Skip to main content
Log in

Abstract

It is tempting to posit an intimate relationship between belief and assertion. The speech act of assertion seems like a way of transferring the speaker’s belief to his or her audience. If this is right, then you might think that the evidential warrant required for asserting a proposition is just the same as the warrant for believing it. We call this thesis entitlement equality. We argue here that entitlement equality is false, because our everyday notion of belief is unambiguously a weak one. Believing something is true, we argue, is compatible with having relatively little confidence in it. Asserting something requires something closer to complete confidence. Specifically, we argue that believing a proposition merely requires thinking it likely, but that thinking that a proposition is likely does not entitle one to assert it. This conclusion conflict with a standard view that ‘full belief’ is the central commonsense non-factive attitude.

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.

Institutional subscriptions

Notes

  1. Grice (1957) went so far as to give a reductive analysis of meaning and assertion in terms of intentional belief transference.

  2. Williamson puts forward this thesis in Knowledge and Its Limits:

    We can suggest something more radical. Belief does not aim merely at truth; it aims at knowledge. The more it is justified by knowledge, the closer it comes to knowledge itself.(208)

    It is plausible… that occurrently believing p stands to asserting p as the inner stands to the outer. If so, the knowledge rule for assertion corresponds to the norm that one should believe p only if one… knows p…. Given that norm, it is not reasonable to believe p when one knows that one does not know p. (256–257)

    We are not entirely sure what is meant by ‘occurrent belief’ as opposed to just plain belief, but it is clear that (recently at least) Williamson endorses the knowledge norm for belief generally. In recent work (2013), he has put forward the knowledge norm for belief as a consequence of his view that ones evidence is just one’s knowledge (‘E = K’):

    Given E = K, the requirement to conform one’s beliefs to one’s evidence is the requirement to conform one’s beliefs to one’s knowledge. Believing only what one knows is then a natural understanding of that requirement. (92)

    The E = K thesis together with requirement that one believe only what one knows entail the requirement that one believe propositions that are entailed by one’s evidence, beliefs that have (‘evidential’) probability 1. Further arguments for knowledge as the norm of belief can be found in Williamson (forthcoming-a, forthcoming-b).

  3. We are not alone in rejecting entitlement equality. Milne (2012), for example, writes, ‘Sincere assertion requires a stronger epistemic standing than belief’ and Stanley (2008) also rejects entitlement equality.

  4. Those who explicitly equate belief with full credence include de Finetti (1990), Levi (1991, 2004), Clarke (2013) and Williamson (though the latter prefers his notion of evidential probability, not credence). In fact, our arguments below will also go against other weaker, but not weak enough, accounts of belief, e.g. Ross and Schroeder (2014) p. 280: “It is rational to believe a proposition p only if one’s evidence significantly favors p over its negation.” Indeed many epistemologists endorse this thesis (at least implicitly) perhaps because like Ross and Schroeder they think it “is difficult to deny.” [ibid.] The requirement that For a recent defense of a view that clearly complies with this general characterization, see, e.g., Leitgeb (2014) or Smith (2010). In general, views that require beliefs to be consistent and sufficiently supported by evidence, will either require beliefs to be fully supported by evidence, or restricted in some other evidentially demanding way.

  5. We should note also that we are not committed here to the idea that there are constitutive norms of assertion at all (see Pagin 2011, for an argument for this negative view). But even if there are not, our arguments still show that in most typical speech-act situations, having a rational belief in p does not entitle you to assert p.

  6. It is not completely clear how requirement on assertion follows from accounts of the evidential norms of assertion such as Williamson’s and Lackey’s. Such data might support a principle that in order to assert something you need to be certain of it, which Stanley (2008) endorses. Williamson would seem to need something like a default principle that if you mention a specific piece of knowledge you are sure of it. (In the case of knowledge, this also might explain why it seems strange to say of one person that he knows that p but isn’t sure that p.) See Williamson (2000, p. 254), for further discussion.

  7. Assuming when you are sure that p and not q you are sure of p and sure that q is false.

  8. These examples are particularly surprising since, as Yalcin (2007) notes, any use of ‘p and might not p’ is infelicitous when embedded under an attitude verb. So we cannot say:

    1. (i)

      ? Tim believes that it’s raining and that it might not be raining.

    Given that (3) is acceptable we should not take (i) to show that belief in p is incompatible with belief in the epistemic possibility that p is false. Rather there may be a semantic or pragmatic explanation of why (i) is bad, as in Yalcin (2007), Klinedinst and Rothschild (2012), Dorr and Hawthorne (2013). Assuming knowledge entails belief, the lack of oddness in ‘Tim believes that it’s raining but he knows that it might not be’ is evidence that this combination is not epistemically defective.

  9. On this Stalnakerian picture, the norms for any given assertion might be relative to the purposes of the conversation. In this case, as Seth Yalcin (p.c.) suggested to us there might not be, as Pagin (2011) argues with regard to norms generally, any general notion of the norms of assertion.

    Note that even on this picture—that may be used to salvage entitlement equality—belief is compatible with rational doubt. Thus, the Williamsonian brand of entitlement equality isn’t saved nor is the thesis that rational belief must be knowledge with (evidential) probability 1.

  10. Williamson (2000, p. 256), for instance, seems committed to the idea that belief that one’s lottery ticket is a loser less than perfectly rational in these cases. Note that he also refers to ‘outright belief’ in this context, which we argue below is a technical concept that does not correspond to belief in the ordinary sense. Cohen and Comesaña (2013) claim that rationality doesn’t permit degrees of irrationality, and so not being perfectly rational is equivalent to being irrational. In any case, in recent work Williamson does not qualify irrationality but instead distinguishes between unjustified belief and unjustified but excusable belief.

  11. A worry about this argument is that due to what is called neg-raising, which we discuss below, negated attributions of belief often are interpreted as attributions of belief in the negation. So, saying ‘I don’t believe it will rain’ seems tantamount to saying ‘I believe it won’t rain.’ This could explain why (5) is contradictory without recourse to the thesis that thinking entails belief. However, this explanation fails to explain a range of other data. First, note that if neg-raising were really obligatory then the following sentence would sound contradictory which it does not:

    1. (i)

      Tim doesn’t believe it’s raining, nor does he believe it’s not raining.

    In addition, even when there is not an explicit sentential negation there is plenty of evidence that ‘think’ is not weaker than ‘believe’. For example, if it were the following sentence would sound coherent:

    1. (ii)

      ?? Martha thinks it’s raining, but Ned believes it’s raining.

    Compare this, for instance, to:

    1. (iii)

      Martha likes to do the dishes, but Ned loves to do them.

    Neg-raising cannot explain this contrast since ‘likes’ is also neg-raising.

  12. Though we note here a number of differences between ‘think’ and ‘believe’. ‘Believe’ but not ‘think’ can be modified by ‘firmly’ and ‘fully’. Having a belief that p seems equivalent to believing p, but thinking p doesn’t seem equivalent to having the thought that p. ‘Think’ also allows a progressive use that belief does not: ‘I was thinking that John would win, but then I realized he wouldn’t.’

  13. We use ‘??’ to indicate more infelicity than ‘?’. For a minimal pair compare to ‘Tim doesn’t actually believe that John stole the painting, but he hopes that he did.’

  14. These examples come from Williamson (p.c.).

  15. Of course, neg-raising in its canonical form only applies when these terms take propositional complements, as in ‘I like/love that John made me breakfast’ or ‘I would like/love to go to the movies’.

  16. The lack of ambiguity in ‘believe’ and ‘think’ may be surprising to those who posit wide-spread semantic or pragmatic ambiguity across natural language predicates.

  17. ‘Think’ seems slightly more natural here, but ‘think’ often feels more natural. Note also it is awful to say, in this case, ‘I think horse A will win but I don’t believe it will’.

  18. We note in both cases that these judgments are not completely universal, but they seem fairly robust. Of course, all these judgements could represent some form of cognitive error (see Windschitl and Wells 1998), but we agree with Yalcin that it is more likely that they reflect somewhat surprising aspects of our notion of likelihood and belief.

