thomblake comments on The noncentral fallacy - the worst argument in the world? - Less Wrong

157 Post author: Yvain 27 August 2012 03:36AM

You are viewing a comment permalink. View the original post to see all comments and the full post content.

Comments (1742)

You are viewing a single comment's thread. Show more comments above.

Comment author: thomblake 30 August 2012 08:06:20PM *  3 points [-]

Usually when we say 'evidence' we mean 'Bayesian evidence'. If you examine arbitrary triangles and they all happen to have one side whose length squared is equal to the sum of the squares of the lengths of the other two sides, then being a triangle is evidence that the shape has this property. It was still evidence even if it turns out the triangle didn't have the property.

Comment author: Dan_Moore 30 August 2012 09:14:07PM 0 points [-]

If you examine arbitrary triangles and they all happen to have one side whose length squared is equal to the sum of the squares of the lengths of the other two sides, then being a triangle is evidence that the shape has this property.

Agreed. But in this example, it's known that the new triangle being considered is different from those previously examined because it's not right. Therefore, the presumption that a sampling of the previously examined triangles is arbitrary, with respect to the larger class that includes the new triangle, is not a rational presumption.

Comment author: thomblake 31 August 2012 01:29:19PM 0 points [-]

But in this example, it's known that the new triangle being considered is different from those previously examined because it's not right. Therefore, the presumption that a sampling of the previously examined triangles is arbitrary, with respect to the larger class that includes the new triangle, is not a rational presumption.

I was assuming that the folks doing the observing did not necessarily realize that all the previous triangles were right and this one is not.

Also, your line of reasoning works equally well if all the triangles you've seen so far were written on paper, and this one (also a right triangle) is scratched in the dirt. But in that case, it would be good evidence. So clearly it's evidence in either case.

Comment author: Dan_Moore 31 August 2012 06:43:56PM 1 point [-]

I was assuming that the folks doing the observing did not necessarily realize that all the previous triangles were right and this one is not.

In this case, the folks doing the observing do realize that this triangle is different from all those previously considered, but they downplay the significance of this fact, perhaps using the justification: a triangle is a triangle is a triangle.

(I'm actually not making this up as I go along. I had worked out this example some time ago to illustrate what I believe to be a widely-held false belief in finance. I believe that WAW is a good description of the thought process behind this belief.)