The use of mantras being prescribed in different circumstances (in ritual, magic, yoga, or spiritual and meditative practices) are quoted, revealed or prescribed in different ways. In ritual, and especially in the course of the mandatory daily worship of a deity, but also in many other rites, such as dīkṣā, pratiṣṭhā, pavitrārohana and so forth, their utterance at various moments of the ritual performance is prescribed by Tantras and other Tantric texts, or in compilations (nibandhas) or manuals (paddhatis) of ritual. The only condition for their use is that they are uttered exactly and as prescribed, by a socially and ritually authorised person. Such has been the case from Vedic times down to the present. It is, one may say, the normal, ordinary form in which the utterance of these constituent elements of all rites is prescribed, the form most often met with. It is, however, not the most characteristically Tantric and most interesting one. More interesting, and typically Tantric, is another way, first of quoting or revealing a mantra, then of examining and selecting it so as to be sure it is well adapted to its user and to the end pursued. This method is used whenever the mantra – which is the phonic, efficacious form of a deity and is therefore not to be revealed to inapt or uninitiated persons – is to be communicated to an adept or devotee. This form of revelation or secret (ritual or non-ritual) communication of a mantra is what is called mantroddhāra, whereas the method for examining and adapting a mantra is called mantravicāra, two practices some forms of which we shall see presently.
A mantra being the phonic form of a deity, replete with all the deity’s power and effectiveness, cannot be freely revealed to just any person. It is to be revealed in a particular, secret (or at least indirect) way. Though being a divine entity, a mantra is also in practice not merely a phonic entity, or a sound, but a verbal form, made up of letters or phonemes, that is, made up of the constituent elements of the Sanskrit alphabet, the varnas. Its revelation will consist therefore in extracting the phonemes of which it is made from the ensemble of phonemes of the alphabet, the varṇasamāmnāya, which, in this context, is not a mere collection of letters, but the phonetic form of the absolute, of vāc, the Word, which, as is said in the R̥gveda, is the mother of all the gods.2
One may say in a general way that the root or fundamental mantra (mūlamantra or mūlavidyā3) of a deity to be worshipped, or the mantra given to an adept for a specific reason, to achieve a particular end, has to be selected in a particular way. The principle underlying this selection is that mantras are basic, efficacious forms or aspects of the Word, of vāc, which is the original, fundamental ever-active form of the supreme godhead, its energy or power, śakti. To be made use of, mantras, which are made of phonemes and/or words, must therefore be taken out of, extracted from, this Word totality – hence the term and practice of mantroddhāra.
The term mantroddhāra (or mantroddhāraṇa) is made up of mantra and uddhāra/uddhāraṇa, from the verbal form ud-DHR̥ (to draw out, extract or select). Mantroddhāra is thus the extraction of the mantra out of the totality of sound or, more precisely, out of the totality of the Sanskrit phonemes, the varṇasamāmnāya (the alphabetical form of vāc4). It is at the same time the selection and composition of the mantra of a particular deity, or of the mantra to be given to a particular person – the notion of extraction being, however, always essential.5 The various forms of mantroddhāra are to be found in a large number of texts, mainly in the ‘digests of mantraśāstra’, to use T. Goudriaan’s words,6 such as the Prapañcasāra (eighth century), or Lakṣmanadeśika’s twelfth century. Śāradātilaka, in manuals or compendiums of Tantric ritual such, for instance, as Mahīdhara’s Mantramahodadhi (sixteenth century), Kr̥ṣṇānanda Āgamavāgīśa’s Tantrasāra (seventeenth century), or the nineteenth-century Mantramahārṇava. There are also, as we shall see, mantrodhāra(ṇa) chapters in most Śaiva āgamas or tantras, in the Pāñcarātra saṃhitās, and, of course, in various Tantric works, all these too numerous to be quoted here.
We shall review rapidly in these pages some typical ritual processes, either simple or complex, which are all practical applications of the general organising principle of mantroddhāra.
