(anti) realism

central question

should we believe that our best scientific theories give approximately true descriptions of unobservable reality (electrons, curved spacetime, dark energy) or should we remain agnostic about unobservables and treat theories merely as instruments for predicting observable phenomena? 

main position for

scientific realism is the general position that we should believe they give approximately true descriptions of unobservable reality…No miracles arg says that our theories have extraordinary predictive success and the best explanation for this success so is that they are at the very least approximately true descriptions of the world. Examples are like Einstein’s light bending that was speculative at the time. Preservative and structuralist are the two versions - preservative says we should believe that the success-generating theoretical constituents are at least approximately true, and structuralism says the same for the mathematical structures preserved through theory change

main argument for

No miracles arg. Of all the possible explanations of science’s success in instances such as Einstein’s prediction of light bending near large masses, the best explanation is approximate truth. Einstein’s general theory of relativity predicted a behaviour that was novel and bold conjecture, but it was proved correct by Eddington’s 1919 experiment. We ought to infer to best explanation because it is a common form of reasoning that is constantly used in science (and everyday life). When a doctor infers a patient has a particular disease from their symptoms, the disease is unobservable but it is still the best explanation. The alternative is that we accept ‘worse explanations’ of science’s success, and this would not promote genuine advances in science. IBE may be circular but this is not necessarily vicious…all epistemic justification has to start somewhere and we cannot justify all reasoning from outside of reasoning

major objection

Chang’s historical critique of caloric theory is a decisive objection. It goes against the genuine success response which is the strongest reply to the pessimistic meta-induction. Caloric theory meets the genuine success criterion precisely as Laplace’s correction of Newton’s speed of sound was novel, mature and dramatically accurate. But the assumptions generating this success were specifically caloric ones (caloric as a self-repulsive fluid was the explanatory mechanism, not idle metaphysics). These assumptions have been rejected by modern science completely, so preservative’s divide and conquer strategy cannot be said to identify with constituents to retain when the false ontological ones are the ones generating success

what do you think?

preservative realism fails to show that the divide and conquer strategy is historically sustainable. Structural realism is the most defensible remaining position which retreats to mathematical structures having approximate success rather than specific theoretical constituents. It can survive Chang’s argument because the claim that mathematical structures survive through scientific theories even when ontological interpretation is abandoned has historical sustainability (maxwell’s equations describing electromagnetic radiation were preserved from classical electrodynamics to quantum electrodynamics). In the case of caloric, some relationships like calorimetric laws were preserved in kinetic theory (though not all..still a point of contention). But it still faces the unresolved challenge that preservation of mathematical structure might reflect human cognitive bias rather than mind-independent reality…so the view is chastened to a position that admits the debate is still open and that the no miracles argument fails in its simple form, but still resists full anti-realism can be resisted because van Fraasen’s arg of constructivism empiricism leaves the novel prediction problem unresolved

realism and anti realism defs

realism - our best scientific descriptions of unobservables are at least approximately true

anti realism - we have no reason to believe the unobservables of our best sciences are real and we should be agnostic about the matter until they prove observable…for practical purposes we proceed as if they do not exist

preservative and structural

pres - parts of our best scientific theories produce a great deal of success even when they describe unobservable aspects of the world

struc - the structural parts of our best scientific theories enjoy a great deal of success, even when they describe unobservable aspects of the world

no miracles arg

P1 - our best scientific theories enjoy a great deal of success even when they describe unobservable features of the world

P2 - the best explanation of such extreme success is that the description of science are at least approximately true

P3 - we ought to infer to best explanation

Conc - Therefore our best scientific descriptions are at least approximately true

van fraassen’s constructive empiricism

Van Fraassen argues we should not believe scientific theories are approximately true — we should only believe they are empirically adequate, meaning they correctly predict all observable phenomena. Science aims to save the phenomena not to describe unobservable reality. We should be agnostic about unobservables — electrons, curved spacetime, dark energy — not because they definitely don't exist but because we have no good reason to believe they do. A theory can be empirically adequate and useful without its unobservable claims being true

underdetermination argument

Premise 1 — The no miracles argument deploys IBE and requires a unique best explanation to be warranted.

Premise 2 — IBE leads to contradiction if two contradictory theories are equally good explanations.

Premise 3 — For any successful theory T and any unobservable proposition P, T+P and T+¬P are both equally successful since they include all the success of T yet are mutually incompatible. The initial observational equivalence of Copernican and Ptolemaic systems illustrates genuine scientific underdetermination.

Premise 4 — If multiple incompatible theories are equally successful IBE cannot identify a unique best explanation.

Conclusion — Therefore IBE cannot establish that any particular theory's unobservable claims are approximately true.

pessimistic meta induction

P1 - success implies truth;

p2 - there are past scientific theories as well as current ones that were successful;

p3 - the successful past and current theories contract one another and are therefore incompatible;

p4 - since success implies truth, both past and current theories must be true but since these theories are incompatible this implies a contradiction;

conc - therefore success does not imply truth since to assume otherwise implies a contradiction

genuine success response

The realist responds that past theories were not genuinely successful in the relevant sense. Genuine success requires novel predictions within a mature theoretical framework — the kind exemplified by Eddington's confirmation of light bending. Crystalline spheres didn't make novel predictions. This criterion defends both preservative realism and structural realism against the PMI by denying that the historical cases Laudan invokes meet the relevant standard of success.

chang’s counter objection

Premise 1 — Preservative realism requires that the success-generating parts of past theories are those preserved in later theories — the divide and conquer strategy identifies theoretical constituents responsible for success and retains those while discarding idle metaphysical parts.

Premise 2 — The caloric theory was genuinely successful in the relevant sense. Laplace used caloric theory to make a dramatically correct prediction of the speed of sound. This is mature novel predictively successful science — exactly what the genuine success criterion requires.

Premise 3 — The assumptions responsible for these successes were the specifically caloric ones — not idle metaphysics but the explanatory core of the derivations. Laplace's correction of Newton's speed of sound required assuming caloric existed as a self-repulsive point-particle fluid. Psillos argues calorimetry show preservation without caloric commitments — but Chang shows calorimetry predates the caloric theory and was never part of it.

Premise 4 — These specifically caloric assumptions have been completely rejected by modern science and are not preserved in kinetic theory. The Laplacian metaphysics of mutually repelling caloric particles has been unhesitatingly rejected.

Conclusion — Therefore the caloric case does not support preservative realism — the success-generating assumptions were precisely the ones later rejected. The divide and conquer strategy cannot identify which constituents to retain when the false ontological ones are the success-generating ones.

structuralist response to chang

Structural realism is the most defensible remaining position which retreats to mathematical structures having approximate success rather than specific theoretical constituents. It can survive Chang’s argument because the claim that mathematical structures survive through scientific theories even when ontological interpretation is abandoned has historical sustainability (maxwell’s equations describing electromagnetic radiation were preserved from classical electrodynamics to quantum electrodynamics). In the case of caloric, some relationships like calorimetric laws were preserved in kinetic theory (though not all..still a point of contention). But it still faces the unresolved challenge that preservation of mathematical structure might reflect human cognitive bias rather than mind-independent reality…so the view is chastened to a position that admits the debate is still open and that the no miracles argument fails in its simple form, but still resists full anti-realism can be resisted because van Fraasen’s arg of constructivism empiricism leaves the novel prediction problem unresolved