Repenser l'évaluation comparative des théories scientifiques - Stabilités, Ruptures, Incommensurabilités ?

 

Aristides Baltas (Athens)
Commentator / Commentateur : Jean-Michel Salanskis (Paris)

Title / Titre : What can Wittgenstein contribute to the understanding of scientific theory change?
Abstract / Résumé : Wittgenstein's views at both ends of his thought (the early of the Tractatus and the late of his On Certainty) can offer some important light of what is really at issue at the semantic level of radical theory change. It will be argued that, at least in some telling cases in the history of physics, the phenomenon of meaning incommensurability is indeed present (and perhaps unavoidable if the change is radical enough), but this does not lead to relativism or total communication breakdown while the notion of progress, if properly clarified, can be 'saved'. 

Anouk Barberousse (Paris)
Commentator / Commentateur : Igor Ly (Paris)

Title / Titre : D'une variante à l'autre. Le changement intra-théorique
Abstract / Résumé : Lorsque l'on cherche à caractériser une théorie scientifique, en particulier celles du passé, on se heurte à l'obstacle suivant : quelle version en choisir ? Une théorie scientifique est en effet un objet éminemment diachronique et évolutif, même si elle conserve une certaine unité qui fait qu'elle peut légitimement garder le même nom au cours des décennies, voire des siècles. La communication tentera de répondre à la question suivante : comment appréhender les changements intra-théoriques ? On les a parfois considérés comme des approfondissements successifs, ou comme des élargissements à des domaines plus vastes ; mais ces descriptions valent également pour certains cas de changement inter-théorique. Quelle est dès lors la différence entre changement intra-théorique et changement inter-théorique ? 

Alexander Bird (Bristol)
Commentator / Commentateur : Paul Hoyningen-Huene (Hanover)

Title / Titre : A Naturalistic Understanding of Incommensurability and World-Change 
Abstract / Résumé : In this paper I propose a way of understanding Kuhn's earlier discussion of incommensurability and world-change that emphasizes and draws upon Kuhn's extensive use of psychological evidence. Kuhn argued that the way we literally see the world (our perceptual experience of the world) can be altered by a shift in paradigms. He believed that this idea could be extended to a more general or metaphorical way in which the way we 'see' the world changes when a paradigm changes. This thought is an important part of Kuhn's earlier understanding incommensurability and world-change. But the idea was never articulated in detail in a way that satisfied critics and Kuhn moved away from a psychological conception of incommensurability to a linguistic, philosophical conception; he also dropped talk of world-change. I aim to show that the earlier approach was on the right lines and that a satisfactory, detailed psychological (and so naturalistic) account of incommensurability and world-change can indeed be found. 

Martin Carrier (Bielefeld)
Commentator / Commentateur : Michel Bitbol (Paris)

Title / Titre : The Aim and Structure of Methodological Theory 
Abstract / Résumé: One of the challenges Kuhn's work poses to philosophy of science concerns the insight that theory-choice and, accordingly, theory-change is governed by a more complex and subtle procedure than anticipated. In particular, this procedure is claimed to inevitably and justifiedly leave room for individual preferences so that theory-choice fails to be determined unambiguously by criteria with epistemic bearing. This methodological uncertainty is now frequently labeled as Kuhn-underdetermination. Unlike Duhem-Quine underdetermination, it does not require empirical equivalence but rather refers to a situation in which alternative theories have their strengths and faults in different areas and in different respects so that no clear overall picture emerges. 
Overarching methodological theories can be construed as attempts to overcome the limits set by Kuhn underdetermination. In this perspective, theories like Lakatosianism and Bayesianism provide rules for epistemic judgments that are intended to make a clear evaluation of the credentials of rivaling scientific theories possible. Among other things, Lakatosianism features the prediction of novel facts, Bayesianiam places emphasis on the increase in the likelihood of the data through the adoption of a hypothesis. The two methodological theories are supposed to serve as guidelines for methodological judgment or at least to explain with hindsight why a particular theory was picked. However, on closer scrutiny the two methodological theories founder in this task of accounting for theory choice decisions. The criteria of excellence they specify are liable to uncertainties of the same sort as the more traditional virtues they are intended to replace. Lakatosianism suffers from an arbitrariness of delineating research programmes and their scopes. The methodological reconstruction is strongly underdetermined by the historical evidence. Bayesianism comes to grief because of the arbitrariness of imputing numerical values to all conditional probabilities needed in order to arrive at hypothesis probabilities.
An alternative picture is that methodological theories suggest general maxims and rules that guide the confirmation process rather than provide criteria for specific theory-choice decisions. Methodological theories serve to connect and unify such maxims and rules. Traditionally, lists of methodological virtues are drawn up ad hoc. One could easily add further criteria or delete others. By contrast, methodological theories provide a coherent approach to appreciating scientific theories and comparing their explanatory achievements. And they give a rationale for why these rules rather than others deserve to be preferred. For instance, Lakatosianism places theoretical progress and the growth of knowledge at center stage. This general commitment ramifies into emphasizing the guidance of theory development by heuristic directives and the prediction of formerly unknown phenomena. Analogously, Bayesianism is led by its commitment to Bayes' theorem to stressing virtues to the effect that a good hypothesis should be plausible in light of the background knowledge and show that the data could have been expected beforehand. Methodological theories serve to give a unified account of such rules. 

Bernard D'Espagnat (Paris)
Commentator / Commentateur : Marcel Weber (Hanover)

Title / Titre : Conflicting theories or mere interpretation problems? / Conflits de théories ou problème d'interprétation ? 
Résumé : Le langage objectiviste est incompatible avec les données quantiques. Il l'est, même dans une épistémologie critique (" antiréaliste ") définissant la vérité à partir de la possibilité de connaissance, sauf à admettre que la prise de connaissance peut créer une propriété et le peut, même à distance. Ceci invite à interpréter la science comme une activité centrée moins sur la description des phénomènes que sur la prévision d'observations. Et, de ce fait, nombre de conflits entre théories scientifiques apparaissent comme n¹étant que des problèmes philosophiquement mal posés.

Abstract: In science statements are either descriptive or predictive of observation. Now it turns out that quantum physics is expressible in terms of the latter but not in terms of the former. Moreover this holds true, even within a "critical", or "antirealist", epistemology (linking truth with "what we can know"), unless we are prepared to admit that getting informed of a property may create it, and may do so even at a distance. If we don't grant that much, we have to interpret physical theories as being essentially predictive of observation. And many of the criticisms epistemologists directed to science may then be shown to just vanish. 

Rom Harré (Oxford). 
Commentator / Commentateur : Roger Pouivet (Nancy)

Title / Titre : Undetermination is a positivist myth 
Abstract / Résumé : Returning to the source of this idea, Clavius preface to his astronomical work, and the writings of N. R. Campbell, we find that the set of possible theories compatible with empirical data is restricted by the natural kind contraints incorporated in the source or sources of the models (iconic) which a theory describes. The belief that theories are selected by reference to predictive/retrodictive success is a consequence of a logicist view of theory construction. 

Stephan Hartmann (London). 
Commentator / Commentateur : Edward Jurkowitz (Chicago)

Title / Titre : Modeling High-Temperature Superconductors: Correspondence at Bay? 
Abstract / Résumé : There are many ways in which scientific theories relate to their predecessors. By examining cases from science, historical and recent, I show that various elements from a predecessor theory may be taken over (such as the vocabulary, operative descriptions, laws, models and mathematical structures) and integrated into a new theoretical scheme. Scientists seem to use a predecessor theory (and maybe other theories as well) like a 'toolbox' that can be employed to construct a new theory. But how much of the predecessor theory is taken over? And in exactly what relation does it stand to its successor? One type of answer to these questions is given by the various correspondence principles offered in the literature. Most ambitiously, Heinz Post's General Correspondence Principle states a formal condition for the relation between anypredecessor theory and its successor. Post also claims that correspondence considerations are heuristically important. I will examine this claim by looking at a case study from contemporary physics. So far, there is no consensus in the relevant scientific community as to what a theory of high-temperature superconductivity should look like; two opposite strategies are followed. The conservative strategy is to retain as much of the structure of the well confirmed theory of ordinary superconductors ("BCS theory") as possible. The revolutionary strategy contends that a radical break with the predecessor theory is unavoidable. In this case, it does not even make sense to call the BCS theory a predecessor theory. The debate between the advocates of both camps has been carried out in major physics journals and, perhaps surprisingly, philosophical argumentation has not been a stranger to it. While reconstructing this debate I will show that the motives that led physicists to follow a certain methodological strategy were subtle and sensitive to the details of the scientific problem in question. Universal methodological principles, such as the General Correspondence Principle, may turn out to be useful only if they are carefully adapted to particular scientific situations. 

Thomas Nickles (Reno)
Commentator / Commentateur : Emiliano Trizio (Paris)

Title / Titre : Disruptive Scientific Change 
Abstract / Résumé : For Thomas Kuhn, incommensurability arises within a field when a new paradigm, after a revolutionary struggle, succeeds in vanquishing the old one. It also arises on a smaller scale (or threatens to arise) whenever an unexpected discovery is made within normal science, for such a development raises a recognition problem of whether it is even a contribution to that field. Although Kuhn does allow paradigm change within subspecialties containing only a few dozen practitioners, his conception of fields is pre-World War II. He does not take fully into account the emergence of new specialties and cross-disciplinary combinations. That may be one reason why he misses another, possibly more radical form of incommensurability or rupture, namely, those cases that stem from apparently extraneous and irrelevant or inappropriate developments that eventually disrupt and displace an established field and its practices. This sort of case is suggested by recent work (by Clayton Christensen) on disruptive technologies in the business world as well as by some of Kuhn s own early historical work.
Most treatments of incommensurability are either theory-centered (focusing on the logic, meaning, and reference of theories) or else methodological (focusing on incompatible goals and differing value commitments), with an epistemic emphasis. By contrast, my focus will be pragmatic and heuristic, i.e., practice-centered and future-oriented. The disruptive model is worth considering as a kind of discontinuity that presents severe recognition or intelligibility problems for specialists, whether or not it is usefully classified as a kind of incommensurability. After presenting the model, I shall briefly consider some historical cases for discussion from the disruptive point of view. 

Robert Nola (Auckland)
Commentator / Commentateur : Stephen Clarke (Canberra)

Title / Titre : The Optimistic Meta-Induction and Ontological Continuity: The Case of the Electron 
Abstract / Résumé : The pessimistic meta-induction attempts to make a case for the lack of ontological continuity with theory change. In contrast, the optimistic meta-induction makes a case for the high frequency of ontological continuity with successive theory changes. However both arguments must presuppose some account of how the theoretical terms of scientific theories relate to reality before issues of ontological (dis)continuity can be settled. The task of this paper is to set out an account of reference fixing that enables referential, and thus ontological, continuity with theory change. The account employs a version of Russell's theory of descriptions, but modified to allow for reference to kinds rather than individuals. It is also modified to allow for reference to items that do not fully satisfy the descriptions employed, and to allow for multiple reference or referential ambiguity. It is also argued that Russell's theory of descriptions, so understood, is a special case of Ramsey's theory of how to define theoretical terms, with Ramsey's theory understood in the way developed by David Lewis. With this background, a case for ontological continuity can be made that can be widely applied in a number of historical cases. 
The case study examined here is that of the electron. It begins with an outline of the first experimental identification of cathode rays by Plücker and Hittorf by means of properties wholly extrinsic to cathode rays, using a Russellian reference-fixer. The stage is then set for obtaining new knowledge about the kind so identified, for introducing names for the kind (e.g., Goldstein's 'Kathodenstrahlen') and for proposing rival theories of the kind such as the aether and particle theories. The rival theories may provide rival concepts of cathode rays, but they are still concepts of the same "something" picked out, in a theory independent manner by the Russellian description.
For our purposes the next salient stage in the history of cathode rays concerns how, through the work of many experimentalists up to J. J. Thomson, the "somethings" parochially identified only in cathode tubes were then claimed to be constituents of all matter. It is argued that the new reference fixer that picks out these universal constituents still refers to the same kind of "something" identified in cathode tubes. However the term 'corpuscle' that Thomson introduced to refer to them depends on a theory-laden reference fixer that in fact picks out nothing. Thomson resisted using the term 'electron' to refer to these "somethings" well into the 1900s. The term 'electron' was introduced to refer to a "vortex atom" in an aether field. But eventually it was used to refer to the ubiquitous "somethings" that Thomson identified, but without the implication that it was an aetherial item or the same as Thomson's own corpuscle. Finally, the paper makes a link to the work of others that makes a case for subsequent continuity in our talk of electrons in the Lagrangian equation for electrons. Once again the theory of reference fixing developed above shows how this is possible without providing support for the pessimistic meta-induction. 

Andrew Pickering (Urbana Champaign, Illinois)

Title / Titre : 'Machinic and 'performative' incommensurabilities 

Léna Soler (Nancy)
Commentator / Commentateur : Howard Sankey (Melbourne)

Title / Titre : A new kind of incommensurability between scientific theories? / Une nouvelle espèce d'incommensurabilité entre théories scientifiques ? 
Abstract: The paper will focus on the idea, introduced in the last decades, of "literal incommensurability" (I. Hacking) or "machinic incommensurability" (A. Pickering). Reconsidering the relations between theories and experimental actions in the light of the works of the 'new experimentalists', I will distinguish and discuss two modalities of 'literal incommensurability', and will argue that literal incommensurability, although naming a potentially interesting problem, does not dissolve the traditional problem of incommensurability.

