That single sentence is the discipline. Faxian has just spent thirty paragraphs describing the five-storey rock monastery at Dakshina — elephant on the ground floor with 500 apartments, lion floor with 400, horse with 300, ox with 200, pigeon on top with 100; a spring at the top whose water flows in spirals through every room; villagers who watch monks fly into the building because "Our wings are not yet fully formed." He goes on for a full chapter. Then, at the end, he tells you: I never saw any of this. I couldn't get there. People who had been there told me. I am writing down what they told me. This is hearsay.1
That's the entire epistemic posture in one move. Faxian sorts every claim he makes into one of three boxes, and he doesn't mix them.
Tier 1: What he saw. Most of the chronicle. "Three or four le from the city" — he was there. "The vihara stands in a square of thirty paces" — he paced it. "Fa-hien had not arrived in time to see the distinguished Sramana alive, and only saw his burial" — note the precision, the witness-line he can claim and the witness-line he can't.2 When he records the Buddha-shadow at Nagara he describes it from the angle he stood at: "Looking at it from a distance of more than ten paces, you seem to see Buddha's real form... the nearer you approach, however, the fainter it becomes, as if it were only in your fancy."3 You can feel him triangulating with his eyes.
Tier 2: What he heard on-site, present as local oral tradition. The miracle stories. The jataka-marker topes. The legends about specific places. The flying monks at Dakshina. These get reported with the same reverent register he uses for what he saw, but the chronicle does not commit to their literal truth. It commits to the fact that this is what the local sangha was teaching at this site in 405 CE. The Buddha-converted-the-wicked-dragon story at Udyana, the Sakra-as-herd-boy story behind Kanishka's tope, the white-eared dragon as monastery danapati at Sankasya — all of these are recorded faithfully because they are part of the local instructional landscape, not because Faxian saw a dragon.
Tier 3: What he could not visit but was told. Dakshina is the named-and-flagged case. There's also the section in Ch. XXXIX where an Indian devotee preaches the alms-bowl-prophecy sermon — Vaisali to Gandhara to Western Tukhara to Khoten to Kharachar to the land of Han to Sinhala to Central India to the Tushita heaven, splitting into four at Maitreya's Buddhahood. Faxian wants to write it down "as a portion of doctrine." The preacher stops him: "This is taken from no Sutra, it is only the utterance of my own mind."4 Faxian records the disclaimer. The eschatology stays in the chronicle, but flagged as one Indian monk's personal speculation, not canonical teaching.
The discipline is that the tier-marker is preserved. Faxian does not laundere tier-3 hearsay into tier-1 witness. He doesn't claim he saw what he didn't see. When he does report his own witness, he means it — and when he doesn't, he says so.
It is not modern source criticism. Faxian is not asking whether the local oral tradition is historically accurate. He is not skeptical about whether Buddha actually converted a wicked dragon at Udyana. The discipline is genre-level, not truth-level: I will record the legend as legend told here, the site as site stood here, my emotional response as my response — and I will not mix the three.
It is not the discipline of a journalist either. A journalist would seek to verify the local oral tradition against other sources. Faxian doesn't. The traditions are received as received. The discipline is about what kind of report he is making about each piece of material — not about whether each piece is true.
What it most resembles is a working scholar's footnote register. Witness: I observed this. On-site oral: this is what the local sangha is teaching. Hearsay: I could not visit, this is what others reported. Faxian is doing the footnote register in flowing prose, four hundred years before footnotes existed as a literary apparatus.
Look at this case carefully because it's where you can see Faxian's plain text getting embellished by later hagiography, and see why his epistemic discipline matters.
Faxian's own narrative of his night on Gridhra-kuta is restrained. He buys incense, flowers, oil and lamps in the New City. He hires two long-resident bhikshus to carry them up. He makes offerings when he arrives, lights the lamps as darkness falls, feels melancholy but restrains his tears, and says: "Here Buddha delivered the Surangama Sutra. I, Fa-hien, was born when I could not meet with Buddha; and now I only see the footprints which he has left, and the place where he lived, and nothing more." He chants the Surangama Sutra. He stays overnight. He goes back to the New City in the morning.5
That's all his text says. No supernatural occurrences. No threats. No tigers.
