跳至正文

一公斤到底多重

目录

本篇四条红线，置于篇首而不置于篇末——四条同时成立，缺一条这篇就是废的。

第一条（防升格，本篇首要红线）：本篇审的是「SI 是怎么把尺子做对的」，不是「SI 是完美的」或「SI 是正确的唯一方式」。 本篇会全程赞美计量学的制度安排——组织化、可比较性、不确定度自觉、定义可复现、失败被看见——但每一句赞美都要在同一段里配上它的边界：SI 手册自己承认非 SI 单位「广泛使用且预计继续多年」；GUM 自己承认 k=2 的 95% 置信只是近似；Le Grand K 的 50 微克漂移至今没有解释；美国 1975 年至今的公制化是立法与行政分支双双承认的未完成工程。任何把「计量学做对了」读成「测量问题已经全部解决」或「凡是报不确定度的就一定是准的」的读法，本篇明确拒绝。

第二条（防虚无化，写在防升格后面）：本篇审的是尺子，不是尺子存在的价值。 SI 的每一次定义更迭、GUM 的每一个条款、BIPM 的每一场关键比对，都是真实运转了一百五十年的制度设施；「尺子会被污染」「测量有不确定度」「Le Grand K 会漂移」这些事实不是「尺子都是假的、测量都是骗人的」的证据——恰恰相反，它们是被这套制度自己测量出来并公开的。任何把本篇的边界清单读成「计量学也不可信、所有测量结果都可以怀疑」的读法，本篇明确拒绝。

第三条（防污名化，涉及唯一一个未完成的制度工程）：本篇谈美国公制化失败时，只引法案原文、总统声明与官方现状，不做动机推断，不评价任何一届政府、任何机构、任何个人的诚信与能力。 1975 年「自愿」公制化是白纸黑字的立法选择，1988 年「preferred system」是白纸黑字的政策指定，公路标志用英里是 1995 年公法的逐字规定——这些是档案里的事实，不是品格问题。同理，本篇不评价 BIPM 或任何国家计量院、不评价任何在世研究者。

第四条（自指与本篇的位置）：本篇是正面范本，但正面范本同样要过自己的纪律。 本库做了一百一十六篇分母账、口径账、尺子账、不确定度账——这些武器的生产厂就是本篇的对象。所以本篇把本库自己的手法逐件拿去与计量学的对应物对照：本库的「分母账」对应计量学的「溯源链」，本库的「口径账」对应「定义与复现分离」，本库的「不确定度标注」对应 GUM，本库的「去重实测」对应「关键比对」，本库的「编号与注册」对应 CGPM 的决议与 SI 手册的版本管理。对照的结果是：每一件本库当武器用的东西，在计量学里都是先当义务用了几十年的制度设施——本篇落点即在此。

核验图例（全篇通用）：[一手逐字] ＝ 取回原文全文或官方页面逐字比对；[摘要逐字] ＝ 官方摘要逐字，正文未取回；[多源交叉] ＝ 权威文献或官方文件转引逐字，原始正文未取回；[我们的断言] ＝ 本篇自己的推论与算术，不冒充文献结论。六捆并行调研代理取回后，主笔对全部承重引用又做了一轮逐字亲核：SI 手册第 9 版 PDF（BIPM 官方，2.0MB）全文文本亲核约 40 处短语（七常数表、四个重新定义单位逐字定义、序言连续性三句、附录 4 历史注释、Table 8 非 SI 单位段），GUM（JCGM 100:2008，1.9MB）全文文本亲核约 20 处短语（2.2.3 定义、3.2.2 NOTE 2、3.3.1、6.3.3、G.6.6），第 26 届 CGPM 会议记录（6.0MB）投票段亲核（54 国唱名、一致通过），VIM3（JCGM 200:2012）2.41/2.42 溯源定义亲核，GUM S1（JCGM 101:2008）引言亲核。本篇所有官方文件均为直取件，无 Wayback 兜底件参与承重。

去重与分界声明（八条）。 本库已有 174 篇，本篇必须先说清自己占哪块地。全库 python 逐词扫描（604 万字符）：国际单位制／国际计量／BIPM／CGPM／计量学／度量衡／原器／坎德拉／SI Brochure／瓦特天平／MRA 全部零专篇命中；命中全部是旁注——拓扑物质篇用量子霍尔电阻标准当「计量铁证」（4 次）、块宇宙篇与测量代理篇用原子钟/秒定义作案例、AI 评测篇用过 PTB 一词。这是处女地，也是本库元科学疆域里第一篇「正面范本」。邻居如下：

  1. 与拓扑物质篇分界：那篇把量子霍尔效应的电阻量子化当「拓扑物质被计量铁证」的证据用；本篇审计量学本身，量子霍尔电阻标准（RK-90 约定值及其 2019 后的地位变化）在本篇是对象不是证据。
  2. 与测量代理性篇分界：那篇是方法学——「当测量被代理化，会发生什么」；本篇是正面范本——「当测量把不确定度变成义务，会建成什么」。秒定义在那篇是 proxy 的例子，在本篇是 SI 定义账的案例。
  3. 与AI 评测榜篇构成对偶：那篇审「尺子被污染、被测方自己修尺子、分数=能力的升格链」（MMLU 29.1% 泄漏、SWE-bench 官方判死刑）；本篇审「尺子被做对」的教科书。那篇的落点「测量学的答案从来不是『测准』，是『知道自己测不准多少』」在本篇是 GUM 的字面操作规程——本篇是那篇落点的制度史。
  4. 与同行评议篇分界：那篇审科学管自己质量的两台机器（评议、引用指标）的越界；本篇审科学管自己尺子的机器（SI、GUM、比对）的建制。那篇是「机器坏了该不该全废」的裁决，本篇是「机器做对了长什么样」的解剖。
  5. 与司法鉴定篇分界：那篇里 DNA 是「高质量单一来源」的校准物、潜指纹等四类是待审方法；本篇不审任何具体测量方法，只审「把测量结果与 SI 连起来的那条链」本身。
  6. 与地震预测篇的 CSEP 对照：CSEP 是前瞻性比对基础设施（「把预测送出去与人比」），本篇的关键比对（KCDB）是同一设计在测量上的应用——本篇只做交叉引用，不重做那篇的裁决。
  7. 本篇不裁任何「某某测量有多准」的具体争议（引力波参数的版本差、微塑料的测量危机、表观时钟的 ICC 都归各自篇章）；本篇只裁「把不确定度报出来、把尺子送出去比」这套制度。
  8. 本篇与「科学是不是进步」「度量衡是不是殖民工具」的宏大叙事无关。法国大革命公制化的政治史（Alder《The Measure of All Things》的史学批判）本篇只作背景交代，不进入裁决；本篇通篇讲的是制度的承重方式。

〇、母裁决：三件事，一个落点

本库做了一百一十六篇「分母账、口径账、尺子账」。这些武器不是本库发明的——它们是从计量学这门学科里借来的，而计量学把这套纪律制度化地运转了一百五十年。本篇不问「尺子准不准」，问「尺子是怎么被做对的」——因为如果这个问题的答案可判定，本库过去一百多篇的整个手法就有了一个可以回去核对的母本。

第一件：SI 是真实的制度成功，而且它的成功是可以逐条拆解的。 1875-05-20 巴黎签署的米制公约建立了一个至今运转的三机构架构（CGPM 决策、CIPM 执行、BIPM 常设实验室）[一手逐字]；1960 年第 11 届 CGPM 决议 12 正式命名 SI[一手逐字]；2018-11-16 第 26 届 CGPM 决议 1 以 54 个出席成员国唱名表决一致通过，把全部七个基本单位锚定到七个定义常数上，「The numerical values of the seven defining constants have no uncertainty」[一手逐字]；2019-05-20 生效，选择该日期的理由是它既是世界计量日、又是米制公约签署纪念日[一手逐字]。全世界 64 个会员国、36 个准会员[一手逐字]，CIPM MRA 从 1999 年 38 家国家计量院起步扩到今天 258 家机构[一手逐字]，KCDB 收录 1,278 项关键比对与 26,270 项校准测量能力[一手逐字]——这些是数得出来的制度资产，不是修辞。

第二件：做对不是「更准」，是四件事的合题——而这四件事每一件都有原件的折痕。 ①定义不依赖唯一实物：SI 手册第 9 版 1.2 节逐字——实物定义「involve the risk of loss, damage or change」，而常数定义让单位「as a matter of principle, be realized independently at any place and at any time」[一手逐字]；2019 年是「For the first time, a complete set of definitions is available that does not make reference to any artefact standards」[一手逐字]。②不确定度是测量结果的组成部分：GUM 3.1.2 逐字——测量结果「is complete only when accompanied by a statement of the uncertainty of that estimate」[一手逐字]；GUM 3.2.2 NOTE 2 逐字——「error」与「uncertainty」「are not synonyms, but represent completely different concepts」[一手逐字]。③尺子要送出去与人比：VIM3 2.41 把计量溯源定义为「a documented unbroken chain of calibrations, each contributing to the measurement uncertainty」[一手逐字]；CIPM MRA 让国家校准证书跨国互认[一手逐字]。④失败要被看见：Le Grand K 与四十只副本间的质量漂移——BIPM 官方数字是中位约 25 微克（第三次周期检定）[一手逐字]，NIST 官方口径是约 50 微克/百年[一手逐字]，而 BIPM 作者自己写下「There is no explanation available for the differences」[一手逐字]。这四件事合起来才是「做对」——单独任何一件都是不够的。

第三件：「做对」是有边界的，边界同样写在原件里。 SI 手册自己承认非 SI 单位「widely used and that this is expected to continue for many years」并给出换算表（升、吨、电子伏、天文单位、道尔顿全在表内）[一手逐字]；美国 1975 年公法把公制化定为「voluntary conversion」[一手逐字]，1995 年公法逐字禁止联邦强迫各州用公制设公路标志[一手逐字]，CIA 官方至今把美国列为三个未全面采用 SI 的国家之一[一手逐字]；GUM 自己承认 k=2 的 95% 置信「approximately 95 percent」且有条件[一手逐字]，其官方补充文件 S1 逐字列出 GUM 框架的失效条件（模型线性化不足、输出分布明显偏离正态）[一手逐字]，JCGM 至今在修订 GUM[一手逐字]。「做对」从来不是「做完」——本篇说的每一句赞美，同一个句号里都要带上一句边界。

落点：尺子做对与做错的分界线不是精度，是三个问题的答案——定义能不能让任何人都能复现（不依赖唯一实物）；测量结果有没有随身携带不确定度（不是测准，是知道自己测不准多少）；造尺人有没有把尺子送出去与人比对（自己说自己准不算数）。 本库一百余篇的分母账、口径账、尺子账、不确定度账，没有一件不是这三个问题的变形；计量学是这些武器的正厂——「知道自己测不准多少」不是谦辞，是 SI 自 1875 年以来运行的完整操作规程，是写在 GUM 第 3.1.2 条里的字面义务。

结构胎记：正面范本——不是「三跳」，是「三步不跳」

本库既往一百一十六篇的结构胎记几乎全是升格链——把「关于我们的话」读成「关于它的话」，把「相关性」读成「因果」，把「一份文件」读成「一个共识」。本篇的结构胎记不是跳，而是不跳——但它不是「没有跳」这种空洞的肯定，而是「跳的口子都被制度堵上了」这种可检验的陈述。拆成三步：

第一步·定义不跳：定义与实物分离，任何人都能复现。 SI 手册 1.2 节逐字说明了这个设计：实物定义「the definition and the realization are equivalent」（1901 年前的一公斤就是「那块铂铱圆柱的质量」），而常数定义把二者分离，「realized independently at any place and at any time」[一手逐字]。判定「定义跳」的标准动作——「唯一的原件在唯一的地方、丢了就没了」——在 SI 里没有对应的主语：2019 年后没有任何一个基本单位的定义挂在任何一件实物上，连 IPK 都只是历史文物（「will not be on display for the general public」[一手逐字]）。

