Abstract
In our view, data confirm a hypothesis just in case they increase its probability; they constitute evidence for one hypothesis vis-à-vis others just in case they are more probable on it than on its available rivals. In subsequent chapters, we go on to clarify and amplify the confirmation/evidence distinction. Before doing so, however, we need to consider various objections that might be made, not to the distinction itself but to the way in which we have formulated its principal elements. Four of these objections are standard in the literature. The first, third, and fourth raise questions concerning our analyses of both confirmation and evidence; the second has to do more narrowly with the application of Bayesian methods. Each suggests a different way in which our intentions in this monograph might be misunderstood.
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.
See, for example, Kitcher (2001, pp. 29–41), whose discussion itself runs together the two concepts we distinguish. “We lack an analysis that will reveal exactly which claims are justified (to what degree) by the evidence available at various stages” (p. 30). The distinction we make implies that no such analysis is possible. Kitcher does not himself advance the view that scientific claims cannot reasonably be appraised, but his discussion is so seamless and subtle that it is often difficult to know which side of the debate he is on.
- 2.
Kuhn (1962/1970/1996).
- 3.
Kitcher (1992, pp. 75–76).
- 4.
“Naturalism” in epistemology to this point remains little more than a program, with few concrete results or basic insights other than those of Tversky and Kahneman mentioned in the preceding chapter. Hatfield (1990) exposes some of the difficulties when the program is actually pursued rigorously, as it was by Helmholtz in his theory of spatial perception.
- 5.
Indeed, he thinks that both traditional and naturalized perspectives have a place in epistemology.
- 6.
See Our Two Accounts and Interpretations of Probability in Chap. 2.
- 7.
Kitcher (2001, p. 37, n. 5). It illustrates the ambiguity of “probability” that what Barnard takes as the misleading character of the expression “the [measurable] probability of a hypothesis” is itself the flip side of the claim that no hypothesis is ever more than merely “probable” [i.e., might be mistaken].
- 8.
During the late 19th century, there were two mutually exclusive and jointly exhaustive theories of the nature of light, the corpuscular and wave theories. Now we know that this way of dividing the alternatives was mistaken, and that light has a wave-particle duality. The more we know about light, the more finely we should be able to partition possible competing theories.
- 9.
Or some other epistemic-pragmatic criterion. See Rosenkrantz in Earman (1983) for an argument in behalf of the “evidential value” of simplicity. See Bandyopadhyay and Brittan (2001) for a survey of several criteria of model-selection and an argument for adopting one in particular. See also Forster (2000) for some novel proposals.
- 10.
- 11.
See Pollock and Cruz (1999, pp. 101ff).
- 12.
- 13.
Pollock and Cruz (1999, p. 105), second emphasis ours.
- 14.
The third assumption is often known as “internalism,” that a person must be in possession of the grounds on which a belief is held if she is justified in believing it. It is captured in our Bayesian account of confirmation. But our account of evidence is “externalist,” i.e., whether or not data constitute evidence is independent of what an agent knows or believes.
- 15.
The line of argument here has been indicated most clearly by Donald Davidson. See, for example, his (1982). Very possibly Wittgenstein had something like this line of argument in mind at Tractatus 5.1362 when he writes “(‘A knows that p is the case’, has no sense if p is a tautology).” Pears and McGuiness translation.
- 16.
This is to take a “free-logical” way with logical truth. See Meyer and Lambert (1968).
- 17.
Although we do so for other reasons. See the section on Absolute and Incremental Confirmation in Chap. 2.
- 18.
In this we follow Howson and Urbach (2006, p. 287): “In our [Bayesian] account there is nothing that demands what [are] taken as data in one inductive inference cannot be regarded as problematic in a later one. Assigning probability 1 to some data on one occasion does not mean that on all subsequent occasions it needs to be assigned probability 1.” See also Levi (1967, p. 209). We use conditionalization as an eminently clear way of making a distinction between confirmation and evidence, not to defend it as a principle of rationality. For objections to it, and to the certainty and fallibility models, see Bacchus, et al. (1990). It should be noted that in fact more and more scientific testing does take data uncertainty into account, and our account is easily accommodated to it. More important conceptually is the possibility of misleading evidence, discussed in Chap. 8.
- 19.
What follows is indebted to Kyburg’s work, in particular to Kyburg (1984). It remains a scandal that so many philosophers of science ignore both the importance of error and the sophisticated statistical techniques that have been developed to deal with it.
- 20.
Ibid., p. 91.
- 21.
As Kyburg notes, one does not have to be a statistician to know that “large errors are much less frequent than small errors; that errors tend to average out; and that [in day to day carpentry, say] an error of a quarter of an inch is relatively infrequent.” Ibid., p. 257. In the case of at least one of the authors, if in a series of measurements of a board to be cut we get the same value for two consecutive measurements, we call it good and proceed to saw away.
- 22.
Patrick Suppes (1984, p. 215), like Davidson and Kyburg one of our mentors, issues a caution that must be taken seriously with respect to our own and others’ examples. “Published articles about experiments may make it seem as if the scientist can in a simple, rational way relate data to theory in the finished form commonly found in the scientific literature. In my own experience, nothing can be further from the truth. In the setting up of experiments and in the analysis of the results there are countless informal decisions of a practical and intuitive kind that must be made”.
- 23.
I.e., measurements that are outside the predicted interval of values for the quantity in question, not measurements that are incompatible with the absolute value in the hypothesis, for only occasionally, and in a statistically predictable way, will measurements coincide with the absolute value.
References
Bacchus, F., Kyburg, H., & Thalos, M. (1990). Against conditionalization. Synthèse, 85, 475–506.
Bandyopadhyay, P., & Brittan, G. (2001). Logical consequence and beyond: a look at model selection in statistics. In J. Woods & B. Hepburn (Eds.), Logical consequence. Oxford: Hermes Science.
Barnard, G. (1949). Statistical inference. Journal of the Royal Statistical Society: Series B, 11, 115–149.
Bozdogan, H. (1987). Model selection and akaike information criterion (AIC)—the general theory and its analytical extensions. Psychometrika, 52, 345–376.
Chisholm, R. (1957). Perceiving. Ithaca, NY: Cornell University Press.
Davidson, D. (1982). Rational animals. Dialectica, 36(4), 317–327.
Earman, J. (Ed.). (1983). Testing scientific theories, Minnesota studies in the philosophy of science (Vol. X). Minneapolis: University of Minnesota Press.
Fishman, Y., & Boudry, M. (2013). Does science presuppose naturalism (or anything at all)? Science & Education, 22, 921–947.
Forster, M. (2000). The new science of simplicity. In H. Kreuzenkamp, et al. (Eds.), Simplicity, inference, and economic modeling. Cambridge: Cambridge University Press.
Hatfield, G. (1990). The natural and the normative. Cambridge, MA: MIT Press.
Howson, C., & Urbach, P. (2006). Scientific reasoning: the bayesian approach (3rd ed.). Chicago and LaSalle, IL: Open Court Publishing.
Jeffrey, R. (1990). The logic of decision. Chicago: University of Chicago Press.
Kitcher, P. (1992). The naturalists return. The Philosophical Review, 101(1), 53–114.
Kitcher, P. (2001). Science, truth, and democracy. Oxford: Oxford University Press.
Kuhn, T. (1962/1970/1996). The structure of scientific revolutions. Chicago: University of Chicago Press.
Kyburg, H. (1974). The foundations of scientific inference. Dordrecht: D. Reidel.
Kyburg, H. (1984). Theory and measurement. Cambridge: Cambridge University Press.
Lele, S., & Taper, M. (2012). Information criteria in ecology. In Hastings & Gross (Eds.), Sourcebook in theoretical ecology. Berkeley: University of California Press.
Levi, I. (1967). Probability kinematics. British Journal for the Philosophy of Science, 18, 200–205.
Lewis, C.I. (1946). An analysis of knowledge and valuation. LaSalle, IL: Open Court.
Meyer, R., & Lambert, K. (1968). Universally free logic and standard quantification theory. Journal of Symbolic Logic, 33, 8–26.
Pollock, J., & Cruz, J. (1999). Contemporary theories of knowledge (2nd ed.). Totiwa, NJ: Rowman and Littlefield.
Rosenkrantz, R. (1983). Why Glymour is a Bayesian. In Earman.
Skyrms, B. (1986). Choice and chance, (3rd ed.). Belmont, CA: Dickenson.
Suppes, P. (1984). Probabilistic metaphysics. Oxford: Blackwell.
Author information
Authors and Affiliations
Corresponding author
Rights and permissions
Copyright information
© 2016 The Author(s)
About this chapter
Cite this chapter
Bandyopadhyay, P.S., Brittan, G., Taper, M.L. (2016). Initial Difficulties Dispelled. In: Belief, Evidence, and Uncertainty. SpringerBriefs in Philosophy(). Springer, Cham. https://doi.org/10.1007/978-3-319-27772-1_4
Download citation
DOI: https://doi.org/10.1007/978-3-319-27772-1_4
Published:
Publisher Name: Springer, Cham
Print ISBN: 978-3-319-27770-7
Online ISBN: 978-3-319-27772-1
eBook Packages: Religion and PhilosophyPhilosophy and Religion (R0)