If you want to know what Tantra looked like before it fractured into rival schools, you read the Niśvāsa-tattva-saṃhitā — "the Collection of Principles Exhaled by God" — the first indisputably Tantric text, composed around 500–550 CE in north India.1 It's the closest thing to a snapshot of Tantra's origin. And the most revealing thing about it is what it lacks: a clear sectarian identity. It nominally belongs to the Siddhānta (the dualist right-current) — is even considered that school's root-text — "and yet it also exhibits some features that would later be associated with very different (transgressive goddess-worshipping) groups."2 In other words, in the earliest phase, "the tradition had not yet differentiated into the two primary currents" (Siddhānta and Kaula). The Niśvāsa is Tantra before the schism — both streams still mingled in one undivided source.
The foundational text-historical anchor for Tantra's origins:3
The Niśvāsa's importance is evidential, and it works on two fronts. First, on dating and origin: it's a century earlier than the first comparable Buddhist Tantric text, and it draws on the older Atimārga — together strong evidence for Śaiva priority and for Tantra evolving out of pre-Tantric Shaivism rather than appearing from nowhere. Second, and more subtly, on the nature of sectarian division itself: the fact that the earliest text mingles features later split between rival schools tells you the schools were later developments, not original essences. The Siddhānta/Kaula divide that came to define the tradition wasn't there at the root — it emerged. The streams diverged from a common source that contained both in undifferentiated form.
This reframes how to think about the whole sectarian landscape. The schools aren't eternal, mutually-exclusive truths that happened to coexist; they're differentiations of an original unity, like a river splitting into branches downstream. And the text's detail "suggests an earlier oral tradition of at least a century or two" — so even the earliest written Tantra points back to an undocumented oral phase, meaning the origin recedes further than the manuscripts can reach. The Niśvāsa is thus a kind of fault-line marker: it shows where the tradition was before its defining split, which both grounds the dating and quietly relativizes the later sectarian certainties — if Siddhānta and Kaula both flow from one pre-sectarian source, their later opposition is a development, not a bedrock fact.
The Niśvāsa anchors the Śaiva-priority origin thesis (earliest, draws on Atimārga) and shows the two streams as later differentiations of an originally-undivided source. Its drawing on the Atimārga connects to the Pāśupata/Atimārga pre-history, and its Nayasūtra's "formless ground the ego can't appropriate" prefigures the essence-nature doctrine and the ahaṅkāra (self-image-manufacturer) teaching.
Wallis quotes the Niśvāsa's Nayasūtra directly, and the passage contains a teaching that looks Buddhist — and the way he handles the resemblance is instructive about reading early texts.10 The Goddess asks how union with the "Independent Deity" comes about. The Lord answers: everything seeable, speakable, thinkable, imaginable "are made aspects of 'I' [appropriated by ego]" — so "one should search for the place in one's own body where there is no such form."
The instruction — find a formless place within that the ego can't appropriate — "shows an apparent overlap with Buddhist ideas," the search for a formless ground beyond ego-grasping.11 A careless reader might conclude the early Śaivas borrowed this from Buddhism (reversing the priority Wallis argues for elsewhere). Wallis blocks that move: "as later scriptures show, the idea of an innate ground of being that is formless and indescribable is just as much a part of Śaiva Tantra as Buddhist."12 The resemblance is convergence, not borrowing — both traditions, working the same contemplative territory, arrive at "a formless ground beyond ego-appropriation," because that's what the territory actually contains, not because one copied the other. The case study models a crucial interpretive discipline: a shared idea between two traditions is not automatic evidence of influence in either direction — it can be independent discovery of the same reality (the same point made at scale in the Dzogchen convergence). And it shows the Niśvāsa, the very earliest Tantra, already teaching the formless-ground doctrine that would become central — so this isn't a late philosophical overlay but present at the origin, in the first text, before the schools split.
For a researcher or serious student, the Niśvāsa is a methodological touchstone. Treat it as the dating anchor. When questions of "how old is Tantra / which came first" arise, the Niśvāsa (~500–550 CE, a century before comparable Buddhist Tantra, drawing on the older Atimārga) is the key textual evidence — cite it for Śaiva priority and Atimārga descent. Read the schools as downstream. Hold the Siddhānta/Kaula divide as a later differentiation of an originally-undivided source (the Niśvāsa contains both), not as eternal opposed essences — which tempers sectarian over-certainty. Mind the oral horizon. The text's detail implies a prior oral tradition of a century or two, so the documented origin is not the actual origin — the roots recede beyond the manuscripts. Don't infer borrowing from resemblance. The Buddhist-looking formless-ground teaching is convergence, not theft — apply the same caution to any cross-tradition resemblance (shared content ≠ influence). Note the practice-first character. As the earliest text, it's cosmology- and practice-heavy (the Nayasūtra yoga chapter), consistent with scriptures being about practice while philosophy came later in the commentaries.
