Satipatthana and The evolution of the Dhamma Theory

Bhikkhu Sujato

 This essay is a response to and partial critique of Rupert Gethin’s article, ‘He who Sees Dhamma Sees Dhammas: Dhamma in Early Buddhism’[1]. Please note that it is a partial critique, for as usual, there is much in Gethin’s article that is to be welcomed. His discussion of the broader Indic connotations of dhamma is valuable, and certainly aids in capturing that ‘certain elusive coherence’. I do not attempt a full analysis of Gethin’s work, but concentrate on two things: firstly, questioning the methodology used, and secondly, considering some possible implications this may have for our understanding of the evolution of the concept of dhamma.

Gethin’s primary sources for early Buddhism are the Pali Nikāyas. He says in a footnote that:

‘… it would seem that any account of early Buddhist thought based on the Chinese Ᾱgamas would be essentially similar to an account based on the Pali Nikāyas. As Étienne Lamotte has observed, the doctrinal basis common to the Chinese Ᾱgamas and Pali Nikāyas is remarkably uniform; such variations as exist affect only the mode of expression or the arrangement of topics.’[2]

While this is generally true, as we shall see, sometimes a rearrangement of topics, though seemingly innocuous, has unexpected implications. In any case, it is not best practice to rely on such a generalization when conducting an inquiry into a specific field or text.


For example, one of the key controversial attributes of dhammas according to the Sarvāstivādins was that dhammas exist in the past, future, and present. It has been noticed that the Sarvāstivādin Ᾱgamas – specifically, the Saṁyukta and Madhyama preserved in Chinese translation[3] – frequently say that various dhammas have existed in the past, will exist in the future, and do exist in the present. While similar phrases are found in several places in the Theravāda also, the Sarvāstivāda suttas seem to use such sayings more frequently, and sometimes with a more pregnant sense. I do not know of a passage where such a formula is applied to the term dhamma specifically, although it is certainly used of things that are dhammas, such as the four noble truths or the five aggregates. For example, Thich Minh Chau points out that the Sarvāstivādin version of the Saccavibhaṅga Sutta (MA 31), has the following phrase, not found in the Theravādin version of this sutta:

‘Thus this noble truth of suffering has existed in the past, is existing in the present, and will be existing in the future… ’[4]

Thich Minh Chau concludes that such statements are sectarian, implying a nascent tendency towards the Sarvāstivādin perspective even within the Ᾱgamas.[5] Another example is found in the Sarvāstivādin version of the Khandha Saṁyutta. Here I repeat Choong Mun-keat’s comments:[6]

Yin Shun maintains that the passages quoted below contain evidence of sectarian teachings. SA 79 (no SN [Pali] counterpart) contains the following wording:[7]

‘ … because material form (also feeling, perception, activities, consciousness) exists in the past … because material form exists in the future … because material form exists in the present … ’

Also, in three other SA discourses, but not in their SN counterparts, the word ‘exist’ is found at the end in a statement that it should be substituted in the text, as follows:[8]

‘ … As for “should be taught”, so also for “exist”, and “should be known”’.

Yin Shun suggests that the above expressions found in the SA are connected with the Sarvāstivāda emphasis on ‘existence’ in past, future, and present time. By contrast, in SN 22.62, the discourse called Niruttipatha ‘Mode of Expression’, the Buddha explains in detail the need to distinguish the three times: any material form (feeling, perception, activities, consciousness) in the past is to be spoken of as ‘existed’ (ahosi); in the future as ‘will exist’ (bhavissati); in the present as ‘exists’ (atthi). Yin Shun ponts out that this discourse, which has no SA counterpart, is affirming the theory of ‘the existence of present time’, a teaching of the Tāmraśāṭīya tradition (Pali Buddhism).

So there is good reason to suspect that differing sectarian perspectives of the ontological status of dhamma may be discerned in the Ᾱgamas and Nikāyas. All Buddhist schools hold that the four noble truths are equally valid at all times. They are, in a sense, permanent, although they are not permanent entities. Like the laws of physics, they describe how things behave, and they tell us that the world tends to work in a predictable manner. While the details of how this manifests will never be identical, still the fundamental pattern remains constant: in the past, suffering was caused by craving; in the future, suffering will be caused by craving; and right now, suffering is caused by craving. But in such passages as the above we can discern a tendency for this ‘timelessness’ of the four noble truths – the ‘Dhamma’ – to be applied to the aggregates – the ‘dhammas’. The ambiguity in the meaning of dhamma directly contributes to the emerging Sarvāstivādin conception of time. It would seem crucial, then, to consider as best we can the different sectarian versions of the early Suttas in this context.


Before dealing with the Satipaṭṭhāna Sutta, Gethin outlines a number of important usages of dhamma. There are a number of points in Gethin’s presentation that I would take issue with. The first is the description of –dhamma as the second element in a compound meaning ‘a particular nature or quality possessed by something’. The only examples Gethin gives, which are the most important doctrinally, concern the ‘nature’ of things to arise and cease. The ‘origination-dhamma’ of something is not a event, not a simple arising, but an inference, a generalized understanding that all phenomena of the same type will behave in a similar way. It strikes me as odd to say that ‘origination’ is a nature ‘possessed’ by something, which seems to reify English grammatical usage in a way not suggested by the Pali (‘The nature of things is to arise and cease’). It would seem more apt to say dhamma here signifies ‘the way things occur’, or more technically: ‘the principle of how phenomena function’. These considerations, of course, are not an attempt at translation, and I am happy to translate samudayadhamma as ‘nature of origination’, as long as we are careful not to take the mere grammar of the phrase too seriously.

A second problem I have is with Gethin’s description of dhamma as a mental or physical state, thing, or quality. None of these terms strike me as particularly useful in this context. I have long been perplexed at why so many translators use ‘state’ for dhamma. My Oxford tells me that the basic meaning of ‘state’ is : ‘the existing condition or position of a person or thing (in a bad state of repair; in a precarious state of health)’. This has nothing to do with any of the meanings of dhamma. Similarly, ‘quality’ is given as, firstly ‘the degree of excellence of a thing’ (of good quality; poor quality)’ or related meanings. These would suit only in contexts involving a value judgement, such as considering good and bad dhammas. The more relevant meaning is ‘a distinctive attribute or faculty; a characteristic trait’. This seems to imply a distinction between a substance and it’s attributes, which is to say the least, problematic, and is certainly not suggested by Gethin’s contexts. As for ‘thing’, really, this is so hopelessly vague – can’t we do better than this? I would prefer to follow the lead of many other translations and use ‘phenomenon’ in this sense. This has a fairly well-defined meaning as an occurrence that is perceived or noticed by the senses or the mind. As long as we do not import more context-specific implications from certain western philosophers, I think ‘phenomenon’ is reasonably accurate in this context.

