Robustness and Hacking’s Scientific Realism

An important step towards answering both objections was taken by the experi­mentalist turn, and notably by Ian Hacking, who implicitly referred to robustness in his criticism of the no miracle argument.

On the one hand, Hacking claims that, in its most common form, the “cosmic accident argument” (which is another name for the no miracle or, to use Smart's expression, the “cosmic coincidence” argument: Smart 1963: 39) is circular:

The cosmic accident argument notes that often in the growth of knowledge a good theory will explain diverse phenomena which had not hitherto been thought of as connected. [...] Once again, this seems to me to beg that realist/anti-realist issue. The anti-realist agrees that the account, due to Einstein and others, of the mean free path of molecules is a triumph. It is empirically adequate—wonderfully so. The realist asks why is it empirically adequate—is that not because there just are molecules? The anti-realist retorts that explanation is no hallmark of truth, and that all our evidence points only to empirical adequacy. In short the argument goes around in circles (as, I contend, do all arguments conducted at this level of discussion of theories). (Hacking 1983: 54-55)

Hacking does not develop this objection as fully as could be wished. But his point is that scientific realism may be defended using experimental practice instead of traditional representational approaches:

Hacking's “Grid Argument” is perhaps the best interpretation of the no miracle argument (or, to use Hacking's expression, of the “cosmic accident argument”) in terms of his ‘interventionist' point of view.

see exactly the same shapes and letters as were originally drawn on a large scale. It is impossible seriously to entertain the thought that the minute disc, which I am holding by a pair of tweezers, does not in fact have the structure of a labelled grid. I know that what I see through the microscope is veridical because we made the grid to be just that way. [...] Moreover we can check the results with any kind of microscope, using any of a dozen unrelated physical processes to produce an image. [...] Is it a gigantic conspiracy of totally unrelated physical processes that the large scale grid was shrunk into some non-grid which when viewed using different kinds of micro-scopes still looks like a grid? (Hacking 1983, 201).

The “Grid Argument” supports realism about entities by insisting on the robustness of something manipulated, and not just ‘observed’, using different and independent techniques. In my opinion, Hacking’s main point is that intersubjective reproducibility, considered from the point of view of a pragmatist conception of knowledge and in relation to our interactions with the surrounding world, may be understood as the most valuable clue to the independent existence of all empirical contents, including theoretical entities. If Hacking is interpreted in this way, I think he is right in advocating a technical-operational ‘criterion’ for the reality of theo­retical entities: true ideas ‘agree’ with reality, in the sense and to the extent that they make it possible to master it.^[51]

However, Hacking did not coherently apply this idea to both senses of robust­ness. Owing to an excessive and one-sided emphasis on the independence of experiment from theory,^[52] in 1983 he is led to attach more importance to robustness-as-stability than to robustness-as-consilience, where robustness- as-stability is understood as consisting primarily in the stability of particular or ‘local’ technical-operational interventions on reality.^[53] His later work, on the con­trary, seems to attach very much more importance to robustness-as-consilience than to robustness-as-stability: the cause of the “stability” in science is found by him in a particular process of reciprocal adaptation between “thoughts, acts and manufac­tures” leading to a body of “types of theory and types of apparatus and types of analysis that are mutually adjusted to each other” (Hacking 1992: 30).

It is clear that this perspective, rejecting the truth of very general theories and making that of “local” laboratory statements relative to a system of self-justifying theoretical presuppositions, leads Hacking to embrace, even though in a somewhat particular sense, the incommensurability thesis.

As a matter of fact, he claims quite explicitly that in science there can be stabilisations of theories that

would not even be comparable, because they would be true to different and quite literally incommensurable classes of phenomena and instrumentation. I say incommensurable in the straightforward sense that there would be no body of instruments to make common mea­surements, because the instruments are peculiar to each stable science (Hacking 1992: 31).

According to Hacking, Newton's and Einstein's theories are incommensurable not because the statements of one could not be expressed in the other, but because one is true “to one body of measurements given by one class of instruments, while the other is true to another” (Hacking 1992: 54).

All these statements could come from The Structure of Scientific Revolutions, including the last one, since Kuhn (even though Hacking seems unaware of this) was explicitly concerned with the problem of the incommensurability of instru­ments relative to different paradigms (cf. Kuhn [1962] 1970: 126).

To sum up these considerations concerning Hacking, it may be said that, on the one hand, he deserves credit for having made evident the epistemological impor­tance of experimenting for robustness. On the other hand, Hacking only goes half way. Instead of reaching a deeper comprehension of both notions of robustness, he reaches only a superficial schema concerning robustness as consilience (which is a merely analogical extension of the statistical meaning to the experimental sciences), to which, in the later work, he does accord a privileged epistemic status.^[54]

We have now to consider how to develop more coherently Hacking's idea, by applying it to both notions of robustness.

4