The heuristic account and the periodic table

In 2001, Worrall and Scerri critically reviewed claims held by philosophers and historians of science (Maher 1988, Lipton 1991, Brush 1996) that the acceptance of Dimitri Medeleev’s periodic table within the scientific community was strongly influenced by the prediction of novel elements made by Mendeleev based on “gaps” within the table. The intuition behind these claims has to do with the justification of theories. If the evidence e temporally precedes the hypothesis h, which explains e, then h can be constructed in such a fashion that it will always account for e; h would be ad hoc. If on the other hand, h precedes e then h, so the story goes, takes a high risk of being refuted in case e is not observed but instead e’ is. Mayo (1991, 1996) has developed this intuition – probably first explicated by Popper (1965) – into an account of severity of test. But as it turned out it seems to be ridiculously hard to find any historical evidence that would support the thesis that a high degree of severity of test would make h more acceptable to the scientific community. This has been shown in numerous studies by the historian Stephen Brush (1989, 1990, 1993a, 1993b) inspired by Worrall (1989) who buried the myth surrounding the major impact of the novel prediction by Fresnel’s theory about the Poisson spot. Yet, Brush (1996) thought to have found the exception to this “rule” in Mendeleev’s periodic table. Partly based on Brush’s research on Mendeleev’s periodic table, Scerri and Worral pointed out that the conclusions Brush drew from the historical material are unwarranted. Despite wrong-headed criticisms by Barnes (2005) and enduring resistance by Brush (2007), I think Scerri and Worrall have shown almost beyond doubt that novel predictions – contrary to all intuitions – did not extraordinarily contribute to the acceptance of the table. However, Scerri and Worrall claimed that their study also gave evidence to the “heuristic account” devised by Zahar (1973) and defended by Worrall (1989). The heuristic account replaces temporal novelty by use-novelty: A fact e, explained by h, is to be considered novel if e has not been used in the construction of h or in the wordings by Mayo (1991): (i) T entails e, but (ii) e is not used in the construction of T. Mayo has called this the use-novelty rule (UN). However, since its invention, the heuristic account has faced the charge from psychologism: one actually needs to get into the head of someone in order to establish whether a particular piece of evidence was used in the construction of h or not. Worrall (1989), however, has always denied