Podcast: Play in new window | Download (Duration: 52:18 — 47.9MB)
On "Falsification and the Methodology of Scientific Research Programmes" (1970), featuring Mark, Wes, Dylan, and Seth. In what way is scientific progress rational?
To understand the state of the debate by Lakatos' time, let's run quickly through some history of the philosophy of science:
Early scientists like Francis Bacon saw science as providing certainty in the progressive accumulation of truths. In the 1930s, Karl Popper argued that while experiments cannot prove a scientific theory to be true, they can and do falsify theories, again with certainty: Science proceeds through a series of creative conjectures that are then refuted by experiments, leading scientists to further conjectures that are more accurate, i.e. not refuted by the experiments undertaken to that point. Thomas Kuhn argued in the 1960s that this was nonsense: When a supposedly falsifying experiment is performed against a theory, the theory is not typically given up; adherents of the theory instead just ignore the experimental results as an anomaly. Science on Kuhn's view is a series of paradigms (world-views that include ways of doing science including for determining what counts as admissible results), and why one paradigm goes out of fashion to be replaced by another is a matter for sociological study with no ultimate rationality behind it.
Lakatos' view was somewhere between Popper's and Kuhn's: He accepted Kuhn's charge that scientists have not historically given up their theories as soon as a refuting experiment comes along, but he nonetheless claimed like Popper than scientific progress is rational, and so we can study it by studying the logic of justification and not just the sociology of intellectual fashion. Instead of talking about paradigms, Lakatos uses the term "research program." Each scientific research program has a firm set of core principles that are resistant to change by experiment, but also includes numerous auxiliary principles that can be changed. An extensively considered example in the text is Newtonian mechanics, which even when it was developed could not explain the motion of the moon. When some particular orbit seemed to defy Newton's laws, people did not give up those laws, but would instead posit that there must be some other body out there distorting that planet's motion.
The rationality of a scientific research program is determined by how "progressive" it is, which according to Lakatos is its ability to continually make new and unexpected predictions. If a program makes a false prediction, well, that's not enough reason to just give it up, but enough false predictions necessitating ad hoc modifications of auxiliary principles mean that the program is "regressive." A scientist working within that program could still rationally keep working on it to try to make it more progressive, but at least in historical hindsight, we can tell a rational story about how Newtonian mechanics explained the world better than Aristotelianism, and how Einstein improved on Newton, etc.
Buy the book that contains this essay (and we also read the intro from 1973 from this collection of papers) or try this online version.
Audio editing by Tyler Hislop.
Leave a Reply