PHI403 Lecture 10 — Is it the Business of Science to Construct Theories?

Addresses the relationship between data and theory. Radical naturalism (Hume, Bacon's Novum Organum) holds that science should stay within empirical observation; theory construction is metaphysical overreach.

The lecture argues this position is untenable:

Underdetermination of theory by data — any data set admits of multiple theories that fit and explain it equally well.
Theory-ladenness of observation (Norwood Russell Hanson, Patterns of Discovery) — "there is more to seeing than meets the eyeball"; an expert and a layperson looking into a microscope do not see the same thing. Seeing is an activity.
Data are made, not given (Sabina Leonelli, Data-Centric Biology) — data are products of complex interactions between researchers and the world; their evidential value depends on context.

Data alone cannot supply explanation, prediction, or policy recommendation. We need a causal theory to move from observed correlation to actionable claim. Without a theory, data are inert — applicable only to the particular recorded cases.

Connections

Back-link: sources/2023-anjum-rocca-phi403-causation-in-science. Concepts: concepts/philosophical-bias · concepts/causation · concepts/mechanisms-in-science.