The Introduction to Legge's 1886 translation of Faxian contains a remarkable third section — pages 196-220 of the source — that has nothing to do with Faxian and everything to do with the religious-statistical politics of late-19th-century Anglo-Protestant scholarship.
The setup: by the late 19th century, several major estimates of world Buddhist population were in circulation. General Alexander Cunningham (1854): 222 million Buddhists. Max Müller (1868): 455 million. T. W. Rhys Davids (Manual of Buddhism): 500 million. Berghaus' Physical Atlas (cited by Müller): Buddhism at 31.2% of world religious population, Christianity at 30.7%. Buddhism, on these estimates, was the world's largest religion.
This was a problem for missionary-minded Anglo-Protestant scholars. James Legge — Reverend, former missionary to Hong Kong of 30+ years, Oxford Professor of Chinese, translator of the Confucian Classics and now of Faxian — was not willing to let this stand. His Introduction part 3 is an extended statistical argument to demote Buddhism from first-place to fifth-place in the world-religion rankings.
His method is single-axis: deduct the Chinese population from each estimate, on the grounds that most Chinese are Confucian, not Buddhist. Cunningham's 222 million has 170 million Chinese — subtract those and Buddhism falls to 52 million globally. Rhys Davids' 500 million has 414 million Chinese — subtract those and Buddhism falls to 85 million globally. Both numbers are much smaller than Christianity (about 270 million in Cunningham's count) or Mohammedanism (which Legge does not separately compute but argues exceeds Buddhism in the same way).
His conclusion at line 220: "It does seem to me preposterous to credit Buddhism with the whole of the vast population of China, the great majority of whom are Confucianists. My own opinion is, that its adherents are not so many as those even of Mohammedanism, and that instead of being the most numerous of the religions (so called) of the world, it is only entitled to occupy the fifth place, ranking below Christianity, Confucianism, Brahmanism, and Mohammedanism, and followed, some distance off, by Taoism."1
Buddhism: fifth place. Christianity, Confucianism, Brahmanism, Mohammedanism all rank above it. Taoism a distant sixth. This is the framework Legge sets up before translating a single sentence of Faxian.
Legge does not have to write this section. Faxian's chronicle is about 5th-century India, not 19th-century world-religion-rankings. The translation could begin without any statistical apologetic.
But Legge does write it. Three full pages of the Introduction are devoted to the population argument. The argument is structured as a careful piece of statistical reasoning. He cites his sources. He shows his math. He acknowledges uncertainty in his estimates ("we have no certain information of the population of China"). He preserves the appearance of scholarly disinterestedness while reaching a conclusion that conveniently demotes Buddhism in the world-religion-rankings.
What this signals: Legge is introducing the chronicle inside a framework that minimizes Buddhism's importance. Faxian's chronicle documents a multi-century Buddhist civilization at peak operation. Legge's Introduction sets up the reader to encounter that civilization as the religious-historical record of a fifth-place religion. The framing matters. The reader is led to approach the magnificent Pataliputra monasteries, the six-century-old metal-plate-grant economy, the 60,000-monk Anuradhapura, and the trans-continental Sangha hospitality network as the products of a tradition that, when accurately counted, is less important than Christianity, Confucianism, Brahmanism, or Mohammedanism.
This is missionary-Sinologist apologetics applied to translation-introduction. The translation that follows is, in fact, of unusual quality and care — Legge is a serious scholar. But the framing is theologically committed.
Read Legge's deduction carefully. He is not arguing that Chinese Buddhists don't exist. He is arguing that the religious affiliation of Chinese people is multi-stratum and that Confucianism is the dominant stratum, with Buddhism as a secondary or emotional layer.
