It is 1924. Maurice Bloomfield, professor of Sanskrit at Johns Hopkins, publishes a long article in the Journal of the American Oriental Society titled "On False Ascetics and Nuns in Hindu Fiction." Bloomfield has been working through the medieval Sanskrit story-corpus — the Kathāsaritsāgara, the Bṛhatkathā, the Pañcatantra, the Vetālapañcaviṃśati — and he has noticed a recurring narrative type. In story after story, a yogi appears who is not what he seems. He pretends to be a holy ascetic. He insinuates himself into a household, a court, a temple. He steals children. He steals women. He steals the body of the king. He performs parakāya-praveśa — entering another's body — for sinister ends.
Bloomfield's reading: this sinister yogi is a literary stock figure. He recurs because narratives need villains, and the ascetic-yogi-villain is one of the medieval Sanskrit imagination's most reliable villain-types. The figure is fictional. Stories use him the way Western stories use the wicked stepmother or the evil sorcerer. He is a narrative trope, not an ethnographic reality. The medieval Indian villagers who supposedly feared yogis were participating in a literary tradition, not reporting an actual social phenomenon.1
It is 2006. Frederick M. Smith, professor of Sanskrit and South Asian Studies at the University of Iowa, publishes The Self Possessed, an enormous study of possession in South Asian religious traditions. Smith treats Bloomfield's literary-stock thesis with care but pulls back from it. The sinister yogi in medieval Sanskrit literature is not just literary. Smith reads the body-stealing narratives alongside epigraphic, ritual, and ethnographic evidence and concludes that the literary type corresponds to documented social practices. The literary tradition is not invention. It is reflection, however stylized, of operative possession-traditions that the Sanskrit story-corpus encodes in particular narrative forms. Smith does not go all the way to claiming the literary yogi is a historical person. But he refuses to dismiss the literary corpus as pure fiction.2
It is 2009. David Gordon White publishes Sinister Yogis and goes all the way. The sinister yogi of medieval Sanskrit literature is a historical reality, ethnographically attested across multiple sources — Mughal-period European traveler accounts, colonial-era administrative reports, contemporary North Indian village ethnography, primary-source Sanskrit and vernacular literature read against grain. White marshals the evidence cumulatively. The yogi who steals bodies, possesses infants, recalls the dead, takes over kings — exists in history, not just in literature. Bloomfield was wrong. The literary-stock thesis was a Victorian-style scholarly maneuver to protect Indian religious history from its weirder operative dimensions by routing them into literary-trope categories.3
This isn't a niche philological argument. The Bloomfield-Smith-White debate is a test case for how Sanskrit scholarship reads its primary sources. The question is: when medieval Sanskrit literature describes a phenomenon repeatedly, in detail, across genres, across centuries — what is the appropriate scholarly default? Default-fictional: the phenomenon is a literary type until proven historical. Default-historical: the phenomenon is reported reality until proven fictional. Default-mixed: the literary corpus reflects social reality through narrative stylization, and the scholar's job is to extract the social signal from the literary form.
Bloomfield, working in the high-Victorian-philological tradition, defaults fictional. His move is to treat Sanskrit literature as a self-contained literary system whose recurring types tell us about the system's narrative grammar, not about Indian society. This is a safe scholarly move. It protects the philologist from the embarrassment of reading credulously, of treating medieval claims about supernatural-yogic powers as historical reports.
White defaults historical. His move is to treat Sanskrit literature as ethnographically informative — as one source among many for reconstructing what medieval and pre-modern South Asians actually did. The literary stylization is real but recoverable; underneath the stylization is reportage of practices, social positions, ritual specialists. The risk of this move is over-reading — treating every literary description as historical fact. The reward is that if the historical phenomena were real, the literary corpus is one of our richest sources for them, and the Bloomfield-default has been systematically blinding generations of scholars to that source.
Smith defaults mixed. He treats the literary corpus as informative but stylized: the social phenomena are real (possession traditions, ritual specialists, marginalized practitioners), but the narrative form reorganizes and dramatizes them. The scholar's task is triangulation — reading literary, epigraphic, ethnographic, and ritual evidence together to recover the operative practice that all the source-types refract differently.4
The debate matters because it determines what the Sanskrit corpus can be. If Bloomfield is right, the corpus tells us about literature. If White is right, the corpus tells us about a documented yogic operative tradition. The same texts read under different default assumptions yield radically different historical pictures.
White's case against Bloomfield is cumulative — no single piece of evidence is decisive; the weight of the case comes from accumulation across source-types. Reading the case requires patience because each source-type, alone, can be dismissed.
The literary corpus itself — Kathāsaritsāgara, Vetālapañcaviṃśati, Bṛhatkathā — recurs to the sinister-yogi figure across centuries and regions. Bloomfield is right that the figure is a narrative type; White's response is that types this persistent, this widely distributed, with this much operational specificity, are unlikely to be pure invention. Other types in the Sanskrit corpus (the wicked merchant, the lustful widow) are clearly stylized but correspond to recognizable social positions; why would the sinister yogi be different?
The European traveler accounts — Marco Polo (1290s), Ludovico di Varthema (1503-08), Duarte Barbosa (early 1500s), Pietro della Valle (1620s), François Bernier (1660s), Jean-Baptiste Tavernier (1670s), Niccolao Manucci (late 1600s) — describe yogis encountered in person across the Indian subcontinent over four centuries. The travelers report on yogi-orders, yogi-economic functions, yogi-political relations with regional courts, yogi-ritual practices including body-magic and corpse-handling. These are firsthand European reports — sometimes credulous, sometimes skeptical, but largely independent of the Sanskrit literary tradition. They corroborate the literary picture from outside.5
The colonial administrative record — especially the British East India Company's 1773 Hastings-era criminalization of fakir and sannyāsī orders — treats yogi-groups as organized military-economic actors whose suppression was a colonial-administrative priority. You do not criminalize a literary trope; you criminalize a social force.
The contemporary ethnography of North Indian village religion — fieldwork by Crooke (1896), Briggs (1938), and more recent scholars — documents yogi-figures who are feared, who are believed to steal souls, who are credited with body-magic. The literary-figure has a living analog in the ethnographic present.
The vernacular oral traditions — Bhojpuri yogi-songs, Hindi bharthari-gopcand romances, Sufi-influenced jogī poetry — preserve yogi-narratives in non-Sanskrit registers that cannot be dismissed as artifacts of Sanskrit literary convention. These traditions encode the same operative-yogi under different aesthetic forms.
Taken together, White argues, the evidence-base supports treating the sinister yogi as a real historical category. The literary form is one source among many; the ethnographic, epigraphic, and administrative sources independently corroborate the literary picture.6
Bloomfield wasn't being unscholarly. He was working within the methodological constraints of 1924 philology — heavy emphasis on text-internal analysis, suspicion of "popular" or "vernacular" sources, distance from ethnographic fieldwork, an inherited Protestant-template orientation that treated religion as belief-doctrine and read everything else as cultural-literary residue. Under those constraints, the sinister yogi as recurring literary type was the natural reading. The fieldwork that would later corroborate the literary picture had not been done yet, and the disciplinary apparatus did not yet integrate text-philology with ethnography.
White's reading is possible now because eight decades of additional scholarship — colonial-archive analysis, North Indian ethnography, Mughal-period European-source studies, vernacular-literature work — have built the corroborative context Bloomfield lacked. White is not smarter than Bloomfield; he has access to an evidence-base Bloomfield did not have. The reading is historically possible in 2009 in a way it was not in 1924.
This matters because it reframes Bloomfield not as wrong-because-confused but as correct-given-his-tools-and-incorrect-now-that-we-have-better-tools. The debate isn't about competing intuitions. It is about what counts as evidence and how source-types integrate. White's win, if it is a win, is the win of a methodological consolidation that took eighty-five years to assemble.7
You are reading a translation of the Kathāsaritsāgara — Somadeva's eleventh-century Sanskrit story-collection. You hit one of the sinister-yogi narratives. A yogi has been frequenting a cremation ground. He performs parakāya-praveśa. He steals the body of a young king. The legitimate king's spirit is now in the yogi's old body — sometimes in the body of a parrot, sometimes a deer. The story-cycle takes off from there.
The Bloomfield-trained reader thinks: literary trope. The cremation-ground yogi is a stock villain; the body-theft is a narrative engine; the parrot-and-deer is fairy-tale typology. The story is fun, important for the literary tradition, but tells you nothing about whether actual yogis did anything like this.
The White-trained reader thinks: recoverable historical practice. The cremation-ground yogi is a real ritual specialist: medieval South Asia had a class of practitioners who frequented cremation grounds, who claimed body-magical powers, who were associated with marginal-yogi orders. The parakāya-praveśa is operative: medieval texts describe it in technical detail, ritual manuals catalog the procedure, European travelers report claims of it, modern ethnography documents fear of it. The parrot-and-deer is narrative stylization, but the body-theft itself is reportage of a claimed practice. The story is partly a literary type, partly a stylized report on operative-yogic activity.
You read with both readings live. The Bloomfield-reading gives you the literary grammar — what the story is doing as story. The White-reading gives you the social-historical signal — what the story is reporting about practitioners and their claims. Both readings are present. Neither is sufficient alone. The integrated reading is the productive one.
A month later you encounter a Mughal-period European traveler account — Bernier, perhaps, on the yogis of Delhi. The traveler describes yogi-claims of body-transmigration in language strikingly similar to Kathāsaritsāgara narratives. Now the cross-source resonance lands. Two independent source-types (Sanskrit literary, European travelogue) describe the same operative-yogic claim in different registers. The Bloomfield-thesis cannot account for this resonance without ad-hoc moves. The White-thesis accounts for it directly: both sources reflect a real yogic-operative tradition observable from inside (Sanskrit) and outside (European).
You are reading a yoga-history book. The author treats every non-meditative yogi practice as "later folk accretion" or "literary type" or "popular distortion." The yogic powers in Vibhūti-pāda are metaphor. The Vikrama Cycle is literary tradition. The Nāth Yogī mercenary networks are "social bandits" who happened to call themselves yogis. The European traveler accounts are credulous miscomprehension. The ethnographic fieldwork is local superstition unrelated to the philosophical yoga tradition.
This is the diagnostic. You have inherited the Bloomfield-default. Every source-type that complicates the meditative-philosophical picture of yoga has been routed into a not-really-yoga category, and the routing has been thorough enough that the residual category "yoga" is now coextensive with the meditative-philosophical strand. The book reads cleanly because everything that doesn't fit has been disposed of.
The corrective is to re-admit the routed material. Read the Vikrama Cycle as social-historical source. Read the European travel accounts as ethnographic source. Read the colonial-administrative record as institutional source. Read the contemporary village ethnography as a continuous historical thread. The book that emerges from the re-admission is a different book — White's book, or a book like it.
The principal tension is methodological: the debate is not finally about evidence but about what counts as evidence. Bloomfield's default-fictional and White's default-historical are interpretive frames that select different evidence as primary. A more cautious scholar might insist that the debate cannot be resolved by accumulating evidence alone — it requires justifying the interpretive frame in advance.
A second tension: even granting that the sinister yogi is historically real, the degree of correspondence between literary representation and operative reality is unclear. Did medieval yogis actually claim to perform parakāya-praveśa? Almost certainly — the claim is in primary-source texts and contemporary ethnography. Did parakāya-praveśa actually work in any sense? That is a different question, and White is wisely silent on it. The historical claim is about the practice and its social position, not about its metaphysical effects.
Open question: where else in Indic religious history does the Bloomfield-default operate? White's case is specifically about the sinister yogi, but the methodological pattern recurs. Are there other phenomena (yoginī cult? warrior-yogi orders? Tantric body-substance ritual?) where contemporary scholarship still defaults-fictional and would benefit from a White-style re-evaluation?
Bloomfield, Smith, and White represent three positions on a methodological spectrum. The convergence is on the fact that the Sanskrit literary corpus contains a recurring sinister-yogi figure. The tension is on what that figure refers to. Bloomfield treats it as referring to literary tradition. Smith treats it as referring to stylized social phenomena. White treats it as referring to operative historical practice. The three positions are not just chronologically successive — they are methodologically distinct defaults that could each be defended in principle, even if the cumulative evidence pushes the field toward the White-default.
In a different convergence, the debate parallels disputes in adjacent disciplines. Was the historical Jesus more than a literary type? The decades-long historical-Jesus debate runs on the same methodological structure: how much does literary representation tell us about historical referent, and what counts as cross-corroboration? Were the Greek mystery cults real esoteric practices or literary tropes about them? Same structure. The Bloomfield-Smith-White debate is one instance of a general methodological problem in religious studies and historical anthropology, and recognizing the pattern lets us see that the debate's stakes extend beyond Indic yoga to historical method as such.8
The Bloomfield-default — if a recurring representation cannot be cross-verified through non-textual sources, treat it as literary type rather than historical reality — is a methodological move that occurs across history, religious studies, and ethnographic anthropology. Recognizing it is a transferable skill.
History: Protestant Template Distortion — Samuel's diagnostic of how Protestant-template scholarship narrowed religion runs on the same methodological move that produced Bloomfield's default. The Protestant template treated belief-doctrine as the religion and ritual-operative material as cultural-residual; Bloomfield treated philosophical-meditative yoga as the yoga and literary-operative material as fictional-residual. Both moves are methodological narrowings that route inconvenient material out of the central category. The cross-domain insight: methodological narrowings tend to recur across fields under different surface vocabularies; learning to recognize the move in one field improves the ability to recognize it in others.
Behavioral mechanics: Strategic-influence scholarship sometimes encounters the same default. Are the "power-moves" in Greene's 48 Laws of Power historical realities or literary tropes constructed from selectively-read primary sources? Greene-critics often deploy the Bloomfield-default: the power-moves are literary types, not operational realities. Greene-defenders deploy something like the White-default: the power-moves are operationally real and the literary stylization reports them in concentrated narrative form. The same methodological dispute — about whether recurring representations in canonical texts correspond to operative reality — plays out in influence-scholarship under different vocabulary. The cross-domain insight: the literary-vs-real debate is the same debate wherever a canonical-text tradition encodes practical material in narrative form.
The Sharpest Implication: If White is right that Bloomfield's default-fictional has been operating quietly across yoga-historiography for a century, then a systematic re-evaluation of every yogi-related Sanskrit narrative is warranted. The Vikrama Cycle is one place where the re-evaluation is now well underway. But other corpora — the Mahābhārata yogi-narratives, the Yogavāsiṣṭha siddha-stories, the Nāth hagiographies — have been read under the Bloomfield-default for generations and have not yet received systematic White-style re-evaluation. There is more historical-yogi-tradition to recover than has yet been recovered. The recovery is ongoing.
Generative Questions: