Finding Easter Eggs in Pāli Texts

I've been studying the Kaccānagotta Sutta (S 12.15) for some time now. We are fortunate to have three extant versions of the text: Pāli (KP), Chinese (KC), and Sanskrit (KS). KC is from one of two Chinese Saṃyuktāgama translations (Taisho 2.99, no.301) related to the Sarvāstivādin School and was translated in the mid 5th century CE. The original language was probably a Sanskritised Prakrit aka Buddhist Hybrid Sanskrit.  KS is from a cache of texts in Turfan from a manuscript copied in the 13th or 14th centuries. There is presently no published English translation of the Sanskrit (a situation I hope to rectify).

The text seems to have been quite important as it is cited directly by Nāgārjuna in his Mūlamadhyamakakārikā (MMK 15.7); and indirectly in the Aṣṭasāhasrikā Prajñāpāramitā Sūtra and the Laṅkavatāra Sūtra. It's also likely that Chandrakīrti who commented on MMK had a different Sanskrit version that the Turfan Ms. 

In this essay I want to explore a single passage which contains an elaborate play on words that gets lost in translation. I call this passage paragraph 5c:

• KP: Tañcāyaṃ upayupādānaṃ cetaso adhiṭṭhānaṃ abhinivesānusayaṃ na upeti na upādiyati nādhiṭṭhāti ‘attā me’ti.
• KS: etāni ced upadhyupādānāni cetaso ’dhiṣṭhānābhiniveśānuśayān nopaiti nopādatte nādhitiṣṭhati nābhiniviśaty ātmā meti |
• KC: 若無此取者，心境繫著使不取、不住、不計我

Translations

• KP: And that obstinate tendency of the mind to attachment and grasping this [noble disciple] doesn’t approach, doesn’t hold, [he] doesn’t insist on ‘the self is mine’.
• KS: And [they] don’t hold this obstinate tendency of the mind to grasp and cling, they don’t accept, [they] don’t insist on or have a tendency to say: ‘this is my self’.
• KC: Not seizing those, they don’t have the obstinate mental state of attachment; they don’t insist on, or think wrongly about ‘I’.”

Buddhaghosa’s commentary on KP throws light on this passage. He says

Tañcāyanti tañca upayupādānaṃ ayaṃ ariyasāvako. (SA 2.33)

'Tañcāyaṃ' means that attachment and grasping, and this noble-disciple.

This makes it much easier to unravel the syntax by supplying a subject who does not insist on the statement ‘the self is mine’, without whom the sentence is puzzling since on the face of it the subject who doesn't hold the wrong view is the same subject as the one bound by attachment and grasping (which is caused by wrong views). The reference to self is part of the oft repeated formula found in Early Buddhists texts regarding wrong views about the self, namely:

rūpaṃ etam mama, eso'ham-amsi, eso me attā ti samanupassati

he regards forms: this is mine, I am this, this is myself.

The formula is repeated for each of the skandhas, and in each case the assutavant is incorrect, where as the sutavant ariyasāvaka knows that it is not true.

What I particularly want to draw attention is a form of syntax which is unusual in English. We can for instance say "I sing a song" but not "I work a work" or "I talk a talk". Mostly this kind of idiom doesn't work in English but it is common in Pāli and Sanskrit. We have several examples here, though in the negative. The Pāli has (with the verbal root of the two words in parentheses):

upayaṃ na upeti (upa√i) - he does not attach the attaching

upādānaṃ na upādiyati (upa√pad) - he does not cling the clinging

adhiṭṭhānaṃ nādhiṭṭhāti (adhi√sthā) - he does not insist the insisting

Compare the Sanskrit:

