Can Dependent Origination Be Saved?

Image courtesy of Keattikorn at

Image courtesy of Keattikorn at

Dependent origination is a conundrum, particularly in its most common and elaborate twelve link formula. On the one hand, it is both historically and philosophically central to the Buddha’s dhamma, on the other hand it is a deeply problematic attempt to reconcile kammic rebirth with a potential awakening into non-self. The subject is so complex and has been written about from so many different angles that there is no way to deal with it all, let alone in a blog post. Nevertheless I will try to sketch out some potential lines of concern and inquiry.


Famously the Buddha said, “One who sees dependent origination sees the Dhamma; one who sees the Dhamma sees dependent origination.” (MN 28.28). In the later tradition it often seems as though dependent origination has usurped the place of the Four Noble Truths, which in the Nikāyas is most central (C.f. MN 28.2). In the dedicatory verse to Nāgārjuna’s Mūlamadhyamakakārikā (arguably the most famous work of Mahāyāna Buddhist philosophy) the Buddha is described as having taught “the doctrine of dependent origination”, rather than (e.g.) the Noble Truths. This should be seen as indicatory of Nāgārjuna’s purpose, which was to view the dhamma through the lens of dependent origination, understood to be the “middle way” between nihilism and eternalism. This gave the name to his Madhyamaka (“middle way”) school.

While the Buddha did identify the dhamma with dependent origination in one famous passage, as we have seen, it is not as widely appreciated that the form of dependent origination the Buddha then went on to describe in that passage is unusual:

And these five aggregates affected by clinging are dependently arisen. The desire, indulgence, inclination, and holding based on these five aggregates affected by clinging is the origin of suffering. The removal of desire and lust, the abandonment of desire and lust for these five aggregates affected by clinging is the cessation of suffering.

The three points he makes in this passage are (1) that our psychophysical selves arise dependent on causes, (2) That suffering arises dependent on desire, and (3) That suffering ceases dependent upon the abandonment of desire. Apart from clinging, none of these links appear (at least explicitly) in the usual formulae. Therefore when the Buddha identifies dependent origination with the dhamma it’s not clear that he is talking about much more than the conditioned nature of the self, and the second and third Noble Truths.


Nowadays people often take dependent origination to mean something rather ordinary or even indisputable such as that all things arise due to causes and conditions, or that all things are causally interconnected. The simplest version of dependent origination in the Nikāyas is indeed broad and indisputable. In the Udāna (1.3) we read,

This being, that is; from the arising of this, that arises; this not being, that is not; from the cessation of this, that ceases.

Sāriputta is said to have gained insight from a similar phrase, but if so it must be because he was already well advanced in training. Simply to say that things arise and cease isn’t particularly interesting or useful on its own. Insofar as the scope of this claim is made explicitly universal (all things arise and cease), the claim is just that of anicca or impermanence, plus an empty gesture at causation.

The Buddha did not mean dependent origination to be descriptively thin, ordinary, or banal. Indeed he made clear to Ānanda (DN 15.1) that it was not nearly as simple as it might seem at first glance. The Buddha was making several complex and substantive claims. In other words, whatever dependent origination is, it cannot simply be the claim that all things arise due to causes and conditions, nor simply that all things are interconnected.

How Many Links?

Perhaps the biggest problem with dependent origination is how complex we are supposed to understand it to be, and which links we are to understand as being essential to it. For as we have just seen, the Buddha does describe dependent origination has having two links: suffering arising from desire. This is just a version of the second Noble Truth.

Richard Gombrich terms the twelve link form of dependent origination “by far the commonest form of this chain” (2009: 132). That is,

Six sense-bases
Aging and death

This is, for example, the form in which we find dependent origination propounded by the Buddha in the first sutta of the Nidānasaṃyutta (SN 12.1), and the form we find explained by Sāriputta in the Sammādiṭṭhi Sutta (MN 9), the sutta on Right View.

There are several apparent repetitions implicit within the chain. For example, craving is the eventual cause of dukkha by the second Noble Truth, and also ignorance is the eventual cause of dukkha by the first link in the chain. Craving and volitions (saṇkhāras) are both listed, while craving itself is a form of volition (saṇkhāra). Consciousness appears as conditioning name-and-form, and also it must be assumed as part of contact, since contact is defined as the meeting of sense bases, sense objects, and consciousness.

