Supervisors info:
Ψύλλος Στάθης, Καθηγητής, Τμήμα Ιστορίας και Φιλοσοφίας της Επιστήμης, ΕΚΠΑ
Αραμπατζής Θεόδωρος, Καθηγητής, Τμήμα Ιστορίας και Φιλοσοφίας της Επιστήμης, ΕΚΠΑ
Στεφάνου Ιωάννης, Επίκουρος Καθηγητής, Τμήμα Ιστορίας και Φιλοσοφίας της Επιστήμης, ΕΚΠΑ
Summary:
All forms of contemporary scientific realism admit that we should not be realists about scientific theories taken as wholes, but that we have to restrict our epistemic commitment to specific theoretical parts – those parts that are “essential” for the empirical “success” of the corresponding theories. This is the basic idea behind the so-called “divide et impera strategy”, which is implemented by S. Psillos (1999) in his effort to support scientific realism against the infamous Pessimistic Meta-Induction. In the first part of this thesis, I offer a short review of Psillos’ approach and I examine the main lines of criticism against it, most prominently that of K. Stanford and T. Lyons. The main objective of this first part is to evaluate the virtues and the shortcomings of Psillos’ divide et impera strategy.
Afterwards, I focus on P. Vickers’ work, in which is found one of the most sophisticated attempts to elaborate Psillos’ approach on behalf of the “new generation realists”. Vickers puts forward a new “recipe” for separating the “essential” from the “idle” parts of scientific theories, which will be called “Criterion V”. At this point a multifaceted evaluation of Criterion V is in order. On the one hand, Criterion V seems absolutely plausible and epistemologically solid. It can provide a basis for answering to most known historical challenges to scientific realism, while it appears to solve a part of the problems of Psillos’ approach. On the other hand, Criterion V faces some problems, among which, one is considered as the most serious: it cannot account for the unificatory aspect of scientific theorizing.
Moreover, Vickers’ account has a consequence that he himself suspects, but that he doesn’t further explore. That is, it seems like the implementation of the criterion V forces us straight to the ground of Epistemic Structural Realism (ESR). Stating it by using Psillos’ characteristic terminology (2001), Vickers is doomed to cross the “downward path” to ESR. Thus, given that Vickers’ account is just a slightly elaborated version of Psillos’ divide et impera strategy, the aforementioned result brings to the forefront the suspicion that the divide et impera strategy (already in its formulation by Psillos) hides in its core “the seed of ESR”. At this point, a major aim of this thesis is to establish that if one wishes to be faithful to the divide et impera strategy, one will have to end up as a kind of structural realist. That is, I extensively argue that the divide et impera strategy is intrinsically linked to ESR and that selective scientific realism can only be structural.
Nevertheless, as is well known, ESR faces the accusation of being trivial. This accusation is raised due to the notorious Newman–problem according to which the main epistemological thesis of pure ESR, i.e. that we can have knowledge of the unobservable realm only “up to isomorphism”, is trivial in the sense that every abstract structure is a priori guaranteed to be instantiated provided that there are enough individual objects. In reference to the Ramseyan version of structuralism, that means that a Ramseyfied theory is trivially true if it is empirically adequate and the cardinality of the theory’s domain is appropriate. This ascertainment sets in doubt the realistic character of ESR and shows that ESR collapses to a form of scepticist empiricism. It is suggested that the only possible way to avoid the accusation of triviality is to adopt the “Restricted Quantifier Response” (“RQR”). That is, by restricting the domain of the second order quantifiers in the Ramseyfied theory (“TR”) to the class of natural properties, one can avoid the trivial satisfaction of TR, for it is not a priori guaranteed that the abstract structure generated by TR is isomorphic to a part of the “natural structure” of the world. Despite the fact that it is unanimously admitted that RQR successfully faces the critique of triviality, RQR is taken to betray ESR because of the admission of the intensionally interpreted, second order, non–logical and non–observational predicate ‘x is a natural property’. My aim in this section is to argue that the form of “slightly impure” ESR, which is the result of the employment of RQR, although strictly speaking “impure”, still maintains the crucial epistemological commitments of ESR. Thus, in the end, Newman–problem seems not even to be a problem, unless we are committed to a superfluous demand for a structuralist “ideological purity”.
The motif emerging from the above summary should be clear. It seems that a faithful-to-the-divide-et-impera-strategy realism has to be structural. Conversely, structural realism, in order to be “realism enough”, has to become “impure”. Given that, a question naturally arises: is there room for a genuine distinction between the supposedly robust scientific realism that Psillos and Vickers aim to serve and a slightly impure Ramseyan structuralism? The main objective of the last section of this thesis is to address this (mostly rhetorical) question.
Keywords:
Scientific Realism, Divide et Impera Strategy, Epistemic Structural Realism, Structuralism, Ramsey-Sentences, Newman-Problem, Theoretical Terms, Pessimistic Meta-Induction, Underdetermination, Theoretical Unification, Rudolf Carnap, Stathis Psillos, Peter Vickers, John Worrall, Ioannis Votsis, Kyle Stanford