The simpler and certainly the most frequent form of mantroddhāra consists in merely giving the syllables in the order which constitutes the mantra. Mantras being the energy aspect, the active form of deities, and being therefore powerful, their syllables are not simply enumerated for everybody to know them, but usually given in such a way that only initiates can identify them. There are, for such purposes, conventional names for the Sanskrit phonemes, which are not the same in all traditions: in practice, the guru will tell his disciple how to understand the secret uddhāra of his initiatory tradition. There are also books which list the conventional names of the Sanskrit letters:7 they can be made use of to decipher an uddhāra – but, as is well known, a mantra learned of in a book or deciphered with the aid of a book is powerless and therefore useless.
Here are a few instances of this simpler form of mantroddhāra:
First, that of the bījā HRĪṂ, of the goddess Tārikā, which I take from a Pāñcarātra text, the Lakṣmītantra (LT), Chapter 44, 7a. It is as follows: “The concentrated [sound of] Sūrya (H) and anala (R) together, joined with Viṣṇu, (Ī) with vyomeśa (Ṃ) at the end.” To take another instance, in the Netra (NT), 2.21, for the Netramantra OṂ JUṂ SAḤ, the uddhāra of the first syllable is given like this: “The first of all, that which comes after the [phonic] form of the universe, the one who kills the universe”. “The first of all”, says the commentary, “is the first phoneme, A; that which comes after the form of māyā (māya being identified with the universe8) which is Ī, is U; what kills this universe is time, expressed by the letter M.” We have thus A+U+M= OM.
A more cryptic uddhāra is that of the ‘heart-mantra’, SAUḤ given in a Śaiva Trika text, the Parātrīśikā. It runs as follows: ‘United with the fourteenth, O Fair One, associated with the end of the Master of the tithis, the third brahman, O fair-hipped Woman, is the heart of the self of Bhairava’:9 the fourteenth is the fourteenth Sanskrit phoneme, AU, the tithis are the sixteen ‘vowels’10 corresponding to Śiva, who is their master, the last of these being Ḥ ; the third brahman is to be understood as the letter SA or, more precisely S. One has thus SAUḤ – provided of course one knows the proper order of the phonemes, and also that ‘third brahman’ means SA.11
Those are uddhāras based on what we may call play of words. Though sometimes very cryptic, they are simple processes. But there are also ritual mantroddhāras, complex processes where the extraction of a mantra consists of several ritual acts to be performed by a master (guru, ācārya or deśika)12 or by a qualified, that is, an initiated adept (a sādhaka). These masters and adepts, being initiated, have already a mantra which they received when initiated and which they have mastered by the process of puraścaraṇa;13 they may however need to use a particular mantra for a specific purpose – in kāmya rituals especially – and would then need to have access to it either in a simple, ‘literary’ fashion, as shown above, or they would have to ‘extract’ it ritually, which is what we shall see now.
The general structure of the process to be used is always the same, referring also to the same basic principle. The performer, in such cases, is first to invoke and to display ritually, according to prescribed rules, the fifty phonemes of the Sanskrit alphabet,14 and pay homage to them; then he will extract, that is, select, from among this phonetic ensemble the phonemes of the mantra. Then, when this totality of the divine power or energy (śakti) of vāc in her phonematic form has been duly, ritually, made present and honoured, the adept picks out of it the elements which, when assembled, will make up the mantra. Being derived from a totality, the mantra may thus appear theoretically as a somewhat less exalted form of phonic power than the whole alphabet (though there are mantras which are deemed to embody this very totality – OṂ notably15), but it is a more efficient, and especially a more usable one – one, also, which can be adapted to particular uses or circumstances.
