Peter Achinstein, ed., Scientific Evidence. Philosophical Theories and Applications, Johns Hopkins University Press, 2005

This volume compiles thirteen papers presented in a conference on scientific evidence held at Johns Hopkins in 2003. Seven are authored by former graduate students of the editor (many of whom are now renowned philosophers) and eight take issue with one or another aspect of his theory of evidence, so the compilation will certainly be of interest for those who are attracted by Peter Achinstein's approach. But it may be equally attractive in various degrees for those who are more generally curious about evidence.

The volume is divided in two parts, of which the first is devoted to philosophical theories of evidence and the second to present six related case studies in different scientific domains. The former are namely those of Achinstein, Mayo, Norton and Whewell, though different quotas are assigned to them. Whewell's concept of consilience is examined and partly vindicated by Laura Snyder, who supports a kind of version of scientific realism, if we interpret it as an inferential pattern that appears in the theoretical unification of various natural kinds. Norton presents his own material theory of induction in a brief but very informative comparison to other standard alternatives. Mayo exposes her severe test approach to evidence and defends it from the criticism received from Achinstein. He summarises, in turn, the view exposed in his The Book of Evidence (2001), refuting four contrary theses. Finally, still in the first part of the book, another three papers discuss various aspects of Achinstein's theory: S. Gimbel defends and illustrates the viability of a threshold concept of evidence within this framework; F. Kronz and A. McLaughin call for its expansion to incorporate some guidelines for the process of evidence-acquisition, drawing on Peirce's proto-economics of science; and S. Roush argues for the positive relevance condition that Achinstein criticized in the name of his own explanatory based view of evidence.

The quality of the discussion is good enough as to deserve attention from those who are generally interested in the topic. However, they may enjoy most the cases discussed in the second part of the book, since they all are worth reading on their own. H. Longino and L. Principe argue for a contextual view of evidence that challenges Achinstein's concept of veridical evidence. Principe shows how the concept of alchemical transmutation came to be rejected sometime around 1700 without the addition of new evidence against it, but as a result of somehow unrelated factors. Longino surveys what is taken to be evidence regarding our behavior according to different theoretical approaches that coexist today at many university departments, claiming that pure evidential considerations will not contribute much to bring them together. A. Rosenberg, to the contrary, ponders quite optimistically the particular contribution that genomic evidence might make to account for human cooperation (though without explicit implications for or against Achinstein's view). In a thorough and subtle analysis, R. Richards reconstructs the epistemic situation (an Achinsteinian concept once more) in which Darwin and Wallace trace the analogy between natural and artificial selection, showing what evidential support could the latter provide to the former. Drawing on another concept by Achinstein, K. Staley proposes a fine analysis of the potential evidence that could have supported the discovery of the top quark in 1994, comparing it with an error statistical account in the spirit of Mayo. Finally, G. Hatfield discusses current epistemological views of introspective evidence that he contrasts with those of some early pioneers in experimental psychology.

The inevitable question for many readers will be to wonder how well Achinstein´s theory of evidence fares in view of all of these arguments. The yardstick might well be the often quoted challenge thrown by Achinstein's dean: Does it really account for the way evidence is actually used in science? There is no doubt that the proposed distinctions are useful to discern and ponder the various roles played by evidence in scientific argumentation, yet we may question what its proper scope is. The empiricist in search of a foundational justification for scientific beliefs will be certainly pleased with Achinstein's theory: if properly expanded, as suggested by Gimbel, Kronz and McLaughin, it will provide a realistic model for rational decision making in science either in the pursuit of new evidence or for the assessment of what we actually have. Yet, as usual, the contributions of Principe, Longino and even Richards cast some doubts about the convenience of such epistemological framework to account for the consensus that scientists eventually achieve. These matters of principle would have deserved a more explicit discussion. On the other hand, even if the second part is titled "Applications", we will certainly miss concrete analyses of how the combination probabilities cum explanatory nexus rationalize current or past scientific consensus. At least for this reader, the discussion suggests that Mayo's approach is more effective in dealing with tests effectively performed today in many disciplines and less committed to a foundationalist stance. Other readers, however, will surely draw different conclusions: the book is rich enough in arguments as to allow this to happen.

{November 2005}
{Metapsychology online reviews}