Full Idea
The inferential path to unobservables is often the same as to unobserved observables. In these two sorts of case, the reason for belief can be equally strong, so the suggestion that we infer truth in one case but not the other seems perverse.
Gist of Idea
The inference to observables and unobservables is almost the same, so why distinguish them?
Source
Peter Lipton (Inference to the Best Explanation (2nd) [2004], 09 'Voltaire's')
Book Reference
Lipton,Peter: 'Inference to the Best Explanation (2nd ed)' [Routledge 2004], p.146
A Reaction
[Van Fraassen 1980 is the target of this] Van F seems to be in the grip of some sort of verificationism, which I always disliked on the grounds that speculation can be highly meaningful. Why embrace something because it 'could' be observed?