23 November 2012

Heart Sutra Syntax

Dr. Edward Conze
UPDATED. This is one of those issues where I just have to write down my thoughts and there's not enough to warrant a published article.

I'm studying the Heart Sutra to practice my Sanskrit, and for the joy of chanting it aloud.  And I've found an interesting puzzle. Like many people I'm familiar with Conze's critical edition in Buddhist Wisdom Books (2nd Ed. 1975); checked against the version in Thirty Years of Buddhist Studies (1967) which contains the variant readings he found, and against Vaidya's edition (which is also online). The part that concerns me is right at the start:
ārya-avalokiteśvaro bodhisattvo gambhīrāṃ prajñāpāramitā-caryāṃ caramāṇo vyavalokayati sma: panca-skandhās tāṃś ca svābhavaśūnyān paśyati sma.
I think this sentence has been parsed incorrectly by Conze. There are three syntactical problems with this rendition.

Firstly the verb vyavalokayati (vi + ava + √lok; conjugated as a 10th class verb) has no object. If we break the sentence when Conze says, then it suggests that the verb vyavalokayati is intransitive. Avalokiteśvara is just looking, not looking at anything. This is a little bizarre, but worse when we consider the rest of the sentence.

The second problem is the place of ca. This is an enclitic particle that follows the word it applies to. Here it joins two sentences. This suggests that the second sentence must begin with tāṃś - an accusative plural 'they' (with a sandhi change from tān to tāṃś to accommodate the following ca). This means that pañcaskandhās must belong to the first sentence, but as it is in the nominative singular case it doesn't fit. In this case tāṃś is referring back to the object of the previous sentence - except that it does not have one.

Thirdly tāṃś (pronoun) and svabhāvaśūnyān (adjective) are both accusative plural forms. With what are they agreeing if not the five skandhas?

Let me just describe what is happening in this sentence for the non-Sanskrit reader. The first sentence has two clauses: The subject of the sentence is the Noble Bodhisattva Avalokiteśvara (ārya-avalokiteśvaro bodhisattvo). He is practising (caramāṇo) the deep (gambhīrāṃ) practice of the perfection of wisdom (prajñāpāramitācaryāṃ). Here caramāno is a secondary verb in the form of a present participle which describes an action taking place at the same time as the main verb. The two words gambhīrāṃ prajñāpāramitācaryāṃ are in the (feminine) accusative singular - gambhira being an adjective which must match it's noun.  As such prajñāpāramitācaryāṃ is the object of the secondary verb, caramāṇo.

While practising Avalokiteśvara did the action of the verb vyavalokayati sma. The particle sma is called a "periphrastic past". To make a present-tense verb past, one has the option of just adding sma after it, as paśyati 'he sees' does in the next phrase. Conze's rendition of this as "looked down from on high" is fanciful and seems to be based on over analysing the word avalokita. Edgerton's Buddhist Hybrid Sanskrit Dictionary is more sober and suggests 'inspecting, examining' which makes more sense. Conze got carried away here in reading vyavalokayati as an intransitive verb and it left him a problem in the form of pañcaskandhās which has no clear syntactic relationship with the other words in either sentence, but is important in the text.

Now we know that extant manuscripts have a lot of variation - hence the need for critical editions. Conze (1967) notes here show that there is a majority with his reading, but some variations amongst the 12 Nepalese manuscripts. Regarding this passage (N stands for Nepalese manuscript).
Ne - omitted;
Nb,c pañca-skandhān svabhāva-śūnyāṇ vyavalokitavyam
Nk (begins) vyavalokitavyam
Nd,l - omitted.
Vaidya has:
... व्यवलोकयति स्म। पञ्च स्कन्धाः, तांश्च स्वभावशून्यान् पश्यति स्म॥
... vyavalokyati sma | pañca skandhāḥ, tāṃśca svabhāvaśūnyān paśyati sma
Vaidya has the same problem: what to do with pañcaskandhāḥ in the nominative plural? Now Conze renders this part of the passage "He looked down from on high, He beheld but five heaps, and saw that in their own-being they were empty".

I think this is a blunder on Conze's part, because the sentence simply can't say this. Pañcaskandhās is unambiguously a nominative plural, and thus the subject of the sentence. In the sentence pañca skandhāḥ, tāṃśca svabhāvaśūnytān paśyati sma it is clearly the pañca-skandhāḥ that did the seeing. Hence the superimposed comma (which Sanskrit entirely lacks) because otherwise it is nonsense.

Conze's reading is also problematic because it suggests that the skandhas have self-existence (svabhāva) which is empty; and as I understand this text (and Nāgārjuna) a central plank of Prajñāpāramitā thinking is that dharmas and skandhas all lack svabhāva. Self-existence is not possible according to Nāgārjuna. Surely the text must be saying that the skandhas are empty of svabhāva, rather than that their svabhāva was empty?

There is a simple solution to this dilemma which is suggested by Conze's Nepalese manuscripts b & c. Which is that pañcaskandhās is in fact an accusative plural: pañcaskandhān. However -ān followed by t undergoes compulsory sandhi change to -āṃs. Thus we expect to see pañcaskandhāṃs tāṃś ca. The difference is more subtle in Indic scripts which indicate the anusvāra with a dot:

पञ्चस्कनधांस्  vs  पञ्चस्कन्धास् 
pañcaskandhāṃs vs pañcaskandhās   

We know that Nepalese manuscripts are often sloppy with anusvāra. It solves both the problem of the lack of object for vyavalokayati and the placing of ca. In favour of this theory is that in Vaidya's edition of the long Heart Sutra we find just this, i.e. pañca skandhāṃstāṃśca.

The amended passage now reads:
āryāvalokiteśvaro bodhisattvo gambhīrāṃ prajñāpāramitācaryāṃ caramāṇo vyavalokayati sma pancaskandhāṃs, tāṃś ca svābhavaśūnyān paśyati sma.
Noble Avalokiteśvara, the Bodhisattva, practising the deep practice of the perfection of wisdom, examined the five skandhas and saw them empty of self-existence.
This simple change also avoids the awkwardness of Conze's translation, and makes more doctrinal sense.

It's important to be aware that the original Sanskrit manuscripts did not have punctuation and the hand writing was often very poor, especially on later Nepalese manuscripts. Conze himself complains of this. A manuscript might have looked a bit like this:

There's little or no punctuation in this style of writing, and no word breaks as each syllable is written as a standalone. For the record pañcaskandhāstāṃśca would look like this in Devanāgarī

पं च स्क न्धा स्तां श्च

paṃ ca ska ndhā stāṃ śca
In this manuscript it would not be hard to confuse stāṃ and ntāṃ, it's only a matter of a single misplaced stroke, or a smudge. In the image below I've taken the syllable from the manuscript above: stā is on the left and is altered to read ntā on the right (this ms. leaves off the anusvāra or ):

However the loss of anusvāra is the simplest answer to the conundrum. I'll see if I have time to check the Chinese at some future date. This section of the text was almost certainly composed in China, and there is no parallel in the Large Perfection of Wisdom Sutra.

Conze, Edward. (1967) ‘The Prajñāpāramitā-Hṛdaya Sūtra’ in Thirty Years of Buddhist Studies: Selected Essays, Bruno Cassirer. p. 147-167. (Originally published in: Journal of the Royal Asiatic Society, 1948, pp. 33-51.)

Conze, Edward. (1975) Buddhist Wisdom Books: The Diamond Sutra and the Heart Sutra. 2nd Ed. London: George Allen & Unwin.

'Facsimile of the Horiuzi Palm-leaf MSS. of Hōryū-ji Monastery' in Buddhist Texts from Japan. (Anecdota Oxoniensia, Aryan series), 1881. Online: http://en.wikipedia.org/wiki/File:Falongsibeiye.png
Vaidya, P.L. (1961) Mahāyāna-sūtra-saṁgrahaḥ ( part 1). Buddhist Sanskrit Texts No. 17. Darbhanga: The Mithila Institute. Online: http://dsbc.uwest.edu/ [This volume contains editions of both long and short versions of the sutra].

Chinese Texts
照見五蘊皆空T 8.251 ST XuanzangHe observed that the fives skandhas are all empty.
照見五隠空T 8.250 ST KumarajīvaHe observed that the five skandhas are empty.
照見五蘊皆空T 8.253 STHe observed that the fives skandhas are all empty.
照見五蘊自性皆空。彼了知五蘊自性皆空:T 8.252 LTHe observed that the five skandhas where empty of self-existence. He knew the five skandhas were empty of self-existence.
照見五蘊自性皆空T 8.254 & 257 LTHe observed that the five skandhas where empty of self-existence.
觀察照見五蘊體性悉皆是空T 8.255 LTHe observed that the five skandhas where empty of self-existence. (with synonyms)

LT = Long text; ST = Short text.

15 November 2012


Just a quick note to let people know that an article of mine has been published in the Journal of the Oxford Centre for Buddhist Studies. Vol 3.

This article explores the plausibility of Michael Witzel’s speculation that the Śākya tribe might have Iranian origins, or at least Iranian connections. Circumstantial evidence suggests that ideas associated with Iran and Zoroastrianism appear in north-east India, especially amongst the śramana groups,and in particular amongst Buddhists, but not in the Brahmanical culture.Whereas Buddhism is frequently portrayed as a response to Brahmanism,or, especially by Buddhists, as ahistorical, Witzel’s suggestion gives us a new avenue for exploring the history of ideas in Buddhism. This essay attempts to show that, at the very least, possible connections with Iran deserve more attention from scholars of the history of ideas in India and especially Buddhism.
Related Posts with Thumbnails