Hedging and the ignorance norm on inquiry


What sort of epistemic positions are compatible with inquiries driven by interrogative attitudes like wonder and puzzlement? The ignorance norm provides a partial answer: interrogative attitudes directed at a particular question are never compatible with knowledge of the question’s answer. But some are tempted to think that interrogative attitudes are incompatible with weaker positions like belief as well. This paper defends that the ignorance norm is exhaustive. All epistemic positions weaker than knowledge directed at the answer to a question are compatible with having an interrogative attitude towards that question. We offer two arguments for this conclusion. The first is based on considerations about the role of hedging in inquiry. The second is conditional on considerations related to the aim of inquiry as a goal-directed activity.

This is a preview of subscription content, access via your institution.


  1. 1.

    A potential example of a different kind of inquiry is discussed by Kelp (forthcoming). Kelp suggests that in addition to inquiry whose target is settling a question, there is also inquiry whose target is understanding a phenomenon. Such inquiry differs potentially along two dimensions: it is phenomenon-directed as opposed to question-directed and not driven by an interrogative attitude but a desire or intention to understand. In this paper, our focus is on the former kind of inquiry. Though one of us is inclined to think that inquiry directed at understanding a phenomenon just is inquiry directed at settling a question, we sidestep getting entangled into issues in the epistemology of understanding.

  2. 2.

    The proceeding discussion does not require us to take a side on what questions are. But it might help some readers to consider a Hamblin-style semantics on which an interrogative denotes a set of propositions that are its candidate answers. A complete answer to a question is then a proposition that entails the truth or falsity of every proposition in that set. In contrast, a partial answer is one that entails the truth or falsity of only some propositions in that alternative set.

  3. 3.

    Whitcomb (2017) goes a step further to defend that IGN is a constitutive norm that individuates a speech act of asking. However, we are generally skeptical that speech acts are individuated by constitutive norms. For reasons to be skeptical that assertion is individuated by a constitutive norm that apply to asking as a speech act, see Maitra (2011) and Kelp and Simion (2020).

  4. 4.

    A nearby notion is well-intentioned termination. Though IGN, as a wide-scope norm, is silent on when to initiate or terminate inquiry, one might propose a norm according to which one must terminate inquiry only when one knows a complete answer. Well-intentioned terminating would then be when one stops inquiring because they believe they know an answer. Whitcomb (2010, pp. 677–678) endorses such a termination norm and convincingly argues that Gettiered inquirers are well-intentioned. They terminate inquiry impermissibly, but their epistemic wrong-doing is excusable because they are justified in believing they know.

  5. 5.

    This reply pairs well with a responsibility-based conception of self-knowledge. If awareness of one’s attitudes involves regarding them as commitments or states for which one is responsible (Burge 1996, Bilgrami 2006), attitudes of which one is unaware are not attitudes that one can attempt to bring into compliance with norms on inquiry.

  6. 6.

    IGN is compatible with an array of stronger epistemic norms such as norms prohibiting inquiry into what we know we know or into what is certain for us. But given standard thinking about higher-order knowledge and certainty, the stronger norms will just prohibit a proper subset of the actions IGN already prohibits. For this reason, we ignore stronger norms.

  7. 7.

    A speaker can hedge without using an attitude verb. Likewise, a speaker can hedge without using believe. We focus on believe to have a clear, unambiguous counterexample to DBI. For example, Millson (forthcoming) presents a challenge for DBI involving confirmation-seeking expressions as opposed to hedging expressions. His way around the challenge is to deny Lockeanism about belief and maintain that confirmation-seeking expressions express credences as opposed to full belief in a proposition. No similar response will be available here because hedging with believe entails speaker belief. Admittedly, hedging with believe is less common than hedging with think. But there is reason to think believe is as weak as think. See Hawthorne et al (2016) and Rothschild (2020). Then think-based examples would be equally problematic for DBI.

  8. 8.

    We take the data to sufficiently motivate that believe in matrix position can be interpreted as a hedge and remain neutral on how best to explain how it is interpreted as such. Presumably, the hedging interpretation is arrived at pragmatically. The agent asking (6) recognizes that (7) is not an answer in a strict sense. Assuming a Hamblin semantics for questions, for example, (7) does not entail the truth or falsity of any of the propositions in (6)’s denotation. The asking agent then attempts to discern why the speaker of (7) would have said what they did. The hedging interpretation will often be the best explanation.