[upayaṃ]* nopaiti  (upa√i)

upādānaṃ nopādatte (upa√pad)

adhiṣṭhānaṃ nādhitiṣṭhati  (adhi√sthā)

abhiniveśaṃ nābhiniviśati (abhi-ni√viś) he does not tend the tendency

We can see that where KP has upaya, KS has upadhi. This is difficult to explain because upadhi means ‘addition, attribute, or ‘condition, support’; so it might mean ‘tendency to grasp at supports where upadhi refers to dvayaṃ niśrito ‘based on a duality’. BHSD s.v. upadhi suggests that S. upadhi = P. upadhi (upa √dhā) ‘foundation, basis’; or upādi = upādāna. So KS could be intending upadhi as a synonym of upādana. However upadhi doesn’t seem to fit here, and from the Pāli we would expect to see upaya. What's more the play on words breaks down with upadhi. So it seems that upadhi is a substitution, though it does occur twice in the text.

Other features of the syntax hide the play on words to some extent. The nouns are all given in advance and some are compounded: upayupādānaṃ cetaso adhiṭṭhānaṃ abhinivesānusayaṃ. We are left wondering about the role cetaso (a past participle in the genitive or dative case). My translation above takes things as they come, but here I'm exploring an alternate possibility. If we take the nouns to go with the matching verbs then we might rearrange things like this:

Tañ ca ayaṃ upayaṃ na upeti, upādānaṃ na upādiyati adhiṭṭhānaṃ nādhiṭṭhāti abhinivesaṃ [abhinivisati] cetaso ānusayaṃ 'attā me’ti.

And he does not grasp the grasping, cling to the clinging, insist on the insisted, incline the inclining, this tendency of the mind [i.e.] 'this is my self'.

Clearly this doesn't work so well in English and there are strong arguments for not trying to use Pāli syntax for English translations. How might we improve it then?

And he does not grasp, cling to, insist on, or incline to this tendency of mind [i.e.] 'this is my self'.

I have taken a liberty here. KS completes the pun by including abhiniviśati where Pāli lacks the parallel. Given the structure I believe it was intended to be included and that the Pāli scribes left it out in error. It completes the picture and it's hard to imagine the author of this play on words missing the opportunity. So the Sanskrit is not an interpolation.

Now one test of this is to look at how the Chinese translators handled it. In Chinese we would expect a phrase like 'he does not cling the cling' to be confusing because the two words would likely be represented the same character.

KC 若無此取者 is literally ‘if not a seizer of those’ (i.e. existence and non-existence). It corresponds closely to KS. etāni ced upadhyupādānāni, but is similar to KP. Tañcāyaṃ upayupādānaṃ when it is read in the light of Buddhaghosa’s commentary. This confirms that Buddhaghosa’s reading is the correct one.

KC 心境繫著使 breaks down as: 心境 ‘mental state’ which renders S. cetaso; 繫著 ‘to be bound, attached’ seems to correspond to KP adhiṭṭhānaṃ abhinivesa and KS adhiṣṭhānābhiniveśa, where abhiniveśa means ‘obstinate or tenacious’; 使 renders S. anuṣaya ‘bias, proclivity, tendency’.

不取、不住、不計 are clearly the equivalent of P. na upeti na upādiyati nādhiṭṭhāti. For 不住 compare P nādhiṭṭhāti (i.e. na adhiṭṭhāti) ‘does not insist’ where adhiṭṭhāti (Skt. adhitiṣṭhāti) is from adhi+√sthā. the character 住 means ‘stopping, settling, staying’ which is Sanskrit √sthā 'stand, remain', so I have read it as Sanskrit adhitiṣṭhati. Re 計 DDB includes the notions of ‘discriminating, construing, and positing’ so there has been a slight reinterpretation here from nādhiṭṭhāti ‘attā me’ti (doesn’t insist on 'this self is mine') to 不計我 ‘does not construe a self’. While a self (P. attā, S. ātman) is not explicitly denied in Pāli Nikāyas, thinking in terms of a self is discouraged in the strongest possible terms. The attitude seems to be that a self is not relevant. However it seems that as Buddhist philosophy moved towards more ontological thinking that the denial of the existence of a self seemed a natural progression from warnings not to think in terms of a self.

This passage in particular shows up the way that an Indic original helps to makes sense of the Chinese. A problem discussed by Bucknell (2010). By contrast previous translators, apparently relying on the Chinese alone have rendered this passage as:

“Suppose one is without this grasping, not grasping at a mental realm which causes suffering, not dwelling, and not discerning a self.” Lapis Lazuli (2010)

“In one who has no such attachment, bondage to the mental realm, there is no attachment to the self, no dwelling in or setting store by self.” Choong & Piya (2004)

Some of the nuances get lost. Clearly “grasping at a mental realm” or “bondage to the mental realm” is far less satisfactory than “mental state of attachment” in Buddhist doctrinal terms.

So the Chinese does not pick up on this elaborate pun that we see in the Indic texts, and lends weigh to my first translation. However the nature of the play on words gives the sentence an added and ingenious structure. We can see that the structure has been marred in both the extant Pāli and Sanskrit, which are, of course, both translations. However the structure gives us what is called a checksum in computer jargon: a way of assessing the fidelity of transmission. Metre is often able to alert a read that a passage has been altered. For example the last verse of the Kāraṇiya Mettā Sutta is in a different metre from the other nine verses suggesting perhaps that it was added later. The structure here allows us to see how the sentence was originally constructed and what it meant. Of course we do not know when or where this sentence was composed, nor by whom, but they were more than averagely clever in this instance. 

When a computer programmer leaves a little message, or even small application that performs simple and usually benign functions, hidden in their code it is called an Easter egg. It is something for later generations of programmers or users to discover and delight in. Here the early Buddhist author has left us an Easter egg, and if one appreciates the subtleties of Indic grammar it is quite delightful. 

Bibliography

KC: CBETA. http://tripitaka.cbeta.org/T02n0099_012 

KP: Chaṭṭha Saṅgāyana Tipiṭaka. Version 4.0. 1995. Vipassana Research Institute 

KS: Tripāṭhī, Chandra. (Ed.) (1962). 'Fünfundzwanzig Sūtras Des Nidānasaṃyukta' in Sanskrittexte aus den Turfanfunden (Vol. VIII). Edited by Ernst Waldschmidt. Berlin: Akademie-Verlag, 1962. [Includes translation into German]: 167-170. 

Bucknell, Roderick S. (2010) ‘Taking Account of the Indic Source-text,’ in Translating Buddhist Chinese: Problems and Prospects. Konrad Meisig (ed.). Harrossowitz Verlag. 

Choong Mun-keat & Piya Tan (2004) ‘Saṃyukta Āgama 301 = Taishō 2.99.85c-86a’. Dharmafarer. Online: dharmafarer.org/wordpress/wp-content/uploads/2009/12/6.13a-Sa-301-Kaccayanagotta-S-rev.pdf (pages numbered 89-91) 

Lapis Lazuli Texts (2011) ‘Saṃyuktāgama 301: Kātyāyana.’ Wikisource. http://en.wikisource.org/wiki/ Saṃyuktāgama_301:_Kātyāyana

Pañca-skandha: Etymology and Dynamics

SKANDHA in the Sanskrit Dictionary (Comparing MW and Apte) primarily means 'bough, branch; shoulder bone (scapular); break, fracture.'  It also means 'a branch of knowledge', 'a chapter in a book', and 'a division of an army' which are obvious metaphorical extensions. Added to this the word apparently means 'a troop or multitude', which is not obvious from the etymology. The supposed root is skandh 'to collect'; sometimes confused with skand 'spring, jump, spurt'  from PIE *skand (AHD) or related to Latin scandare 'to climb, ascend, descend'. Frankly it would seem that the dictionary definition and etymology are inconsistent, except, funnily enough in the Buddhist sense of a 'collection'. However in point of fact the skandhas are not collections at all, and I can't help but wonder if 'collections' have been tacked on here precisely to accommodate Buddhist usage, thus fouling the trail.

