Dhāraṇī - origins, meaning, and usage.

The word dhāraṇī is a characteristically Buddhist term at times synonymous with mantra, and at others seeming to have it's own special significance. In this short essay I want to examine the word, and the main ways it is used.

The word dhāraṇī, according to Edgerton's dictionary of Buddhist Hybrid Sanskrit, does not occur outside of texts written in BHS. This point is sometimes summarised as "does not occur outside of Buddhist texts", but Edgerton's point is more specific and that specificity has some possible consequences. We need to be aware here of the shifting and uncertain nature of BHS. BHS is in fact a Prakrit language that was in the process of being Sanskritised. By which we mean a vernacular North Indian dialect was being regularised in it's grammar to conform to the ideals of linguistic form represented by Classical Sanskrit*. As such BHS shows considerable variation in grammar and spelling especially in the area of inflections - the suffixes added to words to indicate the grammatical relationship between them.

Buddhist texts cover a spectrum:

• Texts written in relatively pure Prakrits (the Gāndhārī texts for instance),
• Texts written in Pāli, a somewhat artificial "church language" constructed from several Prakrits.
• Texts in which the Prakrit has begun to be Sanskritised
• Texts in which the process of Sanskritisation is well advanced
• Texts in more or less pure Classical Sanskrit (e.g. Aśvaghoṣa's Buddhacarita)

In fact there is a word in Sanskrit - dhāraṇa. It means, according to Monier-Williams:

"holding, bearing, keeping (in remembrance), retention, preserving, protecting, , maintaining, possessing, having".

This is so close to the uses of our word that I am somewhat surprised that the literature supplies no argument for distinguishing the two terms. Remember in BHS spelling is variable. In Tibetan the word is frequently translated, again according to Edgerton, as "gzuṅs, literally, "hold, support". This suggests that the Tibetans understood dhāraṇī to by synonymous with dhāraṇa. My linguistic knowledge is not sufficient to press the point, but it seems so obvious that I wonder why no one more qualified has not dealt with this issue.

Jan Nattier suggests that the earliest use of the term dhāraṇī occurs in relation to the Arapacana Alphabet (Nattier : 292) - now known to be the alphabet of the Gāndhārī Prakrit. This alphabet, uniquely in India, was used as a mnemonic device, a kind of acrostic where each letter stood for a keyword, which then became the subject of a phrase. By the time of the Pañcaviṃśatisāhasrikā Prajñāpāramitā Sūtra (ca. 2nd century) this technique was being used as a memory aid for a meditation on aspects of śunyata. From this usage we find the word dhāraṇī associated with mnemonic devices - many writers insist the dhāraṇī is always a mnemonic device. However a glance over some of the many dhāraṇī's preserved in, or as, texts will quickly make this identity much less certain. Most dhāraṇī apparently have no mnemonic features, i.e. they do not appear to stand for other things. They do employ many of the prosodic features of poetry in order perhaps to help them be memorable, but they do not seem to, as some authors would have us believe, "summarise the text to which they are attached". More often a dhāraṇī bears no apparent relationship to a text, even when it is strongly associated with a text - as in the very prominent case of the Heart Sūtra where interpretations of what the mantra means are as numerous as are commentaries on the text. That there is no consistent exegetical tradition associated with any of these dhāraṇī only serves to confirm this impression.

Like mantras dhāraṇī come in a variety of forms. In early Buddhist texts markers at the beginning such as 'oṃ' or 'namaḥ samanta buddhāṇāṃ' are missing. dhāraṇī can be strings of words, frequently all with the same grammatical ending (usually the feminine vocative). An example from the Saddharmapuṇḍarīka Sūtra is:

Anye manye mane mamane citte carite same samita viśānte muke muktame same avishame samasame jaye [kṣāye] akṣāye akṣaīne śānte samite dhāraṇī ālokabhāshe pratyavekṣāṇi nidhiru abhyan taranivishṇe abhyantarapāriśuddhi utkule mutkule araṭe paraṭe sukāṅkṣaī asamasame buddhavilokite dharmaparīkṣaite saṃghanirghoshaṇi [nirghoshanī] bhayā-bhayaviśodhani mantre mantrākṣāyate rule rutakauśalye akṣāye akṣāyavanatāye [vakkule] valoda amanyanatāye [svāhā]. (Bunnō : 329) **

Such strings make frequently use of poetic devices such as alliteration, repetition, and often make use of phonetic variations on a theme. These are clearly visible in the first line of the dhāraṇī above. Alternatively they may be strings of syllables which do not make words. Again from the Saddharmapuṇḍarīka:

iti me iti me iti me iti me iti me ni me ni me ni me ni me ni me ruhe ruhe ruhe ruhe ruhe stuhe stuhe stuhe stuhe stuhe [svāhā]. (Bunnō : 331)

Here the effect is of repeated sounds, which to my ear suggests some kind of sound symbolism. On a Buddhist online forum one member suggested that they represent coded coordinates for some object like a stupa, but as far as I know this is pure speculation. Though the argument is similar to ones made by Subhash Kak about codes in the Ṛgveda.

