It was with some anticipation that I began to read Ronald Davidson's new review article in the Journal of Indian Philosophy on the meaning of dhāraṇī in Mahāyāna Buddhism - a subject in desperate need of an overhaul. However Davidson seems to have misunderstood crucial aspects of the system of practice in which early dhāraṇī was located. My comments will mainly concern his understanding of the Arapacana alphabet, especially as it occurs in the Large Perfection of Wisdom Sutra,  however I flag up wider concerns as well.
Gāndhārī Alphabet in
the Kharoṣṭhī script
Gāndhārī Alphabet in
the Kharoṣṭhī script
Davidson proposes the idea that the main point of the words associated with the Arapacana is to draw attention to how the letters can support (carry √dhṛ) meaning - thus linking dhāraṇī with the type of esoteric speculation in the early Upaniṣads and Tantric Buddhism. He explicitly denies the other and more natural possibility that the letters are mnemonics for words and concepts. His contention seems to be that the relationship must be this way around because in some texts different words are associated with the letters. The existence of variations on the theme in different texts surely suggests a technique widely used in different contexts, rather than incoherence or simple polysemy.
There are two main objections to Davidson's thesis. He argues that the words indicated by the syllables are intended to help the student remember the alphabet. Even if we put aside the fact that the Gāndhārī alphabet continues to be used even when the rest of the work is composed in Sanskrit, and can therefore be of little practical use for learning there are deeper problems with the idea that the Arapacana developed in this way. Davidson uses same example already put forward by B.N. Mukherjee, though he seems unaware of this: a is for apple, b is for bear etc. But stop and think about this. A in the Arapacana, even in the very early versions, is for anutpannatva. This is an abstract noun from anutpanna (not-arising) meaning 'not-arising-ness'. In fact this is one of the most complex abstract ideas of Indian philosophy which cannot be easily understood outside the context of many years of instruction in Buddhist thinking. The other 4o odd letters stand for equally complex abstract concepts. Can Davidson really believe that such an abstruse abstract notion would be of use to a learner trying to memorise the alphabet? Surely this would be an impediment rather than a helpful mnemonic device! When we teach the alphabet we use concrete examples. I note that the children's Devanāgarī chart I picked up last time I was in India uses concrete examples as well: e.g. a is for anāra (pomegranate) and bha is for bhālū (bear).
It makes much more sense to think of the letters as a mnemonic for the concepts, not the other way around. Davidson suggests that literacy in the India world at this time was low, but even if literacy was low in the rest of the world generally, Buddhist monks in Gandhāra probably all learned basic reading and writing, since the reading of texts had by then become a fundamental monastic skill. Indeed Buddhist monks were the primary vector for literacy in most of Central, Southern and South-East Asia as the persistence of Brahmī derived scripts testifies!
More broadly the very presence of such lists and this level of abstraction speak of a written rather than oral culture. I've written about the probable Persian influence the alphabetical list, and that was a literate culture without any doubt, and their writing formed a model for the Kharoṣṭhī script [see: Persian Influences on Indian Buddhism]. However here I'm particularly thinking of the characteristics of oral cultures enumerated by Walter Ong - "an oral culture has no vehicle so neutral as a list... [oral cultures are] situational rather than abstract, unavoidably using concepts but again within situational frames of reference that are 'minimally abstract'." 
The other objection is broader. If we look at the words indicated by the Large Perfection of Wisdom Sutra, and most other versions of the Arapacana  then we see that they are all related to śūnyatā - the notion that experiences are neither existent nor non-existent, that they have no independent existence (svabhāva). This is the heart of the Perfection of Wisdom approach to practice. Indeed taking the Arapacana in context with other statements in the sūtra  we can see that they form the basis of a insight meditation practice - by reflecting on various aspects of śūnyatā one comes to see the true nature of experience, and is liberated. The texts emphasise the sameness (samatā) of each of the syllables, not because of the inherent polysemy of letters making them interchangeable which they plainly are not, but because the concepts which they stand for show the practitioner the truth about experience being śūnyatā - śūnyatā is the common characteristic (i.e. the basis for the sameness) of all experience. Davidson seems to have lost sight of Nāgārjuna's polemics against ontology, not to mention the Buddha's.
What I think Davidson is doing is reading the texts with a particular result in mind, specifically that the word dhāraṇī can best be understood as meaning code/coding. I wholeheartedly agree that other contemporary writers have erred in emphasising the mnemonic function of dhāraṇī generally or in maintaining the fiction that dhāraṇī are somehow 'summaries' of the text they appear in. The mnemonic function is restricted solely to the Arapacana context, though it clearly is a mnemonic in this context contra what Davidson says. I do not believe that I have seen any dhāraṇī that comprehensibly summarises a text - though of course this has not stopped people producing ad hoc/post hoc exegesis on the basis that dhāraṇī are somehow summaries. Witness the many and varied readings of the Heart Sūtra mantra for instance - most of which are mutually contradictory!
A far better attempt, though more limited in scope, was published by Paul Copp in 2008.  Copp explores the way the word is used in Chinese translations of the Bodhisattvabhūmi and the Treatise on the Great Perfection of Wisdom Dazhidu Lun (大智度論). Copp shows that the basic meaning of the word dhāraṇī in these contexts can best be understood as 'grasp' - used in the sense of grasping the meaning, holding in memory, keeping in mind, etc. I concur. Just as the various meanings dharma (which I explored in Dharma - Buddhist Terminology) can be understood in terms of 'foundation' used literally, abstractly and metaphorically. It remains for Copp to show how his ideas fit into a much broader context, but his views seem more promising. I certainly prefer Copp's method of working from the texts to see what the word must mean in context, than Davidson's reading the meaning into the text.
From the point of view of a practitioner Davidson's error is perhaps an understandable one. For him the ideas do not seem to be tied into the practical use that is made of them: Buddhism is an intellectual system to be studied and understood in contemporary Western terms. No doubt he understands that Buddhists practice Buddhism, but the deeper implications of this pragmatism are not apparent. The impracticability of teaching an alphabet with recondite abstractions is only the most obvious sign of this.
One useful thing in Davidson's article is his survey of the history of the Western commentary on dhāraṇī - this threw up a few references I had not come across before. But that history is a bit depressing - it is a history of misunderstandings and the clash of Western preconceptions with Buddhist preoccupations. We're still trying to disentangle ourselves from that train wreck and in my opinion Davidson is pulling in the wrong direction.
- The Large PoW Sutra was translated by Conze but for variety of reasons the translation is less than satisfactory: for instance Conze was not working from an edited text and freely used passages from other versions in 18,000 and 100,000 lines where his manuscript (which itself has many faults) let him down. He also rearranged the text to suit subject headings from the Abhisamayālaṅkāra. Dutt's edition of the Sanskrit is flawed in the Arapacana sections with some doubling of syllables (which may be why Conze did not use it). Dutt was editing the text in the years before Salomon demonstrated that it was a real Alphabet. KIMURA is bringing out an edited Sanskrit text (see below) but the crucial part with the Arapacana is in the volume which has not yet been published. However some other related passages are available and I am working on translated them with my rather haphazard Sanskrit. - return to article
- [my italics] Walter J. Ong. Orality and Literacy, cited in Lopez, Donald S. Elaborations on Emptiness, p.33. - return to article
- The Arapacana in the Gandhavyūha Sūtra is the major exception. In this version the keywords do not relate to the alphabet at all indicating that the point of the exercise has been missed in this case. In this case the exception proves the rule. - return to article
- See for instance passages at p.162, 488-9, and especially p.587 in Conze's translation. I wonder if these scatter references were once more closely associated? - return to article
- Davidson may have been writing before Copp published, but does not seem to be aware of the article. - return to article
- Conze, Edward (trans.) The Large Sutra on Perfect Wisdom. University of California Press, 1975.
- Copp, Paul. Notes on the term Dhāraṇī in Medieval Chinese Buddhist Thought. Bulletin of SOAS. 71 (3) 2008: 493-508.
- Davidson, Ronald. 'Studies in Dhāraṇī Literature I: Revisiting the Meaning of the Term Dhāraṇī'. Journal of Indian Philosophy. 37 (2) April 2009: 97-147.
- Dutt, N. Pañcaviṃśatisāhasrikā Prajñāpāramitā Sūtra: edited with critical notes and introduction. London, Luzac & Co, 1934.
- KIMURA Takayasu : Pañcaviṃśatisāhasrikā Prajñāpāramitā. Tokyo: Sankibo Busshorin 1986. Vols II-V (vol I forthcoming) Online: http://www.sub.uni-goettingen.de/ebene_1/fiindolo/gret_utf.htm#PvsPrp
- Lopez, Donald J. 1996. Elaborations on Emptiness: Uses of the Heart Sūtra. Princeton University Press.
- Mukherjee, B. N. 1999. Arapacana: a mystic Buddhist script in Bhattacharya, N. N. (ed) Tantric Buddhism. New Delhi: Manohar Publishers & Distributors. p.303-317.