) what is thought over. From conceptual proliferation comes reckoning, of one’s conceptual proliferations and apperceptions, and this afflicts a man with regard to the past, present and future forms cognised by the eye. ⁶³

This passage states that the conceptual forms constructed in the cognitive process ultimately cause a person’s suffering. Selfconsciousness arises in this way: according to the Vepacitti Sutta its subjective aspect - the various forms in which the notion “I” is expressed - is a form of conceptual “proliferation” or “manifoldness” (papañcita). This implies, then, that the subject of self-consciousness does not exist beyond particular cognitive events. Such an analysis leaves little room for an inherently real self denoted by the term “I.” Indeed the Vepacitti Sutta’s comprehensive account of the forms in which this notion occurs seems to indicate that there is no true “I” behind its appearances in thought. At the least, there is very little ground on which this case could be made.

That the notion “I am” indicates subjectivity per se, and that this is dependently originated, is made explicitly clear in an important section of the Mahānidāna Sutta which analyses three notions of intrinsic identity. ⁶⁴ The first is the simplest: a self (attan) identical to sensations (vedanā) is dismissed since this would mean that intrinsic identity is changeable, i.e. a contradiction is terms. ⁶⁵ The second and third understandings of the self are more subtle, however. The latter seems to reject the notion that the subjective aspect of self-consciousness is independently real:

‘Therein, Ānanda, to the person who claims that his self (attan) is different from sensation (vedanā) but not without experience (no pi appațisaṃvedano), it being able to sense (attā me vediyati) and having sensations as a property (vedanādharmo hi me attā), one should say: “When sensation has completely and utterly ceased without remainder (vedanā va hi āvuso sabbena sabbaṃ sabbathā sabbaṃ aparisesā nirujjheyum), when there is no sensation whatsoever since it has ceased (sabbaso vedanāya asati vedanānirodhā), is it possible in that state to have the notion ‘I am this’ (ayam aham asmī ti)’?’

‘It is not so, master.’

‘Therefore, Ānanda, it is because of this reason that it is not suitable to think that one has a self different from sensation but not without experience, it being able to sense and having sensations as a property. ⁶⁶

This passage rejects the notion of an independently real subject of perception. The problem with such a notion is that although it is possible to conceive this understanding of individual identity when conditioned experience (sensation: vedanā) functions normally, this is not the case in the absence of these conditions. This critique therefore makes the point that the sense of being an inner perceiver only arises under certain conditions, those that pertain in conditioned experience, and thus rejects the notion that a person’s sense of being an independent subject of perception is ultimately real. Such an analysis leaves little room for any sort of self, for what identity could a person have apart from the subjectivity denoted by the term “I?” The only other possibility is perhaps that a person has a transcendent identity beyond conditioned experience. The problem with such a notion is made clear in the second of the Mahānidāna Sutta’s critiques of intrinsic identity. This section of the text points out the impossibility of integrating the concept of individuality (a necessary aspect of any notion of personal identity) with that of transcendence (which implies a completely impersonal state):

‘Therein, Ānanda, to the person who claims “my self (me attā ) is beyond sensation (na … vedanā) and experience (appațisaṃvedano),” one should say: “Is it possible to have the notion ‘I am’ (asmī ti) when there is no sensation whatsoever (sabbaso vedayitaṃ n’ atthi)?”’

‘It is not so, master.’

‘Therefore, Ānanda, it is because of this reason that it is not suitable to think that one has a self beyond feeling and experience. ⁶⁷

The problem with the notion of a transcendent self, according to this analysis, is that personal identity is conceived in terms of something completely impersonal. But how could it be claimed that a transcendent self is “one’s own” (me attā ) when the individualising factor of self-consciousness (asmī ti) is absent? The notion that something is “one’s own” depends on a person being self-conscious and so able to conceptually appropriate it. The absence of the conditions necessary for self-consciousness, then, would seem to render identification with a truly transcendent state impossible. A state beyond conditioned experience cannot be conceived in terms of personal identity, therefore, the latter pertaining only within certain, limited, cognitive states.

These two critiques consider the same problem from different angles: the third critique points out the problem of hypostasising the inner perceiver into a transcendent entity, whereas the second critique points out the problem of individualising the transcendence of conditioned experience. Since notions of a self within conditioned experience or beyond it are both negated, this dual analysis would seem to leave no room for any sort of intrinsic identity. This suspicion is confirmed by the fact that these critiques respond to similar conceptualisations of intrinsic identity stated in the eighth chapter of the Chāndogya Upaniṣad. In the dialogue between Prajāpati and Indra (CU VIII.7-12), Prajāpati presents various ways of understanding the ātman to his pupil Indra, each of which is in turn rejected and replaced by a higher and more sophisticated understanding. The final two conceptualisations of the ātman match the final two notions of the attan criticised in the Mahānidāna Sutta studied above. Prajāpati first presents the ātman in its transcendent or macrocosmic aspect, i.e. as an intrinsic identity beyond conditioned experience, which Indra rejects as follows (CU VIII.11.1):

‘In the state in which a person falls into deep sleep, so that he becomes whole, completely tranquil, and does not perceive even a dream, this is the self,’ said [Prajāpati] - ‘it is the immortal free from fear, it is brahman.’

Indra then left, his heart fully satisfied. Before reaching the gods, however, he saw this problem: ‘This person certainly does not know himself - nor even these beings here - directly in the form ‘I am this’ (ayam aham asmīti); he has become completely annihilated. I see nothing beneficial in this. ⁶⁸

This conceptualisation of the ātman, and Indra’s reasons for rejecting it, correspond closely to the Mahānidāna Sutta’s second critique of intrinsic identity (attan): both address the notion that a person’s true identity is to be found beyond self-consciousness, the problem being that there is no means of identification with such a state. A similar correspondence can be seen between the Chāndogya Upaniṣad’s final description of the ātman (CU VIII.12.2-5) and the Mahānidāna Sutta’s third and final critique of the attan:

(2). The wind has no body; the clouds, lightning and thunder are also bodiless. Just as these, rising up from space and reaching the highest light, emerge into their true form (3) so too does this tranquil one, rising up from this body and reaching the highest light, emerge into his true form. He is the supreme person and wanders about there laughing, playing and enjoying himself with women, carriages and relatives, without being aware of this appendage of the body. Just as a draught animal is harnessed to a cart, so is this lifebreath harnessed to this body.

(4). Thus the person whose faculty of vision is fixed on space, he is the subject who sees (cākṣuṣah puruṣah): the faculty of vision merely enables him to see. Thus the one who thinks ‘let me smell this,’ he is the self; the olfactory faculty merely enables him to smell. Thus the one who thinks ‘let me utter this,’ he is the self; the faculty of speech merely enables him to speak. Thus the one who thinks ‘let me hear this,’ he is the self; the faculty of hearing merely enables him to hear.

(5). Thus the one who thinks ‘let me think this,’ he is the self, and mind is his divine faculty of vision. It is only this one who - through his mind, the divine faculty of vision - sees the pleasures to be found in the world of Brahma, and enjoys them. ⁶⁹

This passage places the inner perceiver of the Yājñavalkyakānḍa within a more developed Brahminic cosmology, so that after its separation from the body such a self is said to exist in the world of Brahma and enjoy its pleasures. Although this cosmology is entirely absent in the Mahānidāna Sutta, this Buddhist text reports the same idea, i.e. the notion that the inner perceiver constitutes a person’s intrinsic identity. There can be no doubt, then, that ideas have been shared between the two texts. Indeed, the Mahānidāna Sutta’s first critique of the attan also corresponds to the formulation of self that precedes the final two teachings of Chāndogya Upaniṣad VIII. Prajāpati here teaches that the ātman is the bodily self, a notion that Indra rejects as follows:

Just as this self becomes well-adorned when the body is well-adorned, becomes well-dressed when the body is well-dressed, and decorated when the body is decorated, so too does it become blind when the body is blind, become weary when the body is wearied, and crippled when the body is crippled. It is annihilated in consequence of the body’s annihilation. I see nothing beneficial in this. ⁷⁰

This rejection of a bodily self is akin to the Mahānidāna Sutta’s rejection of a self consisting of sensation, a notion that is rejected because such a self would be changeable and subject to suffering. ⁷¹ Both critiques thus point out that the notion of identity with a bodily self is unsatisfactory because of its transience. This correspondence confirms that the Buddhist and Brahminic texts are parallel. If so, it would seem that the Buddhist text has drawn from a Brahminic source, for it goes one step further than the Brahminic parallel by criticising the final formulation of personal identity, and as such seems to extend and supplement an already existent teaching. Furthermore, the order of the Buddhist text is peculiar. It would make better sense if the third critique (the notion that the inner perceiver constitutes a person’s intrinsic identity) preceded the second critique (the notion of a transcendent self beyond conditioned experience), for this order - the denial of a self within conditioned experience followed by the denial of a self beyond it - makes better sense of an analysis that begins with the bodily human being. That this order is not followed is odd, but can be easily explained as a response to an Upaniṣadic teaching in which the order had been determined by the need to claim superiority for the notion of the ātman as an inner perceiver.

It would seem, then, that an important early Upaniṣad has been used to communicate the early Buddhist critique of the ātman/attan. This helps clarify the meaning of the Buddhist teachings on personal identity. The final teaching of Chāndogya Upaniṣad VIII understands the self as an inner perceiver, i.e. the “unseen seer” of the Yājñavalkyakānḍa, and so hypostasises a person’s sense of subjectivity into an intrinsic identity. Since this idea is firmly rejected, based on the argument that it only pertains in a conditioned state of consciousness, there would seem to be no sense in which early Buddhist thought allows a self in the form of the subjective aspect of self-consciousness: a person’s sense of being an independent subject denoted by the term “I” cannot be hypostasised into a self, according to the Mahānidāna Sutta.

This indicates that the early Buddhist critiques of the notions “I” and “I am” ultimately address the problem of subjectivity per se. In the end, the Buddhist teachings therefore point out that the separation of subjective and objective aspects of self-consciousness has no ultimate ontological basis. If so, there would seem to be no room in the early Buddhist analysis for any sort of self. Indeed the second critique of the Mahānidāna Sutta even criticises the notion of a transcendent self, and so suggests the impossibility of there being any such thing as intrinsic identity: such an analysis leaves no room for a self abstracted from causes and conditions. This suggests that the point of the early Buddhist critiques of the ātman, as seen in the Not-Self teaching as well as the various other analyses of self-consciousness, is finally that a person has no intrinsic identity and thus is “selfless.” If so, the difference between the Not-Self teaching and the No-Self teaching of the Vajirā Sutta would seem to be only terminological, the implicit point of the former being explicitly articulated in the latter. But before accepting this interpretation, we must first explain why the “No Self” doctrine was not expressed explicitly in the first place.

5. Cognitive conditioning and personal identity

In the above analysis of early Buddhist texts on self-consciousness, the most crucial evidence is provided by the Mahānidāna Sutta. It is this text more than any other that points towards the No Self doctrine, for if intrinsic identity cannot be found in the inner perceiver or beyond it, what sort of “self” could a person possibly have? The apparently obvious implication of this text is that a person has no intrinsic identity. But if so, why does the text not state this explicitly? Why does it instead point out that it is “not suitable” (na kkhamati) to think that there is neither a self beyond conditioned experience, nor that the inner perceiver is a self? The same feature can be seen in the Not-Self teaching, which similarly asks if it is suitable (kallaṃ nu) to regard the five aggregates as one’s self. This is peculiar: if these early critiques really are based on the No Self doctrine, it is odd that they dodge the issue with such evasive formulations. The form of these teachings must be explained before concluding that the No Self doctrine is implicit in them.

The only possible reason for the failure to articulate an explicit No Self doctrine is that this abstract philosophical issue is avoided for practical purposes. The pragmatic bent of early Buddhism is well attested in the canonical texts, of course: the simile of the raft and the simile of the arrow make it clear that all unnecessary speculation is better off avoided by the person who wishes to attain Nirvana. ⁷² This pragmatic minimalism is even the reason given for the failure to affirm or deny the self’s existence in one canonical text, where the Buddha explains that he did not answer Vacchagotta’s direct questions because he did not wish to confuse him. ⁷³ This approach, in which the psychological well-being of a spiritual seeker is deemed more important than an abstract point, aptly summarises the early Buddhist approach to philosophical discourse: all that is not directly connected to achieving the cessation of suffering is not a proper subject of early Buddhist thought. ⁷⁴ Such a didactic approach would seem to explain the lack of a direct ontological assertion in both the Not-Self teaching and the Mahānidāna Sutta, for the purpose of both is to effect an existential detachment that paves the way for liberation. ⁷⁵ If so, there are good reasons for thinking that the general failure of the early texts to assert the non-existence of the self was due to the pragmatic purpose of early Buddhist discourse, despite the fact that the most relevant teachings point towards this conclusion.

This explanation is not entirely convincing, however. The problem is that there is no obvious reason why denying the existence of the self should be regarded as an unnecessary ontological speculation. For if there is no self, and if the belief in and attachment to the self is the ultimate cause of suffering, it would seem wise to indicate its non-existence as a means of helping others attain liberation. After all, subsequent generations of Buddhist thinkers in India did not have any problem in admitting this fact, and it is hard to see why this should not also have been the case in earlier times. While it might not have been suitable for the Buddha to state the non-existence of the self when asked outright by Vacchagotta, it is easy to imagine that this truth could have been revealed on other occasions when the Buddha’s interlocutors were not so likely to have been confused by the answer. Indeed the Mahānidāna Sutta seems eminently suitable as a context in which to reveal this truth. For after the Buddha has presented an opinion about the self and stated an argument against it, it seems natural to follow up this refutation with the conclusion that a person does not possess such a self: if the teaching does indeed presuppose this, there is no reason why the conclusion should not be “it is because of this reason that a person cannot have a self beyond feeling and experience” rather than “it is because of this reason that it is not suitable to think that one has a self beyond feeling and experience.” In short, if the early teachings claim that belief in the self is the principle cause of suffering, the question of its existence or non-existence is not a pointless ontological question of the kind met with elsewhere in the early Buddhist literature, e.g. whether the world is finite/eternal or not, or whether the Tathāgata exists and so on after death.

