Skip to main content
Log in

How can necessary facts call for explanation?

  • Published:
Synthese Aims and scope Submit manuscript

Abstract

While there has been much discussion about what makes some mathematical proofs more explanatory than others, and what are mathematical coincidences, in this article I explore the distinct phenomenon of mathematical facts that call for explanation. The existence of mathematical facts that call for explanation stands in tension with virtually all existing accounts of “calling for explanation”, which imply that necessary facts cannot call for explanation. In this paper I explore what theoretical revisions are needed in order to accommodate this phenomenon. One of the important upshots is that, contrary to the current consensus, low prior probability is not a necessary condition for calling for explanation. In the final section I explain how the results of this inquiry help us make progress in assessing Hartry Field’s style of reliability argument against mathematical Platonism and against robust realism in other domains of necessary facts, such as ethics.

This is a preview of subscription content, log in via an institution to check access.

Access this article

Price excludes VAT (USA)
Tax calculation will be finalised during checkout.

Instant access to the full article PDF.

Similar content being viewed by others

Notes

  1. For more examples, see Leslie (1989, pp. 8–19), White (2005, p. 3) and White (2018).

  2. Some authors imply that there is a close connection between calling for explanation and improbability, but do not state explicitly that they view improbability as a necessary condition. Examples include Smolin (1997, p. 45), Price (2002, p. 115).

  3. Sometimes the explananda are events rather than facts, sometimes the relevant property is described as "surprise" or "puzzlement", rather than "calling for explanation". For a more nuanced discussion of Horwich see Baras (2019). For discussion of Schlesinger see Baras (unpublished ms.).

  4. Prior to this paper, the possibility that necessary truths can call for explanation has been raised in blogposts by Pruss (2013) and Berry (2017), and in Baras (2017a, p. 203) and Baras (2017b, p. 485) in the context of Hartry Field’s (1989, pp. 25–30) style of argument against mathematical Platonism and parallel arguments against realisms in other abstract domains. As I completed this paper, Berry developed and published her view in Berry (2020). I engage with this literature and with her article in Sect. 5. In the context of philosophy of mind there is a debate about whether identities, such as mind–body identities, can call for explanation and some have made general claims about necessary facts in this context. Examples are Melnyk (2003, p. 52) and Papineau (2011, p. 9) who claim without much argument that necessary facts cannot call for explanation and Morris (2011, p. 378) who responds by citing mathematical explanations to show that they can call for explanation. These references are all to very brief comments. They do not explore the question in depth as I do here. (I thank Assaf Weksler for referring me to this debate). After completing this paper, I discovered that van Cleve (2018) explores at great length the idea that some necessary facts should be explained. Fortunately, there is not much overlap between the papers, so they nicely complement each other.

  5. I thank an anonymous referee for this example.

  6. What precisely is a mathematical explanation and how does it stand in relation to other types of explanation? These are difficult questions that I cannot address. I note though that Lange's (2016) account seems attractive. For a classic, see Steiner (1978). For a recent survey, see Mancosu (2018).

  7. For a recent argument for the necessity of mathematics from mathematical practice, see Yli-Vakkuri and Hawthorne (2018). For doubts about the necessity of mathematics, see Field (1989, Chapter 7).

  8. For a formal argument (unfortunately, not very accessible to non-mathematicians) that mathematical claims can be translated to hypothetical claims in modal logic (S5) about what follows necessarily from certain axioms, see Hellman (1993, 1996). I remain neutral with regard to his further claim that this is all that mathematical claims amount to. (I thank Sharon Berry for discussion and the reference to Hellman).

  9. See Hájek (2019, sec. 1), Talbott (2016, sec. supplement) and Strevens (2017, p. 13).

  10. For examples and some more elaboration, see Baras and Shenker (2020).

  11. In ordinary conversation, “coincidence” is sometimes used as synonymous with “lacks an explanation”. Other times, people use “coincidence” and “explanation” differently, such that a coincidence is considered a legitimate type of explanation. My use of these terms here (and I’m following a core of the literature) is in accordance with the former, not the latter.

  12. There is a theoretical question here, which I will explore elsewhere, whether the claim is just that E must have an explanation, or a stronger claim, that E must have a particular kind of explanation. For the purposes of this article, the distinction will not matter. For preliminary discussion, see Baras (2019) and Baras (2020).

  13. This epistemic principle is sometimes called “the striking principle” because some authors call the property in virtue of which a fact calls for explanation “strikingness”. For references to authors who endorse this principle and arguments that rely on it as a premise, see Baras (2019, 2020).

  14. I thank Assaf Weksler for this suggestion.

  15. Perhaps there are other examples that do support the possibility of mathematical claims calling for explanation in sense (4). Some colleagues have suggested to me that there are arguments against certain mathematical axioms that can be understood as based on the claim that those axioms imply that a mathematical fact that calls for explanation could not be explained. Assessing such proposed examples would require further research.

  16. Other adherents of epistemic interpretations add additional constraints for the priors because they believe that not every probability distribution that conforms to the axioms is rational. Such views are often called “logical” interpretations of probability.

  17. I’m inspired here by Colyvan (2013), who distinguishes three types of motivations for idealization in normative domains. Colyvan lists mathematical simplicity as an independent type of motivation whereas I group it with other pragmatic considerations. (I thank Arnon Levy for acquainting me with this discussion).

  18. For an attempt to develop a formal representation of credences without logical omniscience, see Gaifman (2004). Bjerring (2013) demonstrates difficulties in using an impossible worlds framework for this task.

  19. For a defense of accepting the logical omniscience idealization as a normative idealization without global fact omniscience, see Christensen (2004, pp. 153–157) and Smithies (2015).

  20. I thank Jessica Wilson for raising a suggestion along these lines in conversation.

  21. See Baras (2019, sec. 4) for some discussion.

  22. Recent critiques of this style of response include Schechter (2018), Faraci (2019) and Korman and Locke (forthcoming).

References

Download references

Acknowledgements

I am thankful to Sharon Berry, David Enoch, Arnon Levy, Alexander Pruss, Joshua Schechter, Assaf Weksler, two anonymous reviewers and to the audiences at my presentations at the 22nd Israeli Philosophical Association conference (2019) and the 93rd Joint Session of the Aristotelean Society and the Mind Association (2019) for very helpful comments and discussion. This article was written with the support of the Martin Buber Society of Fellows at the Hebrew University of Jerusalem.

Author information

Authors and Affiliations

Authors

Corresponding author

Correspondence to Dan Baras.

Additional information

Publisher's Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Rights and permissions

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Baras, D. How can necessary facts call for explanation?. Synthese 198, 11607–11624 (2021). https://doi.org/10.1007/s11229-020-02817-7

Download citation

  • Received:

  • Accepted:

  • Published:

  • Issue Date:

  • DOI: https://doi.org/10.1007/s11229-020-02817-7

Keywords

Navigation