He quotes Eitel approvingly: "It is not too much to say that most Chinese are theoretically Confucianists, but emotionally Buddhists or Taoists." Legge then strengthens the claim: "For the 'most' in the former of these two sentences I would substitute 'nearly all;' and between my friend's 'but' and 'emotionally' I would introduce 'many are.'"2
So Legge's model: nearly all Chinese are theoretically Confucianists; many are emotionally Buddhist or Taoist; the theoretical affiliation is the primary one for census-counting purposes; the emotional affiliation is secondary and should not be counted as full religious membership.
This is a specific theological-administrative move. Legge is assigning primary status to one stratum of multi-strata religious identification and discounting the others. The choice is not neutral. If you took the inverse move — primary status to emotional affiliation, secondary to theoretical — you'd get the opposite conclusion (most Chinese are primarily Buddhist or Taoist, with Confucian theoretical-stratification as secondary).
Why does Legge choose Confucian-primary? Because Confucian-primary is what produces the desired statistical outcome (Buddhism fifth place). The choice of which stratum to count primary is the argument's load-bearing decision, and Legge does not justify it independently of the desired outcome.
The deeper move: Legge is operating in a frame where religious affiliation must be single-tradition. People are either Buddhist or Confucian or neither, not both. The premise of mutually-exclusive religious affiliation is itself a Protestant-Christian assumption that does not match East Asian religious practice. East Asian religious practitioners typically operate simultaneously in multiple traditions — the same person consults Confucian ancestor-rituals for family matters, Buddhist temple-rituals for funerals, Taoist or local-religion practices for health and luck. The whole population is multi-strata religious. Legge's Protestant frame requires assigning a primary tradition and counting only that. The Protestant frame is what generates the Buddhism-is-fifth conclusion.
Legge's other introductory move worth noting: he treats Buddhism's specific theological claims with explicit non-equivalence to Christian claims. At Ch. II Legge note 3 (line 294), he refuses to translate "sang" (monk) as "priests": "in Christianity, beyond the priestly privilege which belongs to all believers, I object to the ministers of any denomination or church calling themselves or being called 'priests;' and much more is the name inapplicable to the sramanas or bhikshus of Buddhism which acknowledges no God in the universe, no soul in man, and has no services of sacrifice or prayer in its worship."3
Three claims in one sentence: Buddhism (a) acknowledges no God, (b) acknowledges no soul, (c) has no services of sacrifice or prayer. The triple-negation framing positions Buddhism as structurally inferior to Christianity on the three dimensions Protestant Christianity considers essential to religion: theism, soul-doctrine, sacramental worship.
This is consistent with Legge's broader project. Buddhism is statistically smaller (population argument) AND structurally inferior (no-God, no-soul, no-sacraments). The translation that follows is therefore the work of a generous-spirited scholar engaging with a tradition that, on his framework, is a fifth-place religion lacking the essential structural features of true religion.
Whether Legge would have explicitly endorsed this framing if asked is unclear; the framing operates through his framing-choices in introductory material rather than through explicit thesis-claims about Buddhism's inferiority. But the framing is real and visible.
Take Legge's argument seriously and the wider field of 19th-century Sinology becomes visible. Legge is not unique. The whole generation of British-Protestant-missionary-Sinologists who produced the first wave of English translations of Chinese texts operated within frameworks designed to position Christianity favorably against the Chinese religious traditions they were translating.
The frameworks varied: some scholars treated Confucianism as a "natural religion" that prepared Chinese readers for Christianity (Legge himself wrote in this register about the Confucian Classics). Some treated Buddhism as foreign-derived and therefore not authentically Chinese (a position that downplays Buddhism's centrality in Chinese cultural history). Some applied Protestant-confessional categories (priest, sacrament, scripture) and found Chinese traditions deficient when assessed by those categories.
What the various frameworks share is the incompatibility-of-translation-with-neutral-comparison. The act of translating Chinese religious texts into English in the 19th century was inseparable from the act of evaluating Chinese religious traditions against Christian benchmarks. The translation produced is high-quality scholarly product; the framing the translation arrives within is theologically committed.
