Picture a scholar in his Oxford college rooms in late 2002, preparing the Wilde Lectures. He has a Buddhist sūtra open on his desk. Next to it: a photograph of a Bharhut stūpa railing, an inscription from a Pāla land-grant, an archaeological report from Chandraketugarh, an ethnography of a Tibetan village he lived in twenty years earlier, a paper on the C14 calibration curve for the first millennium BCE.
He is trying to figure out where yoga came from.
The standard move in Indology would be to start with the texts — read the Ṛgveda, then the Brāhmaṇas, then the Upaniṣads, then Patañjali, and trace a line. Geoffrey Samuel refuses that move. He thinks the texts are necessary but radically insufficient. The Vedas tell you what a tiny group of priestly authors thought was worth recording, not what religion looked like for the people who lived it. The archaeology tells you about pottery and walls but not about what was felt at the moment a child was named or a corpse was burned. The iconography tells you what was depicted but not what those depictions meant to the worshippers in front of them. Each evidence-stream is partial; each is essential.
His commitment, stated bluntly: "A religion is not just a body of texts. It is, above all, something that lives and is maintained through the lives of human beings."1
This is the methodological gate that everything else in The Origins of Yoga and Tantra (2008) walks through. It governs which questions Samuel asks, which evidence he weighs, which conclusions he reaches, and — crucially for the vault — which claims by practitioner-lineage tradition he treats as historically verifiable versus historically unverifiable. To engage Samuel's substantive content well, the method has to be in view.
Critical-anthropological method, as Samuel practices it, has five operative commitments.
First: multi-source reading. Texts read alongside epigraphy alongside iconography alongside archaeology alongside ethnography alongside linguistic reconstruction. No single evidence-stream is decisive. The job is to find a picture that fits all the evidence reasonably well, rather than to privilege one stream and torture the others to match. This is harder than it sounds. Most Indologists are textual specialists, comfortable in Sanskrit but not in archaeology; most archaeologists handle pottery sequences but not late Vedic syntax. Samuel works to integrate.
Second: reading practices into life-worlds. Samuel will not analyze a text — say, an Upaniṣadic passage on sacrifice — without trying to reconstruct who wrote it, who heard it, what economic and political pressures they were under, what their daily ritual practice looked like, what the surrounding folk religion was doing. The text-as-utterance only makes sense within the text-producer's life-world. "Ultimately, it is people and their specific life-worlds that we are attempting to understand."2
Third: Weberian elective affinity rather than linear causation. When Samuel asks why śramaṇa orders developed in the fifth century BCE, he does not look for a single cause. He looks for the convergence of conditions — new urban environment, new merchant class, new self-awareness in individuals navigating uncertainty, new state-formation needs, available models from existing tradition — that made the development meaningful and viable for all the parties involved (ascetics, lay followers, rulers). Causation in social analysis is multi-actor, non-linear, and rarely amenable to clean falsification.3
Fourth: skepticism toward strict orthogenetic textual approaches. "Orthogenetic" means tracing later developments to earlier texts (later Vedic → Upaniṣads → Patañjali → modern yoga as a smooth line). Samuel argues this works for some elements but systematically distorts where the actual energy of religious development was coming from. Much of what became important in Indic religion did not derive from the Ṛgveda — it derived from Central Gangetic agrarian populations, yakṣa cults, śramaṇa innovations, Tantric appropriations of pre-Brahmanical material, regional goddess cults. The orthogenetic frame makes Brahmanical-textual material look more central than it actually was for most of the population most of the time.
Fifth: religion as ways-of-living-in-relation-to-the-universe, not religion as belief-system. This is the deepest commitment. Christianity in its Protestant form (which shaped 19th-c. Western religious studies) defined religion through belief-statements held to be true. But asking a Sri Lankan Buddhist villager "what do you believe?" misses the point — Buddhism for that person is a way of living in relation to monks, family, ancestors, local deities, ethical practice, ritual calendar. The Western frame distorts what it claims to describe.
The critical-anthropological method is not arbitrary preference. It emerges from a specific diagnosis of how Indology has gone wrong.