If the argument from pragmatism is not entirely satisfactory, we must try to find another explanation for the refusal to address the issue of the self’s existence. Why is this rather straightforward matter of ontology avoided in both the Not-Self teaching and the Mahānidāna Sutta, two teachings well-suited to make this point? And why is the matter avoided in all the other texts on personal identity studied above? These other texts on personal identity suggest that there is a problem with ontology itself. For we have seen although self-consciousness is sometimes considered in terms of the notion “I” (ahan ti, ahamkāra), it is also considered in terms of “the conceit I am” (asmimāna) or more simply “the notion I am” (asmī ti), the latter form being used in the Khemaka Sutta, the Vīnā Sutta, the Vepacitti Sutta, the Dhātuvibhanga Sutta and the Mahānidāna Sutta. Since the formulation “I am” consists of a subject (the personal pronoun “I”) and a verb (-as, “to be”), it surely implies not just a critique of individual identity but also of the notion of individual existence (“I am”). The Vepacitti and Dhātuvibhanga Suttas even expand their critique to include the notions “I will be” (bhavissan ti), “I will not be” (na bhavissan ti) and various other ways of conceiving individual existence. The emphasis in these texts is as much on individual existence as it is on individual identity. Furthermore, since the notion of existence is inconceivable apart from the existence of individual entities, it would seem that the Buddhist texts on personal identity have a problem with the notion of existence itself. Why is this?

The Vepacitti Sutta states that the various forms in which notions of personal existence are expressed are conceptualisations (maññita). This indicates that such notions - which are forms of conceptual proliferation (papañca) - pertain only under certain cognitive conditions, i.e. that they are dependently originated. This suggests the understanding that “existence” does not pertain independent of human consciousness, but is in fact a reality constructed in the cognitive process, i.e. a conceptual rather than an ultimate truth. While this would seem to provide an extreme solution to the failure to state the non-existence of the self, it at least provides a logical explanation for this peculiarity. For it suggests that the early Buddhist problem with the statement “the self does not exist” is that although it claims to report an ultimate truth (a state of affairs that is true independent of human consciousness), both the notions “self” and “existence” have no reality beyond particular cognitive conditions. And if the terms “existence” and “self” are conceptual constructions rather than ultimate truths, the statement “the self does not exist” cannot describe the way things really are, and must therefore be avoided.

Radical as this idea may seem, the notion that “existence” does not pertain beyond human thought fits the Mahānidāna Sutta’s critique of personal identity very well. As we have seen, the second of these critiques denies the notion of intrinsic identity (attan) in a state beyond conditioned experience. Although it would be easy to conclude from this that there is no self beyond conditioned experience, the teaching steers away from drawing an ontological conclusion of this sort. It instead asks if a person can have the thought “I am” (asmī ti) beyond sensation (vedayita), and since this is not so it concludes that a person’s identification with a transcendent identity as “my self” (me attā ti) is impossible. The point of this is that the very notion of individual identity (attan) depends on particular cognitive events, i.e. the sensations that, arising in the process of conditioned experience, lay the cognitive foundations for self-consciousness (“I am”). In other words, in the process of conditioned experience arises the notion of individual existence, and dependent on that arises the notion of intrinsic identity. But both have no reality beyond the conditioning of human consciousness. The ultimate truth, then, is not that the “self” does not “exist,” but rather that the very notions of individual existence and intrinsic identity are dependently originated, and thus that any articulation of such concepts cannot be ultimately true; such concepts do not correspond to the way things really are. Exactly this point about the dependent origination of consciousness is made in the Brahmajāla Sutta, the first Sutta of the Dīgha-Nikāya.

This discourse presents and criticises an extensive list of views, attributed to various unnamed ascetics and Brahmins, concerning the ultimate nature of the human being (attan) and the world (loka). The large number of views (sixty-two in total) and the complex manner of their presentation can obscure the ultimate point of the critique. Hayes, for example, has focused on the therapeutic aspects of the text and so missed the philosophical point entirely. ⁷⁶ The philosophical point has also been missed by Fuller, who had synthesised the Brahmajāla Sutta with the Sammāditthi Sutta and concluded that the text’s critique of views is concerned only with the “knowledge of the cessation of craving.” ⁷⁷ Even Collins, who has noted that the focus of the Brahmajāla Sutta is the conditioned status of views, believes that this means nothing more than that the views are conditioned by the craving of their proponents. He thus comments that “It is here, par excellence, that the argumentum ad hominem, the denigration of others’ views on the ground of the character of those ‘others,’ and the argumentum ad verecundiam, the appeal to feelings of reverence and respect (for the Buddha), can be seen in Buddhist thinking. ⁷⁸ Although it is true that the ascetics and Brahmins of the Brahmajāla Sutta are criticised for their craving, the focus is more on the cognitive foundations of their views rather than their affective faults. Moreover, the text does not praise the Buddha simply because of his virtuous character, but also because of his understanding and transcendence of views: faith plays no role in this discourse at all.

None of these studies of the Brahmajāla Sutta, nor even that of Rhys Davids (1899: xxv-xxvii), Bhikkhu Bodhi (1978) or the recent analysis of Evans (2009), comment on the fact that the views it criticises are analysed in terms of their connection with the past (pubbanta) or future (aparanta). Although it is easy to overlook this presentation of views according to their temporal significance, it is fundamental to the text’s analysis. Time is an essential aspect of human experience, it being impossible to conceive the fundamental reality of the human being and the world in which he exists apart from the notion of past, present and future. If so, the presentation of views in temporal terms surely indicates that the point is to criticise any attempt to conceptualise the existential reality of the human being and the world. The following statement, occurring immediately after the presentation of views, makes this quite clear:

Whatever ascetics or Brahmins speculate and form views about the past, future or both, declaring various sorts of opinion (adhivuttipadāni) with reference to the past and future, all do so through these sixty-two points or one of them - there is no possibility besides this. ⁷⁹

Whether or not one accepts the comprehensiveness of the views presented, the logic of this statement is that any attempt to understand the reality of the human being and the world in terms of the past and future falls within its critique. This means, then, that the text rejects the notion that the ultimate reality of things can be understood in terms of the concept “time.” For if this were not the case, it would surely be possible to conceive the reality of inner and outer things in terms of some sort of temporal analysis, i.e. another “possibility besides this.” And if the human being and the world cannot be understood in terms of time, their true reality cannot be conceptualised at all. The presentation of views in terms of the past and future is no accidental or convenient way of ordering ideas, then, but is intended to show that any attempt to conceptualise the ultimate existential reality of the human being and the world is impossible, and that the Buddha’s liberated understanding - to which the views are eventually contrasted - is beyond all notions of

There are, bhikkhus, other matters - profound, hard to see and understand, tranquil, supreme, beyond the scope of logic, subtle, to be known by the wise - that the Tathāgata declares through his own understanding and vision. It is because of these that those who speak correctly would speak the true praise of the Tathāgata. ⁸⁰

The account of views is thus intended to elucidate the nature of the Buddha’s awakened understanding through a simple contrast: whatever the Buddha understands, it is entirely different from any attempt to conceptualise the way things are in terms of the past and future. In fact the sixty-two views are dominated by attempts to understand the human being and the world in spatio-temporal terms. This is most obvious with regard to the views about the world, all of which concern its spatial or temporal limits: the world is imagined to have a beginning or not (and so be eternal), or to be with or without spatial limits, or else to be a mixture of the two. ⁸¹ The views about the human being are also concerned with the spatiotemporal reality of individual existence: they either comment on the ultimate temporal existence of a person (e.g. that there is an eternal but transmigratory self, ⁸² or that the self has a beginning since it comes into existence spontaneously, ⁸³ or that the human being finds ultimate felicity within the bounds of this life), ⁸⁴ or his spatial existence (e.g. that there is a self consisting of consciousness within the body, this being the essential subject that experiences sensations), ⁸⁵ or else his spatio-temporal existence (e.g. that the true self - however imagined - is realised after death, ⁸⁶ or that the different constituents of individual existence cease at death). ⁸⁷ The text thus contrasts what it imagines are all conceivable ideas about existence in time and space with the Buddha’s understanding that is “beyond the scope of logic” (atakkāvacara). The text goes on to explain this awakened understanding as follows:

The Tathāgata understands all these [views as follows], bhikkhus: ‘These points of view thus seized and grasped will have such a destiny and such an outcome.’ The Tathāgata understands this, and he understands what is beyond it (uttarītara), but he does not grasp at this understanding, and through not grasping he experiences internal quenching. Having comprehended as it really is the rise and fall of sensations, as well as the pleasure and danger in them and the release from them, the Tathāgata is released without grasping, bhikkhus. ⁸⁸

Rather than present another view about the spatio-temporal reality of the human being or the world, it seems that the Buddha understands “what is beyond” such conceptualisations and so is able to say something objective about them. This objective understanding takes two forms: first, the Buddha understands the effects of the various views in the form of the continued existences to which they lead; and second, the Buddha understands the structure of conditioned experience (the “rise and fall of sensations”), i.e. the cognitive conditions under which views arise. It is the latter aspect of the Buddha’s understanding - the construction and limits of views - that is taken up in the remainder of the text. It is first pointed out that the affective and cognitive state of the various “ascetics and Brahmins” who hold views renders their understanding of primary experience unreliable:

Therein, bhikkhus, whatever ascetics or Brahmins form ideas about the past, future or both, and have views about them, making all sorts of claims by means of these sixty-two statements with reference to the past and future, this is because (tad api) what is sensed (vedayita) by these venerable ascetics and Brahmins, who have no knowledge and vision and are affected by thirst, is subjected to ‘trembling’ and ‘quivering’ (paritasita-vipphandita). ⁸⁹

Although the grammar of this statement is complex, the terms paritasita and vipphandita indicate that views depend on the cognitive processing or elaboration of primary experience (vedayita).⁹⁰ That this cognitive elaboration is spoken of in terms of “trembling” and “quivering” (paritasita-vipphandita) suggests, in fact, that a person’s conceptual grasp of things is in fact a sort of “distortion” of primary experience. Indeed we have seen that the term phandita is synonymous with the term papañcita in the Vepacitti Sutta, where both define the various forms of the notion “I am.” The Brahmajāla Sutta’s critique would thus seem to be concerned with the cognitive differentiation or conceptual proliferation of a person’s basic experience. Just as the Vepacitti Sutta describes the notion “I am” as a “palpitation” or “distortion,” so the Brahmajāla Sutta states that the same is true of all possible views about individual existence in space-time. The text goes on to point out that these statements of view ultimately depend on sense contact (DN I.43.8: tad api phassapaccayā), and that experience only comes about under these particular cognitive conditions (DN I.44.30: te vata aññatra phassā paṭisaṃvedissantī ti n’ etaṃ ṭhānaṃ vijjati). All ideas about the spatio-temporal reality of the human being and the external world are therefore contingent, the implication being that should the cognitive conditions change so too would a person’s grasp of reality. This is exactly what the text implies has happened to the Buddha: his cognitive state and subsequent grasp of reality are so utterly different that it is impossible to capture in terms of ideas about existence in time and space.

Views about the human being and the world are therefore not absolute: they are ideas that pertain only under particular cognitive conditions, those that come about in the process of conditioned experience. It follows from this that the entire content of human consciousness - notions of personal identity, existence, non-existence and even space-time itself - are real only in so far as the cognitive conditions for them pertain, and have no essential reality beyond this dependently originated state of consciousness. ⁹¹ In other words, what human beings assume to be objectively true facts about reality are nothing of the sort. ⁹² The Brahmajāla Sutta thus articulates a philosophy of epistemological conditioning and its transcendence through the doctrine of the dependent origination of consciousness. This constitutes, in other words, a rejection of philosophical realism, as has been pointed out by Ronkin:

What the Buddha rejects is realism, conceptual and ontological alike: the notion that the encountered world is made up of distinguishable substances, and the linguistic theory that words refer to these substances which they represent; the conviction that our language corresponds to or mirrors a mind-independent reality. He points towards conventionalism in language and undermines the misleading character of nouns as substance-words. Whatever we know is part of the activity of language, but language, by its very nature, undermines certified knowledge. ⁹³

The Brahmajāla Sutta’s philosophy of epistemological conditioning explains the early Buddhist teachings on personal identity very well: it explains the dependent origination of notions of individual existence (as found, for example, in the Vepacitti Sutta), and indicates why the Mahānidāna Sutta and the Not-Self teaching do not deny the self’s existence outright. Furthermore, the position of the Brahmajāla Sutta at the beginning of the Dīgha-Nikāya surely indicates that it is foundational for the early texts. This suggests that the philosophy of epistemological conditioning ought to be generalised to the entire edifice of early Buddhist thought, and since this perspective explains the early teachings on personal identity, there can be little doubt that virtually all these teachings fit into a homogeneous understanding, one that can be ascribed to the same thinker(s) or period of thought. This is also indicated by the fact that the Not-Self teaching and the Mahānidāna Sutta both use early Upaniṣadic thought to elucidate new ideas: the Not-Self teaching alludes to the Upaniṣadic ātman in its transcendent or macrocosmic aspect, whereas the Mahānidāna Sutta uses the dialogue between Prajāpati and Indra (Chāndogya Upaniṣad VIII) in order to criticise the Upaniṣadic ātman in both its microcosmic and macrocosmic aspects (as an inner perceiver and transcendent essence). Furthermore, the philosophy of epistemological conditioning is complemented by a consistent understanding of religious means and ends. This understanding can be discerned in most of the texts studied above, especially those passages of Brahmajāla Sutta that deal with the Buddha’s transcendent state and the insight thought to effect it. This doctrine of transcendence is not confined to the texts on personal identity, however, but is assumed by some of the most important early Buddhist teachings.