Modern reception of Legge's translations has had to do double work: reading Faxian (which is good) through Legge (which is excellent translation-craft) while filtering out Legge's introductory and footnote framings (which are missionary apologetics). The Faxian content survives the filtering because the chronicle itself is robust against framing-overlay. Less robust texts have been more damaged by analogous 19th-century framings.
The revisionism page is, in this sense, an artifact of its specific historical moment rather than a contribution to understanding Faxian. It is preserved in the vault as a primary-source datum about how Faxian was made available to English readers in 1886, and what the cost of that availability was.
The primary witness is Introduction part 3, lines 196-220.4 The specific population-deduction reasoning is at lines 198-208. The conclusion is at lines 218-220. The Eitel quotation Legge strengthens is at line 218.
The Christian-comparative footnote refusing "priest" terminology is at Ch. II Legge note 3, line 294.5
Cross-validation from outside the source: the late-19th-century world-religion-statistics debate is well-documented in religious-studies historiography. Legge's specific argument is a less-well-known instance of a broader Anglo-Protestant scholarly pattern. The position that Chinese Confucians should not be counted as Buddhists despite their participation in Buddhist devotional practice was contested at the time and remains contested in religious-studies methodology today.
The most acute tension: Legge's argument depends on a methodological premise that is itself doctrinal. The premise that religious affiliation is single-tradition rather than multi-strata is a Protestant-Christian assumption. Used as a methodological premise to count world religions, it produces statistical outcomes that favor traditions which historically operated under Protestant-influenced single-affiliation frameworks. The methodology and the conclusion are entangled. A neutral methodology that allowed multi-tradition affiliation would produce different counts.
A second tension: even on Legge's own terms, his Confucian-primary assignment for Chinese population is debatable. Many Chinese in the period he was writing would have identified primarily as Buddhist or Taoist rather than Confucian. The Imperial-Confucian-orthodoxy framing was strongest in the educated literati class but weaker in the broader population. Legge's 30 years as missionary in Hong Kong gave him direct exposure to popular Chinese religious practice, which he reports as "many are emotionally Buddhists or Taoists." His own evidence undercuts his framework. The dismissal of "emotional" affiliation as secondary is the move that preserves the framework against his own evidence.
A third tension: Legge's revisionism does not engage with Faxian's chronicle itself. The chronicle documents a Buddhist civilization operating at high institutional density across most of Asia in 405-412 CE. Whatever the population-statistics of 1886, the historical fact of 5th-century Buddhist civilizational success is independent of late-19th-century counts. Legge's Introduction-framing creates an awkward mismatch: the reader is led to think Buddhism is small, then reads a chronicle showing Buddhism operating at metropolitan-civilizational scale across a continent. The mismatch tells you the framing is doctrinal-rhetorical rather than historically-grounded.
Diarmaid MacCulloch tells you upfront that he is a gay liberal Anglican writing church history; he expects you to read his work knowing that. Legge in 1886 does not do this. His Christian-missionary commitments operate behind the Introduction's statistical analysis without ever being named. The reader who knows Legge's biography sees missionary apologetics; the reader who does not sees scholarly analysis. MacCulloch's discipline is upfront declaration. Legge's discipline is implicit operation.
Diarmaid MacCulloch's historian-ethic: declare your biases upfront so the reader can adjust for them. Don't pretend to be neutral; instead, be explicit about where you're coming from.
Legge does not do this. His Introduction part 3 reads as objective statistical analysis. He cites sources. He shows reasoning. He acknowledges some uncertainty. The Protestant-missionary commitments that drive the choice of methodological premises are not declared. They are operating implicitly behind the analysis. A reader who doesn't know Legge's biography would read the Introduction as scholarly analysis. A reader who does know it would read it as missionary apologetics.
