In the MAT the phrase is, in Stephen Hodge's translation:
"I declare that A is the essence of all mantras, and from it arise mantras without number; and it produces in entirety the Awareness which stills all conceptual proliferations".Previous explanations of this phrase are based on two ideas: first that unmodified consonants in the Sanskrit alphabet assume the vowel 'a'; or second, that 'a' added to any adjective or noun causes it to mean the opposite. These don't seem explain the claim that 'a' is the essence of all mantras. The syllable 'a' is not involved either phonetically or graphically in the other vowels sounds, and added to a verb usually indicates the past imperfect tense. I have put forward the theory that this idea makes more sense in an environment in which the Gāndhārī  language and Kharoṣṭhī script were used: where the character for 'a' is modified by diacritic marks to indicate other vowels.
Here I want to explore a link to the Perfection of Wisdom tradition by examining one of the phrases which make up the alphabetic acrostic of the Arapacana poem as found in the Pañcaviṃśatisāhasrikā Prajñāpāramitā Sūtra - the Perfection of Wisdom Sūtra in 25,000 Lines (hereafter the 25kPP). The first five lines go like this:
akāro mukhaḥ sarvadharmāṇāṃ ādyanutpannatvātClearly there is a pattern here. Akāro, repho, pakāro etc are the names of the syllables in Sanskrit (r being irregular). Sarvadharmāṇām is a compound of sarva + dharma in the genitive plural case - roughly 'of all dharmas'. Conze's translation into English remains the only accessible one and he translated the first phrase as: "The syllable A is a door to the insight that all dharmas are unproduced from the very beginning".
repho mukhaḥ sarvadharmāṇāṃ rajo 'pagatatvāt
pakāro mukhaḥ sarvadharmāṇāṃ paramārtha nirdeśāt
cakāro mukhaḥ sarvadharmāṇāṃ cyavanopapattyanupalabdhitvāt
nakaro mukhaḥ sarvadharmāṇāṃ nāmāpagatatvāt
Conze has not just translated the words, he has interpreted them - there is nothing to correspond to "the insight that" in the Sanskrit. The grammatical relationship suggests that the letters are indeed the 'mukhaḥ' of all dharmas, but here we need to tread carefully. Firstly my regular readers will know that dharma is a very ambiguous term that can be translated rather differently under different circumstances. I have pointed out that in many cases that dharmas (plural) should be taken to be what arises in dependence on causes (the primary focus of the Buddha's insights and teaching), and further that it is better to think of dharmas in this sense as the units of conscious experience - they are the building bricks of our subjective 'world'. I think that this definition might apply here also, but before I go into this we need to explore this word 'mukha'.
Mukha is almost a slippery as dharma. Since we know that the language of the Wisdom alphabet was originally a Prakrit rather than Classical Sanskrit we need to consult more widely than Sanskrit dictionaries in defining this word. I have consulted Monier-Williams' Sanskrit Dictionary, Edgerton's Buddhist-Hybrid Sanskrit Dictionary and the Pāli-English Dictionary (PED). Definitions largely overlap except for one specific case. The PED provides the most useful summary of the meanings:
Conze has chosen to render mukha as 'door' and the reason for this may be that in the 25kPP mukha occurs with another term which suggests that they might be synonyms: "akṣaramukham akṣarapraveśaḥ" (25kPP 21.2.08). Akṣara is 'syllable' in both cases. Praveśa can mean "entering, entrance, penetration or intrusion into". It is quite common in Pāli texts to use two synonyms like this for emphasis - although often commentators feel compelled to make hair splitting differences between the two. However 'Door' is not the most obvious translation of mukha even under these circumstances. Salomon translates it 'head' in one of his papers on the Arapacana Alphabet for instance, although I do not think this is right either.
- face, or of the face
- opening, metaphorically a means of income
- cause, ways, means, reason
- front, top, head (and hence:)
- pinnacle, best part, foremost, top most.
Let's step back for a minute and explore the context which in this case is meditation. The words of the acrostic are an aide de memoire for meditation. This is brought our quite clearly in a later passage (420 pages later in Conze's translation!). Here the text makes it clear that the reader should be meditating "on the 42 letters" . If one reads through all of the lines it becomes clear that this is a meditation on emptiness: or to be quite specific it is a meditation designed to reveal that dharmas are empty of svabhāva or independent existence. This is not different from my own approach to dhammas relying on Pāli texts. Because dharmas are the subjective aspects of experience and nothing substantial arises in the process of having an experience, nothing is defiled, nothing is beyond this, nothing ceases, there is nothing to pin a label on (these are rough translations of the first five lines of the Arapacana). That is to say the subject for contemplation is not the nature of Reality, but the nature of experience.
So the letter 'a' reminds us of the word anutpanna (non-arisen) which expands to the line akāro mukhaḥ sarvadharmāṇāṃ ādyanutpannatvāt, and the overall idea is to contemplate the notion that within experience nothing substantial or independent arises. Conze's suggestion, then, that the syllable 'a' is a door, even a door to insight, is not completely implausible. However praveśa suggests not simply an entrance, but a penetration into something - ie an insight - into the meaning of the words. The syllable 'a' certainly provides a reminder, and perhaps we could see it as providing a way into insight. Perhaps then mukha is being used in the sense of 'means' or 'opportunity'? Another possibility comes from the BHS dictionary where Edgerton suggests that another way of reading the word is 'introduction' or 'ingress'. It could be that the meditation practice is seen as having two phases - introduction to the concept, and penetration to the consequences of it.
Conze says that "all dharmas are unproduced from the very beginning", but I don't think this is quite what was intended. Let's take apart this complex compound ādyanutpannatvāt and see what it says: ādi + an + ud + panna + tva + āt. The prefix ādi means 'beginning or commencement'. An + utpanna is just the opposite of utpanna, and utpanna is ud + panna (d changes to t before p) which is 'rising up' or 'arising'. So anutpana is 'not rising up'. Now -tva is a suffix used to form abstract nouns: if god is the noun, then divinity is the abstract noun. You could also translate -tva as -ness. If a stone is hard then it exhibits hardness. And -āt is an ablative suffix - it can express the English 'from' or 'because of'. So putting things back together: anutpannatva means 'having the quality of not arising'. Adding ādi gives us Conze's "from the very beginning".
I would translate the whole phrase:
akāro mukhaḥ sarvadharmāṇāṃ ādyanutpannatvāt
the syllable a is an opening because of the primal quality of not arising of all dharmas.This is not so different to Conze. There is an ambiguity: sarvadharmāṇāṃ is a genitive plural "of all dharmas" and it could mean the 'opening of all dharmas...' or 'the primal quality of not arising of all dharmas.' Conze chose the former, but it occurs to me that the latter needs to be considered as a possibility, and works better in my opinion - I'm a beginner and Conze was a very experienced linguist and translator, but, even so.
It is interesting to note that the text has effectively become esoteric - i.e. it cannot be understood as it stands. One needs a little Sanskrit, and to have studied the text with a view to the Arapacana meditation. It does yield up it's secrets to study, but not to the casual reader. I have examined all of the published occurrences of the Arapacana. I don't have access to the many unpublished manuscripts. The manuscript from Bajaur which will no doubt provide more insights when published as it is the oldest known Arapacana. In my opinion the incorporation of a working Arapacana meditation in the 25kpp links it to the Gandhāra area - recall that no other alphabetical lists are known in ancient Indian texts.
My view is that this tradition represents a continuous line of development from early Buddhism which preserves the essential elements of the original. The crucial notions are that dharmas are units of experience, and that the important thing is to the workings of experience from the subjective pole (as opposed to trying to describe 'reality'). But the particular tradition withers and, I think, dies. Traces of the Arapacana tradition survive for hundreds of years, but are increasingly abstract. Between the Mahāvairocana Abhisaṃbodhi Tantra (ca mid 7th century) and the next major Tantric text, the Sarvatathāgata-Tatvasaṃgraha Tantra (ca late 7th - early 8th century), the whole alphabet gets paired down to just the syllable 'a'. In the 25kPP the meditation is on all of the syllables of the Gāndhārī alphabet - it is a complex task to remember the 42 (or 43 or 44) lines. And the 25kPP itself says that all of these reflections point to the same truth. So the whole thing got pared down to: akāro mukhaḥ sarvadharmāṇāṃ ādyanutpannavāt. As I have remarked elsewhere the line later became embedded in bījas and was turned into a mantra: oṃ akāro mukhaṃ sarvadharmāṇāṃ ādyanutpannatvāt āḥ hūṃ phaṭ svāhā. This form crops up in contexts which appear completely dissociated from its origins in Gandhāra.
 Note that the purpose is to still proliferations. I don't have space to link this with last week's essay on proliferation, but the connection is an interesting one.
My spelling of Gandhāra and Gāndhārī have been somewhat erratic in the past - I think I have it right in this essay and will endeavour to correct it in past essays as time permits.
 The text does indeed say 42, although most versions of the Arapacana have 43 or 44, and the one in this text has 44. It's not clear why this discrepancy exists.
Note: A complete and reliable edited Sanskrit text of the Pañcaviṃśatisāhasrikā Prajñāpāramitā Sūtra is not yet available, and access to manuscripts is out of the question for someone like me. Dutt's edition is complete but unreliable - for instance the Arapacana has two duplications of syllables. Another edition is in the process of being edited by Takayasu Kimura, but the volume which contains the Arapacana is not yet published, although the other related passages are available in Kimura (I haven't had a chance to compare them yet).
image: Seed-syllable āṃḥ - combines the syllables a, ā, aṃ, aḥ which represent the four stages of the path in the Mahāvairocana Abhisaṃbodhi Tantra, and therefore symbolises their culmination and apotheosis as embodied by Mahāvairocana.