Xunzi's Radical Constructivism: Li as Autonomous Normativity Beyond Ren — Epoche C2
The nine characters, and the word that carries them Xunzi's twenty-third chapter opens with nine characters — 人之性惡，其善者偽也 — and the reading defended here turns on the last of them that carries lexical weight: 偽 wei , which does not mean falsehood but means what has been deliberately worked up. The sentence is conventionally rendered "human nature is bad; its goodness is deliberate effort", and treated as the pessimistic counterpart to Mencius's optimism it is a claim about temperament. Read with Xunzi's own definitions in hand it is something else: the conclusion of a compact argument whose premises are partly stipulative and partly empirical, and whose upshot is that ritual — li 禮, the whole graded system of ceremonial, sumptuary and behavioural norms — is not the outward expression of an inner benevolence but a human artefact justified by what it accomplishes. The claim to be defended is the strong one. Ren 仁, ordinarily rendered benevolence or humaneness, is on the dominant Confucian account the foundational virtue, an inner disposition of which li is the outward form and the training. Xunzi reverses the order of grounding: li is prior, its authority derives from a function that can be described without using any moral term at all, and ren is among the things li produces. Xunzi (conventionally c. 310–235 BCE) states this in a text whose received form we owe to Liu Xiang (79–8 BCE), who reports collating 322 bundles of Xunzi material, discarding 290 as duplicates, and fixing the remainder at 32 chapters, the numbering used below. Why the argument of chapter 23 is valid, and where it is vulnerable Chapter 22, "Rectifying Names" (正名 Zhengming ), supplies the definitions, and chapter 23 uses them. Xunzi defines xing 性, nature, as what is so of itself from birth, and wei as what the heart-mind ( xin 心) brings about when it deliberates and the faculties act on the deliberation — with a second and more important clause: when deliberations have accumulated and capacities have been practised, and the result is complete only then, that too is wei . So wei names both the single deliberate act and the sediment such acts leave. Chapter 23 turns this into an exhaustive and exclusive partition of everything present in a person: Feature present in a person Classified as Cannot be learned, cannot be worked at xing 性 Can be learned and mastered, can be worked at and completed wei 偽 The partition is exhaustive because the two conditions are contradictories, and Xunzi says so in terms: this is the division between xing and wei . The argument then runs in three steps. Ritual conduct and yi 義 — righteousness, the sense of what is fitting to one's station — are acquired by learning and completed by effort, and are not found in a person who has done neither. By the definition, therefore, they are wei and not xing . Since moral goodness consists in them, goodness is wei . That much is watertight, because it is nearly analytic. The controversial half of the chapter's title is the other half: not that goodness is acquired, but that xing is bad. This is not a definitional claim at all but an empirical prediction — that human beings left to the spontaneous course of their desires will contend, and that contention issues in disorder. A. C. Graham's analysis of the debate makes the point that xing in these texts is less an essence than the course along which a thing spontaneously develops, and that Xunzi's 惡 e is accordingly a verdict on where that course leads rather than an attribution of wickedness to a newborn. The distinction matters for what follows: an argument that human nature tends toward outcomes we deplore leaves the newborn morally blank, and a blank is exactly what a constructive project requires. The dominant reading, and how much of the Analects actually supports it Before setting Xunzi against the priority of ren , the priority thesis should be stated at its true strength, which is lower than it is usually given. It is commonly said to be rooted in the Analects . Part of the Analects does root it: at 3.3 Confucius asks what a man who lacks ren has to do with ritual, or with music — which plainly treats ren as the condition on which ritual has any value. But at 12.1, answering Yan Hui's question about ren itself, he says that to overcome oneself and return to li is ren , and then specifies the virtue entirely in ritual terms: do not look, listen, speak or move contrary to ritual. That passage characterises ren by way of li , not the reverse. So the Analects contains both orders, and the settled priority of ren is the achievement of the Mencian line and of its later orthodoxy rather than an unambiguous datum of the earliest text. This is worth conceding, because it changes what Xunzi is doing. He is not overturning a fixed reading of Confucius; he is resolving an indeterminacy in one of the two available directions, and doing so with an argument rather than by fiat. The Mencian position he is resolving against is also more specific than "human nature is good". Mencius 2A6 argues from a thought experiment: anyone who suddenly saw a child about to fall into a well would feel alarm and compassion, and would feel it before any calculation of advantage, praise or reputation. The claim is not that people are reliably good but that the reaction demonstrates an incipient tendency — one of the four "sprouts" that, given cultivation, grow into the four virtues. The disagreement with Xunzi is therefore not optimism against pessimism. It is a disagreement about whether what is innate is a tendency , which already points in the moral direction and needs only nourishing, or a bare capacity , which points nowhere and needs to be given a direction from outside. Where li comes from: the apportionment argument Chapter 19, "Discourse on Ritual" (禮論 Lilun ), gives ritual an origin story that contains no moral premise, and this is the load-bearing passage for the whole reading. "Whence does ritual arise? I say: people are