These repetitions give pause, however there is nothing in the formula that should require links to be mutually exclusive. Links in the chain may participate in different causal processes at different times. And we may well say that identifying “the cause of dukkha” depends upon our concerns at the time. It is not strictly speaking false to say that dukkha is caused by both craving and ignorance, just as it is not false to say that a house fire is caused by both bad intentions and a lit match.

The question is more about whether we can make sense of how the conditioning is to take place at each link of the chain. So for example can we understand how it is that ignorance causes volitions? Perhaps so, but only if we make surreptitious modifications to the formula: by “ignorance” we mean something perhaps more like “inversions of perception” (e.g., AN 4.49, Matilal 1980), and by “volitions” we mean “unskillful volitions” or perhaps “kammically potent volitions”. But these are not small changes, nor is it entirely clear they are the appropriate ones to make.

In one of the earliest Buddhist compositions, the Kalahavivāda Sutta of the Sutta Nipāta (Sn. 862-872), we find a preliminary form of dependent origination with six links. Speaking roughly, there we find that name-and-form conditions contact, contact conditions feeling, feeling conditions desire, desire conditions clinging, and clinging conditions quarrels, disputes, lamentations, and grief. In other words, desire conditions dukkha. While this process is not explicitly described as “dependent origination” within the sutta, several of the links overlap with dependent origination, and the structure is essentially isomorphic to fuller and presumably later formulations. Hajime Nakamura (1980) refers to this as the “earliest stage” of dependent origination, and says,

Some of the twelve links of later days are mentioned here. But most of them are mentioned under different names, and their explanations are disorderly, not systematized. This means that they represent the stage prior to the formulation of the Twelve Link theory.

A similar structure akin to dependent origination can also be found in the Madhupiṇḍika Sutta (MN 18.16), perhaps not surprising given the affinities of this sutta with the Kalahavivāda. In the Madhupiṇḍika we find that sense bases and sense objects condition sense consciousness, which is called “contact”, contact conditions feeling and perception, perception conditions thinking, thinking conditions mental proliferation.

Gombrich (2009: 138) is thus likely right that early formulations of dependent origination would have involved fewer than twelve links. Following Frauwallner he argues that originally it would have been a five-link chain, as at SN 12.52, where craving conditions clinging, clinging conditions existence, existence conditions birth, and birth conditions decay and death. Here the initial condition is craving, which we find echoed in the second Noble Truth as the cause of dukkha. It is concerns such as these that lead Johannes Bronkhorst (2009: 44, 57) to claim that the twelve link chain “is not part of the original teaching of the Buddha,” although this appears to be little more than speculation. Significant elaborations can occur in a forty five year teaching career.

The example of the Kalahavivāda Sutta at least suggests that early forms of dependent origination may have included links prior to craving, such as name-and-form, contact, and feeling. The gap between feeling and craving is the fulcrum around which practice rotates, whereby we try to break the tendency to crave those things that feel pleasant, and to become aversive to things that feel painful. Hence it is perhaps unlikely that early formulations would have left that link out.

Three Lifetimes?

Traditionally the twelve link chain is taken to depict a causal process that extends over three contiguous lifetimes. Ignorance in lifetime one conditions volitions in that lifetime that lead eventually to death, then the kamma accumulated by those volitions leads to consciousness (etc.) in lifetime two. Then lifetime two unfolds as a conditioned process up through “existence”, which is taken to denote full participation within the wheel of saṃsāra. That is, “existence” must connote eventual death. Then birth occurs in lifetime three, capped off by decay and death at the end. Of course, then the cycle will begin again, presumably by volitions in lifetime three causing rebirth consciousness in lifetime four, and so on.

The three lifetime interpretation of dependent origination is at times (E.g., Gethin 1998: 150) attributed to the 4th-5th c. CE Mahāyāna commentator Vasubandhu’s Abhidharmakośa (iii.20ff), or the 5th c. CE Theravāda commentator Buddhaghosa’s Visuddhimagga (Ch. XVII). In fact this interpretation can be found significantly earlier, for example in the Paṭisambhidāmagga (I.275), part of the Pāli Nikāyas stemming from the 2nd or 3rd c. BCE. (Warder 2000: 299).