They converge on the principle: historians and translators have biases, and the question is how to handle them. They diverge on the handling. MacCulloch's solution is explicit declaration. Legge's solution (typical of his era) is implicit operation through methodological choices. The handshake produces an insight neither alone states: the implicit-operation pattern is the more common historical practice, and the explicit-declaration pattern is a relatively recent and still-unevenly-adopted alternative. Most 19th-century scholarly work operates like Legge — biases shape methodological choices that produce conclusions, with the biases not declared. Modern historiography has shifted somewhat toward MacCulloch's pattern, but unevenly. Religious-studies, which sits in proximity to confessional commitments, is particularly contested.
The Legge case is a diagnostic example of the implicit-bias pattern operating at high craft-quality. He is not a bad scholar. His translation is excellent. His statistical reasoning is internally coherent. The bias operates through the selection of which questions to ask, which premises to accept as default, which conclusions to find acceptable. The bias is invisible at the surface but consequential. The methodological-improvement that MacCulloch articulates explicitly is the declaration of these selection-biases upfront so the reader can adjust.
The Legge revisionism reaches into three adjacent vault domains.
Creative-practice: Declared Biases as Historian Ethic (MacCulloch) — see Author Tensions section above.
Behavioral-mechanics: rival-religion statistical defense (concept implicit in propaganda-mass-persuasion-hub) — religious traditions in competition for adherents often deploy statistical arguments to demonstrate their relative dominance. The 19th-century missionary-scholar deployment of population-statistics to demote rival traditions is a specific instance of the broader pattern. The structural parallel: statistics-as-rhetoric in inter-religious competition. The handshake produces an insight: religious-statistical arguments often serve rhetorical-positional functions rather than purely descriptive ones. The Buddhist-population-debate of the 1850s-80s was not just a question of correct counting; it was a contest over which religion would be presented to Anglo-Protestant readers as the largest and therefore most-worthy-of-engagement. Legge's specific intervention is one move in this contest.
Psychology: Identity Architecture and Defense Hub — Legge's life-long Christian missionary identity operates as the substrate behind his scholarly choices. The methodological premises that produce the desired statistical outcome track to the identity-substrate. The structural parallel: scholars' deepest professional identities affect what feels like "obvious" methodological choices in their work. Legge's Christian missionary identity makes "religious affiliation is single-tradition" feel like an obvious premise. To a non-Christian scholar from a tradition operating multi-tradition affiliation, the same premise would feel non-obvious. The handshake produces an insight: methodological intuitions are downstream of identity-substrate. Scholars working in good faith reach different methodological conclusions because their identity-substrates make different premises feel obvious. The MacCulloch ethic of declaring biases is a partial solution to this; identity-substrate-aware methodology is a deeper solution.
The Sharpest Implication
Take the Legge revisionism seriously and the standard image of 19th-century scholarship as "more objective than modern scholarship" inverts. Legge is unquestionably learned, careful, and craft-disciplined. He is also operating within explicit missionary commitments that shape his methodological choices in ways the surface scholarship doesn't acknowledge. The third-wire reading: most scholarship operates with substrate-commitments that shape methodological choices invisibly, and the appearance of objectivity is a function of the substrate-commitments being culturally invisible to contemporary readers rather than absent from the work. The Anglo-Protestant scholarly assumptions Legge operated within were the cultural default in 1886 Oxford; they look like assumptions only in retrospect when the cultural default has shifted. Modern scholarship operates within different default assumptions that will be visible as assumptions to future readers but currently look like the obvious correct method. The implication: every reader, including a contemporary reader of contemporary scholarship, needs to develop the capacity to see substrate-commitments behind methodological choices. This is harder than declarative bias-disclosure (MacCulloch's solution) because it requires recognizing what feels obvious as a culturally-contingent default rather than as a transparent truth. The Legge case is a useful exercise because his commitments are now visible enough to be analyzed without partisan pressure. The same exercise applied to contemporary scholarship is harder but structurally similar.
Generative Questions