Summary
Highlights
Sabrina Imobisi introduces a video that complements a chapter on classical philosophy of science, logical positivism, and falsificationism. The video will cover the main questions of classical philosophy of science, responses from logical empiricism and Popper's falsificationism, and the classical approach to scientific explanation.
Classical philosophy of science addresses several core questions: the structure and testing of scientific theories, the criterion of demarcation (distinguishing scientific from non-scientific statements), the distinction between discovery and justification contexts, the justification of empirical theories, the nature of scientific progress, and scientific explanation (answering 'why' questions).
Logical empiricism proposes that scientific statements must be translatable into observational language. They differentiate between scientific statements (which have cognitive content and can be true or false) and non-scientific statements. Scientific statements are further divided into formal sciences (truth/falsity independent of the world) and empirical sciences (judged by their translatability to observational language).
Logical empiricism rejects the idea that hypotheses are discovered through induction, citing issues like the relevance of observations and the lack of theoretical terms in purely inductive hypotheses. They argue that hypothesis generation falls outside the scope of philosophy of science, which instead focuses on the logical analysis of justification.
For logical empiricism, empirical theories are justified by induction. Unfavorable results deductively refute a hypothesis, but favorable results only confirm it, making it probably true. More confirming instances increase probability. Scientific progress is seen as cumulative, with new theories preserving the 'true' content of older ones as science moves towards truth.
Carl Popper's falsificationism completely rejects the role of induction in the philosophy of science. He argues that inductive inferences cannot be logically justified without circularly appealing to induction itself, making them unsuitable for establishing scientific truth.
Popper proposes falsifiability as the criterion of demarcation: a statement is scientific if it can potentially be proven false. This requires the possibility of formulating basic empirical statements (falsifiers) that are logically incompatible with the scientific claim. This criterion excludes formal sciences and probabilistic generalizations.
Similar to logical positivism, falsificationism distinguishes between discovery and justification, focusing on the latter through logical reconstruction. For Popper, justification involves trying to refute hypotheses through testing. If a hypothesis withstands attempts at falsification, it is considered 'corroborated' (not proven true), as all scientific statements are provisional and fallible. The empirical basis itself is fallible, accepted by conventional consensus.
Popper argues that scientific theories cannot be conclusively proven true or even probably true. Instead, scientific progress occurs through conjectures and refutations, where false content is eliminated, and science incrementally approximates the truth.
A comparison highlights key differences: demarcation is via translatability to observational language for logical positivism and falsifiability for falsificationism. Justification is by induction (confirmation) for logical positivism and by deduction (corroboration through non-refutation) for falsificationism. Scientific progress is cumulative for logical positivism, preserving true content, while for falsificationism, it's about eliminating false content through refutation.
Both logical positivism and falsificationism share the covering-law model of scientific explanation. This model states that a phenomenon is explained by showing it falls under a scientific law, which is a general statement describing an empirically confirmed regularity. The explanation consists of an explanans (laws and initial conditions) from which the explanandum (the phenomenon to be explained) can be derived.
The covering-law model includes two types: nomological-deductive explanations, where the explanandum is deductively inferred from universal laws and initial conditions, and statistical-inductive explanations, where the explanandum is inductively inferred from statistical or probabilistic laws. Falsificationism only accepts nomological-deductive explanations as truly scientific, while logical positivism accepts both.