  9. 9.

    A referee suggests an alternative diagnosis. If the negation in the continuation is interpreted in the embedded clause (i.e. neg-raising), then (12) could be paraphrased as this: There is leftover milk, I believe, but I believe there is not leftover milk. This does not express a contradiction. Instead, it expresses that the speaker is in an irrational state. With this diagnosis, one might argue that believe-parentheticals actually expresses an attitude weaker than belief like suspicion. After all, it would still be defective to suspect what you believe is false. However, the neg-raising interpretation can be blocked and a contradiction still results. The easiest way to block neg-raising is to add intonational focus onto the matrix verb (Gajewski 2007; Romoli 2013). When added to (12), infelicity still results.

    (12)* # There is leftover milk, I believe. But I don’t [believe]F that there is leftover milk.

    Additionally, the insertion of a temporal adverb like still between negation and believe also problematizes neg-raising.

    (12)** # There is leftover milk, I believe. But I don’t still believe there is leftover milk.

    Interpreted with neg-raising, the continuation (12)** states that the speaker still believes there is not leftover milk. But that interpretation is hard to access. Finally, we can construct a similar contradiction without negation.

    (12)*** # There is leftover milk, I believe. But I refrain from believing there is leftover milk.

    The phrase refrain from replaces negation in (12***). This phrase does not license a raised interpretation but still yields a contradiction. An agent cannot both V and refrain from V-ing. These considerations show that the alternative diagnosis is not as explanatory as the straightforward proposal that believe-parentheticals express belief.

  10. 10.

    See Asher (2000), Murray (2017), and van Elswyk (forthcoming).

  11. 11.

    It is important to distinguish what hedging can reflect about the speaker’s position and what hedged testimony may entitle for a hearer. We have defended that a speaker can hedge and believe. This is fully compatible with hedged testimony not entitling full belief. For example, Jackson (2020, §3.3) suggests that hedged testimony only entitles a hearer to adjust their credences as opposed to forming a full belief in what is stated.

  12. 12.

    Both of these observations about the speaker and hearer obtain even if the context has incredibly low stakes. This suggests that hedged assertions are not on their own enough to settle inquiry regardless of the stakes. To be sure, Harry could proceed as if Sally’s reply were sufficient. But, in doing so, Harry would be acting in a way that goes beyond what Sally actually contributed to the conversation.

  13. 13.

    Archer (2018) considers discourses like I am wondering whether the bank is open, but I believe it is and argues that they problematize the thesis defended in Friedman (2017) that possessing an interrogative attitude requires suspending belief. We do not share this judgment. We submit that the difference between his discourses and (13) and (14) is that the latter are clearly hedges.

  14. 14.

    Note that (17) improves if and is replaced by but. Since but merely indicates contrast between the conjuncts, (17) is at least partly odd for reasons having nothing to do with inquiry.

  15. 15.

    We will not argue in this paper that one properly terminates an activity if and only if the goal of that activity is attained. Though attainment seems to ensure proper termination, proper termination may not guarantee attainment. For activities with a goal or aim that can be put along a scale, it is plausible that some activities allow proper termination at attainment and at some lower point on the scale. Were DBI correct, one might hypothesize that inquiry fits this description. Knowledge is its aim, but full belief sets the minimum threshold for proper termination. But two observations about this hypothesis are worthwhile. First, the connection lacuna would still remain. An explanation would still be owed for why one can properly terminate inquiry at a lower point on the scale. Second, the hypothesis would need to distinguish quitting from proper termination. To illustrate the distinction, consider a chess player who walks away from a game before its completion because they are bored or see no way to win. Such a player quits as opposed to properly terminates the game. The distinction matters because it is conceivable that agents who stop inquiry at full belief are quitting as opposed to properly terminating: they see no way to arrive at knowledge, or they are practically deterred.

  16. 16.

    Friedman (2013) argues against analyzing IAs like curiosity as metacognitive states like desires to know. We do not assess her case here. But note that even if IAs are not identical with metacognitive states, inquiry might still require a metacognitive state to accompany an IA.

  17. 17.

