(Mis)Understanding scientific disagreement: Success versus pursuit-worthiness in theory choice

https://doi.org/10.1016/j.shpsa.2020.10.005Get rights and content

Highlights

  • Examines core sources of divergence among scientists in contexts of theory choice.

  • Clarifies the nature of theory choice and status of alternative theoretical virtues.

  • What seem like direct disputes about explanatory or epistemic value often are not.

  • Divergent theory choice often stems from differing judgments of pursuit-worthiness.

  • These concern which theory will be best or addresses more crucial data or questions.

Abstract

Scientists often diverge widely when choosing between research programs. This can seem to be rooted in disagreements about which of several theories, competing to address shared questions or phenomena, is currently the most epistemically or explanatorily valuable—i.e. most successful. But many such cases are actually more directly rooted in differing judgments of pursuit-worthiness, concerning which theory will be best down the line, or which addresses the most significant data or questions. Using case studies from 16th-century astronomy and 20th-century geology and biology, I argue that divergent theory choice is thus often driven by considerations of scientific process, even where direct epistemic or explanatory evaluation of its final products appears more relevant. Broadly following Kuhn's analysis of theoretical virtues, I suggest that widely shared criteria for pursuit-worthiness function as imprecise, mutually-conflicting values. However, even Kuhn and others sensitive to pragmatic dimensions of theory ‘acceptance’, including the virtue of fruitfulness, still commonly understate the role of pursuit-worthiness—especially by exaggerating the impact of more present-oriented virtues, or failing to stress how ‘competing’ theories excel at addressing different questions or data. This framework clarifies the nature of the choice and competition involved in theory choice, and the role of alternative theoretical virtues.

Introduction

Confronted with competing theories or research programs, even the best scientists often part ways. The major sources of this common divergence in theory choice can be unclear, though—in part because intuitive notions of ‘theory choice’ can obscure several ambiguities. Central among these are distinct views concerning what it means for scientists to ‘choose’ one or another competing theory, and what it means for theories to ‘compete’ with each other, in the first place. Thus, for instance, it is famously contested whether choosing or accepting a theory involves believing that it is (approximately) true, as scientific realists might propose; or whether, as some anti-realists suggest, acceptance is a less narrowly epistemic attitude that involves commitment to use or develop a theory, without necessarily believing it to be true (e.g. van Fraassen, 1980, 2007; Bueno, 1999).1 Likewise, although less frequently discussed, it is unclear whether theories genuinely compete only when they yield mutually inconsistent claims, as in Copernican and Ptolemaic astronomers debating whether the Earth orbits the Sun, rather than vice versa; or whether theoretical ‘competition’ also encompasses more indirect forms of rivalry for resources.

In this light, I will argue below, using case studies from early modern astronomy and 20th-century geology and biology, that one apparently strong explanation for divergent theory choice is often, in fact, extremely misleading. Namely, many cases in which scientists choose different research programs appear to be rooted in disagreements about which of several rival theories, competing to address shared questions or phenomena, is currently the most epistemically or explanatorily valuable. But many such cases are really more so rooted in divergent assessments of which theory will be best down the line, or which addresses the most significant data or questions. In other words, what appear to be disputes about which theory is most successful, given its current explanatory or epistemic value, are often in fact more so disputes about which is most pursuit-worthy. Earlier commentators have offered valuable accounts of pursuit-worthiness (e.g. Laudan, 1977; Whitt, 1990; Achinstein, 1993; Šešelja & Straßer, 2013; Nyrup, 2015, 2019), but without thus arguing that apparent disputes about success often more directly concern pursuit-worthiness.

I draw several interrelated conclusions from this descriptive thesis about the sources of systematic divergence in scientific theory choice. First, certain common conceptions of theory choice or attitudes like acceptance have unexpectedly limited purchase on scientific practice. This includes both narrowly epistemic accounts and some pragmatically-inflected accounts offered by anti-realists. Insofar as many anti-realists do emphasize the relation between theory choice or acceptance and practical commitment to a given theory's ongoing development, my analysis indirectly illustrates the explanatory power of notions of acceptance or theory choice similar to those advocated by many anti-realists. But anti-realist notions of acceptance or theory choice still often encompass both a theory's current value (including its current usefulness) and its prospective value.2 Hence, I also show that such broad appeals to pragmatic dimensions of theory choice have less explanatory purchase on scientific practice than some anti-realists might hope. I do not aim to supplant existing concepts of pragmatic ‘acceptance’, but the notion of pursuit-worthiness that I stress more directly illuminates certain important dynamics of divergence in theory choice.

Second, however, my argument does not just concern the distinction between a theory's present and future value. Appraisals of pursuit-worthiness also encompass considerations of problem salience and methodological approach, that is, researchers' assessments of which questions or data are most significant, and of how to best approach them.3 This relates to the second kind of ambiguity mentioned above, about what it means for theories to be ‘competing’. At the extreme limit, someone who aimed to ‘choose’ between theories in totally disparate fields of inquiry, like quantum mechanics and the Darwinian theory of evolution by natural selection, could rightly be accused of making a basic mistake. These two theories simply are not competing, in the relevant sense, and it therefore makes little sense to ‘choose’ between them. Intuitively, then, theories can truly compete only if they are about the same thing. But the case studies I develop below indicate that reality is less clear-cut than this intuitive gloss suggests. Competing research programs that are ‘about the same thing’ at coarse-grained levels of description still typically emphasize different finer-grained questions, data, or methods. Scientists thus often ‘choose’ more passively between competing theories, simply by stressing different questions, data, or methods. The result is a kind of practical divergence, driven largely by incompletely-overlapping research priorities.

Third, I suggest that widely shared criteria for pursuit-worthiness function as imprecise, mutually-conflicting values or heuristics. This broadly aligns with Kuhn's analysis of theoretical virtues like simplicity or empirical fit as “values” amenable to variable interpretation and application, rather than “rules” that dictate singular choices (Kuhn, 1977, pp. 320–339; see also Kuhn, 1970, pp. 152–154). Notably, Kuhn uses this contrast between values and rules to frame an influential explanation of rational disagreement in theory choice.4 His thought is that the inherent imprecision of values—compounded by tension between different values and the ambiguity of their relative weighting (Kuhn, 1977, pp. 322–324)—allows scientists to disagree about a given theory's merits, even while sharing the same basic standards for theoretical appraisal. Shared virtues only hold anything close to universal appeal if they remain vague or resistant to clear-cut application; hence, valuing ‘simplicity’ in theories is like valuing ‘freedom’ in politics. Kuhn's framework thus lends simultaneous credence to two intuitive but superficially incompatible notions: that there is fairly broad consensus about what makes for a good scientific theory, but that scientific dissensus is still often rational (if not even beneficial).5 Hence, I aim in part to elaborate a basic Kuhnian model, by isolating a revealing trend in the balance of virtues that underlies divergence in theory choice. What appears to be disagreement concerning virtues like simplicity or empirical fit—and Kuhn's analysis can easily reinforce this impression—often more directly concerns virtues like prospective fruitfulness.6 This addresses an important lacuna, if not a serious limitation, in Kuhn's account.

Fourth, and finally, my analysis clarifies the role of alternative theoretical virtues, like the “feminist virtues” that Longino (1995, 1996, 1997) proposes in response to Kuhn. I suggest that virtues like Longino's ontological heterogeneity and complexity of relationship, which she argues stand in tension with standard virtues like simplicity, are often productively understood as guiding researchers towards certain questions and phenomena, rather than as domain-neutral epistemic or explanatory desiderata. Alternative virtues can act as criteria for pursuit-worthiness, then, even if traditional virtues like simplicity mark explanatory success.

The plan of the paper is as follows. In section 2, I distinguish several common concepts of acceptability, adequacy, and pursuit-worthiness, to clarify my own notions of pursuit-worthiness and success. In section 3, I elaborate Kuhn's basic model of rational disagreement arising from differing interpretations or weightings of shared values, arguing that future-oriented virtues like fruitfulness figure more centrally than he seems to suggest. In section 4, I use a case study from geology to show how problem salience and methodological approach also bear on pursuit-worthiness. In section 5, I relate this to the status of alternative theoretical virtues in Barbara McClintock's work on maize genetics. Section 6 then provides a brief summary and conclusion.

Section snippets

Acceptance, success, and pursuit-worthiness

Before turning to analyze concrete cases of divergent theory choice, a brief discussion of terms like ‘pursuit-worthiness’, ‘adequacy’, and ‘acceptance’ is first in order. Disambiguating several common senses of each will help to clarify my notions of pursuit-worthiness and success.

For some commentators, appraisal of a theory's ‘adequacy’ or ‘acceptability’ includes assessment both of its current epistemic or explanatory value and of its prospective value or suitability for ongoing development.

Kuhn's model and its limits in the Copernican case

With these points of terminology and context in tow, we may now turn to case studies. Here it will prove helpful to begin by considering one of Kuhn's examples: 16th-century astronomers faced a choice between Ptolemaic geocentric theories and heliocentric approaches building upon Copernicus's work.

Kuhn hopes to explain divergent choice between Ptolemaic and Copernican systems by appeal to differing interpretations or weightings of shared ‘values’. For instance, astronomers who valued simplicity

Problem salience and methodological approach in the case of sea-floor spreading

Appraisals of pursuit-worthiness encompass not just direct assessments of a theory's future epistemic or explanatory value, but also those of problem salience and methodological approach—that is, judgments about which questions or data are most significant, and how to best approach them. Again, this relates to the question of what it means for theories to be ‘competing’. Alternative research programs that are ‘about the same thing’ at coarse-grained levels of description, and so intuitively

Alternative theoretical virtues in McClintock's maize genetics

Finally, in this light, we may turn to the relation between criteria for pursuit-worthiness and alternative theoretical virtues. Traditional virtues like simplicity can function as generic criteria for assessing a theory's current explanatory or epistemic value. But, as we have seen, they can also act indirectly as criteria for problem selection or methodological approach—e.g. as injunctions to simplify, rather than to ‘choose’ simpler theories. This helps to clarify the role of non-standard

Conclusion

Many cases in which scientists choose different research programs appear to be rooted in disagreements about which of several theories, competing to address shared questions or phenomena, is most successful—i.e. currently most epistemically or explanatorily valuable. But many such cases are in fact more directly rooted in divergent assessments of pursuit-worthiness, concerning which theory will be best down the line, or which research program addresses the most significant data or questions. I

Acknowledgements

I am grateful to Gordon Belot, Tom Ryckman, and three anonymous reviewers for very helpful comments on previous drafts of this paper. Thanks also to Tom Ryckman and Helen Longino for further valuable feedback on an earlier project in which I began to think about many of the issues that I examine here.

References (68)

  • N. Weinberger et al.

    Making sense of non-factual disagreement in science

    Studies In History and Philosophy of Science Part A

    (2020)
  • P. Achinstein

    How to defend a theory without testing it: Niels Bohr and the “logic of pursuit”

    Midwest Studies In Philosophy

    (1993)
  • J.E. Adler

    Belief's own ethics

    (2002)
  • A. Blair

    Tycho brahe's critique of Copernicus and the copernican system

    Journal of the History of Ideas

    (1990)
  • E. Bonatti

    Origin of offsets of the Mid-Atlantic Ridge in fracture zones

    The Journal of Geology

    (1973)
  • S. Bromberger

    On what we know and don't know: Explanation, theory, linguistics, and how questions shape them

    (1992)
  • O. Bueno

    What is structural empiricism? Scientific change in an empiricist setting

    Erkenntnis

    (1999)
  • R. Carnap

    Logical foundations of probability

    (1950)
  • R. Cifelli

    Age relationships of mid-Atlantic Ridge sediments

    Geological Society of America Special Paper

    (1970)
  • R. Cifelli et al.

    Paleogene sediments from a fracture zone of the Mid-Atlantic Ridge

    Journal of Marine Research

    (1968)
  • L.J. Cohen

    An essay on belief and acceptance

    (1992)
  • N. Copernicus

    On the revolutions of the heavenly spheres

  • F. Dellsén et al.

    Disagreement in science: Introduction to the special issue

    Synthese

    (2020)
  • M. Dorato

    Epistemic and nonepistemic values in science

  • H.E. Douglas

    Science, policy, and the value-free ideal

    (2009)
  • C. Elgin

    True enough

    (2017)
  • K. Elliott et al.

    How values in scientific discovery and pursuit alter theory appraisal

    Philosophy of Science

    (2009)
  • M. Ewing et al.

    Sediment distribution in the oceans: The mid-Atlantic Ridge

    The Geological Society of America Bulletin

    (1964)
  • B. van Fraassen

    The scientific image

    (1980)
  • B. van Fraassen

    From a view of science to a new empiricism

  • H. Frankel

    The development, reception, and acceptance of the Vine-Matthews-Morley Hypothesis

    Historical Studies in the Physical Sciences

    (1982)
  • A.I. Goldman

    Knowledge in a social world

    (1999)
  • S.R. Grimm

    Explanatory inquiry and the need for explanation

    The British Journal for the Philosophy of Science

    (2008)
  • J.R. Heirtzler et al.

    Crustal structure of the mid-ocean ridges: 3. Magnetic anomalies over the mid-Atlantic Ridge

    Journal of Geophysical Research

    (1965)
  • C.G. Hempel

    Valuation and objectivity in science

  • D. Hull

    Science as a process

    (1988)
  • E.F. Keller

    A feeling for the organism: The life and work of Barbara McClintock

    (1983)
  • E.F. Keller

    Reflections on gender and science

    (1985)
  • T. Kelly

    Disagreement and the burdens of judgment

  • C. Kelp et al.

    Sustaining a rational disagreement

  • P. Kitcher

    The division of cognitive labor

    Journal of Philosophy

    (1990)
  • A. Koyré

    Galileo and the scientific revolution of the seventeenth century

    Philosophical Review

    (1943)
  • J. Kries

    Die principien der wahrscheinlichkeitsrechnung

    (1886)
  • T.S. Kuhn

    The Copernican revolution: Planetary astronomy in the development of western thought

    (1957)
  • Cited by (8)

    • Pursuitworthiness between daring conservatism and procrastination: Wheeler and the path towards black holes

      2022, Studies in History and Philosophy of Science
      Citation Excerpt :

      Wheeler's storyline across the years that we have considered also saw, for many aspects, the dawning of such situation, in which the claim of something being pursuitworthy seems to have been capable of lasting for dilated durations, without empirical interplay. We can thus get at least some insights on the uses and risks of a notion like pursuitworthiness, as well as on its possible role (underscored in other contexts by Lichtenstein, 2021) in establishing a scenario of non-dialogical disagreement among different approaches, as epitomized today by various quantum gravity “tribes” in conflict (or ignoring and dismissing each other – and this, in a form of collective enterprise such as science, is a form of conflict as well). The impression that is left is that Wheeler's persuasive and argumentative strategies, as seen in the heuristic path examined in this paper, can work as a pharmakon, in its double meaning.

    • Pursuit and inquisitive reasons

      2022, Studies in History and Philosophy of Science
      Citation Excerpt :

      If the theory were true, one would expect the fact constituting the anomaly not to obtain. Despite anomalies serving as disfavoring evidence, Lichtenstein (2021) has recently argued that anomalies provide scientists with reasons for pursuing a theory. Specifically, he argues that the presence of anomalies for a theory provides guidance for future research into the theory.

    • Mechanistic inquiry and scientific pursuit: The case of visual processing

      2022, Studies in History and Philosophy of Science
      Citation Excerpt :

      Our project in this paper is thus to provide such an integrative account by explicating how pursuitworthiness and mechanistic inquiry are linked. While the pursuit literature speaks of theories from rivaling research traditions (e.g. Laudan, 1977; Lichtenstein, 2021), the mechanistic literature largely focuses on models being pursued (Colombo et al., 2015; Hochstein, 2016). For our analysis here we focus on the pursuit of mechanistic models describing entities and their causal interactions (activities) which are responsible for the phenomenon to be explained (Craver & Darden, 2013).

    • Why do experts disagree? The development of a taxonomy

      2023, Public Understanding of Science
    View all citing articles on Scopus
    View full text