Louis Vervoort

We review publications in automated scientific discovery using deep learning, with the aim of shedding light on problems with strong connections to philosophy of science, of physics in particular. We show that core issues of philosophy of science, related, notably, to the nature of scientific theories; the nature of unification; and of causation loom large in scientific deep learning. Therefore, advances in deep learning could, and ideally should, have impact on philosophy of science, and vice versa. We suggest lines of further research, and highlight the role ‘theory-based’ AI could have in future developments of the field.

The aim of this article is to present a variant of epistemic relativism that is compatible with a language practice especially popular among scientists. We argue that in science, but also in philosophy, propositions are naturally ‘relativized’ to sets of hypotheses or theories, and that a similar language practice allows one to interpret canonical problems of epistemology. We apply the model to Gettier’s problem, and derive a condition under which counterexamples à la Gettier to Plato’s account of knowledge do not arise. We argue that these findings give further content to a well-known result by Zagzebski (1994). Our interpretation points to a type of epistemic relativism having links with contextualism in epistemology, and perspectivism in philosophy of science.

RÉSUMÉNous proposons une théorie compatibiliste du libre arbitre, dans la tradition de la philosophie naturalisée, qui tente : 1) de fournir une synthèse de théories bien connues, capable de résoudre certains problèmes de ces dernières; 2) de tenir compte du fait que le libre arbitre a des degrés; 3) d’établir des liens avec la neurobiologie. Nous arguons que le libre arbitre d'un agent varie par degrés en fonction de la capacité de l'agent à faire des hypothèses et à utiliser des théories. Notre modèle met en évidence que des actions «libres» sont consciemment surveillées par l'agent, à travers des croyances, des hypothèses et finalement des théories — d'où l'expression «modèle CMT» (pour Conscious-through-Monitoring-through-Theories).

Here we propose a compatibilist theory of free will, in the tradition of naturalized philosophy, that attempts: 1) to provide a synthesis of a variety of well-known theories, capable of addressing problems of the latter; 2) to account for the fact that free will comes in degrees; 3) to interface with natural sciences, especially neurobiology. We argue that free will comes in degrees, as suggested by neuroscience. We suggest that a concept that can precisely ‘measure’ the variability of free will is that of ‘theory’, or rather, the capacity to make assumptions and use theories. Our model, then, highlights that free-willed actions are consciously monitored by the agent, through beliefs, assumptions and ultimately theories – hence the CMT model (for Conscious-through-Monitoring-through-Theories). The ultimate goal of this attempt at synthesis is to make the comparison with a variety of well-known accounts, and to show that our model can address several of their problems.

We describe here a series of experimental analogies between fluid mechanics and quantum mechanics recently discovered by a team of physicists. These analogies arise in droplet systems guided by a surface (or pilot) wave. We argue that these experimental facts put ancient theoretical work by Madelung on the analogy between fluid and quantum mechanics into new light. After re-deriving Madelung’s result starting from two basic fluid-mechanical equations (the Navier-Stokes equation and the continuity equation), we discuss the relation with the de Broglie-Bohm theory. This allows to make a direct link with the droplet experiments. It is argued that the fluid-mechanical interpretation of quantum mechanics, if it can be extended to the general N-particle case, would have an advantage over the Bohm interpretation: it could rid Bohm’s theory of its strongly non-local character.

Bell’s theorem admits several interpretations or ‘solutions’, the standard interpretation being ‘indeterminism’, a next one ‘nonlocality’. In this article two further solutions are investigated, termed here ‘superdeterminism’ and ‘supercorrelation’. The former is especially interesting for philosophical reasons, if only because it is always rejected on the basis of extra-physical arguments. The latter, supercorrelation, will be studied here by investigating model systems that can mimic it, namely spin lattices. It is shown that in these systems the Bell inequality can be violated, even if they are local according to usual definitions. Violation of the Bell inequality is retraced to violation of ‘measurement independence’. These results emphasize the importance of studying the premises of the Bell inequality in realistic systems

The principle of causality or of determinism, and the notion of cause, are studied in the light of recent results in quantum mechanics. A definition of the concept of cause, loosely related to Lewis' counterfactual approach, is proposed. Then the question 'has every (physical) event a cause ?' is investigated. According to the orthodox quantum theory the answer to above question is negative. However, it is argued that there exist at least as many arguments in favor of a 'yes' (and thus of determinism) as of 'no' – arguments that are fully compatible with quantum theory. Therefore the above question remains philosophical, and cannot be settled by scientific arguments, i.e. experiments or calculations, as many researchers believe.

In this text the ancient philosophical question of determinism (“Does every event have a cause ?”) will be re-examined. In the philosophy of science and physics communities the orthodox position states that the physical world is indeterministic: quantum events would have no causes but happen by irreducible chance. Arguably the clearest theorem that leads to this conclusion is Bell’s theorem. The commonly accepted ‘solution’ to the theorem is ‘indeterminism’, in agreement with the Copenhagen interpretation. Here it is recalled that indeterminism is not really a physical but rather a philosophical hypothesis, and that it has counterintuitive and far-reaching implications. At the same time another solution to Bell’s theorem exists, often termed ‘superdeterminism’ or ‘total determinism’. Superdeterminism appears to be a philosophical position that is centuries and probably millennia old: it is for instance Spinoza’s determinism. If Bell’s theorem has both indeterministic and deterministic solutions, choosing between determinism and indeterminism is a philosophical question, not a matter of physical experimentation, as is widely believed. If it is impossible to use physics for deciding between both positions, it is legitimate to ask which philosophical theories are of help. Here it is argued that probability theory – more precisely the interpretation of probability – is instrumental for advancing the debate. It appears that the hypothesis of determinism allows to answer a series of precise questions from probability theory, while indeterminism remains silent for these questions. From this point of view determinism appears to be the more reasonable assumption, after all.

Recent experiments have shown that certain fluid-mechanical systems, namely oil droplets bouncing on oil films, can mimic a wide range of quantum phenomena, including double-slit interference, quantization of angular momentum and Zeeman splitting. Here I investigate what can be learned from these systems concerning no-go theorems as those of Bell and Kochen-Specker. In particular, a model for the Bell experiment is proposed that includes variables describing a ‘background’ field or medium. This field mimics the surface wave that accompanies the droplets in the fluid-mechanical experiments. It appears that quite generally such a model can violate the Bell inequality and reproduce the quantum statistics, even if it is based on local dynamics only. The reason is that measurement independence is not valid in such models. This opens the door for local ‘background-based’ theories, describing the interaction of particles and analyzers with a background field, to complete quantum mechanics. Experiments to test these ideas are also proposed