Non DubitoEssays in the Self-as-an-End Tradition
← The Wire: The City, the Positions, and the Game← 《火线》:城市、位置与游戏
||
Essay I of IX · The Wire《火线》 · I / IX

The Stat

统计数字

How a Number Devours What It Was Meant to Measure

一个数字如何吃掉它要衡量的东西

Sep 1, 2026Han Qin (秦汉)Independent English edition中文原作 · 简 / 繁

Edition note: An independent English rewriting for English-language readers, preserving the Chinese argument while reorganizing context, sequence, and emphasis. Full-series spoilers.

What the Test Is Testing

In a Baltimore middle-school staff meeting, the teachers are told to stop their normal lessons and prepare students for the state test. Roland Pryzbylewski—Prez, a first-year mathematics teacher after a disastrous police career—asks the obvious question. If they teach the questions on the test, what will the test still measure about the children?

Grace Sampson gives him the answer of someone who has survived the system longer than he has: nothing. The test is measuring us. If the scores rise, the school can report progress. If they do not, it cannot.

Prez recognizes the operation at once. He has seen robberies reclassified as larcenies, rapes made to disappear, and commanders promoted when the numbers moved in the required direction. Sampson listens and says, in effect, that it is the same everywhere.

The exchange lasts less than a minute, yet it supplies a key to the whole series. A statistic begins as a description of something beyond itself: whether students learned, whether a neighborhood is safer, whether public work succeeded. Once it becomes the only result that is rewarded, it detaches from the thing it described. The test no longer measures the children; it measures the adults who report it. The clearance rate no longer tells the city how many murders were solved; it tells commanders whether their careers are safe.

No villain is required. The principal follows district policy, the district follows state requirements, the state answers to federal law. Each level can produce a document proving compliance. The Wire is less interested in the evil order than in the bad outcome completed by people who can all plausibly say they did their jobs.

Red Names, Black Names

Homicide keeps a board of the year's cases. Solved murders turn black; open murders remain red. The percentage that changes color is the clearance rate, and pressure travels down the hierarchy with it.

Marlo Stanfield discovers a way to kill without damaging the number. His people put bodies in vacant rowhouses, cover them with lime, and nail the buildings shut. The victims have not entered the police department's official reality. They are merely missing.

Lester Freamon notices that the housing department uses screws while some doors have been closed with nails. He opens one house, then another. Jay Landsman orders him to stop. The year is almost over, the unit's clearance rate is already poor, and moldy bodies with no witnesses will remain red forever. Landsman is not concealing murder for money. He is protecting a measure on which his unit is judged.

That is enough to create an extraordinary alignment. Marlo wants the bodies hidden so he cannot be prosecuted. Homicide benefits when the same bodies stay hidden because unsolved murders hurt the rate. A serial murderer and the department charged with investigating murder arrive at the same practical preference without conspiracy, contact, or corruption between them.

Bunk Moreland never accepts the calculation. To him, one body is one murder and somebody owes that dead person an answer. The show respects this old-fashioned standard while showing how institutionally inconvenient it has become. Around him everyone asks what will happen to the number. He keeps asking what the work itself requires.

The Fourteen Percent

When commanders learn that bad statistics will cost them their positions, most begin to juke the numbers. Major Bunny Colvin, nearing retirement after thirty years in the Western District, tries something else. He moves street dealing into three abandoned zones and announces that police will tolerate sales inside them while enforcing the law outside. The zones become known as Hamsterdam.

Violent crime in the district falls fourteen percent. This is not a reclassification or a case postponed until next year. Fewer people are shot; residents can use their streets; officers have time for other work. Colvin has delivered the result every public speech says the city wants.

The institution destroys him for it. He is demoted, forced into early retirement, and made to take a smaller pension. The reason is not that the number is false. It is that the statistic was never simply evidence that the district had become safer. It was evidence that a commander had performed the approved actions and obtained an acceptable report.

A good result produced by an unauthorized method is dangerous because it becomes an accusation. It suggests that the prescribed method—and therefore the authority of everyone promoted through it—may not have been necessary. City leaders debate concealment, reelection, and political advantage. Almost no one asks the elementary question: were fewer people hurt?

The measure was supposed to answer that question. Once the institution owns the measure, the question itself drops out.

What Cannot Enter the Form

Back in the classroom, Prez first complies with test preparation. The students withdraw. Then he takes out dice and teaches probability through the street game they already know. The room wakes up. Children who have been calculating odds every day discover that what they know has a mathematical name.

In the same school, Colvin joins an academic project that removes the most disruptive students from ordinary classes. Instead of forcing another lesson plan on them, the adults ask how fights begin, how money and status work on the corner, and what the children expect from their lives. The change is visible. For the first time, some of the students are being addressed as people whose reasons matter.

