---
{
  "id": "ASD-EXP-TESTS-FALSIFICATION",
  "path": "/tests/falsification",
  "title": "分层证伪与影响范围",
  "title_en": "Falsification Tree and Scope",
  "category": "实验研究计划",
  "summary": "把装置映射、本构假设、连续理论与微观候选分开，声明每次排除影响哪些主张。",
  "template": "test",
  "version": "1.0.0",
  "updated": "2026-09-22",
  "date_published": "2026-09-22",
  "canonical": false,
  "source_tier": "published",
  "theory_status": "OPEN",
  "evidence_status": "UNTESTED",
  "locale": "zh",
  "module_key": "experimental-program",
  "module_navigation": [
    {
      "path": "/tests",
      "label": "实验计划",
      "en": "Program"
    },
    {
      "path": "/tests/predictions",
      "label": "预测登记",
      "en": "Registry"
    },
    {
      "path": "/tests/falsification",
      "label": "证伪范围",
      "en": "Falsification"
    },
    {
      "path": "/tests/priority-experiments",
      "label": "优先实验",
      "en": "Priorities"
    },
    {
      "path": "/tests/reproducibility",
      "label": "复现标准",
      "en": "Reproducibility"
    }
  ],
  "canonical_scope": "实验研究计划；不新增正式物理结论",
  "source_chapter": "Existing canonical module snapshots / 现有正式模块快照",
  "source_reference": "Predictions & Falsification v1.0 · review 2026-09-22",
  "source_version": "EXP-7ca61744dcd2",
  "source_sha256": "7ca61744dcd2129e7fcd6902223c71166f3ededeb84b599d544cb20dba88130a",
  "source_kind": "manifest",
  "source_manifest": "predictions-source-manifest.json",
  "source_manifest_url": "/downloads/ASD_Predictions_Source_Manifest.json",
  "last_reviewed": "2026-09-22",
  "depends_on": [],
  "related": [
    "/tests",
    "/tests/predictions",
    "/tests/priority-experiments",
    "/tests/reproducibility"
  ],
  "open_problems": [],
  "equations": [],
  "sources": [
    {
      "label": "Predictions source and scope manifest",
      "kind": "research-program",
      "url": "/downloads/ASD_Predictions_Source_Manifest.json"
    },
    {
      "label": "Ota et al. (2022), Nature Physics 18, 1436–1440",
      "url": "https://www.nature.com/articles/s41567-022-01767-w",
      "kind": "external-reference"
    },
    {
      "label": "PDG (2024), Cosmic Rays review",
      "url": "https://pdg.lbl.gov/2024/reviews/rpp2024-rev-cosmic-rays.pdf",
      "kind": "external-reference"
    },
    {
      "label": "Wolf et al. (2006), Cold Atom Clock Test of Lorentz Invariance in the Matter Sector",
      "url": "https://arxiv.org/abs/hep-ph/0601024",
      "kind": "external-reference"
    },
    {
      "label": "BACON Collaboration (2021), optical frequency ratio network",
      "url": "https://www.nist.gov/publications/frequency-ratio-measurements-18-digit-accuracy-using-network-optical-clocks",
      "kind": "external-reference"
    }
  ],
  "outline": [
    "预先固定什么",
    "空结果怎样限制模型",
    "不同层级的失败",
    "发布正结果、空结果与失败记录"
  ],
  "seo_title": "分层证伪与影响范围 | ASD Predictions & Falsification v1.0",
  "canonical_url": "https://asd-physics.org/tests/falsification",
  "translation_path": "/en/tests/falsification",
  "mathematical_review": "INCOMPLETE",
  "evidence_note": "尚无闭合定量判别预测；工作台为规划或比较模型。"
}
---
<FalsificationTree />

## 预先固定什么

一次可解释的排除需要固定：理论及版本、适用域、允许参数域、校准数据、目标观测量、选择规则、误差与相关性、似然或检验统计量，以及拒绝域和检验功效。若参数允许把振幅任意调到零，一次空结果只能限制参数，不能排除整个模型族。

## 空结果怎样限制模型

设固定模型预测非零 $\Delta O=X$。若数据在完整不确定度下与零相容，同时置信区间排除 $X$，且共同背景、校准和选择偏差已检验，则在预声明置信水平上排除这一模型与装置映射的组合。检验标准不能只写“精度小于 X”：还要考虑相关噪声、参数退化、多重检验与检测功效。

相反，偏离零不自动支持 ASD：还须比较 ASD 的符号、相位、形状和幅度，排除普通系统效应，并检验比较模型。系统协方差没有闭合时，显著性也没有闭合。

## 不同层级的失败

| 被检验对象 | 直接影响 | 不自动否定 |
| --- | --- | --- |
| 指定 LVE—钟频率映射 | 该钟种、跃迁与链路的构成 | 所有 LVE 实现 |
| 指定电子候选 | 作用量、拓扑识别及其粒子输出 | 宏观引力、计量定义 |
| 指定 FCC 微观模型 | 离散耦合、连续对应与该候选 | 独立连续纵横核心 |
| 指定探测统计机制 | 其事件率与统计极限 | Fourier 数学或所有有效包络 |

依赖图中的“重新核查”比“全部否定”更准确。某个数学恒等式与物理识别条件必须分开记录。

## 发布正结果、空结果与失败记录

正结果报告可识别性和对替代模型的比较；空结果报告排除参数域；同形结果报告装置不可区分；模型无法闭合时报告缺失方程。全部进入 [Research Ledger](/research/research-ledger)，避免事后改变待检验主张而保留原先预测标签。