Norms of Inquiry

This article provides an overview of recent work on norms of inquiry. After some preliminaries about inquiry in §1, I discuss in §2 the ignorance norm for inquiry, presenting arguments for and against, as well as some alternatives. In §3, I consider its relation to the aim of inquiry. In §4, I discuss positive norms on inquiry: norms that require having rather than lacking certain states. Finally, in §5, I look at questions about the place of norms of inquiry within normative epistemology


Inquiry
It will be useful to start with an account of inquiry.I will take inquiry to be the pursuit of an answer to a question.That is, when one inquires, one inquires into a particular question, and one's aim in doing so is settling the question, or getting an answer to the question. 1 You might, for instance, inquire into where your keys are, when the next train is, or why cats purr.In each case, you're inquiring into a particular question (where are my keys?, when is the next train?, why do cats purr?), and the inquiry is settled once you have an answer (the keys are in ..., the next train is at ..., cats purr because ...).
Answers are propositions.Answers to Where are my keys?, for example, include: the keys are in the drawer, the keys are in the backpack, the keys are on the kitchen counter, and so on.One of these is true (insofar as the question has a true answer at 1 Perhaps there's also inquiry into a phenomenon, rather than into a question.For instance, inquiring into the fall of Rome, climate change, and so on.(See Kelp (2021a).)I'm inclined to think that these can also be understood in terms of inquiring into a question or a set of questions: what causes climate change?how can it be mitigated, or stopped, if at all? and so on.At any rate, the literature on norms of inquiry has focused on inquiring into questions, and so I'll set this issue aside here.all), while all other answers are false.Each of these answers is a complete answer to the question, as opposed to a partial one.A partial answer is a proposition that rules out one or more complete answers, or one that supports one or more complete answers over others.For example, the keys are either in the drawer or in the backpack is a partial answer to Where are my keys?. 2 To 'have' an answer to a question is to stand in a particular relation to it: belief, knowledge, understanding, and so on.
Which one of these, if any, is correct is controversial and is part of what is at issue in the norms of inquiry debate.
I will also take inquiry to require an inquiring attitude, such as being curious, wondering, intending to find out, and so on. 3Such attitudes are directed at particular questions: you can wonder where your keys are, be curious why cats purr, intend to find out who's at the door, and so on. 4Inquiry also requires acting on one's inquiring attitude in some way.Just being curious about whether there's beer in the fridge isn't inquiring into the matter; you also need to check the fridge, or ask someone.
Likewise, opening the fridge doesn't suffice for inquiring into its contents if you only do so to exercise your arm.

The Ignorance Norm
Much of the debate on the norms of inquiry has focused on the ignorance norm.
Where p Q is a complete answer to a question Q, the ignorance norm can be formulated as follows: Ignorance Norm One ought not: know p Q and inquire into Q. 5 Ignorance Norm is a 'wide-scope' norm, where the 'ought' takes scope over the conjunction of knowing and inquiring. 6As such, it forbids the combination of inquiring into a question while at the same time knowing its answer.
Inquiring and knowing, on this view, are conflicting states.In inquiring, one treats 2 See Wiśniewski (2015) for a useful overview of the semantics of questions.
a question as open or unsettled.But in knowing, one treats the question as closed or settled. 7The conflict is more pronounced if one agrees with Friedman (2017) that in inquiring into Q one suspends judgment about Q. 8 Inquiring into Q while knowing its answer then means suspending judgment about Q while knowing its answer.Such a combination of attitudes seems irrational.
Agents who inquire while knowing the answer which they seek seem irrational or not epistemically ideal.Suppose, for instance, that Ann knows that there's beer in the fridge.Nevertheless, she opens the fridge to check if there's beer there.Intuitively, Ann's behavior seems irrational: why would she check whether there's beer when she already knows that there is? Ignorance Norm provides an explanation: Ann violates a norm of inquiry. 9 Relatedly, it is incoherent to say any of the following: (1) #I know it's raining, but I wonder whether it's raining.
(2) #I know there's beer in the fridge, but is there beer in the fridge?
(4) #There's beer in the fridge, but is there beer in the fridge?
Here, too, Ignorance Norm provides an explanation.In the case of (1) and (2), the speaker self-ascribes a violation of Ignorance Norm.Alternatively, we can say that, given Ignorance Norm, a speaker who asks a question or self-ascribes an inquiring attitude represents themself as not knowing the answer to the question.So, in the case of (1) and (2), the speaker both asserts that they know and represents themself as not knowing.In the case of (3) and (4), a similar explanation can be given if we also assume a knowledge norm for assertion: that a speaker permissibly asserts p only if they know p. 10 That is, the speaker both represents themself as knowing the answer and represent themself as not knowing it. 117 Friedman (2017, 311). 8For criticism and discussion of this claim, see Archer (2019Archer ( , 2022)), Masny (2020), andMcGrath (2021).
9 See Friedman (2017Friedman ( , 2019b) ) and van Elswyk and Sapir (2021) for an argument of this sort. 10See e.g.Williamson (2000).The knowledge norm of assertion is controversial.See e.g.Weiner (2005) and Lackey (2007) for counter arguments.See Benton (forthcoming) for a recent overview of the debate.