  19. Note, however, that any partition relativity in ‘think’ or ‘believe’ cannot be too easily switched or it would allow contradictory sentences such as ‘I think horse A will win but I think horse A will lose’ to be acceptable.

  20. Significance might be very weak, but we do not want to say just more likely because it does not seem like only a minuscule difference is sufficient, as Yalcin also observes.

  21. If we accept this equivalence of thinking p and thinking p likely a new problem arises. For in this case believing that p is likely would require believing that p is likely is itself likely. This iteration effect might seem to make belief even weaker than we propose. One option is to adopt a non-standard semantics for belief attributions according to which believing p is likely just requires having a belief state that makes p likely, e.g. Yalcin (2010) and Rothschild (2012).

  22. Note that this view also explains saying ‘Tim doesn’t think/believe Bill will win, but he doesn’t think/believe Bill won’t win’ suggests that Tim gives roughly even chances or has no information at all.

  23. For example, Wedgwood (2012) writes ‘we typically assume that a sincere assertion of a proposition p expresses an outright belief in p.’ So, he takes the close relation to assertion as an essential mark of outright belief.

  24. In particular, those who stress the importance of outright belief tend to do so against a Bayesian eliminativism about the notion (e.g. Sturgeon 2008). In this literature it is generally taken for granted that the notion of outright belief is the commonsense one. Our point here is that this is far from clear.

  25. Similarly, we might think that the mental act verb ‘judge’ is not weak in the way belief is. Williamson considers belief to be the inner analogue to assertion, but it seems that judging is a closer analogue as, like assertion, it is an act. Note also that ‘judge’ is not neg-raising: saying you don’t judge that it’s raining doesn’t suggest you judge it’s not raining. Also it is not clearly contradictory to say of someone that they believe p but do not judge that p. It as at least plausible that the evidential warrant required for judging that p is stronger than that for believing p. Tellingly, perhaps, the verb ‘judge’ in this general sense has little currency outside of philosophical contexts.

  26. It is clear, however, that locutions like ‘firm belief’ do not pick out the right notion. There is nothing at all defective about firmly believing your lottery ticket will lose, even if there are only probabilistic grounds for this.

  27. Of course, one way out of this dilemma is to use the conversational strategy sketched on page 4 to explain (1) and (3), so maintaining that FB is compatible with doubt (this is the strategy pursued by Stanley 2008).

  28. Thanks to Philippe Schlenker for suggesting this line of thought to us. Stanley (2008) pursues a related thought in his argument for a certainty norm for assertion.

References

  • Bouvier, J. (1856). A law dictionary, Adapted to the Constitution and Laws of the United States.

  • Clarke, R. (2013). Belief is credence one (in context). Philosopher’s Imprint, 13(11), 1–18.

    Google Scholar 

  • de Finetti, B. (1990). Theory of probability. New York: Wiley.

    Google Scholar 

  • Dorr, C., & Hawthorne, J. (2013). Embedding epistemic modals. Mind, 122, 867–913.

    Article  Google Scholar 

  • Dudman, V. H. (1992). Probability and assertion. Analysis, 52, 204–211.

    Article  Google Scholar 

  • Fillmore, C. (1963). The position of embedding transformations in grammar. Word, 19, 208–231.

    Article  Google Scholar 

  • Grice, P. (1957). Meaning. Philosophical Review, 66, 377–388.

    Article  Google Scholar 

  • Klinedinst, N., & Rothschild, D. (2012). Connectives without truth-tables. Natural Language Semantics, 20(2), 137–175.

    Article  Google Scholar 

  • Lackey, J. (2007). Norms of assertion. Noûs, 41, 594–626.

    Article  Google Scholar 

  • Leitgeb, H. (2014). The stability theory of belief. Philosophical Review, 123(2), 173–204.

    Article  Google Scholar 

  • Levi, I. (1991). The fixation of belief and its undoing: Changing beliefs through inquiry. Cambridge: Cambridge University Press.

    Book  Google Scholar 

  • Levi, I. (2004). Mild contraction: Evaluating loss of information due to loss of belief. Oxford: Oxford University Press.

    Book  Google Scholar 

  • Milne, P. (2012). Belief, degrees of belief, and assertion. Dialectica, 66, 331–349.

    Article  Google Scholar 

  • Pagin, P. (2011). Information and assertoric force. In J. Brown & H. Cappelen (Eds.), Assertion: New philosophical essays. Oxford: Oxford University Press.

    Google Scholar 

  • Prince, E. F. (1976). The syntax and semantics of neg-raising, with evidence from French. Language, 52(2), 404–426.

    Article  Google Scholar 

  • Ross, J., & Schroeder, M. (2014). Belief, credence, and pragmatic encroachment. Philosophy and Phenomenological Research, 88(2), 259–288.

    Article  Google Scholar 

  • Rothschild, D. (2012). Expressing credences. Proceedings of the Aristotelian Society, 112(1.1), 99–114.

    Article  Google Scholar 

  • Smith, M. (2010). What else justification could be. Noûs, 44(1), 10–31.

    Article  Google Scholar 

  • Stalnaker, R. (1974). Pragmatic presuppositions. In M. K. Munitz & P. K. Unger (Eds.), Semantics and philosophy. New York: NYU.

    Google Scholar 

  • Stanley, J. (2008). Knowledge and certainty. Philosophical Issues, 18, 35–57.

    Article  Google Scholar 

  • Sturgeon, S. (2008). Reason and the grain of belief. Noûs, 42, 139–165.

    Article  Google Scholar 

  • Swinburne, R. (1983). Faith and reason. Oxford: Oxford University Press.

    Google Scholar 

  • Wedgwood, R. (2012). Outright belief. Dialectica, 66, 309–329.

    Article  Google Scholar 

  • Williamson, T. (1996). Knowing and asserting. The Philosophical Review, 105, 489–523.

    Article  Google Scholar 

  • Williamson, T. (2000). Knowledge and its limits. Oxford: Oxford University Press.

    Google Scholar 

  • Williamson, T. (2013). Response to Cohen, Comesaña, Goodman, Nagel, and Weatherson on Gettier cases in epistemic logic. Inquiry, 56(1), 15–29.

    Article  Google Scholar 

  • Williamson, T. (forthcoming-a). Acting on knowledge. In Carter, J. A., Gordon, E., & Jarvis, B. (eds.), Knowledge-first. Oxford: Oxford University Press.

  • Williamson, T. (forthcoming-b). Justifications, excuses, and sceptical scenarios. In Dutant, J., & Dohrn, D. (eds.), The new evil demon. Oxford: Oxford University Press.

  • Windschitl, P. D., & Wells, G. L. (1998). The alternative-outcomes effect. Journal of Personality and Social Psychology, 75(6), 1411–1423.

    Article  Google Scholar 

  • Yalcin, S. (2007). Epistemic modals. Mind, 116, 983–1026.

    Article  Google Scholar 

  • Yalcin, S. (2010). Probability operators. Philosophy Compass, 5(11), 916–937.

    Article  Google Scholar 

Download references

Acknowledgments

We are grateful to Guillermo Del Pinal, Keith DeRose, Jeremy Goodman, Avishai Margalit, Mike Martin, Peter Pagin, Philippe Schlenker, Amia Srinivasan, Seth Yalcin and an anonymous referee for this journal for comments and discussion. Special thanks to Timothy Williamson for detailed comments on an earlier draft.

Author information

Authors and Affiliations

Authors

Corresponding author

Correspondence to Daniel Rothschild.

Rights and permissions

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Hawthorne, J., Rothschild, D. & Spectre, L. Belief is weak. Philos Stud 173, 1393–1404 (2016). https://doi.org/10.1007/s11098-015-0553-7

Download citation

  • Published:

  • Issue Date:

  • DOI: https://doi.org/10.1007/s11098-015-0553-7

Keywords

Navigation