第二步·不确定度不跳：测量结果的完整性义务由 GUM 承担。 本库一百余篇反复抓的毛病是「把测量读成确切数字、把代理读成本体」——GUM 3.1.2 逐字给出反方向的制度答案：「the result of a measurement … is complete only when accompanied by a statement of the uncertainty」[一手逐字]。在计量学里，「报不确定度」不是可选的严谨性，是「测量结果」的定义性组成部分；ISO/IEC 17025 把这条义务延伸到校准实验室（7.6 节要求识别所有显著不确定度贡献），ILAC P14 政策进一步规定校准证书上必须写 y ± U 与覆盖因子[一手逐字/多源交叉]。

第三步·比对不跳：自己说自己准不算数。 VIM3 2.41 的溯源定义逐字要求「a documented unbroken chain of calibrations」[一手逐字]；CIPM MRA 与 KCDB 把这个要求制度化——国家计量院的关键比对、校准测量能力清单、互认协议[一手逐字]。本库的「去重实测」「交叉验证」「编号注册」在计量学里都有建制化的对应物：关键比对、CMC 清单、CGPM 决议与 SI 手册版本管理。

而这三步的每一步，都伴随一个「跳了会怎样」的对照组：法国大革命前约 25 万种度量衡单位（当代人估计）[多源交叉]是「没有定义账」的世界；Le Grand K 的 50 微克漂移无解释[一手逐字]是「定义挂实物」的必然代价；美国公制化失败[一手逐字]是「组织化缺席」的后果。本篇的骨架就是这三步不跳 + 三个对照组——正面范本的意义恰恰在于它给出了「跳」的成本可以被量化的实例。

一、守真锚：SI 是真实的制度成功，且成功可逐条核数

本篇是正面范本，所以第一层不是「现象是不是真的」，而是「制度成功是不是真的、大到什么程度」——这层也必须像任何守真锚一样逐条给原件，因为正面范本最忌讳的就是把「制度成功」当成不用证明的前提。

公约与机构：1875-05-20，巴黎，17 国。 BIPM 官方页逐字：「The Convention to ‘assure the international unification and improvement of the metric system’ and its annexed regulations (commonly known as the ‘Metre Convention’), was signed on 20 May 1875, and amended in 1921. It is an international treaty, the purpose of which was the creation of an international organization called the BIPM.」并列出 17 个创始签署国：德国、阿根廷、奥匈、比利时、巴西、丹麦、西班牙、美国、法国、意大利、秘鲁、葡萄牙、俄罗斯、瑞典与挪威、瑞士、土耳其、委内瑞拉[一手逐字]。公约正文第一条（官方英译）逐字：「The High Contracting Parties undertake to create and maintain, at their common expense, a scientific and permanent International Bureau of Weights and Measures with its headquarters in Paris.」[一手逐字]。

三机构架构至今原样运转。 BIPM 治理页逐字：「The governance structure established by the Metre Convention follows the classical three-body institutional architecture of most international organizations. The Director of the BIPM and the staff, is subordinated to the CIPM, which is subordinated to the CGPM.」[一手逐字]；SI 手册第 9 版「The BIPM and the Metre Convention」章逐字：「Delegates from all Member States attend the General Conference, which normally meets every four years」「The CIPM has eighteen members each from a different State」[一手逐字]。会员规模（2026-08-02 实查）：「There are currently 64 Member States and 36 Associate States and Economies.」[一手逐字]。

SI 的命名：1960 年第 11 届 CGPM 决议 12。 BIPM 决议数据库逐字：「1. the system founded on the six base units above is called the ‘Système International d’Unités’; 2. the international abbreviation of the name of the system is SI」[一手逐字]。

SI 手册：权威文本与其版本管理。 BIPM 官方页逐字：「The SI is defined by the SI Brochure, published by the BIPM.」「From 20 May 2019 all SI units are defined in terms of constants that describe the natural world.」[一手逐字]；第 9 版 2019 年发布（当前 V4.01，2026-06），第 8 版 2006-05，BIPM 官方页确认「The 9th edition of the SI Brochure is the current valid version」[一手逐字]——注意「版本管理」本身就是制度的一部分：SI 手册有版本、有 DOI、有引用格式，本库的编号与注册纪律在这里有建制化的对应物。

七个定义常数（SI 手册 Table 1 逐字段，本库主笔亲核）[一手逐字]：

定义常数 固定数值 单位
铯-133 基态超精细跃迁频率 ΔνCs 9 192 631 770 Hz
真空光速 c 299 792 458 m/s
普朗克常数 h 6.626 070 15 × 10⁻³⁴ J s
基本电荷 e 1.602 176 634 × 10⁻¹⁹ C
玻尔兹曼常数 k 1.380 649 × 10⁻²³ J/K
阿伏伽德罗常数 NA 6.022 140 76 × 10²³ mol⁻¹
频率 540 × 10¹² Hz 单色辐射的光视效能 Kcd 683 lm/W

该表正下方的句子（SI 手册逐字，本篇的守真锚之锚）：「The numerical values of the seven defining constants have no uncertainty.」[一手逐字]——「定义常数没有不确定度」这句话不是狂妄，而是定义系统的自洽性陈述：不确定度被全部赶到「复现」一侧去了，而这正是下一层的主题。

这一层能数出来的全部家当：64 会员国＋36 准会员、三机构架构运转 151 年、SI 手册第 9 版现行、7 个定义常数、1,278 项关键比对、26,270 项校准测量能力、258 家互认机构[一手逐字]——SI 的成功是「可以被审计的家当清单」，不是「感觉上很成功」。本库的方法在这里找到了第一个对应物：守真锚不是修辞，是清单。

二、做对之前：法国大革命度量衡——「没有定义账」的世界长什么样

正面范本需要一个对照组，否则「做对」就成了没有参照的形容词。前 SI 世界就是对照组：约 25 万种单位（当代人估计）[多源交叉]、每一个镇子有自己的尺、度量衡是领主权力的一部分。这一层的任务不是审判历史，是问「没有定义账的世界，代价是什么」。

「约 25 万种」的出处与口径。 Ken Alder《The Measure of All Things》(2002) 序章逐字：「Contemporaries estimated that under the cover of some eight hundred names, the Ancien Régime of France employed a staggering 250,000 different units of weights and measures.」[多源交叉]；Agnoli & d’Agostini (arXiv 2004，后刊于 Metrologia 2005) 转述：「about 800 different names for measures have been estimated, whose units varied in different towns, for a total of about 250000 differently sized units」[多源交叉]。注意口径：这是「contemporaries estimated」——当代人自己的估计，不是现代史学的普查，本篇如实标注，不把它当成精确统计。同样真实的是史学界的另一侧：Kula《Measures and Men》(1986) 的核心论点恰是革命前的度量衡「有与其用途相称的精度与表征功能」[多源交叉]——「混乱」是制度视角的叙述，不是价值判断；本篇引 Alder 的数字只为说明「没有一个统一定义账」的世界，不为说旧制度「落后」。

1795-04-07（法历共和三年芽月 18 日）国民公会法令：定义账第一次被写下来。 该法令原件被 UNESCO 世界记忆计划收录，其档案描述逐字：「a more thorough law was published on 18 Germinal, year III (7 April 1795): this was the law (‘decree’) (e) on the decimal metric system, which affirmed the need to make a provisional platinum standard metre」[一手逐字]；法令英译逐字：「Meter, the measure equal to one-ten millionth of the arc of the terrestrial meridian included between the north pole and the equator;」以及「There is only one standard of weights and measures for the entire Republic; there shall be a platinum ruler on which will be marked the meter」[一手逐字]；法语原文（Gallica 一手印刷本）逐字：「on appellera Mètre, la mesure de longueur, égale à la dix-millionième partie de l’arc du méridien terrestre, compris entre le pôle boréal et l’équateur」[一手逐字]。注意：米最初的定义是「地球子午线四千万分之一」——把定义挂在行星上，就是第一代「不依赖唯一实物」的努力，而它立刻撞上了实测困难（子午线弧长测量误差），于是 1799 年退回实物原器。

1799 年铂原器入藏：定义退回实物，日期两说。 SI 手册第 9 版附录 4 逐字：「The creation of the decimal metric system at the time of the French Revolution and the subsequent deposition of two platinum standards representing the metre and the kilogram, on 22 June 1799, in the Archives de la République in Paris, which can be seen as the first step that led to the present International System of Units.」[一手逐字]；法国科学院 Comptes Rendus 论文（Débarbat 2019）同口径：「le mètre définitif en platine, qui a été déposé aux Archives nationales le 4 messidor an VII (22 juin 1799)」[一手逐字]。另一家（Johnson Matthey 技术评论 2000）写「On 22nd July 1799 the definitive standards … were ceremonially deposited in the French National Archives」[一手期刊，日期与主流口径分歧，本篇并列如实标注]。

1812 年拿破仑恢复习惯单位：定义账第一次被撤。 1812-02-12 皇帝敕令原件（法国矿业年鉴影印件）逐字：「Il est permis d’employer pour les usages du commerce … une mesure de longueur égale à deux mètres, qui prendra le nom de toise, et se divisera en six pieds」——「toise」（英寻）被复活为「恰好两米、分成六法尺」的商用单位[一手逐字]。1837-07-04 法律逐字：「Le décret du 12 février 1812, concernant les poids et mesures, est et demeure abrogé.」并把旧器具的使用宽限到 1840-01-01[一手逐字/多源交叉]。1812 到 1837：一次定义账的撤销与恢复，全程 25 年。

这一层的三个观察（本库断言）：①「没有定义账」的代价不是「测量不准」，是每次交易的尺子都是谈判的一部分——贸易纠纷的源头不是误差而是「谁的尺」；② 第一代公制定义（挂行星）比第二代（挂实物）更符合「可复现」理想，却死在实测精度上——「定义不依赖实物」不是一蹴而就的原则，是被两百年试错试出来的；③ 1795/1799/1812/1837 的来回证明「定义账」是一种需要政治制度维持的东西，SI 后来的 151 年稳定不是自然现象，是米制公约这个组织化决策的产物——下一层就是它的定义账。

三、定义账：六次更迭，每一次都把「实物」换掉一步

这是本篇的第一笔独占账：SI 的每一次定义更迭都是同一个方向的移动——把「定义挂在实物上」换成「定义挂在常数上」，而每一次的决议文本都逐字可查。SI 手册第 9 版附录 1 收录了 1889 年以来全部 CGPM/CIPM 决议原文[一手逐字]，本篇按此亲核。

1889 年：原器时代开始。 第 1 届 CGPM 决议「Sanction of the international prototypes of the metre and the kilogram」逐字：「The Prototype of the kilogram adopted by the CIPM. This prototype shall henceforth be considered as the unit of mass.」[一手逐字]；1901 年第 3 届 CGPM 声明逐字：「The kilogram is the unit of mass; it is equal to the mass of the international prototype of the kilogram」[一手逐字]。注意这个定义的自洽性缺陷：它把「单位」定义在「一件物品」上——千克的「定义」与「实物」合一，这正是 SI 手册 1.2 节后来写明的「artefacts involve the risk of loss, damage or change」[一手逐字]。

1960 年：米先离开实物。 第 11 届 CGPM 决议 6 逐字：「The metre is the length equal to 1 650 763.73 wavelengths in vacuum of the radiation corresponding to the transition between the levels 2p10 and 5d5 of the krypton 86 atom.」[一手逐字]——米从「一根铂铱棒」变成「一种光的波长数」，定义第一次不依赖实物。

1967 年：秒变成原子频率。 第 13 届 CGPM 决议 1 逐字：「The second is the duration of 9 192 631 770 periods of the radiation corresponding to the transition between the two hyperfine levels of the ground state of the caesium 133 atom」[一手逐字]。1968 年起，秒成了第一个完全挂在原子跃迁上的基本单位——铯原子的基态超精细跃迁频率后来成为 2019 年七个定义常数中的第一个（ΔνCs）。

