---
name: path-dependence
description: >-
  Reference card for path dependence: how early choices and increasing returns
  lock systems into trajectories that are costly to reverse. Load when mapping
  lock-in, sunk commitments, or why "better" options stay unused.
id: path-dependence
layer: reference
scope: official
disciplines:
  - economics
  - history
language: zh
tags: [history, institutions, lock-in, increasing-returns]
version: 0.1.0
metadata:
  openwisdom: true
---

# 路径依赖 / Path Dependence

## Definition

路径依赖描述的是：早期条件、偶然事件与自我强化机制，使系统沿着某条轨迹推进，后续切换成本上升，即便事后看存在更优选项。  
Path dependence means early conditions, contingencies, and self-reinforcing mechanisms push a system along a trajectory where switching becomes costly—even when better options appear later.

它不是「历史决定论」：强调的是**递增回报、转换成本与协调惯性**，而不是命运不可改写。  
It is not historical fatalism: the claim is about increasing returns, switching costs, and coordination inertia—not destiny.

经典文献方向：David (1985) 技术锁定；Arthur (1989, 1994) 递增回报与 QWERTY 类案例；Pierson (2000) 政治与政策中的路径依赖。

## When to use

- 场景 `macro-scan` 标注制度节点、否决点与「为何现状难改写」时。  
  In `macro-scan`, when marking institutional nodes, veto points, and why the status quo is hard to rewrite.
- 比较政策、技术标准、组织惯例或个人长期承诺的可逆性时。  
  When comparing reversibility of policies, standards, routines, or long-horizon personal commitments.
- 需要区分「暂时惯性」与「真正锁定」（沉没成本 + 网络效应 + 学习曲线 + 政治否决）时。  
  When you must separate temporary inertia from real lock-in (sunk costs, network effects, learning curves, political vetoes).

**不要用于：** 把一切失败归因「路径依赖」而跳过激励与行动者分析；或用它为现状辩护、否定改革可能性。

## Core claims

1. **早期事件可被放大** — 在递增回报下，小的早期优势或偶然选择可被正反馈放大成长期主导。  
   Early events can be amplified under increasing returns.
2. **锁定不等于最优** - 路径依赖允许「次优但稳固」的均衡：效率论证不能从「存在」直接推到「最优」。  
   Lock-in ≠ optimality: stable sub-optimal equilibria are possible.
3. **机制要写具体** — 至少点名一类机制：沉没投资、网络/协调效应、学习与适应、制度互补、政治否决联盟。  
   Name mechanisms: sunk investments, network/coordination effects, learning, institutional complementarities, veto coalitions.
4. **切换成本是可观察的** — 锁定强度可用转换成本、协调人数、专用资产、法律/声誉成本来检验，而不是口号。  
   Lock-in strength is checkable via switching costs, coordination scale, specific assets, legal/reputational costs.
5. **临界点与断裂可以发生** - 外部冲击、技术代差、联盟重组或规则改写可打开新路径；路径依赖解释惯性，不禁止断裂。  
   Critical junctures and breaks remain possible; path dependence explains inertia, not impossibility of change.

## Limits

- **过度决定论**：历史路径说明约束，不取消能动与策略。  
  History constrains; it does not erase agency.
- **事后合理化**：凡是「变成这样了」都贴路径依赖标签，会掏空概念。  
  Not every outcome is path dependence—name the reinforcing loop.
- **尺度混淆**：技术标准锁定 ≠ 个人习惯 ≠ 国家政体路径；机制与证据层级要对齐。  
  Align scale: tech standards ≠ personal habits ≠ regimes.
- **政策滥用**：以路径依赖为由拒绝比较选项，是修辞，不是分析。  
  Using path dependence to shut down option comparison is rhetoric, not analysis.
- **证据边界**：多数现实案例是机制推断，不是实验室证明；证据不足处写不确定。  
  Most cases are mechanism inference; mark uncertainty.

## Notes

- 分析在用户自己的 coding agent 中运行，不在 Openwisdom 服务器上。  
  Analysis runs in the user's coding agent, not on Openwisdom servers.
- 本文件是 **reference**（理论卡片），由场景 skill 引用；不单独冒充完整工作流。  
  This is a reference card cited by scenario skills—not a full workflow.
- 不编造数据、客户或安装热度；引用经典方向时不伪造页码或 DOI。  
  Do not invent data, customers, or install heat; cite classic directions without fake page numbers or DOIs.