In both programming and mathematics, there’s a sense that only 3 numbers need no justification (0,1, infinity). Everything else is messier.
Unfortunately something similar is true for arguments as well. This creates a problem.
Much of the time, you want to argue that people underrate X (or overrate X). Or that people should be more Y (or less Y).
For example, people might underrate human rationality. Or overrate credentials. Or underrate near-term AI risks. Or overrate vegan food. Or underrate the case for moral realism. Or overrate Palestine’s claims. Or underrate Kendrick Lamar. (These are all real discussions I’ve had).
Much of the time, if a writer thinks their readers are underrating X, they’ll make an argument in favor of X. (Sounds obvious, I know).
But X and Y are usually not precise things that you can measure, never mind ascertain a specific value to it.
So if a writer argues for X, usually they don’t have a good sense of what value the reader assigns X (in part because of a lack of good statistics, and in part because a specific reader is a specific person with their own idiosyncratic views). Nor does a writer have a precise sense of what the optimal value of X ought to be, just that it’s higher (or lower) than what others think.
This creates major problems for both communication and clarity of thought!
One solution of course is to be an extremist. But this is a bad solution unless you actually think maximal (or minimal) X is good.
Sometimes either the structure of reality, or the structure of our disagreements, create natural mid-points while we can explicate their disagreements. For example, in my debate with BB, a natural midpoint is (we believe[1]) whether bees have net positive or net negative welfare. “0” is a natural midpoint. In my second post on the “rising premium of life”, I can naturally contrast my preferred hypothesis (premium of life rising) against the null hypothesis that the premium of life is mostly unchanged, or against the alternate hypothesis that it’s falling.
But reality often doesn’t give us such shortcuts! What are natural midpoints to argue for in terms of appropriate levels of credentialism? Or appropriate faith in human rationality? Or how much we should like Kendrick Lamar?
I don’t want to give people the illusion of an answer here, just presenting the problem as-is.
In both programming and mathematics, there’s a sense that only 3 numbers need no justification (0,1, infinity). Everything else is messier.
Unfortunately something similar is true for arguments as well. This creates a problem.
Much of the time, you want to argue that people underrate X (or overrate X). Or that people should be more Y (or less Y).
For example, people might underrate human rationality. Or overrate credentials. Or underrate near-term AI risks. Or overrate vegan food. Or underrate the case for moral realism. Or overrate Palestine’s claims. Or underrate Kendrick Lamar. (These are all real discussions I’ve had).
Much of the time, if a writer thinks their readers are underrating X, they’ll make an argument in favor of X. (Sounds obvious, I know).
But X and Y are usually not precise things that you can measure, never mind ascertain a specific value to it.
So if a writer argues for X, usually they don’t have a good sense of what value the reader assigns X (in part because of a lack of good statistics, and in part because a specific reader is a specific person with their own idiosyncratic views). Nor does a writer have a precise sense of what the optimal value of X ought to be, just that it’s higher (or lower) than what others think.
This creates major problems for both communication and clarity of thought!
One solution of course is to be an extremist. But this is a bad solution unless you actually think maximal (or minimal) X is good.
Sometimes either the structure of reality, or the structure of our disagreements, create natural mid-points while we can explicate their disagreements. For example, in my debate with BB, a natural midpoint is (we believe[1]) whether bees have net positive or net negative welfare. “0” is a natural midpoint. In my second post on the “rising premium of life”, I can naturally contrast my preferred hypothesis (premium of life rising) against the null hypothesis that the premium of life is mostly unchanged, or against the alternate hypothesis that it’s falling.
But reality often doesn’t give us such shortcuts! What are natural midpoints to argue for in terms of appropriate levels of credentialism? Or appropriate faith in human rationality? Or how much we should like Kendrick Lamar?
I don’t want to give people the illusion of an answer here, just presenting the problem as-is.
[1] This is disputed, see here.