1983 年：米变成光速。 第 17 届 CGPM 决议 1 逐字：「The metre is the length of the path travelled by light in vacuum during a time interval of 1/299 792 458 of a second」[一手逐字]，并逐字废止 1960 年定义：「The definition of the metre in force since 1960, based upon the transition between the levels 2p10 and 5d5 of the atom of krypton 86, is abrogated.」[一手逐字]——「abrogated」这个词是 SI 定义账的标准动作：每一次更迭都逐字废止旧定义，没有一次是「悄悄并存」。

1990 年：电学单位的约定值——「定义账」的一个例外形态。 CIPM 1988 年决议 1/2 逐字：「that 483 597.9 GHz/V exactly be adopted as a conventional value, denoted by KJ-90 for the Josephson constant」与「that 25 812.807 Ω exactly be adopted as a conventional value, denoted by RK-90, for the von Klitzing constant」[一手逐字]。NIST 官方特刊（Taylor 1989）逐字：「KJ-90 = 483 597.9 GHz/V exactly … to come into effect starting January 1, 1990」[一手逐字]。注意：这是「约定值」（conventional value）——用约瑟夫森效应和量子霍尔效应的量子化台阶定义电压与电阻的实现方式，但当时没有把它们钉进基本单位定义；它本质上是「定义账的过渡形态」，2019 年重新定义后 e 成为定义常数，KJ-90/RK-90 的「约定」地位被吸收（SI 手册附录 1 有记录）[一手逐字]。

2011 → 2014 → 2017 → 2018：重新定义的四年四步。 第 24 届 CGPM（2011）决议 1 逐字：「takes note of the intention of the International Committee for Weights and Measures to propose a revision of the SI as follows: … the Planck constant h is exactly 6.626 06X × 10−34 joule second」——X 是待定数字，决议自己把未定值写进决议文本（这是「先立定义账、后填数」的制度设计）[一手逐字]。第 25 届 CGPM（2014）决议 1 逐字：「that despite this progress the data do not yet appear to be sufficiently robust for the CGPM to adopt the revised SI at its 25th meeting」并要求「complete all work necessary for the CGPM at its 26th meeting to adopt a resolution」[一手逐字]——2014 年说「数据还不够稳健、这次不采」：定义账的推迟本身被写进决议，这是「做对」的一部分（不急着交卷，把没准备好写进文件）。2017 年 CIPM 决定 10 逐字：「The CIPM noted that the agreed conditions for the redefinition are now met and decided to submit draft Resolution A to the 26th meeting」[一手逐字]。

2018-11-16：第 26 届 CGPM 决议 1——本次表决实况（本库亲核会议记录）。 决议 1 核心句逐字：「decides that, effective from 20 May 2019, the International System of Units, the SI, is the system of units in which: the unperturbed ground state hyperfine transition frequency of the caesium 133 atom ΔνCs is 9 192 631 770 Hz, the speed of light in vacuum c is 299 792 458 m/s, the Planck constant h is 6.626 070 15 × 10−34 J s, the elementary charge e is 1.602 176 634 × 10−19 C, the Boltzmann constant k is 1.380 649 × 10−23 J/K, the Avogadro constant NA is 6.022 140 76 × 1023 mol−1, the luminous efficacy of monochromatic radiation of frequency 540 × 1012 Hz, Kcd, is 683 lm/W」[一手逐字]。旧千克定义的废止措辞逐字：「the definition of the kilogram in force since 1889 (1st meeting of the CGPM, 1889, 3rd meeting of the CGPM, 1901) based upon the mass of the international prototype of the kilogram is abrogated」[一手逐字]。投票实况（第 26 届 CGPM 会议记录，英文版 pp. 465–466）逐字：「the voting on Draft Resolution A … would be carried out by a roll-call of the Member States … each of the 54 Member States that were present will be called on to vote in order … Any Member State that wishes to vote no or abstain should simply remain silent. Following the vote, Draft Resolution A, was adopted unanimously as Resolution 1」[一手逐字]。

任务书纠错一条（本库纪律）：「需要 2/3 多数」无一手依据。 本篇任务书曾预写「投票需 2/3 多数」，亲核后纠正：BIPM Compendium 表决规则逐字为「The general majority rule at the meetings of the CGPM is that of absolute majority of represented voting States」[一手逐字]；26 届会议记录显示本次为出席 54 国唱名、一致通过，无弃权记录[一手逐字]。同理，「the International Prototype of the Kilogram shall be retired」不是决议原文——决议用「abrogated」，而「will now be retired」出自 BIPM 2018-11-16 新闻稿[一手逐字/镜像]；本篇按决议文本承重，新闻稿措辞只作登记。

2019-05-20 生效与「世界计量日」的选择。 SI 手册序言逐字：「The changes were agreed by the CGPM in November 2018 with effect from May 20th 2019, a date chosen because it is World Metrology Day, the day when the Metre Convention was signed in 1875.」[一手逐字]——生效日选在 1875 年签约纪念日，这是制度把自己写进历史的动作，也说明「定义账」与「组织账」（米制公约）在计量学里是同一件事的两面。

重新定义的动机：SI 手册 1.2 节逐字（本篇最重要的动机原件）[一手逐字]：

「In the case of an artefact, the definition and the realization are equivalent – a path that was pursued by advanced ancient civilizations. Although this is simple and clear, artefacts involve the risk of loss, damage or change. The other types of unit definitions are increasingly abstract or idealized. Here, the realizations are separated conceptually from the definitions so that the units can, as a matter of principle, be realized independently at any place and at any time. In addition, new and superior realizations may be introduced as science and technologies develop, without the need to redefine the unit. These advantages – most obviously seen with the history of the definition of the metre from artefacts through an atomic reference transition to the fixed numerical value of the speed of light – led to the decision to define all units by using defining constants.」

序言那句总纲（SI 手册逐字）[一手逐字]：「For the first time, a complete set of definitions is available that does not make reference to any artefact standards, material properties or measurement descriptions.」

连续性账：量值不变是明示义务。 定义更迭最容易被问的问题是「2019 年 5 月 20 日零点，一公斤变了吗？」SI 手册序言逐字：「great attention has been paid to ensure that these definitions are consistent with those in place at the time the change was implemented」[一手逐字]；「This consistency of reference to the SI reflects the efforts of the CGPM and the CIPM to ensure the continuity of the values of measurements expressed in SI units through each change that has been made」[一手逐字]；千克衔接的数值账逐字：「The number chosen for the numerical value of the Planck constant in this definition is such that at the time of its adoption, the kilogram was equal to the mass of the international prototype, m(K) = 1 kg, with a relative standard uncertainty of 1 × 10−8, which was the standard uncertainty of the combined best estimates of the value of the Planck constant at that time」[一手逐字]；决议附录 2 同口径逐字：「the mass of the international prototype of the kilogram m(K) is equal to 1 kg within a relative standard uncertainty equal to that of the recommended value of h at the time this Resolution was adopted, namely 1.0 × 10–8」[一手逐字]。连续性不是口号，是选了「h 的推荐值」这种衔接动作做出来的：旧世界说「一公斤=原器的质量、h 是测出来的」，新世界说「h 钉死、原器质量是测出来的」，交接点是 h 的 CODATA 推荐值与 1×10⁻⁸ 的原有不确定度。

本层小结（本库断言）：六次定义更迭（1889 原器→1960 米离开实物→1967 秒原子化→1983 米光速化→1990 电学约定值→2019 全常数化）呈现一条单调方向——每一代都把「定义」向「任何人可复现」推一步，且每一步都有决议文本、废止文本与连续性声明。「做对」在定义层的含义：不是永远不改，是每次改都改得可以审计。

四、IPK 账：Le Grand K 的五十微克——「定义挂实物」的代价被公开测量

这是本篇的第二笔独占账，也是全书最有戏剧性的一层：一个「做对了的领域」最著名的失败——千克原器不可解释的漂移——恰恰是被做对的方法测量出来的。这一层的全部数字都以官方口径为准，且要指出官方口径之间的差异本身就是证据的一部分。

IPK 的制造与副本（官方口径）。 BIPM 官方 IPK 页逐字：「The IPK has been conserved at the BIPM since 1889, when it was sanctioned by the 1st General Conference on Weights and Measures (CGPM). It is of cylindrical form, with diameter and height of about 39 mm, and is made of an alloy of 90 % platinum and 10 % iridium.」[一手逐字]——注意官方口径是「约 39 mm」，39.17 mm 的精确值只见于二手转述，本篇按官方口径。副本数量与分配：BIPM 历史页逐字：「Forty similar prototypes were made at about the same time and these were all machined and polished to have closely the same mass as the international prototype. At the 1st CGPM (1889), after calibration against the international prototype, most of these ‘national prototypes’ were individually assigned to Member States」[一手逐字]；NIST 期刊文献（Jabbour & Yaniv 2001）逐字：「In 1884, 40 replicas of the kilogram were delivered from Johnson Matthey; they were compared to the mass of the IPK in 1888. In 1889, 34 of these replicas were distributed to the signatories of the Meter Convention who requested them.」[一手逐字]。制造者：Johnson Matthey 公司史逐字：「Then agreements made in 1882 with the French government led to Johnson, Matthey & Co producing 30 standard metres and 40 standard kilograms, all in the 10 per cent iridium-platinum alloy.」[一手逐字]；UK Metric Association 转述铸件由 George Matthey 在伦敦完成、在法国由 M Collot 精加工[多源交叉]。任务书纠错一条：「皇家造币厂 / George Johnson」不成立——一手来源只有 Johnson Matthey 一家。

存放与保管（官方口径的边界）。 BIPM 历史页逐字：「This was, and still is, kept at the BIPM under the conditions specified by the 1st CGPM when it sanctioned the prototype」[一手逐字]；NIST 官方逐字：「Kept in a triple-locked vault on the outskirts of Paris, the platinum-iridium cylinder was officially called the International Prototype of the Kilogram (IPK)」[一手逐字]；NIST 2011 官方新闻稿逐字：「The International Prototype Kilogram is stored in an arrangement of three nested bell jars.」[一手逐字]——「地窖保险柜、三把钥匙、每四十年取出一次」的完整叙述只见于二手（维基聚合），官方没有这一整句，本篇按官方口径分段标注。

三次周期检定（日期口径的分歧本身就是证据）。 BIPM 官方历史页对第三次的表述逐字：「the third verification carried out from 1989 to 1991, the median difference being about 25 micrograms for the set of original prototypes sanctioned by the 1st CGPM (1889)」[一手逐字]；第 19 届 CGPM（1991）会议记录（法文）逐字：「Je rappelle pour mémoire que les deux précédentes vérifications périodiques ont eu lieu de 1899 à 1911 et de 1946 à 1953. La troisième vérification périodique des prototypes nationaux du kilogramme s’effectue en deux étapes」[一手逐字]；NIST 官方口径：「These ‘periodic verifications’ occurred in 1899–1911, 1939–1953, and 1988–1992」[一手逐字]。注意三分歧：第一次 1899–1911（BIPM 官方）vs 1889（二手）；第二次 1946–1953（BIPM）vs 1939–1953（NIST）；第三次 1988–1992（NIST）vs 1989–1991（BIPM）。本篇按 BIPM 官方口径（1899–1911／1946–1953／1988–1992）承重，其余口径并列注明——口径分歧不是疏漏，是「同一个事件在不同档案里的坐标」，记录本身就是史料。

漂移的数字：25 微克中位数 vs 50 微克/百年。 BIPM 官方（SI 手册附录 4）逐字：「By the time of the second verification of national prototypes in 1946 it was found that on average the masses of these prototypes were diverging from that of the international prototype. This was confirmed by the third verification carried out from 1989 to 1991, the median difference being about 25 micrograms」[一手逐字]；NIST 官方（Kilogram: The Present）逐字：「The trend during the past century had been for most of BIPM’s official copies to gain mass relative to the IPK, although by somewhat different amounts, averaging around 50 micrograms (millionths of a gram) over 100 years. But an alternative explanation is that the IPK was losing mass relative to its copies. Even more likely, it was a combination of both.」[一手逐字]；NIST 期刊文献（Jabbour & Yaniv 2001）第三轴：「The results of the third periodic verification demonstrated a long-term instability of the unit of mass on the order of approximately 30 µg/kg over the last century; this instability is attributed to surface effects that are not yet fully understood」[一手逐字]。三个数字（25 中位/50 百年/30 每公斤）是同一现象的三个统计口径，本篇并列注明，不归一化。

