Open almost any introductory book on ancient India and you'll find the same picture, stated with total confidence. Here's the historian Gavin Flood's version, typical of the genre: "Vedic society... was divided into four classes (varnas), the Brahmans, the Nobles or Warriors (rajanya, kshatriya), the Commoners (vaishya) and Serfs (shudra)."1
Four boxes, hierarchically ranked, foundational to how "ancient Indian society" gets described in textbook after textbook, museum placard after museum placard, popular-history book after popular-history book. It's one of those pieces of received knowledge confident enough that most readers never think to ask where, exactly, it comes from — the fourfold picture just arrives, pre-assembled, the way "ancient Rome had senators" arrives.
That confidence is worth noticing before getting to the contrary evidence, because the size of the gap between "universally taught" and "actually attested" is part of what makes this page's finding land as more than a minor scholarly footnote.
Now go read the largest surviving body of contemporary evidence anyone has for exactly this period — the actual inscriptions of the man who ruled most of the Indian subcontinent in the third century BCE, extensively documenting social relationships, obligations, obedience, and hierarchy. The word varna does not appear. Not once. Neither does kshatriya. Neither does vaishya. Neither does shudra.
The entire fourfold structure that every textbook tells you organized ancient Indian society is completely, totally absent from the one contemporary voice we actually have.1
Olivelle's method here (see Olivelle's Portrait-Not-Biography Method) is to take the silence seriously as data rather than dismissing it as coincidence. Ashoka discusses social relationships extensively — his conception of dharma is built entirely around a structured web of relational obligations (see the forthcoming Dharma Pentagon page).
If the varna system genuinely organized how people in his empire thought about their social position, it would be strange for a king this preoccupied with proper relational conduct to never once invoke it, even in passing, even accidentally.
The one member of the fourfold system Ashoka does mention is the Brahmin — but crucially, never as part of the varna hierarchy. He appears exclusively in the context of religious organizations, the Pasandas, always paired with the Sramanas (wandering ascetics) as one of two categories of religious professional.
Olivelle's key observation: "If the varna system was internalized by the population or the ruling elite, or if it was a way for people to think about social roles and hierarchies, then it would have been natural for Ashoka, at least sometimes, to pair Brahmin with the other three varnas, something he never does. It is always the Sramana who is paired with the Brahmin."1
The pairing pattern is the tell. A king who thought in varna terms would occasionally, even unconsciously, slot the Brahmin into that fourfold frame. Ashoka's Brahmin never appears there — only ever alongside a category (Sramana) that has nothing to do with the varna hierarchy at all. That's the difference between an argument from silence and an argument from substitution: Ashoka doesn't just fail to mention varna, he consistently reaches for a different, non-hierarchical pairing every single time the topic of religious professionals comes up.
Olivelle doesn't stop at documenting the silence — he asks who benefits from the picture the silence complicates. "We should begin to question whether the sources that present this fourfold social division, sources that are the basis for contemporary textbook descriptions of ancient Indian society, are reliable. They are, for the most part, authored by Brahmins, who have historically been placed at the summit of this fourfold hierarchical structure. Are they presenting an ideology garbed in the clothes of sociology? Is the varna system aspirational and prescriptive rather than descriptive?"1
This is a genuinely pointed methodological reframing. The textbook picture of ancient Indian society as varna-organized comes almost entirely from Brahmanical textual sources — texts written by the very group the system places at the top. Olivelle's suggestion is that historians have been treating a self-interested aspiration (this is how society should be organized, with us at the summit) as though it were a neutral description (this is how society actually was organized).
Ashoka's silence — a contemporary, non-Brahmin-authored, empire-wide administrative and moral document with every reason to reference social structure if it mattered to governance — becomes a rare independent check on a historical picture that has otherwise gone almost entirely unchallenged by contemporary evidence, because almost no other contemporary evidence from this period survives at all.
The strongest move Olivelle makes is refusing to let this rest on Ashoka's silence alone. He brings in the historian Johannes Bronkhorst's independent finding: the varna-absence isn't unique to Ashoka's corpus — it's a pattern across an entire category of evidence.
"The brahmanical vision of society is largely absent in South Asian inscriptions that are not in Sanskrit and whose makers or instigators have no association with Brahmanism. It is absent from the inscriptions of Ashoka. They refer to none of the four varnas except the Brahmins, nor to the system as a whole. The same is true of the early Tamil inscriptions."1
This converts a single suggestive silence into a documented pattern. Every pre-2nd-century-CE inscription not produced by or for Brahmanical interests — Ashoka's included, early Tamil inscriptions included — is silent on the same fourfold structure.
If varna were simply the operative social reality of the period, this pattern of independent silence across unrelated inscriptional traditions would be a remarkable coincidence. Read the other way, the pattern makes far more sense: varna, in this period, was a Brahmanical textual-ideological project, actively promoted in Brahmanical literature, but not yet — or not evenly — the lived, self-describing social reality of the wider population whose own inscriptions simply don't reach for the concept.
Consider how far apart Ashoka's Magadha-centered court and the early Tamil-speaking south actually sit — different language family entirely, different political tradition, no direct evidence of shared court culture at the time these inscriptions were produced. Two evidentiary streams this unconnected converging on the identical silence is a much harder pattern to explain by coincidence than either silence would be alone.
Push on that pairing pattern harder than a single observation allows, because "it is always the Sramana who is paired with the Brahmin" is a claim about consistency across an entire multi-decade corpus, and consistency claims deserve to be examined for how they'd actually fail if they were wrong.
Think through what the alternative would look like. If Ashoka genuinely carried some latent varna framework, even unconsciously, you'd expect at least occasional slippage — a passage where he lists Brahmin alongside kshatriya or vaishya in some other context, a moment where the fourfold structure leaks through even without being named directly, the way a person raised inside a framework sometimes reveals its shape through offhand phrasing even when they're not consciously invoking it. Olivelle's claim is that this leakage never happens. Every single occurrence of Brahmin in the corpus sits inside the same one relationship: paired with Sramana, inside the Pasanda religious-organization frame.
That's a stronger evidentiary shape than a single well-chosen quote can convey. A pattern that holds with zero exceptions across an entire reign's worth of inscriptions, touching religious policy repeatedly across multiple edicts and multiple years, is much harder to explain as coincidence than a pattern that holds "usually" or "mostly." The absence of counter-examples is doing real work here — it's the kind of negative evidence that would be easy for a less careful reading to miss, because there's no single dramatic quote that proves a zero-exceptions pattern. It has to be established by checking every occurrence and confirming none of them break the rule.
This is also where the case for aspirational-rather-than-descriptive varna gets its sharpest edge. A social category doesn't just have to be invisible in the positive sense (never mentioned) — it has to be invisible in this specific structural sense (never even slipping into the vocabulary someone would use for genuinely adjacent purposes) before the silence stops looking like an accident of genre and starts looking like evidence about what categories were actually live in the culture producing the text.
Step back from Ashoka specifically and ask the more general question Olivelle's reframing implicitly raises: how does a hierarchy description written by the group that benefits most from it end up treated by later historians as a neutral sociological fact rather than as evidence of that group's own interests?
Part of the answer is simple asymmetry of survival. Brahmanical textual production was systematic, self-reinforcing, and institutionally supported across many centuries — dharma texts, legal treatises, ritual manuals, all describing and re-describing the same fourfold structure with mutual cross-reference.
Non-Brahmanical evidence from the same period, like Ashoka's inscriptions or the early Tamil corpus, survives in isolated, non-cumulative fragments — a single reign's edicts, a scattered handful of early inscriptions, nothing like the self-reinforcing textual ecosystem the Brahmanical tradition built for itself.
When later historians went looking for a picture of ancient Indian society, the Brahmanical textual tradition offered something the other evidence didn't: systematic completeness. A confident, internally consistent, endlessly cross-referenced description is simply easier to build a textbook chapter around than a scattered, silence-heavy epigraphic record that mostly tells you what a period's evidence doesn't say. The volume and coherence of the self-interested description didn't make it more true. It made it more usable, and usability quietly got mistaken for reliability across generations of historical writing until scholars like Olivelle and Bronkhorst started asking who actually produced the picture and why.
Olivelle is disciplined here, and this discipline should be preserved whenever this page is cited: "Simple silence, the absence of any reference, cannot by itself prove that the institution of varna did not exist during the time of Ashoka."1
This is not a claim that varna is a total fiction, invented wholesale by later Brahmanical writers with no third-century-BCE referent at all. It is a narrower and more defensible claim: the evidence quality behind the standard textbook picture is weaker than its confident, unhedged presentation suggests, and a genuinely contemporary, non-Brahmanical-authored source that had every occasion to invoke the system simply never does — which should at minimum downgrade the certainty with which the fourfold-varna picture gets asserted as settled ancient-Indian social fact.
The distinction matters practically for how this page gets cited elsewhere in the vault: the correct claim is "the evidence for lived varna-as-social-reality in Ashoka's period is thinner than commonly presented," not "varna didn't exist in the third century BCE." Collapsing the second, stronger claim out of the first, more careful one would be exactly the kind of overreach Olivelle's own methodological discipline is trying to guard against.
You're building a table of every social-category term Ashoka uses across the corpus, and you start, as you should, by looking for what's present rather than what's absent — that's the easier, more concrete task. Slaves and servants show up. Ethnic and regional groupings show up. Religious-professional categories (Brahmin, Sramana) show up. The table fills in reasonably fast.
Then you go back and check it against the standard textbook list — Brahmin, kshatriya, vaishya, shudra — and three of the four rows in your table stay completely empty no matter how many passages you search. You resist the urge to conclude too quickly from this. An absence in one document proves very little on its own; plenty of real, contemporaneous social categories go unmentioned in any single text just because that text wasn't the right genre to mention them in.
So you go looking for a second, independent body of evidence that could tell you whether this is really Ashoka-specific or something broader. You find Bronkhorst's cross-check against early Tamil inscriptions — a completely different region, completely different linguistic tradition, no plausible direct link to Ashoka's own court. The same three empty rows show up there too. Now you have two independent silences pointing the same direction, and two independent silences are a meaningfully stronger claim than one.
You write the finding down carefully, matching Olivelle's own hedge rather than overstating it: not "varna didn't exist," but "the evidence for varna as lived social reality in this period is thinner and more one-sided than its confident textbook presentation implies." You close the file with a transferable habit: when a claim rests on a genre of source that has an obvious interest in the claim being true, go looking specifically for independent, disinterested corpora that had occasion to make the same claim and see whether they do. The gap between "everyone said it" and "everyone who benefited from it said it" is often the whole story.
The core evidentiary claim — total absence of varna and the three non-Brahmin varna-terms from Ashoka's corpus, and the consistent Brahmin/Sramana pairing pattern instead of Brahmin/varna pairing — is a straightforward, checkable textual observation, strengthened considerably by Bronkhorst's independent cross-corpus corroboration. The interpretive leap (from "absent from this evidence" to "possibly aspirational rather than descriptive ideology") is more contestable, and Olivelle flags it as exactly that — a question raised, not a claim proven.
An open tension worth naming: if varna truly was, as this reading suggests, primarily an elite Brahmanical textual project rather than the lived organizing structure of the broader population, when and how did it become the socially operative reality it clearly was by later centuries? The silence documents an absence at one point in time; it doesn't by itself explain the subsequent process by which the system became descriptively real.
A further methodological tension worth flagging: this page's own case-study section (Counting the Brahmin/Sramana Pairing) treats a zero-exceptions pattern as unusually strong evidence, which is a reasonable inference but still an inference from a finite, non-exhaustive corpus. A single undiscovered or misread edict fragment breaking the pattern wouldn't overturn the broader argument, but it would soften the specific "always" claim this page and Olivelle both lean on.
