The six acts

The grid everyone reaches for

In 1588, at Varanasi, a scholar named Mahīdhara finished a book of some three thousand eight hundred verses and called it the Mantramahodadhi, the great ocean of mantras. He wrote a commentary on his own book as well and called that the Naukā, the boat. It is the kind of joke a confident man makes about his own work.

Then he had to decide what went where.

He arranged the ocean deity by deity. Gaṇeśa opens it. Kālī and Tārā take four chapters between them. Then the Yakṣiṇīs, then Bālā-Tripurā, then Annapūrṇā, then Tripurasundarī, then Hanumant, Viṣṇu, the nine planets, Śiva as conqueror of death, Kubera, Gaṅgā, Kārtavīrya Arjuna, Kālarātri and the fierce goddesses. Then the yantras, the ritual diagrams. Then daily worship. Then Kṛṣṇa. Then the generalities of mantravāda, the doctrine of mantras.

The six acts went last. Twenty-fifth of twenty-five.

Goudriaan and Gupta record the same shape in the Mantraratnākara of Yadunātha Cakravartin, a thousand verses in ten waves: Gaṇeśa, Dhūmāvatī, Bhairava, Tripurā, Kālī, Tārā, and the six acts at wave ten of ten. Two of the great mantric digests, and in both of them the famous classification of magical ritual sits where an appendix sits.

The classification is the ṣaṭkarma, the six acts: appeasement, subjugation, immobilization, enmity, eradication, liquidation. It is the grid everybody reaches for when this kind of material comes up. Ask how Indian magical ritual is organised and the six are the answer that arrives. They are old, they are widespread, and a great deal of the Sanskrit mantric literature does work with them.

Lay them over the Śābara corpus and they do not describe it.

That is the finding of this chapter, and it is a negative one, which is why it has to be built slowly. The grid does not fail against the vernacular material because the vernacular material is disorderly. It fails because the two bodies of text are sorted on different axes. The six acts classify by what is being done to a third party. The Śābara material classifies by whose warrant is being claimed. There is no angle at which those two grids meet.

Nobody has said so before, and the reason is part of the finding. Nobody has laid them side by side.

Say plainly, before any of it, what this chapter will and will not do. It names the six acts and describes how none of them is performed. Where a source gives a procedure, this chapter says that the procedure is given and stops there. No mantra appears below, no seed syllable, no rite, no material, no time of day, no repetition count, no assignment of a deity to a purpose, no correspondence table, no substitution. That is a real cost and it is paid deliberately. A thinner chapter is the correct outcome.

The six, and the one that is not like the others

Gudrun Bühnemann, writing on the six rites of magic, states that although they are frequently mentioned in Tantric literature, the names used for them vary. Her set is the Mantramahodadhi’s, and her one-line definitions are conceptual rather than procedural, which is why they can be repeated here.

Śānti, appeasement, is the curing of disease. Vaśya or vaśīkaraṇa, subjugation, brings a person under the practitioner’s control; stambhana, immobilization, stops a person’s activity; vidveṣa, enmity, creates dissension between two people attached to each other. Uccāṭana, eradication, deprives a person of his place. Māraṇa, liquidation: taking life.

P. V. Kane, writing in 1962 in the History of Dharmaśāstra, glosses them in substantially the same terms and calls them the six cruel or terrible magic acts. We will come back to his adjective.

One member of the six is benevolent. Five are destructive. Bühnemann says so in exactly those terms, and the ratio is the first thing anyone notices about the scheme.

In its developed form the six acts are a grid rather than a list. Mahīdhara states that a practitioner must know nineteen items before performing a rite; he enumerates them at Mantramahodadhi 25.4 to 5 and expounds them one by one from 25.6 to 65ab. Six acts, nineteen variables, every cell filled. Bühnemann renders eight of those rows as tables, and translates the whole of Mantramahodadhi 25.1 to 102.

This chapter gives no cell of that grid, and no line of that translation. The architecture is the argument here. The values are the payload, and the payload stays in the book.

Nobody dates it

Here is the odd thing about a classification this famous. No source consulted for this chapter says when it was fixed.

Bühnemann gives no origin date. Her operative anchor is that the definitions are already in place in the Śāradātilaka of Lakṣmaṇadeśika, which she dates to the tenth or eleventh century, at 23.121 to 125, and that the Mantramahodadhi of 1588 is a late and unusually orderly restatement. Kane and Goudriaan both cite the same passage as 23.122 onward. One verse of difference, unresolved here, because the Sanskrit text of the Śāradātilaka was not obtained and every reference to it in this chapter is at second hand.

Goudriaan is more explicit. In Māyā Divine and Human he says his four-type classification of the enumerations is not meant to establish a historical sequence, and he offers it, in his phrase, for the sake of showing the floating character of the tradition. He declines the developmental question outright.

So the tenth-to-eleventh-century compendium is a terminus ante quem. It is not an origin. That absence is itself a finding and it should be printed as one: the standard grid for Indian magical practice has no datable birth in the scholarship that uses it.

There is one place the answer might be sitting. H.-G. Türstig published a study called “The Indian Sorcery Called Abhicāra” in the Wiener Zeitschrift für die Kunde Südasiens in 1985. It is the only monographic treatment of abhicāra in the literature, Bühnemann lists it in her own further reading, and it was not obtained for this chapter. If the dating argument exists anywhere, it is most likely there. This chapter cannot say that it does. It can only say that it did not look.

What can be said about the prehistory is structural. Goudriaan’s position is that the older enumerations are pre-systematic, that they name abhicāra as one item standing beside the others rather than as a category containing them, and that they reduce to the classical threefold aim of all ritual action: śānti, pacification; puṣṭi, welfare and acquisition; abhicāra, harming others. On that reading the six re-cut the three. Śānti survives as member one. Puṣṭi drops out of the Hindu set. And abhicāra is opened out into five.

The compendia cannot agree what the six are

Goudriaan’s survey of how the six get classified is the only systematic comparison of the enumerations that exists, and what it documents is a canon that cannot hold its own membership.

At least eight distinct orderings are attested, and the inventory is his. It runs from the classic order, which Śāradātilaka, the Kāmaratna, the Ṣaṭkarmadīpikā, Narasiṃha’s Tantrasāra and the Mantramahodadhi share and which, he notes, is very nearly the only company of sources that actually uses the phrase ṣaṭ karmāṇi at all, out to a whole class of lists that drops śānti entirely, so that the set reads as wholly coercive, and admits mohana, delusion, or ākarṣaṇa, attraction, in its place. The Bagalāmukhīrahasya labels a list ṣaṭkarman in the preceding verse and then prints ten items.

Kane records that the Guhyasamāja knows the six and substitutes ākarṣaṇa for vidveṣaṇa. In the Jaina Bhairavapadmāvatīkalpa, māraṇa is regularly replaced by niṣedha, the forbidden act, and Goudriaan attributes the substitution to the weight of ahiṃsā, non-injury. That is the sharpest evidence in the whole file that the taxonomy carries an ethics inside it. A tradition that cannot accept the sixth member does not drop it. It renames it.

And then there is the witness that settles the question of whether the tradition knew. Śaktisaṃgamatantra I.38.103 to 105 sets out three different opinions on what the six are, in three consecutive verses, attributed to three named authorities. The text is not confused. The text is reporting a disagreement.

Even the modern scholarship reproduces the instability. Goudriaan makes the classic order standard in 1978; the 1981 book he wrote with Sanjukta Gupta gives the most common enumeration with members four and five transposed. Both readings come from scans and should be checked against the printed books before the discrepancy is pressed. The same scholar, in two books, moves two members of the six.

A canon that cannot hold its own membership is not a description of practice. It is a scholastic settlement, repeatedly resettled.

A minority principle in its own literature

Before this chapter claims the grid fails against the vernacular, it has to record how far it already fails inside the Sanskrit.

Goudriaan and Gupta call the Uḍḍīśatantra a difficult case for philologists and describe a cluster rather than a text: around eighty manuscripts carrying the name, many popular editions, not one of them critical. Within that cluster the Uḍḍāmareśvara Tantra uses ṣaṭkarman terminology frequently and its divisions are, they say, not based on it. Its sibling, the Uḍḍīśatantra proper, is divided mainly by that terminology and shows traces of systematisation. One recension of a named text-group organises by the six acts and the recension beside it explicitly does not. The Vīrabhadratantra, in the same cluster, is a mass of mantras with no apparent principle of arrangement.

The Kāmaratna is the clearest case. It opens with Śiva, gives a one-verse introduction, treats the general rules of the six acts briefly, gives a chapter on vaśīkaraṇa, and then abandons the frame. The bulk of the book, Goudriaan and Gupta report, is kautuka, curiosities, with aphrodisiac and medical material, cures of disease and infertility, and assistance in childbirth. Material sorted by ailment. At least sixteen popular editions of it were in print.

And the Vārāhī Tantra, listing the seven marks of an Āgama, includes execution of the six acts as one of the seven. The scheme is canonised there as a topic a Tantra ought to contain. It is not canonised as the shape a Tantra ought to have.