But the Memoirs of Eminent Monks (compiled 519 CE, a century after Faxian) adds an embellishment: while Faxian is on the peak that night, two black lions crouch before him, "licking their lips and waving their tails." Faxian's resolve so impresses them that "they drooped their heads, put down their tails, and prostrated themselves before him." Beal's note (the previous English translator) imported this story, even saying Faxian was "attacked by tigers."6
Legge catches the import. His footnote on the passage corrects: "But 'the high priests' merely means distinguished monks, 'eminent monks'... Nor was Fa-hien 'attacked by tigers' on the peak. No 'tigers' appear in the Memoir. 'Two black lions' indeed crouched before him for a time this night... This of course is not an historical account, but a legendary tribute to his bold perseverance."7
So we have three layers. Faxian's chronicle (414 CE): restrained, melancholy, no supernatural element. Memoirs of Eminent Monks (519 CE): legendary lion-test as canonization. Beal's 1869/1884 English translation: imports the lion-test back into Faxian's own account, mistranslates lions as tigers, and treats the legend as part of the primary record. Legge 1886: separates the layers and identifies which is which.
The original epistemic discipline is Faxian's. The hagiographical layering happens after his death. What the Gridhra-kuta case shows you is how fast a witness chronicle gets converted into a saint's life once the witness is gone. Faxian's own register would not have allowed the lions to appear unless he saw them. The later sources have no such constraint. The discipline is not transmissible — it has to be re-imposed at every reading.
Three structural places in the text where Faxian flags his tier-shifts explicitly:
Dakshina (Ch. XXXV): "The kingdom of Dakshina is out of the way, and perilous to traverse. There are difficulties in connexion with the roads; but those who know how to manage such difficulties and wish to proceed should bring with them money and various articles, and give them to the king. He will then send men to escort them. These will (at different stages) pass them over to others, who will show them the shortest routes. Fa-hien, however, was after all unable to go there; but having received the (above) accounts from men of the country, he has narrated them."8 The disclaimer comes at the end, after the description. The full picture is preserved and the source-tier is marked.
The alms-bowl prophecy (Ch. XXXIX): "(Such was the discourse), and Fa-hien wished to write it down as a portion of doctrine; but the man said, 'This is taken from no Sutra, it is only the utterance of my own mind.'"9 Faxian records both the prophecy and the preacher's disclaimer. He doesn't suppress the speaker's caveat to make the prophecy look canonical.
The hunter encounter on the Shan-tung coast (Ch. XL): When the merchants find Faxian on what they think is the Kwang-chow shore, they bring back two hunters and ask Faxian to question them. Faxian asks "Who are you?" They reply "We are disciples of Buddha." He asks what they are looking for. They lie — they say they are looking for peaches to present to Buddha. Faxian catches the lie. Legge's footnote: "I suppose these men were really hunters; and, when brought before Fa-hien, because he was a Sramana, they thought they would please him by saying they were disciples of Buddha. But what had disciples of Buddha to do with hunting and taking life? They were caught in their own trap, and said they were looking for peaches."10 Faxian, even in the moment of receiving information that would lead him home to the prefect Le E, records the contradiction in the hunters' story rather than smoothing it over.
That last case is the discipline in its purest form. He has nothing to gain by recording the lie. He records it anyway because that's what reportage is.
The discipline is real but not perfect. Faxian also reports without disclaimer things he plainly could not have witnessed: Asoka's spirit-built palace at Pataliputra ("all made by spirits which he employed, and which piled up the stones, reared the walls and gates, and executed the elegant carving and inlaid sculpture-work — in a way which no human hands of this world could accomplish"), the bowl-shaped depression where Sakra traced 42 questions with his finger at Rajagriha ("the prints of his tracing are still there"), the Buddha-shadow at Nagara, the bhikshu-suicide-as-three-stage-attainment at Devadatta's rock dwelling. These are reported in the same matter-of-fact register as the geography. Whether Faxian believed them literally, believed them as local oral tradition, or believed them as devotional emblems — he doesn't say. The reader has to do that tier-assignment now.
So the discipline is best stated as: Faxian marks the tier whenever it would matter to a careful reader of his own time. He doesn't bother marking the tier of standard Buddhist legend, because his sangha audience would already know it was legend. He does mark the tier of the Dakshina-monastery report (because his reader might mistake it for direct witness) and the alms-bowl-prophecy (because his reader might mistake it for canonical Sutra) and the hunters' lie (because his reader might mistake the lie for verified information).
The discipline is reader-sensitive, not absolute. He's marking the tier for the use the reader will make of the material, not for some abstract standard of evidentiary purity.
Faxian writes in 414 CE, marking each claim with its source-tier. Legge translates him in 1886 and writes another layer of marks — "I cannot tell," "Beal renders this differently," "I have not met with the particulars." Both men are doing the same operation, but their registers nest like Russian dolls — Legge's 1886 register wraps Faxian's 414 register wraps the 5th-century-CE on-site oral traditions Faxian was receiving.