Both efforts are stopped. Test season returns the students to conventional classes, and the pilot cannot produce the kind of result that protects a program inside the school system. The most important developments—attention, trust, the first experience of being taken seriously—have no secure box on the form.

The tragedy is not merely that institutions measure the wrong thing. It is that what can be counted can be defended in budgets and meetings, while what cannot be counted depends on someone continuing to see it. When that person leaves, the work disappears as though it never happened.

The Man Who Did Not Exist

In the final season, Baltimore cannot afford police work. Overtime is frozen, surveillance money is gone, and the investigation of Marlo—responsible for more than twenty bodies—stalls. McNulty responds by manipulating the corpses of unhoused men and inventing signs of a serial killer.

The press repeats the story. City Hall notices. Money appears. McNulty diverts the resources assigned to the imaginary killer toward the real investigation.

The direction has reversed, but the operation is familiar. For years the department turned events into non-events: robbery into theft, murder into a missing person, an open case into next year's problem. McNulty turns natural deaths into a sensational crime. He does not fabricate for promotion. He fabricates because, within this system, counterfeit police work has become the only way to fund real police work.

Bunk is appalled. He tells McNulty that using the dead this way has made the job into something else. His anger matters because he is one of the few characters who still carries a standard the institution did not give him and cannot easily revise.

“Data manipulation,” “metric distortion,” and “bureaucratic formalism” all name what happens here. But those phrases can let us leave too quickly. The harder question is what the number does to the person holding it. Prez, Landsman, Colvin, and McNulty differ in character, rank, and intention. Each is gradually asked to answer for the visible result instead of the work.

Given enough time, the question “Did I do this well?” becomes “Does it look complete?” Only the second can be answered entirely by someone else.

What is the number in your own work? If it vanished tomorrow, would you still know how to judge whether the work was good? And when did the standard you once used move out of your hands?

The Wire, created by David Simon, aired from 2002 to 2008 across five seasons and sixty episodes. This essay relies on the series itself.

版本说明:中文原作;简体与繁体同页切换。正文依底本完整排入,保留分节、强调与引文。全剧剧透。

1.1 "它评估的是我们"

一间中学的教师会议室。

校方在传达新的教学安排:接下来一段时间,各科课时全部让给州统考的模拟题训练。请翻到第十一页。

坐在后排的是罗兰·普莱兹博,八年级数学老师,任教第一年。他举手问了一个问题:

我不太明白。这么做只是为了统考分数高一点?如果我们教的就是考卷上的题目,那这个考试到底在测量学生的什么?

旁边的格蕾丝·桑普森老师回答了他。

她说:什么也不测量。它测量的是我们。

分数上去了,他们就能说学校在进步。分数不上去,他们就说不了。


普莱兹博愣了一下,然后说了三个字:

做数字。

桑普森说:什么?

他说:把抢劫案改记成盗窃案。让强奸案消失。数字做够了,少校就能升上校。

他说:这个我干过。

桑普森老师看了他一眼,说了一句:走到哪儿都一样。


这一段对话不到一分钟,是整部《火线》的入口。

值得先停在桑普森那句回答上。

那个考试被设立出来,本来是为了衡量孩子有没有学到东西。

而她说得很准确:它现在衡量的是老师。

一个数字被造出来的时候,它指向别的东西——学生学得怎么样、街区安不安全、这座城市的人过得好不好。等它变成唯一被奖励的那样东西,它就脱钩了。它不再指向那件事,它开始指向报告它的那些人。

一个数字最初是用来描述某件事的。等它变成唯一被计分的东西,它就变成了那件事本身。

而说出这句话的人,先后在两个毫无关系的机构里工作过——警察局和公立学校,中间隔了十年。他在一场例会上,用了不到一分钟就认了出来。


还有一点必须先说清楚,否则后面九章都会被读歪。

那间会议室里没有一个坏人。

校长在执行学区的要求。学区在执行州里的要求。州里在执行联邦法案的要求。每一层都在合规。每一层都可以拿出文件来证明自己做的事情是对的。

普莱兹博自己也会照做——他后来确实教了几个星期的模拟题。

这一章要讲的东西,全程没有一个人在作恶。

这是《火线》和大多数讲机构的作品最大的区别。它不去找那个下命令的坏人,因为它认为那个人不存在。它感兴趣的是:当没有任何人作恶的时候,一件坏事是怎么被完整地做完的。


1.2 红字与黑字

凶杀组的墙上有一块板子,上面是这一年的所有案子。

破了的案子写成黑字。没破的写成红字。

红转黑的比例,叫破案率。

这个数字往上走,中队长有面子,副局长有面子,市长有面子。这个数字往下走,从上到下都要挨骂。整部剧里,从副局长罗尔斯到中队长兰德斯曼,压力沿着这条线一层一层往下传。


