Introduction

Three main senses of the term “robustness” are often distinguished: (1) Robustness of models; (2) Robustness as stability or insensitivity of output as against variations in parameter values; (3) Robustness as consilience of results from different and independent hypotheses, procedures or sources of evidence.

M. Buzzoni (is)

University of Macerata, Macerata, Italy e-mail: marco.buzzoni@unimc.it © Springer International Publishing AG 2017

E. Agazzi (ed.), Varieties of Scientific Realism,

DOI 10.1007/978-3-319-51608-0_7

In Sect. 2, I shall characterize the two types of robustness that are most important for understanding experimental sciences and for the issue of scientific realism. Section 3 examines the important step towards answering both objections that was taken by Ian Hacking, who implicitly referred to robustness in his criticism of the no miracle argument. As we shall briefly see, Hacking only went half way, and finally fell into a more traditional view of robustness, which cannot avoid sceptical con­clusions, akin in some respects to that of Kuhn's incommensurability thesis.

This will lead, in Sect. 4 to think in terms of the two kinds of robustness working together in such a way that we are able to cope with some difficulties concerning robustness-as-consilience that have indirect consequences for the scientific realism debate. According to one of the central points emphasized throughout this paper, the two senses of the term “robustness” have to be distinguished in order to avoid confusion, but, on the other hand, they are intimately connected with one another. Robustness-as-stability and robustness-as-consilience are two specifications of the same experimental and intersubjective reproducibility, which, in this sense, may be plausibly presented as the substantial common core of robustness in the empirical sciences.

As we shall see in Sect. 5 there is no vicious circle here, but a technical-practical synergy (or, if you prefer, a robustness of higher order)—which is at the heart of the experimental method and which can help us out of the two main problems for robustness-as-consilience. If the different notions of robustness-as-stability and robustness-as-consilience are not only distinguished, but also connected with one another, and if their intimate relationship is recognised, we find a notion of robustness of higher order. As we shall see, on the basis of this notion and by drawing on Hacking's hint to improve the no miracle argument, the spiral synergy of robustness-as-stability and robustness-as-consilience provides some support for sci­entific realism, at least in the minimal senses that (1) it furnishes an epistemic warrant for a reality which is independent of us, and (2) it puts a great number of reliable findings at our disposal, which we are not willing, at least provisionally, to put in question. Thus, from a technical-operational point of view, we may express differently one of the best known pragmatic theories of truth, by saying that the truth is something like ‘what shows itself most robust' in the widest context and the longest run.

2