Abstract
We (2013, 2014) argued that explanatoriness is evidentially irrelevant in the following sense: Let H be a hypothesis, O an observation, and E the proposition that H would explain O if H and O were true. Then our claim is that Pr(H | O & E) = Pr(H | O). We defended this screening-off thesis (SOT) by discussing an example concerning smoking and cancer. Climenhaga (Philos Sci, forthcoming) argues that SOT is mistaken because it delivers the wrong verdict about a slightly different smoking-and-cancer case. He also considers a variant of SOT, called “SOT*”, and contends that it too gives the wrong result. We here reply to Climenhaga’s arguments and suggest that SOT provides a criticism of the widely held theory of inference called “inference to the best explanation”.
Notes
McCain and Poston (2014) miss this point. They grant that SOT is true in certain cases, but contend that even in those cases explanatoriness is evidentially relevant because it affects what they term “weight of evidence”, which they say bears on a probability’s “resilience”.
We do not mean “realistic in all respects” since we assume that logical omniscience holds in the cases in question. However, this idealization is harmless, for if the assumption of logical omniscience were dropped, then a thesis similar to SOT would hold. See Roche and Sober (2014, p. 195) for discussion.
We assume that explanatoriness can be informative to logically omniscient subjects. This should be uncontroversial. E is not a logical truth in the example about Joe. Note that in cases where E is a logical truth, SOT is trivially true.
All references to Climenhaga are to Climenhaga (forthcoming).
Climenhaga’s official formulation of SOT* (p. 5) involves the expression “for all K” where K is the background information. We take this to be a slip, for SOT* thus formulated is trivially false.
The same is true of any alternative assumption on which Pr(~C 1 | Ca) equals 0 and Ca entails Sm.
It is controversial how estimation problems should be solved. Broadly speaking, there are Bayesian and frequentist approaches. See Howson and Urbach (1993) for a Bayesian perspective and Romeijn (2016) for discussion of maximum likelihood estimation. For discussion of statistical decision theory, see Vassend, Sober, and Fitelson (2017).
Climenhaga notes in effect (footnote 9) that it is not true in general that Pr(Sm | Ca) equals the sample frequency of people who smoke among people who get cancer. We are not claiming otherwise. Our claim, rather, is that Pr(Sm | Ca) equals your estimate of the population frequency of people who smoke among people who get cancer.
It might seem that Sober (2001, p. 343) believes otherwise. But there he has in mind cases where Pr(CC) ≥ Pr(SC).
The same is true with respect to CSOT and a theory defended by Douven and Wenmackers (2015, p. 5) according to which, roughly, if you learn O, and also learn that H is the best explanation (in a partition of hypotheses) of O, then H gets a probabilistic bonus in that your new probability for H should exceed your old conditional probability for H given O.
References
Climenhaga, N. (forthcoming). How explanation guides confirmation. Philosophy of Science.
Douven, I., & Wenmackers, S. (2015). Inference to the best explanation versus Bayes’s rule in a social setting. The British Journal for the Philosophy of Science. doi:10.1093/bjps/axv025.
Howson, C., & Urbach, P. (1993). Scientific reasoning: The Bayesian approach (2nd ed.). Chicago: Open Court.
Lipton, P. (2004). Inference to the best explanation (2nd ed.). London: Routledge.
Lycan, W. (2002). Explanation and epistemology. In P. Moser (Ed.), The Oxford handbook of epistemology (pp. 408–433). Oxford: Oxford University Press.
McCain, K., & Poston, T. (2014). Why explanatoriness is evidentially irrelevant. Thought, 3, 145–153.
Psillos, S. (2007). The fine structure of Inference to the Best Explanation. Philosophy and Phenomenological Research, 74, 441–448.
Roche, W., & Sober, E. (2013). Explanatoriness is evidentially irrelevant, or inference to the best explanation meets Bayesian confirmation theory. Analysis, 73, 659–668.
Roche, W., & Sober, E. (2014). Explanatoriness and evidence: A reply to McCain and Poston. Thought, 3, 193–199.
Romeijn, J. (2016). Philosophy of statistics. In E. Zalta (Ed.), The Stanford encyclopedia of philosophy (Spring 2016 ed). http://plato.stanford.edu/archives/spr2016/entries/statistics/.
Sober, E. (2001). Venetian sea levels, British bread prices, and the principle of the common cause. British Journal for the Philosophy of Science, 52, 331–346.
Sober, E. (2008). Evidence and evolution: The logic behind the science. Cambridge: Cambridge University Press.
Sober, E. (2015). Ockham’s razors: A user’s manual. Cambridge: Cambridge University Press.
Vassend, O., Sober, E., & Fitelson, B. (2017). The philosophical significance of Stein’s Paradox. European Journal for Philosophy of Science. doi:10.1007/s13194-016-0168-7.
Vigen, T. (2015). Spurious correlations. New York: Hachette Books.
Acknowledgments
Thanks to an anonymous referee for a helpful comment on a prior version of the paper.
Author information
Authors and Affiliations
Corresponding author
Rights and permissions
About this article
Cite this article
Roche, W., Sober, E. Is Explanatoriness a Guide to Confirmation? A Reply to Climenhaga. J Gen Philos Sci 48, 581–590 (2017). https://doi.org/10.1007/s10838-016-9357-5
Published:
Issue Date:
DOI: https://doi.org/10.1007/s10838-016-9357-5