The 19th-c. construction of "Indic religion" was a collaboration between Western scholars (mostly Protestant Christian in formation) and Asian reformers (especially the Brahmo Samaj on the Hindu side, the Mahabodhi Society on the Buddhist side). Both sides had stakes in producing a sanitized, intellectually elevated, text-centered account that would be respectable in European terms. The result was the "standard story" Samuel critiques in Chapter 2: Hinduism as a development from the Vedas; Buddhism and Jainism as reactions against it; Tantra as awkward late corruption; Theravāda Buddhism as pure original Buddhism, with everything else as decay.
This story made some people comfortable. It made the texts look orderly. It made the lineages look continuous. It made the scholarly access seem complete. It also bore little resemblance to actual religious life in South Asia.
Samuel's method is the way out. By insisting on multi-source reading, life-world reconstruction, elective affinity, anti-orthogenetic skepticism, and ways-of-living definition, he can ask new questions: What was happening in the population while the priests were composing the Vedas? Why were Buddhist monks settling at sites associated with the cult of the dead? Why did Tantric ritual emerge when and where it did? What did kingship have to do with deity-conception? Where did yoga actually come from?
The method is the price of admission to those questions.
The critical-anthropological method threads through every Samuel page in the vault and gives them their distinctive epistemic shape.
It connects directly to Protestant Template Distortion of Asian Religions — the methodological commitment is what enables the diagnostic. It connects to Orthogenetic vs Heterogenetic Indic Religious History — the choice to read against strict orthogenesis is the operative consequence. It connects to Chronological Uncertainty of Indic Religious Texts — the willingness to say "we do not know when this was written" comes from refusing to privilege textual self-presentation.
Across the vault, the method also lights up a productive disagreement with practitioner-lineage epistemology. The Sanjay-Rath direct-disciple lineage represented by Foss's Sanjay-Rath-Direct-Disciple Lineage Positioning operates on a transmission-canon principle: this teaching has been passed down through unbroken lineage from named teacher to named teacher, and that transmission is the warrant. Samuel's method operates on a critical-historical principle: the evidence we have for any chronological claim is the evidence we have, and the absence of independent corroboration limits what we can assert. These are not opposed claims about the same question. They are two different epistemic vantages on what counts as warrant for religious knowledge. Both can be true at once.
The clearest illustration of Samuel's method in operation is his treatment of the famous Mohenjo-Daro M-304 "Proto-Śiva" seal.
The seal shows a horned, possibly three-headed figure seated in what John Marshall (1931) called a yogic posture, surrounded by animals. Marshall called this proto-Śiva — Lord of the Beasts (Paśupati). The interpretation has been repeated for nearly a century as evidence that yoga originates in the Indus Valley (2600-1900 BCE).
A textual-tradition scholar might either accept Marshall's reading (the orthogenetic move: trace yoga to the deep Indic past) or politely sidestep it. Samuel does neither.
Instead, he lines up six competing interpretations of the same image. Hiltebeitel reads the head as a buffalo, the surrounding animals as Vedic gods of the four directions. The Allchins read the figure as ithyphallic. Sullivan and Atre argue it is a female deity, not a male one. Dhyansky and McEvilley read the posture as mūlābandhāsana, an advanced Tantric posture. Parpola suggests the posture imitates Proto-Elamite seated bulls. Each interpreter brings their training, their textual repertoire, and their cultural assumptions to a visually ambiguous image. Each reading is plausible on the evidence available. None can be confirmed.
Samuel's conclusion: "The only reasonable conclusion is that we do not actually know how to interpret the figure, nor do we know what he or she represents."4
This is the method working. He has read what each interpreter wrote. He has examined the image. He has asked what evidence would discriminate between interpretations. He has found that the evidence does not discriminate. He stops there, rather than picking a winner.
The consequence is important for the vault. Several vault pages on yoga history, Tantra origins, and Indic deity-archetype assume an unbroken thread from Indus Valley imagery to later practice. Samuel's reading does not foreclose that possibility — he agrees continuity may exist — but it refuses to use the proto-Śiva seal as evidence for it. The interpretive load that seal has carried for a century turns out to be load it cannot actually support.
It is 9:14 in the morning. You are about to write a concept page on the historical origins of a yogic practice. Open at hand: a passage from a Sanskrit text, a translation by a 19th-c. Indologist, a practitioner-lineage commentary, an archaeological article on a relevant site, an ethnographic study of a contemporary community where the practice is still performed.
You read the Sanskrit text first. You make a note of what the text claims: this practice was given by such-and-such a sage in such-and-such a year.