Clearly the main meaning of 'branch' (or a tree or a line) is not related semantically to 'collect', but etymologists also draw attention to semantic similarities with Greek κλάδος (klados) 'branch, shoot, an offshoot' (CSED); from PIE *kldo- from *kel 'to cut' and giving rise to English 'calamity, clade, clast, gladiator' etc. The problem here is that, while the semantic field is very similar, the phonetics don't seem to be related.

Another possibility is that skandha comes from a PIE root *(s)k(h)ed or *sked (AHD). It has a form  nasalised form *skend and partly covers the semantic field 'to split' - which would connect it semantically with *kel 'to cut' This root is not well represented in Sanskrit, but is the root behind English words such as 'scatter, shatter and shingle'. If this is the PIE root then it is not well represented in Sanskrit either. I think we have skandha < *skand 'to split off, to branch' and that skandhayati is a denominative. I don't see any other related forms. This rarity of form may have led to confusion with Sanskrit skandh and (homonymous) skand 'to jump'. However I can't be sure of this without a lot more research.

Thus the etymology of the word is obscure. However I think we can agree that 'aggregate' as a translation is madness, and that other modern translations of Buddhist usage (group, mass, heap) are almost equally unhelpful. If indeed skandha were related to Greek klados, at least semantically, then we might translate skandha as 'branch'. The skandhas, then, might be thought of as the five 'branches of experience'. This might be interesting in light of the word prapañca (On the origin of the word prapañca from the branching of the hand into five fingers see: Translating Prapañca). We might postulate that prapañca (becoming fivefold) is related to the five branches of experience, though clearly some work would be required to establish this speculation. Since this is a blog rather than an academic publication I can take liberties and so for this essay I translate skandha as 'branch'; as one of five branches (pañcaskandha) of experience; it's justified by the dictionary and, as we will see, it works pretty well here. I would note that the branching of the end of limbs into five digits is universal amongst animals. What appear to be exceptions (horse hooves etc) turn out to have have fused two or more of the five digits and embryonic forms often begin with five distinct digits. This ought to be a very common basis for metaphors and we should not be surprised to find it.

Richard Gombrich (What the Buddha Thought, 2009) has suggested with reference to the term pañc'upādāna-kkhandhā, that Pāli upādāna-kkhandha should be read as a contraction of upādāna-aggi-kkhandha where aggi-kkhandha means 'a blazing mass' but c.f. Sanskrit skandhāgni 'fire made with thick logs'. (MW) I now wonder whether aggi-kkhandha just means 'a burning bough'. The phrase pañca upādānakkhandha is often rendered 'five aggregates of clinging', where we might read it as 'five branches which are fuel [for the fire]'; where upādāna means 'fuel' and 'the fire' is 'the fire of being'. I find the connection with the extended fire metaphor entirely plausible (see Playing with Fire; and Everything is on Fire!) and I suggest that it works even better when khandha is understood as 'branch'.

Having looked, somewhat inconclusively, at the meaning of skandha, let us now examine a passage from the Pāli Mahāpuṇṇama Sutta (M 109; iii.15ff ) which tells us something about the relationships of the skandhas:

The four elements are the reason for, the condition for the manifestation of the branch of form. 

Contact is the reason for, the condition for the manifestation of the branch of sensations. Contact is the reason for, the condition for the manifestation of the branch of recognition. Contact is the reason for, the condition for the manifestation of the branch of volitions. 

Name & form is the reason for, the condition for the manifestation of the branch of cognition.

Here someone has tried to show the dynamics of the skandhas. The form branch (rūpa-kkhandha) of experience is made up from the four elements: earth, water, fire and wind (catur mahādhātu: paṭhavī, āpo, tejo, vāyo;); or in experiential terms: resistance, cohesion, heat and movement. There's every reason to believe that just as cakkhu 'the eye' stands for the visual faculty; that paṭhavī stands for the experience  of resistance. Before looking at the other branches let us look at an interesting passage (SN 35.93) which sheds further light on this. 