Another kind of dhāraṇī reads like a poem or prayer to a particular deity. These are more like the Vedic mantra in literary character - here we could translate dhāraṇī as "hymn" just as many Vedic scholars do for mantra. These dhāraṇī are part of an extant Buddhist tradition which is rooted in Pure Land ideas: chanting the dhāraṇī invokes the saving power (or vow) of the Buddha or Bodhisattva, delivering the chanter either from some immediate misfortune, or ultimately from the suffering of saṃsara altogether.

One oddity of the way the word dhāraṇī is used is that it can be both the means to the goal, and the goal itself. One chants a dhāraṇī in order to be protected or gain insights; however some texts talk about the acquisition of dhāraṇī as one of the results of the Bodhisattva's practice. The Lotus Sūtra deities offer dhāraṇī to be memorised and chanted for protection, while the Pañcaviṃśatisāhasrikā has the Bodhisattva attaining dhāraṇī : in this case dhāraṇī almost seems to be synonymous with samādhi, and note that this is sometimes how the word dhāraṇa is used in the Upaniṣads (see for example Deussen : 389f)

It is frequently assumed that dhāraṇī, and the Pāli paritta texts, are simply precursors to mantra. However I'm not convinced that there is a continuity here. Some of the popular dhāraṇī texts did end up being considered to be "kriya tantras" by later Tibetan exegetes, but there is nothing in the content of these dhāraṇī, nor in the context in which they occur, to suggest that they function like mantras in the Tantric sense. This identification has lead some scholars, for instance Robert Thurman, to argue for very early dates for Tantric texts, when other evidence makes it seem very unlikely.

A more thorough exploration by a qualified scholar is eagerly awaited, although I am not presently aware that any scholar of Buddhism is taking an interest. I speculate that a closer analysis of the evidence will reveal a more subtle interplay of religious ideas and impulses at work, and make it clear that dhāraṇī and paritta are not in origin at least, simply mantra by another name. The word dhāraṇī came into play in a time and place of innovation: in the 1st-2nd centuries in Gāndhāra, under foreign rulers (the Kuṣans), during which period also the first images of the Buddha were made, and the Mahāyana began to be mainstream. However it was quickly taken up by the Buddhist world - new ideas appear to have spread quickly at this time, perhaps due to extensive trading networks. The term then appears to have undergone a process of evolution over several centuries until the advent of Buddhist mantra proper, probably in the 7th century, when it was subsumed under that rubric. Traditional explanations of what makes dhāraṇī distinctive lack this historical perspective, while contemporary accounts have jumped too quickly to the conclusion that similarity equals sameness.



Notes.

• * Where I do not qualify it the word Sanskrit will mean specifically Classical Sanskrit from now on.
• ** Square brackets in this and the next quote indicate that the author has reconstructed the Sanskrit original from a Chinese text, and it is speculative. Note there is a Sanskrit text but it is very late and not necessarily more accurate.
  

Bibliography.

• Bunnō, K. et al. 1986. The threefold Lotus Sutra : Innumerable meanings ; The lotus flower of the wonderful law ; meditation on the bodhisattva Universal Virtue. Tokyo : Kosei Publishing.
• Deussen, Paul. 1906. The philosophy of the Upanishads. (trans. by Geden, A. S.) New York, Dover Publications, 1966.
  
• Nattier, Jan. 2003 A few good men : the Bodhisattva path according to 'The inquiry of Ugra' (Ugraparipṛcchā). University of Hawai'i Press.

image: dhāraṇī in the Siddhaṃ script, calligraphy by Jayarava.

Non-lexical utterances, stobhas, and mantra

In researching the background to Buddhist mantra I inevitably began to read about Vedic mantra. There is a lot more research on Vedic mantra and on the whole it is more interesting than research on Buddhist mantra, so far. Reading up on the Vedic tradition has given me an appreciation of the Vedic literature which is of surpassing beauty and profundity at times. I think we Buddhists tend to write off the other Indian scriptures but that is our loss. The Vedic tradition stands in relation to Indian culture rather like the Ancient Greeks do to Europe.

If you do read up on Vedic mantras you will find that mantra originally meant one of the hymns to the gods as exemplified and recorded in the Ṛgveda. The date of this text is disputed rather vigorously and sometimes hotly, but it seems likely that it was compiled around 1500-1200 BCE, probably out of an already existing oral literature. As the verses (or ṛk) began to be used ritually two things happened. Firstly an exegetic literature began to be composed to explain how, where, and when to use the verses in the rituals; and secondly the verses themselves were reframed. For my purpose today I want to draw attention to the Sāmaveda which reframes the Ṛgvedic verses by setting them to music. Verses sung or chanted to these rhythms and tunes as called sāman.