「没有解释」的官方逐字——本篇最锋利的一格。 BIPM 2015 年 Metrologia 论文（BIPM 作者）逐字：「There is no explanation available for the differences in mass change after cleaning and washing between the IPK and its official copies, or between nos. 25 and 73.」[一手逐字]；BIPM 2017 年 Metrologia 论文逐字：「Observations of unexplained instabilities of mass standards were presented and discussed at a workshop … at the BIPM in November 2012.」[一手逐字]；BIPM 官员 Davis 接受 BBC 采访（2007）逐字：「The mystery is that they were all made of the same material, and many were made at the same time and kept under the same conditions, and yet the masses among them are slowly drifting apart. We don’t really have a good hypothesis for it.」[多源交叉]。任务书纠错一条：「cannot be attributed」无官方逐字，官方对应表述是上述两句——本篇按官方逐字承重。

美国副本 K20/K4：同一个制度下的具体账。 NIST 官方逐字：「Copies of the IPK were sent to the countries which had signed the agreement. In 1890, the United States received the fourth and 20th copies — K4 and K20. They served as official national kilogram standards and check standards」[一手逐字]；「The K20 mass served as the primary prototype national standard. The other, called K4, served as a ‘check’ standard that monitors the constancy of K20’s mass.」[一手逐字]；检定数值（Jabbour & Yaniv 2001 表格逐行）：K20 相对 IPK 的偏差 1889 = −39 µg、1937 = −21、1948 = −19、1984 = −22、1992 = −21、1999 = −39；K4 自 1889 至 1999 相对 IPK 变化 41 µg[一手逐字]；其国家基准名称逐字（NBS Circular 570）：「The primary standard of mass for this country is United States Prototype Kilogram 20」[一手逐字]。

2019 年 IPK 的退役与「继续存在」：一个制度性动作的两面。 决议 1 废止定义逐字（见第三章）[一手逐字]；BIPM SI 修订 FAQ（官方 PDF）逐字：「There are no plans to change the storage conditions for the International Prototype of the Kilogram (IPK). It will remain at the BIPM and it will not be on display for the general public. The IPK will retain a bit of metrological interest and therefore it will be monitored very sporadically in the future to avoid as much as possible any surface damage. Measurements of the mass stability of the IPK in the future may help us extrapolate its mass stability in the recent past.」[一手逐字]；BIPM《Mise en pratique》4.1 节逐字：「the IPK no longer retains a special status or a dedicated role in the mise en pratique … For that reason, the IPK and its six official copies are conserved at the BIPM under the same conditions as they were prior to the redefinition.」[一手逐字]。

本层小结（本库断言）——为什么「漂移无解释」是正面范本的证据而不是反证：① 漂移被测量出来，是因为制度要求周期检定——1889 年批准原器的同一届 CGPM 就设定了对照副本的校准制度，没有副本链就没有「漂移」这个概念；② 漂移被公开，是因为 BIPM 把「no explanation available」写进论文——本库一百余篇抓的「造尺人不报自己的失败」，在这里是反过来的；③ 漂移被当问题而非被掩盖，是因为它直接构成「定义挂实物」的代价论证——SI 手册 1.2 节那句「artefacts involve the risk of loss, damage or change」不是泛泛而谈，后面站着一个 50 微克的案例；④ 50 微克的量级（约 5×10⁻⁸ 相对）——人类最大的贸易、制药、科研体系对「一公斤」的要求是一致性远好于十亿分之一，而实物标准自己在一百年里漂掉了 50 微克——这就是「定义必须离开实物」的经验理由。

五、溯源账：校准金字塔、关键比对与 MRA——「自己说自己准不算数」

本库一百余篇反复抓一个毛病：同一批人既做试验又定标准、既打分又退役排行榜。计量学对这个问题给出的制度答案叫「溯源链」与「关键比对」——这是本篇的第三笔独占账。

溯源链的定义（VIM3 2.41 逐字，本篇最核心的一条制度定义）：「metrological traceability — property of a measurement result whereby the result can be related to a reference through a documented unbroken chain of calibrations, each contributing to the measurement uncertainty」[一手逐字]；配套条目 2.42「metrological traceability chain」逐字：「sequence of measurement standards and calibrations that is used to relate a measurement result to a reference」＋「NOTE 1 A metrological traceability chain is defined through a calibration hierarchy.」[一手逐字]。注意这一定义的两个硬点：①「documented unbroken」（有记录、不间断）——缺环即断链；②「each contributing to the measurement uncertainty」（每一环都贡献不确定度）——链的每一环都往结果里注入不确定度，所以「溯源到 SI」永远伴随着一个不确定度预算，而不是一个「绝对准」的承诺。

溯源链的顶端（SI 手册 1.3 节逐字）：「The definitions of the SI units, as decided by the CGPM, represent the highest reference level for measurement traceability to the SI. Metrology institutes around the world establish the practical realizations of the definitions in order to allow for traceability of measurements to the SI. The Consultative Committees provide the framework for establishing the equivalence of the realizations in order to harmonize traceability world-wide.」[一手逐字]——「最高参照层级」是 CGPM 决议，不是任何一家实验室；「Consultative Committees 提供等价性框架」就是关键比对的组织化。

校准金字塔（通俗图式的官方对应）：NIST 官方（Calibration of Mass Standards 页）逐字：「NIST mass disseminations are traceable to the International System of Units (SI) through the Consensus Value (CV) of the kilogram … The CV is then transferred to the NIST Pool of Mass Standards … The pool of artifacts is also used to create working standards which form the beginning of the NIST mass scale.」[一手逐字]。任务书提示的「primary realization → national → secondary → working」四层金字塔图式，官方文本里对应的正式词是 VIM 的 calibration hierarchy 与 SI 手册 1.3 节——「金字塔」是通俗表述，本篇按官方术语承重。

CIPM MRA（1999）：把「我的校准证书」变成「我们互认的校准证书」。 BIPM 官方背景文件逐字：「The CIPM Mutual Recognition Arrangement … was signed in Paris on 14 October 1999 by the Directors of 38 National Metrology Institutes (NMIs) and two international organizations.」[一手逐字]；MRA 文本（CIPM-MRA-2003.pdf）开篇逐字：「This is an arrangement between national metrology institutes which specifies terms for the mutual recognition of national measurement standards and for recognition of the validity of calibration and measurement certificates issued by national metrology institutes.」[一手逐字]；当前规模（BIPM 2021 官方公告）逐字：「The Arrangement was originally signed in 1999 by 38 Member States and two international organizations and has since matured into a well-recognized pillar of the international quality infrastructure which now covers 258 institutes: signatories from 62 Member States, 40 Associates of the CGPM and four international organizations as well as 152 further institutes designated by the signatory bodies.」[一手逐字]。MRA 的制度含义：一家计量院出的校准证书，被另一国的监管与贸易体系接受，靠的不是名气，是它参与关键比对并在 KCDB 登记的记录。

KCDB（关键比对数据库）——「送出去比」的实物证据。 2026-08-02 实查（BIPM 官方 KCDB 首页）：「Currently in the KCDB there are 259 CIPM MRA participants / 1278 key comparisons / 26270 CMCs / 775 supplementary comparisons」[一手逐字，实时快照]。关键比对的本质：几家独立的国家计量院各自用自己的复现装置测同一个东西，比对结果公开登记——这是「自己说自己准不算数」的建制化。

ISO/IEC 17025 与 ILAC P14：不确定度义务延伸到校准实验室。 ISO 官方标准页逐字：「ISO/IEC 17025:2017 — General requirements for the competence of testing and calibration laboratories」[一手逐字]；其 7.6 节要求（ANAB 官方解释文件转述逐字）：「ISO/IEC 17025:2017 requires laboratories to identify all contributions to measurement uncertainty that are of significance, including those arising from sampling (7.6.1)」[多源交叉，ISO 正文付费墙未取回]；ILAC P14:09/2020 政策正文逐字：「5.1 The Accreditation Body shall ensure that an accredited calibration laboratory reports the measurement uncertainty in compliance with the GUM.」「5.2 … In calibration certificates the measurement result should be reported as y ± U … The coverage factor and the coverage probability shall be stated on the calibration certificate.」[一手逐字]。「校准证书上必须写 y ± U 与覆盖因子」——不确定度义务的最终落点是证书文本。

千克的两种复现路线：Kibble 天平与硅球原子计数（XRCD）——「任何能连上 h 的物理都可以当实现」的活例。 SI 手册第 9 版 kg 实现一节逐字：「For the kilogram, the unit whose definition has undergone the most fundamental change, realization can be through any equation of physics that links mass, the Planck constant, the velocity of light and the caesium frequency. One such equation is that which describes the operation of an electro-mechanical balance, previously known as a watt balance, more recently known as a Kibble balance … Another method that can be used for a primary realization of the kilogram is through the determination of the number of atoms in a silicon sphere」[一手逐字]——注意「realization can be through any equation of physics」：定义与实现分离的实质好处在这里——两条物理上完全不同的路线都合法，且互相校核。

Kibble 天平（原名瓦特天平）：名字与出处。 NIST 官方逐字：「Named after its inventor, Bryan Kibble at the U.K.’s National Physical Laboratory (NPL), who conceptualized it in 1975」[一手逐字]；NPL 官方历史页逐字：「He realised a limitation of the current balance, the instrument used to define the ampere, in the early 1970s and began work on a method for improving it. Kibble’s new measuring device, the moving-coil watt balance, first made in 1978 … After his death in 2016, the international metrology community made the decision to rename the moving-coil watt balance the Kibble balance」[一手逐字]。原理一句（NIST 官方）：「When this equation is combined with the one above for force on the coil, the problematic B and L cancel out, leaving IV = mgν. (That is, electrical power equals mechanical power, both expressed in watts.)」[一手逐字]；NPL 官方补充：「It allows the SI unit of mass to be realised in terms of quantum electrical units; Josephson voltage and quantum Hall resistance, in relation to a fixed value: the Planck constant.」[一手逐字]。

各国实现的不确定度账（全部官方数字）：NIST-4 提交给 2019 年重新定义的 h 测量：「NIST’s final h measurement submitted for the redefined SI had an uncertainty of 13 parts in a billion. Another Kibble balance measurement, from the National Research Council (NRC) of Canada, had an uncertainty of just 9.1 parts per billion.」[一手逐字]；NIST-4 论文（Haddad et al. 2017）逐字：「The relative standard uncertainty of this result is 13 × 10−9」「This yields a total relative standard uncertainty of 25 × 10−9」（24 小时测量）、「Integrating four days of data will reduce the total relative uncertainty to below 20 × 10−9」[一手逐字]；中国计量院 NIM-2（焦耳天平，Kibble 思路的变体）：英文官方页逐字「In May 2017, the NIM-2 result was submitted to Metrologia and accepted by the CODATA Task Group on Fundamental Constants (TGFC) as input data」「By 2018, the Type A uncertainty of NIM-2 has been decreased to 5×10-8 (k=1)」[一手逐字]、中文官方页逐字「截止到2019年10月，NIM-2能量天平装置的A类不确定度已经达到2×10-8（k=1）」[一手逐字]；德国 PTB 的 Planck-Balance 官方页逐字：「Setting up of a ‘Planck-Balance’ according to the Kibble principle. After the planned redefinition of the International System of Units (SI), the balance will make it possible to realize the kilogram via the Planck constant h.」[一手逐字]。

XRCD（阿伏伽德罗路线）：28 硅球与它的造价。 NIST 官方逐字：「Beginning in the early 1990s, members of a worldwide collaboration called the International Avogadro Coordination (IAC) — now known as the International Avogadro Project — have been working … a uniform crystal of silicon-28, carefully machined into spheres with a mass of 1 kg」「These super-spheres … each cost about $3.2 million and had to be handcrafted by a master lens maker.」[一手逐字]；PTB 官方（协调方）逐字：「The crucial phase of the long-term Avogadro project – which is coordinated by PTB – started in 2003 … the first measurements have been completed on the two 1 kg spheres of 28Si – which had been polished in Australia」[一手逐字]；澳大利亚 Powerhouse 博物馆官方藏品档案逐字：「Australia and the CSIRO have played an integral part in this, being currently the only laboratory producing the silicon spheres that are used in the project.」[多源交叉]。