One might say that the mantra thus made to appear out of the totality of the alphabet by a selection of its constituent elements brings together in a more practically usable and more effective form the individually distinct phonemic entities of the alphabet. To quote Arthur Avalon, ‘In the mātr̥kās, the mantra lies scattered. Mantroddhāra is the formation of the mantra by selection of the mātr̥kās’.16 Analogously, and infinitely more authoritatively, Abhinavagupta’s disciple, the Kaśmirian Kṣemarāja (11th c.), said in his Vimarśinī on Vasugupta’s Śivasūtra, 2.3: ‘the secret essence of the mantras is none other than the blissful [Mātr̥kā], “whose nature”, as has been shown, “is that of cognition”. And this is the reason why in every āgama the presentation of the extraction of mantras is preceded by the spreading out (prastāra17) of the mātr̥kā or of the mālinī’.18 Mātr̥kā is the ‘normal’ order of the Sanskrit phonemes, from A to KṢA, mālinī is an apparently haphazard disposition of the phonemes, the first being NA and the last one PHA: nādiphāntam, as is sometimes said. Without considering these two terms here, we shall note that mātr̥kā (in the singular when applied to the whole collection of phonemes, or in the plural if these are considered separately, and in that case the word can very appropriately be translated by ‘little mother’) is used for the varṇas to underline their character as mothers (or as a Great Mother) who bring into existence the whole cosmos, and, more specifically, the mantras. As is often said in Tantric works, she is the unknown mother of all the mantras, or, to quote Kṣemarāja’s commentary on the Svacchandatantra,19 2.30–31, ‘the universal mother, unknown, of all bound souls, begetter of all mantras and tantras, going from A to KṢA’.20 She is the original ground or substrate (pūrvabhitti) of all words and objects.
Many quotations from various sources could be adduced as proof of the validity of Kṣemarāja’s formula as well as of the general frequency of the mantroddhāra practice. For example, we may look at a passage of the first chapter, on the extraction of mantras (mantroddhāraprakaraṇam), of the kriyāpāda of the Mr̥gendrāgama (Mr̥g), with its commentary by Nārāyaṇakaṇṭha. The latter, commenting on sūtra 2 of this chapter, repeats (after the āgama) that the Power of the Lord is kuṇḋalinī as the supreme word (parāvāc). In her, a first form of subtle sound, a supreme resonance (paradhvani), the nāda, appears, which then condenses as a drop (bindu) of phonic energy (described by Nārāyaṇakaṇṭha as an inner murmuring – antarsam. jalpa). From it comes the akṣara, the Imperishable, the syllable OṂ, which begets the mātr̥kā, the mother of all words, consisting of the Sanskrit alphabet from A to KṢA from which all the mantras are born. The chapter then goes on enumerating the various mantras of this Śaiva tradition, which are displayed in form of a diagram. The mantras are thus deemed to be ‘extracted’ in that they derive ontologically from the phonemes of the alphabet. No actual method for ‘extracting’ them is described, but the principle is clearly there.
This being so, and the principle being always the same, in actual practice the ritual of mantroddhāra is performed in different ways, details varying from text to text, but the overall pattern remaining always the same. The adept is always to spread out21 the mātr̥kā in front of himself, to pay homage to or worship it, then to extract from it the letters which make up the mantra.
Thus, the SvT (1.30ff.) says that:
[Is to be prepared an area of] white, red, yellow or black earth, purified by sprinkling [water on it], without any twigs, endowed with all [possible favourable] characteristics and able to bring about all that is to be wished, perfumed with pleasant-smelling ointments, adorned with flowers and beautified above by a canopy. [There,] the master, purified, smeared with sandal-and aloe-paste and made fragrant with incense, his mind calm and pleasant, facing the east or the north, with one-pointed22 and concentrated mind, is to spread out the mātr̥kā in the order from A to KṢA
The tantra goes on to say that the sixteen vowels are Bhairava and the consonants are Bhairavī: the supreme god and his energy whose conjunction engenders the universe. The officiating master is to worship (prapūjayet) the god Mātr̥kābhairava with the vowels (avargeṇa), and his consort Bhairavī with the groups of consonants (kādinā): the tantra (1.34–36) enumerates eight goddesses, the eight Mothers, the aṣtamātr̥kā,23 that are to be worshipped each with each of the eight groups of phonemes (the vargas). The pūjā once performed, the tantra adds, the master is to extract the mantra in the proper order as prescribed. Kṣmarāja explains this injunction as follows: ‘At the end of this mātr̥kāpūjā [the mantras are to be extracted] in the proper order, that is successively beginning with the āsanamantra, [then] the mūrtimantra,24 etc. This is to be done as prescribed, fixing one’s attention very carefully on the deity expressed by each mantra.’25 This last prescription refers to the rule that the mantras of deities to be worshipped are to be enunciated beginning with the mantra of the throne (the āsana) on which the deity is to be placed (or, before it, the root-mantra, the mūlamantra), then the mantra of the form or body of the deity (mūrtimantra), then the other ones. These others are the mantras of the deity’s ‘limbs’ (aṅga) and ‘ancillary limbs’ (upāṅga) and, finally, those of other deities.26
The NT, a work from the same tradition as the SvT though probably slightly later, explains in its second chapter (2.17–21) that on a levelled, pure area, anointed with sandal and aloe and perfumed by different substances, the master, anointed, perfumed and bejewelled, with peaceful mind, must first draw the outline of a lotus whose eight petals point towards the four cardinal and the four intermediate directions of space. Having placed (nyasya) in the centre of the lotus the mantra OṂ, he writes or draws out (likhet) each of the eight groups (varga) of letters of the mātr̥kā on each of the eight petals, starting from the one pointing east. Then, ‘having worshipped with utmost devotion (pūjayet parayā bhaktyā) the goddess mother of all mantras, and having offered her flowers, incense, etc., he is to extract the mantradevatā’ the phonemes of which are then enumerated in a covert fashion.
