Bayesian Charity

post by Wes F (wfenza) · 2021-01-22T14:49:59.517Z · LW · GW · 5 comments

Cross-posted from Living Within Reason

In philosophy, the Principle of Charity is a technique in which you evaluate your opponent’s position as if it made the most amount of sense possible given the wording of the argument. That is, if you could interpret your opponent’s argument in multiple ways, you would go for the most reasonable version.

– UnclGhost [LW · GW]

There is a corollary to the Principle of Charity which I’m calling Bayesian Charity. It says that, in general, you should interpret your opponent’s wording to be advocating the most popular position.

This is implied by Bayes Theorem. Your prior for whether someone believes something unpopular should be related to its popularity. Since an unpopular belief, by definition, has few believers, your prior for whether someone believes an unpopular position should be lower the less popular the position is, and you should update your prior based on how clear their statement was.

Extraordinary claims require extraordinary evidence. If someone said something to me suggesting that the moon landing was real, I wouldn’t think twice about it. Most people believe the moon landing was real. However, if someone said something suggesting that the US has an underground base on Mars, I’m going to ask for clarification.

Context matters, of course. If you’re at a flat Earther convention, your prior should be much higher that any random person believes in a flat Earth. But in general, if someone says something that almost no one believes, it’s probably a good idea to confirm you understood what they meant before assuming they hold such a fringe opinion.

5 comments

Comments sorted by top scores.

comment by Dagon · 2021-01-22T18:25:36.975Z · LW(p) · GW(p)

Wait.  Interpreting by (your perception of) popularity seems to be less truth-seeking than the most sensible OR that strongest (see https://www.lesswrong.com/tag/steelmanning [? · GW] ).  

If you're talking about using it as evidence that at least one person makes this claim, then I see your point - you should consider all possible interpretations of an argument, weighted by likelihood that the person is making that specific claim.  That likelihood is subject to Bayesean calculations.  "popularity" is a fine prior, but you ALWAYS have additional evidence to alter your weighting of probability from that point.

Replies from: wfenza, ChristianKl
comment by Wes F (wfenza) · 2021-01-24T15:38:05.335Z · LW(p) · GW(p)

I'm not understanding your disagreement. Of course popularity is just a prior. The less popular a given position, the stronger your prior should be against the other person holding it. Doing that will lead you to be less wrong about what the other person means by what they're saying.

What is the difference between what you said and "your prior for whether someone believes an unpopular position should be lower the less popular the position is, and you should update your prior based on how clear their statement was"?

comment by ChristianKl · 2021-01-23T14:17:49.540Z · LW(p) · GW(p)

If you're talking about using it as evidence that at least one person makes this claim, then I see your point 

The concept Principle of Charity cames from linguistics and was created to speak about what people actually mean (which claims they make).

Replies from: Dagon
comment by Dagon · 2021-01-23T16:04:03.671Z · LW(p) · GW(p)

It's slightly different in philosophy and in linguistics/conversational norms (as I understand).  Conversationally, the principle of charity is to assume the kindest interpretation, in order not to derail the conversation.  Philosophically, it's to assume the strongest interpretation, in order to understand what claims about the world make sense.

In neither case would I recommend "popularity" as the basis for interpretation, except as a prior before you know anything about the topic or claimant.

Replies from: ChristianKl
comment by ChristianKl · 2021-01-23T18:16:09.827Z · LW(p) · GW(p)

The term was proposed by Wilson in the paper Substances without Substrata for a way to figure out what people actually mean. It's a way to decide that the claim Caesar was a Roman dictator is supposed to refer to Julius Caesar and not some other individual named Caesar. In practice Philosophers do use the term to mean a variety of different things but the original definition was about that.