The production of cognition is conditioned by a pair. Visual-cognition arises conditioned by the eye and forms. The eye is impermanent, reversible, and unstable. Forms are impermanent, reversible, and unstable. So this pair is transient, wavering, impermanent, reversible, and unstable.

Visual-cognition is impermanent, reversible, and unstable. Whatever the reason (hetu) or condition (paccaya) for the arising of visual-cognition, it is impermanent, reversible, and unstable. So how could visual-cognition ever be permanent?

These three things [i.e. forms, eye, & visual-cognition] coming together, encountered and co-occurring are called visual-contact (cakkhu-samphassa). Visual contact is impermanent, reversible, and unstable. Whatever the reason or condition for the arising of visual-contact, it is impermanent, reversible, and unstable. So how could visual-contact ever be permanent?

The contacted (phuṭṭha) is sensed (vedeti); the contacted is willed (ceteti); the contacted is recognised (sañjānāti). So these things (dhammā) are transient, wavering, impermanent, reversible, and unstable.

[The other senses are outlined in identical terms.]

So now our model has rūpa being made up of the elements (dhātus); and contact (phassa) is the condition for the next three branches. S 35.93 fills in a gap here. Contact comes about when rūpa, cakkhu (the eye, or visual faculty) and cakkhuviññāna 'visual cognition' come together; the latter in fact being conditioned by the former pair. And then contact is sensed, willed, and recognised. 

The last three mental actions are equivalent to the three middle branches. We can match up the verbal and nominal forms: vedanā vedeti 'he senses sensations'; saṅkhārā ceteti 'he wills volitions'; and saññā sañjānāti 'he recognises names' or 'he recognises recognitions'. However note that the three verbal forms are in a different order than the nominal forms. And this is unusual. Sue Hamilton's comprehensive survey of the skandhas in Pāli finds that they always occur in the same order (Early Buddhism, p.72). 

Next we add the three branches of vedanā, saññā and saṅkhārā to rūpa to make up nāmarūpa; and nāmarūpa is the condition for the arising of viññāna.

The astute reader will have already spotted a problem here: viññāna (associated with the six kinds of senses) is the condition for the arising of vedanā, saññā and saṅkhārā and thus indirectly the condition for  nāmarūpa. Thus we have two models in which the direction of the relationship between nāmarūpa and viññāna goes in opposite directions. How can we have two mutually exclusive models which are both  canonical? In the  Mahānidāna Sutta (D 15) nāmarūpa and viññāna condition each other? If this was the original idea it might have survived in two fragmentary and contradictory forms with unidirectional conditionality. Or perhaps the Mahānidāna Sutta sought to harmonise the two different models by combining them. Certainly the Mahānidāna Sutta says that nāmarūpa is root, cause, origin and condition for contact (tasmāt iha ānanda, eseva hetu etaṃ nidānaṃ esa samudayo esa paccayo phassassa, yadidaṃ nāmarūpaṃ). Nāmarūpa is quite a problematic term in its own right: see Nāmarūpa. And in another model we find "nāma is 'sensation, recognition, impulse, contact and attention'." (vedanā, saññā, cetanā, phasso, manasikāro – idaṃ vuccati nāmaṃ. S ii.3). And this in a model where phassa is a stage in its own right that comes after nāmarūpa. 

It must be the case that viññāna is being used in at least two different senses: cakkhuviññāna (and the other sense-viññānas) and viññāna (as a standalone) cannot be referring to the same process or even the same kind of process. The standard explanation is that viññāna 'furnishes bare cognition of the object' (Nyanatiloka. Buddhist Dictionary) but this is still contradictory. In the model of the five branches we're looking at viññāna simply cannot amount to 'bare cognition' since it is preceded by vedanā, saññā and saṅkhārā. Without 'bare cognition of the object' how could these three exist? Is the problem with the phrase 'bare cognition'? 

Elsewhere I have pointed out that viññāna is said in Pāli to be always related to the sensory stimulus that conditioned it. For example in the Mahātaṇhāsaṅkhaya Sutta. (MN 38. PTS M i.259):

Bhikkhus from whatever condition viññāṇa arises, it is called that kind of viññāṇa. Viññāṇa arising with the eye and form as condition, is called eye-viññāna (i.e. visual cognition).