第四季,马洛·斯坦菲尔德接管了西区。

他杀人的方式和以前的人不一样。他不把尸体扔在街上。他的人把尸体搬进空置的排屋,撒上生石灰,然后把门窗封死。

于是那些人没有死。

他们只是不见了。没有报案,没有现场,没有编号,没有出现在那块板子上。

莱斯特·弗里蒙注意到了一件事:市住房局封空屋用的是螺丝,而有些屋子是用钉子封的。他开始一间一间地撬。

第一具,第二具,第三具。

然后兰德斯曼赶到现场,命令他停下。

理由说得非常清楚:还剩三个星期就到年底,全组破案率不到五成。我们不主动去找尸体,尤其是发霉的无名尸。我们不自愿往板子上添红字。

弗里蒙说,这些人是马洛杀的。

兰德斯曼说:那他们就归他。

然后他撂下一句命令:你他妈不许再撬一块木板。


这里必须公平一点。

兰德斯曼不是坏人。他风趣、老练、护着自己的手下,在这部剧里是个相当可爱的人。

而他说的话,在这个部门自己的逻辑里,是完全正确的。那些尸体是破不了的——没有目击者,没有现场,没有时间线。它们会在板子上一直红下去,把这一年的数字彻底毁掉。

他不是在掩盖谋杀。他是在保护一个数字。

而在这个位置上,这两件事已经不容易分开了。


现在看这件事的形状。

马洛把尸体藏起来,是为了不被起诉。

凶杀组需要尸体不要浮出来,是为了保住破案率。

一个连环杀人的毒贩,和负责调查凶杀案的警察,在"这些尸体不要被找到"这件事上,利益是一致的。

没有人安排过这件事。没有人受贿,没有人通气,双方甚至不知道对方存在。他们的目的完全相反,而中间那个数字,让他们走到了同一个结论上。

这个共识维持了将近一年。

二十多个年轻人从西巴尔的摩消失,而这座城市的官方立场是:什么也没发生。


这部剧里有一个人始终没有接受这套算法。

邦克·莫兰。

他办案的方式很老派:一具尸体就是一桩凶杀案,得有人为它负责。他会为了一个跟破案率毫无关系的案子熬几个星期,会因为同事把案子往下推而当场翻脸。

剧集对他的态度很微妙——它尊重他,也把他拍成一个不合时宜的人。整个部门都在按数字运转,只有他还在问那个更老的问题:

这件事,做得对不对。


1.3 百分之十四

第三季,压力到了顶点。

局长伯勒尔把东区的一位分局长撤了,理由是数字太难看。信号非常清楚地传下去:数字不好看的人会被撤职。

于是各区的分局长开始做数字。

只有西区的分局长邦尼·科尔文没有做。他照实上报,报出来的结果是重罪上升百分之二。

他快退休了,还剩一年。他在这条街上当了三十年警察。

然后他做了一件事。


他在自己辖区里划出三块地方,全是空屋区,然后把话传下去:所有街角的人都搬到这三块地方去卖,在里面卖,警察不管;在外面卖,就抓。

那三块地方后来被人叫作"仓鼠特丹"。

结果是这样的:

暴力犯罪在西区下降了百分之十四。


请注意这个数字的性质。

它是真的。它不是把抢劫改记成盗窃,不是把尸体封进空屋,不是把案子推到明年。西区确实有更少的人挨枪子,确实有更少的老太太不敢出门,确实有一批警察终于腾出手来做别的事。

百分之十四,是所有人在公开场合都说自己想要的那个东西。

而当这件事被捅出来之后,科尔文完了。

降级,被迫提前退休,退休金按降下来的级别算。三十年,最后一年,全部作废。


他给了这个系统一个它公开索要的数字,然后被这个系统毁掉了。

为什么?

因为那个数字从来不是重点。

那个数字是一份合规证明。它证明的不是"这个区更安全了",而是"这个人执行了规定动作,并且执行得不错"。

一个用规定动作以外的方法做出来的数字,不是一个更好的数字。

它是一份指控。

它说的是:那套规定动作其实是不必要的。这句话如果成立,那么上面所有靠执行规定动作升上去的人,位置都成了问题。

所以科尔文必须被毁掉,而且要毁得彻底,好让这个数字连同它的来路一起作废。


值得看一遍那间屋子里的反应。

伯勒尔的第一反应是压住,不能让报纸知道。市长罗伊斯犹豫过——如果真的管用,能不能悄悄留着,反正他要竞选连任。竞选对手卡塞蒂拿到消息之后,把它当成了一枚炸弹。

这些反应各不相同,而它们有一个共同点:

从头到尾,没有一个人问过那个问题——

那边受伤的人是不是变少了?


