PESSIMISTIC META-INDUCTION AND THE EXPONENTIAL GROWTH OF SCIENCE

Abstract

In my talk, I aim to defend scientific realism against the pessimistic meta-induction (PI, for short). Scientific realism, as I define it, endorses the success-to-truth principle, i.e., the principle that if a scientific theory is successful, then it is (approximately) true. The PI, then, consists in pointing out that the history of science is full of theories that were once successful for a while, but later shown to be false. These theories constitute counterexamples to the success-to-truth principle, and seem to refute it. To rebut the PI, I start from the observation that the notion of success is graded, that the degree of success of a theory increases, when the cases of fit between its predictions and observations grow in number, diversity and precision. The main thesis of my talk is that among theories with very high degrees of success (e.g., our current most successful theories) almost no refutations have occurred, and that practically all successful refuted theories enjoyed rather low degrees of success. I support this thesis with two observations from the history of science. First, the degree of success of the most successful theories has by and large grown exponentially, so that the greatest growth of success occurred in the last few decades. I support this claim by considering various indicators of success, such as amount of data, computing power, scientific manpower, etc. Second, in the recent past practically no theory changes occurred among our most successful theories.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 21 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.