Projecting the schism backward. You read the later Siddhānta/Kaula division into the origin, assuming Tantra was always split into these rival camps — missing that the earliest text mingles both and the schools emerged. The tell: treating the sectarian divisions as primordial essences rather than later differentiations of a common source.
Inferring borrowing from resemblance. You take the Niśvāsa's Buddhist-looking teachings (or any cross-tradition similarity) as proof of influence/copying, jumping from "looks alike" to "borrowed from." The tell: confident claims of derivation where the evidence supports independent convergence on the same contemplative reality — the same error that, run the other way, would wrongly make Śaiva Tantra a Buddhist derivative.
Cutting-edge text-historical scholarship (the Niśvāsa's dating and significance are recent findings; Wallis notes "scholarship in this area is still in its early days").13 The genuine tension is between the textual dating (~500–550 CE, evidence-based) and the oral-tradition claim the text's own detail implies (a prior century or two) — plus the explicit note that "the oral tradition would like to date the beginning of Tantra far earlier." So there are competing timelines: the conservative textual one (mid-5th c. for Tantra proper), the inferred-oral one (a century or two earlier), and the tradition's own self-dating (far earlier still). The open question is how much weight to give the inferred and traditional earlier dates against the hard textual evidence — and, given that only one manuscript of the Niśvāsa survives (in Nepāl), how much of the "earliest Tantra" picture rests on a single fragile witness whose survival is partly luck, raising the question of what didn't survive and whether the picture would shift if it had.
Wallis here writes squarely as the Sanderson-trained critical historian, and this is a place of clean convergence with the vault's critical-historical corpus (Samuel, White) rather than tension — the Niśvāsa dating and Śaiva-priority thesis are mainstream current scholarship, and Wallis presents them as settled ("recent scholarship has sealed the deal," "this controversy can now be dispelled"). But there's a quiet internal tension between Wallis-the-historian and the oral tradition he's part of as a practitioner: the historian dates Tantra proper to the mid-5th century on textual evidence, while "the oral tradition would like to date the beginning of Tantra far earlier" — and Wallis, the scholar-practitioner, sits astride both. He sides clearly with the textual evidence here (the historian wins), which is notable given that elsewhere he defends insider/practitioner warrants against pure critical history. What that reveals is the discipline of his scholar-practitioner method on display: on matters of dating and historical fact (Part-2 territory), he defers to the textual evidence even against his own tradition's preferred earlier dating, reserving the insider warrant for matters of doctrine and practice (Part-1/3 territory). The same author who will assert nondual doctrine as a devotee will, on the question of when Tantra began, overrule the oral tradition's wishes with the manuscript evidence — exactly the Part-by-Part epistemic split his method prescribes.
The plain link: this is a case study in reconstructing origins from fragmentary early evidence and distinguishing convergence from borrowing — connecting to the psychology of how traditions remember (and mis-date) their own origins, and the behavioral mechanics of how sectarian identities form from a common source.
The Sharpest Implication If the earliest Tantra we possess mingles, in one undivided text, the features that later schools would split into bitter opposition — dualist Siddhānta orthodoxy and transgressive Kaula goddess-worship together — then the sectarian divisions that came to define the entire tradition were not there at the root. They emerged. Which means the certainties each later school held about being the true Tantra, opposed to the others, were certainties about a differentiation mistaken for an essence. The destabilizing part generalizes far beyond Tantra: the deepest, most identity-defining divisions within any tradition — the ones whose partisans experience them as primordial and absolute — may turn out, at the documented origin, to be downstream branchings of a common source that contained both possibilities undivided. The river was one before it forked, and the forks, looking back, each claim to be the original stream. The Niśvāsa is the rare surviving glimpse of the water before the fork — and what it shows is that the thing you take to be the eternal opposition at the heart of a tradition may be a family quarrel among siblings who've forgotten they share a parent, the parent being a single text, exhaled (the tradition says) by God, before anyone had decided which kind of Tantra was the right kind.
Generative Questions