Today's post, My Kind of Reflection was originally published on 10 July 2008. A summary (taken from the LW wiki):

A few key differences between Eliezer Yudkowsky's ideas on reflection and the ideas of other philosophers.

Discuss the post here (rather than in the comments to the original post).

This post is part of the Rerunning the Sequences series, where we'll be going through Eliezer Yudkowsky's old posts in order so that people who are interested can (re-)read and discuss them. The previous post was The Fear of Common Knowledge, and you can use the sequence_reruns tag or rss feed to follow the rest of the series.

Sequence reruns are a community-driven effort. You can participate by re-reading the sequence post, discussing it here, posting the next day's sequence reruns post, or summarizing forthcoming articles on the wiki. Go here for more details, or to have meta discussions about the Rerunning the Sequences series.

New Comment
3 comments, sorted by Click to highlight new comments since: Today at 10:03 AM

if you start claiming that the universe isn't actually regular, that the answer to "Does induction work?" is "No!", then you're wandering into 2 + 2 = 3 territory.

The regular universe existed before any (known) reasoning being was around, and there wasn't any induction going on. That's not 2 + 2 = 3 territory. The regular universe exists now with some reasoning beings around, and induction still has its problems. No amount of confirmation will tell you that an inductive theory is regularly true, but one falsification will tell you it is regularly false. That's also not 2 + 2 = 3 territory.

Induction works great much of the time, but falsification works a wee bit better and a tiny few more times.

The original point was that the fact that induction works as well as it does would be extremely confusing if there wasn't regularity in the universe. Despite the inherent incompleteness of induction, claiming that induction isn't practically useful is essentially equivalent to asserting the deduction is not practically useful.

I also don't understand your comparison between falsification and induction. Falsifiability is a useful label for dividing pseudo-scientific predictions from scientific predictions, not a way of discovering new truth. Essentially all scientific predictions are inductive.

I've always considered induction to be the most important part of reasoning. Without induction, it is impossible to actually have a premise in way except the most hypothetical. While I enjoy coming up with unsupported premises and seeing where they go deductively, they are not at all useful if you can't figure out at some point what is actually the case. An argument is only sound if it is valid and the premises are true. Induction allows you to actually stop at some point and say which premises are true.

Occam's razor is an example of a good inductive tool. It isn't strictly correct, but it works pretty well. I agree strongly with the idea that you use the best of your current capability to determine how to improve. You just have to hope you are starting at a good enough position that it ends up going in the right direction.