The Krama: the school that worshipped a sequence

Some time between the years 939 and 948 of the Common Era, in Kashmir, a king had a brahmin’s forehead branded with the shape of a dog’s foot.

The king was Yashaskara. He had committed himself, in the words of the chronicler who recorded it, to the vigorous supervision of the caste-classes and the brahmanical disciplines, and he had discovered that an ascetic named Chakrabhanu had broken those disciplines at something called a chakramelaka, a gathering of a circle. What happened at such gatherings we will come to. It was enough to mark a man’s face for the rest of his life.

The story did not end with the branding. Chakrabhanu’s maternal uncle was Viranatha, who happened to be the king’s own minister of war and peace. The gurus of the tradition, the chronicler tells us, still boast that Viranatha performed a rite of chastisement against the king in his fury, and that the king was dead within seven days.

The chronicler is Kalhana, writing his Rajatarangini, the River of Kings, in Kashmir around 1148. He has a source-critical habit unusual in Sanskrit literature, and he does not believe a word of the last part. The rendering below is Sanderson’s, who emends the verse against Stein’s edition and reads chakramelaka as the name of a rite, where Stein had tentatively suggested it might be a place-name:

“The story that they themselves have put about is that the king died at the end of seven days. But how can that be reasonable when what killed him was the illness from which he had long been suffering?”

That single sceptical sentence is one of the most valuable things we have. It is an outsider’s notice, dated by a reign, of a tradition that otherwise speaks about itself almost entirely in its own voice.

One caution before we build on it. Sanderson proposes that this Chakrabhanu is the Chakrabhanu of the Krama lineage, and that Viranatha, the king’s minister, is the man the Krama’s own literature calls Hrasvanatha, Chakrabhanu’s guru. The identification rests on a shared alternative name and a chronology that fits. It is a proposal. It is a strong one, and much of this section depends on it.

If it holds, this is what the Krama looked like from the outside: disreputable, punishable, and confident enough that a court minister was named in a story about killing his king.

A school with fourteen names

Start with what to call it, because the tradition could not settle that either.

Alexis Sanderson, whose 2007 study The Saiva Exegesis of Kashmir remains the most detailed treatment in print, collects the self-designations in a single footnote, each with a citation. Kramashasana. Kramartha. Kramadarshana. Kramanaya. Mahartha, the Great Meaning. Mahanaya, the Great Way. Mahamnaya. Mahashasana. Kalinaya, the Way of Kali. Kalikakrama. Devatanaya. Devinaya. Atinaya. Rahasyamnaya, the Secret Transmission. Fourteen, and his list opens with “e.g.”

Navjivan Rastogi, whose The Krama Tantricism of Kashmir of 1979 is the standard monograph and an independent reconstruction that disagrees with Sanderson at several points, gives the naming problem an entire chapter and adds more that most accounts never mention: Auttara Krama, from the Northern Seat; Chakrakrama-sampradaya; Navachakra-sampradaya.

And Rastogi notices something sharper. No traditional compendium of Indian philosophy available to him notices the Krama at all. The Sarva-darshana-sangraha, the great fourteenth-century survey of the schools, does not have it. The compendia that catalogue Indian philosophy for Indian readers, in Sanskrit, across centuries, did not see a Krama school to catalogue.

That is worth sitting with. The modern habit of speaking about “the Krama school” as a discrete unit with a doctrine and a membership is doing something the tradition’s own classifiers did not do.

The two scholars who put the name into modern circulation are both Indian. K. C. Pandey added a chapter on the Krama to the second edition of his study of Abhinavagupta in 1963. Gopinath Kaviraj noticed it in a paragraph or two. Rastogi, writing in 1979, names them as his only predecessors. Everything that follows in English, including this article, stands downstream of those three.

A word that does two unrelated jobs

One clarification has to be made early, because it trips careful readers.

Krama means sequence, step, succession, from the root kram, to step or proceed. The same ordinary word names a completely different thing in the Vedic world: krama-patha, the step-recitation, one of the modified recitation patterns by which the Samhitas were preserved. A reciter takes the words of a line numbered one, two, three, four and says them in overlapping pairs. One-two. Two-three. Three-four. It is a mnemonic device of extraordinary power, and we have written about it in the account of how the Vedas were preserved.

No source connects it to the school in this article, and no line of descent runs between them. It is one word, from one root, with one core sense of stepwise succession, pressed into service by two technical traditions that had nothing to do with each other. An account that derives the Kali-krama from Vedic recitation is inventing a link.

The sequence itself

What, then, is being sequenced?

The answer is the reason the tradition is interesting, and also the reason it is difficult. The Krama sequences neither a cosmic era nor a ritual calendar. Its subject is the inside of a single act of perception.

A collection of aphorisms compiled by the Kashmirian guru Bhattashri Shitikantha, which Sanderson says he has seen only in a manuscript whose copyist warned that his exemplar was fragmentary, glosses the school’s name outright. Consciousness, it says, flows eternal like a river, and manifests four states in every cognition: emission, and the rest. The teaching of the Krama is the sequence of those states.

The four are srishti, emission or emergence; sthiti, stasis, the holding of the thing in view; samhara, withdrawal, the drawing of it back in; and then a fourth, which is where the tradition becomes itself. The fourth is anakhya. The Nameless.

It is nameless for a precise reason. It is the phase in which withdrawal itself is withdrawn. There is no residual object left over for a name to fasten to. Rastogi puts the doctrinal version plainly: the ultimate reality, being free of temporal and spatial affiliation, lies beyond the reach of words, and that is the exact implication of the word anakhya.

Rastogi then does something the article ought to report rather than quietly use. He suggests that anakhya may be a positive Krama reading of the Madhyamika Buddhist shunyata, emptiness, reaching the Krama through esoteric Buddhism rather than from the Madhyamikas directly. That is his interpretation and not an attested Krama self-statement. It matters here because it puts the tradition’s crowning concept inside the Shaiva and Buddhist question we come to later, on the opposite side from where a Shaiva reader might expect.

Jnananetra, the earliest teacher whose lineage every Krama author traces, put the phase as praise rather than definition. In his Kalikastotra he salutes the Goddess resting in the ground that lies within the state in which withdrawal itself is withdrawn, untouched by the existent and non-existent that she projects, and manifest in everything at every moment.

Then the structure does something that repays attention. Emission, stasis and withdrawal each contain all four again. Three phases, each fourfold, and you have twelve. The twelve are the Kalis.

The sub-phases have names, and the names are not stable across the tradition. In Jnananetra’s own hymn the fourth of them is vishrama, rest. The later Trivandrum Mahanayaprakasha gives a different and more dramatic set: udaya, emergence; avabhasana, manifestation; charvana, which means chewing, the relishing or internalising of what has been perceived; and kalagrasa, the devouring of time. Kshemaraja, Abhinavagupta’s pupil, uses a fourth version: udyoga, the initial exertion; avabhasana; charvana; vilapana, dissolution.

That same Trivandrum text maps the whole cycle onto the standard Indian epistemological triad. Prameya, the object known. Pramana, the means of knowing. Pramatri, the knower. Emission is the reaching out towards the object, stasis its manifestation, withdrawal its reversion into the perceiver’s awareness of himself as the perceiver. And anakhya is what remains when even that subjectivity has been devoured.

It would be convenient to say the Krama got there first. It did not. Sanderson dates the Trivandrum Mahanayaprakasha after Abhinavagupta, holds that it has adopted his way of construing the twelve, and calls the substitution of this triad for the older ones one of the most striking characteristics of Abhinavagupta’s Trika exegesis. On this particular point the traffic ran from the Trika to the Krama.

On most points Sanderson argues it ran the other way. The Krama, he writes, exercised a profound influence on the Trika and, through Kshemaraja, on a much wider range of Shaiva texts, while influence in the reverse direction is far less evident. A tradition can be the creditor in general and the debtor in a particular.

