DisruptiveInnovation.Org - Latest Discussions

Feynman on the Scientific Method

David Sundahl, 10/6/2016


One of the things that hamstrings the social sciences is that they are stuck on the problem of induction--first and best characterized by David Hume. Hume points out that just because we have all our lives seen sequences of events of the form A then B--the sun rises in the east and travels westward and sets--doesn't show us causality. We do not observe, we cannot observe causality. We see regularity but causation, ug it's merely a heuristic we employ or concept we impose on the observed regularities. The problem is that the work of smart statistitcians, philosophers and mathematicians was constantly working on cleverer and cleverer ways to analyze data (AKA "observations"). In the absence of high-quality supplementary experimental work, correlation could never lead to causation. (In a future entry, I'll explain why Simpson's Paradox demonstrates this to be so.) But for now what matters is that the social sciences tend to start with analysis.

My treat for you today is a link to Richard Fyenman on the scientific method. NB: Scientific education--no matter what discipline should be about having high-quality ignorance, asking high-quality questions.

Feynman on the Scientific Method