One of the key features of sāmans is the insertion of syllables to alter the metre of the original. These syllables are called stobha. Stobha can be one or two syllables. One list of stobha is:

ā (e)re hā-u is phat as hā hṃ iṭ pnya auhovā hahas ho-i kāhvau um bhā hai hum kit up dada hā-i hup mṛ vava (e)bṛ ham hvau nam vo-I (e)rā has ihi om (Staal Vedic Mantras p.61)

Recently I was revisiting some websites about the sounds that people make during conversations - which the researchers call "non-lexical utterances" or "conversational grunts". The interest in these sounds came out of research into human-computer interfaces. Here is a list of non-lexical utterances on one site:

ai hh-aaaah iiyeah okay nuuuuu ukay uam uumm yeahh am hhh m-hm okay-hh nyaa-haao um uh uun yeahuuh neeu ao hhh-uuuh mm ooa nyeah um-hm-uh-hm uh-hn uuuh yegh nuu aoo hhn mm-hm ookay o-w umm uh-hn-uh-hn uuuuuuu yeh-yeah ohh aum hmm mm-mm oooh oa ummum uh-huh wow yei yeah eah hmmmmm mmm ooooh oh unkay uh-mm yah-yeah yo ehh hn myeah oop-ep-oop oh-eh unununu uh-uh ye yyeah achh h-nmm hn-hn nn-hn u-kay oh-kay uu uh-uhmmm yeah ah haah huh nn-nnn u-uh oh-okay uuh uhh yeah-okay ahh hh i nu u-uun oh-yeah uum uhhh yeah-yeah (Reponsive Systems Project)

The list could be supplemented from popular music (think James Brown for instance!), or for that matter from serious vocal music, which also use non-lexical syllables to pad sentences or verses to fit a metre. These non-lexical sounds function as feedback to the speaker, and are uttered in concert with the speaker in order to let them know that they are being heard and understood. A lot (but not all) of the information conveyed by these non-lexical sounds is contained in the prosodic aspects of speech - tone of voice, inflection - along with non-verbal signals such as facial expression, hand gestures, and body posture. These can indicate the attitude of the listener to what is being said, and how they feel about it.

While we cannot confirm this, it seems reasonable to surmise that stobhas were drawn from non-lexical sounds amongst Vedic speakers at the time. This further suggests that stobhas not only help a verse to conform to a metre or rhythm, but may also have served another pragmatic function when chanted in sāmans. They may have been imitating prosodic elements of speakers of the time, incorporating information about responses to the sāman within it. It may be possible for a suitably qualified person to test this idea.

It is the conclusion of some researchers into mantra, Fritz Staal being the leading light, that because mantra contain non-lexical sounds, that they are "meaningless". We would have to agree that sounds like oṃ, āḥ, and hūṃ do not have dictionary definitions, they do not refer to any "thing". However it's clear that Staal et al have been too narrowly focussed on semantics. Languistics may be focussed on words, but human communication involves very much more, and a great deal of communication may take place without any words at all. We can even make words mean the opposite of their dictionary meanin: I can say "I like your new haircut", while implying the exact opposite in an unequivocal way through the use of facial expression and vocal inflection for instance. (This is known technically as conversational implicature)

After the Ṛgvedic period mantras began to make more use of non-lexical sounds. Staal sees this as a persistence of primitive pre-linguistic sounds into the present: they are like bird song, animal noises, or the burbling of infants, and quite meaningless. They are the caveman grunts of popular imagination, retained by Indian religious leaders for ritual purposes. If we for a moment accept Staal's hypothesis his analysis of those kinds of sounds is grossly oversimplified since all three of these phenomena are far from meaningless if one knows how to listen. Worse still Staal appears to be making some unfortunate, rather "orientalist", implications about the subjects of his studies. This inelegant hypothesis is untestable, and does not open the way to further research. It certainly does not chime with the experience of mantra. Kūkai goes to the other extreme and counts every mantric syllable as being infinitely meaningful, and being the starting point for elucidating all knowledge and experience. In this he is adopting a world view which has its basis in the Avataṃsaka Sūtra. A full explanation of Kūkai idea deserves it's own essay, and goes beyond what I am suggesting here. Not that I disagree, but I am looking for intermediate steps that make sense in a contemporary context.

Stobhas used in sāmans may well have been the model for the use of non-lexical syllables in mantra although this would be difficult to prove. They do bear a resemblance to non-lexical sounds used meaningfully in conversation by contemporary English speakers (and others). But even if they did not what it suggests to me is that we can look for meaning in ways that might not be obvious, and still not have to stray into metaphysics and mysticism. It may be that no explanation in these terms can fully comprehend mantra. That is not a problem. But in attempting such an explanation I think we can shed a lot more light on this subject, and make it more accessible in the process. The "mantras are meaningless" mantra is a dead end as far as research goes, and as far as elucidating the persistence of mantra over several millennia in Indian religious contexts.



References

• The responsive systems project. [website]
  
• The divākīrtya sāmans. mAnasa-taraMgiNI [blog]
  
• Staal, F. 1989. "Vedic Mantras" in Alper, H. P. (ed) Mantra. Albany : State University of New York Press.

image: The Reading Genie - "Say ah!"