You stop. You do not transcribe the claim into your page yet. You ask: what would I have to know to evaluate this claim?
You read the 19th-c. translation. The translator is a Protestant clergyman writing in 1872. He has translated three key Sanskrit terms with English Christian equivalents that subtly distort their semantic range. You notice and flag this. You do not throw the translation away — it is still useful, partially. You note where the distortion lies.
You read the practitioner-lineage commentary. The commentator is a 21st-c. teacher in direct lineage from a 15th-c. master. The commentary asserts an unbroken transmission. You take the claim seriously as a claim about this lineage's understanding of its teaching, without yet accepting the implied historical chronology.
You read the archaeological article. The site dates to roughly the right period. The material culture is consistent with the text's geographic context but does not specifically attest the practice in question.
You read the ethnography. The contemporary community performs a practice that looks visually similar to what the Sanskrit text describes. The ethnographer notes that the community traces its practice to a 17th-c. teacher — not to the Sanskrit author the text names.
You sit with all of this for ten minutes.
What you can confidently write: a description of what the Sanskrit text says, with the text's date roughly identifiable from internal evidence within a couple-of-centuries window. A note that the practice has continuous attestation from at least a known later period. A note that the lineage-tradition holds an earlier origin which is not independently corroborated by archaeology or ethnography. A note that the practice today exists in a form whose continuity with the Sanskrit text is plausible but unverifiable.
What you should not write: that the practice was given by the sage in the year the lineage-tradition asserts. That the practice has been performed continuously since that time. That the modern form is identical to the original.
The discipline is to write what you can warrant from the evidence, mark where warrant fails, and preserve the lineage-tradition claim as a claim of that lineage rather than a claim about history.
Samuel's method represents one of three currently-operative epistemic positions on how to think about Indic religious history.
Witzel and the Vedic textual scholars represent a second position: textual evidence read with the highest available philological sophistication, with archaeology and ethnography brought in primarily where they support textual reconstruction. Witzel's "tape recording" claim — that present-day Ṛgveda recitation is essentially the same text as composed three thousand years ago5 — would be unacceptable in Samuel's frame because it privileges textual self-presentation over the kinds of corroboration Samuel demands.
Practitioner-lineage scholars like Foss, Larsen, Pai, DiCara, and Rao represent a third position: the lineage transmission is the warrant. The Sanjay-Rath lineage's claim that Vimshottari is ancient and was transmitted through specific named teachers is treated as historically reliable because the lineage stands behind it. This works for the lineage's own internal purposes and offers something academic critical-history cannot — the warrant of unbroken practice.
The convergence among all three is that texts are real, transmission is real, and Indic religious life has continuity worth tracing. The tensions are about what counts as warrant for a claim and how to handle the inevitable gaps in the evidence. Each position is correct within its own epistemic vantage. The vault preserves all three because the productive disagreement among them produces an insight none generates alone: religious knowledge has multiple kinds of warrant, and the same fact can be true under one warrant-frame while remaining unverifiable under another.
The vault page on Foss's Mantra-Primary Remedy Hierarchy vs Other Vault Lineages already preserves a Tier-1 disagreement among four practitioner-lineages. Samuel adds a Tier-1 disagreement at the meta-level, between practitioner-lineage warrant and critical-historical warrant. Both kinds of disagreement deserve preservation.
Samuel's critical-anthropological method generates structural insights that travel well beyond the eastern-spirituality domain.
The handshake to history is direct. Critical-anthropological method overlaps significantly with the historiographic methods used in Weatherford's Weatherford Five-Discipline Research Method for the Genghis Khan corpus and in Kershaw's "working toward the Führer" reconstruction of Nazi mass-mobilization in the vault history domain. All three involve multi-source reading, reconstruction of life-worlds, and refusal to privilege any single evidence-stream. The insight neither domain generates alone: the same methodological commitments that produce reliable history of distant Indic religion also produce reliable history of 20th-c. political mass-movements; the method is portable because the underlying epistemic problem (reconstructing distant or alien life-worlds from partial evidence) is the same.
The handshake to psychology is more subtle but real. The vault holds extensive material on depth-psychology mechanisms (projection, dissociation, etc.) and on object-relations work on encompassing-the-good-and-bad-parent. Samuel's diagnosis that "Indic deities encompass the bad father or mother as well as the good father or mother" is itself a psychological observation produced by anthropological method. The insight neither domain generates alone: the cultural specificity of Western religion's good/evil dichotomization vs Indic religion's encompassment is observable to anthropology but its psychological consequence (capacity to integrate split-off material) is observable to psychology. Samuel uses anthropological method to identify what depth-psychology then has to explain. The two are mutually reinforcing rather than competing.
