19 February 2021

Modern Interpretations of the Khandhas: Saṅkhāra

This is the fourth installment in my series of essays on two modern interpretations of the skandhas. Tilmann Vetter's The Khandha Passages in the Vinayapiṭaka and the Four Main Nikāyas and Sue Hamilton's Early Buddhism: A New Approach were both published in the year 2000. After the rūpa essay was published I found Rupert Gethin's article: “The Five Khandhas: Their Treatment In The Nikāyas And Early Abhidhamma” (1986) which is a valuable document for being a succinct description of what the Sutta-Piṭaka and Abhidhamma say without much extra exegesis or interpretation. 

When I embarked on this task I was optimistic and I thought it might help clarify some things that had long been unclear for me. Sadly, this has not been the case. I am no longer optimistic; very little has been clarified, the methods used by the two main authors leave much to be desired. This realisation comes just as Vetter is hitting his stride and his attempt to explain saṅkhārā and his translation, "impulses", takes up more than half of his entire summary of the khandhas (36 pages!).

In any case, we now come to the khandha of saṅkhārā (The only khandha that is routinely plural). The problem we have with saṅkhārā is that everyone agrees how the word is used, but no one can make sense of why this word is used this way. As Gethin (37) says, the saṅkhārā are "primarily defined in terms of will or volition (cetanā)" and we can relate this to the famous phrase "cetanā is what I call kamma" (AN 6.63). Thus the saṅkhārā are somehow connected to karma (I will say something about this below). However, Gethin also points out that the Nikāyas "describe [saṅkhārā] as putting together (abhisaṃkharonti) each of the khandhas in turn into something that is put-together (saṃkhata) (37). The latter is a reference to the Khajjanīya Sutta that I will also address below. 

Bodhi (2000: 45) tells us that the word saṅkhārā can be analysed as a verb karoti "make" (√kṛ) with a prefix saṃ- that usually means "together" or "complete". In Latin, "to make" is facere and the cognate prefix is con-, so saṅkhāra is very like our word confection. Given the difference between the etymological meaning and the usage in Buddhism, especially in the khandhas, we ought to be alert for our old friend the etymological fallacy (which has been a feature of these essays). Bodhi points to five different contexts in which the word is used. As a khandha, Bodhi translates saṅkhārā as "volitional formations". Another common translation is "karmic formations". Translators who use these phrases as translation are trying to link the two different meanings of the word—volition and confection—in one phrase. The result is rather awkward.

A quick digression on the kha in saṅkhara. If we switch to Sanskrit, the skandha is saṃskārāḥ and through abhisaṃskaronti we get things that are saṃskṛta. But the verb is karoti (√kṛ). The extra s in saṃ-skāra and saṃ-skṛta is there because in Indo-European the root is *sker. The s is largely dropped in Vedic and Sanskrit but reappears when some suffixes are added. Sanskrit /sk/ becomes /kh/ in Pāli, so saṃskārāḥ becomes saṇkhārā but /k/ is unchanged so a Sanskrit word like sūtrakāra "the author of the sutra" (where -kāra is also from √kṛ) becomes suttakāra in Pāli. 



Vetter's method is now settled. He begins by consulting the Khajjanīya Sutta (SN 22.79). And this is vexing because, as I have now repeatedly shown, this text is not a reliable source. It is fool's gold, shiny but not valuable. For the first time, Vetter moves deeper into the Khajjanīya Sutta and considers another  passage. Vetter declares there is no good translation into English and instead gives a German translation that he doesn't bother to translate into English. Fortunately, in that very same year, 2000, Bhikkhu Bodhi's translation of the Saṃyutta Nikāya came out:

Kiñca, bhikkhave, saṅkhāre vadetha? Saṅ­kha­ta­mabhi­saṅ­kha­ron­tīti kho, bhikkhave, tasmā ‘saṅkhārā’ti vuccati. Kiñca saṅ­kha­ta­mabhi­saṅ­kha­ronti? Rūpaṃ rūpattāya saṅ­kha­ta­mabhi­saṅ­kha­ronti, .... (SN III.87). 
“And why, bhikkhus, do you call them volitional formations? ‘They construct the conditioned,’ bhikkhus, therefore they are called volitional formations. And what is the conditioned that they construct? They construct conditioned form as form...” (Bodhi 2000: 915)

Note that what is explained here is not the volitional part, only the formations part. Gethin points to a few other suttas that have this formula. But it's a pericope—the same passage repeated—rather than more on the theme. Vetter's explanation looks post hoc to me. The heart of it is: 

saṅ­kha­ta­m abhi­saṅ­kha­ron­tī ti kho, bhikkhave, tasmā ‘saṅkhārā’ti vuccati. 

They construct (abhisaṁkharonti) the constructed (saṅkhata) therefore (tasmā) they are called (vuccati) the constructs (saṅkhāra). 

It doesn't take much thought to see that this is completely unrelated to karma. Vetter says this is difficult to translate. The first part is not. What is difficult is the repeated formula that comes next. In Bodhi's translation: "They construct conditioned form as form" (rūpaṃ rūpattāya saṅ­kha­ta­m abhi­saṅ­kha­ronti). 

We know the agent must be saṅkhārā since the verb is in the 3rd person plural, which makes rūpaṃ the patient and saṅkhatam an adjective or predicate. So the basic sense is that the "volitions construct appearance" ([saṅkhārā] rūpaṃ abhisaṅkharonti). But this is nonsense, even in Buddhism. We have volitions in response to sense experience, especially in response to the positive and negative hedonic qualities of experience. This is saying the opposite, i.e. that volition is what makes form. To repeat, this is nonsense. 

Worse we still have yet to explain rūpattāya. One of the difficulties is that rūpattā is an abstract noun, which I am at a loss to translate directly into English: "form-ness"? We tend to treat rūpa as abstraction in the first place, both as the traditional translation "form" and in my preferred translation "appearance". We know this partly because when we use form in this sense we don't include articles. It is not "a form" or "the form" but just "form". My grammar checker doesn't deal with abstract nouns very well and pings me on this every time. And I think this might explain why Bodhi's translation does not distinguish between rūpa and rūpattā. Effectively they mean the same thing in Buddhist English. I have my doubts that they mean the same thing in Pāli, but I don't understand what the distinction is. 

Unfortunately, rūpattāya is a degenerate case ending that could be instrumental, dative, ablative, or genitive. Not all of which can fit, though, so we have some options: the saṅkhārās construct form through form-ness (ins), for form-ness (dat), from formness (abl), or saṅkhārās construct form that is formness (indirect object). The latter is apparently how Bodhi reads it, but as I say he translates both rūpa and rūpattā as "form". His explanation follows the commentary (2000: 1071-2 n.113), i.e. the passage means that form "becomes conditioned form in accordance with its nature". This doesn't seem remotely connected to the khandhas or their functions. Nor does it link saṅkhārā to karma. Form is conditioned when considered as the object of the eye, but we are talking khandhas. In what sense does the rūpakhandha become a conditioned thing? It is conditioned.

Once again the Khajjanīya Sutta (SN 22.79) is tempting because it appears to offer an explanation of the khandhas that is lacking elsewhere, but once again the explanation turns out to be impenetrable or nonsensical. My view is that the suttakāra did not know why the word saṅkhārā was used in this context and tried to make something up based on the obvious etymology. But this does not shed any light on how the word saṅkhārā is used in Pāli. But it did start an long running attempt to shoehorn the literal meaning of saṅkhārā into explanations of the khandhas

Vetter carries on for 36 pages of learned discourse, looking at literally all the ways in which the word is used, with examples. But he never manages a convincing explanation. Working through all of this has become intensely irritating because it appears to be wilfully ignorant of the manifold textual and linguistic problems. Bodhi can be forgiven because he is a senior Theravādin monk with responsibilities and is thus naturally an apologist for the tradition. Vetter has no such excuse for failing to read this material critically. 

So I'm just going to move on. And yet this brings me no joy.