两条路线在 2017 年的收敛——「不同方法独立收敛」的官方记录。 CODATA 2017 特别平差论文（Newell et al. 2018）逐字：「five have relative standard uncertainties ur at or below 50 × 10−9, with two at or below 20 × 10−9, where the latter includes results from both the Kibble balance and the x-ray crystal density (XRCD) methods.」[一手逐字]；NIST 团队论文（Possolo et al. 2018）逐字：「In fact there are now six experiments with relative standard uncertainties smaller than 50 × 10−9 … Four of these have relative standard uncertainties smaller than 20 × 10−9: IAC-15, NRC-17, IAC-17, and NIST-17」「The level of agreement between the most accurate determinations of h available today is simply remarkable, especially considering that they involve multiple, different, and independent versions of two radically different approaches to its measurement.」[一手逐字]。两个物理机制完全不同的方法（电学功率平衡 vs 数原子）各自做到十亿分之二十以内，且互相吻合——这正是「不依赖唯一实物」在证据层的兑现。

本层小结（本库断言）：溯源账是「自己说自己准不算数」的完整制度答案——① 溯源链的定义要求「有记录、不间断、每环贡献不确定度」；② 关键比对让独立的复现互相校核并把结果公开登记（KCDB）；③ MRA 把互认变成证书层面的法律事实；④ ISO 17025/ILAC P14 把「报不确定度」变成证书义务；⑤ 定义与实现分离让「任何能连上 h 的物理」都可以当实现，于是有了两条独立路线收敛到 20 ppb 以内的证据。对照本库：本库的「去重实测」是关键词比对，「主笔亲核」是把链上每一环的不确定度都检查一遍，「编号注册」是 MRA——本库这些动作全是从这层抄的。

六、不确定度账：GUM 的诞生、内容与自我局限——「测不准多少」的制度史

这是本篇的第四笔独占账，也是与AI 评测榜篇落点（「测量学的答案从来不是测准，是知道自己测不准多少」）直接对接的一层：「知道自己测不准多少」在计量学里不是谦辞，是 GUM 里的字面义务。

GUM 的来历（1993 七组织联合发布 → 1995 重印 → 2008 JCGM 100 重新发布）。 GUM 2008 版前言逐字：「The following seven organizations* supported the development of this Guide, which is published in their name:」其后列出 BIPM／IEC／IFCC／ISO／IUPAC／IUPAP／OIML[一手逐字]；2008 版脚注逐字：「In 2005, the International Laboratory Accreditation Cooperation (ILAC) officially joined the seven founding international organizations.」[一手逐字]；历史起点（GUM 前言逐字）：「In 1977, recognizing the lack of international consensus on the expression of uncertainty in measurement, … the CIPM requested the BIPM to address the problem …」→ BIPM 问卷 32 个国立实验室 → 1980 年 INC-1 Recommendation →「The CIPM approved the Recommendation in 1981 and reaffirmed it in 1986.」[一手逐字]；1993 年份的直接逐字（GUM S1 前言）：「a Joint Committee for Guides in Metrology (JCGM), chaired by the Director of the Bureau International des Poids et Mesures (BIPM), was created by the seven international organizations that had originally in 1993 prepared the ‘Guide to the expression of uncertainty in measurement’ (GUM) and the ‘International vocabulary of basic and general terms in metrology’ (VIM). The JCGM assumed responsibility for these two documents from the ISO Technical Advisory Group 4 (TAG4).」[一手逐字]。

不确定度的定义（GUM 2.2.3 逐字）：「uncertainty (of measurement) — parameter, associated with the result of a measurement, that characterizes the dispersion of the values that could reasonably be attributed to the measurand」[一手逐字]——注意「could reasonably be attributed」（可以合理地归属于）——不确定度是对「被测量可能取哪些值」的刻画，不是对「误差有多大」的声明。

不确定度 vs 误差：本库最常引的区分在这里有逐字。 GUM 3.2.2 NOTE 2 逐字：「In this Guide, great care is taken to distinguish between the terms ‘error’ and ‘uncertainty’. They are not synonyms, but represent completely different concepts; they should not be confused with one another or misused.」[一手逐字]；GUM E.5.4 逐字：「this Guide’s concept of uncertainty eliminates the confusion between error and uncertainty … this Guide’s operational approach, wherein the focus is on the observed (or estimated) value of a quantity and the observed (or estimated) variability of that value, makes any mention of error entirely unnecessary.」[一手逐字]；3.3.1 NOTE 逐字：「the uncertainty of the result of a measurement should not be confused with the remaining unknown error.」[一手逐字]；VIM3 前言逐字：「The change in the treatment of measurement uncertainty from an Error Approach (sometimes called Traditional Approach or True Value Approach) to an Uncertainty Approach necessitated reconsideration of some of the related concepts」[一手逐字]。

「测量结果必须带不确定度才算完整」——本篇的落点句在 GUM 里的原文。 GUM 3.1.2 逐字：「In general, the result of a measurement (B.2.11) is only an approximation or estimate (C.2.26) of the value of the measurand and thus is complete only when accompanied by a statement of the uncertainty (B.2.18) of that estimate.」[一手逐字]。

类型 A 与类型 B（GUM 逐字）：「Type A evaluation (of uncertainty) — method of evaluation of uncertainty by the statistical analysis of series of observations」「Type B evaluation (of uncertainty) — method of evaluation of uncertainty by means other than the statistical analysis of series of observations」[一手逐字]；且 3.3.4 明确：「The purpose of the Type A and Type B classification is to indicate the two different ways of evaluating uncertainty components and is for convenience of discussion only; the classification is not meant to indicate that there is any difference in the nature of the components resulting from the two types of evaluation.」[一手逐字]——分类只是便利，不是本体区别：这正好是「口径账」在 GUM 里的对应物（名字怎么分，不影响成分本身的性质）。

不确定度预算：多分量（GUM 3.3.1 逐字）：「The uncertainty of the result of a measurement reflects the lack of exact knowledge of the value of the measurand … The result of a measurement after correction for recognized systematic effects is still only an estimate of the value of the measurand because of the uncertainty arising from random effects and from imperfect correction of the result for systematic effects.」[一手逐字]；3.3.2 列出十个不确定度来源（测量对象定义不完整、采样不具代表性、环境影响、仪器分辨率、标准物质/常数不准、方法近似与假设、重复观测涨落等）[一手逐字]。

覆盖因子 k=2（GUM 6.3.3 逐字，注意「近似+有条件」）：「one can assume that taking k = 2 produces an interval having a level of confidence of approximately 95 percent, and that taking k = 3 produces an interval having a level of confidence of approximately 99 percent.」[一手逐字]；G.6.6 惯例总结逐字：「adopt k = 2 and assume that U = 2uc(y) defines an interval having a level of confidence of approximately 95 percent; or, for more critical applications, adopt k = 3 … approximately 99 percent.」[一手逐字]；而 6.2.3 的自我谦辞逐字：「in most cases the level of confidence p (especially for values of p near 1) is rather uncertain, not only because of limited knowledge of the probability distribution … but also because of the uncertainty of uc(y) itself」[一手逐字]——GUM 自己说 k=2 的 95% 是「近似且有条件」，本库引它时必须保留这个限定。

GUM Supplement 1（JCGM 101:2008，蒙特卡洛）——GUM 框架的官方补丁。 引言逐字：「The described Monte Carlo method is a practical alternative to the GUM uncertainty framework [GUM:1995 3.4.8]. It has value when a) linearization of the model provides an inadequate representation, or b) the probability density function (PDF) for the output quantity departs appreciably from a Gaussian distribution or a scaled and shifted t-distribution, e.g. due to marked asymmetry. In case a), the estimate of the output quantity and the associated standard uncertainty provided by the GUM uncertainty framework might be unreliable. In case b), unrealistic coverage intervals … might be the outcome.」[一手逐字]；并逐字声明：「Whereas there are some limitations to the GUM uncertainty framework, the propagation of distributions will always provide a PDF for the output quantity that is consistent with the model of the measurement and the PDFs for the input quantities.」[一手逐字]；8.1.1 验证条款逐字：「The GUM uncertainty framework can be expected to work well in many circumstances. However, it is not always straightforward to determine whether all the conditions for its application (see 5.7 and 5.8) hold. … Since the domain of validity for MCM is broader than that for the GUM uncertainty framework, it is recommended that both the GUM uncertainty framework and MCM be applied and the results compared.」[一手逐字]。

GUM 的自我局限清单（官方自己的认知，诚实空位补位）：① 非线性：S1 引言 NOTE 2 逐字「The GUM provides an approach when linearization is inadequate [GUM:1995 5.1.2 note]. The approach has limitations: only the leading non-linear terms in the Taylor series expansion of the model are used, and the PDFs for the input quantities are regarded as Gaussian.」[一手逐字]；② 相关输入＋非线性缺口：S1 9.4.3.1.1 逐字「the GUM uncertainty framework with higher-order terms is not applied, no counterpart being provided in the GUM for the formula containing higher-order terms when the xi have associated non-zero covariances (see 5.8).」[一手逐字]；③ 置信水平本身不确定：6.2.3（见上）[一手逐字]；④ 官方仍在修订：BIPM JCGM-WG1 官方页证实「JCGM 100:2008 Amendment 1: Non-linearity in measurement models」在案，JCGM 2024 年官方文件把 GUM 升级为文档套件、S1 将修订为 GUM-7[一手逐字]。任务书纠错两条：「error approach」与「integral part」两句在 GUM 正文无逐字——本篇按上述等价逐字承重，不伪称 GUM 原文。

本层小结（本库断言）：不确定度账是「知道自己测不准多少」的制度史——① GUM 把「报不确定度」写成测量结果完整性的定义性条件；② 它把「误差」与「不确定度」的区分写成硬性概念纪律；③ 它用 Type A/B、合成、覆盖因子给出可操作的口径；④ 它对自身的局限（非线性、非正态、相关输入）有官方自认与官方补丁（S1 蒙特卡洛）与官方修订流程（JCGM WG1）。对照本库：本库的核验图例（一手逐字/摘要逐字/多源交叉/我们的断言）就是 Type A/B 的分类纪律加上「分类只是便利、不是本体区别」的自觉；本库的保守降级闸门就是「k=2 的 95% 只是近似」的元版本。

七、传播账：SI 到你手里——GPS、油枪、血糖报告与「全世界」

正面范本需要回答「这套制度到底传到了多远」——这一层把 SI 的传播拆成四条线下可查的链：时间、贸易、医学、法律。

时间链：GPS 是全世界最大的 SI 秒分发器，而相对论修正是官方的字面账。 NIST 官方逐字：「These satellites hurtle so fast through space that special relativity indicates they should fall behind earthbound clocks by 7 microseconds (millionths of a second) per day. Meanwhile, the lower gravity in medium Earth orbit should speed them up by 45 microseconds per day compared to clocks on Earth. Combining the two effects, atomic clocks aboard GPS satellites run 38 microseconds per day faster than earthbound clocks.」[一手逐字]——38 微秒/天 = 狭义相对论 −7 + 广义相对论 +45，两个效应方向相反，合成值官方逐字给出。GPS 卫星钟的构成与溯源（美国海军天文台官方）逐字：「Each Block IIR/IIR-M satellite contains three Rubidium (Rb) atomic clocks. Each Block IIF satellite contains two Rb and one Cesium (Cs) atomic clocks.」且「The system time, in turn, is referenced to the Master Clock (MC) at the USNO and steered to UTC(USNO)」[一手逐字]；BIPM 侧的对应（USNO 官方页面转述）逐字：「BIPM then applies frequency corrections (‘steers’) to EAL, based on measurements from primary frequency standards and intended to keep the International System’s basic unit of time, the second, constant.」[一手逐字]；NIST 官方出版物逐字：「the Global Positioning System (GPS) is also the world’s primary system for the distribution of accurate (sub-microsecond) time」「it is inherently traceable to the SI and UTC, because it is referenced to UTC(USNO)」[一手逐字]。边界（任务书纠错一条）：手机是 GPS 秒的用户，不是原子钟的拥有者。 NIST 官方逐字：「GPS satellites — the ones that control the blue dot on your phone … all carry atomic clocks」[一手逐字]；而 NIST 官方同时逐字：「An ultimate solution would involve every smartphone or smartwatch containing its own atomic clock. That would require atomic clocks much smaller, more energy efficient and cheaper than any on the market today. While scientists are developing such clocks, they likely remain years away.」[一手逐字]——「每台手机背后是原子钟」对，「每台手机里有原子钟」错，两句话都按官方逐字。