A third handshake to behavioral-mechanics sits in the background. Critical-anthropological method shares with practitioner influence corpora like Hughes's Six-Minute X-Ray a commitment to reading behavior in its full context rather than as decontextualized signal. Hughes's elicitation method, like Samuel's method, refuses to interpret a single utterance without reconstructing the speaker's situational and motivational frame. The insight neither domain generates alone: Hughes is operating in real-time on living conversational partners; Samuel is operating retrospectively on partially-recovered ancient life-worlds. The two are scale-different versions of the same core epistemic stance — that behavior or text means what it means in context, and removing the context to extract the signal is the methodological error that produces almost all bad readings.
The method has characteristic ways of going wrong. Recognizing them is part of using it well.
Over-skeptical paralysis. If multi-source reading is taken to mean "no claim can be made without corroboration from all evidence-streams," nothing can be said about anything before about 500 CE. Samuel does not slip into this — he is willing to make claims that the evidence supports while flagging where it does not — but a less-disciplined practitioner of the same method can produce a paralyzed scholarship that refuses to commit. A page that ends with "we cannot know" on every question is a failure of the method, not a faithful execution of it.
Smuggled orthogenesis. A scholar can profess anti-orthogenetic principles and then quietly assume a continuous chain of transmission anyway, because the alternative is too unsettling. Samuel mostly avoids this but occasionally slips — for example, when he allows that "elements of these practices going back to the Vedic culture of Kuru-Pañcāla" without quite acknowledging that "going back to" reinscribes the orthogenetic frame he has critiqued. Watch for this in your own writing: did you slip a chain-of-transmission claim into a sentence whose evidence does not support it?
Anthropological reductionism. The opposite failure mode. If the method becomes "religion is only social practice, the texts and the inner experiences are epiphenomena," it cannot account for what practitioners say they are doing. Samuel is explicit that internal-experiential perspective and external-anthropological perspective both have truth-content and must be integrated. The reduction to either side alone is a methodological collapse.
Confusing critical-historical warrant with practitioner-lineage warrant. This is the one most likely to affect vault use. If you read Samuel and conclude that the practitioner-lineage Sanjay-Rath corpus is "wrong" because Samuel could not historically verify its antiquity claims, you have misunderstood the epistemic structure. Samuel produces historical-critical claims. Practitioner-lineage produces transmission-warranted practice claims. These are not in competition for the same epistemic slot. They occupy different slots. Both can be load-bearing in different contexts.
The strongest evidence for the method's productivity is the result it generates: a coherent account of Indic religious history from 500 BCE to 1200 CE that handles multi-source evidence integratively, identifies the two-worlds structure that makes the emergence of śramaṇa orders intelligible, identifies the sāmantisation pattern that makes the rise of state-Tantric ritual intelligible, identifies the kula→kaula→Phase-3 trajectory that makes the privatisation of Tantra intelligible. These accounts are difficult or impossible to reach via textual method alone.
The strongest evidence against an overly-strong reading of the method is that on key questions, multi-source reading does not actually converge. Samuel himself acknowledges that we cannot reach confident conclusions on Indus Valley religion, on early Indo-Aryan migrations, on the dating of major Brahmanical texts, on the historicity of Pārśva. The method can identify what we do not know; it cannot manufacture knowledge where the evidence is absent.
Open questions worth flagging: How do we systematically handle the productive disagreement between critical-historical warrant and practitioner-lineage warrant when both bear on the same vault concept? Is there a vault-pith convention for marking which warrant-frame a claim is operating under? Should pages have explicit warrant-frame frontmatter? These questions are not in Samuel's book — they emerge from the work of integrating Samuel into a vault that already holds extensive practitioner-lineage content.
Sharpest implication: When the vault carries practitioner-lineage chronology claims at Tier 1 and Samuel's chronological-uncertainty claims at Tier 1 about the same object, both are operating correctly within their warrant-frames. The vault's job is not to pick one. The vault's job is to preserve the productive disagreement-architecture between them. Future synthesis pages should explicitly name the warrant-frame each claim derives from.
Generative questions: