“Defining the Law”: holding names to account for reality

“Defining the Law” is Han Fei’s comparison and synthesis of Shen Buhai’s “technique” (shu) and Shang Yang’s “law” (fa). “Holding names to account for reality” (xun ming er ze shi) can be read as a principle of performance assessment: first define what each post should do (the name), then reward or punish according to what was actually done (the reality). This links directly to the metric problems of Chapter 6: assessment presupposes that the “name” can be clearly defined, and once the “name” is used to decide life, death, reward and punishment, it comes under exactly the kind of pressure Goodhart described.

Note also the difference between technique and law. Law is “displayed in the government offices”, public; technique is held by the ruler alone. The chapter “Difficulties III” (Nan san) of the Han Feizi puts it more directly: “Technique is hidden in the breast, used to match the many threads of affairs and to steer the ministers in secret” (术者，藏之于胸中，以偶众端而潜御群臣者也) (陈奇猷 校注 2000). In form this resembles the “unannounced random audit” of Chapter 6, but the purpose is entirely different. A modern random audit usually makes its rules public in advance (there will be checks, and these are the penalties) and conceals only the timing and targets of the checks; the point of “technique” is that the ministers should be unable to read the ruler himself.

“The Two Handles”: words and deeds, names and results

The received text reads 言异事也 in the second sentence; some editions emend it to 言与事也, “that is, words and deeds”, and the translation follows that reading.

To a modern reader the most interesting feature of this passage is that it punishes both overstatement and understatement. If only “large words, small results” were punished, ministers would understate their commitments, keep something in reserve and then claim credit for exceeding them; punishing deviations in both directions amounts to requiring the declaration itself to be accurate. In form this resembles the modern approach of making people “report truthfully”: a declaration that departs from reality in either direction carries a cost.

But there are three differences. First, modern incentive compatibility makes truthful reporting serve the participant’s own interest on condition that participation is voluntary; the ministers of “The Two Handles” had no option of leaving, and the penalty was death. Second, real outcomes are subject to luck: a rule that strictly punishes any deviation makes ministers (or employees) avoid anything uncertain, because even an honest declaration may turn out wrong. In modern contract theory, when outcomes are noisy, optimal incentives must tolerate deviations (Book 16 of this series discussed the trade-off between risk and incentives). Third, “small words, large results are also punished” would discourage initiative in a modern organisation: it ranks control above output, which is exactly Han Fei’s choice of values.

This story is often placed in “Defining the Law”; in fact it comes from “The Two Handles”. From the point of view of mechanism design, Marquis Zhao worried about two things. One is that once the boundaries of responsibility blur, the “name” can no longer be used to assess the “reality”: the keeper of robes failed, but the failure was covered by the keeper of caps, and the assessment system would never see it. The other is that helping beyond one’s post lets officials build relationships of mutual cover, the “factions” of the story’s last line. This resonates with the fragility of VCG against coordination among losers (Chapter 2) and with the ease of gaming a single metric (Chapter 6): for an assessment system to work, it must prevent those being assessed from coordinating in ways that destroy information.

The cost is equally clear. The keeper of caps did something good for the ruler; punishing it tells everyone that doing something useful for the organisation that is not on one’s list of duties is dangerous. Modern multitask models tell us that much valuable work is precisely the unmeasurable kind that is on nobody’s list; a system that rewards only “names matching reality” will squeeze it out.

Guanzi, “Making the Law Clear”: not letting intentions roam outside the law

The author’s draft quoted “when the former kings governed the state, they did not let their intentions roam outside the law, nor grant favours within the law”, which agrees with the received text except that the received text opens with “therefore” (shi gu). It should be noted that the Guanzi also contains a chapter “Explaining ‘Making the Law Clear’” (Ming fa jie), a line-by-line commentary on “Making the Law Clear”; the two are different chapters, and this sentence comes from Ming fa. The Han Feizi chapter “Having Standards” (You du) contains an almost identical sentence: “do not let their intentions wander outside the law, nor grant favours within the law; no action departs from the law” (不游意于法之外，不为惠于法之内，动无非法) (陈奇猷 校注 2000).

The sentence stresses the certainty of rules and the exclusion of private discretion. In the language of mechanism design this relates to “commitment”: a set of rules works as participants expect only if the designer is also bound by it and will not change it afterwards for some individual. Chapter 2 said that the virtue of the second-price auction is that “your bid only decides whether you win”; if the auctioneer could change the pricing rule after seeing the bids, that virtue would vanish. “Not granting favours within the law” (not giving anyone extra favours inside the rules) is one way of saying “do not change the rules after the fact”.

Confucians: “to govern is to correct” and “competing for profit”

These Confucian passages stand in sharp contrast to the Legalists. The basic assumption of mechanism design is that participants pursue their own interests, and the designer’s task is to align self-interested action with a common goal. Mencius warns instead that if everyone, high and low, starts from “profit”, the whole community is endangered. “Lead them with decrees and keep them in line with punishments, and the people will avoid punishment but have no sense of shame” can be read beside the crowding-out effect discussed in Book 24 of this series: rules based mainly on punishment may secure compliance while weakening people’s own commitment to the norm. That is an empirically disputed phenomenon with some supporting evidence, and Confucius’s words cannot be treated as evidence for it. “To govern is to correct” reminds us that whether rule-makers themselves keep the rules affects the credibility of the rules, which approaches from another direction what the Guanzi says with “not granting favours within the law”.

Complementarity, tension and boundaries

Reading these texts beside modern mechanism design reveals a kind of complementarity. The Legalists cared about the enforceability of rules, about assessing whether names match reality, and about preventing coordination among subordinates; the Confucians cared about the example set by those who make the rules and about moral motivation outside the rules. Modern research offers more precise tools on both fronts: the models of Chapter 6 show why “name and reality” assessment fails under pressure and how random audits help; research on crowding out shows under what conditions punishment-based rules weaken intrinsic motivation.

But three boundaries must be stated. First, the purposes differ. Legalist reward and punishment were instruments for a ruler to control ministers and subjects, and “technique” called for secret steering that no one could read; modern mechanism design addresses equal participants, requires participation to be voluntary and rules usually to be public, and aims at the welfare of participants and the public. Second, the methods differ. The ancient texts are normative arguments and stories; modern mechanism design consists of mathematical proofs under explicit assumptions and testable empirical research. A formal resemblance between “holding names to account for reality” and incentive compatibility does not mean the ancients proposed the same theory. Third, the value premises differ. When Mencius criticised “competing for profit”, he was not discussing how to design incentive-compatible rules but arguing for the priority of morality; reading him as saying that “incentive design must prevent conflicts of local interest” is a modern translation, not his meaning. The plain-language readings in this chapter are the author’s; the commentarial tradition often reads these passages differently.

Chapter summary

“Holding names to account for reality” in the Han Feizi chapter “Defining the Law” is a principle of assessing actual work by the name of the post, close in spirit to modern performance assessment and subject to the same pressure towards gaming. “The Two Handles” punishes both “large words, small results” and “small words, large results”, formally requiring accurate declarations; the story of Marquis Zhao “punishing both the keeper of robes and the keeper of caps” (from “The Two Handles”, not “Defining the Law”) shows the cost of punishing helpful overreach to keep names and reality assessable. The Guanzi line “not letting intentions roam outside the law, nor granting favours within the law” stresses the certainty of rules and the exclusion of after-the-fact discretion. The Analects on “to govern is to correct” and “lead them with virtue”, and Mencius on “competing for profit”, point to example and moral motivation beyond rules. Ancient Legalism served the ruler’s centralised power, whereas modern mechanism design serves voluntary contracts between equals and public welfare; formal resemblances cannot hide fundamental differences of purpose, method and value premise. The next chapter offers a checklist and a trial protocol.