Abstract
It is often argued that while biases routinely influence the generation of scientific theories (in the ‘context of discovery’), a subsequent rational evaluation of such theories (in the ‘context of justification’) will ensure that biases do not affect which theories are ultimately accepted. Against this line of thought, this paper shows that the existence of certain kinds of biases at the generation-stage implies the existence of biases at the evaluation-stage. The key argumentative move is to recognize that a scientist who comes up with a new theory about some phenomena has thereby gained an unusual type of evidence, viz. information about the space of theories that could be true of the phenomena. It follows that if there is bias in the generation of scientific theories in a given domain, then the rational evaluation of theories with reference to the total evidence in that domain will also be biased.
This is a preview of subscription content, log in via an institution to check access.
Similar content being viewed by others
Notes
Longino‘s elements are (1) research practices, (2) research questions, (3) research data, (4) specific background assumptions, and (5) general background assumptions.
Reiss and Sprenger’s stages are (1) choosing a research problem, (2) gathering evidence, (3) accepting a theory, and (4) the proliferation and application of results. Reiss and Sprenger are concerned with threats to scientific objectivity due to the influence of various moral, social and political values, but these values are more or less equivalent to what I define as a ‘bias’ below (see §2). Another example of discussions of objectivity and values that steers clear of theory generation is Elliot (2017).
Similar arguments are often advocated by scientists themselves. Witness, for example, Carl Sagan in his popular television program Cosmos: “There are many hypotheses in science which are wrong. That’s perfectly all right […] To be accepted, new ideas must survive the most rigorous standards of evidence and scrutiny” (Malone 1990).
Note that my definition of ‘bias’ here differs from Antony’s ‘empiricist’ definition of bias as “possession of belief or interest prior to investigation” (Antony 1993: 188). For the purposes of this paper, a more ‘operational’ definition is appropriate, i.e. one on which bias can be identified in terms of the agent’s dispositions to behave in certain ways rather than her belief or interests (which may or may not be manifested in the agent’s behavior). Furthermore, it is not at all clear whether, or how, Antony’s definition could be made to subsume implicit biases, which are usually taken to be non-doxastic states (i.e. not beliefs) and which clearly need not line up with the agent’s interests.
Thus, in what follows, the term ‘bias’ should always be taken to refer to biases that are grounded in contextual values.
Many thanks to an anonymous reviewer for suggesting that I contrast the Implication Thesis with this more innocuous point.
Indeed, it seems to me that this point would not threaten the Confinement Defense of scientific objectivity at all, at least not on a plausible construal thereof. After all, a proponent of the Confinement Defense could respond that, even granting this point, each individual theory that has in fact been generated and evaluated (positively or negatively) would be evaluated in just the same way regardless of whether it and its competitors were generated in a biased way. In particular, such a proponent could argue that a theory that is sufficiently positively evaluated to be accepted would still be evaluated in just the same way regardless of whether its generation was biased or unbiased. Thus the fact that only theories that have been generated could be (positively) evaluated would not threaten the central contention of the Confinement Defense that biases in theory generation do not undermine our reasons for accepting the theories that we do in fact accept.
Another argument that comes close to the one I will make below is sketched briefly by Elliot and McKaughan (2009: 607–608), who argue that proposing new theories “can transform what appeared to be irrelevant facts into crucial pieces of evidence” (Elliot and McKaughan 2009: 608). However, it is not clear from Elliot and McKaughan’s brief discussion what it is for evidence to be ‘transformed’ in their sense, especially since they appear to deny that this type of transformation “alter[s] the evidential relationship between the available theories and data” (Elliot and McKaughan 2009: 608). By contrast, I argue below that proposing new theories can make it rational to evaluate old theories less favorably (e.g., by it becoming rational to assign a lower probability to them), even when there is no change in the relevant empirical data. In this sense, pace Elliot and McKaughan, I maintain that proposing new theories can alter the evidential relationship between available theories and data.
This idea is reminiscent of some early conceptions of Inference to the Best Explanation (IBE), where the evaluative step merely renders the comparative verdict that one theory provides a better explanation than available alternatives (e.g., Harman 1965; Thagard 1978). However, as I explain below, Okruhlik’s model is much more radically comparative than standard conceptions of IBE.
Thus Okruhlik would have to deny, for example, that the best explanation of one’s evidence is probably true. On her view, assuming she accepts some form of IBE (see previous footnote), we could at most assert that it is more probably true, or perhaps more approximately true, than available rival explanations.
Douven (2017b) refers to this problem as the asymmetry problem. On Douven’s description of the problem, the issue is that most formulations of IBE “license an inference to an absolute verdict—that a given hypothesis is true—from what will typically only be a relative judgment, namely, that the hypothesis is the best explanation among those on the table” (Douven 2017b: 9). It is perhaps worth noting that some influential conceptions of IBE propose to avoid this problem by including an ‘absolutist’ requirement on the conditions for IBE to the effect that the inferred explanation should not merely be the best, but also “satisfactory” (Musgrave 1988) or “good enough” (Lipton 2004); see also Dellsén (2017, 2018) for a different approach to the problem.
The type of situation described here is in a sense the converse of the current situation in particle physics, in which repeated failed attempts to come up with a plausible alternative to string theory has arguably contributed to scientists becoming quite confident that no such alternative exists (Dawid 2013; Dawid et al. 2015).
I choose to focus on BCT in what follows in part because it is by far the most widely endorsed framework for rational theory evaluation among philosophers of science (see footnote 13); in part because BCT clearly provides the means to evaluate theories in an absolute—i.e. not merely comparative—manner (in contrast to Okruhlik’s ‘irreducably comparative’ model); and in part because BCT has well-known prima facie difficulties in handling the epistemic impact of new theories.
Any general Bayesian solution must at least allow for this possibility. Indeed, this is exactly the sort of situation in which the problem of new evidence arises (see, e.g. Earman 1992: 196–197).
To see this, note first that since Cnew is (by construction) incompatible with Tnew, the probability of their disjunction is equal to the sum of their individual probabilities: Pafter(Tnew or Cnew) = Pafter(Tnew) + Pafter(Cnew). So the situation we are focusing on is one where:
-
(1)
Pbefore(C) > Pafter(Tnew) + Pafter(Cnew)
Now, as noted, the probability axioms demand that the probabilities before and after both sum to unity, i.e. that:
-
(2)
Pbefore(T1) + ··· + Pbefore(Tk) + Pbefore(C) = Pafter(T1) + ··· + Pafter(Tk) + Pafter(Tnew) + Pafter(Cnew) = 1
(1) and (2) jointly entail that Pbefore(T1) + ··· + Pbefore(Tk) > Pafter(T1) + ··· + Pafter(Tk), as desired.
-
(1)
Indeed, as noted in footnote 16, I have chosen to discuss how to model this type of evidence within BCT in part because it is initially not at all obvious that BCT could accommodate evidence of this type at all.
Recall the definition of ‘generation-bias’ at the beginning of Sect. 2.
Or at a minimum to try to minimize the effects of forces that cause scientific communities to become more homogenous in this regard, such as what Holman and Bruner (2017) call “industrial selection”.
References
Antony, L. (1993). Quine as a feminist: The radical import of naturalized epistemology. In L. Antony & C. Witt (Eds.), A mind of one’s own (pp. 185–225). Westview: Boulder.
Barnes, B. (1972). Sociological explanation and natural science: A Kuhnian reappraisal. Archives Européens de Sociologie, 13, 373–393.
Chihara, C. S. (1987). Some problems for Bayesian confirmation theory. The British Journal for the Philosophy of Science, 38, 551–560.
Dawid, R. (2013). String theory and the scientific method. Cambridge: Cambridge University Press.
Dawid, R., Hartmann, S., & Sprenger, J. (2015). The no alternatives argument. The British Journal for Philosophy of Science, 66, 213–234.
Dellsén, F. (2017). Abductively robust inference. Analysis, 77, 20–29.
Dellsén, F. (2018). The heuristic conception of inference to the best explanation. Philosophical Studies, 175, 1745–1766.
Douglas, H. (2000). Inductive risk and values in science. Philosophy of Science, 67, 559–579.
Douglas, H. (2009). Science, policy, and the value-free ideal. Pittsburgh: University of Pittsburgh Press.
Douven, I. (2017a). Abduction. In E. N. Zalta (Ed.), The stanford encyclopedia of philosophy (Summer 2017 Edition). Retrieved November 5, 2019
Douven, I. (2017b). Inference to the best explanation: What is it? And why should we care? In K. McCain & T. Poston (Eds.), Best explanations: New essays on inference to the best explanation (pp. 7–24). Oxford: Oxford University Press.
Earman, J. (1992). Bayes or bust? A critical examination of Bayesian confirmation theory. Cambridge: MIT Press.
Einstein, A. (1905). Zur Elektrodynamik bewegter Körper. Annalen der Physik, 17, 891–921.
Elliot, K. C. (2017). A tapestry of values. Oxford: Oxford University Press.
Elliot, K. C., & McKaughan, D. J. (2009). How values in scientific discovery and pursuit alter theory appraisal. Philosophy of Science, 76, 598–611.
Fischer, R. A. (1959). Smoking: The cancer controversy; some attempts to assess the evidence. Edinburgh: Oliver and Boyd.
Glymour, C. (1980). Theory and evidence. Princeton: Princeton University Press.
Godfrey-Smith, P. (2003). Theory and reality: An introduction to the philosophy of science. Chicago: University of Chicago Press.
Harman, G. (1965). The inference to the best explanation. The Philosophical Review, 74, 88–95.
Hempel, C.G. (1965). Science and human values. In Aspects of scientific explanation and other essays in the philosophy of science (pp. 81–96). New York: The Free Press.
Hempel, C. G. (1966). Philosophy of natural science. Englewood Cliffs: Prentice-Hall.
Holman, B., & Bruner, J. (2017). Experimentation by industrial selection. Philosophy of Science, 84, 1008–1019.
Intemann, K. (2005). Feminism, underdetermination, and values in science. Philosophy of Science, 72, 1001–1012.
Intergovernmental Panel on Climate Change. (2014). Climate change 2014 synthesis report: Summary for policymakers. Retreived November 5, 2019 from http://www.ipcc.ch/pdf/assessment-report/ar5/syr/AR5_SYR_FINAL_SPM.pdf.
Kantorovich, A. (1993). Scientific discovery—Logic and tinkering. New York: State University of New York Press.
Knorr-Cetina, K. (1981). The manufacture of knowledge. Oxford: Pergamon Press.
Kuhn, T. S. (1962). The structure of scientific revolutions. Chicago: University of Chicago Press.
Kuhn, T. S. (1977). Objectivity, value judgment, and theory choice. In The essential tension (pp. 320-39). Chicago: University of Chicago Press.
Lipton, P. (2004). Inference to the best explanation (2nd ed.). New York: Routledge.
Lloyd, E. A. (1993). Pre-theoretical assumptions in evolutionary explanations of female sexuality. Philosophical Studies, 69, 139–153.
Lloyd, E. A. (2005). The Case of the female orgasm: Bias in the science of evolution. Cambridge: Harvard University Press.
Longino, H. (1990). Science as social knowledge: Values and objectivity in scientific inquiry. Princeton: Princeton University Press.
Longino, H. (1994). Gender and racial biases in scientific research. In K. Shrader-Frechette (Ed.), Ethics of scientific research (pp. 139–151). Lanham: Rowman and Littlefield.
Longino, H. (1996). Cognitive and non-cognitive values in science: Rethinking the dichotomy. In L. H. Nelson & J. Nelson (Eds.), Feminism, science, and the philosophy of science (pp. 39–58). Dordrecht: Springer.
Longino, H. (2002). The fate of knowledge. Princeton, NJ: Princeton University Press.
Longino, H., & Doell, R. (1983). Body, bias, and behavior: A comparative analysis of reasoning in two areas of biological science. Signs: Journal of Women in Culture and Society, 9, 206–227.
Maher, P. (1995). Probabilities for new theories. Philosophical Studies, 77, 103–115.
Malone, A. (1990). Cosmos: A personal voyage (Episode 4: Heaven and Hell). Los Angeles: PBS and KCET.
Musgrave, A. (1988). The ultimate argument for scientific realism. In R. Nola (Ed.), Relativism and realism in science (pp. 229–252). Dordrecht and Boston: Kluwer.
Norton, J. (2018). The material theory of induction. Unpublished manuscript. Retreived November 5, 2019 from https://www.pitt.edu/~jdnorton/homepage/research/ind_material.html.
Okruhlik, K. (1994). Gender and the biological sciences. Canadian Journal of Philosophy, 24, 21–42.
Reichenbach, H. (1938). Experience and prediction. Chicago: University of Chicago Press.
Reiss, J., & Sprenger, J. (2017). Scientific objectivity. In Zalta, E. N. (Ed.), The stanford encyclopedia of philosophy (Winter 2017 edition). Retreived November 5, 2019 from https://plato.stanford.edu/archives/win2017/entries/scientific-objectivity/.
Rudner, R. (1953). The scientist qua scientist makes value judgments. Philosophy of Science, 20, 1–6.
Solomon, M. (2001). Social empiricism. Cambridge, MA: MIT Press.
Stolley, P. D. (1991). When genius errs: R. A. Fischer and the lung cancer controversy. American Journal of Epidemiology, 133, 416–425.
Strevens, M. (2017). Notes on Bayesian confirmation theory. Unpublished manuscript. Retreived November 5, 2019 from http://www.nyu.edu/classes/strevens/BCT/BCT.pdf.
Symons, D. (1979). The evolution of human sexuality. New York: Oxford University Press.
Thagard, P. (1978). The best explanation: Criteria for theory choice. Journal of Philosophy, 75, 76–92.
van Fraassen, B. C. (1989). Laws and Symmetry. Oxford: Clarendon Press.
Wenmeckers, S., & Romeijn, J. W. (2016). New theory about old evidence. Synthese, 193, 1225–1250.
Author information
Authors and Affiliations
Corresponding author
Additional information
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Rights and permissions
About this article
Cite this article
Dellsén, F. The epistemic impact of theorizing: generation bias implies evaluation bias. Philos Stud 177, 3661–3678 (2020). https://doi.org/10.1007/s11098-019-01387-w
Published:
Issue Date:
DOI: https://doi.org/10.1007/s11098-019-01387-w