Let us now look at a work from a different Tantric (Vaiṣṇava) tradition, the Jayākhyasaṃhitā (JayS) of the Pañcarātra, whose sixth chapter describes the extraction of the main mantras (mukhyamantroddhāra) of that system. The description is very detailed. The ritual practice is more complex than the one we have just seen; it enters into more particulars, but the organising principle is the same. An area is first to be prepared – that is, flattened, cleaned, smeared with cow dung, sprinkled with the five products of the cow (pañcagavya27), anointed with sandal, perfumed with incense, etc. Then, having paid homage to this area with white flowers and a mantra, the yogin28 is to draw with clay a round or square throne for the mātr̥kā (mātr̥kāpīṭha) on which he spreads out (prastāra) the phonemes of the alphabet. While doing so he is to realise that the mātr̥kā is nothing else than the multiple aspects of the supreme deity as it manifests the cosmos. The phonemes he has written out are therefore no mere letters but forms of the divine energy. He is then to place by nyāsa29 the mātr̥kā on his body which is thus pervaded by her power as well as inhabited by the fifty mātr̥kās as deities. Now he is to draw a large eight-pronged figure (mahācakra) on whose centre he writes OṂ, and in the eight directions the eight vargas. There follows a ritual worship (pūjā), to be done with devotion (bhaktyā30) and with the usual ritual offerings, of the mātr̥kā written on the consecrated area and of the one placed on his body, a worship during which the officiant is to invoke by their name31 and worship separately all the fifty phonemes, ‘visualizing them as shining and blazing like the midday sun’. ‘These sounds (śabda)’, the saṃhitā (6.58b–60a) adds, ‘are fragments (aṃśā), shining aspects (bhāsvaravigrahā) of the Lord. Increased by the power of the Lord, they are the cause of all the mantras, whose birth (utpatti) they bring about by their mutual conjunctions. In this moving and unmoving world, there is nothing that is not brought about by them.’32 Then, the main mantras, beginning with the mūlamantra of Viṣṇu followed by the mūrtimantra, are to be extracted, these mantras being, of course, not spelled out directly: their constituent phonemes are designated by their names and their place in the varṇasamāmnāya order (śl. 61–69). This whole section (śl. 5 to 69 of the chapter), which I have very much summarized,33 is interesting in that, while following the general organizing ritual pattern previously outlined, it underlines the meaning of the process, namely the passage, through a process of progressive condensation and precision, from the undifferenciated energy of the deity, first to a more concentrated but still diffuse aspect, that of the mātr̥kā, then finally to the discrete form of the mantras. It also underlines the necessary identification of the Tantric performing adept with the power he both manipulates and worships.
A different process of uddhāra is shown in another Pāñcarātra text, the Sanatkumārasaṃhitā (Indrarātra, 2.55ff.). There, the mantra Oṃ namo nārāyaṇāya is given, but its eight syllables are to be placed in eight of the sixty-four compartments of a diagram where fifty-one letters34 and twice the mantra OṂ are displayed, the extraction of this mantra being followed by its resorption, together with that of all mantras, into the bosom of Viṣṇu.