6. The transcendence of cognitive conditioning

The means of attaining liberation suggested by the Brahmajāla Sutta consists of a discipline in which a person directs attention towards the process of conditioned experience (“the rise and fall of sensations”). A similar means is envisaged by both the Khemaka and Aggivacchagotta Suttas (“form is thus, its arising is thus, its fading away is thus” etc.), this being a more experiential elaboration of the Not-Self teaching: the goal of both is the correct comprehension of conditioned experience. Such a discipline is, of course, consistent with the Brahmajāla Sutta’s philosophy of epistemological conditioning. For if a person’s experiential condition and the suffering encountered therein is ultimately due to the dependent origination of consciousness, a person must understand how this state of affairs comes about in order to be released from it. The end result of such practices is the attainment of a transcendent state, that which the Aggivacchagotta Sutta states is beyond concepts (maññita) and self-consciousness (the notions “I” and “mine:” ahiṃkāra, mamimkāra). The Brahmajāla Sutta similarly understands that this state is beyond views (tato utarītaraṃ) and the scope of logic (atakkāvacarā). Although such statements on religious means and ends indicate that the liberated state is beyond conceptualisation, they also imply that it is beyond conditioned experience per se. Indeed the Brahmajāla Sutta states that the Tathāgata is released not only because he “understands as it really is” (yathābhūtaṃ viditvā) the rise, fall, pleasure and danger of “sensations,” but also because he undestands the release from them (nissarana). ⁹⁴ This point is made explicit towards the end of the text when a bhikkhu is said to be liberated by understanding the “rise, fall, pleasure, danger and release from the six spheres of contact.” ⁹⁵ Perhaps the meaning of this is that in such a mindfully aware person, there is no distortion of primary experience. The implication of this is that transitive consciousness (viñ̃̃āna) must be transcended too, since this is an essential aspect of sense-contact (phassa). ⁹⁶ This idea is expressed in the Alagaddūpama Sutta - certainly the most important canonical discourse on the Not-Self teaching ⁹⁷ - when the Buddha describes the transcendent state of the liberated bhikkhu who has understood the Not-Self teaching as follows:

Therefore, bhikkhus, I say that when Indra, Brahma, Prajāpati and the gods search for the bhikkhu thus released in mind (evam vimuttacittam), they cannot establish that ‘the consciousness of the Tathāgata is located here.’ What is the reason for this? As soon as the teaching is realised, bhikkhus, I say that a Tathāgata is untraceable. ⁹⁸

This statement would seem to indicate the understanding that the liberated bhikkhu is devoid of transitive consciousness, this being the reason for the failure of the gods to locate him. Such a statement on the inability to find the liberated bhikkhu is akin to the notion that the Tathāgata is indefinable, as implied, for example, by the refusal to answer certain questions about his existential state. These questions form the final four of the well-known set of ten unanswered (avyākata) questions: the first four concern the eternality (or not) and finitude (or not) of the world, ⁹⁹ the next two ask whether the soul or life principle (jīva) is the same as the body (or not), and the final four are concerned with the Tathāgata’s existential status after death (whether he exists, does not exist, both exists and does not exist, or neither exists nor does not exist). ¹⁰⁰ According to Collins these questions are not answered since they are “linguistically ill-formed,” the problem being that they use

personal referring terms, which according to Buddhist thinking have no real referrent; hence, any answer given directly to them would confirm the misleading presupposition that such terms do refer to some real individual. ¹⁰¹

In explaining that the problem posed by these questions is that there is “no real referrent” to terms such as “soul” and “Tathāgata,” Collins reads the classical “No Self” doctrine into the Buddha’s failure to answer them. On the other hand, however, he indicates that the failure to answer questions about the ontology of the world is based on a different reason, i.e. that they are pragmatically pointless. ¹⁰² This explanation thus assumes the classical ontology in which the term “world” refers to a reality independent of consciousness, whereas personal terms have no ultimate referent in this world. This is problematic not only because such an ontology is not made clear in the early texts, but also because a common set of questions is explained differently. A single explanation should be found for all the questions: if there is a “linguistic problem” with the questions, as maintained by Collins, this should similarly apply to all the points they cover.

A single, coherent explanation for the linguistic problem is provided, however, by the Brahmajāla Sutta’s philosophy of epistemological conditioning. According to this philosophy, all linguistic formulation - even basic concepts such as “space-time” and “existence” - have no reality beyond a person’s dependently originated state of consciousness. According to this perspective, the problem with questions about the ontology of the world and the human being is they that assume the mind-independent reality of “spacetime” and “existence,” and to answer them would be to subscribe to such notions. But as Hamilton has pointed out, the unanswered questions are based on false premises:

If, however, … space and time are part of the structural characteristics of the experiential world, and that this is cognitively dependent, then one can see that the presupposition of the transcendental reality of time and space is false, and that the fundamental premises on which the questions rest are therefore also false. What this means is that though the questions are meaningful within a conceptual framework which assumes that space and time are transcendentally real, if space and time are not transcendentally real the questions are in effect unanswerable if one wishes to be truthful. Any formulation of a response within the same conceptual framework as the questions would not truthfully reflect a reality which does not conform to that conceptual framework. ¹⁰³

The unanswering of these questions thus indicates the Buddha’s transcendence of dependently originated states of consciousness, such that notions of “space-time” and “existence” have no ultimate reality to him. It would seem, then, that the philosophy of the Brahmajāla Sutta thus provides a coherent explanation for all of the unanswered questions. Indeed this understanding of the unanswered questions is articulated in at least two canonical texts. In the Avyākata Vagga of the Añguttara-Nikāya (Sattaka Nipāta 51), the Buddha explains that each of the final four of these questions is a view endowed with thirst (taṇhāgata), apperception (or ideation: sañ̃̃āgata), conceptualisation (maññita), conceptual diffuseness (papañcita), attachment (upādānagata) and regret (vippaṭisāra). ¹⁰⁴ In other words the unanswered questions are conceptualisations that have no correspondence with reality. A similar explanation is found in the Mahānidāna Sutta. After describing the liberation of the bhikkhu who understands its teachings on personal identity (those studied in section four above), ¹⁰⁵ it states that it is unsuitable (akalla) to consider this bhikkhu “thus liberated in mind” (evam vimuttacittam) in terms of questions about his existence, non-existence, existence and non-existence, and neither existence nor non-existence, i.e. the final four of the ten unanswered questions. ¹⁰⁶ The reason for this is given as follows:

The extent of articulation and its range, of utterance and its range, of designation and its range, of understanding and its scope, of existence and its ‘movement’ - the bhikkhu is released from all this through higher understanding. ¹⁰⁷

This account of transcendence is in agreement with the Brahmajāla Sutta’s philosophy of epistemological conditioning outlined above: questions about the ontological state of the liberated bhikkhu cannot be answered because he has transcended the cognitive conditions on which they are founded. In other words, the problem here is with conceptuality per se: the liberated bhikkhu transcends the “path of articulation” (adhivacanapatha), “the path of designation” (paññattipatha), “the scope of understanding” (paññāvacaram) and even the notion of “existence” (vatṭa) and its “movement” (vațtaṃ vattati), i.e. time. Indeed, the equating of the two latter concepts (“existence” and “time”) with those of “articulation,” “designation” and “understanding” indicates the understanding that space-time is only conceptually real.

That the problem with the unanswered questions is as much with the notion of “existence” as it is with the notion of “personal referring terms” is similarly suggested in the Aggivacchagotta Sutta. For this text likens the inexplicability of a Tathāgata’s liberated condition to that of an extinguished flame, the point being that the fuel through which both can be designated - the five aggregates for the Tathāgata, grass and firewood for the flame - has ceased. The Tathāgata cannot be conceptualised, therefore, because he has gone beyond the experiential conditions through which an individual is normally understood (the five aggregates). And in this transcendent condition, the notions of “arising” (or being reborn: upapajjati), non arising, both arising and non-arising, and neither arising nor non-arising have no relevance. This account indicates that the conceptual problem is not simply one of defining the Tathāgata, but in conceiving his individual existence.

A similar problem with applying the concept of “existence” to the liberated person can be seen in the Buddha’s dialogue with Upasīva in the Pārāyanavagga (Sn 1069-1076): in response to Upasiva’s question about whether the liberated sage (muni) exists eternally or does not after death (Sn 1076), the Buddha states that he cannot be measured (na pamānam) because the means of speaking about him have ceased. ¹⁰⁸ In other words the liberated sage has gone beyond the concepts by which the existence of anything can be known (yena nam vajju tassa natthi).

What all this means is that the apophatic strand strongly evident in early Buddhist thought can be explained by the notion that space-time is a relative truth transcended by a Tathāgata, the one whose condition is therefore incomprehensible, i.e. “like that.” ¹⁰⁹ The philosophy of epistemological conditioning and its transcendence thus explains some of the more puzzling aspects of early Buddhist thought: critiques of personal identity that do not commit to any ontology, the refusal to comment on the ultimate reality of the world, the avoidance of questions about the existential status of the liberated sage, the simile of the extinguished flame and so on.

The Brahmajāla Sutta therefore provides a coherent explanation for the early Buddhist teachings on personal identity as well as the apophatic strand in early Buddhist thought, and so underpins a religio-philosophical understanding found consistently and extensively throughout the early Buddhist texts. Although this is not a philosophy of realism since it assumes that the reality of space-time does not extend beyond a person’s cognitive conditioning, this does not imply an idealist understanding. For idealism is still an ontology of sorts, and indeed one that can only be imagined under particular cognitive conditions; the authors of the Mahānidāna Sutta would no doubt object to this by pointing out that there would be no means of conceiving such an understanding beyond “sensation” and self-consciousness. Indeed idealism is, basically, an hypostasisation of a person’s subjective awareness into a mind-independent, ultimate reality, an idea which the Mahānidāna Sutta rejects of this system of thought is rather that the way things really are is unspeakable and unthinkable. In other words, the Brahmajāla Sutta’s philosophy of epistemological conditioning implies that reality is ultimately ineffable, as is the state of the person who realises it by escaping his cognitive conditioning.

7. The development of reductionistic realism: Abhidharma origins

This conceptual clarification allows us to see that the metaphysical assumptions of the Brahmajāla Sutta differ considerably from those of the first attempt to systematise the Buddha’s teachings, i.e. the reductionistic realism articulated in the various Abhidharmas. This reductionism is based on the “No Self” doctrine: it assumes that although a person “exists” in the mind-independent reality of the world, he is made up of impermanent “existents” or “events” (dharmas) which lack self. The human being and the world are reduced to their constitutent parts, therefore, these being thought to lack essence but exist transiently in the objectively real domain of space-time.

Traces of the change towards a proto-Abhidharma reductionism can be found in the early texts. It is clear, for example, that the Vajirā Sutta is both reductionistic as well as realistic, for it speaks of the aggregates “existing” (khandhesu santesu) and of the failure to “find” an essential being in them (na yidha sattūpalabbhati). This goes beyond the Not-Self teaching and the philosophy of the Brahmajāla Sutta by assuming that ultimate truth can be spoken of in terms of the concepts “existence” and “non-existence.” If so, it follows that the Not-Self teaching and the Vajirā Sutta are separated by an important philosophical change: whereas the former is based on the doctrine of epistemological conditioning and the relative truth of space-time, whereas the latter is based on the realistic assumption that space-time exists independent of human consciousness. This change in thought was probably complex and multi-faceted, but in principle can be simply explained. For it requires only that certain Buddhist thinkers focused on the Not-Self teaching (and related ideas) at the expense of the philosophical framework provided by the Brahmajāla Sutta. In such a scenario the Not-Self teaching could easily have been taken in a realistic sense, and this would have set the foundations for the emergence of the Abhidharmic reductionism.

The beginnings of such a development can perhaps be seen in the Khemaka Sutta. As we have seen, the bhikkhu Khemaka was of the opinion that “I see no sort of self (attan) or its property (attaniya) in these five aggregates of attachment, venerable sirs” (imesu khv āhaṃ āvuso pañcasu upādānakkhandhesu na kiñci attānaṃ vā attaniyam vā samanupassāmi). Although this does not indicate any departure from the philosophy of the Brahmajāla Sutta, it certainly paves the way for a new sort of enquiry, one in which the contemplative bhikkhu analyses the different aspects of his being in the attempt to find a self. Such an approach leaves Buddhist thought on the verge of an important conceptual change: through this enquiry, factors of experience (the five aggregates) which were thought unsuitable to be regarded as “self” begin to look like impermanent factors of being which lack self. The change from the teaching that “form is not self” (this being an unsuitable conceptualisation) to the similar but subtly different idea that “no self can be found in form” (the latter being something that exists whereas the former does not) is easy to imagine, therefore, on the basis of the Khemaka Sutta. Textual evidence for this change is in fact contained in the Mahāhatthipadopama Sutta. This discourse begins with the simile of an elephants’ footprint but soon turns into an analysis of the first aggregate - form - in terms of the four material elements (earth, water, fire and wind). The Not-Self teaching is applied to these four material elements in the following manner:

What, venerable sirs, is the earth element? It might be internal or external. And what is the internal earth element? That which is internal and personal, i.e. that which is solid, hard and materially derivative, namely: head-hair, bodily hair, nails, teeth, skin, flesh, sinew, bones, bone-marrow, kidney, heart, liver, membrane, spleen, lungs, bowels, intestinal tract, stomach, faeces, and whatever else is internal and personal, i.e. that which is solid, hard and materially derivative, this, venerable sirs, is said to be the internal earth element. This very internal earth element and the external earth element are simply the earth element, which should be seen with correct understanding as it really is: “This is not mine, I am not this, this is not my self.” Once this has been seen with correct understanding as it really is, one becomes disillusioned with the earth element, one cleanses one’s mind of passion for the earth element. ¹¹⁰

This teaching does not focus on the process of conditioned experience and its conceptual appropriation, but is rather concerned with a virtually exhaustive analysis of the physical constituents of a human being. ¹¹¹ This is a considerable departure from the Not-Self teaching. The point is no longer that the concept “self” arises in connection with a dynamic process of experience and is ill-suited to it, but rather that when a human being is broken down into his constituent parts all is found to be lacking in self. The Mahāhatthipadopama Sutta thus assumes a realist ontology in which the four elements really “exist” but the self does not. Furthermore, the ultimate truth of things is here captured in words, and is not something beyond logic and the conceptual construction of consciousness, as stated in the Brahmajāla Sutta. Concepts thus capture the fact that the human being really exists in the mindindependent reality of space-time, albeit as an aggregate of various elements in which a self cannot be found. The simile of the house, found towards the end of the text, makes this reductionistic realism quite clear:

Venerable sirs, just as an enclosed space is designated ‘house’ dependent on logs, creepers, grass and clay, so too is an enclosed space designated ‘form’ dependent on bones, sinew, flesh and skin. ¹¹²

This is a statement of the ultimate truth that a person’s physical being is nothing more than an accumulation of different parts in which, as the text goes on to state, are to be found sensation, apperception, volitions and consciousness. ¹¹³ The reference to an enclosed space (ākāso parivārito) shows that the authors of this text regarded the human being as a construction in space-time, albeit one that lacks intrinsic identity, an idea that comes very close to the chariot simile of the Vajirā Sutta. The truth to be known is here that the human being exists but is an aggregate lacking essence, a fact that the bhikkhu should come to understand by the following means:

He understands thus: ‘Thus indeed is the coming togther, collection and accumulation of the five aggregates of attachment.’ But the Blessed One has said this: ‘The one who sees Dependent Origination sees the Dhamma, and the one who sees the Dhamma sees Dependent Origination.’ These very things are dependently originated, that is to say the five aggregates. ¹¹⁴

In this passage the five aggregates (upādāna) really are “aggregates,” i.e. parts out of which a person is made, these being impermanent (dependently originated) and so not intrinsically real. This shows that a realistic reading of the Not-Self teaching did indeed lead to reductionism, i.e. the notion that a person is an accumulation (samavāya) of impermanent, causally-connected elements. This understanding is entirely different from that articulated in the Brahmajāla Sutta and most of the other texts on personal identity: there is no trace in the Mahāhatthipadopama Sutta of the notion that the dynamic processes of conditioned experience is not to be understood in terms of intrinsic identity. The point, rather, is that intrinsic identity cannot be found in the different existential factors of the human being. How is this conceptual divergence to be explained?

