# Analogies should be used as thinking tools but not as evidence An analogy used as a thinking tool looks like Einstein’s dream about riding a light beam. “Moving at the speed of light would be like riding a light beam - what if this were the case? What are the consequences?” Or Francis and Crick’s mechanical model of DNA - it allowed them analogize between the model and the data they were seeing. ==This example isn’t as good as I’d like== An analogy used as an evidence looks like “X is like Y, Y has property Z, therefore X also has property Z.” The current period of history looking like <the Fall of Rome> <Weimar Germany> <the 1970s> etc. Unless the analogy is perfect, the deduction of property Z is not necessarily true — the world is complex and chaotic which means that small differences of inputs can lead to huge difference of outcomes. The line between the two is subtle — a historical analogy *can* be used as a thinking tool if it generates questions like “what would we expect to happen that would either increase or decrease the resemblance between these two time periods?” Maybe generating questions vs conclusions is the discriminator between the two cases: it’s legitimate to use analogies to generate questions but not conclusions 🤔 ### Related * [[People coming up with analogies during a conversation is like a lookahead search]] * [[Structuring knowledge is expensive]] * [[chanSOLVENTMixedInitiative2018]] [^1]: Math is a cool special case, where if you are confident that the analogy between the mathematical representation of something and the thing itself holds, then showing that it works in math is actually strong evidence that it is true in the thing. (This of course assumes that the world is not, in fact, just math, which some philosophers have suggested.) The power of mathematical analogies as evidence obviously depends on how good the analogy is: we obsess about closed-form representations because when they work (as in with light waves) they work *very* well. [Web URL for this note](http://notes.benjaminreinhardt.com/Analogies+should+be+used+as+thinking+tools+but+not+as+evidence) [Comment on this note](http://via.hypothes.is/http://notes.benjaminreinhardt.com/Analogies+should+be+used+as+thinking+tools+but+not+as+evidence)