This material sits entirely outside Lahiri's existing coverage, which doesn't isolate the varna-silence as its own analytical unit. It's methodologically a close cousin of the vault's existing genealogical-silence material (see Ashoka's Deliberate Genealogical Silence) — both pages read an absence in Ashoka's corpus against an assumed baseline to extract historical meaning — but the two silences work in opposite directions.
The genealogical silence is read as a deliberate, personal choice Ashoka made about his own self-presentation. The varna silence is read as evidence about the surrounding society, not primarily about Ashoka's personal choices — he isn't hiding something he knew; the absence more plausibly reflects a social reality (or at least a non-elite linguistic reality) genuinely different from what later Brahmanical sources project backward onto this period.
That distinction — personal-choice silence versus societal-evidence silence — is worth keeping sharp whenever the vault's growing collection of Ashokan silences gets read together. Not every absence in the corpus is doing the same interpretive work, and collapsing them into a single undifferentiated "Ashoka's silences" category would blur exactly the distinction that makes each one analytically useful on its own terms.
Psychology: Identity Architecture and Defense Hub — There's a recognizable psychological mechanism underneath the "aspirational vs. descriptive" reframing Olivelle proposes: a social group experiencing anxiety about its actual position writes confident, totalizing descriptions of a hierarchy that places it securely at the top, and those descriptions, repeated and canonized, eventually get mistaken by later readers (including modern historians) for neutral sociological fact rather than a specific interest group's self-serving wish.
This is structurally the same move an individual makes when a confidently-stated self-narrative substitutes for an anxious, contested reality underneath — except performed at the scale of an entire textual tradition rather than a single psyche, with a much longer half-life for the resulting distortion because the "aspirational" text gets treated by history itself as documentary evidence.
Behavioral Mechanics: Propaganda & Mass Persuasion Hub — A social hierarchy repeatedly described in confident, systematic textual detail by the group that benefits most from it, absent independent corroboration from outside that group's own textual tradition, is a template that recurs across history well beyond ancient India — caste systems, aristocratic genealogies, and professional-guild hierarchies alike often rest on self-description by the interested party rather than externally verified social reality.
The insight neither domain produces alone: the durability of a hierarchy's textbook description often has less to do with how widely the hierarchy was actually lived, and more to do with which group controlled the surviving textual record describing it — a documented ancient case study for a mechanism modern propaganda analysis names but rarely gets to test against this much independent, contemporary counter-evidence.
Most modern cases of this pattern (corporate mission statements versus lived workplace culture, an institution's official history versus its archives) can't be checked against genuinely independent contemporary evidence, because the interested party usually also controls or dominates the surviving record. Ashoka's corpus is the rare case where an independent, contemporaneous, non-Brahmanical voice happens to survive at scale — which is exactly what makes it useful as a documented instance of the mechanism rather than just a plausible-sounding theory about how such distortions might work.
Sharpest Implication
If Olivelle and Bronkhorst are right, then one of the most confidently taught "facts" about ancient Indian civilization — the fourfold varna social order — rests on evidentiary ground considerably shakier than its textbook-level certainty implies, and the strongest available contemporary counter-evidence has been sitting in plain sight, in the largest surviving body of non-Brahmanical primary evidence from the period, for nearly two centuries since Prinsep's 1837 decipherment.
The historical discipline's continued confident presentation of varna-as-descriptive-fact, despite this, suggests how powerfully a well-organized textual tradition's self-description can outcompete contradictory primary evidence simply by being more systematic, more available, and produced by the group with the most cultural capital to canonize its own account.
That's a sobering methodological lesson well beyond this one case: even a field with rigorous standards can inherit and repeat an interested party's self-description for generations, simply because the alternative evidence is scattered, silence-based, and harder to build a confident narrative around.
Generative Questions
A closing scope note: everything above traces to the specific inscriptional and comparative evidence Olivelle and Bronkhorst cite. Where this page extends into its own analysis — the pairing-pattern case study, the survival-asymmetry mechanism — that extension is flagged as vault-level synthesis, not attributed to either scholar's own stated argument, and any future page building on this material should preserve that same distinction.