The presiding goddess of the whole Kalikula is Kalasamkarshini, whose name means something like she who draws in, or devours, time. She should be kept distinct from the goddess who crowns the circle of the Nameless, who is Mahabhairavachandograghorakali, counted thirteenth at the centre in the Krama’s own scheme and twelfth in Abhinavagupta’s. Of the Goddess in the general sense the Trivandrum Mahanayaprakasha says something with a care that is almost embarrassed: the Lord’s repose within himself is the highest state of self-awareness, and yet by the finest of distinctions there shines a state higher still, a Goddess-ground in which even the Lord cannot see his way.

A note for anyone chasing the name through the manuscripts. The later Kashmiri authors write Kalasamkarshini; the scriptural manuscripts consistently write Kalasamkarshani. Both are the same goddess.

The worship that abolished the apparatus

Here is where the Krama stops resembling what most readers expect a tantric school to be.

Read the standard tantric ritual manuals of the period and you find an enormous machinery: initiation and consecration rites, mandalas laid out in coloured powder, hand-postures, elaborate visualisations of the deity’s form and ornaments and weapons, fire-sacrifice, post-initiatory ascetic observances measured out in days and syllables. The Jayadrathayamala, one of the great scriptures of the Kalikula, the Kali-centred stream within the Bhairava canon, is full of it. It also carries Krama material, which Sanderson notes sits embedded there in a mass of chapters of the ordinary type.

The Krama scriptures proper are a different matter. Of four of them, the Kalikulapanchashataka, the Kalikulakramasadbhava, the Devidvyardhashatika and the Yonigahvara, Sanderson writes that they “differ in fundamental respects from the Mantramargic standard,” teaching “no rituals of initiation and consecration,” and:

“no hand-postures (Mudras), no visualizations of the constituent deities’ forms, no fire-sacrifice, and none of the elaborate ascetic observances…”

Initiation, in the post-scriptural tradition, was the simple act of consuming a sacrament, the Kaula nectars, an act called charuprashana. That is the whole of it.

The scriptures announce the negation programmatically. In one Krama text, Bhairava asks for the teaching that is free of restriction by lunar day and auspicious hour, without appointed place or time, free of the conventions of sacred sites, without hand-postures and mantras and coloured powders, without fire-sacrifice or oblations of grain and sesame, without the summoning of deities, without post-initiatory observance, and above all free of the exhalations and inhalations and retentions of breath-control.

And note who is asking. The Krama scriptures are framed as instruction given by the Goddess Bhairavi to Bhairava. The god asks. The Goddess teaches. Sanderson calls this “an inversion of the normal gender of teacher and pupil,” and it is structural rather than decorative: it is how the texts are built.

There is a further ranking that popular accounts almost always miss, though it should be attributed to its source rather than to the tradition at large. The anonymous author of the Trivandrum Mahanayaprakasha teaches that contemplative worship is the lowest of three paths. Above it stands the path of oral instruction, kathana, transmitted through terse mystic aphorisms called chumma, which aims at sudden enlightenment. Above that stands attainment with no instruction at all, arriving spontaneously or through some wordless stimulus such as the guru’s glance. A commentary preserving one hundred and five of these aphorisms in Old Kashmiri survives, with the Vatulanathasutra alongside it, and both are scarcely mentioned outside specialist work.

Even the order of worship was a screen. The tradition distinguished pujakrama, the sequence in which the phases are worshipped, from samvitkrama, the sequence in which they actually unfold in consciousness. The scriptures and the gurus taught the first, according to Jayaratha, in order to conceal the second.

So what was the chakramelaka that cost Chakrabhanu his face?

It was a Kaula rite, a gathering of a circle of male initiates and yoginis, women understood as embodiments of the Goddess’s powers, with what Sanderson describes as a strongly orgiastic character. The sources are specific about who those women were: sixteen, eight or four Yoginis of named castes, mostly or entirely low, below the rank of Shudra. Matangi, Charmakari the leather-worker, Dhivari the fisherwoman. That detail belongs in any honest account of the rite, and it should be allowed to sit there without being turned into a compliment to the school. We do not know how these women understood their own participation. Judit Torzsok’s caution about the whole genre applies here with force: these texts are normative and describe an ideal state of things, and they are not a census.

Some Krama gurus were ascetics of the Kapalika observance, and Sanderson’s description is worth quoting rather than paraphrasing: they went about “decking themselves with ornaments of human bone, carrying a human skull as a begging bowl, and living in the cremation ground.” Where Abhinavagupta’s Trika made the rejection of such socially distinctive externals a principle of its universalism, the Krama and the Kalikula kept their distance from ordinary society.

This site does not give instruction in any of it, and nothing here should be read as one. But an account that describes the Krama as a serene philosophy of consciousness and stops there has described half of it, and not the half that got a man branded.

The women at the source

Every Krama author who speaks of the lineage traces it back to one teacher: Jnananetra, also called Jnananetranatha and Shivananda.

That is not quite the same as saying he began it, and Rastogi does not think he did. Working from a Kaula line of Siddhas that the Devipanchashatika and the Yonigahvara record, he argues that Nishkriyananda should be recognised as the system’s first thinker, that Krama seeds were sown much earlier than one is apt to believe, and that in a sense it is the earliest of the monistic Shaiva systems of Kashmir. Sanderson’s reconstruction begins with Jnananetra. Rastogi’s reaches behind him. What follows is the lineage as the tradition itself tells it.

He was active in the ninth century. Beyond that the dating is a reconstruction, and we will come to how shaky a one.

What the tradition says of him is that he received the teaching at Uddiyana, directly from the Yoginis of the place, who are called the Pithesvaris, the goddesses of the seat, or from their leader Mangala.

He is said to have initiated a number of disciples. A lost lineage-record quoted by Jayaratha says seventeen; Jayaratha himself raises the count to at least nineteen and remarks that the source failed to count two of them. Three of these disciples are the ones that matter for everything afterwards, and all three are women: the Yoginis Keyuravati, Madanika and Kalyanika. From them the teaching passed to three men, Govindaraja, Bhanuka and Eraka.

Later gurus trace the whole line back through Keyuravati alone. She is also called Kakaradevi, the Goddess K. Sanderson describes her as occupying a position of particular reverence in the tradition as the conduit through which Jnananetra’s teachings reached all subsequent gurus. He also observes, elsewhere, that in this environment no other tradition claims to have descended from a woman.

That is a real and unusual fact, and it deserves to be stated. It also deserves not to be inflated, and it routinely is.

Sanderson’s own caution, in the same paragraph that establishes Keyuravati’s centrality, is the sentence popular retellings drop. Abhinavagupta’s account, he writes, though it adds Madanika and Kalyanika, claims no independence for them as sources of individual lineages, merging them somewhat implausibly with Keyuravati as co-initiators. And then: no work is attributed to her in spite of her eminence, or to any other of the known female gurus of this tradition.

Rastogi takes the other side, carefully. The importance of the female teachers, he writes, may be assessed from the fact that the system is said to have originated from the mouth of the Yoginis, and while it is difficult to establish cause and effect between the tradition’s orientation towards Shakti and its early preaching by female ascetics, it is a factor to be reckoned with. Note the hedge. It is his, and it is usually cut.

The claim sometimes made, that the Krama was the first tradition in India to grant women full guru-hood, is not supported by any specialist source we have read. Sanderson’s claim is the narrower one about this Kashmiri environment. The narrower claim is remarkable enough.

A Buddhist pilgrim checks the map

Uddiyana, where the revelation is said to have happened, was a real place: a small kingdom north of Peshawar in what is now Pakistan, known to Chinese travellers as Wuzhangna and to Tibetans as U rgyan. It lay about 150 miles from the Kashmiri capital as the crow flies. Proposals locating it in eastern India, at Kanchi, at Kashgar or at Urgench have been made, and Sanderson dismisses the evidence for them as negligible.

Uddiyana is also one of the great seats of Buddhist Vajrayana. This is where the story acquires an unexpected witness.

Sanderson checks the Kashmiri claims about the site against the travel record of U rgyan pa rin chen dpal, a Tibetan Buddhist pilgrim who lived from 1230 to 1309 and went there. The pilgrim attests the goddess Mangala directly. He also records a cremation ground east of the town under a name, bhir sma sha, which Sanderson reads as a Tibetan deformation of Karavirashmashana, the Karavira cremation ground of the Krama’s own account. The first is an independent attestation. The second is a reading of a Tibetan transcription, and a reader should know the difference.

Sanderson’s conclusion about the geography is that what the Kashmirian Krama knows of the place is exact. His conclusion about the man is a good deal drier, and the article that quoted only the first half would be selling something:

“It is not implausible that Jnananetra should have propitiated Kali in Uddiyana. There is no very strong evidence that he did…”

Jnananetra’s own words say only that he had his revelation in the great cremation ground, mahashmashana, which Sanderson notes may have been meant esoterically rather than geographically.

The revelation story remains a revelation story. What can be said is narrower and still worth something: the place existed, its landmarks are independently reported, and the tradition’s account of them is accurate.

The wider question of what passed between Shaiva and Buddhist tantra at sites like this is among the most contested in the field, and it should be named rather than settled. Sanderson has argued at length that the Buddhist Yogini-tantras took much of their material directly from Shaiva sources, and in 2009 he set out concordance tables to show it. The objection, which he states himself and attributes to Ronald Davidson and David Gray, is that this violates the text critic’s maxim lectio difficilior potior, that the more difficult reading is to be preferred. Sanderson does not concede the point; he replies that the objection misunderstands how the rule is to be applied. Shaman Hatley has since argued that at least one passage travelled the other way, from a Buddhist text into a Shaiva one. Davidson has since proposed a different model again, in which non-literate practitioners fed both traditions independently. That last is reported at second hand here and has not been read at source.

We do not adjudicate this. Readers should know the question is live, that serious scholars hold each position, and that any account presenting one of them as settled is not reporting the state of the field. Remember also that Rastogi, from inside the Krama’s own scholarship, suspects Buddhist influence running into the Krama at the level of its central term.

One thing should be said plainly, because it circulates: the case Sanderson builds rests on the Chakrasamvara and Hevajra material. It is not a Kalachakra argument. Accounts that tie the Krama to the Kalachakra Tantra are asserting something the specialist literature does not.

The goddess who left the record

Now the thing this article exists for.

In the culminating phase, the phase of the Nameless, the Krama worships a circle of Kalis. How many?

The scriptural mainstream says thirteen. That is the count in the Kalikulapanchashataka, in Jnananetra’s own Kalikastotra, and in all three of the texts titled Mahanayaprakasha. One scripture, the Kalikulakramasadbhava, stands apart and teaches seventeen.

Abhinavagupta presents twelve.

The thirteenth, in the tradition’s own account, is the twelve seen as one, worshipped at the centre of their circle. The Trivandrum Mahanayaprakasha is careful about its status: the ground in which the separate identities of the twelve are dissolved, it says, is what we call the thirteenth goddess figuratively, upacharatah, and it is the highest object of worship because it is the goal.

So where did the thirteenth go in Abhinavagupta’s version? It did not. Something stranger is at issue, and here the two authorities part company.

The Kali who disappears is one from the middle of the sequence. Her name is Sukali, and Yamakali immediately precedes her.