贸易链：油枪与秤是法定计量的日常现场。 SI 手册 1.3 节逐字：「Individual countries have established rules concerning the use of units by national legislation, either for general use or for specific areas such as commerce, health, public safety and education. In almost all countries, this legislation is based on the SI. The International Organization of Legal Metrology (OIML) is charged with the international harmonization of the technical specifications of this legislation.」[一手逐字]；OIML 官方定位逐字：「The mission of the OIML is to enable economies to put in place effective legal metrology infrastructures that are mutually compatible and internationally recognized, for all areas for which governments take responsibility, such as those which facilitate trade」[一手逐字]；其建议书（R 系列）逐字：「International Recommendations (OIML R), which are model regulations that establish the metrological characteristics required of certain measuring instruments and which specify methods and equipment for checking their conformity. OIML Member States shall implement these Recommendations to the greatest possible extent.」[一手逐字]；燃油加油机属 OIML R 117 管辖（官方文本逐字）：「This Recommendation specifies the metrological and technical requirements applicable to dynamic measuring systems for quantities (volume or mass) of liquids other than water subject to legal metrology controls.」＋「5 Requirements specific to certain types of measuring systems … Fuel dispensers …」[一手逐字]；美国侧贸易衡器依据（NIST 官方 FAQ）逐字：「Weighing and measuring instruments used for trade must comply with the requirements in NIST Handbook 44」[一手逐字]。

医学链：临床检验的 SI 化是一段有日期的历史。 IFCC 与 IUPAC 联合推荐（1973，官方出版物）逐字：「The first seven are the so-called International System of Units (SI units) (cf. 8.12).」[一手逐字]；2006 年 IFCC 血糖报告建议逐字：「The document recommends reporting the concentration of glucose in plasma (in the unit mmol/L), irrespective of sample type or measurement technique.」[一手逐字]；体外诊断校准溯源标准 ISO 17511:2020 官方授权样本 PDF 逐字：「This document specifies technical requirements and documentation necessary to establish metrological traceability of values assigned to calibrators, trueness control materials and human samples for quantities measured by IVD MDs. … Metrological traceability of values for quantities in human samples extends to the highest available reference system component, ideally to RMPs and certified reference materials (CRMs).」[一手逐字]——注意 ISO 17511 的「ideally」：医学量并非全部能溯到 SI（标准本身承认存在 Non-SI traceable 的 IVD 器械），这是医学链的边界，本篇如实标注。

法律链：全世界谁用 SI、谁不用。 CIA World Factbook 官方（2022 档案版）逐字：「At this time, only three countries – Burma, Liberia, and the US – have not adopted the International System of Units (SI, or metric system) as their official system of weights and measures. … The US is the only industrialized nation that does not mainly use the metric system in its commercial and standards activities, but there is increasing acceptance in science, medicine, government, and many sectors of industry.」[一手逐字]；NIST 官方博客口径逐字：「While it’s true that metric use is mandatory in some countries and voluntary in others, all countries have recognized and adopted the SI, including the United States.」[一手逐字]；美国法律现状（NIST 官方）逐字：「Currently, U.S. law permits the use of either the SI (International System of Units) or the U.S. customary measurement system.」[一手逐字]。任务书纠错一条：「96% 以上国家采用公制」无官方一手统计——CIA 的「三国未采用」与 NIST 的「所有国家都承认并采用」才是官方口径；「95% 世界人口」出自计量史学者 Vera 2011 的论文（维基引为 p. 49），属学术估算，本篇不引用为官方数字。

非 SI 单位的现实（SI 手册 Table 8 逐字，正面范本自己写的边界）：「It is recognized that some non-SI units are widely used and that this is expected to continue for many years. It is therefore important to recall the values of these non-SI units in terms of SI units, because the SI is the internationally agreed reference with respect to which all other units are defined. A non-exhaustive list of non-SI units is given in Table 8」[一手逐字]——表内含分/时/日（1 d = 86 400 s）、度分秒、升（1 L = 1 dm³）、吨（1 t = 10³ kg）、电子伏（1 eV = 1.602 176 634 × 10⁻¹⁹ J）、天文单位（1 au = 149 597 870 700 m）、道尔顿（1 Da = 1.660 539 068 92(52) × 10⁻²⁷ kg）、海里/节、巴、分贝等[一手逐字]——「SI 承认非 SI 单位会继续存在」写在 SI 自己的手册里：做对不等于消灭其他单位制。

本层小结（本库断言）：传播账说明「做对」的成效是可数的——38 微秒/天的相对论账有官方逐字、油枪有 OIML R 117、血糖有 IFCC 的 mmol/L、法律上有 190+ 国家的 SI 立法基础——同时边界同样可数：手机没有原子钟、美国允许两套单位并行、ISO 17511 承认非 SI 溯源存在、SI 手册自己收编非 SI 单位表。传播与边界是同一枚硬币的两面，本篇两面都数。

八、边界账：美国公制化失败——「做对」的反面案例里的制度痕迹

正面范本必须有能力解剖失败，否则就成了宣传。美国的公制化是全世界最著名的度量衡失败案例，而它的失败恰恰留下了完整的制度档案——每一道该做的步骤都做了（立法、行政令、执行机构），每一步的结果都被记录。这一层全部引官方原件，不做动机推断。

1975 年 Metric Conversion Act（公法 94-168）：自愿。 法案原文 SEC. 3 逐字：「It is therefore declared that the policy of the United States shall be to coordinate and plan the increasing use of the metric system in the United States and to establish a United States Metric Board to coordinate the voluntary conversion to the metric system.」[一手逐字]；福特总统签署声明逐字：「It is important to stress that the conversion contemplated in this legislation is to be a completely voluntary one.」[一手逐字]——「completely voluntary」是总统签署声明里的原话：立法把公制化定为「完全自愿」，这就是失败的第一个制度原因，写在一手档案里。

1988 年修订：政策转向「preferred system」（任务书纠错一条：正确条款是 15 U.S.C. §205b(1)，不是 §205a）。 现行法典汇编（COMPS-1851）逐字，§205a(2)（国会发现段）：「Although the use of metric measurement standards in the United States has been authorized by law since 1866 (Act of July 28, 1866; 14 Stat. 339), this Nation today is the only industrially developed nation which has not established a national policy of committing itself and taking steps to facilitate conversion to the metric system.」[一手逐字]；§205b(1)（政策声明段）逐字：「to designate the metric system of measurement as the preferred system of weights and measures for United States trade and commerce; (2) to require that each Federal agency, by a date certain and to the extent economically feasible by the end of the fiscal year 1992, use the metric system of measurement in its procurements, grants, and other business-related activities」[一手逐字]。

1991 年 EO 12770：行政分支跟进。 行政令正文逐字：「and in order to implement the congressional designation of the metric system of measurement as the preferred system of weights and measures for United States trade and commerce, it is hereby ordered as follows: Section 1. … The Secretary of Commerce (‘Secretary’) is designated to direct and coordinate efforts by Federal departments and agencies to implement Government metric usage in accordance with section 3 of the Metric Conversion Act (15 U.S.C. 205b) … Sec. 2(a) use, to the extent economically feasible by September 30, 1992 … the metric system of measurement in Federal Government procurements, grants, and other business-related activities」[一手逐字]。

1995 年公法 104-59 §205(c)：公路标志的英制被立法保护。 法律原文逐字：「The Secretary shall not require the States to expend any Federal or State funds to construct, erect, or otherwise place or to modify any sign relating to a speed limit, distance, or other measurement on a highway for the purpose of having such sign establish such speed limit, distance, or other measurement using the metric system.」[一手逐字]；FHWA 官方《统一交通控制设施手册》(MUTCD) 第 11 版附录逐字：「Throughout this Manual all dimensions and distances are provided in English units.」[一手逐字]。

机构的两种命运：Metric Board 已停办，Metric Program 仍在。 1975 年设立的美国公制委员会（U.S. Metric Board）于 1982 年停办（现行法汇编本已删去其章节）[一手逐字]；NIST Metric (SI) Program 页面现状（2026-08-02 实查）逐字：「The Metric Program helps implement the national policy to establish the SI (International System of Units, commonly known as the metric system) as the preferred system of weights and measures for U.S. trade and commerce.」且「As of August 16, 2023 the physics.nist.gov historic SI Units site has permanently retired. This page and complete Metric (SI) Program contains current SI information.」[一手逐字]。

美制单位与 SI 的关系（官方口径，边界也是制度）：NIST 官方 FAQ 逐字：「Non-SI units are defined in terms of the SI. NIST recommends the use of the term ‘U.S. customary system of measurement’ to describe the collection of non-SI measurement units used in the United States」[一手逐字]——磅、加仑、华氏度今天的合法定义全部是 SI 值的换算：即使美国在贸易与日常生活中不用公制，它的非公制单位本身也是由 SI 定义的。这是「定义账」在失败案例里的残余：制度没有把单位换成 SI，但制度让非 SI 单位的定义归 SI 管。

本层小结（本库断言）：美国公制化失败不是一个「没有努力」的案例，是一个「每一步都做了、每一步都做成了半截」的案例——1975 自愿、1988 preferred、1991 行政令、1995 公路标志立法保护、1982 委员会停办、Metric Program 至今仍在。它的档案价值在于：当组织化、强制力与日常习惯三方不一致时，「做对」的制度链条在哪一环断的，每一环都有文件可查。本库不做任何「美国应该公制化」的政策主张——这只说明一个测量学事实：SI 的全球传播不是自动的，是组织账、定义账、强制账三本账共同运转的结果，三本账中任何一本缺席，传播就停在那里的档案里。

九、母题收口：正面范本给我们什么——本库一百余篇手法的「正厂」

本篇是正面范本，收口层不做新的裁决，只做一件事：把本库的方法逐件拿去与计量学的对应物对照，并指出每个对应物都是「先当义务、后当武器」——本库的武器是从这里借的，这里把武器造出来的时候先造的是纪律。

对照表（本库断言，逐行可回查）：

本库的武器 计量学的对应物 对应物的一手原文
分母账（还原分子分母） 溯源链（每环贡献不确定度） VIM3 2.41「each contributing to the measurement uncertainty」
口径账（换标准重算） 定义与复现分离（定义是决议、复现是任意物理） SI 手册 1.2「realized independently at any place and at any time」
尺子账（量具是否被污染） 关键比对（KCDB 公开登记） KCDB 官方页 1,278 项关键比对
不确定度标注（保守降级闸门） GUM（不确定度是完整性的定义条件） GUM 3.1.2「complete only when accompanied by a statement of the uncertainty」
核验图例（一手/摘要/多源/断言） Type A/B 分类（分类只是便利） GUM 3.3.4「is for convenience of discussion only」
去重实测（全库扫描） 溯源链「documented unbroken」 VIM3 2.41
编号与注册（机制裁决编号） CGPM 决议 + SI 手册版本管理 SI 手册第 9 版 V4.01 与决议 DOI
自指（尺审自己） GUM S1 验证条款 + JCGM 修订流程 S1 8.1.1「it is recommended that both the GUM uncertainty framework and MCM be applied and the results compared」
失败公开（记录被自己推翻的假设） Le Grand K 的 50 微克漂移无解释公开 BIPM 2015「There is no explanation available」
退役作废（旧排行榜退役声明） 决议「is abrogated」（逐字废止旧定义） 26 届 CGPM 决议 1「is abrogated」

