Predictions of Mendeleev

Elements and Discovery Date Properties Predictions of Mendeleev in 1871 Determined Values... 

Estimated predictions of Mendeleev about gallium and its determined values. 

The priority dispute took place in the Berichte derDeutschen Chemischen Gesellschaft and could be easily read by the chemical community, which at the time was dominated by organic chemists. How did the chemical community respond to the priority dispute At first, no other papers could be found that explicitly discussed the priority dispute. It cannot be excluded that the correspondences reflected the dispute." But it is obvious that the authors who dealt directly or indirectly with the questions of atomism, elements, and their classification or the determination of atomic weights often mentioned the successful predictions of Mendeleev and showed some reticence concerning Meyer s work. It seems that the success of (some) predictions of new elements influenced the process of forgetting about Meyer s precautionary contribution and overemphasizing the role of the predictions in the process of acceptance. ... 

The periodic system was compatible with the chemistry done in Sweden. That the predictions of Mendeleev were corroborated by empirical analysis, both in chemistry and in physics, contributed to its acceptance. The more difficult empirical tests a theory could withstand, the more it could be accepted as a theory—such was the view of the atheoretical scientists in Sweden. Mendeleev s system passed that test relatively easily, but chemists in Sweden never did, thanks to their atheoretical attitude, ask for an explanation as to why regularities occurred in Mendeleev s system. To give such an explanation would be to propose a hypothesis. To the empirical chemists in Sweden the system remained a description of empirical facts, and a description that did not have to connect with any unproved metaphysical hypothesis on the ultimate structure of matter. 

E. Barnes, On Mendeleev s Predictions Comment on Scerri and Worrall, Studies in History and Philosophy of Science, Part A, 36 801-812, 2005 S. Schindler, Use-novel predictions and Mendeleev s Periodic Table Response to Scerri and Worrall, Studies in History and Philosophy of Science, Part A, 39 265-269, 2008 D. Harker, On the Predilections for Predictions, British Journal for the Philosophy of Science, 59 429—453, 2008 E. R. Scerri, Response to Barnes Critique of Scerri and Worrall, Studies in History and Philosophy of Science, 36 813-816, 2005. 

The direct focus of the present paper, however, is not this general issue, but rather a famous particular case from the history of science that has become embroiled in it. The episode involves Mendeleev and the prediction of the existence of hitherto unknown elements on the basis of his celebrated periodic table. According to an account that has widespread currency, Mendeleev s table was given little or no general credit by his contemporary scientists in virtue of its accommodation of the already known elements. What really told with Mendeleev s peers, according to this account, was the fact that gaps in the table were used as the basis of predictions of the existence of hitherto unrecognised elements, that turned out really to exist. So, for example, Isaac Asimov writes ... 

And Maher goes on explicitly to underline the conclusions about confirmatory weight that he sees as illustrated by this episode. He claims that Mendeleev s prediction of the existence of the third of the new elements, eka-silicon (aka germanium), was initially regarded as quite unlikely to be true but then later, with the discovery of the first two new elements (gallium and scandium), confidence in the prediction of the existence of the third new element became so high that its eventual empirical confirmation was widely regarded as a matter of course. Maher writes ... 

As this indicates, Brush s view of other historical cases (see for example Brush, 1989) has been that, whatever philosophers of science may claim about the relative merits of prediction and accommodation, there is no historical evidence that the scientists concerned in these cases gave any special weight to predictive success. But he feels forced by the evidence in the case of Mendeleev to conclude that there—perhaps uniquely—special weight was indeed given to predictions. We will argue that the historical evidence that Brush himself alludes to in fact, on careful analysis, fails to support this conclusion. 

No one could deny, of course, that it is one thing for a theory to make predictions of the existence of hitherto unknown elements and quite another for it to make successful, empirically verified predictions. But neither Maher nor Lipton—nor, so far as we can see, anyone else—cites any substantial evidence for the sceptical attitude of Mendeleev s fellow scientists in 1869 or 1871 and the only evidence they cite for the increased confidence in the theory as a result of the successful predictions is the award to Mendeleev of the Davy Medal in 1882-— and this, as we show in the next sub-section, turns out to be (worse than) unconvincing. 

Maher draws attention to the 1874 paper of Lecoq de Boisbaudran in which he announced the discovery of gallium, the first of Mendeleev s predictions to be confirmed. And this is, indeed, an important paper. One possible misconception should, however, be quashed immediately. De Boisbaudran did not discover gallium as a result of testing Mendeleev s prediction. Instead he operated quite independently by empirical means in ignorance of Mendeleev s prediction and he proceeded to characterise the new element spectroscopically. De Boisbaudran s findings were published in the Comptes Rendus. 

Lipton talks of the predictive successes with gallium and scandium as giving greater credence to Mendeleev s theory (a troublesome notion as already noted and as we shall see later in more detail) and only thence to the further prediction of germanium from that general theory. But Maher talks more directly of the impact of the successes with the first two new elements on the confidence that chemists... 

The answer is patently no —not because of any general sceptical-philosophical scruples about induction, but because of the particular fact that the impression of consistent predictive success for Mendeleev s scheme is a complete misrepresentation of history a classic example of an effect (Mendeleev s predictive success ) created by selection-bias. Mendeleev made any number of predictions on the basis of his scheme (or rather schemes—there are at least 65 versions of Mendeleev s table, published and unpublished). Many of these predictions (depending on how exactly they are individuated)—perhaps a majority—were unsuccessful. 

Calling two such elements x and y—both lighter than hydrogen and x being the element that consituted the optical ether—Mendeleev based his predictions of their properties on numerical relations between atomic weight ratios in the following periodic table which he devised in 1904 ... 

Mendeleev predicted that eka-silicon is a refractory substance as predicted by Mendeleev. In fact it melts at the relatively low temperature of 950°C.16... 

Stephen Brush has, in several papers (for example Brush, 1989), examined different episodes from the history of science and argued that the scientists involved in them did not in fact regard (temporal) predictive success as carrying an epistemic premium. However, in a recent paper (Brush, 1996) on the reception of Mendeleev s ideas, he claims that this case is exceptional—here temporal novelty of predictions really did count. 

After spending considerable time perusing the crumbling pages of late nineteenth-century chemistry journals and textbooks, I have confirmed the traditional account Mendeleev s periodic law attracted little attention. . . until chemists started to discover some of the elements needed to fill the gaps in his table and found that their properties were remarkably similar to those he had predicted. The frequency with which the periodic law was mentioned in-journals increased sharply after the discovery of gallium, most of that increase was clearly associated with Mendeleev s prediction of the properties of the new element, (p, 617)20... 

But, to repeat, the extent of the diffusion of Mendeleev s scheme is not the issue here—no one disputes that this was widely (though not universally) accepted by 1890 in the scientifically advanced countries. The question at issue is the relative strength of the roles played in this diffusion by evidence of various sorts and in particular by the successful predictions of new elements. 

---