Where the grid does not reach

Now the corpus this treatise is about.

Goudriaan and Gupta treat the Śābara or Śavara Tantras, the texts that carry the name of barbarian speech, immediately after the Uḍḍīśa cluster and explicitly as a group of their own. Chapter 3 built the corpus out of that single paragraph. Their account is short. Not many texts are known to belong to the group. The contents are usually restricted to enumerations of incantations, often in Sanskrit alternating with Hindī. The titles are confused and may follow one another inside a single manuscript. Their two main subjects are mastery of the mantra of Śabara, a minor deity of magical potency, and mantras of a variety of other gods, Yoginīs, Kṣetrapāla, Gaṇeśa, Kālī, for magical purposes. And the Śābaracintāmaṇi, ascribed to Ādinātha, carries vernacular versions of mantras said to have been proclaimed by his five pupils in the languages of Gauḍa, Kerala, Karṇāṭa, Andhra and Gurjara.

Read that as evidence and the organising axes are visible. The corpus is arranged by deity and by lineage and language of revelation. Whose mantra it is, from whom it came, and in whose tongue it was given.

The six acts are never mentioned in connection with it.

That silence is not an argument from nothing, because the same two authors, in the same run of neighbouring entries, use the phrase “describes the ritual of the Six Acts” for the Bhūtaḍāmara, the Uḍḍāmara, the Dattātreyatantra and the Kāmaratna. The Śābara paragraph is the one paragraph in that stretch which does not say it.

The chronology points the same way. Chapter 4 established that śābara-vidyā is already in the record in Uddyotanasūri’s Prakrit Kuvalayamālā, dated to 779 CE, roughly two centuries before the earliest fixed enumeration of the six that anyone can point to. And chapter 5 set out the rival filing, in which a Śābaratantra is classed among the Gāruḍa scriptures, the stream that deals with poison and its counteraction, which is a taxonomy of currents rather than of acts. That classification is reached through a secondary route and is used here as corroboration only.

Then there is a gap in the scholarship, and it is worth naming. No Tier 0 to 2 scholarship on the popular Śābara compilations exists at all, meaning nothing at the level of a primary text, a classical commentary or a named modern scholar. Searches for academic treatment of the modern Hindī corpus return astrologer pages, vaśīkaraṇa services and content farms. G. W. Briggs, whose 1938 study of the Kānphaṭā yogis would be the obvious place, does not use the word once. The classificatory literature and the vernacular corpus have not been brought into contact by anyone.

Which means the comparison has to be made from the ground up, and this treatise has the count. Chapters 10 through 13 worked through a popular compilation and found a register of appeal and truth-assertion. Chapter 10 also read the tables of contents of eight printed Śābara compilations and classified the entry titles of the largest of them, and it arrives at the architecture Goudriaan and Gupta describe by a route that has nothing to do with them. These books organise by named purpose and by affliction, and the commonest single kind of entry title is the name of an ailment or a trouble to be got rid of. One twelve-part series organises by contributor instead, which is the authority axis again. And ṣaṭkarma vocabulary does appear, on a minority of entry titles, as adjectives. As labels attached to individual items. Never as a section division, and never as the frame the book is built on. The counts are in chapter 10.

The words are there. The grid is not.

Two axes that do not meet

State the reason plainly, because it is structural and not accidental.

The six acts classify by intended effect on a third party. Cure, control, arrest, estrange, displace, kill. Every member of the set answers the question: what is being done to whom.

A register of appeal and truth-assertion classifies by the speaker’s relation to an authority. Chapter 11 set that out: who is being invoked, on whose warrant, and what is being asserted as true. The Śābaracintāmaṇi’s five pupils and five regional tongues are exactly that architecture at the level of a whole book. The mantra’s warrant is whose it is and in whose language it was given.

These are orthogonal axes. A corpus built on the second cannot be sorted onto the first without discarding the feature that organises it.

That is why the grid does not fit. The fit was never poor. The fit was never attempted, by anyone, because the question was never asked.

The tradition’s own objection

The ethics of the six is not a modern apology bolted on. It is in the sources, it is older than the compendia, and it is load-bearing.

Manusmṛti 11.63 places abhicāra and mūlakarman among the upapātakas, the minor sins. Medhātithi, commenting, glosses abhicāra as killing one’s enemies by curses or incantations and mūlakarman as using incantations to gain control over persons. In the vocabulary of the later scheme those are māraṇa and vaśīkaraṇa, members six and two. Manusmṛti 9.290 then makes them punishable by the king with a fine of two hundred, and the commentarial tradition notes that where such a rite causes death the penalty rises to that for homicide.

Centuries before any compendium arranged them into a set of six, the two acts that give the scheme its coercive character were a named sin and a named offence with a fixed fine. The taxonomy did not create the moral problem. It inherited one that had already been adjudicated, and organised it.

The compendia then hedge.

They classify the six under kāmya, optional and desire-motivated ritual, which is the lowest of the three tiers of their own ritual taxonomy, below the obligatory and the occasional, and which carries a doctrinal ceiling: the fruit of such a rite is the fruit of that rite and nothing beyond it. They append authorial disclaimers. Mahīdhara declares at Mantramahodadhi 25.72 to 76ab that he has set the desire-oriented rites out only for the sake of persons attracted to sense pleasures, and on the authority of earlier teachers, and then extols desireless worship and the highest knowledge from 25.76cd to 81, and returns to the same advice at 25.97 to 101. Narasiṃha’s Tantrasāra makes the same move. They build in an expiation: at Mantramahodadhi 25.49cd to 55ab, a blessing obtained inside the rite itself, by a means the verses specify and this chapter does not, is said to destroy the sin that accrues from the performance of destructive magic. The text asserts that sin accrues, and then neutralises it without leaving the page. And the Purāṇas, where they name the group as a whole, call it kṣudra, petty, vile.

Kane’s own vocabulary is the reception datum. He calls the acts cruel and terrible. He observes that even the comparatively sober Śāradātilaka mentions them, defines them, assigns presiding goddesses and lays down periods and seasons, and this chapter records that it does so and withholds every assignment. And he registers astonishment that a work attributed to Śaṅkarācārya, the Prapañcasāra, should describe at length a mantra for accomplishing the six cruel acts at 23.5. The most authoritative modern Indian treatment of the normative literature reads the six acts as a scandal inside the tradition.

Goudriaan’s rider has to be printed alongside all of it, because it is the honest end of the paragraph. Pious utterances of this kind, he observes, did not stop the authors from setting out the magical science at length.

The disclaimer is at the end of the chapter. The chapter is still there.

The bridge to the ledger, and how thin it is

Chapter 20 found that no documented harm attaches to the Śābara corpus by name, and that what is documented attaches to a ritual-specialist role and to an accusation. There is one place in the scholarship where the classical taxonomy is joined to that modern world, and it is a single paragraph.

David Gordon White, introducing the volume Bühnemann’s chapter sits in, argues that with the disappearance of the Tantric polity the rites of subjugation, immobilization and annihilation, which he glosses in the running text as the six rites of magic with a reference to Bühnemann, passed to individuals working for their own profit or for other individuals on a for-cash basis. Absent state patronage, he says, deploying them amounts to little more than black magic, and it is in this context that many Hindus in India today deny the relevance of Tantra to their tradition and dismiss what they call tantra-mantra.

That is the bridge. Now measure it against the statute.

The Schedule to the Maharashtra Act of 2013 runs on karṇī, bhānāmatī, cheṭuk, jāraṇ-māraṇ, mantra-tantra and gaṇḍā-dorā. Six terms and one derogatory label, and only one of them has a classical ancestor. Māraṇa survives, inside a Hindi compound, at items 4 and 7. Śānti, vaśīkaraṇa, stambhana, vidveṣaṇa and uccāṭana appear nowhere in the Schedule at all. Item 9, which criminalises diverting a person bitten by a dog, snake or scorpion from medical treatment, names a practice that in the classical filing belongs to the Gāruḍa stream, the same stream under which a Śābaratantra is classed.

The categories the compendia built and the practices the law names are largely separate worlds. The vernacular world, the corpus, the trade and the statute that polices it, has consistently declined the six-act grid.

The shape of the finding

The scheme has one durable modern life, and it is worth ending on because it is exactly the right size. The Ṣaṭkarmadīpikā is still in print, under a title that advertises the six acts by name. That last observation comes from a bookseller’s catalogue entry, which is a fact about a title page and not scholarly evidence, and it is flagged as such.

A grid that cannot fix its own membership, that nobody can date, that sits as the last chapter of books organised on another principle, that one recension of a text uses while its sibling refuses it, that the vernacular corpus never adopted and the statutes never learned. And it is still the first thing anyone reaches for when this material comes up.

The tradition wrote the disclaimer and then wrote the chapter. This treatise has written the chapter and withheld the contents.

The grid was always a way of seeing. It was never the shape of the thing seen.