Sanderson’s account is that the loss is editorial and that the mechanism is recoverable at the level of a single word. In the hymn that names her, the reading sukalim was expunged and bhavanim put in its place: a generic epithet of the Goddess, which allowed the verse to be read as a second verse on Yamakali. One name out, one non-name in, and the count falls from thirteen to twelve. He attributes the change to Abhinavagupta or to a predecessor and does not claim to know which.

Jayaratha, writing two centuries later, defends the twelve, and he has three arguments. Told that the substitution disrupts the hymn’s pattern of one verse per goddess, he replies that there is no such pattern, since a second verse is given to another Kali too. Told that the offending verse is an interpolation, he replies that he has seen it in Hrasvanatha’s own autograph of his commentary on the hymn. And he argues that Jnananetra taught both counts, twelve in one verse of the Kalikastotra and the thirteenth in another, so that both traditions are valid options.

Sanderson answers the third and then delivers the verdict this article was built to reach:

“Since Jayaratha does not cite any textual evidence that the tradition of twelve occurs earlier in this line of transmission—he would surely have done so if any were available, since that would greatly have strengthened his position as Abhinavagupta’s defender—one is bound to suspect that all this is special pleading to justify a major innovation.”

Fairness requires the next sentence too, and it is Jayaratha’s strongest ground. Sanderson grants that the fact Jayaratha can cite a passage from scripture appearing to list only twelve Kalis would at least show that Abhinavagupta’s view has a basis in the tradition of the Kalikula, even if not in the line descending from Jnananetra. He goes on to suspect that this passage has itself been modified. Suspicion, twice, in a scholar who says so.

And now the disagreement that a piece leaning only on Sanderson would hide.

Rastogi, working from the same Jayaratha material, does not tell a story of excision at all. On his reading the Kramastotra itself excluded Sukali and numbered the goddesses twelve, while the Panchashatika, the Kalikulapanchashataka under its shorter name, retained her: two positions in the literature rather than one act of editing. He points to a commentary called the Vivarana as evidence that the thirteen-party had a different text of the hymn, or a different reading of it, with the verses in a different arrangement. And he sets the problem inside a decline: by Jayaratha’s time the Panchashatika’s own tradition had suffered a great setback, he writes, and its text had started to carry variants and had naturally become liable to varying interpretations.

Two scholars, the same evidence, two accounts. One sees a philosopher adjusting a hymn to fit a system. The other sees a damaged transmission throwing up rival recensions, with nobody to blame. This article has followed Sanderson because his reconstruction identifies a mechanism at the level of a word, which is the kind of claim that can be checked. That is a reason for a preference. It is not a proof, and Rastogi is not a lesser authority.

What both agree on is that something moved.

Rastogi also maps the numbers onto something larger than a textual accident. Krama authors who make Shiva ultimate tend towards twelve Kalis and four phases; those who make Kali ultimate tend towards thirteen and five, the fifth being bhasa, which Sanderson renders Pure Light. The count is a theology in miniature. He immediately warns that the correlation overlaps and that no uniform rule can be laid down, which is exactly the kind of hedge that makes him trustworthy.

The disagreement about a goddess’s name is a disagreement about whether the last word belongs to the god or to her.

What the lamp was carried in

It is worth being explicit about how thin the evidence is.

The Kramastotra, the hymn at the centre of the dispute, was honoured with the title Kramastotrabhattaraka, an honorific otherwise reserved for scripture. Who wrote it is itself disputed, and the dispute is a small model of the larger one. Sanderson attributes it to Eraka. Rastogi, reading the same page of Jayaratha, declines that identification and argues for a Siddhanatha whom he places in the first half of the tenth century, treating Eraka only as the point beyond which the hymn cannot be later. Two scholars, one page, two authors. No manuscript of it has reached us. Abhinavagupta’s Kramakeli, the commentary from which the core of the lineage account comes, is also lost. Both survive as quotations inside Jayaratha’s commentary on the Tantraloka.

The scriptures themselves survived by leaving. Most of the principal Krama texts come down to us in Nepalese manuscripts. A practical manual was written around 1200 by Vimalaprabodha, teacher to the Nepalese king Arimalla. The major later exegete, Maheshvarananda, was South Indian and wrote his verses in Prakrit. A tradition everyone calls Kashmiri is preserved substantially outside Kashmir.

Here, too, is the promised reckoning with the dates. There are three fixed dates in the Kashmirian Shaiva literature of this period, all of them Abhinavagupta’s, and everything else is counted backwards from them by assuming twenty-five years to a teaching generation. Sanderson, who makes the assumption explicit, calls the result “this crude calculation” and puts Jnananetra at roughly 850 to 900. Rastogi runs the same method from different anchors and arrives fifty years earlier, at 800 to 850. Neither is a date in the sense a reader might assume. They are arithmetic performed on a lineage list.

In Kashmir itself the tradition thinned. Sanderson, who was in the valley between 1972 and 1977, records that he encountered no sign that any practical knowledge of the ritual aspects of the Trika or the Krama had reached the present. In a separate essay on Swami Lakshman Joo he is careful about what that means and does not call anybody dishonest: systems of this kind are unusually subject to loss, and when they fragment and contract, their teachers are barred from seeing it by their faith in what their own teachers gave them.

Even the name of the larger field is modern. J. C. Chatterji, first director of the research centre that produced the Kashmir Series of Texts and Studies, appears to have been the first to write “Kashmir Shaivaism”, in a book of that title in 1914, and Dyczkowski notes that he did so simply to distinguish it from other Shaivisms, and that the term appears nowhere in the works of the authors it describes. Sanderson calls the label “doubly misleading,” because it excludes the Shaiva Siddhanta and because it fosters “the illusion that non-Saiddhantika Saivism was a single doctrinally unified whole.”

Which returns us to where we began, and to the trap on the other side. If the umbrella term flattens several traditions into one, the correction is not to seal them into sects that never spoke. Sanderson warns against that too, hoping to explore “the co-functionality of these Saiva systems that we have too readily seen as though they were those of distinct schools or sects.” Rastogi, arguing hard for the Krama’s independence, notes in the same book that Jayaratha sometimes refers to the Krama as the Trika as though the two were one.

Both things are true. They were not one system. They were not several sealed ones either.

The book the south produced

Rastogi names three authors for the sheer density of historical allusion packed into their work: Abhinavagupta, his commentator Jayaratha, and a man who lived sixteen hundred miles south of them.

At a date recent scholarship puts around 1300, Goraksha, son of Madhu or Madhava, composed seventy-one verses in Maharashtri Prakrit and then wrote a long Sanskrit commentary on his own verses. The verses are the Maharthamanjari, the flower-cluster of the mahartha. The commentary is the Parimala, its fragrance. Goraksha’s initiatory name, given him by his teacher Mahaprakasha, was Maheshvarananda. Sanderson, Cox and Wenta all place him at Chidambaram in the Cola country of the Tamil south. Rastogi, who is otherwise the fullest guide to him, puts the Cola country in modern Karnataka, which is a mistake, and records a page later that Maheshvarananda travelled far and wide in almost every direction before he began to write.

The Prakrit has a story attached, and the commentary tells it. A yogini appeared to him in a dream, and because she expressed her will in Prakrit he wrote in Prakrit. He took the dream to Mahaprakasha, who interpreted it, and on that interpretation he began. Rastogi, reporting this, prefixes it with “the story has it,” and the hedge is worth keeping. A. N. Upadhye judged the result mechanical, the work of a man who drafts in Sanskrit and then renders it into Prakrit. Rastogi concedes the point about the Prakrit and adds that philosophically the text is mature.