三个具体对照（本篇的灵魂三句）：

对照一·退役账：SWE-bench 退役 vs IPK 退役——同一动作，两种制度。 AI 评测榜篇记录了 OpenAI 2026-02-23 官方宣布停止报告 SWE-bench Verified（「all frontier models we tested were able to reproduce the … gold patch」），理由是造榜人自己承认榜被刷穿；本篇记录了 IPK 的退役——26 届 CGPM 决议逐字「is abrogated」，退役当天新定义生效、连续性声明、h 的数值选在旧定义的不确定度带宽内（1×10⁻⁸）、FAQ 写明 IPK 保留在 BIPM 且「monitored very sporadically」。两个退役都是「造尺人自己作废旧尺」，但计量学的退役是：逐字决议、提前四年预告（2011→2018）、数值衔接账公开、退役后用途公开；SWE-bench 的退役是：2023-10 发布→2026-02 判死刑，中间没有一次「我们正在修尺子」的预告决议。本库不据此判 SWE-bench 的退役「不诚实」——判的是制度形态：计量学证明了「尺子作废」可以是一件有预告、有账本、有交接的公开事件。

对照二·不确定度账：MMLU 分数 vs 校准证书——同一件事，两种义务。 AI 评测榜篇的落点是「AI 从来不缺分数，缺的是对分数的效度论证」；GUM 3.1.2 逐字把同一件事写成字面义务：「the result of a measurement … is complete only when accompanied by a statement of the uncertainty」；ILAC P14 把义务落到证书格式「y ± U」＋覆盖因子。计量学的答案不是「每张证书都准」，是「每张证书都带着自己的不准」——而一个不带不确定度的 MMLU 分数，按 GUM 的定义根本不是一份完成的测量结果。

对照三·定义账：Le Grand K 的 50 微克 vs 排行榜的静默退役——「看不见的漂移」的两种命运。 IPK 的漂移被看见，是因为制度要求周期检定并把「no explanation available」写进论文[一手逐字]；排行榜的分母漂移（测试集混入训练语料）被看见，往往要等第三方探针（选项乱序掉 77%、GSM1k 掉 8%）[多源交叉]。同样的「尺子在漂」，一个被制度例行地看见，一个靠外人偶然地戳穿——差的不是测量技术，是「有没有人把『检查尺子』当成例行义务」。

落点（与 AI 评测篇的落点接上，本篇再往前推一步）：「测量学的答案从来不是『测准』，是『知道自己测不准多少』」——这句话在计量学里不是结论，是操作规程：定义可复现（不依赖唯一实物）、不确定度随行（y ± U 是义务）、尺子送出去比（关键比对+互认）。本库一百余篇做分母账、口径账、尺子账，做的都是计量学 151 年前就开始做的同一件事；本篇的价值不是赞美，是把借来的武器还回它的正厂看一眼——看它被造出来的时候，每一件都是先当义务、后当武器。

十、诚实空位与不确定点（本篇自己的不确定度账）

按本库纪律，正面范本同样要报自己的不确定度：

  1. 「约 25 万种度量衡单位」无一手普查：是「contemporaries estimated」的当代估计（Alder 2002 转述），非现代史学统计；Kula 1986 的史学批判（旧制度单位有其功能精度）与之并存，本篇按「多源交叉」降档引用，不做精确数字。
  2. 1799 年铂原器入藏日期两说：SI 手册与法国科学院（Débarbat 2019）用 1799-06-22（牧月 4 日），Johnson Matthey 技术评论 2000 用 1799-07-22；本篇按主流口径承重，并列标注。
  3. 三次周期检定的日期各官方口径不一：BIPM（1899–1911／1946–1953／1988–1992）、NIST（1899–1911／1939–1953／1988–1992）、二手（1889／1948／1989）三分歧；本篇按 BIPM 口径承重并并列注明。
  4. 漂移数字三口径：BIPM 25 微克（第三次检定中位数）、NIST 50 微克/百年（均值）、Jabbour & Yaniv 约 30 µg/kg；本篇并列，不归一化。
  5. ISO/IEC 17025:2017 正文未取回（付费墙），7.6 节内容经 ANAB 官方文件转述承重（多源交叉），不承重未核原文的细节。
  6. ILAC P14 正文经 IAAC 官方镜像取回（ilac.org 直链为动态列表），文件内容为 ILAC 官方政策文本，标注镜像。
  7. 「需要 2/3 多数」「retired 措辞」「39.17 mm 尺寸」「a safe in a basement」「皇家造币厂」五处任务书预写内容均被亲核推翻或修正——本库纪律生效，全部按官方逐字改写。
  8. KCDB 数字为 2026-08-02 实时快照，随季度更新变动。
  9. 「96% 以上国家采用公制」无官方一手统计，本篇改用 CIA「三国未采用」与 NIST「所有国家都承认并采用」两个官方口径。
  10. 「SI 是单位制、不是测量学科本身」无一句现成官方原话：由 BIPM「recommended practical system of units」定位与 GUM/VIM 分工支撑，属整合表述，本篇已标注。
  11. GUM 1993 年首版的年份：GUM 2008 文本自身只标「GUM 1995 with minor corrections」，1993 年份的直接逐字来自 GUM S1 前言与 VIM3 前言，本篇按此承重。
  12. 本篇自己的未竟之处：像本库既往各篇一样，本篇没有在开工前写下完工判据（与抗衰路线图篇、动物意识篇自查出的欠账同构）；本篇的补救是事后写下这十二条不确定度账——弱补救，如实登记。