Résumé : L'intervention examinera l'idée, introduite dans les dernières décades, d'incommensurabilité " littérale " (I. Hacking) ou " machinique " (A. Pickering). Elle interrogera, à la lumière des travaux des 'nouveaux expérimentalistes', les rapports entre théories et actions expérimentales. Elle distinguera et discutera deux modalités d'incommensurabilité littérale, et elle arguera que l'incommensurabilité littérale, tout en désignant un problème épistémologique potentiellement intéressant, ne dissout toutefois pas le problème traditionnel de l'incommensurabilité. 

Friedrich Steinle (Stuttgart)
Commentator / Commentateur : Scott Walter (Nancy)

Title / Titre : Forming Incommensurable Frameworks: the Case of Electrodynamics 
Abstract / Résumé : Competing scientific theories may differ from each other in a wide range of aspects, such as differences in focus, or different, but translatable assumptions about microscopic entities, or even fundamentally different conceptual frameworks. In my talk, I shall focus on situations in which constellations of the latter type (often called "incommensurabilities") may come about. The case of 19th ct. electrodynamics is illustrative, in which two theoretical systems - action-at-a-distance theory and field theory - coexisted for several decades. Though they dealt largely with similar empirical effects, the two systems differed drastically in their mathematical toolbox and, most important, in the basic concepts by which they were formulated. Moreover, it is striking to see that at the origin of the dichotomy, the two main actors - Ampère and Faraday - started out from the same electromagnetic experiments in the 1820s, and introduced, in the context of intense experimental work, new but essentially different concepts. The incommensurability was created step by step. To understand this peculiar development, one has to ask questions about the specific uses of these concepts, about the reasons for their introduction, and about the actors' epistemic goals in their experimental activities. Going far beyond the specific historical case, these questions open new perspectives on the relation between experiment, concept formation, and the origin of incompatibilities. 

Paul Teller (San Francisco)
Commentator / Commentateur : Ronald Giere (Minneapolis)

Title / Titre : Of Course Idealizations are Incommensurable! 
Abstract / Résumé : The conference call cites a number of problems circling around the idea of Kuhnian incommensurability: Scientific rationality, realism, relativism, underdetermination, stability of reference, variable methodological norms. All these problems are real, but I suggest, their apparent intractability is an artifact of thinking of science as being in the business of finding comprehensive, exact truths. Instead, think of science, along with Cartwright and Giere, as being in the business of systematic development of idealized models, always limited in scope and never completely accurate. Bear clearly in mind that what model is appropriate depends sensitively on interests and other contextual features, and that evaluation always needs to be relativized to the limits of human capacities. Now reconsider the issue of incommensurability and its associated problems, thinking in terms of the simple example of alternative idealized accounts of water in terms of continuum hydrodynamics and particle statistical mechanics. Treatment of theory pairs such as classical and quantum mechanics will, in many respects, be analogous. Also keep in mind that the humanly accessible information in, for example, quantum field theory is extraordinarily limited, in particular affording us no humanly accessible information about the fluid or dispersive properties of water. Neither model of water tells us what water “really is” not exactly! But both tell us real things about water. Evaluation is relative to our predictive, explanatory and other interests, both practical and intellectual. If the best that is humanly possible are alternative idealizations, we expect the phenomena to be usefully approached though a variety of theoretizations and underdetermination suddenly looks not only natural, but an asset. Intertheoretic comparisons of reference is certainly a problem, but, if viewed in terms of comparison of idealizations, a naturally intelligible one. Relativization of choice of idealization to interests dictates that norms of evaluation with be analogously variable. Of course idealizations are incommensurable! There is much interesting work to be done in understanding how dealizations are to be compared, probably work that must be done on a local, case by case basis; but the issues now look intelligible and useful. 

Hervé Zwirn (Paris)
Commentator / Commentateur : Soazig Le Bihan (Nancy)

Title / Titre : Monde classique versus monde quantique : les implications pour le réalisme 
Abstract / Résumé : Les difficultés à concilier la description du monde que donne la mécanique quantique avec son apparence classique, à laquelle nous sommes accoutumés, semblent être une indication de ce que le réalisme au sens naïf du terme n'est plus une position philosophiquement tenable. Quelles sont les conséquences qu'il faut tirer d'un tel constat et permettent-elles d'aboutir à des conclusions solidement fondées ?