However in the Paṭisambhidāmagga the formula is not precisely the same as the standard twelve links: certain of them are repeated, probably to clarify the causal picture. I do not, however, believe that makes the comparison inapt. There must be a good deal of process overlap in the twelve link chain. It cannot, for example, be the case that birth and death only happen in the third lifetime, nor that consciousness only occurs in lifetime two. Therefore the twelve link formula must be seen as something more akin to an impressionistic attempt at systematizing causal processes than a complete analysis thereof.

The Problem of Rebirth

Then there is the problem of rebirth, which I dealt with in a prior piece and will not repeat here. One typical contemporary response is simply to deny that the chain has anything much to do with multiple lifetimes. As Dhivan Thomas Jones (2011: 55) puts it,

However, while [the three lifetimes] interpretation has been worked out over many centuries in the Buddhist tradition, it is not to be found in the early Buddhist teachings, which mainly present the twelve links just as explaining the arising and ceasing of dukkha.

While this may be correct in a sense, in that one cannot find any explicit reference to three lifetimes within the earliest discussion of dependent origination, it is nevertheless misleading in suggesting that the early Buddhist teaching of dependent origination was focused on a single lifetime. For it is quite clear in the early texts that this was to be understood as a process occurring over multiple, kammically linked lifetimes.

To take an example, birth and death (the last two links in the chain) are nowadays often taken to be metaphors for the temporary arising and passing away of mind-states within a single lifetime. While the early Buddhist texts do not rule out such a reading, the definitions of “birth” and “aging and death” in the Sammādiṭṭhi Sutta do not leave much room for interpretation. There birth is defined as “precipitation [in a womb], manifestation of aggregates, obtaining the [six sense-]bases for contact.” (MN 9.26) Aging is defined as: “brokenness of teeth, greyness of hair, wrinkling of skin, decline of life, weakness of faculties”. Death is defined as “dissolution, disappearance, dying, laying down of the body.” (MN 9.22). (Also, SN 12.2). There is no indication that these descriptions are to be taken as metaphors.

Consciousness and Name-and-Form

Perhaps the most confusing parts of the twelve link chain are at the beginning. To start, Sāriputta tells us that ignorance in this context means ignorance of the Four Noble Truths (MN 9.66). This should not be taken to mean a simple unawareness of such truths, of the kind that might be overcome by reading a book about Buddhism. It means something deeper: a lack of having trained oneself to think and behave in accord with such truths. We may understand how this kind of ignorance (or its associated wrong views: Matilal 1980) may produce in us unskillful volitions.

The step between volitions and consciousness however leaves us uncertain. How do volitions condition the arising of consciousness? Is there a way to reasonably interpret it without relying on a multi-lifetime interpretation?

Some (e.g., Jones 2011: 66-67) interpret consciousness’s role in this-life terms. They see consciousness, brought about by reactive volitions, as dividing the world into a problematic duality of subject and object. Taking this point further, consciousness might therefore be seen to provide the conceptual framework that gives rise to the “name” of name-and-form. (E.g., as in the Vedic picture provided in Jurewicz 2000). However whatever role consciousness might play as third link in the chain, it cannot be that of providing a cognitive framework of subject and object, because that is a conceptual role, and perception is responsible for such conceptual arisings, not consciousness. (E.g., AN 6.63).

True, in early Buddhism all consciousness is consciousness of something or other. (E.g., Gombrich 2009: 120). But this should not be seen as a faulty product of reactive volitions; it is simply the way consciousness operates, at a precognitive level. The model of consciousness as inherently “non-dual”, and as ordinarily covered over by an unskillful conceptual overlay of subject/object duality, did not arise until much later.

To clarify the situation between volitions and consciousness let us look at the next pair, the conditioning that is said to occur between consciousness and name-and-form, in the Mahānidāna Sutta. The Buddha says,

“It was said: ‘With consciousness as condition there is name-and-form.’ How that is so, Ānanda, should be understood in this way: If consciousness were not to descend into the mother’s womb, would name-and-form take shape in the womb?”

