Abstract
According to Wimsatt’s definition, the robustness of a result is due to its being derivable from multiple, partially independent methods, and increases with the number of such methods. In the case of the experimental sciences, the multiple methods will amount to different types of experiments. But clearly, this holds only if the convergent derivations involved are genuine arguments, that is, if each of them can be considered as sufficiently reliable or solid. Thus, the issue of the robustness of results inevitably leads to a reflection on the solidity of methods. What is, then, that makes a method, and in particular an experimental procedure solid? Despite the possible worries of circularity, part of the answer lies, without doubt, in a sort of reversed formulation of Wimsatt’s definition: the solidity of a method will increase with the number of independent results, previously established as robust, that it will enable to be derived. But this seems to be only a part of the answer. Intuitively at least, it is expected that the solidity of a method could also be linked to specific properties of this method, to features that are more ‘intrinsic’ than the results it allows to derive. In this chapter, I try to probe into the nature of these ‘intrinsic’ characters, through a discussion of an example connected to the discovery of weak neutral currents in particle physics. More precisely, the method that will be investigated is an experimental procedure developed at the beginning of the 1970s, which uses a giant bubble chamber named Gargamelle, and which is commonly believed to have contributed to establishing the existence of weak neutral currents. I analyze the content of the Gargamelle experimental ‘proof’ and bring to light its internal architecture. Then I examine the relations between this architecture and the wimsattian scheme of invariance under multiple determinations. Thereafter, I specify this scheme, and draw some general conclusions about the solidity of methods and results. Finally, some implications with respect to the issues of scientific realism and the contingency of scientific results are sketched.
This is a preview of subscription content, log in via an institution to check access.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Notes
- 1.
Pickering (1984). See also Galison (1983). This historical case is very interesting from a philosophical point of view (and has led to controversial interpretations). I discussed it with respect to the issue of a possible incommensurability at the experimental level in Soler (2008c), and with respect to the issue of contingentism in Soler (201X).
- 2.
For the sake of simplification, I only take into account the supportive side of an argumentative line. But actually the concept of an argumentative line, as I conceive it, is intended to be of broader scope, and to encompass, as well, the negative arguments, that is, arguments that play against a result R. So at the most general level, an argumentative line is any argument which is relevant to R.
- 3.
For more details about this interaction see below Section 10.5.
- 4.
See Chapter 1, Section 1.8 for more developments on this distinction between two kinds of independence.
- 5.
One reason to think that the leap is questionable is the multiplicity of the historical cases in which schemes of robustness indeed obtained at a given research stage S 1 have been dissolved at a subsequent stage S 2. In that movement, the node R first taken to be objective/true on the basis on the robustness scheme indeed available at S 1, is, subsequently at S 2, re-described as an artifact, a human mistake or the like. For insights about this configuration and a historical example, see the paper of Stegenga, Chapter 9 (as he rightly concludes: “concordant multimodal evidence can support an incorrect conclusion”. For other kinds of possible arguments against the jump from robustness to realism, see the concluding section of this chapter, and Chapter 1, Section 1.8.3.
- 6.
Hasert (1974). See also Hasert (1973b) (which is roughly the same paper but with fewer details), (Benvenutti 1974) (which target the same weak neutral interaction through an electronic experiment – what I have called above the NAL line), and (Hasert 1973a) (which is about another kind of weak neutral current, the muon-neutrino electron scattering, investigated with a bubble chamber – what I have called above the Aachen line). I do not claim, of course, that these papers exhaust the publications commonly considered to announce the discovery of weak neutral currents.
- 7.
A neutrino is a lepton, that is, a particle subject to weak interaction, whereas hadrons are particles subject to the strong interaction. To be more precise, the incident particles involved in the interaction are muonic neutrinos. The reaction corresponds to a neutral current, since no change of charge happens: both the incoming and the outgoing leptons are neutral particles.
- 8.
In what follows, the “Gargamelle line” refers to the road that goes from the machinic outputs (the photographic images on the film) to the conclusions about ‘what they say’ in terms of the NC/CC rate. What happened before in order for the experimenters to be in a position to obtain and trust the corresponding pictures (the history of the construction of the bubble chamber, the history of the knowledge that was required in order to conceive something like a bubble chamber and so on…), is not taken into account. In other words, the Gargamelle line as it is analyzed in what follows is restricted to what is often called “data analysis”. Obviously, the Gargamelle experimental derivation could be understood in a broader sense, including elements of the anterior history of science that are presupposed in order to take what I called the ‘degree zero’ of the experimental data (what appears on the film) as reliable data. In relation to the delimiting choice made in the present chapter and to the adopted re-description of the Gargamelle line so delimitated as a four-floors edifice, we could say that this four-floors building is not suspended in the void but rests on a deep and structured underground.