A passage of the LT (chapter 24) is also worth quoting, not so much for its ritual details as for the way in which the uddhāra is presented which, putting it in its divine-cosmic context, underscores remarkably the meaning of the process. Here is, briefly, the content of the first thirty-two stanzas of the chapter:35 the supreme brahman, the absolute I, the omnipresent supreme soul, Lakṣmī-Nārāyaṇa, takes on, so as to help all living beings, the nature of the mātr̥kā made of śabdabrahman, and more specifically the nature of mantras. The master is to invoke the Supreme as the saviour mantra (tārakamantra), which here is OṂ. He is to take the three phonemes of OṂ, A, U, M, conjoin them, adding (‘beautifying them with’) bindu and then with the phonic resonance, the nāda, which dissolves eventually in the origin of sound. He is now to meditate on the Absolute as made up of these three letters which are identified with Aniruddha, Pradyumna and Saṃkarṣaṇa – the Vyūhas of Viṣṇu. Meditating this mantra, identifying with it, the adept unites with the eternal brahman and with the energies that emanate from it and which sustain and animate the universe. Mantra and the universe are then identified. Later on, the master will place ritually (by nyāsa) the mantra with its twelve aṅgas and upāṅgas on his body, which identifies him with the deity as master of the cosmos, then, continuing, he fuses mentally with the divine transcendent formless Absolute whence however he will come back to an ordinary state of consciousness so as to be able to transmit the tāramantra to his disciple. We have here therefore not merely the extraction of a mantra, but a fusion of the master with the Absolute which enables him to transmit this reality to his disciple. This explains the complexity and intensity of the meditative activity of the master who is both to identify with a mantra and then to transfer its power.
A number of other texts prescribing or describing different forms of mantro-ddhāra could still be cited: the Kiraṇāgama (Kir), 12, for instance, or the Rauravāgama (RAU), kriyāpāda 2b–4, or the Pūrvakāmika, 2, or the Īśanaśivaguru-devapaddhati (Īśgp) II, kriyāpāda 6, etc. But this is not necessary since many texts are not very explicit, therefore not worth quoting, while in other texts we would find the same general method, or very similar ones, for extracting mantras.36
As regards the diagrammatic display of phonemes for the mantroddhāra, a practice found in some texts, notably in some Śaiva Kaula traditions, is worth mentioning. The diagrams used in such cases are called prastāra or gahvara, that is, display or cavern. The letters to be chosen are in effect spread out (pra√STR̥) in these diagrams, but this usually in such a way that only initiates can find them: they can therefore be considered as being hidden in the diagram as it were, as in a cave or a cavern (guha or gahvara).
We find such a prastāra, for instance, in chapter 1.83–85 of the Vāmakeśvarīmata (VM), a tantra of the Tripurā (or Śrīvidyā) tradition, in a passage which gives the uddhāra of the vidyā of the goddess Karaśuddhikarī, AIṂ KLĪṂ SAUḤ. The tantra quotes the letters of the vidyā by referring to their place in a prastāra of a triangular shape divided into forty-nine triangular compartments, each with one of the forty-nine Sanskrit phonemes inscribed in it in their order from A to HA (it is printed on p. 45 of the Kashmir Series edition of that text). The AIṂ, in that case, is described in the tantra as ‘the bīja which is between E and O, the one which is alone and related to the word’: the phoneme AI is in effect placed between E and O on the first line of the prastāra, the commentary explaining that ‘the one which is alone’ is bindu. The same method is used for the seven other letters of the vidyā. Jayaratha, in his comment on this passage, quotes two texts, the Nityākulatantra and the Rasamahodadhi, which describe the same sort of prastāra. Those, as well as the gahvaras, can be of varied shapes with phonemes disposed in a variety of orders. The phonemes in the diagrams, too, can be those of the ‘normal’ alphabet, the mātr̥kā (or the śabdaraśi which is another name for the same alphabetical order when it is considered as the phonemic form of Bhairava), or that of the mālinī.37 The prastāra-gahvara system is likely to be found in different Tantric traditions since all insist on the secret nature of their teaching.