As Hamilton says, "The order of the khandhas is never explained, but they are almost invariably (the single exception being to accommodate the metre of a verse) in the order given above; that is, body, sensations, apperception, volitional activities, consciousness." (72). We are slightly surprised, then, when in the succeeding pages she describes her understanding of body, sensations, apperception, and consciousness, and then introduces "The fifth khanda of volitional activities" (78). Because it is not the fifth khandha, it is the fourth khandha

Hamilton describes saṅkhārā as "one's affective response to what one is experiencing" (78). She rightly notes the centrality of karma in Buddhist soteriology and of volitions to karma: "The aim of the path of Buddhism is to arrive at a point when the fuel of continuity is blown out, and it is volitions that are that fuel." (78). 

But affect is not equivalent to volition: affect is a very much broader category. Whether used in a technical or lay sense, affect is a general way of talking about emotional or felt responses to sensory experience (the word literally means "to have an effect"). Volition is related in the sense that emotions usually lie behind actions. The problem is that Pāli doesn't really have this EMOTIONS ARE AGENTS metaphor.  In fact, as I have noted elsewhere, Pāli doesn't consider emotions separately from what we would call cognitive activity. Pāli makes a distinction between physical (kāyika) and mental (cetasika) events, but does not distinguish emotions as a distinct category. 

Hamilton then draws attention to none other than the passage from the Khajjanīya Sutta just discussed. In her account, abhisamkharonti means "to volitionally construct". Except that this is not what the word means or how it is used. Here volitionally is an adverb that Hamilton has tacked onto the verb, in a way that looks tendentious - it suits her argument to translate it this way. 

This is not the only problem we encounter in the details of Hamilton's account of saṅkhārā. In her view: 

"the khandhas that are described as being volitionally constructed need to be interpreted in the sense that together they represent the entire human being. So it is one's volitional activities that determine one's future coming-to-be in its entirety. (80). 

And yet two pages earlier:

Understanding them as the individual physical and mental 'parts' of which a human being is comprised misses two crucial points. First, that it is collectively that they operate, and second, perhaps even more importantly, that what they represent is one's cognitive system: the apparatus by means of which we have all our experiences. The point is not to offer an analysis of all that we are... Rather, they are what one needs to understand about oneself in one is to achieve liberation from the cycle of lives as the Buddha did. (78) 

These two passages seem to contradict each other. And the main drift of Hamilton's argument is in favor of the second reading. The khandhas do not represent the entire human being, they represent what we need to understand in the pursuit of liberation. Specifically they tie the perceptual/cognitive process to karma. 

But of course, Hamilton has the same problem as Vetter in using this passage, i.e. that it is deceptive because the khandhas are not "volitionally constructed". Neither rūpa nor vedanā are volitional. Except, while dwelling in emptiness (suññatā-vihāra), we are bound to experience both without having any say in the matter. We have free won't over rūpa and vedanā if we practice meditation, but we don't have free will.

And at this point I feel like I'm getting nowhere. So again, I'm going to move on. 

How Does Saṅkhārā Relate to Karma?

As we saw above we can make the link between saṅkhārā and karma via cetanā. This is because saṅkhārā are often described in Pāli in terms of the six kinds of cetanā. For example, in the Upādānaparipavatta Sutta (SN 22.56):

katame ca, bhikkhave, saṅkhārā? chayime, bhikkhave, cetanākāyā – rūpasañcetanā, saddasañcetanā, gandhasañcetanā, rasasañcetanā, phoṭṭhabbasañcetanā, dhammasañcetanā. (SN III.60)

And what, monks, is volition? Monks, there are six kinds of intention: intention towards appearance, sounds, smells, tastes, tangibles, and mental events. 

The text switches from cetanā to sañcetanā but there is no discernible change in meaning. But there is no logic here. No one, as far as I can see, can explain why a word the denotes confection can connote volition. Where is the link? The connection between cetanā and karma is much easier. Which brings us back to the Nibbedhika Sutta (AN 6.63):

cetanāhaṃ, bhikkhave, kammaṃ vadāmi. cetayitvā kammaṃ karoti – kāyena vācāya manasā. (A III.409)

Monks intention is what I call karma. Having intended one acts, with body, voice, or mind. 

Hence, volition is karma. As Vetter notes, this is only one way of talking about karma and not the most common. The equation cetanāhaṃ kammaṃ vadāmi occurs only once in the Nikāyas though it is clearly important as it is picked up by later writers, especially Nāgārjuna (in his MMK chapter on karma). 