来源清单（编号与正文对应）

  1. BIPM 官方页·Metre Convention（签署日期、17 国名单、1921 修订）— https://www.bipm.org/en/metre-convention
  2. BIPM 官方页·Official texts — https://www.bipm.org/en/official-texts
  3. 米制公约英文全译本（T. J. Quinn 译）— https://e-seimas.lrs.lt/rs/lasupplement/TAD/f6e781b07f0511e6a0f68fd135e6f40c/6e2637ccbaee11e6a3e9de0fc8d85cd8/
  4. BIPM 官方页·Member States（64 会员国 36 准会员，2026-08-02 实查）— https://www.bipm.org/en/member-states
  5. BIPM 官方页·Governance（三机构架构）— https://www.bipm.org/en/governance
  6. BIPM 官方页·28th CGPM（CGPM 职责与会议周期）— https://www.bipm.org/en/cgpm-2026
  7. BIPM Compendium（CGPM 表决规则「absolute majority」）— https://www.bipm.org/documents/20126/44107685/compendium.pdf/5d927b43-bc26-2ba7-59f2-e6040e60333b
  8. SI 手册第 9 版（BIPM 官方 PDF，V4.01）— https://www.bipm.org/documents/20126/41483022/SI-Brochure-9-EN.pdf
  9. BIPM 官方页·SI Brochure 出版物页 — https://www.bipm.org/en/publications/si-brochure
  10. BIPM 官方页·SI Brochure 第 8 版存档页 — https://www.bipm.org/en/si-brochure-8
  11. BIPM 官方页·The SI — https://www.bipm.org/en/measurement-units
  12. BIPM 决议数据库·Resolution 12 of the 11th CGPM (1960) — https://www.bipm.org/en/committees/cg/cgpm/11-1960/resolution-12
  13. BIPM 决议数据库·Resolution 1 of the 13th CGPM (1967) — https://www.bipm.org/en/committees/cg/cgpm/13-1967/resolution-1
  14. BIPM 决议数据库·Resolution 1 of the 17th CGPM (1983) — https://www.bipm.org/en/committees/cg/cgpm/17-1983/resolution-1
  15. BIPM 决议数据库·Resolution 1 of the 26th CGPM (2018) — https://www.bipm.org/en/committees/cg/cgpm/26-2018/resolution-1
  16. BIPM 决议数据库·Resolution 1 of the 24th CGPM (2011) — https://www.bipm.org/en/committees/cg/cgpm/24-2011/resolution-1
  17. BIPM 决议数据库·Resolution 1 of the 25th CGPM (2014) — https://www.bipm.org/en/committees/cg/cgpm/25-2014/resolution-1
  18. CIPM 1988 决议 1（KJ-90）— https://www.bipm.org/en/committees/ci/cipm/77-1988/resolution-1
  19. CIPM 1988 决议 2（RK-90）— https://www.bipm.org/en/committees/ci/cipm/77-1988/resolution-2
  20. 第 26 届 CGPM 会议记录（投票实况 54 国一致通过）— https://www.bipm.org/documents/20126/35655029/CGPM26.pdf/f1d8d7e6-2570-9479-9ee9-bb8d86138fe9
  21. BIPM 官方页·IPK — https://www.bipm.org/en/mass-metrology/ipk
  22. BIPM 官方页·History: kilogram — https://www.bipm.org/en/history-si/kilogram
  23. 第 19 届 CGPM 会议记录（三次周期检定日期，法文）— https://www.bipm.org/documents/20126/33145745/CGPM19.pdf/3f65ecf1-f86e-b86c-5c28-709d9f12eec4
  24. BIPM SI 修订 FAQ（IPK 现状逐字）— https://www.bipm.org/documents/20126/41483022/SI-FAQs-EN.pdf
  25. BIPM《Mise en pratique of the definition of the kilogram》4.1 节 — https://www.bipm.org/documents/20126/28428218/working-document-ID-6763/23933265-3f75-42d7-f1d5-fa20b90a4052
  26. BIPM 2015, Metrologia 52, 310（「There is no explanation available」）— https://iopscience.iop.org/article/10.1088/0026-1394/52/2/310
  27. Stock et al. 2017, Metrologia 54（「unexplained instabilities」）— https://google.iopscience.iop.org/article/10.1088/1681-7575/aa8d2d
  28. Davis 2003, Metrologia 40, 299（IPK 形态 90/10、约 39 mm）— https://iopscience.iop.org/article/10.1088/0026-1394/40/6/001
  29. Jabbour & Yaniv 2001, J. Res. NIST 106, 65（1884 交付、1888 比对、三次检定、K20/K4 数值）— https://nvlpubs.nist.gov/nistpubs/jres/106/1/j61jab.pdf
  30. NIST 官方·Kilogram: The Past（K20/K4）— https://www.nist.gov/si-redefinition/kilogram/kilogram-past
  31. NIST 官方·Kilogram: The Present（50 微克/百年、三次检定日期）— https://www.nist.gov/si-redefinition/kilogram/kilogram-present
  32. NIST 官方·Kilogram: Introduction（triple-locked vault）— https://www.nist.gov/si-redefinition/kilogram-introduction
  33. NIST 官方新闻稿 2011-11（three nested bell jars）— https://www.nist.gov/news-events/news/2011/11/redefining-si-base-units
  34. NIST 官方·Kilogram: The Kibble Balance — https://www.nist.gov/si-redefinition/kilogram/kilogram-kibble-balance
  35. NIST 官方·Kilogram: Mass and Planck’s Constant（NIST-4 13 ppb、NRC 9.1 ppb）— https://www.nist.gov/si-redefinition/kilogram/kilogram-mass-and-plancks-constant
  36. NIST 官方·Kilogram: Silicon Spheres and the International Avogadro Project（IAC、28Si 球、$3.2M）— https://www.nist.gov/si-redefinition/kilogram/kilogram-silicon-spheres-and-international-avogadro-project
  37. Haddad et al. 2017, Metrologia 54, 633（NIST-4 论文，13×10⁻⁹）— https://tsapps.nist.gov/publication/get_pdf.cfm?pub_id=923579
  38. NPL 官方·Bryan Kibble 生平 — https://www.npl.co.uk/about-us/history/famous/bryan-kibble
  39. NPL 官方·Kibble Balance — https://www.npl.co.uk/research/mass-mechanical-measurement/kibble-balance
  40. PTB 官方·Planck-Balance — https://www.ptb.de/cms/en/ptb/fachabteilungen/abt1/fb-11/ag-115.html
  41. NIM 官方英文页·Joule Balance（NIM-2 提交 CODATA）— https://en.nim.ac.cn/node/387
  42. NIM 官方中文页·NIM-2 不确定度 2×10⁻⁸ — https://www.nim.ac.cn/node/1189
  43. PTB-News·Avogadro 项目协调与澳大利亚抛光 — https://www.ptb.de/cms/en/presseaktuelles/journals-magazines/ptb-news/ptb-news-ausgaben/archivederptb-news/ptb-news-20103/the-new-kilogram-is-approaching.html
  44. Powerhouse 博物馆官方藏品档案（CSIRO 唯一生产球体）— https://collection.powerhouse.com.au/object/539785
  45. Newell et al. 2018, Metrologia 55, 125（CODATA 2017 特调，两方法 ≤20×10⁻⁹）— https://iopscience.iop.org/article/10.1088/1681-7575/aa950a
  46. Possolo et al. 2018, Metrologia 55, 29（六实验 <50×10⁻⁹，两路线收敛）— https://iopscience.iop.org/article/10.1088/1681-7575/aa966c
  47. GUM（JCGM 100:2008，BIPM 官方 PDF）— https://www.bipm.org/documents/20126/2071204/JCGM_100_2008_E.pdf
  48. GUM Supplement 1（JCGM 101:2008，BIPM 官方 PDF）— https://www.bipm.org/documents/20126/2071204/JCGM_101_2008_E.pdf
  49. VIM3（JCGM 200:2012，BIPM 官方 PDF；2.41/2.42 溯源定义）— https://www.bipm.org/documents/20126/2071204/JCGM_200_2012.pdf
  50. BIPM 官方·JCGM WG1 GUM 页（修订在案）— https://www.bipm.org/en/committees/jc/jcgm/wg/jcgm-wg1-gum
  51. JCGM News May 2024（GUM 升级文档套件、S1→GUM-7）— https://www.bipm.org/documents/20126/58020018/News-from-JCGM-May-2024.pdf
  52. NIST TN 1297（Taylor & Kuyatt，k=2 与 95%）— https://nvlpubs.nist.gov/nistpubs/legacy/tn/nbstechnicalnote1297.pdf
  53. ISO 官方页·ISO/IEC 17025:2017 — http://web.archive.org/web/20260722045935id_/https://www.iso.org/standard/66912.html
  54. ANAB 官方文件（17025 7.6 转述逐字）— https://anab.qualtraxcloud.com/ShowDocument.aspx?ID=23186
  55. ILAC 官方·政策系列页（P14:09/2020）— https://ilac.org/publications-and-resources/ilac-policy-series/
  56. ILAC P14:09/2020 正文（IAAC 官方镜像，y ± U 逐字）— https://iaac.org.mx/wp-content/uploads/2025/05/ILAC_P14_09_2020-1.pdf
  57. CIPM MRA 官方背景文件（1999-10-14 签署、38 NMI＋2 组织）— https://www.bipm.org/documents/20126/27085544/bipm%20publication-ID-2845/180c7d3a-ad83-8fd3-9a50-928acea67363
  58. CIPM MRA 文本（CIPM-MRA-2003.pdf）— https://www.bipm.org/documents/20126/43742162/CIPM-MRA-2003.pdf
  59. BIPM 官方·CIPM MRA 页 — https://www.bipm.org/en/cipm-mra
  60. BIPM 官方·2021 MRA 修订公告（258 机构）— https://www.bipm.org/en/-/cipm-mra-documents-1
  61. BIPM KCDB 首页（2026-08-02 快照：1278/26270/775）— https://www.bipm.org/kcdb/
  62. NIST 官方·Calibration of Mass Standards（溯源链实例）— https://www.nist.gov/programs-projects/calibration-mass-standards
  63. NIST 官方·Putting Einstein to the Test（GPS 38 微秒/天）— https://www.nist.gov/atomic-clocks/a-powerful-tool-for-science/putting-einstein-test
  64. NIST 官方博客·Taking Measure（卫星携带原子钟、手机边界）— https://www.nist.gov/blogs/taking-measure/putting-einstein-test-worlds-most-accurate-clocks
  65. NIST 官方·Knowing Where We Are（手机原子钟仍是未来目标）— https://www.nist.gov/atomic-clocks/a-technology-powerhouse/knowing-where-we-are
  66. USNO 官方·GPS Time Transfer（卫星钟构成与溯源）— http://web.archive.org/web/20211202033115id_/https://www.cnmoc.usff.navy.mil/Organization/United-States-Naval-Observatory/Precise-Time-Department/Global-Positioning-System/USNO-GPS-Time-Transfer/
  67. USNO 官方·International Time Scales and the BIPM — http://web.archive.org/web/20260730082035id_/https://www.cnmoc.usff.navy.mil/Our-Commands/United-States-Naval-Observatory/Precise-Time-Department/The-USNO-Master-Clock/International-Time-Scales-and-the-BIPM/
  68. Lombardi, NIST 官方存档（GPS 是主要时间分发器、可溯源）— https://tf.nist.gov/general/pdf/3140.pdf
  69. OIML 官方首页（使命）— https://www.oiml.org/en/front-page
  70. OIML B 006-2（Recommendations 定义）— https://www.oiml.org/en/files/pdf_b/b006-2-e19.pdf
  71. OIML R 117-1:2019（燃油加油机）— https://www.oiml.org/en/files/pdf_r/r117-1-e19.pdf/@@download/file/r117-1-e19.pdf
  72. NIST 官方 FAQ·International Legal Metrology（Handbook 44）— https://www.nist.gov/pml/owm/faqs/international-legal-metrology-faqs
  73. IFCC/IUPAC 1973 推荐（SI 单位在临床化学）— https://old.iupac.org/publications/pac/1974/pdf/3704×0517.pdf
  74. IFCC 2006 血糖报告（mmol/L 建议）— http://web.archive.org/web/20251019043823id_/https://www.degruyterbrill.com/document/doi/10.1515/CCLM.2006.275/html
  75. ISO 官方页·ISO 17511:2020（IVD 校准溯源，ideally 限定）— https://cdn.standards.iteh.ai/samples/69984/6f1630afed4349189ad6fc31b46b8444/ISO-17511-2020.pdf
  76. CIA World Factbook 官方档案版·Weights and Measures — https://www.cia.gov/the-world-factbook/about/archives/2022/references/weights-and-measures/
  77. NIST 官方博客·Busting Myths About the Metric System — https://www.nist.gov/blogs/taking-measure/busting-myths-about-metric-system
  78. NIST 官方·Metrication in Law（两套单位并行）— https://www.nist.gov/pml/owm/metrication-law
  79. NIST 官方 FAQ·U.S. Metrication（美制单位由 SI 定义）— https://www.nist.gov/pml/owm/faqs/us-metrication-faqs
  80. NIST 官方·Metric (SI) Program（现状与旧站退休）— https://www.nist.gov/pml/owm/metric-si
  81. 1975 Metric Conversion Act（公法 94-168，STATUTE-89 原件）— https://www.govinfo.gov/content/pkg/STATUTE-89/pdf/STATUTE-89-Pg1007.pdf
  82. 福特签署声明（completely voluntary）— https://www.presidency.ucsb.edu/documents/statement-signing-the-metric-conversion-act-1975
  83. 15 U.S.C. §205b（preferred system）— https://www.govinfo.gov/link/uscode/15/205b
  84. 15 U.S.C. 汇编 COMPS-1851（§205a(2) 国会发现逐字）— https://www.govinfo.gov/content/pkg/COMPS-1851/pdf/COMPS-1851.pdf
  85. EO 12770（Federal Register）— https://www.federalregister.gov/executive-order/12770
  86. EO 12770 附件（NIST SP 811 版本）— https://www.nist.gov/document/peo-12770-nist-sp-811-september-1998pdf
  87. PL 104-59 §205(c)（公路标志英制保护）— https://mutcd.fhwa.dot.gov/pdfs/11th_Edition/appendices.pdf
  88. MUTCD 第 11 版附录（English units 逐字）— https://uscode.ecfr.io/title/23/section/109
  89. Ken Alder《The Measure of All Things》序章（25 万种估计）— http://www.kenalder.com/measure/excerpts.htm
  90. 国会图书馆出版社样页（同句逐字）— https://catdir.loc.gov/catdir/enhancements/fy0705/2002070267-s.html
  91. Agnoli & d’Agostini 2004（arXiv physics/0412078，25 万转述）— https://arxiv.org/pdf/physics/0412078
  92. UNESCO 世界记忆·1795 法令档案描述 — https://media.unesco.org/sites/default/files/webform/mow001/france_decimal_metric_system.pdf
  93. 1795 法令英译（George Mason CHNM）— https://revolution.chnm.org/d/463
  94. 1795 法令法文原件（Gallica）— https://gallica.bnf.fr/ark:/12148/bpt6k62126153/texteBrut
  95. 1812 敕令原件（法国矿业年鉴）— https://annales.ensmp.fr/articles/1812-1/235-245.pdf
  96. 1837 年法律条文（法文原件转引）— http://reglementation-polmer.chez-alice.fr/Textes/loi_du_04.07.1837.htm
  97. histoire.du.metre（1837 法律英译）— http://histoire.du.metre.free.fr/en/Pages/Sommaire/13.htm
  98. Débarbat 2019（法国科学院，1799-06-22 入藏日期仲裁）— https://comptes-rendus.academie-sciences.fr/physique/item/10.1016/j.crhy.2018.12.002.pdf
  99. Johnson Matthey 技术评论 2000（1799-07-22 分歧方）— http://web.archive.org/web/20251119200415id_/https://www.technology.matthey.com/article/44/3/125-134/
  100. Johnson Matthey 公司史 1973（1882 合同、40 千克）— http://web.archive.org/web/20260208142436id_/https://technology.matthey.com/content/journals/10.1595/003214073X1726668
  101. UK Metric Association（George Matthey 铸件）— https://ukma.org.uk/what-is-metric/first-international-prototypes/
  102. NBS Circular 570（US-20 名称）— https://nvlpubs.nist.gov/nistpubs/Legacy/circ/nbscircular570.pdf
  103. BBC 2007（Davis 谈漂移之谜）— http://news.bbc.co.uk/2/hi/sci/tech/7084099.stm
  104. NIST 官方新闻稿 2018-11-16（60 国口径）— https://www.nist.gov/news-events/news/2018/11/historic-vote-ties-kilogram-and-other-units-natural-constants
  105. BIPM 新闻稿 2018-11-16（「will now be retired」，镜像）— https://www.keiryou-keisoku.co.jp/pdf/1810BIPM_Press%20release_After_FINAL.pdf
  106. WMD 2019 官方新闻稿 — https://2019.worldmetrologyday.org/press_release.html
  107. NIST 官方·A Turning Point for Humanity — https://www.nist.gov/si-redefinition/turning-point-humanity-redefining-worlds-measurement-system
  108. NIST SP 330-2019（美国版 SI 手册）— https://nvlpubs.nist.gov/nistpubs/SpecialPublications/NIST.SP.330-2019.pdf
  109. NIST JRES 1994/95 Taylor（KJ-90 生效）— https://nvlpubs.nist.gov/nistpubs/jres/094/jresv94n2p95_A1b.pdf
  110. NIST SP 958（JVS 史）— https://nvlpubs.nist.gov/nistpubs/sp958-lide/315-318.pdf
  111. VIM3 在线版（2.9 measurement result）— https://jcgm.bipm.org/vim/en/2.9.html

尾注：本篇的编号与署名

本篇为机制裁决第 117 篇·对称双向第 112 篇·全库 175 篇。正面范本不改变编号规则——它同样是一次机制裁决，裁的对象从「某个主张」换成「一门学科的制度形态」。全部承重引用经六捆并行调研与主笔逐字亲核；本篇所有官方文件均为直取件（SI 手册第 9 版 2.0MB、GUM 1.9MB、GUM S1 1.5MB、VIM3 4.0MB、CGPM26 6.0MB、Compendium 361KB，全部 200 直取）。正文、来源清单与数字口径以本文件为准；双版 HTML 由同目录生成器从本文件生成。

标签: