Analogical Reasoning
Analogical Reasoning: When facing a hard problem, ask: "What other domain has solved a structurally similar problem?" Then adapt the solution. Gutenberg didn't invent printing from scratch β he adapted wine-press mechanics. Darwin didn't invent evolution from scratch β he adapted Malthusian population dynamics. Most breakthrough solutions are structural analogies from unexpected domains.
What Is Analogical Reasoning?β
Analogical reasoning is the process of solving problems by identifying structural parallels between a new problem and one already solved in another context. Unlike surface-level similarity (a new social app is "like Facebook"), analogical reasoning seeks deep structural similarity β the underlying relationships between elements are the same, even if the surface features differ dramatically.
The cognitive science research (Gentner, Holyoak) distinguishes two types of similarity: surface similarity (same features β same industry, same materials, same customers) and structural similarity (same relationships between elements β even in completely different domains). Analogical reasoning exploits structural similarity; pure pattern-matching uses surface similarity.
Deep analogies have driven some of history's most important innovations. Gutenberg's printing press adapted the screw mechanism from wine and olive presses β same physics, completely different purpose. Darwin's natural selection adapted Malthus's economic insight about population and resources to biological organisms β same logic of variation + selection pressure, completely different domain. Shannon's information theory adapted Boltzmann's thermodynamic entropy β same mathematical structure, different physical substrate. The ability to see structural similarities across domain boundaries is a signature characteristic of exceptional problem-solvers.
The practical implication: when stuck on a problem, look for analogies rather than working harder within the current domain. Where has a similar structure been solved? In biology, physics, economics, history, military strategy, or game design? Often, the hard work has already been done β just in a different domain.
How It Worksβ
Step 1: Abstract the problem to its structural features
β Strip away domain-specific details
β What are the key relationships, constraints, and goals?
β "We have limited resources, many competing priorities, and need to allocate optimally"
Step 2: Search for structural analogies in other domains
β What other systems have faced the same structural challenge?
β Consider: biology, physics, military, economics, games, architecture
Step 3: Identify the deep structural match
β How closely do the relationships map?
β What must be true for the analogy to hold?
Step 4: Examine the solution in the source domain
β How is the analogous problem solved there?
β What are the key mechanisms and principles?
Step 5: Adapt the solution to your domain
β What translates directly? What must be modified?
β What breaks down (limits of the analogy)?
Step 6: Test and iterate
β Does the adapted solution actually work in the target domain?
Three Real-World Examplesβ
Amazon Supply Chain from Military Logisticsβ
Amazon's early supply chain thinking drew explicit analogies from military logistics β specifically the US Army's experience managing complex, time-critical, geographically distributed supply chains under resource constraints. The structural analogy: both Amazon and military logistics face the problem of getting the right item to the right location at minimum cost within a time constraint, at massive scale. Military logistics had developed solutions for hub-and-spoke distribution, buffer stock optimisation, and real-time inventory tracking that directly informed Amazon's fulfillment architecture.
Epidemiology Applied to Computer Virus Spreadβ
Early researchers studying how computer viruses spread recognised that the dynamics were structurally identical to biological disease epidemiology: infection rates, recovery rates, immunity, and population density all had direct computational analogues. The SIR model (Susceptible-Infected-Recovered) from epidemiology was adapted directly to model computer worm propagation. This analogical transfer gave the cybersecurity field decades of mathematical tools without having to develop them from scratch.
Lean Manufacturing Principles Applied to Softwareβ
Toyota's lean manufacturing principles β waste elimination, just-in-time production, continuous improvement β were developed for physical manufacturing. Mary and Tom Poppendieck recognised the structural analogy to software development: both involve complex processes with variable inputs, potential for waste (unnecessary features, defects, handoffs), and improvement cycles. This produced the "Lean Software Development" movement and directly influenced Agile methodologies. The analogy wasn't surface-level (software and car manufacturing look nothing alike) but structural (both are knowledge-work processes requiring waste reduction and flow optimisation).
When to Use Itβ
β Analogical Reasoning is most valuable for:
- Innovation challenges where incremental thinking within the domain has stalled
- Complex problems where the domain lacks established solutions
- Cross-disciplinary research and product development
- Generating multiple solution options quickly
β Exercise caution when:
- The analogy is surface-level rather than structural (danger of misleading parallels)
- The solution domain has very different constraints that make transfer invalid
- Over-reliance on analogies substitutes for deep understanding of the actual problem
| Pairs well with | Why |
|---|---|
| First Principles | First Principles abstracts to fundamentals; Analogy searches for equivalent solutions elsewhere |
| Reframing | Reframing and Analogy both break domain-specific constraints |
| Thought Experiment | Thought experiments often use analogical structures to test ideas |
| Lateral Thinking | Lateral thinking and analogical reasoning are complementary creative strategies |
Common Misuses and Limitationsβ
Mistaking surface analogy for structural analogy. "This is just like Uber for X" is often surface analogy β same two-sided marketplace surface features, without the same structural economics (Uber's driver supply elasticity, rider frequency, and unit economics don't transfer to "Uber for [rare service with infrequent use]"). Test analogies by mapping specific structural relationships, not just overall patterns.
Forcing analogies that don't fit. Once you've found an analogy you like, confirmation bias leads to forcing it even where it breaks down. Every analogy has limits β the discipline is knowing exactly where those limits are.
Transferring the solution without adapting it. Military logistics principles don't translate directly to e-commerce without adaptation β the constraints differ (no bullets, no strategic secrecy requirements, different geography). Successful analogical transfer requires both recognising the structural similarity and carefully adapting for domain differences.
Related Modelsβ
| Model | Relationship |
|---|---|
| First Principles | First Principles rebuilds; Analogical Reasoning transfers |
| Reframing | Both techniques escape current-domain thinking |
| Thought Experiment | Thought experiments often use analogical structures |
| Abstraction Laddering | Abstraction enables finding analogies by stripping domain-specific surface features |
Frequently Asked Questionsβ
How do you find good analogies for a problem?
Three strategies: (1) Abstract the problem to its structural essence (what are the key relationships, constraints, and goal dynamics?), then deliberately search multiple domains for the same structure; (2) Use a "solution library" approach β study how different fields (biology, physics, military, economics, architecture) solve recurring problems (resource allocation, information transmission, defence, scaling) and ask which applies; (3) Consult domain outsiders β someone from a different field will naturally see structural similarities invisible to domain insiders.
What makes an analogy "deep" vs "superficial"?
Deep analogies map relationships between elements, not just surface features. "Neural networks are like the brain" is superficially analogous (both process information) but structurally limited (the mechanisms differ dramatically). "Natural selection is like economic competition" is structurally deep β both involve variation, differential reproduction/growth based on fit with environment, and iterative selection pressure. The deeper question: if you map element A in domain 1 to element A' in domain 2, does the relationship between A and B in domain 1 hold between A' and B' in domain 2? If yes, the analogy is structurally deep.
Can analogical reasoning lead you astray?
Yes β "Lure of the Analogy" is a documented cognitive bias. Once a compelling analogy is found, people over-apply it, mapping elements that don't actually correspond and drawing false conclusions. Finance analogies from physics (Black-Scholes assumes normally distributed returns like Brownian motion) have produced famous failures when the analogy broke down in the tails. The discipline is testing the analogy explicitly at each mapping point, identifying where it holds and where it doesn't, and not trusting conclusions derived from the parts of the analogy that don't map cleanly.
Further Readingβ
- Gentner, D. & Holyoak, K.J. (1997). "Reasoning and Learning by Analogy." American Psychologist
- Holyoak, K. & Thagard, P. (1995). Mental Leaps: Analogy in Creative Thought
- Dunbar, K. (2001). "The Analogical Paradox" β how scientists use analogy in real research
Apply with AIβ
π Find analogies for your problem with MindMax β
This page is part of the MindMax Mental Models Knowledge Base.