1.4 教室

回到那间中学。

模拟题训练开始之后,普莱兹博试着照做。他念题,学生趴着。他讲解,学生看窗外。

后来他放弃了课本。

他从口袋里掏出两个骰子。

他开始讲概率——两个骰子扔出七的可能性有多大,扔出蛇眼的可能性有多大,为什么街上那个赌局长期看你一定输。

教室活了。

这些孩子里有一半在街角混,天天在算这个东西,只是没人告诉过他们这叫数学。他们开始抢答,开始争,开始自己算。

那是这部剧里为数不多的、能让人高兴一下的场面。

它不出现在任何一份报表上。


同一所学校里,科尔文——退休之后到大学当了研究助理——和一个学者做了一个试点项目。

他们把最难管的那批孩子从普通班里挑出来,单独放进一间教室。不上课本,就是坐下来说话:为什么打架,怎么算钱,家里什么情况,将来打算干什么。

那间教室里发生了真实的变化,肉眼可见。

项目被叫停了。

不是因为有人反对,是因为统考快到了,那些孩子必须回到原来的班上去考试;也因为这个项目产出的东西,学校数不出来。

一间教室里的十几个孩子第一次被人当人看——这件事没有对应的表格。


能被计量的东西才能被保护。而最要紧的那几样,恰好数不出来。

这条线在第五章会专门展开。这里只需要记住它第一次出现的样子:

两件真正有效的事情,同时发生在同一所学校里。

一件是自发的,一件是设计好的。

两件都被停掉了,而且都不是因为有人认为它们不好。


1.5 编出来的那个人

第五季,警局没钱了。

市里的财政被学校的赤字掏空,加班费停了,办案经费砍了,弗里蒙那条查马洛的线断了——没有钱做监听。

一个杀了二十多人的毒贩就在那儿,而这座城市付不起查他的钱。

于是麦克纳尔蒂做了一件事。


他去了几具无人认领的流浪汉尸体的现场。他动了现场。他在死者脖子上留下痕迹,让自然死亡看起来像勒杀。他前后弄了几次,把这些案子串成一个模式。

然后他放出风声:城里有一个专杀流浪汉的连环杀手。

报纸接上了。市长接上了。钱下来了。

一大笔专案经费,一支专案组,全部用来抓一个不存在的人。

而麦克纳尔蒂把这笔钱挪去做了别的事——监听马洛。

那条真正的线,靠一个虚构的凶手养活了。


这件事在剧里被拍成了一个荒唐的、越滚越大的灾难。但它的结构值得看清楚。

这个部门花了五年时间,把抢劫改记成盗窃,把尸体留在空屋里,把案子推到下一年。

麦克纳尔蒂把"什么都没发生"改记成了一个连环杀手。

方向反了,动作是同一个。

当一个数字变成唯一被奖励的东西,最有效的做法就不再是把事情做好,而是直接生产那个数字。

而麦克纳尔蒂这件事最难受的地方在于:他不是为了升职。他是为了办那个真案子。

也就是说,在这个系统里,做真正的警察工作,唯一的路径已经是造假

同一季,报社里有个记者在编造引语和人物。他靠这个拿了普利策奖。这件事第六章会专门讲。


这部剧里有一个人当面说了出来。

邦克对麦克纳尔蒂说的那段话,语气里没有一点玩笑。他说这他妈是最下作的事,说你在拿死人开玩笑,说你把这份工作变成了别的东西。

他说完就走了,很长时间不跟他说话。

在一部几乎没有道德说教的剧里,这是少有的一次。而说这话的人,正是那个还在问"这件事做得对不对"的人。


我想留一个问题。

这一章讲的现象有很多现成的说法——数据造假、形式主义、指标异化。这些说法都对,也都太好用了,用完之后不会剩下什么。

我更想留下的是另一件事:一个数字对拿着它的那个人做了什么。

普莱兹博、兰德斯曼、科尔文、麦克纳尔蒂,这四个人处境完全不同,能力和品性也完全不同。而他们每一个都被要求对一个数字负责,而不是对自己的工作负责。

这件事持续得足够久,会发生一个很安静的变化:

一个人渐渐不再问"我这件事做得好不好",只问"看起来做完了吗"。

这是两个不同的问题。

而其中只有一个,是别人能替你回答的。


所以:

你的工作里,有没有一个数字?

如果明天那个数字取消了,你还知道用什么来判断自己这件事做得好不好吗?

如果这个问题让你卡了一下——那么可以再想一想:

你原本用来判断的那个标准,是什么时候从你手里挪走的?


下一章讲的地方,看起来是这座城市里最没有规矩的地方。

它有等级,有分工,有纪律,有继承,有对背叛的惩罚,也有对忠诚的回报。它运转良好,出了事有人扛,赚了钱按规矩分。

它比这部剧里任何一个合法机构都更像一个组织。

第二章:街角。


《火线》,David Simon 创作,2002 至 2008 年播出,共五季六十集。

← Series overview← 系列总览Next: The Corner →下一篇:街角 →