Scientific Representation: Paradoxes of Perspective (review)

For three decades, Bas van Fraassen’s constructive empiricism has been the most influential anti-realist position in philosophy of science. When logical empiricism lost its place as the leading philosophy of science, support for the anti-realism associated with it also waned. Van Fraassen’s brand of empiricism, although significantly different from that of the logical empiricists, has kept alive the anti-realist spirit of the earlier enterprise. In Scientific Representation: Paradoxes of Perspective, the author develops and defends a position he calls ‘empiricist structuralism.’ He agrees with the slogan that ‘all we know in science is structure.’ His challenge is to explain how this slogan can be meaningfully interpreted in the absence of commitment to any corresponding real structure in the world. Because my disagreements with Scientific Representation are fairly deep and because this is not an easy book, I shall begin with an account of its contents and leave the assessment to a subsequent section.

I Summary

Some Background: Many readers will know about van Fraassen’s constructive empiricism chiefly through The Scientific Image, published in 1980. There he argued that the aim of science is empirical adequacy rather than truth:

Science aims to give us theories which are empirically adequate; and acceptance of a theory involves as belief only that it is empirically adequate. This is the statement of the anti-realist position I advocate; I shall call it constructive empiricism.

(1980, 10)

He went on to say that ‘a theory is empirically adequate if what it says about the observable things and events in the world is true — exactly if it “saves the phenomena”’ (1980, 12). The notion of ‘observability’ employed here is narrow insofar as it is tied to our sensory apparatus; it should be understood as perceptibility. Detectability with the aid of instruments is not observability; so accepting a theory as empirically adequate does not require us to believe in any of the theoretical entities it postulates, even those that seem to make themselves known to us through instruments. On the other hand, ‘observability’ is broad enough to cover dinosaurs and distant stars that no one has or ever will observe — because if members of our epistemic community were properly situated, those members would be able to perceive such things without augmenting their sensory apparatus.

Somewhat more technically, constructive empiricism has been developed within the context of a semantic view of theories so that:

To present a theory is to specify a family of structures, its models; and secondly, to specify certain parts of those models (the empirical substructures) as candidates for the direct representation of observable phenomena.

(1980, 64)

Given that empirical adequacy is the sole aim and criterion of science, theory acceptance does not require belief that the theory is true. The only belief required by acceptance of a scientific theory is belief that its observational consequences are true. Acceptance also requires that we commit to using the theory, but that is a pragmatic rather than an epistemic consideration. To the predictable complaint that his view strips science of its explanatory power, van Fraassen has consistently replied that explanation is no part of the aim of science. Explanation serves pragmatic rather than epistemic ends, and the search for explanation tends to produce metaphysics rather than science.

If you personally want to believe that the theories you accept are true and that their unobservable components refer to real things in the world, you are free to do so. Constructive empiricism is permissive in that respect. It does not advocate agnosticism about theoretical entities or about the truth of theories (although it has often been understood that way). Constructive empiricism simply says that once you decide to become a believer, you have gone beyond science into the realm of interpretation or metaphysics.

Van Fraassen has modified some of his views since 1980. Two of these changes bear mentioning here. One concerns his understanding of what it is to be an empiricist. He now describes empiricism as a stance rather than...