Use Judgement Not Proofs

Context: You are either studying or using a human discipline that involves systems whose behavior depends on information flow. Most of the knowledge in this discipline is either unteachable (art), teachable but not writable (craft), or writable but not proveable (soft-science).

Problem: How do you objectively work in this discipline?

Forces: Since most of the knowledge in the discipline is unprovable, it can be called opinion. At first, opinions might not seem objective. But there are three quantifiable things about opinions: 1) do you believe them, 2) how many others believe them (usually expressed in percentage), and 3) how much do they overlap other opinions.

Solution: Collect and analyze opinions. Use your judgement to decide whether or not you believe in an opinon. Observe if others believe in it and use it. A rule of thumb is that if you can say something better, you don't believe in it; if you can't say it better, you believe in it. Continually look for better ideas, and update your opinion list often, again using your judgement.

I find rejection of the scientific process a bit frightful. It risks all kinds of harmful GroupThink. If we fail to try to articulate our opinions, then places like wiki become nothing more than a vote tallying system. Just because the scientific process is not easy in our discipline is not a reason to give up. I invite everyone to try harder to try to explain and justify their favorite approaches rather than settle for AdVerecundiam as "good enough".

In safety critical projects..... Judge first, prove second, and then get some third party opinions on your so-called proofs. See ImreLakatos's thesis.

CategoryProof