“Certainly not, venerable sir.” …

“If the consciousness of a young boy or girl were to be cut off, would name-and-form grow up, develop, and reach maturity?”

“Certainly not, venerable sir.” … (Bodhi 1995: 51, amended translation of nāmarūpa).

The key stage for consciousness is during “descent into the womb” (cf. MN 38.26), conditioned prior to that descent by volitions and ignorance.

The natural way to read this passage is as involving rebirth from one life to another. Consciousness works as an intermediary between lives, and also as a kind of vital force that allows development of the psychophysical organism. This is not an odd picture at all, it is just factually incorrect. It is a kind of premodern, vitalist biology wedded to a notion of rebirth. Thus the problem with consciousness for the Buddha is not so much that it is outward looking as that it is the product of certain kinds of kammic volitions: those that cause rebirth.

Yes, Three Lifetimes

The traditional interpretation has a lot going for it. Given the material we find in the Sammadiṭṭhi and Mahānidāna Suttas, the twelve link chain is best interpreted as extending over three contiguous lifetimes, or if we are not to understand the chain as literally progressive, over more than one such lifetime. While the three lifetimes interpretation may not be explicit nor consistent in the Nikāyas, it appears at least to be reasonably implicit in the twelve link formula. And while, for example, the Mahānidāna Sutta omits three of the links in its discussion, that ends up having no particular bearing on whether it nevertheless affirms rebirth. It clearly does.

To put it another way, to interpret all twelve links in the chain as occurring solely within a single lifetime would seem to be an exercise in secular apologetics. It would seem to start from the premise that the formula of dependent origination must be true, and work from there to figure out how it could be so, consistent with a rejection of literal rebirth. If instead we start by questioning whether dependent origination is true, whether all twelve links in the chain are really valid, we may end up with a different answer.

This is not to say that even traditional Buddhists didn’t have a this-life interpretation of the twelve link chain: such an interpretation was laid out in the Vibhaṅgha (144-145), which Anālayo (2008: 94) says is “probably the earliest book of the Abhidhammapiṭaka.” That said, the this-life version of dependent origination differs in two crucial respects from its multi-lifetime counterpart: in place of name-and-form, only name (= feeling, perception, and volition) is conditioned by consciousness, and in place of the six sense-bases only the sixth sense base (mind) is conditioned by name. So the picture then is that consciousness → feeling/perception/volition → mind sense. That is to say, consciousness does not condition physical form, nor does the mind (separate from the body) condition the five external sense bases. It only conditions the internal (mind) sense base. This is because there is no literal rebirth of a physical body in the case that we are talking only about mind-moments.

To make a relatively obvious but perhaps overlooked point, since on the traditional interpretation all of the labeled processes must happen in some lifetime or other, and since the processes overlap, there is no problem in our saying that in some sense they all must happen in every lifetime. Some processes happen at birth, others happen at death. But while this is one reasonable perspective, it makes a hash of certain of the links, particularly those between volition and consciousness, and existence and birth, which are most naturally interpreted as extending between rather than within single lifetimes. Or to put it another way, this perspective substantially confounds the chain. Rather than a single chain of twelve, it begins to look as though we are talking of three separate, shorter, overlapping chains.

Within a traditional context apologetics is to a certain extent to be expected. And within a context where rebirth is assumed true, making as it were off-label use of the formula as a guide to contiguous mind-moments may seem apt. But in a context where we make no such assumption, it’s not clear why we ought to look to something so questionable to ground our this-life causal claims.

The Buddha was intelligent, so it’s reasonable to assume that he must have had good reasons for his proposals. But in this case, when it is clear that he proposed something problematic (viz., a process extending over multiple lifetimes), it may simply be better to pare the thing back to a useful core rather than try to figure out what the Buddha must have meant by something that to all appearances is ill-considered.

The Core

The central six links in the chain, from name-and-form through clinging, unproblematically take place within a single lifetime. Further, if we understand “existence” to denote the arising of a deluded sense of self (which may indeed be within the spirit of early Buddhist teaching, though cf. AN 3.76), and if we understand the last three (existence, birth, aging and death) as conjointly denoting the dukkha of saṃsāra, we can capture most of the twelve in a single lifetime. This is about as far as we can go with a relatively accurate secular reconstruction.