- 9.
For more considerations about the epistemological status of these logical steps, especially with respect to their relation to the chronology of actual scientific practices, see below Section 10.11.
- 10.
The experimenters don’t give any further explanation, but I will consider in more detail below (Section 10.13) the content of the process and the motivation that might have led them to retain this value ‘10’ from the three values 9, 11 and 11.
- 11.
As I developed in Chapter 1 (Section 1.8), the clause of independence is highly problematic. Clearly, in practice, independence is very often assumed intuitively and tacitly (without any systematic discussion or attempt of explicit clarification). The issue of independence is discussed further in this volume, notably in the contributions of Stegenga (Chapter 10), Nederbragt (Chapter 5) and Trizio (Chapter 4).
- 12.
Such claims involving a reference to a ‘true value’ are rarely made explicit, especially in published chapters, but I take them to be common intuitive ways of thinking among practitioners. In particular, such a framework commonly underlies the way practitioners understand and treat a series of actual measurement results associated with one and the same targeted variable when the results are obtained with one and the same instrument at different moments. Regarding this point, an interesting document is the 2008 version of the International Vocabulary of Metrology – Basic and general concepts and associated terms (VIM, 3rd edition, http://www.bipm.org). The document provides a unified vocabulary about “metrology, ‘the science of measurement and its application’ ” (p. vii), with the aim of being “a common reference for scientists and engineers (…) as well as for both teachers and practitioners involved in planning or performing measurements”, and “to promote global harmonization of terminology used in metrology” (p. 1), but it is of course not just a question of words. The elaboration of the final text has required an analysis of what it means to measure in the empirical sciences (an analysis of the different kinds of measurements, of the calibration procedures, of the basic principles governing quantities and units…), “taken for granted that there is no fundamental difference in the basic principles of measurement in physics, chemistry, laboratory medicine, biology, or engineering” (p. vii). Now in the final text, we read, in the introduction: “Development of this third edition of the VIM has raised some fundamental questions about different current philosophies and descriptions of measurement”. Two approaches are then contrasted: the “Error Approach (sometimes called Traditional Approach or True Value Approach)”; and the “Uncertainty Approach”. “The objective of measurement in the Error Approach is to determine an estimate of the true value that is as close as possible to that single true value. The deviation from the true value is composed of random and systematic errors. (…) [the two kinds of errors] combine to form the total error of any given measurement result, usually taken as the estimate.” (p. vii). It is not my aim here to explain the second approach, which is meant to get rid of the idea of a true value. I just want to stress that the first “traditional” approach coincides, in its fundamental features, with the one I have in mind in my analysis above. The second approach has been elaborated in response to the increased awareness that the traditional one – the only one involved in the previous versions of the VIM – was actually problematic (this point is still clearer in the 2004 first draft of the 3rd edition that has been submitted for comments and proposals to the eight organizations represented in the Joint Committee for Guides in Metrology (JCGM) and then revised according to their reactions; see the first paragraph of the Foreword). I take the fact that the “true value approach” has been the first one identified by the VIM, joined to the fact that it is subsequently described as the “traditional” one, as support in favor of the claim that it is indeed an intuitive, widespread largely tacit framework through which practitioners read the relation between different quantitative values obtained through different ‘derivations’ (in the VIM case: measurements) for one and the same targeted quantity. Since the 2008 edition is the result of a cooperation between numerous international experts, and since it has been approved by each of the eight member organizations of the JCGM, we can bet that it is representative enough with respect to claims about ‘widespread intuitive commitments’.
- 13.
See for example (Bachelard 1927) and (Hacking 1990). As Bachelard stresses in his book, even the nowadays pervasive and obvious idea that an average value is a good way to represent a set of numbers has, historically, been the object of important discussions (see especially chapter VII). In a similar vein, see Buchwald (2006), a very interesting chapter on the developing methodology of taking statistical averages, from scientists in France and elsewhere, around 1800. I thank Thomas Nickles to turn my attention to that work.
- 14.
In the present historical episode, the existence of individual variations of judgments at many levels is largely documented and attested. See for example (Galison 1983, 1997; Rousset 1996; Schindler 201X). Moreover, with respect to pragmatic judgments about which derivations are reliable/unreliable, trustworthy/not trustworthy, and as a limiting case, worth mentioning or even considering as an argument, it has to be stressed that what appears in a published paper is the result of an antecedent invisible ‘pre-selection’ introduced by practitioners. In the paper, the reader finds three highly convergent derivations of the muon noise. But he knows nothing about other possible derivations elaborated during the investigation but finally put aside as ‘unconvincing’ and not mentioned in the paper. Moreover, all other things being equal concerning the reliability evaluations of the methods involved in a given derivation, the fact that the result of this derivation appears to be very far from the results of several others, can itself work – in cases where the ‘non conformist’ derivation under scrutiny is not based on already well-established approaches – as a reason to reject this derivation as unconvincing. This is another kind of operation through which the Ris can be said to be ‘mutually adjusted’ in the course of the construction of a unique emergent R (in this case: because one of the Ris does not fit with the others, it is eliminated – which means that the imputation of the discrepancy is directed toward its derivation: both the particular Ri and its derivation are discarded altogether as ‘unreliable’, ‘too uncertain’, etc.).
- 15.
- 16.
This is just to sketch the general principle of some possible implications. In the present case, R is not meant to become an invariant physical quantity that will be subsequently measured by multiple instruments. Moreover, the kinds of implications just mentioned will only exist under the condition that the difference between R = 11 and R = 9 indeed makes a difference with respect to the aims of the investigation. In our example, this condition would primarily mean that the difference between 11 and 9 would engender, at the level of the two numbers of the NC-candidates left after the subtraction of the pseudos, a difference that would lead to cross the frontier between a ‘yes’ and a ‘no’ answer to the question of the experimental detection of the NCs.
- 17.
As we saw, the jump can appear more or less creative according to the case. In the case of the ‘Muon noise’ module, the jump would certainly not be perceived as creative by anybody. Actually, it is even difficult to see that there is any jump. This could lead to discard this example as a good means to give credit to the general epistemological point at stake, namely that the R must be considered as a significantly different result with respect to the Ris. Indeed, I concede that other examples would help to understand better and reinforce the point (see just below for references to examples in the present book).
But my strategy has precisely been to show that even in cases in which the passage from the Ris to the emergent R might seem to be completely automatic and uniquely imposed, this passage nevertheless involves operations of conversion that have to be recognized. Indeed, the fact that they are not recognized – or in other words the reference to configurations in which they are almost invisible like in the example of the ‘Muon noise’ module – is precisely what fuels the no-miracle argument that lurks behind the robustness scheme of the three arrows converging on one unique R and makes it appear so convincing (see Section 10.3). Maybe, in some historical cases, we can talk as if the operations of conversion involved are not significant (are indifferent with respect to certain aims). But in order to be in a position to draw this conclusion legitimately, we have first of all to recognize the very existence of such operations and examine the kind of work they accomplish in each case.
For other examples developed in this volume which could help to understand better and reinforce the point here put forward, see notably the Chapter 9 of Stegenga Section 9.4 (about the transmission of a virus in epidemiology), the contribution of Trizio Chapter 4, Section 4.3 and the article of Allamel-Raffin and Gangloff Chapter 7, Section 7.6 (about the production of maps in astronomy). Through the analyses proposed in the latter article in particular, we clearly see how the images first obtained with different kinds of telescopes have to be manipulated and transformed before they can appear ‘essentially similar’ one to the others. As a result, a new and still questionable couple ‘derivation-result’ is viewed to be ‘in essential agreement’ with more ancient and already taken-as-established ones. This harmony works as an argument in favor of the new derivation-result couple under discussion. Because the new result is seen as ‘the same’ as already taken-as-robust old ones, it follows that, jointly, the new derivation is taken as solid and the new derived result is taken as robust. Once this has been achieved, the situation is re-described as: multiple derivations lead to one and the same invariant result. But as soon as we examine the details of the historical process, we find that the new result and the ancient ones were not immediately ‘the same’ from the start. The ‘initial’ images indeed have been transformed, through certain specifiable operations, in order to become comparable.
- 18.
Regarding this point, the example of the 1974 paper about NC is interesting. In my account above, I simplified the presentation in many respects. One of these is that the actual analyses of the 1974 article are in fact constituted of two parallel investigations: the one devoted to the neutrino-induced interaction (on which I have exclusively focused above), and another one devoted to the similar anti-neutrino-induced interaction (i.e., an anti-neutrino interacts with a nucleon, leading to a positive muon and a shower of hadrons). Essentially the same treatment is applied to one and the other case (or in other words: the anti-neutrino case can be reconstructed through essentially the same structural architecture than the one I sketched for the neutrino case). No doubt, the harmony obtained, in terms of the totalizing outputs of the different modules on each floor of the architecture, between the neutrino case on the one hand and the anti-neutrino case on the other hand, also contributes, to the practitioners’ eyes, to reinforce the confidence in the final conclusion of the chapter in favor of the plausibility of weak neutral currents. Now, this reinforcement configuration, primarily based on treatment similarities (by opposition to different, independent treatments), does not correspond to a Wimsattian robustness scheme.
- 19.
It is worth noting that such a characterization of the situation is a conceptualization of the analyst (the historian or the philosopher of science), and a reconstruction which, with respect to its faithfulness to real cases, might be difficult to establish and remains highly conjectural by nature. This is because crucial ingredients of the story which are supposed to determine the ‘fluxes of solidity’, such as “the ‘solidity values’ that are initially attributed by practitioners”, “an argument first viewed as fragile taken in isolation”, “an argument already taken from the beginning as especially solid” and other appreciations of this kind, most of the time remain tacit and opaque to practitioners themselves (this is what I called the opacity of experimental practices with respect to description and justification). See Soler (2011). Hence they are not the kind of things to which the analyst has a transparent and unproblematic access. Actually, even their very existence can be questioned. It is discussable that these kinds of ingredients, postulated by the analyst in order to make a historical episode more understandable, can be equated to well-determined empirical facts (typically to well-defined stable states of the mind of each real subject of knowledge, or to a collective ‘tacit basis’ shared by the members of a scientific community). In any case, practitioners are usually not well-aware of such states, and when the sociologist of science asks them questions about their confidence in this or that ingredient of their science, the answers are not something like unproblematic numbers or unambiguous sentences that the analyst could immediately identify as such, without any discussion, to the ‘solidity values’ according to which real practitioners indeed rated, in the actual historical sequence, the different ingredients involved.
- 20.
This is congruent with Stegenga’s conclusion in this volume: “robustness-style arguments do not tell us what to believe in situations of evidential discordance.”
- 21.
Realist and inevitabilist commitments, although analytically distinguishable, very often go hand in hand concretely, since most of the time, inevitabilism is endorsed as a result of a realist stance: Such or such ingredient of our science is thought to be inevitable because it is taken as a faithful description of a bit of a unique world which is what it is once for all independently of scientists. See Soler (2008a, Section 3).
- 22.
On this constitutive role, see Pickering (1984) for more elements.
- 23.
This is perhaps even possible on the basis of fixed derivations (although this will certainly appear more questionable), if we admit that the totalizing unique output is a calibrating re-description that is not uniquely imposed by the multiple results obtained as the outputs of the lower-levels sub-modules.
- 24.
Thomas Nickles’ chapter – through an analysis which, although inspired by a quite different literature from the one which inspired my own chapter, also seeks to study the implications of the structural features of the humanly designed epistemic systems – draws congruent conclusions about correspondence realism and contingentism.
References
Allamel-Raffin, Catherine. 2005. “De l’intersubjectivité à l’interinstrumentalité. L’exemple de la physique des surfaces.” Philosophia Scientiae 9(1):3–31.
Bachelard, Gaston. 1927. Essai sur la connaissance approchée. Vrin; 6th ed., 1987.
Benvenutti, A. 1974. “Observation of Muonless Neutrino-Induced Inelastic Interactions.” Physical Review Letters 32(14):800–3.
Buchwald, Jed Z. 2006. “Discrepant Measurements and Experimental Knowledge in the Early Modern Era.” Archives for History of Exact Sciences 61:1–85.
Galison, Peter. 1983. “How the First Neutral-Current Experiments Ended.” Review of Modern Physics 55(2):477–509.
Galison, Peter. 1997. Image and Logic, A Material Culture of Microphysics. Chicago: The University of Chicago Press.
Hacking, Ian. 1990. The Taming of Chance (Ideas in Context). Paperback.
Hacking, Ian. 1992. “The Self-Vindication of the Laboratory Sciences.” In Science as Practice and Culture, edited by A. Pickering, 29–64. Chicago and London: The University of Chicago Press.
Hacking, Ian. 1999. The Social Construction of What? Cambridge, MA: Harvard University Press.
Hacking, Ian. 2000. “How Inevitable are the Results of Successful Science?” Philosophy of Science 67:58–71.
Hasert, F.J. et al. 1973a. “Search for Elastic Muon-Neutrino Electron Scattering.” Physical Letters 46B:121–4.
Hasert, F.J. et al. 1973b. “Observation of Neutrino-Like Interactions Without Muon or Electron in the Gargamelle Neutrino Experiment.” Physical Letters 46B:138–40.
Hasert, F.J. et al. 1974. “Observation of Neutrino-Like Interactions Without Muon or Electron in the Gargamelle Neutrino Experiment.” Nuclear Physics B 73:1–22.
Kuhn, Thomas. 1970. The Structure of Scientific Revolutions. 2nd ed. Chicago: The University of Chicago Press.
Kuhn, Thomas. 1973. “Objectivity, Value Judgment, and Theory Choice.” In The Essential Tension, Selected Studies in Scientific Tradition and Change, 320–39. The University of Chicago Press, 1977.
Pickering, A. 1984. Constructing Quarks, a Sociological History of Particle Physics. Chicago and London: The University of Chicago Press.
Rousset, André. 1996. Gargamelle et les Courants Neutres, Témoignage Sur Une Découverte Scientifique. Presses de l’Ecole des Mines de Paris.
Soler, Léna. 2008a. “Are the Results of Our Science Contingent or Inevitable? Introduction of a Symposium Devoted to the Contingency Issue.” Studies in History and Philosophy of Science 39:221–29.
Soler, Léna. 2008b. “Revealing the Analytical Structure and Some Intrinsic Major Difficulties of the Contingentist/Inevitabilist Issue.” Studies in History and Philosophy of Science 39:230–41.
Soler, Léna. 2008c. “The Incommensurability of Experimental Practices: The Incommensurability of What? An Incommensurability of the third-type?” In Rethinking Scientific Change and Theory Comparison. Stabilities, Ruptures, Incommensurabilities? edited by L. Soler, H. Sankey, and P. Hoyningen, 299–340. Springer, Boston Studies for Philosophy of Science.
Soler, Léna. 2011. “Tacit Aspects of Experimental Practices: Analytical Tools and Epistemological Consequences.” European Journal for the Philosophy of Science (EJPS) 1(3):394–433 (Spécial issue directed by the Society for Philosophy of Science in Practice, Mieke Boon, Hasok Chang, Rachel Ankeny et Marcel Boumans).
Soler, Léna. 201X. “A General Structural Argument in Favor of the Contingency of Scientific Results.” In Science as it Could Have Been. Discussing the Contingent/Inevitable Aspects of Scientific Practices, edited by Léna Soler, Emiliano Trizio, and Andrew Pickering. In progress.
Wimsatt, William. 1981. “Robustness, Reliability and Overdetermination.” In Scientific Inquiry and the Social Sciences, edited by M.B. Brewer and B.E. Collins, 125–63. San Francisco, CA: Jossey-Bass. Reprinted in Re-Engineering Philosophy for Limited Beings, Piecewise Approximations to Reality, 43–71. Cambridge, MA, and London, England: Harvard University Press, 2007. Reprinted in this volume, Chapter 2.
Acknowledgements
I am especially indebted to Emiliano Trizio for multiple helpful feedbacks and extended discussions about the claims of this chapter. I am also grateful to Cathy Dufour, Thomas Nickles and Jacob Stegenga for their useful comments. Finally, many thanks to T. Nickles, J. Stegenga and E. Trizio for their corrections and suggestions of improvement concerning the English language. The end result is, of course, my own responsibility!
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2012 Springer Science+Business Media B.V.
About this chapter
Cite this chapter
Soler, L. (2012). Robustness of Results and Robustness of Derivations: The Internal Architecture of a Solid Experimental Proof. In: Soler, L., Trizio, E., Nickles, T., Wimsatt, W. (eds) Characterizing the Robustness of Science. Boston Studies in the Philosophy of Science, vol 292. Springer, Dordrecht. https://doi.org/10.1007/978-94-007-2759-5_10
Download citation
DOI: https://doi.org/10.1007/978-94-007-2759-5_10
Published:
Publisher Name: Springer, Dordrecht
Print ISBN: 978-94-007-2758-8
Online ISBN: 978-94-007-2759-5
eBook Packages: Humanities, Social Sciences and LawPhilosophy and Religion (R0)