From the same sutta we know what viññāna is not: "that which speaks and feels, that which experiences the good and bad." (yvāyaṃ, bhante, vado vedeyyo tatra tatra kalyāṇapāpakānaṃ kammānaṃ vipākaṃ paṭisaṃvedetī ti) This is cited as an example of a wrong view. We also know that Buddhists do not posit a 'theatre of consciousness', a metaphorical container in which experiences happen, since viññāna is seldom if ever used in the locative case, and where it is the text is usually arguing against a wrong view.

Idiosyncratic, but none the less insightful, bhikkhu Ñānavīra, says

Consciousness (viññāṇa) can be thought of as the presence of a phenomenon, which consists of nāma and rūpa. Nāmarūpa and viññāṇa together constitute the phenomenon 'in person'--i.e. an experience. The phenomenon is the support (ārammaṇa) of consciousness, and all consciousness is consciousness of something (viz, of a phenomenon). [Notes on Dhamma, p.81]

I don't think Ñāṇvīra has solved the problem I have identified here, i.e. the role of viññāṇa in the khandhas.  Since nāma depends on sense-viññāṇa. However he does add an interesting caveat to this discussion.

Consciousness, it must be noted, is emphatically no more 'subjective' than are the other four upādānakkhandha (i.e. than nāmarūpa)... It is quite wrong to regard viññāṇa as the subject to whom the phenomenon (nāmarūpa), now regarded as object, is present. [Notes on Dhamma, p.82]

Back in April 2012 I argued that Westerner terms like subjective and objective only obscure the discussion. In the Buddhist model of consciousness, subject and object are not relevant. This is a corollary of the idea that existence and non-existence don't apply to experience. We only get confused trying to marry the two modes of thinking about our experience of being aware. Indeed to think of viññāṇa as 'consciousness' is demonstrably wrong (see  The 'Mind as Container' Metaphor.) Nāṇavīra himself implies this, but doesn't not make the leap to rejecting the translation outright. We have our discussions as if Western concepts like consciousness, subject, object, etc., are givens. But they are not. The early Buddhists don't seem to have thought in these terms. Where they encounter this style of thinking, they tend to treat it as irrelevant to the task at hand. Ontological questions that fascinate Westerners, are just a distraction to early Buddhists. We are not trying to understand our self or our consciousness in the abstract, we are trying to understand why we suffer. Of course early Buddhists had a raft of assumptions about their experience of the world and we need to tease these assumptions out in order to understand their worldview. But imposing modern philosophical jargon often obscures more than it reveals.

This essay has at least shown how translating skandha as 'branch' and pañcaskandha as 'the five branches [of experience]' might work, and might be more comprehensible than present alternatives. It is intrinsically interesting that at least one of the strands that made up the Pāli Canon attempted to give the five branches of experience a temporal sequence, though whether it works is moot. My previous understanding, based on reading Sue Hamilton was that the skandhas did not form a sequence, but were to be taken collectively as the "apparatus of experience". That this attempt breaks down with the inclusion of viññāṇa is a puzzle that I'm sure I will come back to.

We have considerable work to do still to untangle early Buddhist ideas about why we suffer from Western ideas about our existence and the ontology of the self with which they seem to have been snarled. The first step in any comparative philosophy will be to understand early Buddhism on it's own terms and I don't think we have done this yet.

~~oOo~~

Note 5/5/13

In article by Collete Cox I found this the re the *Mahāvibhāṣā "In a discussion of the implications of the various meanings of the term "aggregate," or skandha, as a heap (rāśi), a bundle, an assemblage, or as a collection..." Thus the usage 'heap' predates Conze!
Cox, Collett. (2004) 'From Category To Ontology: The Changing Role of Dharma in Sarvāstivāda Abhidharma.' Journal of Indian Philosophy. 32: 543-597.