To be made use of, mantras are not only to be conferred on somebody or to be constituted by being extracted from the Sanskrit alphabet. It is also necessary for mantras to be adapted to the person to whom they are given and to the use they are to be put. This adaptation is checked through a process of examination or investigation (vicāra or parīkṣa) of the mantra. The idea on which such an adaptation is based is that for the mantra to be effectively used there should exist between it and its user a specific link or relationship: this is expressed by the composite term mantrāṃśa, aṃśa being taken in the sense of share or portion, the user being considered as a portion of, or as participating in, the mantra, as having a particular affinity with it. This commonality between the mantra and the mantrin permits the latter to use fully and effectively the former. This control of mantrin–mantra commonality is effectuated by diverse divinatory methods which vary according to traditions or circumstances, their diversity corresponding also to the wish to be able to use different methods in case the one first resorted to proves unsatisfactory, or to the possibility of chosing the method one prefers.38
Here too we shall limit ourselves to a few instances, quoting a few Śaiva and Vaisṇava texts, so as to show the general structure or the spirit of these aṃśa checking processes, rather than describing many particular practices. We shall first refer to such texts as the SvT and NT, adding a few other references. On the Vaiṣṇava side, instances will be taken from the LT and the Kramadīpikā.39 These processes are in fact to be found in all ritual manuals, ancient or modern, where mantric practices are expounded, and they are always very similar.
One of the best Śaiva description of the amśa practice is to be found in the eighth chapter of the SvT.40 It expounds the means with which to ascertain the participation, relationship or affinity (aṃśa) of an adept with a particular deity or supernatural entity and therefore with the mantra of that deity which that adept will be given by his master. Six different cases are enumerated.
There are, first, two simpler ones, where the disciple spontaneously feels devotion for a particular deity. This is called bhāvāṃśa if it is due to his natural (sahajam) feeling or emotion (bhāva) for a deity on which his mind dwells constantly (anusmaraṇa). If this is due, in a more or less similar fashion, to his innate disposition or impulse (svabhāva), if for instance he feels constantly impelled to worship a liṅga,41 the case will be that of the svabhāvāṃśa (śl. 2 and 12).42 If there is no such natural affinity, the disciple and his master will have to resort to divinatory means. The most usual one is for the disciple, blindfolded, to throw a flower (puṣpa) onto a maṇḋala on which are drawn or symbolically placed Bhairava or other forms of Śiva and of his retinue of deities: the mantra chosen will be that of the deity on which the flower falls. This is the puṣpapātāṃśa, it is mentioned in śl. 13, chapter 8, of the SvT.43 ‘This mantra’, the tantra adds, ‘will be effective if it is worshipped according to the prescriptions of the śāstras.’ The throwing of a flower on a ritual diagram is a fairly common ritual practice. It is described in several texts, ancient or modern. It is also used in other contexts, for instance in initiation, to decide the name of an initiand (see SP, vol. 3, p. 103).
It may happen that the above methods for choosing a mantra (or those we shall see later) do not satisfy the adept, who wishes to receive a particular mantra. He may then try to obtain the desired aṃśa by offerings of human flesh in the sacrificial fire, offerings to which the tantra (14b–15a) alludes indirectly, but which the commentary names clearly: naramāṃsa. Kṣemarāja uses also (p. 9) another, more frequent, conventional term, vīradravya, the heroic [ritual] ingredient,44 while explaining how this offering is to be made by the sādhaka or the ācārya.
If the preceeding methods to secure the aṃśa remain unsuccessful, the disciple can still obtain the desired aṃśa by worshipping the mantra and making oblations in the fire (yāgaṃ purā kr̥tvā agnau homaṃ kārayet – 16a), this having been done according to the prescribed rules, the disciple ‘being united, thanks to a full oblation, with the plane of the Eternal (pūrṇāhutiprayogena yojayec cchāśvate pade), setting his heart upon that plane, will succeed in what he wishes’ (16b–17a).
Now remain to be seen the more current ways for the adept to get or determine the mantrāṃśaka that are to be found in a number of Tantric texts as well as in ancient or modern manuals.
The SvT, chapter eight (śl. 20ff.) and the NT, mention only one method, consisting of writing on two lines, one under the other, the mantra to be checked and the name of the adept who is to receive it, then to count in the traditional Indian way, on the phalanxes of the right hand, the number of letters in the normal order of the Sanskrit alphabet (never in reverse) which separate the initial letter of the name of the disciple from the first one of the mantra. Four aṃśas may thus appear: siddha (which one may translate as ‘obtained’, or ‘effective’), sādhya (‘attainable’), susiddha (‘very efficacious’ or ‘completely atttained’), and ari, śatru, or ripu (‘hostile’, ‘enemy’), according to whether the counting stops on the three phalanxes of the ring-finger or on one of the four that follow, etc. Only those stopping on the first and third are favourable (siddha and susiddha) and allow the mantra to bear fruit – they are even (ibid. śl. 24b) bhuktimuktiphalaprada: their fruit is liberation and rewards. In the two other cases (second phalanx, or the fourth – on the little finger), the mantra is unfavourable or dangerous and one must strictly abstain from giving it to the adept or initiand. Kṣemarāja, whose commentary on śl. 20–24 explains how to proceed, adds (perhaps needlessly) that this aṃśakaparīkṣa does not apply to monosyllabic mantras nor to mālāmantras.45 This being done, the main deity (Bhairava Kapālīśa in this case) is to be worshipped with his retinue.
There are still other methods both to make sure that the mantra is perfectly adapted to its potential user and to distinguish more precisely (notably by determining sub-categories of the four aṃśas) between the effects of the mantra and the methods for using it. As a result, the adept is sure to get a mantra which will be perfectly adapted to him and which will respond to his needs, whilst he will also know beforehand what effects he can expect from its use, and on what conditions.
One will thus check if the adept and the mantra belong to the same ‘family’ (kula) by comparing the first letter of the mantra with the initial of the mantrin’s name, the phonemes of the alphabet being, for this purpose, displayed on a plate with four columns (a kulākulacakra) considered as being mutually either compatible or incompatible: for the kula to exist, the two initials must be in two compatible columns.
One can also, using a square or a round diagram, compare the rāśi, the lunar mansion or zodiacal sign conventionally attributed to the adept and to the mantra: according to the rāśi where they will be, one will know if the mantra would or would not be favourable, or what result one may expect from its use. The same process is made use of to compare the nakṣatra, the constellation or asterism, of the adept with the nakṣatra assigned to the first letter of the mantra, these being in two different sections of the diagram: some coincidences are deemed favourable (siddha, susiddha), others are bad (sādhya or ari).46
Another predictive method consists in using a diagram called akaḋamacakra because the four phonemes A, KA, ḌA and MA are placed in its upper central division. The letters of the alphabet are placed in it by groups of four forming twelve radiating columns in a circle or twelve lines along the sides of a rectangle.47 According to the distance between the first phoneme of the mantra and the initial letter of the adept’s name, the mantra will be considered as siddha, sādhya susiddha, or ari. There is also the r̥ṇidhānicakra, the diagram of the debtor and the creditor, with which – by a rather complicated calculation – one will see if the mantra is favourable (for which it must be the adept’s debtor), or else holds a debt claim against him and is thus unfavourable.48
The method most frequently used, I believe, consists in tracing an oriented square surface, divided into four quarters, each one being divided also into quarters: there are thus sixteen small squares in groups of four where the phonemes of the Sanskrit alphabet are written either in the normal alphabetical order (the akṣarasamāmnāya) or in a different order corresponding to particular symbolical values given them (in this case, it is often called akathahacakra, from the names of the letters placed in the first little square (on the left). The four squares divided into four are assigned, going clockwise, to each of the aṃśa: siddha on top left (northeast), sādhya on top right (south-east), susiddha down left (north-west), and ari down right (south-west). The aṃśa is determined, not by the distance separating the square containing the first letter of the name of the disciple from the square containing the first letter of the mantra, but by their juxtaposition, and it will be expressed by associating the categories assigned to the two squares. The aṃśa would thus be, say, siddha-ari or sādhya-sādhya, or siddha-susiddha, etc., therefore dangerous, useless or favourable, etc.: there are sixteen possibilities instead of four only.49 The validity of a mantra for an adept can also be checked using the kūrmacakra, a diagram in the shape of a tortoise (kūrma) which is to be drawn by the adept who is then to write the letters of the Sanskrit alphabet, by groups, in nine different sections of the body of the tortoise (the vowels in the central part, the consonants grouped by vargas, on top, bottom, sides, ‘hands’ and paws. According to the sections where the initial letters of the mantra and of the adept’s name will be, and by some calculations, the mantra will be found to be siddha, susiddha, sādhya or ari.50
One sees from the above that all possible guaranties are sought for the mantra to be perfectly adapted to its user. The need for such an adaptation is self-evident in the initiation – dīkṣā or abhiṣeka – of the sādhaka, since he is a bubhukṣu, a seeker of bhukti, of supernatural accomplishments, to be attained through the mantra which has been given him and which is to be mastered – it is a sādhyamantra – this mastery being the first thing he is to do, through the ritual of mantrasādhana,51 after been initiated. A mantra adapted to its user is also a necessity in the larger context of desire-oriented (kāmya) rites. These kāmyakarmāṇi, as they are called, are rites practised so as to obtain particular results, notably – but not necessarily – through the cult of particular deities (especially of power-yielding goddesses such as the Yoginīs52) who will bestow on their worshipper boons of different sorts.53 Among those cults are the so-called ‘six [magical] acts’, ṣaṭkarmāṇi, all of which are performed with mantras54 (there are also the, less frequent because condemned, practices of sorcery, abhicāra). But the field of kāmyakarmāṇi is infinitely larger than magical actions: it includes everything in religious life that is done to satisfy a human wish, be it for one’s mundane satisfaction or for one’s salvation – which is more than anything else something devoutly to be wished.
It is to be noted, too, that however important it may be in some contexts, mantravicāra or mantraparīkṣa is not always necessary. It is even sometimes prohibited. As already mentioned, monosyllabic mantras55 and the longer ones, the so-called mālāmantras, cannot be examined in such a way. There are also mantras which are not submitted to aṃśavicāra or to diagrammatic parīkṣa because of their high status: these mantras act directly everywhere and without limitation, such as the divine śakti whose phonic aspect they are. This is the case of the great Śaiva mantras such as the para-or hr̥daya-mantra SAUḤ, or of HAUṂ, the prāsādamantra, which many texts consider as the mūlamantra of Śiva; or of the mūlamantra of the goddess, or of that of Viṣṇu, for Vaiṣṇava traditions, and so forth.56 Such mantras are said to be common to all, universal (sāmānya or sādhāraṇa). There is, finally, what appears to be, as already mentioned, the most frequent case, that is, the cases where the uddhāra of a mantra is given without any ritual practice being mentioned, its constituent letters being simply given in a more or less covert manner or even overtly quoted. There are also, mainly, all the innumerable cases where mantras are to be uttered during a ritual, where, evidently, the performer cannot examine their adequacy before using them, which does not prevent them from being perfectly effective: they effect what they say, or make present the deities they ‘express’ (being their vācaka). There are, as is well known, ritual situations or processes which do not imply any outward action, where merely mantras are uttered (for instance the ātma- or āntara-pūja which, in Tantric practice, always precedes the visible, material, ‘external’, bāhya, worship). There is no rite without mantras, and these are effective without any previous ritual fuss being necessary.
To conclude, we can, I believe, confirm the principle mentioned at the beginning of this study, namely that the mantroddhāra or mantravicāra process is linked to the nature of mantras as it is described in Tantric texts, or, more generally, to the Tantric conception of the power of the word. Mantras being the efficient and utilisable form of this word-power, they must be adapted both to the aim pursued and to their user. This user, being himself animated by a power that is fundamentally of the same nature as that of the mantra, can put the mantra in effective use only by fusing his own energy with that of the mantra, which is posssible only if there is adequacy between them.
It appears therefore in all that has been seen here (as in what is studied in other articles in this volume) that one cannot examine particular aspects of mantras without touching directly or indirectly upon the problem of their nature, which is ambiguous. Though they are forms of the supreme Word, vāc, though they are deities, devatās, in their own right, though existing therefore on a transcendent plane, mantras exist also empirically when – as (in principle at least) utterable phonetic units – they are used by human practitioners. They are both divine in essence and forms of human speech. They can, or must, therefore, when being made use of on the human plane, be submitted to ritual control procedures. These technical mantric rites meant to ensure their adequacy are inseparable from the sacred or magical Word, vāc, while being the external social acts that, by guaranteeing the proper use in context of the forms of vāc that mantras are, ground and ensure their authority on the social plane. The efficacy of mantras, whether one believes it to be real or symbolic, cannot be dissociated from its ritual, social context.