A part of the scientific literature consists of intermediate results within a longer project. Scientists often publish a first result in the course of their work, while aware that they should soon achieve a more advanced result from this preliminary result. Should they follow the proverb “a bird in the hand is worth two in the bush”, and publish any intermediate result they get? This is the normative question addressed in this paper. My aim is to clarify, to refine, and (...) to assess informal arguments about the choice whether to publish intermediate results. To this end, I adopt a rational decision framework, supposing some utility or preferences, and I propose a formal model. The best publishing strategy turns out to depend on the research situation. In some simple circumstances, even selfish and short-minded scientists should publish their intermediate results, and should thus behave like their altruistic peers, i. e. like society would like them to behave. In other research situations, with inhomogeneous reward or difficulty profiles, the best strategy is opposite. These results suggest qualified philosophical morals. (shrink)
Les théories physiques sont aujourd'hui très mathématisées, et ce que les scientifiques manipulent pour décrire, prédire et contrôler les phénomènes, ce sont (entre autres) des équations, comportant de nombreux symboles mathématiques. Ces objets mathématiques n'ont pas de signification physique en eux-mêmes : ils ne « parlent » pas d'eux-mêmes des phénomènes. Une interprétation est nécessaire. Ce qui nous intéresse dans cet article est ainsi l'interprétation dont une théorie physique doit faire l'objet pour remplir son rôle. Nous commençons par expliciter une (...) distinction traditionnelle : l'interprétation « pauvre » (simple instrument permettant d'assigner aux symboles de la théorie un sens physique strictement limité aux résultats des expériences) diffère de l'interprétation « riche » (laquelle compose une image du monde compatible avec la façon dont la théorie décrit mathématiquement les résultats des expériences). Notre but dans cet article est de montrer que cette distinction doit être amendée. Nous nous appuyons sur l'exemple de la mécanique quantique, mais la distinction se veut valable en général pour toute théorie physique. (shrink)
The coexistence of several interpretations of one theory is considered through the example of non-relativistic quantum mechanics. The problem considered is whether physicists manage to work properly in spite of the several interpretations. The criterion adopted is the possibility of re-using others' works for another research: this is called "fruitfulness of works". It is argued that such a fruitfulness is possible between works made in different quantum interpretations.
In this paper, I address a question in social epistemology about the unity of a scientic community to- wards its inner groups (teams, labs...). I investigate the reasons why these groups might want to \go it alone", working among themselves and hiding their discoveries from other groups. I concentrate on the intermediate results of a longer project, where the first steps can help to achieve a more advanced result. I study to what extent the isolation of research groups might be (...) damaging to the epistemic progress of the whole community. In order to get an idea of the reason why scientists might not share their intermediate results with others, consider for example Schawlow and Townes in the summer of 1958 (the historical material is drawn from Broberg, 1991, chap. 3): together they had just designed the principle of functioning of what would be known as the laser, with the intention of constructing the rst laser, relying on their theoretical ideas. Should they have published their theoretical discovery, or rather taken advantage of their advance on possible oppo- nents and start the experimental work? They chose to publish their theoretical ideas in a scientic journal, and they also led a patent application. On the one hand, it was a valuable choice: the patent was granted and their paper was published, making them soon famous. Further- more, another scientist called Gould was to le a simi- lar patent application a few months later, in March 1959; it was refused, because of the priority of Schawlow and Townes. If the latter had not submitted their idea in 1958, the patent would certainly have been acknowl- edged to Gould. In this sense, publishing their intermediate result was a worthy strategy. On the other hand, history showed that their publication had also some drawbacks for them. Schawlow and Townes' publication triggered a race among many American laboratories for the rst experimental construction of the laser. In addition to bringing them competitors, their publication helped these competitors in crucial ways; so much so that it was another scientist, Maiman, and not Schawlow and Townes, who was able to build the rst working laser, in May 1960, and to publish the result. If Schawlow and Townes had not published their theoretical idea, there are chances that they would have been the first ones to build a laser. Their publication of the first theoretical result certainly prevented them from getting the second experimental one. In short, the best publishing strategy about an intermediate result is no easy question for a scientist. Consider now the whole community, concerned with the results available to all and with the pace of the epistemic progress. Schawlow and Townes' publication enabled the rst laser to be built quicker, as other teams could rely on their theoretical work. More generally, being able to reuse other's works seems to be a rm basis of scientic progress. Thus, for the whole community, it seems that there are only benets if intermediate results are published. As this is not necessarily the best strategy for those who reach these intermediate results, there might be a conflict between the interests of individuals and those of the whole community. The question Schawlow and Townes faced was whether they should make their discovery publicly known, or keep it for themselves to progress alone. This alternative can also arise for an entire lab, when competing with other labs on some long-run project: should it make its discoveries known to its opponents, or only share them internally among its members? At a higher level still, this question arises for what can be called scientic schools, which are groups of labs sharing some particular approach|methodological, experimental or interpre- tational, for example. At each level within the scientic community, there are competitive groups of scientists which may decide to go it alone and to prevent their results from being accessible to others. If they had good reasons for doing so, there would be dramatic consequences for the whole community: other groups would not be able to reuse their results, labour force would be wasted in achieving them again, and the progress of the eld would be slowered. 9 The aim of this paper is to investigate this risk of division: to assess the conditions under which it may oc- cur, its damage for the community, and to draw the limit of an acceptable pluralism. The reasoning with benets and drawbacks which has been sketched for Schawlow and Townes seems to be implicitly an informal economic argument. In order to clarify and to rene such an argument, I will investigate it in a rational decision framework, supposing some utility or preferences for the groups of scientists, and I will propose a formal model, in game theory. In social epistemology of science, formal models have been proposed for various questions (for example, Kitcher 1990, 1993, Goldman 2009, Strevens 2003, 2006, Weisberg and Muldoon, forthcoming, Zollman 2007, 2008, 2009, 2010), but not for the question of the division of the scientic community. I will concentrate on the normative aspect of the question, ignoring the descriptive and historical question of whether groups do go it alone or not. The model I propose is sequential: two groups are competing on a research program, which consists of several results which have to be obtained in order. The groups always carry out some research for the next result, which they reach with some probability per unit of time. When a group, say A, reaches a result, it can either make it public for the other group, or continue its research. After a result is made public, the competitor B can start from it so as to continue its own research. Thus, when A publishes it loses its lead, and its competitor will be able to compete with it in the near future. But it is also recognized as the rst to reach this step. I consider two dierent cases, one in which the number of steps is limited, another where the limit concerns the number of temporal intervals. In both cases, assuming that all steps are equally dicult and equally rewarding, it turns out that the best strategy for a group concerned with its own reward is to make public every step it takes. Thus, there is no risk that groups might separate: a scientist, whatever its group, will publish for all. Finally, I turn to more complex situations, where results have varying size or diculty: there, groups can indeed be better o going it alone. (shrink)