But how are saṅkhārā equated with volition? 

I think there is a more direct link between saṅkhārā and karma, and this draws inspiration from work by Joanna Jurewicz (2005) on how Pāli might have borrowed terms from Vedic religion. This is made easier if we move the discussion into Sanskrit again, in which case saṅkhārā becomes saṁskārāḥ.

Saṃskāra is a word familiar to anyone who knows about Vedic culture, where it means a rite of passage. In this context saṃskāra means something like occasion or consecration. It is a feature of words from √kṛ that their meaning is highly context dependent. The prefix saṃ- does not just mean "together" it also means "complete" and from this we get the sense of "perfected". So saṃskāra can mean "finishing, refining, perfecting" and it is used in many figurative ways such as "purification, cleansing, preparing, or rearing of animals". 

 There are typically sixteen rites of passage (saṃskārāḥ) for a twice-born Hindu (for an overview see for example, Klostermaier 1994: 183-92). In fact, different texts prescribe different numbers of sacraments, but sixteen is the most common number. The saṁskāra are rituals conducted at certain points in the life of orthodox Hindus, including birth, naming, first haircut, marriage, and death. The rite of passage puts the finishing touch on the stage of life. It helps to refine the recipient's life and helps them attain the purity required for mokṣa

As Klostermaier says, "It is through the performance of saṃskāras that all Hindus practice the karma-mārga, the Path of Works" (1994: 192). The karmamārga  was previously introduced by Klostermaier in two chapters (9 and 10) and is contrasted with the jñānamārga (path of knowledge). The karmamārga is prescribed for people (men) who follow the householder's life of that Zorba the Greek called the "full catastrophe": wife, children, house, property, business. The jñānamārga is for those who renounce the household life and seek mokṣa or liberation from rebirth. 

Thus a saṃskāra is an occasion on which karma is performed by those pursuing the karmamārga as prescribed by the Dharmaśāstra texts outlining the dharma or religious duties of someone who follows some form of Vedic religion. Buddhists borrowed this idea and incorporated it with little change into their own doctrines. It has been changed mainly by the emphasis on the internalisation and ethicisation of religion; i.e. the focus for Buddhists was not the external performance of ritual actions aimed at manipulating the world, but was instead internal—volitional—and aimed at the world of experience. If one makes no effort at purifying one's intentions, one performs karma that leads to rebirth, i.e. one follows the karma-mārga. Buddhism holds out the possibility that by purifying one's intentions one does not perform karma, and one eventually brings rebirth to a halt. Not being reborn, one is free from suffering, since suffering is associated with embodiment.

The attempts—ancient and modern—to fit the semantic meaning of "confections" or "formations" to saṅkhārākhandha seem to me to be a red herring based on the etymological fallacy (the idea that the meaning of a word can only be obtained by examining the morphology, the past use, and the meaning of the parts of the word. Because the word is not defined this way in this case, such attempts are at best confusing and unconvincing. At worst they are nonsensical.

In this context saṅkhārā seems to straight-forwardly mean volitions (cetanā), i.e. contributions to karma that fuel rebirth. And since understanding how we create karma is central to Buddhist soteriology, this makes sense.



Bodhi. 2000. The Connected Discourses of the Buddha: A Translation of the Saṃyutta Nikāya. Wisdom.

Gethin, Rupert. 1986. “The Five Khandhas: Their Treatment In The Nikāyas And Early Abhidhamma.” Journal Of Indian Philosophy 14(1): 35-53.

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. London: Routledge.

Jurewicz, Joanna. (2005) "Playing with Fire: The pratītyasamutpāda from the perspective of Vedic thought." In Buddhism: Critical Concepts in Religious Studies, Vol. I, edited by Paul Williams. Psychology Press. Originally published 2000 Journal of the Pali Text Society 26 pp. 77-103.

Klostermaier, Klaus K. 1994. A Survey of Hinduism. 2nd Ed. State University of New York Press.

Vetter, Tilmann. 2000. The Khandha Passages in the Vinayapiṭaka and the Four Main Nikāyas. Wien Verlag der Österreichischen Akademie der Wissenschaften.

Related Posts with Thumbnails