← 返回
AI智能 中文

Analogical Reasoning

Apply analogical reasoning to transfer knowledge from familiar domains to unfamiliar ones. Use when the user needs creative problem-solving by finding struct...
运用类比推理将知识从熟悉领域迁移至陌生领域。适用于用户需要通过发现结构相似性进行创造性解决问题的场景。
wanikua
AI智能 clawhub v1.0.0 1 版本 100000 Key: 无需
★ 0
Stars
📥 672
下载
💾 9
安装
1
版本
#latest

概述

Analogical Reasoning

Analogical reasoning transfers knowledge from a familiar domain (the "source") to an unfamiliar one (the "target") by identifying structural similarities. It's how humans naturally make sense of the new — by connecting it to the known. Used brilliantly by scientists (Rutherford: atom is like a solar system), entrepreneurs (Uber for X), and legal scholars (case law precedent). But analogies can also mislead when surface similarities mask deep structural differences. The key is knowing when the mapping holds and when it breaks.


Analyze the current topic or problem under discussion using analogical reasoning. Find illuminating parallels, map them carefully, and extract transferable insights — while being honest about where the analogy breaks down. Apply this framework to whatever the user is currently working on or asking about.


Step 1: Understand the Target Domain

First, deeply understand the problem you're trying to solve.

  • What are the key elements of this problem? (Actors, relationships, dynamics, constraints, goals)
  • What makes this problem hard? Where is the core difficulty?
  • What is unknown or uncertain about this domain?
  • What structure underlies the problem? (Causal relationships, feedback loops, trade-offs)
  • Temporarily set aside domain-specific details — focus on the abstract structure.

Step 2: Generate Source Analogies

Find domains that share structural features with your target.

Search broadly across domains. For each, briefly state the analogy:

Near Analogies (same general field)

  • What similar problem in a related domain has been solved before?
  • What does the nearest competitor or adjacent industry do?

Far Analogies (completely different fields)

  • Nature/Biology: What organism, ecosystem, or evolutionary process mirrors this?
  • History: What historical event or era parallels this situation?
  • Engineering/Physics: What physical system behaves similarly?
  • Games/Sports: What game or sporting strategy has this structure?
  • Medicine: What medical condition or treatment protocol is analogous?
  • Military: What military strategy or campaign matches?
  • Art/Music: What creative process or composition mirrors this?
  • Economics: What market or economic phenomenon has the same dynamics?

Generate at least 5 source analogies, with at least 2 from distant domains. Far analogies are often more creative and insightful than near ones.

Step 3: Deep Mapping — Structure the Best Analogies

For the top 3 most promising analogies, perform a detailed structural mapping:

Analogy: [Source Domain] → [Target Domain]

Source ElementTarget ElementMapping Strength
---------
[Actor/component in source][Corresponding actor in target]Strong/Moderate/Weak
[Relationship in source][Corresponding relationship]Strong/Moderate/Weak
[Dynamic/process in source][Corresponding dynamic]Strong/Moderate/Weak
[Constraint in source][Corresponding constraint]Strong/Moderate/Weak
[Outcome in source][Predicted outcome in target]Strong/Moderate/Weak

Key questions for each mapping:

  • Is the correspondence structural (deep) or merely surface (superficial)?
  • Is the causal mechanism the same, or just the appearance?
  • Does the mapping scale appropriately?

Step 4: Extract Transferable Insights

For each strong analogy:

  • What solutions or strategies worked in the source domain?
  • What principles underlie those solutions (not the specific details — the abstract principles)?
  • How would those principles translate to the target domain?
  • What predictions does the analogy make about the target? (These are testable!)
  • What pitfalls were discovered in the source domain that the target should avoid?
  • What timeline or trajectory did the source domain follow? Does the target follow a similar path?

Step 5: Identify Where the Analogy Breaks Down

This is the most important step. All analogies are wrong; some are useful.

For each analogy:

  • Where do the structural correspondences fail?
  • What key features of the target domain have no counterpart in the source?
  • What features of the source domain are irrelevant or misleading in the target?
  • Where does the analogy predict something false about the target?
  • What is the disanalogy — the most important difference?
  • How might relying on this analogy lead you astray?

Rate the analogy's overall reliability:

  • High fidelity: Core structure maps well, breakdowns are in peripheral details
  • Medium fidelity: Structure partially maps, some important differences
  • Low fidelity: Surface similarity only, deep structure differs significantly

Step 6: Triangulate Across Analogies

  • Where do multiple analogies converge on the same insight? (High confidence)
  • Where do they diverge? (Indicates complexity or important nuance)
  • What insight appears in the far analogies that's invisible in the near ones?
  • What composite analogy (combining elements from multiple sources) best captures the target?

Step 7: Generate Novel Solutions

Based on the analogical analysis:

  • What specific solutions or approaches from the source domains could be adapted?
  • What novel combination of source-domain strategies creates something new?
  • What would a practitioner from the source domain suggest if they saw this problem?
  • What experiment or test would validate whether the analogical transfer actually works?

Synthesis

  • State the most illuminating analogy and the key insight it provides.
  • Acknowledge the limits of the analogy explicitly.
  • Recommend concrete actions inspired by the analogy, adjusted for where it breaks down.

George Pólya said: "Analogy pervades all our thinking." The art is not in finding analogies — the human mind does that instinctively. The art is in testing them rigorously: mapping the structure, checking the correspondence, and being honest about where the parallel fails.

版本历史

共 1 个版本

  • v1.0.0 当前
    2026-03-29 19:25 安全 安全

安全检测

腾讯云安全 (Keen)

安全,无风险
查看报告

腾讯云安全 (Sanbu)

安全,无风险
查看报告

🔗 相关推荐

ai-intelligence

ontology

oswalpalash
类型化知识图谱,用于结构化智能体记忆与可组合技能。支持创建/查询实体(人员、项目、任务、事件、文档)及关联...
★ 711 📥 243,740

mempalace

wanikua
MemPalace — 本地 AI 记忆,召回率 96.6%。对过去对话进行语义搜索,知识图谱记录时序事实,宫殿架构(翼/房间...)
★ 4 📥 1,574
ai-intelligence

Proactive Agent

halthelobster
将AI智能体从任务执行者升级为主动预判需求、持续优化的智能伙伴。集成WAL协议、工作缓冲区、自主定时任务及实战验证模式。Hal Stack核心组件 🦞
★ 834 📥 213,027