Then the part that bears on everything above. The Parimala quotes, and much of what it quotes is gone. Its author lists eleven works of his own, possibly twelve. Two survive, the verses and the commentary on them, and the rest are known only from his own references, three of those without so much as an extract. His teacher Mahaprakasha was described as a composer of tantras and nothing of his survives on its own, so that of the Anandatandavavilasastotra, a hymn to Nataraja’s dance of bliss at Chidambaram, Aleksandra Wenta can print one stanza. The Kramasiddhi is an agama that none of Abhinavagupta, Kshemaraja or Jayaratha knew, and it is that silence which makes Rastogi read it as a late work; four extracts in the Parimala are its entire surviving record. Abhinavagupta’s Kramakeli, which this article has already met in pieces inside Jayaratha, is quoted on eight pages of the Parimala.

Rastogi puts him beside Jayaratha without hedging:

“Like Jayaratha, he has secured an enviable position, since he ensures a monolith of information for the historical reconstruction of the Kashmir Saivism in general and the krama system in particular.”

The same book is harder on him later. Where Maheshvarananda ranks the Mahartha above every other discipline, Rastogi calls that hardly a scientific approach and follows Jayaratha instead.

His date is unsettled in the way the rest of this article’s dates are unsettled, and by a full century. Rastogi places him at 1175 to 1225, reasoning that Jayaratha alludes to his grand-teacher. Sanderson in 2007 placed him at about 1275 to 1325. Whitney Cox dated him early in his 2006 dissertation and moved to Sanderson’s range afterwards; Wenta writes thirteenth to fourteenth century. The trap for anyone citing this is that Sanderson’s own 1987 encyclopedia entry on Krama Shaivism gave 1175 to 1225, and that entry is still online, so a writer can cite Sanderson for a date Sanderson abandoned.

One more thing, because the usual version of it is wrong in both directions. Maheshvarananda’s grand-teacher was the Shivananda whom Rastogi numbers second to keep him apart from the ninth-century founder of the lineage, and who wrote the Rjuvimarshini, a commentary on the principal Sri Vidya scripture. Maheshvarananda ascribes six works to him in all, five of which, the Rjuvimarshini among them, Rastogi classes under the Tripura system. The lineage link is firm, then, and so is the traffic afterwards: Rastogi holds that Maheshvarananda considerably influenced the tantric literature that followed him, that Rameshvara refers to him on no less than three occasions in his commentary on the Parashurama Kalpasutra, and that Kaivalyashrama quotes the Parimala itself in the Saubhagyavardhini on the Anandalahari. The book itself, meanwhile, is classed as a Krama work by Rastogi, Sanderson and Cox alike. Its central figure is the Vrindachakra, and Rastogi’s reading of that figure cuts against a simple story in either direction: he thinks the Srichakra may have inspired it, and he thinks the idea was exported from the Krama into the Tripura system. Influence is documented. Its direction is not.

Rastogi’s own summary of the whole episode is worth the detour. As the Krama was decaying in Kashmir, Maheshvarananda and his teachers were kindling it on the banks of the Kaveri, and between them, he says, they succeeded in evolving a sort of southern school of the thing, though it is nowhere stated or claimed.

What survives

Here is what can be said with confidence.

A Shakta tradition existed in Kashmir from around the ninth century which held that the whole of reality is legible in the structure of a single act of noticing something. It worshipped that structure as a sequence of goddesses. It stripped away the ritual apparatus that every comparable school considered indispensable. It put the Goddess in the teacher’s seat and the god in the pupil’s. It traced its authority back through a woman whose own writings, if she had any, are gone. Some of its teachers lived in cremation grounds, and one of them was branded by royal order for what he did at a circle-rite.

And in the passage from the ninth century to the eleventh, the goddess Sukali left the record.

Whether that was a philosopher bending a hymn to a system, or a damaged manuscript tradition producing rival counts, the sources do not let us say. Sanderson suspects the first and says he suspects it. Rastogi argues the second. Neither of them saw a manuscript of the Kramastotra, because none survives. Rastogi counts about twelve of its verses recoverable from Jayaratha’s quotations, and thinks it once held a few more.

That this article has tried to map the sequence does not mean the sequence needs the map.

The Nameless was never going to answer to a name.