The year Zhang Juzheng imposed his scrutiny system, the Ming treasury was empty and the bureaucracy moved at a crawl. What he wanted was concrete: money, and speed.
In the first year of Wanli he gave every official in the empire a set of books. Each task got a deadline; three registers were kept, one in the ministry, one with the Six Offices of Scrutiny, one with the Grand Secretariat; anything unfinished when the date fell due was pursued up the chain, struck off by the censors, checked by the cabinet. The record he left looks excellent. The Ming History says the granaries held enough grain for ten years and the Court of the Imperial Stud had piled up over four million taels in horse-tax silver. The numbers were real. So was the trouble.
I am not going to call the scrutiny system a bad institution. Quite the opposite: it was one of the few attempts in the Ming to attach responsibility to a specific date. Earlier edicts mostly ended at three words — “let it be carried out” — leaving blank who would do it, by when, and what happened if it was not done. Zhang filled in exactly those three blanks. A state whose orders change from adjectives into verbs deserves some respect for that step alone.
The problem was the fourth blank: who checks the words written in those three registers.
Once a metric becomes the only passport, people stop working for the thing and start working for the metric. A tax call becomes a quota, the quota becomes a chase, and the chase ends in selling grain and pawning land. The local magistrates did not suddenly turn wicked; they learned a different arithmetic. Keeping their own column pretty was cheaper and safer than keeping a county alive. And the man in the cabinet, turning pages, saw only three clean ledgers.
This is not a Ming disease. Today’s weekly reports, paper counts and photo-op checklists are all descendants of the scrutiny system, and look more like the ancestor than the ancestor did. Nor do I want to write numbers off entirely: their greatest service is that they let someone a thousand miles away understand. The diligence of a man, the life of a county — a grand secretary in the capital cannot see these, but he can see “how much of the levy is still owed.” A number is a stand-in for trust. Stand-ins work at a distance and lie up close. That is their temperament, not their crime. I am not arguing for scrapping assessment — an edict without a deadline really does rot in a drawer, and on that point I believe Zhang. What I object to is treating assessment as governance itself. Assessment answers “was it done,” never “was it done right.” Only people willing to answer for outcomes can answer the second question, and people are not manufactured by forms.
I have done this to myself. Some years ago I set myself a thousand words a day; for a fortnight I hit the number every day, and looking back it was all rubbish, harder to delete than it had been to write. I changed it to “understand one thing properly each week.” Output fell by half, and more of it was worth reading. Wherever you put the metric, that is where people’s cleverness grows. That is instinct, and no one should be blamed for it. The blame belongs to whoever set the metric without asking what he actually wanted: the thing, or the number.
After Zhang’s death his estate was seized and the scrutiny system quietly lapsed. People like to say that the man died and the policy died with him. I think what lapsed was not the law but the man willing to open three registers every day and check each deadline line by line. The expensive part of an institution is never the design; it is the maintenance — somebody has to verify, year after year, tediously, with no credit. Nobody will do that for long, so the loveliest institution loosens into a net with very wide mesh.
Numbers grow themselves. What grows cannot be eaten. To judge whether a place is really getting better, do not read the figures on the ledger. Look behind them, and see whether anyone is still doing that one extra step of checking.
卜世 撰
说些什么吧！