When people hear "nine traditions of Tantra," they picture nine rival sects, each claiming to be right, fighting for converts. That's a Western, monotheistic projection. Wallis is careful: the word is sampradāya — "that which is handed down," a lineage — not "sect." There was no holy war between them, no heresy trials, no battle for souls. They're nine branches of one family tree, each handing down its own deities, mantras, and scriptures, often overlapping, sometimes nested, mostly coexisting.1
What makes one sampradāya different from another isn't a creed you sign. It's three concrete things: which deity you worship, which mantra you're given, and which scriptures you hold as revealed.2
Wallis surveys the medieval Śaiva Tantric world as nine principal sampradāyas, arranged along the left-right spectrum from tame to wild:3
The deity/mantra/scripture criterion is doing real work, and it explains why these nine could coexist without the violence of competing creeds.
In a creed-based religion, traditions fight because they make incompatible truth-claims — if my doctrine is right, yours is heresy. But sampradāyas aren't organized around belief; they're organized around practice and transmission. What you "are" is defined by whom you worship, what mantra your guru gave you, which texts your lineage hands down — not by a proposition everyone else must reject. Two sampradāyas worshipping different deities aren't contradicting each other any more than two families with different surnames are. You can even belong to a "higher" lineage that includes a lower one's revelation as a valid-but-partial truth (the "hierarchy of revelation"). This is why the tradition could hold nine lineages, overlapping and nested, without holy war: difference of lineage is not difference of creed, and medieval India asked "whom do you worship and what do you practice?" not "what do you believe?"4
This also explains the placement logic. The nine aren't a flat list — they're arrayed on the spectrum, and a sampradāya's position predicts its features. Knowing the Krama is far-left tells you it's nondual, goddess-centered, transgressive, cremation-ground, women-friendly — all of it — before you read a word of its texts.
The nine sampradāyas are the skeleton of Wallis's Part-2 history; each one cross-links into the yoga-and-tantra-history-hub. They're placed by the left-right spectrum and the two streams. Several connect to existing vault pages: the Śaiva Siddhānta page, the Krama/12 Kālīs page, the Netra Tantra page, the Tumburu/Vīṇāśikha page. This survey gives the vault the frame those scattered pages sit in — turning a pile of named sects into a navigable family tree.
The nine make their best lesson in their deaths. Of the nine, eight are extinct or fragmentary — known mostly through manuscripts, reconstructed by scholars like Sanderson. Only one, Śrīvidyā, survived as a living, intact tradition. Why?
Wallis's answer is a case study in what it costs a wild tradition to survive: Śrīvidyā survived because it got sanitized. Tamil brāhmins absorbed the goddess Tripurasundarī and her Śrī Cakra into a respectable, Vedānticized, purity-compatible form — they kept the deity and the yantra and quietly dropped or coded-over the transgressive, left-current, cremation-ground material. The sampradāya that lived is the one that moved right on the spectrum far enough to be acceptable to orthodox brāhminical society. The far-left lineages — the Krama especially, the most radical and nondual — didn't survive, precisely because their wildness couldn't be domesticated without ceasing to be themselves.
The lesson cuts two ways. It tells you that survival selected for tameness: what we inherit as "living Hindu Tantra" is disproportionately the right-shifted, sanitized end, while the most radical material is exactly what died — which means the living tradition systematically under-represents how wild the medieval reality was. And it sharpens the Wallis↔White tension: if the surviving record is skewed tame by the survival filter, then arguments about "what Tantra was really like" are arguing over a sample that history pre-selected. The eight dead sampradāyas are the missing evidence both the insider and the historian are reconstructing.
You meet a Tantric deity, mantra, or scripture and want to locate it. Run the sampradāya triangulation. Ask: which deity-family? (Sadāśiva → Siddhānta; a fierce skull-bearing goddess → Yāmala/Krama; Tripurasundarī → Śrīvidyā; the three goddesses → Trika; Svacchanda → Mantrapīṭha). What's the mantra's character? (universal/nonsectarian → Amṛteśvara; goddess-bīja → a left lineage). What scripture-class? Then place that sampradāya on the spectrum, and its full feature-set unfolds. The nine-fold map turns an isolated deity-name into a located branch with predictable neighbors, ancestors, and tilt.
Reading them as warring sects. You import the monotheistic "rival religions fighting for truth" frame and miss that these are coexisting lineages defined by practice, not creed. The tell: you look for the doctrinal war and can't find it, and assume the sources are incomplete.
Mistaking the survivors for the whole. You treat living Śrīvidyā (and sanitized modern Tantra) as representative of medieval Tantra, forgetting that survival selected hard for tameness. The tell: you picture the tradition as more respectable and right-shifted than the manuscript record of the eight dead lineages shows it was.
The nine-sampradāya survey is critical-historical (Sanderson's reconstruction, Wallis's Part-2 footing).5 The deepest open question is the survival bias itself: since eight of nine are reconstructed from fragmentary manuscripts and only the sanitized one lives, how confident can anyone be about the proportions — how much of the tradition was far-left versus middle? Both Wallis (mainstream-was-liberation-seeking) and White (mainstream-was-power-seeking) are estimating the shape of a whole from a damaged, non-random sample. The number "nine" is also a convenience of organization more than a hard count; the boundaries between lineages blur, nest, and overlap.
Wallis's nine-sampradāya map and the vault's Sanderson hierarchical-succession page draw on the same scholar (Sanderson) and converge almost entirely — which is the point: Wallis's Part-2 is Sanderson, transmitted by a devoted student. The interesting tension isn't between them but between this critical-historical survey and Wallis's own Part-1 insider theology. Here, in Part 2, he's a rigorous philologist counting lineages and dating texts; in Part 1 he's a devotee asserting the View. The nine-sampradāya survey is where you can watch the seam in the scholar-practitioner most clearly — the same author, in the same book, switching from "here is the dated manuscript evidence for nine lineages" to "here is the timeless truth the tradition teaches." Holding both registers is the whole method, and the sampradāya survey is the empirical floor the theology stands on.
A family tree of lineages defined by transmission rather than creed is a structure with analogues in how traditions and schools propagate.
The Sharpest Implication If survival selected hard for tameness — if the one sampradāya that lived is the one that got sanitized, and the wildest, most nondual lineage (the Krama) died precisely because it couldn't be domesticated — then everything modern seekers inherit as "authentic Tantra" is systematically the right-shifted remnant. The radical core didn't get refuted or outgrown; it got filtered out by the brute physics of what orthodox society would tolerate. Which means recovering the actual depth of the tradition isn't a matter of going to the living lineages — it's a matter of reconstructing the dead ones, the manuscripts, the eight branches that didn't make it. The most alive teaching may be the most extinct.
Generative Questions