    Desires conflict. The desire to know is no different. We submit that cases where an agent is inquiring into a question for which they do not want to know the answer—whether a partner is faithful, whether the calorie count of kefir is too high—are cases of conflicted desires. In inquiring, the agent desires to know the answer to their question, but they also desire to not know the answer because they are afraid of what it might be.

  18. 18.

    There are other ways on which knowledge might be question-relative which do not require a function-theoretic approach to epistemology. See Hookway (1996) and Schaffer (2005, 2008, 2015). The significance to IGN and the connection lacuna is presumably the same. For reasons of space, we overlook these theories.

  19. 19.

    See Friedman (2017, 2019a), Staffel (2019), Mirrachi (2019), McGrath (forthcoming), and Raleigh (forthcoming) for recent work linking inquiry to suspension and the nature of belief.

  20. 20.

    For helpful comments and conversation, we thank participants at the Nature of Inquiry Workshop at Agnes Scott College. In particular, we are indebted to Jane Friedman, Dennis Whitcomb, Jared Millson, Dunja Šešelja, Will Fleisher, Zach Barnett, and Arianna Falbo. Thanks are also due to Joshua Spencer, Chris Willard-Kyle, and two anonymous referees at this journal.


  1. Archer, A. (2018). Wondering about what you know. Analysis, 78(4), 596–604.

    Article  Google Scholar 

  2. Adler, J. (2002). Belief’s own ethics. Cambridge: MIT Press.

    Google Scholar 

  3. Asher, N. (2000). Truth Conditional Discourse Semantics For Parentheticals. Journal of Semantics, 17(1), 31–50.

    Article  Google Scholar 

  4. Benton, M. A. (2011). Two more for the knowledge account of assertion. Analysis, 71(4), 684–687.

    Article  Google Scholar 

  5. Benton, M. A., & van Elswyk, P. (2020). Hedged assertion. In S. Goldberg (Ed.), The oxford handbook of assertion. Oxford: Oxford University Press.

    Google Scholar 

  6. Bilgrami, A. (2006). Self-knowledge and resentment. Cambridge: Harvard University Press.

    Google Scholar 

  7. Burge, T. (1996). Our entitlement to self-knowledge. Proceedings of the Aristotelian Society, 96, 91–116.

    Article  Google Scholar 

  8. Craig, E. (1990). Knowledge and the state of nature: an essay in conceptual synthesis. Oxford: Oxford University Press.

    Google Scholar 

  9. DeRose, K. (1992). Contextualism and knowledge attributions. Philosophy and Phenomenological Research, 52(4), 913–929.

    Article  Google Scholar 

  10. DeRose, K. (2002). Assertion, knowledge, and context. Philosophical Review, 111(2), 167–203.

    Article  Google Scholar 

  11. Fraser, B. (2010). Pragmatic competence: the case of hedging. In G. Kaltenöck, W. Mihatsch, & S. Schneider (Eds.), New approaches to hedging (pp. 15–34). Emerald.

  12. Friedman, J. (2013). Question-directed attitudes. Philosophical Perspectives, 27, 145–174.

    Article  Google Scholar 

  13. Friedman, J. (2017). Why suspend judging? Noûs, 51(2), 302–326.

    Article  Google Scholar 

  14. Friedman, J. (2019a). Inquiry and Belief. Noûs, 53(2), 296–315.

    Article  Google Scholar 

  15. Friedman, J. (2019b). Checking again. Philosophical Issues, 29, 84–96.

    Article  Google Scholar 

  16. Gajewski, J. R. (2007). Neg-raising and polarity. Linguistics and Philosophy, 30(3), 289–328.

    Article  Google Scholar 

  17. Hawthorne, J., Rothschild, D., & Spectre, L. (2016). Belief is weak. Philosophical Studies, 173(5), 1393–1404.

    Article  Google Scholar 

  18. Hookway, C. (1996). Questions of context. Proceedings of the Aristotelian Society, 96, 1–16.

    Article  Google Scholar 

  19. Jackson, E. (2020). Belief, credence, and evidence. Synthese, 197, 5073–5092.

    Article  Google Scholar 

  20. Kappel, K. (2010). On saying that someone knows: Themes from Craig. In A. Haddock, A. Millar, & D. Pritchard (Eds.), Social epistemology. Oxford: Oxford University Press.