But it is hard to know what to do with the trio of volitions, consciousness, and name-and-form on anything like a contemporary understanding of the world. More: what is the point of trying? The essence of the Buddha’s teaching (pace Nāgārjuna) was captured in the Noble Truths, and the key realizations therein are reflected this core of links. This is roughly the portion that Frauwallner, Nakamura, and Gombrich believe may have been part of its earliest formulation, although one must always be wary of reading back into the texts what we find most congenial.

The Buddha’s world view contains a central, apparent inconsistency, reflected in both doctrine and practice. That is, we are beings who accumulate kamma that is supposed to determine how we are reborn into future lives, and yet an enduring self who accumulates that kamma is nowhere to be found. We are beings who are to act ethically in the world, and yet we are to pursue the cessation of all kamma (AN 3.111). The Buddha saw dependent origination as playing a key role in knitting together this odd system. (Cf. Gethin 1998: 156-7). That is, he intended the twelve step version to explain how it was that kamma conditioned beings from one lifetime to the next, as well as how that same kamma resulted in unskillful mental and bodily actions within a single lifetime that led to dukkha and eventual rebirth, all without the need for a grounding in a permanent, continuing self. In large part perhaps the complexity that the Buddha lamented about to Ānanda stems from the detailed, even unwieldy metaphysical role dependent origination was meant to play within the Buddha’s dhamma.

As metaphysics eventually took precedence over practice in scholarly Buddhist circles, so too dependent origination bested the Noble Truths in common understandings of what the Buddha taught.

Is Dependent Origination Salvageable?

All philosophical systems can be nitpicked, therefore so can Buddhism. But it seems dependent origination is particularly nit ridden: held as central by much of the later tradition, it is in fact broken, at least by the light of contemporary belief and practice. Bronkhorst believes that in its most expansive, twelve link form it doesn’t even go back to the Buddha. But even if we shy away from such a strong claim, it appears to have been too ambitious: a diachronic account of ethical causation extending over beginningless lifetimes, meant in some sense to substitute for the Brahmanical notion of the ātman or permanent soul.

Above we observed how the twelve link formula was more an impressionistic attempt at describing these kammic processes than a plain analysis thereof. This lends some credence to Joanna Jurewicz’s (2000) suggestion that it was intended as a kind of rhetorical ploy, intended to upend Brahmanical notions of the role of the ātman by, in Gombrich’s term, “ironizing” (2009: 138) Vedic imagery.

Jurewicz’s approach has historical merit, however unfortunately it neither helps the Buddha nor us. As regards the Buddha, whatever may be the Vedic connotations of dependent origination, the Buddha was not a Brahmin and he certainly did not recognize the authority of the Vedas. That is to say, we cannot argue from the fact that words or phrases had certain Vedic connotations to the claim that therefore the Buddha would have intended those connotations to be denotative in his own work. He may have been ironizing the Vedas, but if so only by stating something non-Vedic in so doing. We know this well in the case of the word “kamma” for example, which meant “action” in traditional Vedic Brahminism, but which the Buddha used to mean “intention”. So his system of dependent origination stands or falls on its own merit, ironies aside. As regards us today, ancient ironies are even less pertinent now than before. As Gombrich (2009: 137) says, “irony does not weather well.”

Cutting the Knot

The central teaching of dependent origination is one captured very roughly in what may be its earliest formulation in the Kalahavivāda Sutta, with links originating in the psychophysical organism and culminating in lamentation and grief. This is arguably is what the Buddha meant when he identified dependent origination with the dhamma. At its most plain and direct, dependent origination is simply the second and third Noble Truths, highlighting the key cause of dukkha and thereby its cessation. Making much more of it takes us off the rails.

Looking at the twelve link version, perhaps we ought simply to dispense with the first three links of the chain. The matter begins with name-and-form, and the associated six senses. The last three links in the chain can be understood as our deluded sense of self (existence) arising out of craving and clinging, and itself conditioning our dukkha-filled interaction with saṃsāra: that is, birth, aging and death. This interpretation, although only a sketch, is clear and does no damage to contemporary sensibilities. It also does not seem to stray terribly far from early Buddhist interpretations, although it does make metaphorical use of the concept birth, and semi-metaphorical use of the concepts aging and death. Semi-metaphorical because part of dukkha is confronting our own quite literal aging and death, as well as the metaphorical aging and death we confront with every change in our lives.

I have argued against taking literal passages metaphorically, and also argued for taking certain passages metaphorically. To stave off the appearance of inconsistency, let me explain. It is incumbent upon us as much as humanly possible to distinguish apologetics from scholarship. It is enough that we note the original texts appear to take the concepts of birth, death, and so on literally. We do scholarship no damage if we note the problems in the texts, and present a constrained and conditional use of metaphor in the case that real advantage to contemporary practice may prove possible in so doing. So long as we do not insist on reading contemporary metaphors back into the texts we are on firm ground.

Dependent origination can be saved through drastic surgery. But in so doing we are likely to find it saying something not so different from the Noble Truths themselves: much of its excess complexity turns out to be incompatible with a contemporary approach to the dhamma. This makes a message Ānanda might have found easier to digest. While undoubtedly something important to the Buddha has been lost in our treatment, I think we should shy away from assuming that dependent origination is hopelessly complex or obscure. Although the Buddha did chide Ānanda, there is little reason to suspect that he would have done the same to Sāriputta, and Sāriputta is at least purported to be author of the Samādiṭṭhi Sutta that as we have seen gives a lucid description of the links in the chain. Whether his description is compatible with contemporary belief and practice is our question, however.

Doubtless human creativity being what it is, contemporary Buddhists and secular Buddhists will continue to mine all twelve links for inspiration and guidance, under the tacit assumption that there must be something to the complete formula even if it is not immediately clear what. Indeed, obscurity and complexity themselves may seem advantageous, clarity limited or naïve in comparison. Ideologies have thrived under such forms of thought for millennia.




Bhikkhu Anālayo (2008). “Rebirth and the Gandhabba”, Mahachulalongkornrajavidyalaya University Journal of Buddhist Studies, Vol. 1: 91-105.

Bhikkhu Bodhi (Various). Various Nikāyas (Boston: Wisdom).

Bhikkhu Bodhi (1995). The Great Discourse on Causation (Kandy, Sri Lanka: BPS).

Johannes Bronkhorst (2009). Buddhist Teaching in India (Boston: Wisdom).

Rupert Gethin (1998). The Foundations of Buddhism (Oxford: Oxford U. Press).

Richard Gombrich (2009). What the Buddha Thought (London: Equinox).

John Ireland (1997). The Udāna and the Itivuttaka (Kandy, Sri Lanka: BPS).

Dhivan Thomas Jones (2011). This Being, That Becomes: The Buddha’s Teaching on Conditionality (Cambridge: Windhorse).

Joanna Jurewicz (2000). “Playing With Fire: The pratītyasamutpāda from the perspective of Vedic thought”, Journal of the Pali Text Society 26: 77-103.

Bimal Krishna Matilal (1980). “Ignorance or Misconception? A Note on Avidyā in Buddhism”, in Buddhist Studies in Honour of Walpola Rahula (London: Gordon Frasier).

Hajime Nakamura (1980). “The Theory of ‘Dependent Origination’ in its Incipient Stage”, in Buddhist Studies in Honour of Walpola Rahula (London: Gordon Frasier).

KR Norman (2001). The Group of Discourses (Sutta-Nipāta), 2nd Ed. (Oxford: PTS).

Bhikkhu Ñaṇamoli (1999). The Path of Purification (Onalaska, WA: BPS Pariyatti).

Leo Pruden, et al. (1988). Abhidharmakośabhāṣyam (4 Vols.) (Berkeley, CA: Asian Humanities Press).

Mark Siderits and Shōryū Katsura (2013). Nāgārjuna’s Middle Way (Boston: Wisdom).

Sayadaw U Thittila (1969). The Book of Analysis (Vibhaṅga) (Oxford: PTS).

AK Warder (2000). Indian Buddhism, 3rd Ed. (Delhi: Motilal Banarsidass).