Whereas the philosophy of epistemological conditioning can be generalised to a great number of early texts, the same cannot be said of the reductionistic realism espoused in the Vajirā and Mahāhatthipadopama Suttas. This indicates that the ideas of the latter texts were marginal in the early period, which can only be explained in two possible ways: either these ideas circulated among a small sub-section of the early sangha, and so were not recorded in most of the early texts, or they belong to a later stage of speculation than that recorded in most of the early texts. In support of the latter hypothesis, we can note that the Atṭhakavagga and Pārāyanavagga, which are certainly among the earliest Buddhist texts, are generally in agreement with the philosophy of epistemological conditioning. ¹¹⁵ In the Kalahavivāda Sutta, for example, the Buddha states (Sn 870) that existence and non-existence depend upon sense contact:

The pleasant and unpleasant originate in sense-contact, but do not arise when there is no sense-contact. I say to you that the fact of nonexistence and existence also originates in this. ¹¹⁶

Later on in the same dialogue ( Sn 874 ) the Buddha explains that “form” (rūpa) is a conceptual proliferation that depends on conceptualisation (or apperception, saññā) and can thus disappear for the religious adept:

Not cognisant of conceptualisation, not cognisant of misconceptualisation, not uncognisant but not cognisant of what is untrue: form disappears for the one who has reached this state, for the discernment of manifoldness (papañcasañkhā) originates in conceptualisation (saññānidānā). ¹¹⁷

According to this enigmatic statement of the Buddha, a person’s physical being is not ultimately real, but depends on the tendency to conceptualise reality in terms of a manifold world of diversity. Elsewhere in the Atṭhakavagga, the seeds of the Buddha’s teachings on personal identity can be seen at the beginning of the Tuvataka Sutta (Sn 915-916), when the Buddha responds to a question about the attainment of Nirvana:

‘I ask you, kinsman of the sun, great sage, about detachment and the state of peace: with what sort of vision is a bhikkhu quenched, so that he grasps at nothing in the world?’ (915)

The Blessed One said: ‘The contemplative (mantā) should put a complete stop to the notion ‘I am’ (asmī ti), which is the root cause of discerning manifoldness (mūlaṃ papañcasañkhāyā). He should ward off whatever inner thirst he has, training himself to be ever mindful.’ (916)¹¹⁸

The notion of individual existence is here said to be the root cause of a person’s diverse perceptions (papañcasañkhā), which is simply a more emphatic way of stating the teaching of the Vepacitti Sutta, i.e. that the notion “I am” is a conceptual proliferation (papañca).

Besides these Atṭhakavagga teachings on existential matters, in the Purābhada Sutta the Buddha explains (Sn 849) that the liberated sage is released from the very notion of time:

‘Devoid of thirst even before death,’ said the Blessed One, ‘not dependent upon the past, immeasurable in the middle, for him nothing is fashioned with regard to the future. ¹¹⁹

The statement that the liberated sage is “immeasurable” in the present is but a poetic way of describing the complete transcendence of time. Indeed in the Atṭhakavagga the notion of being “independent” or “unnattached” (anissita) can refer not just to having no inclination or fondness for something, but also to being completely devoid of the notion of something. When the Paramatṭhaka Sutta states that the bhikkhu should have no dependency on knowledge (Sn 800: ñāne pi so nissayaṃ no karoti), for example, the point is that he should transcend it. Indeed the Kalahavivāda Sutta claims that those who have different opinions about liberation are “dependent” (Sn 877: upanissitā), whereas the Buddha is released, the implication being that he is completely beyond such views. It seems, then, that when the Purābheda Sutta states that the sage is not attached (anissito) to the past, it means that he is free from the very notion of it. Indeed this text goes on to state ( Sn 851 ) that the sage’s freedom from the past and future is connected to his transformed cognitive state and lack of views:

He is without attachment for the future and does not grieve over the past. Perceiving detachment, he is not led into sense-contacts and views. ¹²⁰

All of this evidence is in agreement with the Brahmajāla Sutta’s philosophy of epistemological conditioning: notions of existence, non-existence and time are said to be dependent on a person’s cognitive functioning, the release from which implies the cessation of a person’s awareness of individual existence in space-time. There are reasons for believing that such teachings go back to the Buddha himself, not only because of the Atṭhakavagga’s antiquity and originality, ¹²¹ but also because related teachings in the Pārāyanavagga correspond very closely to the few historical facts about the Buddha that can be deduced from the early literature. ¹²² Similar observations suggest that the Not-Self teaching is also to be ascribed to the Buddha, for it is hard to explain this highly original teaching - especially its occurrence in Alagaddūpama Sutta - as an abstract formulation of later Buddhist teachers. ¹²³ A similar antiquity cannot be assumed of the Vajirā and Mahāhatthipadopama Suttas, however. It is surely important that the Buddha does not feature in either text, and that the orator in the Mahāhatthipadopama Sutta is Sāriputta, the patron saint of the Abhidharma. Furthermore, in its statement that ” … the Blessed One has said this: The one who sees Dependent Origination sees the Dhamma …,” the Mahāhatthipadopama Sutta even seems to speculate on the meaning of the Buddha’s teaching of Dependent Origination. This text’s highly complex and artificial style of analysis also suggests that it is a proto-Abhidharmic work, a text composed as Buddhist thought progressed towards a fully developed philosophy of reductionistic realism.

Further evidence in support of this chronological stratification is found in the various Vinayas accounts of the beginning of the Buddha’s ministry. These texts state that the Buddha’s teaching to the five disciples concluded with the Not-Self teaching, the understanding of which triggered their instantaneous liberation. But it can be shown that the account as a whole draws upon and adapts the earlier account of the Ariyapariyesana Sutta. ¹²⁴ The doctrinal understanding that underpins this adaptation can be seen at the conclusion of the Second Sermon, when it is stated that the minds of the five disciples of the Buddha were “released from the corruptions without grasping” (Vin I.14.34: pañcavaggiyānaṃ bhikkhūnaṃ anupādāya āsavehi cittāni vimuccimsu). Although this formulation of liberating insight occurs throughout the early Buddhist literature, it deviates from the Ariyapariyesana Sutta’s account of how the five bhikkhus “realised Nirvana” (MN I.173.7ff.: pañcavaggiyā bhikkhū … asaṅkilitthaṃ anuttaraṃ yogakkhemaṃ nibbānaṃ ajjhagamaṃsu). The authors of the Second Sermon thus drew upon the Ariyapariyesana Sutta but replaced its apophatic account of the five bhikkhus’ liberation with an apparently reductionistic one, the subject of liberation being the minds of the bhikkhus rather than the bhikkhus themselves. This deviation from an older, apophatic description, one that sees no problem in speaking of the human being as a whole realising Nirvana, and its replacement with a reductionistic formulation - which in Buddhist literature past and present complements the “No Self” doctrine - implies that the authors of the Second Sermon believed in the non-existence of the self. It would even seem that in this case the authors of this Vinaya account read the No Self doctrine into the Not-Self teaching. This is strong evidence for a doctrinal change from a philosophy of ineffability to that of reductionistic realism.

Regardless of the strength or weakness of this theory of doctrinal change, a couple of facts can hardly be denied. First, a doctrine of ineffability is certainly contained in the early Buddhist texts, and attempts to deny this are implausible. ¹²⁵ And second, this philosophy is incompatible with the philosophy of reductionistic realism later outlined in the various Abhidharmas, and anticipated in the Vajirā and Mahāhatthipadopama Suttas. It can hardly be the case that both philosophies were devised by a single thinker or the same group of thinkers. One must have developed from the other, and if the philosophy of epistemological conditioning and its transcendence can be ascribed to the very beginnings of Buddhism and perhaps to the Buddha himself reductionistic realism must belong to a later period. This is in fact the most logical explanation of the evidence. The philosophy of epistemological conditioning radically subverts all our most basic presuppositions about life: it is difficult to believe that space-time is merely conceptual rather than an objective, mind-independent reality. It is not hard to imagine that this challenging philosophy was misunderstood and replaced by a sophisticated but simpler realistic philosophy. Indeed teachings based on the philosophy of epistemological conditioning, such as the Mahānidāna Sutta’s three critiques of personal identity, could easily be misunderstood in a realistic manner unless the underlying philosophy is made clear. But it is hard to see how the philosophy of epistemological conditioning could have emerged from that of reductionistic realism.

In conclusion, the evidence studied here suggests that the difference between the Not-Self teaching and the Vajirā Sutta is philosophical rather than terminological. It follows from this that although the early Buddhist teachings were not presented in the form of a philosophical system, they are at least philosophically grounded. This is not to say that the Buddha should be regarded as a philosopher, however. ¹²⁶ For although the evidence suggests he had worked out a coherent world-view, it also indicates that he applied it as and when he saw fit, i.e. according to the pragmatic demands of the situation. ¹²⁷ Such an interpretation seems to fit with one of the Buddha’s most famous statements on his approach to teaching: in the fifth book of the Saṃyutta-Nikāya (Saccasaṃyutta IV: Simsapāvanavagga, no. 1), the Buddha states that although his knowledge is as vast as the numbers of leaves in a forest grove, the teachings he had revealed were comparable to just a few leaves. ¹²⁸ According to the texts studied above, this statement would seem to mean that although the Buddha had worked out a coherent philosophy, he did not teach it directly because of his pragmatic interest in helping others attain the cessation of suffering.

This explains why the most important teachings studied above, such as the “Not-Self” teaching and the Mahānidāna Sutta’s critiques of personal identity, are philosophical in their method and argumentation but ultimately avoid a direct statement of philosophical truth. The same is true of the Brahmajāla Sutta: it is philosophical without stating a philosophy directly. All this seems to be the work not of a philosopher interested in abstract ideas for their own sake, but rather of a religious teacher keen to apply a philosophical understanding in order to help his followers achieve the best possible spiritual result.

Early Buddhist thinkers were less philosophically parsimonious, however. In contemplating the Not-Self teaching, they came to believe in the non-existence of the self - against the explicit warnings of the Buddha. For in the Alagaddüpama Sutta, perhaps the single most important canonical exploration of the Not-Self teaching, the Buddha describes how others responded to his teachings by weeping, beating their breasts and thinking ‘I will be annihilated!’. ¹²⁹ Such people concluded that the Buddha had taught the non-existence of the self, although the Buddha rejected this charge. ¹³⁰ It is ironic that within probably a few generations of his death, the Buddha’s followers had drawn exactly the same conclusion, even if they did so with a little more composure and meditative calm.

Abbreviations

All Pāli citations are from Pali Text Society editions.

AN	Añguttara-Nikāya
AV	Atharva Veda
BCA	Bodhicaryāvatāra (see Tripathi)
BU	Bṛhadāraṇyaka Upaniṣad (see Olivelle)
CSCD	Chattha Sañgāyana: CD-ROM version of the Burmese Tipiṭika, Rangoon 1954. Dhammagiri: Vipassana Research Institute, version 3.
CPS	Catuṣpariṣatsūtra (see Waldschmidt)
DN	Dīgha-Nikāya
MN	Majjhima-Nikāya
MMW	A Sanskrit-English Dictionary, Monier Monier-Williams. Oxford: Clarendon Press (1899).
Mvu	Mahāvastu (see Senart)
OED	Oxford English Dictionary, Second Edition, prepared by J. A. Simpson and E. S. C Weiner. Oxford: Clarendon Press (1989).
PTS	Pali Text Society
RV	Ṛg Veda
SN	Saṃyutta-Nikāya
SbhV	Sañghabhedavastu (see Gnoli)
Ud	Udāna
Vin	Vinaya
Vism	Visuddhimagga (see Warren and Kosambi)

References

• Bodhi, Bhikkhu. 1978. The All Embracing Net of Views. The Brahmajāla Sutta and its Commentaries. Kandy: Buddhist Publication Society.
• Bodhi, Bhikkhu. 2000. The Connected Discourses of the Buddha. A New Translation of the Samyutta Nikāya (2 Volumes). Oxford: Pali Text Society.
• Bronkhorst, Johannes. 2007. Greater Magadha. Studies in the Culture of Early India. Leiden, Boston: Brill.
• Bronkhorst, Johannes. 2009. Buddhist Teaching in India. Boston: Wisdom Publications.
• Châu, Bhikshu Thích Thiên. 1999. The Literature of the Personalists of Early Buddhism. English translation by Sara Boin-Webb. New Delhi: Motilal Banarsidass.
• Collins, Steven. 1982. Selfless Persons: Imagery and Thought in Theravāda Buddhism. Cambridge: Cambridge University Press.
• Dalai Lama. 1994. The Way To Freedom: Core Teachings of Tibetan Buddhism. San Fransisco: Harper Collins.
• De Jong, J. W. 2000. “The Buddha and His Teachings.” In Wisdom, Compassion, and the Search for Understanding. The Buddhist Studies Legacy of Gadjin M. Nagao, pp. 171-180. Ed. Jonathan A. Silk. Honolulu: University of Hawai’i Press.
• Evans, Stephen A. 2009. “Epistemology of the Brahmajāla Sutta.” Buddhist Studies Review, Vol. 26(1), pp. 67-84.
• Fuller, Paul. 2005. The Notion of ditṭhi in Theravāda Buddhism. The Point of View. London, New York: RoutledgeCurzon.
• Gethin, Rupert. 1986. “The Five khandhas: Their Treatment in the Nikāyas and Early Abhidhamma.” Journal of Indian Philosophy, Vol. 14(1), pp. 35-53.
• Gnoli, Raniero. 1978. The Gilgit Manuscript of the Sanghabhedavastu, Being the 17^th and Last Section of the Vinaya of the Mūlasarvāstivādin, Part I. Roma: Istituto Italiano per il Medio ed Estremo Oriente.
• Gombrich, Richard F. 1992. “Dating the Historical Buddha: A Red Herring Revealed.” In The Dating of the Historical Buddha Part 2, pp. 237-259. Ed. Heinz Bechert. Göttingen: Vandenhoeck & Ruprecht.
• Gombrich, Richard F. 1996. How Buddhism Began: The Conditioned Genesis of the Early Teachings. London: Athlone Press.
• Gombrich, Richard. 2009. What the Buddha Thought. London: Equinox.
• Gómez, Luis O. 1976. “Proto-Mādhyamika in the Pali Canon.” Philosophy East and West, Vol. 26, pp. 137-165.
• Hamilton, Sue. 1996. Identity and Experience: The Constitution of the Human Being According to Early Buddhism. London: Luzac Oriental.
• Hamilton, Sue. 2000. Early Buddhism: A New Approach. Richmond, Surrey: Curzon Press.
• Hayes, Richard. 1988. Dignāga on the Interpretation of Signs. Dordrecht: Publisher.
• Kornfield, Jack. 1996. Living Dharma. Teachings of Twelve Buddhist Masters. Boston and London: Shambhala.
• La Vallée Poussin, Louis de. 1917. The Way to Nirvāna: Six Lectures on Ancient Buddhism as a Discipline of Salvation. (The Hibbert Lectures, 1916) Cambridge: The University Press.
• Malalasekera, G. P. 1957. The Buddha and his Teachings. Colombo: Buddhist Council of Ceylon.
• N̄ānamoli, Bhikkhu and Bodhi, Bhikkhu. 1995. The Middle Length Discourses of the Buddha. A Translation of the Majjhima Nikāya. Boston: Wisdom Publications.
• N̄ānananda, Bhikkhu. 1971. Concept and Reality in Early Buddhist Thought: An Essay on Papañca and Papañca-Sañña-Sankhā. Kandy: Buddhist Publication Society.
• Norman, K. R. 1981. “A Note on attā in the Alagaddūpama Sutta.” In Studies in Indian Philosophy (Memorial Volume for Pandit Sukhlalji Sanghvi), pp. 19-29. Eds. Dalsukh Malvania and Nagin J. Shah. Ahmedabad: L. D. Institute of Indology.
• Oetke, Claus. 1988. Ich und das Ich. Analytische Untersuchungen zur bud-dhistisch-brahmanischen Ātmankontroverse. Stuttgart: Franz Steiner.
• Olivelle, Patrick. 1998. The Early Upanisads. Annotated Text and Translation. Oxford: Oxford University Press.
• Pérez-Remón, Joaquin. 1980. Self and Non-Self in Early Buddhism. The Hague: Mouton Publishers.
• Rahula, Walpola. 1959. What the Buddha Taught. Bedford: Gordon Fraser.
• Ronkin, Noa. 2005. Early Buddhist Metaphysics. The Making of a Philosophical Tradition. London and New York: RoutledgeCurzon.
• Rhys Davids, T. W. 1877. Buddhism: Being a Sketch of the Life and Teachings of Gautama, the Buddha. London: Society for Promoting Christian Knowledge. Republished, New Delhi (2000): Asian Educational Services.
• Rhys Davids, T. W. 1899. Dialogues of the Buddha Part 1. London: Luzac.
• Rhys Davids, T. W. and Stede, William. 1921-1925. Pali-English Dictionary. London: Pali Text Society.
• Senart, Émile. 1897. Le Mahāvastu. Texte Sanscrit, Volume III. Paris: Imprimerie Nationale.
• Siderits, Mark. 2007. Buddhism as Philosophy: An Introduction. UK: Ashgate Publishing Limited.
• Tripathi, Sridhar. 1988. Bodhicaryāvatāra of Śāntideva with the Commentary Pañjikā of Prajñākaramati. (Buddhist Sanskrit Texts No. 12). Darbhanga: The Mithila Institute.
• Vetter, Tilmann. 1988. The Ideas and Meditative Practices of Early Buddhism. Leiden: E. J. Brill.
• Waldschmidt, Ernst. 1952. Das Catuspariṣatsūtra. Teil I: Der Sanskrit Text im Handschriftlichen Befund. Berlin: Akademie-Verlag.
• Warder, A. K. 1991. Introduction to Pali (Third Edition). Oxford: Pali Text Society.
• Warren, Henry Clarke and Kosambi, Dharmananda. 1989. Visuddhimagga of Buddhaghosācariya. (Harvard Oriental Series Vol. 41). New Delhi: Motilal Banarsidass.
• Williams, Paul and Tribe, Anthony. 2000. Buddhist Thought: A Complete Introduction to the Indian Tradition. London and New York: Routledge.
• Wynne, Alexander. 2005. “The Historical Authenticity of Early Buddhist Literature: A Critical Evaluation.” Wiener Zeitschrift für die Kunde Südasiens, Band XLIX, pp. 35-70.
• Wynne, Alexander. 2007. The Origin of Buddhist Meditation. London, New York: Routledge.
• Wynne, Alexander. 2009a. “Early Evidence for the ‘no self’ Doctrine? A Note on the Second anātman Teaching of the Second Sermon.” Thai International Journal for Buddhist Studies, Vol. 1, pp. 64-84.
• Wynne, Alexander. 2009b. “Miraculous Transformation and Personal Identity: A Note on the First anātman Teaching of the Second Sermon.” Thai International Journal of Buddhist Studies, Vol. 1, pp. 85-113.
• Wynne, Alexander. 2010. “The Buddha’s Skill in Means and the Genesis of the Five Aggregate Teaching.” Journal of the Royal Asiatic Soicety, Series 3, Vol. 20(2), pp. 191-216

Footnotes

1. Malalasekera 1957: 33. ↩

2. Rahula 1959: 51. ↩

3. On the Pudgalavādins see Châu 1999 and Williams and Tribe 2000: 124-28. ↩

4. Dalai Lama 1994: 111. ↩

5. La Vallée Poussin 1917: 34. ↩

6. Collins 1982: 95. ↩

7. De Jong 2000: 177. ↩

8. SN I. 296 (v. 553-55): kin nu satto ti paccesi Māra diṭṭhigatan nu te, suddhasañkhārapuñjo ‘yaṃ na yidha sattūpalabbhati. (553) yathā hi an்gasambhārā hoti saddo ratho iti, evaṃ khandhesu santesu hoti satto ti sammuti. (554) dukkham eva hi sambhoti dukkhaṃ tiṭṭhati veti ca, nāññatra dukkhā sambhoti nāññatrā dukkhā nirujjhati. (555) Buddhaghosa cites some of these verses in his Visuddhimagga (XVIII.25, 27; Warren and Kosambi 1989: 508. ↩

9. Rhys Davids 1877: 94. ↩

10. Assuming that the Buddha’s teaching career began at around 450 BCE, and that the Buddha died in about 404 BCE, i.e. about 136 years before Aśoka’s inauguration (Gombrich 1992: 246). On the pre-Aśokan date of canonical Pāli Suttas, see Wynne 2005. ↩

11. Although see section 7 below on the possibility that the doctrine is assumed by the Mahāhatthipadopama Sutta. ↩

12. A minority but not insignificant view is that of Pérez-Remón (1980), who has argued that since the early texts do not deny the self they must in fact presuppose it; for more scholars who believe that the early Buddhist teachings presuppose a self, see Collins 1982: 3-10. Against this view, Vetter (1988: 41, n.10) has argued that although the early texts do not deny the existence of the self, they do not presuppose it. Oetke’s detailed study (1988) also argues that the early texts neither deny nor affirm the existence of the self. ↩

13. According to Olivelle (1998: 22) the term is used in two other senses, i.e. as a simple reflexive pronoun and in denoting “a living, breathing body.” The three usages are also well attested throughout Sanskrit literature. According to Monier Williams, in its simplest sense the term ātman refers to the bodily person: “the person or whole body considered as one and opposed to the separate members of the body” (MMW s.v.). In a similar fashion, the term ātman is often used as a reflexive pronoun: “ātman in the sg. is used as reflexive pronoun for all three persons and all three genders, e.g. ātmānaṃ sā hanti ‘she strikes herself’” (MMW s.v.). Monier Williams also cites the expressions ātman (Ved. loc.) dhatte/karoti: “he places in himself, makes his own,” and atmanā akarot: “he did it himself;” he also suggests that something like “essence” is the oldest and most basic definition of the term ātman as “the breath, the soul, principle of life and sensation (RV, AV).” ↩

14. That the ātman is a spiritual principle is made clear in the numerous references to its immortality, e.g. BU IV.3.12, BU IV.4.16, BU IV.4.17, BU IV.4.25. That it is a spiritual principle distinct from the body is made clear in those passages which describe its reincarnation, e.g. BU IV.3.8 and especially the detailed account of BU IV.4.3: … ayam ātmedaṃ śarīraṃ nihatyāvidyāṃ gamayitvānyam ākramam ākramyātmānam upasaṃharati: “Once this ātman has struck the body down and rendered it unconsciousness, it approaches another station and draws itself towards it.” See also BU IV.4.7: tadyathā ahinirvlayanī valmīke mṛtā prayastā śayīta, evam evedaṃ śarīraṃ śete. athāyaṃ aśarīro ‘mṛtah prāṇo brahmaiva teja eva. ↩

15. BU IV.4.2: tam utkrāmantaṃ prāṇo ‘nūtkrāmati, prāṇam utkrāmantaṃ sarve prāṇā anūtkrāmanti. ↩

16. E.g. BU III.7.23: adṛ̣̣̦to draṣtrāśrutaḥ śrotāmato mantāvij̃̃āto vijñātā (see also BU III.8.11); see also BU IV.5.15 which makes it clear that the ātman, as the perceiver, cannot be perceived (vijñātāram are kena vijānīyād: “By what means might one perceive the perceiver?”). ↩

17. BU IV.3.7: yo ‘yaṃ vijñānamayaḥ prāneṣu; BU IV.4.22: sa vā eṣa mahān aja ātmā vijñānamayah prāneṣu. ↩

18. Although Williams states that “very little” of his discussion about Western concepts of the soul “is relevant to the Buddha,” (Williams and Tribe 2000: 56), his definition of the Cartesian position shows striking similarities with Yājñavalkya’s understanding of the ātman (ibid.): “As is well known, Descartes identified that which gives life to the body, and survives death, with the mind, and he also identified this mind-soul as the true self, of an intrinsically different stuff from the body. The mind-soul is the factor in which lies the identity of the person over time and change.” ↩

19. BU IV.3.32 states that the ātman is the “highest bliss” (parama ānandah), BU IV.5.22 states that the ātman is nondual consciousness (vijñānaghana), BU IV.5.11 states that it is macrocosmic (mahābhūta), and BU IV.5.12 likens the person who unites with the ātman in deep sleep to a “single ocean” (salila ekah), a state equated with the “world of brahma” (brahmaloka). ↩

20. BU IV.5.6: ātmani khalv are dṛ̣̣te śrute mate vijñāta idaṃ sarvaṃ viditaṃ: “When the ātman is seen, heard, contemplated and cognised, the whole world is known.” For a parallel teaching see BU II.4.5: ātmano vā are darśanena śravanena matyā vijñānenedaṃ sarvaṃ viditam. ↩

21. Vin I.13.18: rūpaṃ bhikkhave anattā. rūpañ ca h’ idam bhikkhave attā abhavissa, na yidaṃ rūpaṃ ābādhāya saṃvatteyya, labbhetha ca rūpe: evaṃ me rūpaṃ hotu, evaṃ me rūpaṃ mā ahosī ti. yasmā ca kho bhikkhave rūpaṃ anattā, tasmā rūpaṃ ābādhāya saṃvattati, na ca labbhati rūpe: evaṃ me rūpaṃ hotu, evaṃ me rūpaṃ mā ahosī ti. vedanā anattā, vedanā ca h’ idam bhikkhave attā abhavissa … For this teaching in Buddhist Sanskrit texts, see Mvu III.335.12, SbhV I.138.10 and CPS 15.2 (Waldschmidt 1952: 162). ↩

22. Collins 1982: 97 has suggested that this teaching is directed against the Brahminic notion of the atman as the “microcosmic reflection of the macrocosmic force of the universe (brahman).” But the teaching does not presuppose that the atman should be an inner controller, and if so it would not appear to be a “microcosmic reflection” of a world-controlling force. The teaching instead states that if the five aggregates were atman/attan, a person should be able to change them as he wants. The argument is thus directed against the notion that the five aggregates constitute the atman/attan, and not against the notion that there is ātman/attan controlling them from within, as is also assumed by Siderits (2007: 46ff.). For a full study of the conceptual and historical implications of this teaching see Wynne 2009a; especially 86-88. ↩

23. Vin I.14: taṃ kiṃ maññatha bhikkhave: rūpaṃ niccaṃ vā aniccaṃ vā ti? aniccaṃ bhante. yaṃ panāniccaṃ dukkhaṃ vā taṃ sukhaṃ vā ti? dukkhaṃ bhante. yaṃ panāniccaṃ dukkhaṃ vipariṇāmadhammaṃ, kallan nu taṃ samanupassituṃ: etaṃ mama, eso ‘haṃ asmi, eso me attā ti? no h’ etaṃ bhante. For the various Sanskrit versions of this teaching, see Mvu III.337.11, SbhV I.138.21 and CPS 15.6 (Waldschmidt 1952: 164). ↩

24. Gethin 1986: 49: “the five khandhas, as treated in the Nikāyas and early Abhidhamma, do not exactly take on the character of a formal theory of the nature of man. The concern is not so much the presentation of an analysis of man as object, but rather the understanding of the nature of conditioned existence from the point of view of the experiencing subject. Thus at the most general level rūpa, vedanā, saññā, saṃkhārā and viññāṇa are presented as five aspects of an individual being’s experience of the world …” Hamilton (2000: 27) has similarly written that the five aggregates are “not a comprehensive analysis of what a human being is comprised of … Rather they are factors of human experience.” ↩

25. Norman 1981: 22. To this we might add that by equating “impermanence” (anicca) with being “subject to change” (viparināmadhamma), the Buddha recalls a key feature of the self according to the Yājñavalkyakānda (e.g. BU IV.5.15), i.e. that it is unchangeable. Bronkhorst (2007: 233) has noted that BU IV.5.15 “introduces the notion of the immutability of the self.” ↩

26. See BU II.1.19-20, BU III.9.28, BU IV.3.32-33, TU II.5-9, TU III.6. ↩

27. E.g. BU II.4.12: idaṃ mahad bhūtam anantam apāraṃ vijñānaghana eva; BU IV.5.13: ayam ātmānantaro ‘bāhyaḥ krrtsnaḥ prajñānaghana eva. ↩

28. Opposition to the notion that this teaching denies “soul” has been made by Gombrich (1996: 15): “In Western languages, the Buddha is presented as having taught the doctrine (vāda) of ‘no soul’ (anātman). What is being denied - what is a soul? Western theologians are at home in the Christian cultural tradition. Christian theologians have differed vastly over what the soul is. For Aristotle, and thus for Aquinas, it is the form of the body, what makes a given individual person a whole rather than a mere assemblage of parts. However, most Christians conceive of the soul, however vaguely, in a completely different way, which goes back to Plato: that the soul is other than the body, as in the expression ‘body and soul,’ and is some kind of disembodied mental, and above all, moral, agent, which survives the body after death. But none of this has anything to do with the Buddha’s position. He was opposing the Upaniṣadic theory of soul.” A similar point has been made by Williams (Williams and Tribe 2000: 56). ↩

29. This is how the teaching is presented in its most important occurrence in the Pāli discourses, the Alagaddūpama Sutta (MN I.135.27ff.). ↩

30. I follow the definition of the OED. ↩

31. It is perhaps possible that the compound tanhāvicarita is to be read as a dependent determinative (tatpurusa) in the dative case, i.e. “movements towards thirst.” But the compound tanhāvicarita is more likely to be read as a dependent determinative in the instrumental case: a past participle preceded by a substantive suggests some sort of conceptual activity (vicarita) prompted by a cause (tanhā). If so, self-consciousness would seem to be caused by an underlying affective state termed “thirst.” ↩

32. AN II.212.13: katamāni atṭthārasa tanhāvicaritāni ajjhattikassa upādāya? asmī ti bhikkhave sati … ↩

33. MN I.47.21: yato kho āvuso ariyasāvako evaṃ akusalaṃ pajānāti, evaṃ akusalamūlaṃ pajānāti, evaṃ kusalaṃ pajānāti, evaṃ kusalamūlaṃ pajānāti, so sabbaso rāgānusayaṃ pahāya, pațighānusayaṃ pațivinodetvā, asmī ti ditṭhimānānusayaṃ samūhanitvā avijjaṃ pahāya vijjaṃ uppādetvā, diṭṭheva dhamme dukkhass’ antakaro hoti. ettāvatā pi kho āvuso ariyasāvako sammāditṭhi hoti … ↩

34. Ud VI. 6 (v. 70.23): atha kho Bhagavā etam atthaṃ viditvā tāyaṃ velāyaṃ imaṃ udānaṃ udānesi: ahaṅkārapasutā ayaṃ pajā paraṃkārūpasaṃhitā, etad eke nābbhaññaṃsu, na naṃ sallan ti addaṃsu. etaṃ ca sallaṃ paṭigacca passato, ahaṃ karomī ti na tassa hoti, paro karotī ti na tassa hoti. mānupetā ayaṃ pajā mānaganthā mānavinibaddhā, diṭṭhīsu byārambhakatā, saṃsāraṃ nātivattatī ti. ↩

35. SN III.235.22: idhāhaṃ āvuso vivicc’ eva kāmehi vivicca akusalehi dhammehi savitakkaṃ savicāraṃ vivekajaṃ pītisukhaṃ paṭhamajjhānaṃ upasampajja viharāmi. tassa mayhaṃ āvuso na evaṃ hoti: ahaṃ paṭhamajjhānaṃ samāpajjāmī ti vā, ahaṃ paṭhamajjhānaṃ samāpanno ti vā, ahaṃ paṭhamajjhānā vuṭṭhito ti vā ti. tathā hi panāyasmato Sāriputtassa dīgharattaṃ ahaṃkāramamaṃkāramānānusayā susamūhatā, tasmā āyasmato Sāriputtassa na evaṃ hoti: ahaṃ paṭhamajjhānaṃ samāpajjāmī ti vā ahaṃ paṭhamajjhānaṃ samāpanno ti vā ahaṃ paṭhamajjhānā vutṭhito ti vā ti. The PTS reading pahatmajjhānā at the end instead of paṭhamajjhānā is clearly an error. ↩

36. SN II.275.1: tathā hi panāyasmato Sāriputtassa dīgharattaṃ ahañkāramamañkāramānānusayā susamūhatā. ↩

37. SN IV.40.16: etha me āvuso imaṃ kāyaṃ mañcakaṃ āropetvā bahiddhā nīharatha. purāyaṃ kāyo idh’ eva vikirati, seyyathā pi bhūsamutṭhī ti. ↩

38. SN IV.40.16: etha me āvuso imaṃ kāyaṃ mañcakaṃ āropetvā bahiddhā nīharatha. purāyaṃ kāyo idh’ eva vikirati, seyyathā pi bhūsamutṭhī ti. ↩

39. SN IV.40.20: evaṃ vutte āyasmā Sāriputto āyasmantaṃ Upasenam etad avoca: na kho pana mayaṃ passāma āyasmato Upasenassa kāyassa vā aññathattaṃ indriyānaṃ vā viparināmaṃ. atha ca panāyasmā Upaseno evam āha: etha me āvuso imaṃ kāyaṃ mañcakaṃ āropetvā bahiddhā nīharatha; purāyaṃ kāyo idh’ eva vikirati, seyyathā pi bhūsamutṭhī ti. ↩

40. AN IV.40.29: yassa nūna āvuso Sāriputta evam assa: ahaṃ cakkhun ti mama cakkhun ti vā … la … ahaṃ jivhā ti vā, mama jivhā ti vā … ahaṃ mano ti vā mama mano ti vā. tassa āvuso Sāriputta siyā kāyassa vā aññathattaṃ indriyānaṃ vā vipariṇāmo. mayhañ ca kho āvuso Sāriputta na evaṃ hoti: ahaṃ cakkhun ti vā, mama cakkhun ti vā … la … ahaṃ jivhā ti vā mama jivhā ti vā … ahaṃ mano ti vā mama mano ti vā. tassa mayhañ ca kho āvuso Sāriputta kiṃ kāyassa vā aññathattaṃ bhavissati indriyānaṃ vā vipariṇāmo ti? ↩

41. SN IV.41.11: atha kho te bhikkhū āyasmato Upasenassa kāyaṃ mañcakaṃ āropetvā bahiddhā nīhariṃsu. atha kho āyasmato Upasenassa kāyo tatth’ eva vikiri, seyyathā pi bhūsamutṭhī ti. ↩

42. MN III.32.32ff. ↩

43. AN I.133.1: idh’ Ānanda bhikkhuno evaṃ hoti: etaṃ santaṃ etaṃ paṇītaṃ yadidaṃ sabbasañkhārasamatho sabbūpadhipaṭinissaggo taṇhākkhayo virāgo nirodho nibbānan ti. evaṃ kho Ānanda siyā bhikkhuno tathārūpo samādhipaṭilābho yathā imasmiñ ca saviññānake kāye ahañkāramamañkāramānānusayā nāssu bahiddhā ca sabbanimittesu ahañkāramamañkāramānānusayā nāssu, yañ ca cetovimuttim paññāvimuttiṃ upasampajja viharato ahañkāramamañkāramānānusayā na honti, tañ ca cetovimuttim paññā̄vimuttiṃ upasampajja vihareyyā ti. ↩

44. AN IV.53.7: dukkhe anattasaññāparicitena bhikkhave bhikkhuno cetasā bahulaṃ viharato imasmiñ ca saviññānake kāye bahiddhā ca sabbanimittesu ahaṃkāramamaṃkāramānāpagataṃ mānasaṃ hoti vidhāsamatikkantaṃ santaṃ suvimuttam. ↩

45. MN III.18.32: yaṃ kiñci bhikkhu rūpaṃ - atītānāgatapaccuppannaṃ, ajjhattaṃ vā bahiddhā vā, olārikaṃ vā sukhumaṃ vā, hīnaṃ vā paṇītaṃ vā, yaṃ dūre santike vā - sabbaṃ rūpaṃ: n’ etaṃ mama, n’ eso ham asmi, na m’ eso attā ti, evam etaṃ yathābhūtaṃ sammappaññāya passati … MN III.19.7: evaṃ kho bhikkhu jānato evaṃ passato imasmiñ ca saviññānake kāye bahiddhā ca sabbanimittesu ahaṃkāramamaṃkāramānānusayā na hontī ti. For the same teaching see SN II.252.16, 253.11; SN III.80.7, 81.1, 103.12, 136.4, 136.24, 169.12, 170.7. ↩

46. I give the standard translation of the compound upādānakkhandha, but for a more detailed historical explanation see Gombrich 1996: 67 and Wynne 2007: 84. ↩

47. SN III.128.29: pañc’ ime āvuso upādānakkhandhā vuttā Bhagavatā, seyyathīdam: rūpupādānakkhandho … pe … viññāṇupādānakkhandho. imesu khv āhaṃ āvuso pañcasu upādānakkhandhesu na kiñci attānaṃ vā attaniyam vā samanupassāmi, na c’ amhi arahaṃ khīnāsavo. api ca me āvuso pañcasu upādānakkhandhesu asmī ti adhigataṃ, ayam aham asmī ti na ca samanираззāmī ti. ↩

48. SN III.130.28: kiñcāpi āvuso ariyasāvakassa pañcorambhāgiyāni saññojanāni pahīnāni bhavanti atha khv assa hoti: y’ eva pañcasu upādānakkhandhesu anusahagato asmī ti māno asmī ti chando asmī ti anusayo asamūhato. so aparena samayena pañcasu upādānakkhandhesu udayabbayānupassī viharati: iti rūpaṃ, iti rūpassa samudayo, iti rūpassa atthagamo; iti vedanā … iti saññā … iti sañkhārā … iti viññānaṃ, iti viññāṇassa samudayo, iti viññāṇassa atthagamo ti. tass’ imesu pañcasu upādānakkhandhesu udayabbayānupassino viharato, yo pi ‘ssa hoti pañcasu upādānakkhandhesu anusahagato asmī ti māno asmī ti chando asmī ti anusayo asamūhato, so ‘pi samugghātam gacchati. ↩

49. Pérez-Rémon 1980: 237, on which see Oetke 1988: 107-109. ↩

50. SN III.130.13: seyyathā pi āvuso uppalassa vā padumassa vā puṇdarīkassa vā gandho. yo nu kho evaṃ vadeyya: pattassa gandho ti vā, vaṇnassa gandho ti vā, kiñjakkhassa gandho ti vā, sammā nu kho so vadamāno vadeyyā ti? no h’ etaṃ āvuso. yathākathaṃ panāvuso sammāvyākaramāno vyākareyyā ti? pupphassa gandho ti kho āvuso sammāvyākaramāno vyākareyyā ti. evam eva khv āhaṃ āvuso na rūpaṃ asmī ti vadāmi, na pi aññatra rūpā asmī ti vadāmi, na vedanam … na saññam … na sañkhāre … na viññāṇam asmī ti vadāmi, na pi aññatra viññāṇā asmī ti vadāmi. api ca me āvuso pañcasu upādānakkhandhesu asmī ti adhigataṃ, ayam aham asmī ti na ca samanupassāmi. In the second sentence reading pattassa gandho ti vā, vanñassa gandho ti vā, kiñjakkhassa gandho ti vā with CSCD instead of the PTS pattassa gandho ti, vanñassa gandho pi, kiñjakkhassa gandho ti vā. ↩

51. SN III.105.10: Punṇo nāma āvuso āyasmā Mantāniputto amhākam navakānaṃ sataṃ bahūpakāro hoti. so amhe iminā ovādena ovadati: upādāya āvuso Ānanda asmī ti hoti, no anupādāya. kiñ ca upādāya asmī ti hoti, no anupādāya? rūpaṃ upādāya asmī ti hoti, no anupādāya. vedanaṃ … sañ̃̃aṃ … sañkhāre … viññānaṃ upādāya asmī ti hoti, no anupādāya. ↩

52. SN III. 105.25. ↩

53. SN IV.197.11: ayaṃ kho bhante vīṇa nāma anekasambhārā mahāsambhārā anekehi sambhārehi samāraddhā vadati, seyyathidam: doṇiñ ca paṭicca, cammañ ca paṭicca, daṇdañ ca paṭicca, upaveṇañ ca paṭicca, tantiyo ca paṭicca, koṇañ ca paṭicca, purisassa ca tajjaṃ vāyāmam paṭicca evāyaṃ bhante vīṇa nāma anekasambhārā mahāsambhārā anekehi sambhārehi samāraddhā vadatī ti. Following the translation of Bodhi (2000: 1254). On this simile see Collins 1982: 101. ↩

54. SN IV.197.25: evam eva kho bhikkhave bhikkhu rūpaṃ samanvesati yāvatā rūpassa gati, vedanaṃ samanvesati … pe … saññaṃ … sañkhāre … viññāṇam samanvesati yāvatā viññāṇassa gati. tassa rūpaṃ samanvesato … pe … sañ̃̃aṃ … sañkhāre … viññānaṃ samanvesato yāvatā viññā̄nassa gati, yam pi ‘ssa taṃ hoti ahan ti vā maman ti vā asmī ti vā, tam pi tassa na hotī ti. Reading samanvesati and samanvesato with CSCD for PTS samanesati and samanesato respectively. ↩

55. See n. 42 above. ↩

56. MN I.486.10: atthi pana bhoto Gotamassa kiñci ditṭhigatan ti? ditṭhigatan ti kho Vaccha apanītam etaṃ tathāgatassa, ditṭhañ h’ etaṃ Vaccha tathāgatena: iti rūpaṃ, iti rūpassa samudayo, iti rūpassa atthagamo; iti vedanā, iti vedanāya samudayo, iti vedanāya atthagamo; iti sañ̃̃ā, iti sañ̃̃āya samudayo, iti saññāya atthagamo; iti sañkhārā, iti sañkhārānaṃ samudayo, iti sañkhārānaṃ atthagamo; iti viñ̃̃ānaṃ, iti viññāṇassa samudayo, iti viññāṇassa atthagamo ti. tasmā tathāgato sabbamaññitānaṃ sabbamathitānaṃ sabba-ahimkāramamimkāramānānusayānaṃ khayā virāgā nirodhā cāgā paṭinissaggā anupādā vimutto ti vadāmī ti. ↩

57. SN IV.202.6: yadā ca kho bhikkhave Vepacittissa asurindassa evaṃ hoti: dhammikā kho devā, adhammikā asurā, idh’ eva dānāhaṃ devapuraṃ gacchāmī ti, atha kaṇthe pañcamehi bandhanehi muttam attānaṃ samanupassati, dibbehi ca pañcahi kāmaguṇehi samappito samaṅgībhūto paricāreti. yadā ca kho bhikkhave Vepacittissa asurindassa evaṃ hoti: dhammikā kho asurā, adhammikā devā, tatth’ eva dānāhaṃ asurapuraṃ gamissāmī ti, atha kaṇthe pañcamehi bandhanehi baddham attānaṃ samanupassati, dibbehi ca pañcahi kāmaguṇehi parihāyati. evaṃ sukhumam kho bhikkhave Vepacittibandhanam, tato sukhumataraṃ Mārabandhanaṃ. maññamāno kho bhikkhave baddho Mārassa, amaññamāno mutto pāpimato. ↩

58. SN IV.202.20: asmī ti bhikkhave maññitam etaṃ, ayam aham asmī ti maññitam etaṃ, bhavissan ti maññitam etaṃ, na bhavissan ti maññitam etaṃ, rūpī bhavissan ti maññitam etaṃ, arūpī bhavissan ti maññitam etaṃ, saññ̄̄ bhavissan ti maññitam etaṃ, asaññ̄̄ bhavissan ti maññitam etaṃ, nevasaññ̄̄ nāsaññ̄̄ bhavissan ti maññitam etaṃ. maññitaṃ bhikkhave rogo, maññitaṃ gaṇdo, maññitaṃ sallaṃ. tasmāt iha bhikkhave amaññamānena cetasā viharissāmī ti, evañ hi vo bhikkhave sikkhitabbaṃ. Reading saññ̄̄ bhavissan ti maññitam etaṃ with CSCD instead of PTS saññ̄̄ bhavissan ti; rogo with CSCD instead PTS rāgo; and amaññamānena with CSCD instead of PTS amaññitmānena. ↩

59. MN III.246.11: asmī ti bhikkhu maññitam etaṃ, ayam aham asmī ti maññitam etaṃ, bhavissan ti maññitam etaṃ, na bhavissan ti maññitam etaṃ, rūpī bhavissan ti maññitam etaṃ, arūpī bhavissan ti maññitam etaṃ, saññ̃̃ bhavissan ti maññitam etaṃ, asaññ̃̃ bhavissan ti maññitam etaṃ, nevasaññīnāsaññ̃̃ bhavissan ti maññitam etaṃ. maññitaṃ bhikkhu rogo, maññitaṃ gando, maññitaṃ sallaṃ. sabbamaññitānaṃ tv eva bhikkhu samatikkamā muni santo ti vuccati. ↩

60. SN IV.202.28: asmī ti bhikkhave iñjitam etaṃ, ayam aham asmī ti iñjitam etaṃ, bhavissan ti iñjitam etaṃ, na bhavissan ti iñjitam etaṃ, rūpī bhavissan ti iñjitam etaṃ, arūpī bhavissan ti iñjitam etaṃ, saññ̃̃ bhavissan ti iñjitam etaṃ, asaññ̃̃ bhavissan ti iñjitam etaṃ, nevasaññīnāsaññ̃̃ bhavissan ti iñjitam etaṃ. The text then repeats this passage but replaces iñjitam etaṃ with phanditam etaṃ, papañcitam etaṃ and finally mānagatam etaṃ. ↩

61. MMW s.v.: ‘expansion, development, manifestation … manifoldness, diversity.’ See also Gombrich 2009: 205-206. ↩

62. The standard study is that of Ñānananda 1971. ↩

63. MN I.111.35: cakkhuñ c’ āvuso paṭicca rūpe ca uppajjati cakkhuviññānaṃ, tinnnaṃ sañgati phasso, phassapaccayā vedanā, yaṃ vedeti taṃ sañjānāti, yaṃ sañjānāti taṃ vitakketi, yaṃ vitakketi taṃ papañceti, yaṃ papañceti tatonidānaṃ purisaṃ papañcasaññāsañkhā samudācarati atītānāgatapaccupannesu cakkhuviññeyyesu rūpesu. It is not clear exactly how the compound papañcasaññāsañkhā is to be taken. Ñāṇamoli and Bodhi (1995: 203) translate it as “perceptions and notions [born of] mental proliferation,” although this misses the fact that the compound is declined in the singular number. ↩

64. For an analysis of these teachings see Oetke 1988: 130ff. ↩

65. DN II.67.12: aniccaṃ sukhadukkhavokinnaṃ uppādavyayadhammam attānaṃ … ↩

66. DN II.67.25: tatr’ Ānanda yo so evam āha: na h’ eva kho me vedanā attā, no pi appațisaṃvedano me attā, attā me vediyati, vedanādharmo hi me attā ti, so evam assa vacanīyo: vedanā va hi āvuso sabbena sabbaṃ sabbathā sabbaṃ aparisesā nirujjheyum, sabbaso vedanāya asati vedanānirodhā, api nu kho tattha ayam aham asmī ti siyā tī? no h’ etaṃ bhante. tasmāt ih’ Ānanda etena p’ etaṃ na h’ eva kho na kkhamati ‘me vedanā attā, no pi appațisaṃvedano me attā, attā me vediyati, vedanādharmo hi me attā’ ti samanupassitum. ↩

67. DN II.67.17: tatr’ Ānanda yo so evam āha: na h’ eva kho me vedanā attā, appațisaṃvedano me attä’ ti, so evam assa vacanīyo: yattha pan’ āvuso sabbaso vedayitaṃ n’ atthi, api nu kho tattha asmī ti siyā’ ti? no h’ etaṃ bhante. tasmāt ih’ Ānanda etena p’ etaṃ na kkhamati ‘na h’ eva kho me vedanā attā, appațisaṃvedano me attä’ ti samanupassitum. On this teaching see Collins 1982: 99 . ↩

68. CU VIII.11.1: tad yatraitat suptah samastah samprasannah svapnaṃ na vijānāty eṣa ātmeti hovāca, etad amṛtam abhayam etad brahmeti. sa ha śāntahṛdayaḥ pravavrāja. sa hāprāpyaiva devān etad bhayaṃ dadarśa. nāha khalv ayam evaṃ saṃpratyātmānaṃ jānāty ayam aham asmīti. no evemāni bhūtāni. vināśam evāpīto bhavati. nāham atra bhogyaṃ paśyāmi. ↩

69. CU VIII.12.3-5: aśarīro vāyuh, abhram vidyut stanayitnur aśarīrāny etāni. tadyathaitāny amuṣ̣ād ākāśāt samutthāya param jyotir upasampadya svena rūpenābhiniṣpadyante (2). evam evaiṣa saṃprasādo ‘smāc charīrāt samutthāya paraṃ jyotir upasampadya svena rūpenābhiniṣpadyate. sa uttamapuruṣah. sa tatra paryeti jakṣat krīdan ramamānah strībhir vā yānair vā jñātibhir vā nopajanāṃ smarann idaṃ śarīram. sa yathā prayogya ācaraṇe yukta evam evāyam asmiñ charīre prāno yuktaḥ (3). atha yatraitad ākāśam anuviṣannam cakṣuh, sa cākṣuṣah puruṣo darśanāya cakṣuh. atha yo vededaṃ jighrānīti sa ātmā gandhāya ghrānaṃ. atha yo vededam abhivyāharānīti sa ātmābhivyāhārāya vāk. atha yo vededaṃ śrnavānīti sa ātmā śravanāya śrotram (4). atha yo vededaṃ manvānīti sa ātmā, mano ‘sya daivaṃ cakṣuh. sa vā eṣa etena daivena cakṣuṣā manasaitān kāmān paśyan ramate ya ete brahmaloke (5). ↩

70. CU VIII. 9.1: yathaiva khalv ayam asmiñ śarīre sādhvalaṅkṛte sādhvalaṅkrto bhavati suvasane suvasanaḥ pariṣkṛte pariṣkrta, evam evāyam asminn andhe andho bhavati srāme srāmo parivṛkne parivṛkno. asyaiva śarīrasya nāśam anv eṣa naśyati. nāham atra bhogyam paśyāmi. ↩

71. DN II.67.12: aniccaṃ sukhadukkhavokiṇnaṃ uppādavyayadhammam attānaṃ … ↩

72. The simile of the raft is found at MN I.134.30, on which see Gombrich 1996: 23-25. The simile of the arrow is found at MN I.429.2ff., on which see Gethin 1998: 66-67. ↩

73. SN IV.400-401, on which see Gethin 1998: 161. ↩

74. This approach is aptly summarised when the Buddha compares what he has taught to the few leaves in his hand, whereas what he knows is compared to all the leaves in a forest grove (SN V.437-438). For a different interpretation of this simile, see p. 167 below. ↩

75. See the concluding parts of both teachings at MN I.139.11 and DN II.68.4. ↩

76. Hayes (1988: 48) seems to view the text as some sort of ancient self-help manual: “In an era in which various teachers are gathering disciples around them and making claims of supernatural powers and access to cosmic information that is beyond the ken of ordinary mortals, the superior knowledge of the Tathāgata consists in no more than a full awareness of his own feelings (vedanā) and the realization that tranquility is possible only by giving up being attached to them.” For a detailed study of the form of the Not-Self teaching in the Alagaddūpama Sutta see Wynne 2010: 210-211. ↩

77. Fuller 2005: 115. Although Fuller does not clarify his understanding of the historicity of the early Buddhist literature, his attempt to synthesise the Brahmajāla Sutta with the Sammāditthi Sutta is based on the presupposition that the teachings in the Pāli Nikāyas constitute a homogeneous whole; this is also indicated by statements such as “the Pāli canon teaches …” (Fuller 2005: 157). This is seriously misconceived, however, for there is much evidence in the Pāli Nikāyas for divergent views and even debate. For a sample of such views see Wynne 2007: 117ff. and 2009a. ↩

78. Collins 1982: 129. ↩

79. DN I.39.14: ye keci bhikkhave samaṇa vā brāhmaṇa vā pubbantakappikā caaparantakappikāca pubbantāparantakappikāca pubbantāparantaditṭhino pubbantāparanta ārabbha anekavihitāni adhivuttipadāni abhivadanti, sabbe te imeh’ eva dvāsaṭṭhiiyā vatthūhi etesaṃ vā aññatarena, n’ atthi ito bahiddhā. temporal existence. An indication of this is given in the following statement that precedes the presentation of views: ↩

80. DN I.12.18: atthi bhikkhave aññ’ eva dhammā gambhīrā duddasā duranubodhā santā pañītā atakkāvacarā nipunā paṇ̂itavedanīyā, ye Tathāgato sayam abhiñña sacchikatvā pavedeti, yehi Tathāgatassa yathābhuccam vanṇaṃ sammā vadamānā vadeyyuṃ. ↩

81. DN I.22.17ff. ↩

82. DN I.13.11ff. ↩

83. DN I.28.25ff. ↩

84. DN I.36.23ff. ↩

85. DN I.21.16ff. ↩

86. DN I.31.6ff., DN I.32.10ff., DN I.33.1ff. ↩

87. DN I.34.6ff. ↩

88. DN I.39.20: tayidaṃ bhikkhave Tathāgato pajānāti: ime ditṭ̣hitṭthānā evamgahitā evamparāmaṭ̣thā evamgatikā bhavissanti evamabhisamparāyā ti. tañ ca Tathāgato pajānāti tato ca uttarī̄taraṃ pajānāti, ca pajānanaṃ na parāmasati, aparāmasato c’ assa paccattạ̣ yeva nibbuti viditā. vedanānaṃ samudayañ ca atthagamañ ca assādañ ca ādīnavañ ca nissaraṇañ ca yathābhūtaṃ viditvā anupādā vimutto, bhikkhave, Tathāgato. ↩

89. DN I.41.29: tatra bhikkhave ye te samaṇabrāhmaṇā pubbantakappikā ca aparantakappikācapubbantāparantakappikāca,pubbantāparantānuditṭhino pubbantāparantaṃ ārabbha anekavihitaṃ adhivuttipadāni abhivadanti dvāsaṭṭhiyā vaṭṭhūhi, tad api tesaṃ bhavataṃ samaṇabrāhmanānaṃ ajānataṃ apassataṃ vedayitaṃ taṇhāgatānaṃ paritasitavipphanditam eva. ↩

90. Fuller (2005: 115) translates ajanatam apassatam vedayitaṃ taṇhāgatānaṃ paritasitavipphanditam eva as “only the feeling of those who do not know and do not see […]; only the agitation and vacillation of those immersed in craving.” More recently Evans (2009: 71) writes that “each of the views is merely the feeling [vedayitam] of those who do not know and see, the worry and vacillation of those immersed in craving.” It is unlikely, however, that the past participles in this construction can be taken in a nominal sense: it is more likely that the placing of the “ascetics and Brahmins” in the genitive case (samanabrāhmanānaṃ) indicates that they are agents of the verb expressed by the past participles, so that they are what Warder (1963: 57) has called “agent-genitives.” Collins’ translation (1982: 128) is closer to this, for he takes the term vedayitam as a past participle and understands the correlative clause beginning tad api as follows: ”… something experienced by these ascetics and Brahmins, who neither know nor see, and are subject to craving.” This problem with this translation is that it conventiently avoids the final words paritasita-vipphanditam eva. Moreover, to say that views are “experienced” is odd, and does not add anything to the analysis; it is also implausible to translate tad api as “something,” i.e. as a correlative pronoun with an indefinite sense.
    The main grammatical problem posed by this sentence is the correlative phrase tad api. The word tad cannot be a correlative pronoun, since such a pronoun would have to be in the plural rather than the singular number, there being a large number of diverse views held by various ascetics and Brahmins. Indeed, the fact that the phrase tad api appears in the subsequent expression tad api phassapaccayā (e.g. DN I. 42.2 and following) indicates that tad is not to be taken as a neuter pronoun in agreement with vedayitam. If, then, the word tad is not a correlative pronoun that picks up a noun in the preceding clause and agrees with the term vedayitam that follows, it is more plausible to take the phrase tad api as an adverbial correlative construction. Such a construction could be used to explain the reason for a preceding state of affairs, i.e. the fact that various ascetics and Brahmins state different views. In other words, it seems to have a meaning close to the adverbial sense noted by noted by Rhys Davids and Stede (ta, s.v. 4c: “therefore … that is why, now, then”); a close translation could be something like “in this case too …” The point is thus that the ascetics and Brahmins are able to state various views because their primary experience (that which they sense: vedayita) is subjected to “quivering” and “trembling.” ↩

91. Hamilton (2000: 169-170) has suggested something similar to this: “What is more difficult to grasp, or what is even less obvious, is that if the structure of the world of experience is correlated with the cognitive process, then it is not just that we name objects, concrete and abstract, and superimpose secondary characteristics according to the senses as described. It is also that all the structural features of the world of experience are cognitively correlated. In particular, space and time are not external to the structure but are part of it.” ↩

92. As Ronkin has pointed out (2005: 244), according to this understanding “the boundaries of one’s cognitive process are the boundaries of one’s world: the latter is the world of one’s own experience, dependent on the workings of one’s cognitive apparatus.” Evans (2009: 80) has made the similar point that “by insisting that the 62 positions are vedayita, conditioned by phassa, leading to vedanā and so on, the Sutta implies that they are causally conditioned hence lack fully definite truth-value.” See also Hamilton 2000: 107-108: “the Buddha metaphorically relates the different aspects of what we think as the world around us to one’s subjective experience. In explaining how the khandhas work, he focuses in particular on the fact that we cannot have access to anything else: all of our experience is mediated to us by means of them. And our ‘world’ is simply that. We cannot have access to an ‘external’ world because we cannot get outside of out experience. Our experience, then, is our world.” ↩

93. Ronkin 2005: 245. ↩

94. See n. 97 above. ↩

95. DN I.45.22: yato kho bhikkhave bhikkhu channaṃ phassāyatanānaṃ samudayañ ca atthagamañ ca assādañ ca ādīnavañ ca nissaranañ ca yathābhūtaṃ pajānāti, ayaṃ imehi sabbeh’ eva uttarītaraṃ pajānāti. ↩

96. According to the standard explanation of the Madhupindaka Sutta (MN I.111.35ff.), conditioned experience begins as follows: cakkhuṇ c’ āvuso paṭicca rūpe ca uppajjati cakkhuviññānaṃ, tinnnaṃ sañgati phasso, phassapaccayā vedanā. On this passage see n. 72 above. ↩

97. For the argument that this text is probably the source of the Not-Self teaching, see Wynne 2010. ↩

98. MN I.140.3: evaṃvimuttacittaṃ kho bhikkhave bhikkhuṃ sa-Indā devā sa-Brahmakā sa-Pajāpatikā anvesaṃ nādhigacchanti: idaṃ nissitaṃ Tathāgatassa viñ̃̃̃anann ti. taṃ kissa hetu? diṭṭhe vāhaṃ bhikkhave dhamme Tathāgataṃ ananuvejjo ti vadāmi. ↩

99. On the first four questions, and the textual tradition regarding the unanswered questions, see Collins 1982: 131, n. 1. ↩

100. Collins 1982: 131-133. ↩

101. Collins 1982: 133. ↩

102. Collins 1982: 132 describes these questions as “a standard type of ‘pointless speculation’.” ↩

103. Hamilton 2000: 174. ↩

104. AN IV.68.33ff. ↩

105. DN II.68.4ff. ↩

106. DN II. 68.11 ff . ↩

107. DN II.68.18: yāvat’ Ānanda adhivacanaṃ yāvatā adhivacanapatho yāvatā nirutti yāvatā niruttipatho yāvatā paññatti yāvatā paññattipatho yāvatā pañ̃̃ā yāvatā paññāvacaraṃ yāvatā vatṭam yāvatā vatṭam vattati, tad abhiñña vimutto bhikkhu. ↩

108. On this see Wynne 2007: 90ff. ↩

109. Gombrich 2009: 151. ↩

110. MN L185.14: katamāc’ āvuso paṭhavīdhātu? paṭhavīdhātu siyāajjhattikā siyā bahirā. katamā c’ āvuso ajjhattikā paṭhavīdhātu? yaṃ ajjhattaṃ paccattaṃ kakkhaḷaṃ kharigataṃ upādiṇnaṃ, seyyathīdaṃ: kesā lomā nakhā dantā taco maṃsaṃ nahāru aṭ̣̣hī aṭ̣̣himiñjā vakkaṃ hadayaṃ yakanam kilomakaṃ pihākaṃ papphāsaṃ antaṃ antagunaṃ udariyaṃ karīsaṃ, yaṃ vā pan’ aññam pi kiñci ajjhattaṃ paccattaṃ kakkhaḷaṃ kharigataṃ upādiṇṇaṃ, ayaṃ vuccat’ āvuso ajjhattikā paṭhavīdhātu. yā c’ eva kho pana ajjhattikā paṭhavīdhātu, yā ca bahirā paṭhavīdhātu, paṭhavīdhātur ev’ esa. taṃ: n’ etaṃ mama, n’ eso ‘ham asmi, na m’ eso attā ti evam etaṃ yathābhūtaṃ sammapaññāya daṭ̣̣habbaṃ. evam ataṃ yathābhūtaṃ sammapaññāya disvā paṭhavīdhātuyā nibbindati, paṭhavīdhātuyā cittaṃ virājeti. ↩

111. Hamilton (1996: 10) is correct to point out of that this list of bodily items in the Mahāhatthipadopama Sutta is “manifestly not comprehensive,” but this does not mean that the point is to “indicate examples of the characteristics being described” in order to emphasise “the characteristics and processes which enable the living body of the human being to function.” (Hamilton 1996: 12-13. The passage is entirely lacking in any words indicating that the contemplative bhikkhu’s attention should be focused on how the human body functions. The point is rather to go into enough detail so that the bhikkhu gets the reductionistic point that a self is not found in the body, despite the fact that every known part of it is not listed. ↩

112. MN I.190.15: seyyathā pi āvuso kaṭṭhañ ca pațicca valliñ ca pațicca tinañ ca pațicca mattikañ ca pațicca, ākāso parivārito agāran t’ eva sañkhaṃ gacchati, evam eva kho āvuso aṭthiṃ ca pațicca nahāruñ ca pațicca maṃsaṃ ca pațicca cammañ ca pațicca, ākāso parivārito rūpan t’ eva sañkhaṃ gacchati. ↩

113. MN I.190.28ff. ↩

114. MN I.190.35: so evaṃ pajānāti: evaṃ kira ‘mesaṃ pañcannaṃ upādānakkhandhānaṃ sañgaho sannipāto samavāyo hotī ti. vuttaṃ kho pan’ etaṃ Bhagavatā: yo pațiccasamuppādaṃ passati so dhammaṃ passati, yo dhammaṃ passati so pațiccasamuppādaṃ passatī ti. pațiccasamuppannaā kho pan’ ime, yadidaṃ pañca’ upādānakkhandhā. ↩

115. On the antiquity of these texts see Wynne 2007: 73. ↩

116. Sn 870: phassanidānaṃ sātam asātam, phasse asante na bhavanti h’ ete. vibhavaṃ bhavañ cāpi yaṃ etaṃ atthaṃ, etaṃ te pabrūmi itonidānaṃ. ↩

117. Sn 874: na saññasaññī na visaññasaññī, no pi asaññī na vibhūtasaññ̄̄: evaṃsametassa vibhoti rūpaṃ, saññānidānā hi papañcasañkhā. This statement reminds one of the similar statement in the Aggivacchagotta Sutta that the Buddha has “annihilated” form (rūpaṃ … anabhāvakatam, MN I.487.33). On this passage see n. 134 below and Wynne 2007: 95-96. ↩

118. Sn 915-916: pucchāmi taṃ Ādiccabandhuṃ, vivekaṃ santipadañ ca Mahesiṃ: kathaṃ disvā nibbāti bhikkhu, anupādiyāno lokasmiṃ kiñci? (915). mūlaṃ papañcasañkhāyā ti Bhagavā, mantā asmī ti sabbam uparundhe. yā kā ci taṇhā ajjhattaṃ, tāsaṃ vinayā sadā sato sikkhe. (916). ↩

119. Sn 849: vītatanho purā bhedā ti Bhagavā, pubbam antam anissito, vemajjhe nūpasamkheyyo tassa n’ atthi purekkhatam. ↩

120. Sn 851: nirāsatti anāgate atītaṃ nānusocati, vivekadassī phassesu ditṭhīsu ca na niyyati. ↩

121. Gómez 1976: 139: “When I first read the Mahāviyūha Sutta of the Suttanipāta I was impressed not only by its freshness and directness, but also by its originality.” ↩

122. See Wynne 2007: 127 for a summary of the evidence. ↩

123. See Wynne 2010. ↩

124. For this analysis see Wynne 2009a. ↩

125. Siderits (2007: 70-73) has argued that the Aggivacchagotta Sutta assumes a doctrine of realistic reductionism rather than ineffability. But this understanding cannot be derived from the Pāli text. Most importantly, when the Buddha claims that he, the Tathāgata, has “annihilated” the five aggregates (MN I.487.31), and so cannot be defined, instead of reading tathāgata Siderits reads the term arhat (2007: 71), and understands that it refers to a dead Buddhist saint. Thus he translates the Pāli yena rūpena Tathāgataṃ paññāpayamāno paññāpeyya taṃ rūpaṃ Tathāgatassa pahīnaṃ as follows: “all rūpa by which one could predicate the existence of the arhat, all that rūpa has been abandoned.” By shifting the focus from the living Tathāgata - i.e. the Buddha himself - to the dead arhat, the meaning of the passage is substantially changed. For it would now seem to be saying that nothing can truly be said about the dead saint because the five aggregates of which he was formerly constituted no longer exist. Thus Siderits notes that ” [t] he word ‘arhat is a convenient designator,’ just like ‘fire.’ So nothing we say about the arhat can ultimately be true. The only ultimately true statement about the situation will be one that describes the skandhas in a causal series … Does this mean that the arhat is annihilated - that nirvāna means the utter extinction of the enlightened person? No. There is no such thing as the arhat, so it lacks meaning to say that the arhat is annihilated. And for exactly the same reason, it lacks meaning to say that the arhat attains an ineffable state after death” (2007: 73). But since the passage is concerned with the living Buddha (Tathāgata) rather than the dead saint (arhat), it therefore is attempting to describe the state of being alive and yet liberated. Its negations, and the simile of the extinguished flame, indicate a completely apophatic understanding of religious experience that points towards a doctrine of ineffability. ↩

126. For a recent discussion of the philosophical value of early Buddhist thought, see Bronkhorst 2009: 1-7. ↩

127. Richard Gombrich has recently argued something along these lines (2009: 164): “I do not feel that Buddha was interested in presenting a philosophically coherent doctrine: the evidence that his concern was pragmatic, to guide his audience’s actions, is overwhelming. On the other hand, I have also concluded that the evidence that he had evolved such a structure of thought and that it underpinned his pragmatic advice is no less compelling.” ↩

128. SN V. 437 ff . ↩

129. MN I.137.1ff: tassa evaṃ hoti: ucchijjissāmi nāmassu, vinassissāmi nāma ‘ssu, n’ assu nāma bhavissāmīti. so socati kilamati paridevati urattāḷiṃ kandati sammohaṃ āpajjati. ↩

130. See the section at the beginning of MN I.140, where the Buddha rejects that he is a nihilist (venayika). ↩