Both of them are doing source-criticism, but from opposite directions. Faxian sorts what he is reporting into witness / on-site oral / hearsay. Legge sorts what Faxian reports plus what later sources added into authentic-Faxian / later-hagiography / translator-error. Faxian's register is about clean intake. Legge's register is about clean transmission. They converge on the principle that the source-genre matters more than the content's plausibility — neither asks whether the dragons are real, both ask whether the dragon report comes from the right place. They diverge on where they sit in the chain: Faxian is one stage upstream of Legge.
What the convergence reveals — that neither states alone — is that the entire reliability of Buddhist textual transmission across fifteen centuries depends on discipline-of-this-kind being repeatedly imposed at each step. Faxian's care at intake bought one stage of clean transmission. Legge's care at re-reading bought another. The intervening twelve centuries — the Memoirs of Eminent Monks layer where lions appear, where Beal's translation imports tigers — show what happens at the stages where the discipline is not maintained. The chronicle survives across time only as well as its successive readers re-impose the witness/hearsay distinction Faxian originally maintained. It is not a property of the text. It is a property of the reading practice.
Faxian's discipline is structurally identical to several already-mapped practitioner methodologies in creative-practice, but with a difference of stakes — Faxian was practicing it on devotional material across an oral-tradition civilization, not on prose craft.
Creative-practice: "May Have" Discipline — Degrees of Confidence (MacCulloch) — Diarmaid MacCulloch's historian's-ethic of marking the confidence-tier of every claim ("we may say" / "we can say" / "we know"). Faxian is doing this same operation 1,500 years earlier in a non-academic genre. The convergence: both treat source-tier-marking as the professional virtue of the genre, not as squeamishness or hedge. They diverge on what they're marking against: MacCulloch against the temptation to fabricate vivid detail in historical narrative; Faxian against the temptation to inflate his own travel-witness for status. Together they show: the discipline of marking-the-tier is genre-portable. It works in 5th-century Chinese chronicle and 21st-century Tudor history because the underlying problem is the same — readers will believe what you put on the page, and the only ethical move is to tell them what kind of claim you are making.
Creative-practice: Reader as Proxy — Writer One Step Ahead (Orlean) — Susan Orlean's stance that the writer is the reader's fellow-traveler, one step ahead but not infinitely more knowledgeable. Faxian does this constantly. He is the reader's proxy on the road — he tells you what he saw at the pace and angle of a fellow-pilgrim, not from the elevated viewpoint of an authority. When he can't get to Dakshina, he says so. When he doesn't know who the envoy was that he accompanied out of T'un-hwang, the chronicle says: "Who this envoy was, and where he was going, we do not know."11 (That's Legge's footnote summarizing Faxian's silence.) The structural parallel: reader-as-proxy reportage demands tier-marking, because a one-step-ahead writer cannot pretend to know what they don't. The two together show: the genre we now call "literary nonfiction" or "long-form reportage" has its 5th-century Chinese instance in Faxian. The mechanism is the same.
History: Lahiri's Historiographic Method for Ashoka (existing Lahiri ingest) — Nayanjot Lahiri's landscape-mediates-meaning approach to Asokan inscriptions treats each inscription as situated in a specific physical and political context, refusing to abstract them into a single Asoka-monologue. Lahiri is doing tier-marking on the interpretive frame, not on the witness-status. The convergence with Faxian: both refuse to flatten their material. The divergence: Lahiri's tiers are interpretive (what does this inscription mean in this place?); Faxian's tiers are testimonial (did I see this, or did someone tell me?). Together they suggest that the historian's discipline has at least two layered tier-systems that both must operate: tier-of-witness and tier-of-interpretation. Neither alone is enough.
The Sharpest Implication
If you take Faxian's discipline seriously, your own reading practice for any source — chronicle, news article, biography, recommendation — has to do the tier-sort as you read. Most readers do not. They read until something seems implausible, then dismiss the whole source as unreliable. Faxian's move is finer: I will accept witness-tier reports at witness weight, on-site-oral reports at oral-tradition weight, hearsay at hearsay weight, even within the same chronicle. The lions on Gridhra-kuta and Faxian's restrained-tears moment can coexist on the same page without contaminating each other, if the reader does the sort. The uncomfortable implication: most of what we now call "fact-checking" is binary (true / false) when the more useful sort is tier (witness / oral / hearsay / speculation / hagiography). Binary fact-checking lets us reject whole sources. Tier-sorting forces us to use sources at the right grain — accepting what they can tell us at the level they can tell it.
Generative Questions