In both of these cases, Gethin’s wording is suggestive of the classical distinction between a substance and its attributes. A dhamma is a nature ‘possessed’ by a thing; or a dhamma is a quality (of a thing?). Such usages are difficult to avoid, and as Gethin does not make a substance/attribute distinction explicit, I presume this is a more-or-less casual use of language. But it is precisely such casual usages that, when assumed as ‘canonical’, get taken with absolute literallness and are read back into reality.


Of more urgent practical relevance is the treatment of dhamma within the fourth satipaṭṭhāna, ‘contemplation of dhammas’. Almost every scholar who has studied satipaṭṭhāna has committed the unforgivable heresy of regarding ‘the’ Satipaṭṭhāna Sutta as the be-all and end-all of satipaṭṭhāna. Here Gethin follows suit, and uses the presentation of dhammas in the ‘contemplation of dhammas’ section of ‘the’ Satipaṭṭhāna Sutta to draw important conclusions for his work. But a series of scholars (Warder, Schmithausen,[9] Bronkhorst,[10] and others) have noticed important differences in the various versions of the Satipaṭṭhāna Sutta, including the section under consideration here. The question has been of such contemporary interest that the Chinese versions of the Satipaṭṭhāna Sutta have been translated into English at least four times. The field continues to generate interest: Anālayo considers the three sutta versions in his forthcoming comparative study of the Pali and Chinese Majjhimas; my own A History of Mindfulness considers the topic in detail; and I am aware of several forthcoming theses on the topic.

There are a few questions about the contemplation of dhammas in satipaṭṭhāna that we can clear up immediately. The first is: has there been historical change in the text? The answer is yes. The Burmese version of the Pali canon has displaced the Satipaṭṭhāna Sutta (M10) with the Mahāsatipaṭṭhāna Sutta (D22). The Satipaṭṭhāna Sutta no longer exists in the Burmese canon. This difference is mentioned in the PTS Pali edition, published in 1888, so stems from the 19th Century at least; perhaps the change was made at the Burmese so-called ‘Fifth Council’. The second question is this: has the change any sectarian agenda? Again, the answer is yes. As is well known, those who have presided over the revision and distribution of the Burmese version of the Pali canon are enthusiastic advocates of a meditation technique that they call vipassanā, which claims the (Mahā) Satipaṭṭhāna Sutta as its primary source text. Clearly, the insertion of the Mahāsatipaṭṭhāna Sutta into the Majjhima Nikāya was intended to bolster the prestige of this school by making its text the longest in the Majjhima.

These rather disturbing considerations make it imperative that we should clarify exactly what we are talking about when we refer to ‘the’ Satipaṭṭhāna Sutta. The first point is that there are several versions of ‘the’ Satipaṭṭhāna Sutta, and there is no a priori reason why any one should be more reliable than any other. They can be summed up as follows. For the Pali, the references are to the Pali Text Society’s Roman script editions. For the Chinese, I use the Taisho references in the CBETA digital Tripitaka. For the Prajñāpāramīta, I will have to remain content with giving the reference to Conze’s translation, as it is beyond me to chase down all the references in the vast mass of Prajñāpāramīta literature.







Satipaṭṭhāna Sutta



MN 10 (except Burmese Tripitaka)

Majjhima Nikāya 1.55-61

Mahāsatipaṭṭhāna Sutta



DN 22 (also MN 10 in Burmese Tripitaka)

Dīgha Nikāya 2.290-315

Smṛtyupasthāna Sūtra



MA 98

T I 582b-584b

Ekāyana Sūtra



EA 12.1

T II 568a-569b





Chapter 7, pg. 192




T 1537

T XXVI 475c24-477c9




T 1548

T XXVIII 612b26-616c8

Prajñāpāramīta Sūtra


Chinese, Tibetan, Sanskrit


Ch 16, pp153-5, Conze’s trans.


Each of these contains recognizable features that enable them to be clearly recognized as a version of the ‘Satipaṭṭhāna Sutta material’. While some, such as the Prajñāpāramīta, come from later works, the satipaṭṭhāna material has clearly been ‘cut-&-pasted’ with a minimum of alteration, so there is no reason why it should not preserve genuine early features. However, the Prajñāpāramita only treats body contemplation in detail, so for the purpose of this essay it may be left aside.

We are then left with seven sources for the contemplation of dhammas. We may treat the Theravādin medium length Satipaṭṭhāna Sutta and long Mahāsatipaṭṭhāna Sutta together, as the list of contents in dhamma-contemplation is identical, the only difference being that the long version greatly expands the description of the four noble truths. Here are the lists of contents in each version:



(Mahā-) Satipaṭṭhāna Sutta

Sarvāstivāda Smṛtyupasthāna Sūtra


Mahāsaṅghika Ekāyana Sūtra

Theravāda Abhidhamma Vibhaṅga

Sarvāstivāda Abhidharma Dharmaskandha

Dharmaguptaka Abhidharma Śāripūtrābhidharma


6 sense fields






5 hindrances




5 hindrances


5 hindrances

(in introduction and conclusion)

5 hindrances


5 hindrances


5 hindrances


5 aggregates






6 sense fields




6 sense fields


6 sense fields


7 awakening-factors

7 awakening-factors

7 awakening-factors

7 awakening-factors

7 awakening-factors

7 awakening-factors

4 jhanas

4 noble truths

4 noble truths


Each of these contains several items under ‘contemplation of dhammas’, and these items share much in common. Clearly, the sects were talking about the same kinds of things. Yet each version has its own peculiarities. There are two possible explanations. One is that there is simply some flexibility as to what this section includes. There is every reason to suppose that the Buddha in his long career taught satipaṭṭhāna from several perspectives and the traditions from the earliest times may have recorded different lists of specifics in this context, without implying that one or other was final and definitive. The alternative hypothesis is that there was one version that was the ‘original’ version of contemplation of dhammas, and that the other versions represent developments by the schools. While it is never possible in such cases to come to an absolute decision between these alternatives, I believe there are a number of considerations that, taken together, make it very likely that in this case the second alternative is correct, and moreover, that we can reconstruct the original form with a reasonable degree of certainty.


We should first get some idea of what is at stake. Gethin takes for granted the Theravāda Satipaṭṭhāna Sutta in presenting dhammas as including the five hindrances, the five aggregates, the six sense fields, the seven awakening-factors, and the four noble truths. Now, in the Theravāda Satipaṭṭhāna Sutta, each of the meditation practices is followed by a ‘vipassanā refrain’, which urges us to contemplate each object in terms of its dhamma (nature) to arise and cease. Gethin connects this meaning of –dhamma as the second element of a bahuvrīhi compound in the sense of ‘a particular nature or quality’ with the meaning in the contemplation of dhammas. He argues that dhammas in satipaṭṭhāna cannot mean ‘teachings’, ‘practices’, ‘truths’, or ‘laws’. It can only mean ‘basic qualities, both mental and physical’, or else a ‘mental or physical state or thing’. And it is on this basis that he suggests that this is the most prevalent usage in the Nikāyas, closely anticipating the similar though more technical usage in the Abhidhamma.

One of Gethin’s assumptions seems to be that dhammas in the ‘contemplation of dhammas’ has essentially a uniform meaning. This is by no means necessarily the case – the term dhamma is notoriously ambiguous, and might have been chosen for this very reason. In his preliminary discussion, Gethin settles on about six major meanings of dhamma: teaching; good conduct or behaviour; truth; nature; natural law; mental or physical state or thing. He assigns the dhammas of the Satipaṭthāna Sutta to the category of ‘mental or physical state or thing’.

This causes a rather curious ambiguity in the treatment of the seven awakening-factors. Gethin extends his category of dhamma as ‘good conduct or behaviour’ to include some or all of the dhammas that come under the heading of the 37 bodhipakkhiya dhammas.[11] Gethin says that when the Buddha is referring to the 37 bodhipakkhiya dhammas, he is not talking about ‘teachings’, but about ‘practices’. The quote he gives in support of this is the famous exhortation by the Buddha in the Mahāparinibbāna Sutta for the monks to maintain the holy life through the practice of these dhammas. But while it is quite true that these things are ‘practices’, it is not so clear that they are not also ‘teachings’. In fact, at least two Sanskrit versions of the same passage have the Buddha exhorting the monks to ‘recite’ these dhammas.[12] This is obviously referring to a formalized body of scripture. Currently, almost all the discourses dealing with the 37 bodhipakkhiya dhammas are collected in the Mahā Vagga (or ‘Magga Vagga’) of the Saṁyutta Nikāya/Āgama. I have argued elsewhere that these passages, whether authentic sayings of the Buddha or not, are likely to have been inserted for the purpose of authorizing an early collection of discourses from which the existing Saṁyuttas are descended. So here, as so often, we see a tendency for the different aspects of dhamma to be conflated, to the endless frustration of we who would like it all to be nice and tidy!

Despite relying on the above-mentioned Mahāparinibbāna Sutta passage, which forcefully instructs the monks to ‘practice’ the 37 bodhipakkhiya dhammas, Gethin asserts that the seven awakening-factors, which are included in the 37, are not practices.[13] This conclusion is somewhat bizzare. Of course, the awakening-factors are also ‘mental qualities’, and can equally well be described as ‘phenomena’, but Buddhist meditation is precisely concerned with ‘practicing’ to develop appropriate ‘mental qualities’ and to understand ‘phenomena’. These meanings are all overlapping, and the fact that a given dhamma is describable in one way does not preclude it also fitting elsewhere. The confusion here seems to stem from Gethin’s assumption that all the categories listed under ‘contemplation of dhammas’ actually belong there, and moreover, that they are all dhammas in the same sense, the sense required to support his argument.

The problem is resolved when we recognize that the list of factors in ‘contemplation of dhammas’ falls into three rather different categories:

  1. ‘Qualities of mind’: The hindrances and awakening-factors are qualities of mind, imbued with an ethical valuation. They are either good or bad, and we are given a prescription for what to do about them – get rid of the bad, and develop the good.
  2. ‘Phenomena’: The aggregates and sense fields are alternative descriptions of reality as it is experienced. The prescriptive aspect is implicit, but the main emphasis is to provide the meditator with a cool, clear analytic tool for vipassanā meditation. Even though most translators render dhammas here as ‘mental objects’, the aggregates and sense fields clearly include both physical and mental phenomena, and the mental phenomena encompass both the objective and subjective aspects of mind.
  3. ‘Natural principles’: The four noble truths are an all-encompassing group that incorporate both a description of how reality is, and a value-based prescription of what to do about this situation. They are set of universal laws or principles (although not the kind of law that controls phenomena, making them happen in a certain way). Within the four noble truths, the hindrances and awakening-factors pertain to the fourth noble truth, the path, and hence are ‘to be developed’; while the aggregates and sense fields pertain to the first truth, the reality of suffering, and hence are ‘to be fully known’.

Now of the three meanings which, if any, is more intrinsic to the contemplation of dhammas? Consider the following points.


  1. While all the versions differ to some degree, they all contain the hindrances and awakening-factors. Moreover, while the other items vary in their positions, these two always occur in the same sequence. There is one exception to this: the Ekāyana Sūtra has the four jhanas instead of the hindrances. But this version places the hindrances at the beginning and the end of the sutta, and I believe that this represents a later, independent development, emphasizing the importance of eradicating the hindrances as the ‘alpha & omega’ of satipaṭṭhāna practice.
  2. One version, the Theravāda Abhidhamma Vibhaṅga, contains only the hindrances and awakening factors. This is the simplest of all the versions in its basic specification. For example, it omits the introductory and closing questions: ‘And how does bhikkhu dwell contemplating a dhamma in the dhammas in terms of the five hindrances?’ The specification of body contemplation is similarly primitive. Of course, this simplicity only applies to the basic specification stripped of the developed abhidhamma analysis.
  3. The hindrances and awakening-factors are found together with satipaṭṭhāna in many places, including the Satipaṭṭhāna Saṁyutta, and are a special feature of the Bojjhaṅga Saṁyutta.[14] But a survey of the relevant Saṁyuttas reveals that there is no close relationship between satipaṭṭhāna and the aggregates, sense fields, or truths. In fact, whether in the Saṁyutta or elsewhere, satipaṭṭhāna is almost never mentioned in these contexts.
  4. The Samudaya Sutta (SN 47.42/SA 609) of the Satipaṭṭhāna Saṁyutta (both Pali and Chinese) says that the origin of dhammas in satipaṭṭhāna is attention (manasikāra). While this could apply to the truths,[15] it obviously cannot apply to the aggregates and sense fields. On the other hand, the Bojjhaṅga Saṁyutta contains many suttas that say that unwise attention gives rise to the hindrances, and wise attention gives rise to the awakening-factors. Clearly, this is exactly what the Samudaya Sutta is implying. In case we have any doubt on this point, the commentary explains it in just this way:

Originated by attention’ means: here, due to the origination of wise attention is the origination of the awakening-factor dhammas, due to the origination of unwise attention is the origination of the hindrance dhammas.[16]

This last point is really the clincher. The other considerations carry weight, though they might also be interpreted differently; but I don't see how anyone could reasonably object to the commentarial interpretation. I can only conclude that the original specification of contemplation of dhammas was the hindrances and awakening-factors. While such conclusions can never be certain, this confluence of strong indicators is about as good as we can hope for.

Notice that the hindrances and awakening-factors fall together under the first meaning of dhamma mentioned above: good and bad qualities of mind. This is another reason for thinking they belong together here. Another meaning of dhamma, however, also emerges from the manner in which these factors are treated: each is considered in terms of how it comes to arise and pass away. This implies that dhamma here merges into the meaning of ‘natural principle’, the ‘laws’ of cause and effect. It would seem, then, that the other meaning of dhamma – the phenomena of reality, as exemplified especially by the aggregates and sense fields – is secondary and probably a later interpolation. This leads to a number of conclusions.


The Theravāda assigned their Satipaṭṭhāna Sutta to the Majjhima, and an expanded version to their Dīgha. The Sarvāstivāda also put their Smṛtyupasthāna Sūtra in their Madhyama, while one version found its way into an Ekottara (=Aṅguttara). It would seem, then, that its location within the Nikāyas/Āgamas is entirely promiscuous. On the basis of other research, however, I believe that many, perhaps all, of the suttas that deal with Saṁyutta-style topics were originally included in the (proto) Saṁyutta and were later moved out. In terms of textual interpretation, we have already found that the key to interpreting the ‘contemplation of dhammas’ is to be found in the Saṁyuttas. There are other specific textual reasons for believing that the Satipaṭṭhāna Sutta, at an earlier stage, belonged with the other discourses on satipaṭṭhāna that are today found in the Saṁyutta.

The clearest is that the basic satipaṭṭhāna formula in the Sarvāstivāda Smṛtyupasthāna Sūtra is abbreviated. The Theravāda suttas on satipaṭṭhāna invariably include a standard ‘auxulliary formula’ describing the state of mind of the satipaṭṭhāna meditator. Here is the standard Pali version:

… kāye kāyānupassī viharati: ātāpī, sampajāno, satimā, vineyya loke abhijjhādomanassaṁ.

‘One dwells contemplating a body in the body: ardent, clearly comprehending, mindful, having removed covetousness and displeasure for the world.’

But the Sarvāstivāda suttas usually do not include any corresponding phrase. The Sarvāstivāda Smṛtyupasthāna Sūtra version is very terse, just six characters:


‘The satipaṭṭhāna of contemplating body as body.’

The explanation for this difference is found in the Sarvāstivāda Smṛtyupasthāna Saṁyukta. There, the description of satipaṭṭhāna is usually brief, as given above. But some suttas give in full a formula that corresponds with the Pali, though it is not necessarily identical. Existing Sanskrit (probably Sarvāstivāda) versions of the full formula are, however, very similar to the Pali,[18] so, given that the Saṁyukta and Madhyama found in Chinese translation are also from the Sarvāstivāda or a closely related school, it seems likely that the differences in the Chinese formulas are the result of translation and/or abbreviation.

At the end of one of the suttas of the Smṛtyupasthāna Saṁyukta (SA 612) there is the following note:

In all suttas on the four satipaṭṭhānas, the following phrase should be generally applied, that is to say: ‘Monks, regarding the practice of the four satipaṭṭhānas, you should give rise to higher aspiration and diligent effort, with right mindfulness and right knowledge [clear comprehension]. This is how you should train.’[19]

Now, the Saṁyukta has over fifty discourses on satipaṭṭhāna one after the other, all featuring the same formula. It makes perfect sense to abbreviate the repetitious part of the suttas, then give instructions on how they should be expanded in full (even though these instructions are buried away rather curiously in the middle of the collection). But in the Majjhima, the Smṛtyupasthāna Sūtra is isolated, and the satipaṭṭhāna formula occurs rarely if ever in the rest of the Āgama. There is no reason why the formula should be abridged, and no guidance as to how to expand it. This anomaly is easily explained if we postulate that the Smṛtyupasthāna Sūtra originally belonged in the Saṁyukta, but was later moved and ‘padded out’ for inclusion in the Majjhima. At the very least, this suggests we should look in the Saṁyutta for other clues for interpreting satipaṭṭhāna.


In terms of doctrinal evolution, we have discovered that the original meaning of dhammas in satipaṭṭhāna was good and bad qualities of mind or principles. This later moved towards the meaning of ‘phenomena’; in other words, it tended towards the Abhidhamma meaning. The Abhidhamma connection is, in fact, quite explicit. The Mahāsatipaṭṭhāna Sutta is distinguished from its shorter Majjhima brother by the inclusion of a long explanation of the four noble truths. This explanation is derived from the Saccavibhaṅga Sutta (MN 141), except the treatment of the second and third noble truths is considerably expanded. There is no exact version of this passage in the Nikāyas, and for parallels, we must look to the Abhidhamma Vibhaṅga.[20] There the exposition of the truths is essentially identical with that in the Mahāsatipaṭṭhāna Sutta (though there are a couple of minor points expanded further in the Abhidhamma). This makes it certain that the contemplation of dhammas in satipaṭṭhāna was, from early times, tied up with the evolving Abhidhamma conception of dhammas.

Now, to appreciate what this means we must first consider what kind of text the Mahāsatipaṭṭhāna Sutta really is. Is it, as so often claimed, the most important of all discourses, the greatest of the Buddha’s authentic meditation teachings, the ‘heart of Buddhist meditation’? Sadly, the answer must be no. In fact, we must question whether it is an authentic discourse at all. The Dharmaguptaka version of the Dīgha, preserved in Chinese translation, has no equivalent for the Mahāsatipaṭṭhāna Sutta. They certainly had a ‘Satipaṭṭhāna Sutta’, for their Abhidharma preserves a version of the satipaṭṭhāna material (which is close to the Pali). The newly discovered Sarvāstivāda Dīrgha also lacks a Mahāsatipaṭṭhāna Sutta. The absence from these two collections, of schools quite closely related to the Theravāda, strongly suggests that the Mahāsatipaṭṭhāna Sutta is an independent Theravāda development, a compilation of the Satipaṭṭhāna Sutta with proto-abhidhamma material. The extra material is entirely included in the contemplation of dhammas. Thus to use this text to argue for the similarity between the Nikāya and Abhidhamma usage of dhamma is flawed.


In fact, our historical analysis leads us to exactly the opposite conclusion. While not denying Gethin’s conclusion that the Nikāyas/Ᾱgamas frequently use dhamma to mean something like a basic quality (although as noted we prefer ‘phenomenon’), the usage in satipaṭṭhāna reveals an increasing tendency for this meaning to override what, in this context, is the earlier and more important meaning. The original specification of contemplation of dhammas, according to us, was an ethical injunction: it was a practice, an ought-to-do, based on an understanding of how things work, the principles of causality. Later a more ontological emphasis emerged: a highly detailed and technical listing of what there is. This is precisely where the argument about dhammas raged in ancient India. The critics alleged that the Buddha’s original injunction to seek freedom through meditation and examining one’s own mind had been usurped by a nit-picking, pedantic obsession with categorizing and sorting out the details of the elements of so-called ‘ultimate reality’. In embryonic form, this is exactly the kind of movement we have seen in the ‘contemplation of dhammas’ section of the Satipaṭṭhāna Sutta. The connection is implicit in this key passage from the Saṁdhinirmocana Sūtra, a central source work for the Yogacāra school of Mahāyāna, probably composed in the Second Century CE:

‘The Venerable Subhūti addressed the Blessed One and said: “World-honored One, in the world of sentient beings I know a few who state their understanding without pride, but I know innumerable, untold sentient beings who cherish their pride and state their understanding in a prideful manner. World-honored One, once I was dwelling in a grove in a forest. A large number of monks lived nearby. I saw them assemble after sunrise to discuss various issues and to propose their own understandings, each according to his insight.

“Some proposed their understanding of the aggregates, their descriptive marks, their arising, their exhaustion, their destruction, and the realization of their destruction. Other, in like fashion, proposed their understanding of the sense fields… dependent origination… sustenance… truth… realms [dhātus]… the stations of recollection [satipaṭṭhānas], their descriptive marks, the states they are able to control, their cultivation, their arising from a state of being non-arisen, their assured non-forgetfulness after arising, and their increase from repeated practice. [also the right efforts, etc.]…

‘… Each one of them cherished their pride… and were unable to comprehend the one universal taste of the truth of ultimate meaning… [21]

As with many Mahāyāna sūtras, this makes little sense considered as an authentic teaching of the Buddha (because some of the terminology is clearly later and derived from the Abhidhamma), but makes complete sense if considered as a critique of the academic Buddhist scene that dominated around the start of the Common Era. Whether coincidence or not, the phrasing of this passage closely reflects the phrasing of the contemplation of dhammas in the Satipaṭṭhāna Sutta. As I have shown elsewhere, the list of topics here is derived specifically from the Sarvāstivāda Saṁyukta Ᾱgama; but the manner of phrasing may well be influenced by the Satipaṭṭhāna Sutta. If it is meant as a critique of the developed conception of satipaṭṭhāna as represented by the existing texts, this raises the question whether the Satipaṭṭhāna Sutta may be considered a meditation text at all, or if it is really a doctrinal compilation.


The difficult question of dates may offer some clarification here. We have noticed that the Dīghas of the Dharmaguptakas and the Sarvāstivādins lack a Mahāsatipaṭṭhāna Sutta. The absence from the Dharmaguptaka is particularly important. The most persuasive presentation of the origin of this school was given by Frauwallner.[22] He argued that several of the main ‘Vibhajjavāda’ schools originated as missions sent out from central India around the Aśokan period. Among others, Mahinda and Saṅghamittā were sent to the south and established the Mahāvihāravāsin school in Sri Lanka, while Yonaka Dhammarakkhita went to the West and founded the Dharmaguptaka school. This suggests that there was never a real schism dividing these schools, and substantive differences between the texts of the two schools may be post-Aśokan. In the case of the Mahāvihāravāsins, these may be Sri Lankan in origin. In fact, the Sinhalese commentaries openly admit that some of the latest additions to the Dīgha were made in Sri Lanka, specifically the closing verses of the Mahāparinibbāna Sutta.

While the details are too complex to go into here, I think there is some reason to believe that the final redaction of the Mahāsatipaṭṭhāna Sutta was as late as 20 BCE, when, according to the Chronicles, the canon was written down in the Ᾱloka Vihāra. It was there that the Mahāvihāravāsins made the momentous decision, never since reversed, that their main emphasis would be on preservation, study, and teaching of texts, rather than the practice of their contents. Thus from the Mahāvihāravāsin’s own assertion we find confirmation of the accusations leveled by such texts as the Saṁdhinirmocana Sūtra. And if I am correct in deducing that the Mahāsatipaṭṭhāna Sutta dates from the same period and the same group of monks, it seems that here we are dealing with an essentially scholastic exposition of meditation.


Now it is clear from the foregoing that the contemplation of dhammas has been an important part of the development of the conception of dhamma, for ancient Buddhists as well as modern scholars. This process continued as the Abhidhamma evolved. In the early Sarvāstivāda Abhidhamma period, the Dharmaskandha correlates the four satipaṭṭhānas with the five aggregates: body = form; feeling = feeling; mind = consciousness; and dhammas = perception and activities (saññā, saṅkhārā).

This last equation is incongruous to say the least, if considered as a genuine description of contemplation of dhammas as found in the existing suttas. How can, say, the eye, or the four noble truths, be construed as perception and activities? Obviously they can’t, and this suggests an early confusion as to what exactly is going on. It would seem that the correlation of the aggregates with the satipaṭṭhānas is not really intended to explain the satipaṭṭhānas, but as part of the Abhidhamma project of mapping the various teaching frameworks on top of one another.

What is odd here is that the five aggregates pertain to the first noble truth, suffering, while the satipaṭṭhānas pertain to the fourth noble truth, the path. But here the satipaṭṭhānas are being gradually subsumed under the first noble truth, that is, moving from a description of how to practice to a description of what really exists.

This definition was later adopted across the traditions, including the Theravāda, which caused a further problem. For the Theravāda, alone among the existing versions, the contemplation of dhammas itself includes the five aggregates: how then can it be said to be just two of them?

The whole situation seems odd, and we might wonder how it arose. Again, the history of the texts provides a good answer. The definition first occurs in the Dharmaskandha, which lists under dhamma-contemplation just the hindrances, sense fields, and awakening-factors. Thus it does not include the aggregates within the contemplation of dhammas, escaping at least the obvious incongruity of explaining five aggregates as equal to two aggregates. The Dharmaskandha exposition is very similar to the Vibhaṅga, which has just the hindrances and awakening-factors. We may suspect that dhammas were first defined as perception and activities when the content of the section was just the hindrances and awakening-factors. While this is not obviously ‘correct’ as an interpretation, at least it is less incongruous.

A further significance of this passage in the Dharmaskandha is this: the Dharmaskandha follows each exercise in satipaṭṭhāna with a ‘vipassanā refrain’, seeing each object as impermanent, suffering, a disease, etc. Now, the Theravāda Satipaṭṭhāna Sutta also includes a vipassanā refrain, contemplating the ‘origination-dhamma’, ‘cessation-dhamma’, etc., following each exercise. This vipassanā refrain is not found in all versions, and we conclude that it is a later insertion, formed by splicing in a passage from a different sutta in the Satipaṭṭhāna Saṁyutta (SN 47.40, which has no Sarvāstivādin counterpart; the vipassanā refrain is found instead in the Sarvāstivādin version of the Samudaya Sutta, SA 609). There is no vipassanā refrain in the Sarvāstivāda Smṛtyupasthāna Sūtra, and the vipassanā refrain in the Abhidhamma of the same school is clearly later, having probably been imported from the Khandha Saṁyutta. It is quite different to the vipassanā refrain in the Theravāda and cannot have come from the same source. Nevertheless, we see an increasing tendency to see satipaṭṭhāna in terms of vipassanā. This parallels the change in meaning of dhamma. As the objective ‘what there is’ meaning of dhamma gains in importance, the sutta becomes more and more oriented towards vipassanā. (The major theme of A History of Mindfulness is that the primary context of satipaṭṭhāna in the Nikāyas/Ᾱgamas is samatha rather than vipassanā.)


Gethin wishes to counteract the famous Nagarjunian critique of the developed Abhidhamma conception of dhammas, especially in the Theravādin context, and possibly generally. He quotes the commentarial definition that dhammas are so-called because ‘they maintain (dhārenti) their own particular natures (sabhāva), or because they are maintained (dhārīyanti) by causal conditions’. Gethin says that this definition should be understood as a ‘direct and deliberate counter to the idea of dhammas as “particular natures” that are “maintained” by an underlying substance (dharmin) distinct from themselves: it is not intended to define dhammas as ontologically irreducible entities.’[23] He also says that this definition ‘is not a statement about their ontological status’.

Gethin’s position cannot be sustained in light of how the commentarial tradition itself describes sabhāva. Basic information on this can be gleaned most easily from Ñāṇamoḷi’s footnote 68, chapter 8 of the Path of Purification, his translation of the central Theravāda text, the Visuddhimagga. The Theravāda considers both sabhāvadhammas and asabhāvadhammas. The sabhāvadhammas are the familiar dhammas of Theravādin theory, the mental and physical elements. In other words, the sabhāvadhammas exist in the ‘ultimate’ (paramattha) sense. The asabhāvadhammas include space, and especially concepts, including all objects of meditation that can produce jhana; generally, then, these equate with those things that are considered to exist only ‘conventionally’ (sammuti).

Sabhāva itself is glossed with such terms as bhavana, vijjamānatā, labbhamānarūpa, and especially: A dhamma’s ‘own essence or existing essence’ (sako vā bhāvo samāno vā bhāvo). These terms all have an ontological import, and none of the definitions quoted by Ñāṇamoḷi describe sabhāva as merely characteristics or qualities (nimitta, liṅga, etc.).

The sabhāvas are frequently mentioned in meditation contexts, and sometimes this reinforces our impression that they are primarily ontological. For example, the Visuddhimagga discusses which meditation signs can be ‘extended’; that is, consciously enlarged due to the manipulative power of the yogi’s concentration. But ‘it is not possible to extend a sabhāvadhamma’.[24] Why might that be? We can only make sense of such statements through an understanding of the underlying assumptions. In philosophical perspectives of the relation between the mind and its objects, there are three basic positions:

  1. The mind exists ‘in here’ and the objects exist independently ‘out there’ (‘naïve realism’).
  2. The mind is all there is and objects are an illusion (‘idealism’).
  3. The mind and its objects occur interdependently (which, in my opinion, is the Buddha’s position).

In the second and third cases there would seem to be no problem in extending the object of meditation, since one is, after all, only altering the mode of perception. Only if the object has an independent, inherent existence ‘out there’ would it be the case that one could not extend the object.

This is, of course, the basic ontological presupposition of most science (with the notable exception of quantum theory): the object measured is independent of the measuring apparatus. In science, the measurer is ultimately the scientist themselves; in meditation it is the mind of the yogi. Just as the scientist believes they cannot alter the results of their experiment merely by altering their consciousness, Buddhaghosa believes that the size of a meditation object that is a sabhāvadhamma cannot be altered merely by the yogi altering their consciousness. Just as the scientist’s chemicals and test-tubes really exist ‘out there’, so the yogi’s meditation object really exists as a sabhāvadhamma. Thus, such contexts clearly assume some kind of ‘naïve realism’ in their ontology.


There is a serious problem with the sabhāva theory when we come to consider the status of Nibbana. It is a sabhāvadhamma, that is, it really exists in the ultimate sense. Nibbana could hardly be excluded from those things that exists ‘ultimately’, for in the Suttas it is the only thing described as ultimate truth (paramattha sacca), although there the meaning is rather ‘highest goal’; or else it is true in the sense of the ending of delusion. Be that as it may, the commentaries are in the unenviable position of putting Nibbana on essentially the same ontological footing as conditioned dhammas. The sub-commentary addresses this issue by appealing to the profundity of Nibbana, which explains nothing, merely draping the problem in a mystical shroud.

But the idea of Nibbana as sabhāvadhamma directly contradicts the basic commentarial definition of dhamma we started with. To repeat: dhammas are so-called because ‘they maintain (dhārenti) their own particular natures (sabhāva), or because they are maintained (dhārīyanti) by causal conditions’. The second part of this explanation appeals to the conditioned nature of dhammas, which obviously cannot apply to Nibbana. Hence this definition cannot be consistently applied to those things that are regarded by the commentary as sabhāvadhammas. The second part of the definition is dispensable and is not intrinsic to the definition itself.

From a Nagarjunian perspective one would simply scoff and say, of course! The definition was utterly incoherent from the beginning. How can we, on the one hand, assert the individual maintenance of the sabhāvas, and on the other, say they are conditioned? The two notions are essentially incompatible, and the notion of sabhāva is irrelevant to a description of the four noble truths and dependent origination.

Gethin further supports his non-ontological interpretation of sabhāva by pointing out the functional definition of many dhammas in the commentaries. It is contact (phassa) because it contacts (phusati), it is will (cetanā) because it wills (cetāyati), etc. This mode of definition is, of course, derived from the Suttas, and in using it the commentaries are doing what they should: taking the Buddha’s mode of exposition and applying it in contexts not literally spelt out in the existing texts. But this does not free them from the charge of ontological essentialism; in fact, it confirms the charge. For the commentarial tradition worked out a scheme that grades the various kinds of definitions that it uses.[25] The above kind of definition, which explains a noun with a verb, is exlicitly stated to be provisional, since it relies on a merely verbal distinction. The only ultimately valid definition is the definition in terms of sabhāva; that is, not functional, but ontological.

Thus we must accept that of the two aspects of the basic definition of a sabhāva, the second, being inapplicable to Nibbana, is superfluous. It was the Abhidhamma schools that first placed Nibbana and conditioned dhammas (=samsara) on the same ontological footing. When Nagarjuna’s followers with their ‘emptiness’ critique tended to equate both Nibbana and samsara as empty, they merely mirror the position they are critiquing.


To find an elegant approach to this question we must look outside the Theravāda. Vasubandhu with his ‘three sabhāvas’ has articulated a clever solution.[26] Vasubandhu is a post-Nagarjuna Śuññatāvādin of the Yogacāra school. Unlike your typical abhidhammika, he is actually possessed of a sense of humor, and his use of sabhāva after Nagarjuna is downright cheeky. He clearly does not use it with any ‘realist’ ontological implications, indeed, his first sabhāva is the ‘imagined’. It seems rather that Vasubandhu is playing with terminology, evolving a respectable sense in which philosophical discourse can be reinvented following Nagarjuna’s devastating attacks.

Rather than dividing reality into ‘really real’ and ‘not really real’, as proposed by the abhidhammikas and refuted by Nagarjuna, he suggested a threefold division: the conceived or constructed or imagined (parikalpita); the interdependent (paratantrika); and the consummated (pariniṣpanna). Vasubandhu’s work is difficult and subtle, and I don’t necessarily wish to accept all of Vasubandhu’s conceptual apparatus, such as the ālāyavijñāna. But as I understand it, put simply the three mean something like this. The ‘conceived’ refers to the normal world of conceptually filtered experience. (This is what the abhidhammikas call ‘conventional reality’; the remaining two would both be classed by the abhidhammikas as ‘ultimate reality’.) The ‘interdependent’ refers to the conditioned phenomena as experienced by the insight yogi, which, though closer to the truth than the ‘conceived’, is still tainted by ignorance and so not seen with complete clarity. And the ‘consummated’ is Nibbana, or the experience of enlightenment, which is ultimately true due to the ending of ignorance. Thus there is a vital distinction preserved between Nibbana and conditioned phenomena, a distinction that is intrinsic to the conception, not tacked on as an afterthought. In Vasubandhu’s own words from the beginning of the Trisvabhāvanirdeśa:

“That which appears” is the interdependent [paratantrika]

“how it appears” is the conceived [parikalpita]

Through the former’s state of developing subject to conditions

And the latter’s being mere conception.

The constant state-of-not-being-found

of “how it appears” in “that which appears”

Can be known as the consummated [pariniṣpanna] sabhāva

Because of its state of non-otherness.

Here there is truly no hint of ontology. The conceived is called such because of it’s being mere conception; the interdependent is so-named because it is subject to conditions (this is equivalent to the second half only of the standard Theravāda commentary definition of sabhāva we have considered above); while the consummated is called such because of its ‘non-otherwiseness’, a Sutta term that points to the non-delusory nature of Nibbana.

Although ontology is not central to the doctrine of the three sabhāvas, Vasubandhu does consider their ontological status in terms of ‘existing-and-not-existing’, a choice of terminology that reflects his post-Nagarjunian sensitivity to the issue:

The conceived sabhāva is perceived as existent,

Yet it is complete non-being,

So it is thought of as having an existent and non-existent characteristic.

The interdependent exists, but not in the way that it appears,

There being confusion there,

So it is thought of as having an existent and non-existent characteristic, too.

The consummated sabhāva exists through non-duality,

But is simply the non-being of ‘two’,

So it is thought of as having an existent and non-existent characteristic, too.

So all three can be looked at in different ways, as either existing or non existing. The interdependent ‘exists’, but this is immediately qualified by reminding us that as long as there is delusion, the world of conditions can never be seen with complete clarity. Thus Vasubandhu, like the Suttas, is happy to speak of something that ‘exists’, without insisting on any ‘ultimate’ sense. Even Nibbāna ‘exists’ in the most tenuous imaginable way, as the absence of duality. Throughout, Vasubandhu works in this way, setting up distinctions even as he undermines them. The only unambiguous distinction he draws is that the conceived and the interdependent are both ‘with defilements’, while the consummated is ‘alleviation’, or ‘purity’. Again, this exactly mirrors the fundamental sutta statements on Nibbana.


So in this essay I have criticized some of the methods and conclusions used by Gethin in his exposition of dhammas. I should re-emphasize that there is much of value in Gethin’s article, and judging from the article itself as well as a private communication on the matter, there is ultimately not much difference between Gethin and myself in our understanding of the nature of dhammas in the Suttas and the subsequent evolution of the term. I do think it is counterproductive to marginalize the importance of historical evolution in Buddhist thinking. While it may be true – and I think this is where Gethin is coming from – that some modern interpreters have been too rash and sweeping in their dismissal of traditional interpretations, there is nevertheless much to learn from historical analysis. Not least, we can free ourselves from the simplistic idea that there ever was a uniform interpretation of all aspects of Dhamma adhered to and consistently expounded throughout the life of any school of Buddhism.

At least part of the difference in perspective between Gethin and myself seems to reflect our different environments. Gethin, coming from his western academic background, suggests that some remarks about dhammas, presumably by modern scholars, ‘are perhaps often viewed too much in the light of later controversies about the precise ontological status of dharmas and the Madhyamaka critique of the notion of svabhāva in the sense of “inherent existence”’.[27] I have shown that this ontological critique of the Madhyamaka is mirrored by a practice-based critique from the Yogacāra. And while Gethin may feel that modern scholars over-emphasize Nagarjuna, this is certainly not the case in contemporary Theravāda meditation circles.

If you talk with any Theravādin today, he will insist that there are dhammas that exist ‘ultimately’, that are ‘really real’, as opposed to the merely conventionally existing things we take for the truth. This, for him, is a key distinctive feature of the Buddha’s ‘higher teachings’, and is essential in order to understand the basic problem of how the Buddha could have spoken of ‘persons’, while still preserving the doctrine of ‘not-self’. The program of study in most Abhidhamma courses in Theravāda countries consists precisely of memorizing, analyzing, and classifying these dhammas. There is little consideration of the philosophical problems with this approach, and no meaningful response to the Nagarjunian critique. Surely, this situation, observable today, must have arisen from somewhere: if not from the texts, that is, the commentarial Abhidhamma works, then from where? While it is true that the texts do at times exhibit a degree of philosophical sophistication, for the most part they assume this rather simplistic ontology, an ontology that, we should not need reminding, cannot be traced to the Suttas.

These philosophical suppositions, inherited from the tradition and largely unexamined, underlie and inform the major schools of contemporary Theravāda meditation. Meditators practice precisely in order to see the elements of ‘ultimate reality’. The prime source text for this approach is the ‘Mahāsatipaṭṭhāna Sutta’, which we suggest would be better called the ‘Piltdown Sutta’. Is it too much to hope that the revelation that this is one of the latest and least authentic of all the texts in the Nikāyas will cause such meditation schools to question their own assumptions and methods?



[1] Rupert Gethin, ‘He Who Sees Dhamma Sees Dhammas; Dhamma in Early Buddhism’, Journal of Indian Philosophy 32: 513-542, 2004

[2] Gethin, note 1, pg. 537

[3] The Sarvāstivāda Dīrgha has been recently discovered in ancient Sanskrit manuscript. See Jens-Uwe Hartmann, ‘Contents and Structure of the Dīrghāgama of the (Mūla)Sarvāstivādins’, Annual Report of the International Institute for Advanced Buddhology at Soka University 7, 2004, 119-137.

[4] Minh Chau, Chinese Madhyama Āgama and the Pali Majjhima Nikāya, Motilal Banarsidass, 1991, pg. 124

[5] Minh Chau, pg. 124

[6] Choong Mun Keat, The Fundamental Teachings of Early Buddhism, Harrasowitz Verlag, 2000, pp. 71-72.

[7] T 2, p.20a

[8] SA 69-71, T 2, p 18b-c

[9]L. Schmithausen, ‘Die vier Konzentration der Aufmerksamkeit’, Zeitschrift fur Missionwissenschaft und Religionwissenschaft, 60, 1976.

[10] Johannes Bronkhorst, ‘Dharma and Abhidharma’, Bulletin of the School of Oriental and African Studies 48, 1985.

[11] Gethin, pg. 517.

[12] WALDSCHMIDT 1951: 102-304, 356-452. : [19.7] tasmāt tarh(i) bhikavo ye te dharm(ā) dṛṣṭadharmahitāya sava(r)tante dṛṣṭadharmasukhāya saparāy(ahitāyasapa)rāyasukhāya te bhikubhir udghya paryavāpya (tathā ta)th(ā) dhārayitavyā grāhayitavyā vācayitavyā  yatheda brahmacarya cirasthitika syāt tad bhavi(yati bahu)janahitāya bahujanasukhāya lo(kānukam)pāyārthāya hitāya sukhāya devamanuā [19.8] katame te dharmā dṛṣṭadharmahitāya (savartante d)ṣṭadharmasukhāya sa()parāyahitāy(a saparā)y(a)s(u)khāya te bhikubhir udghya pūrvavad yāvad devamanuam? [19.9] tadyathā catvāri smtyup(asthānāni catvāri) s(a)myakprahāāni catvāra ddhipādā pañcendri(yāi pañca ba)lāni sapta bodhyagāny āryāṣṭ)go mārga

                WALDSCHMIDT 1968a: 6 -16: 17. : (tasmā)t (ta)rhi bhikavo ye mayā dharmā sva(yam abhijñāya) sākīktvopasapadya praveditās tadyathā catvāri smtyupasthānāni catvāri samya(kprahāāni) catvāro ddhipād(ā pañcendriyā)i pañca balāni sapta bo(dhyagāny āryā)ṣṭāgo mārga tatra va sarvai sahitai samagrai samodamānair avivadadbhir ekāgrair ekoddeśai(r) ekakīrodakabhūtai ... dīpayamānai() sukhaspa(rśa vihartavyam.

[13] Gethin, pg. 520

[14] Eg. SN 46.2, 5, 7, 23, 24, 33, 34, 37, 39, 40, 49, 51, 52, 53, 55, 56.

[15]Eg. MN 2/MA 10/T31/EA 18.3 Sabbāsava Sutta: “So ‘ida dukkhan’ti yoniso manasi karoti, ‘aya dukkhasamudayo’ti yoniso manasi karoti, ‘aya dukkhanirodho’ti yoniso manasi karoti, ‘aya dukkhanirodha-gāminī paipadā’ti yoniso manasi karoti.”

[16]Manasikārasamudayāti ettha pana yonisomanasikārasamudayā bojjhaṅgadhammānaṁ samudayo, ayonisomanasikārasamudayā nīvaraṇadhammānaṁ.

[17] CBETA, T01, no. 26, p. 582, b18-19. The Saṁyukta version is even briefer: 身身觀念處 , CBETA, T02, no. 99, p.171, a1

[18]WALDSCHMIDT 1951: 102-304, 356-452, 14.25: iha bhikur (a)dhy(ā)tm(a) kāy(e) kāyānupaśyī viha(raty) ātāpī saprajāna smtimā(n) vinīyābhidhyā loke daurmanasyam.

[19]一切四念處經。皆以此總句。所謂是故。比丘。於四念處修習。起增上欲。精勤方便。正念正智。應當學 (CBETA, T02, no. 99, p. 171, c19-21)

[20]Chapter 4, pp. 98-106

[21] The Scripture on the Explication of the Underlying Meaning (translation of T 676 Saṁdhinirmocana Sūtra), John P. Keenan, Numata Center, 2000, pg. 22.

[22] E. Frauwallner, The Earliest Vinaya and the Beginnings of Buddhist Literature, Roma Is. M. E. O., 1956, Chapter 1.

[23] Gethin, pg 534

[24] PP 3.115

[25] Unfortunately, I don’t have the exact details or references to hand.

[26] Stefan Anacker, Seven Works of Vasubandhu, Motilal Banarsidass 1998, Chapter 9. I have changed a couple of Anacker’s renderings.

[27] Gethin, pg. 533