I've wondered if one could arrive at a good decision using no evidence or facts, but only based on how the different sides debate the issue. The idea is you could determine which party is right based solely on how they say things, rather than how strong the evidence is. You look at how different parties debate. You then compare their styles or types of arguments with debates from the past that have since been settled.
It's Bayesian spam filtering, applied to debates. I don't know how accurate it would be, or if it would be easy to game, but it seems useful for educating guesses.
Here's a rule-of-thumb that I go by: Whoever takes the other side into account best is the winner. For example:
Side 1: Language x sucks because of y!
Side 2: While it's true that x potentially sucks because of y, y is easily mitigated by z, and y still offers benefits a, b, & c that most other languages don't.
Even if you have only a shallow understanding of the argument, it's easy to see that in most real-world arguments Side 2 should win.
It's related to who gets angriest because people usually get angry because they have run out of other options.
What if the other person is just being an idiot? What if you think he's not capable of understanding? What if he's a child? What if you think he's trolling? Is better argument still the way to go?
I am a parent, so I can answer your question with respect to children: raising your voice does not do anything except escalate the discussion into an emotional melt-down.
With respect to people you deem to be idiots and trolls, we have known the answer for a very long time. One of the more amusing bits of advice about the matter came from Mark twain:
"Never argue with a fool, onlookers may not be able to tell the difference."
I agree that yelling at children doesn't help. But I'm wondering if you would debate with a child, or try to persuade him with good ideas, or would use some other sort of approach that isn't yelling or persuading?
This is a good hypothetical. I like it. The "easy" win in such a procedure is to decide which debater is self-consistent. Unfortunately a consistent argument based on faulty premises yields false conclusions. However, since we're allowed to look at past debates for context, I think we can sidestep that issue. We need to determine which past debate-winning arguments led to objectively advantageous actions. This is a non-trivial problem. Let's additionally sidestep it by allowing the human being to weight outcomes based on past arguments.
Aside: Are we defining a 'settled' debate as one where the resultant action came from one side of the debate over the other, or are we additionally taking into account whether we consider the outcome of said action is positive or negative?
I think that it would be simple to identify logical arguments (well, fallacies anyway) which imply that the position being asserted is actually untenable. Identifying arguments which satisfy the inverse - that is, that the style of argument correlates with an advantageous action as a result, falls back into the trap where false premises lead to false conclusions.
Aside: I wonder if there are arguments where objectively false conclusions still lead to advantageous actions? Now there's an interesting question.
I suppose that if you could semantically dissect an argument and then correlate its structure against weighted outcomes of previous arguments this system could yield intriguing results. I would very much like to see what logical constructions correlate with poor outcomes. Perhaps it would reveal self-defeating human tendencies, or arguments which push aside rational behavior?
Then again, the strength of the Bayesian filter approach is that it is dumb. It doesn't have to do a semantic analysis. So we avoid going down the path of strong-ai or special-cased rules when using it. Does the purely textual content of a debate correlate strongly enough with the semantic content to distinguish between good arguments and bogus ones? I think the only way to tell would be to try.
I suspect that many of the most vigorous debates would end up revealing that the relative advantagion or bogosity of a given argument correlate most strongly with what your individual desired outcome is. Then we've just led ourselves to the question of which outcome is superior to the other.
Thanks for an interesting few moments of introspection!
I like the idea, but if it were applied, I bet Scientologists would win the debate against the over emotional "in your face" techniques used by their opponents.
A pretty good test is to ask each side to summarize the other side's position in a way the other side would endorse.
If one side succeeds, and the other fails, that reveals who understands both positions, and who doesn't.
Of course they can still make excuses, such as saying that they don't understand the other guy's position because he didn't explain it clearly. And they can say they asked questions about it to try to understand, but he still didn't answer well. That could even be true. But it usually isn't.
It's Bayesian spam filtering, applied to debates. I don't know how accurate it would be, or if it would be easy to game, but it seems useful for educating guesses.