    Google Scholar 

  21. Kelp, C. (2011). What’s the point of “knowledge” anyway? Episteme, 8, 53–66.

    Article  Google Scholar 

  22. Kelp, C. (2014). Two for the knowledge goal of inquiry. American Philosophical Quarterly, 51, 227–232.

    Google Scholar 

  23. Kelp, C., & Simion, M. (2020). The C account of assertion: A negative result. Synthese, 197, 125–137.

    Article  Google Scholar 

  24. Kelp, C. (forthcoming). Inquiry, knowledge and understanding. Synthese.

  25. Loewenstein, G. (1994). the psychology of curiosity: A review and reinterpretation. Psychological Bulletin, 116(1), 75–98.

    Article  Google Scholar 

  26. Maitra, I. (2011). Assertion, norms, and games. In J. Brown & H. Cappelen (Eds.), Assertion: New philosophical essays. Oxford: Oxford University Press.

    Google Scholar 

  27. McGrath, M. (forthcoming). Being neutral: Agnosticism, inquiry and the suspension of judgment. Noûs.

  28. McCready, E. (2014). Reliability in pragmatics. Oxford: Oxford University Press.

    Google Scholar 

  29. Millar, A. (2011). Why knowledge matters. Proceedings of the Aristotelian Society, 85, 63–81.

    Article  Google Scholar 

  30. Millson, J. (forthcoming). Seeking confirmation: A puzzle for norms of inquiry. Analysis.

  31. Mirrachi, L. (2019). When evidence isn’t enough: Suspension, evidentialism, and knowledge-first virtue epistemology. Episteme, 16, 413–437.

    Article  Google Scholar 

  32. Murray, S. (2017). The semantics of evidentials. Oxford: Oxford University Press.

    Google Scholar 

  33. Peirce, C. S. (1998). The essential Peirce: 1893–1913. Bloomington: Indiana University Press.

    Google Scholar 

  34. Raleigh, T. (forthcoming). Suspending is believing. Synthese.

  35. Romoli, J. (2013). A scalar implicature-based approach to neg-raising. Linguistics and Philosophy, 36(4), 291–353.

    Article  Google Scholar 

  36. Rooryck, J. (2001). Evidentiality, Part I. Glot International, 5, 125–133.

    Google Scholar 

  37. Rothschild, D. (2020). What it takes to believe. Philosophical Studies, 177(5), 1345–1362.

    Article  Google Scholar 

  38. Rysiew, P. (2012). Epistemic scorekeeping. In J. Brown & M. Gerken (Eds.), Knowledge ascriptions. Oxford: Oxford University Press.

    Google Scholar 

  39. Schaffer, J. (2005). Contrastive knowledge. Oxford Studies in Epistemology, 1, 235–237.

    Google Scholar 

  40. Schaffer, J. (2008). Knowledge in the image of assertion. Philosophical Issues, 18(1), 1–19.

    Article  Google Scholar 

  41. Schaffer, J. (2015). Lewis on knowledge. In J. Schaffer & B. Loewer (Eds.), A companion to David Lewis. New York: Wiley.

    Google Scholar 

  42. Slote, M. A. (1979). Assertion and belief. In J. Dancy (Ed.), Papers on language and logic. Keele: Keele University Library.

    Google Scholar 

  43. Staffel, J. (2019). Credences and suspended judgments as transitional attitudes. Philosophical Issues, 29(1), 281–294.

    Article  Google Scholar 

  44. van Elswyk, P. (forthcoming). Representing knowledge. The Philosophical Review.

  45. Whitcomb, D. (2010). Curiosity was framed. Philosophy and Phenomenological Research, 81(3), 664–687.

    Article  Google Scholar 

  46. Whitcomb, D. (2017). One kind of asking. Philosophical Quarterly, 67(266), 148–168.

    Article  Google Scholar 

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

    Google Scholar 

Download references

Author information



Corresponding author

Correspondence to Peter van Elswyk.

Additional information

Publisher's Note

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

Penultimate Draft, Forthcoming in Synthese

Rights and permissions

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

van Elswyk, P., Sapir, Y. Hedging and the ignorance norm on inquiry. Synthese (2021). https://doi.org/10.1007/s11229-021-03048-0

Download citation


  • Inquiry
  • Interrogative attitudes
  • Hedging
  • Knowledge