Inaccessible knowledge
One question about Ignorance Norm is whether it is even possible to do what it prohibits.That is, is it even possible to inquire into Q while knowing p Q ?There are perhaps reasons to think it's not: in inquiring into Q one treats Q as open, while in knowing p Q one seems to treat Q as closed. 12Can one treat the same question as both open and closed at the same time?If it isn't possible to inquire while knowing the answer, then perhaps there's no need to posit Ignorance Norm. 13We don't need a norm to tell us not to do the impossible.Friedman (2017), however, argues that inquiring and knowing are compossible.She appeals to cases such as the following: I know that my colleague Alice is on leave in Paris this term, she told me last month that she would be.Still, yesterday I seemed to have no memory of that conversation with Alice and I wondered why she hadn't been to the talk last week; today I remembered.Yesterday I knew why Alice hadn't been to the talk, and yet I wondered why she'd not been there at the same time.(310) Friedman argues that such cases are commonplace, and gives further examples.An important assumption here is that it's possible to have inaccessible knowledge: the inquirer in the above case knows where Alice is, but has no access to that knowledge.
One argument against Ignorance Norm, by Archer (2018), takes the following form.When knowledge is accessible, it's impossible to inquire while knowing; when knowledge is inaccessible, it's possible to inquire while knowing, but then it can be permissible to do so.So as a claim about accessible knowledge, Ignorance Norm has no force, since it forbids the impossible, while as a claim about inaccessible knowledge it is simply false.The former claim seems true if cases of inaccessible knowledge provide the only reason for positing the compossibility of knowing and inquiring.For the claim that it can be permissible to inquire while having inaccessible knowledge of the answer, consider one such case.Suppose that Beth knows where her keys are, but her knowledge is temporarily inaccessible to her.She needs to find her keys, and so she wonders where they are and starts looking for them.Should we say that Beth is acting irrationality in looking for her keys, or in wondering where they are?Because her knowledge is inaccessible, Beth may be not criticizable for looking for her keys.Moreover, as Archer (2018) notes, inquiring into where hey keys are may be Beth's only way of regaining access to her knowledge.If so, he argues, her inquiry is rationally permissible.
In reply, defenders of Ignorance Norm invoke a familiar distinction between primary and secondary propriety. 14An act has primary propriety when it is done in accordance with its norms, and it has secondary propriety when the acting subject believes with good reason that she is acting in accordance with the norms.van Elswyk and Sapir (2021) apply this distinction to cases such as Beth's.They call cases where the agent violates Ignorance Norm primarily but not secondarily 'wellintentioned inquiry'.In such cases, the agent believes they don't know the answer, and in that sense is doing their best to comply with the norm.Beth, therefore, has secondary though not primary propriety.
A counter-reply is that there may be cases where agents like Beth, who have inaccessible knowledge, lack a belief in their own ignorance, which is required for the above reply to work.In fact, someone like Beth may be aware of their predicament and believe that they do know the answer to their inquiry, though they nevertheless cannot access that knowledge.This may occur in tip-of-the-tongue or feeling-ofknowing cases.Beth might say to herself, for instance, "I know where the keys are, I just can't recall it at the moment", and then proceed to wonder where they are, trying to recall.In such a case, Beth's inquiry violates Ignorance Norm not just in the primary but also in the secondary sense, and the appeal to the primary/secondary propriety distinction won't help.15

Double-checking
A second, related, objection to Ignorance Norm comes from cases of confirmation and double-checking.This objection, too, appeals to the compossibility of knowing and inquiring.Suppose that Ann just left her apartment.She knows she locked the door, yet she might want to double-check that the door is locked.Intuitively, it may seem rationally permissible for Ann to double-check that her door is locked.But this is in conflict with Ignorance Norm: Ann's double-checking is a kind of inquiry into whether the door is locked-an inquiry she conducts while knowing that the door is locked.
Defenders of permissible double-checking, such as Falbo (2023a) and Woodard (forthcoming), further motivate this idea as follows.Plausibly, knowledge is not the strongest possible epistemic state: higher-order knowledge, knowing with a higher degree of certainty, and knowing with stronger justification are examples of stronger epistemic states.So even though Ann knows her door is locked, she may seek a stronger epistemic state: she knows, but she double-checks just to be sure. 16ases of double-checking have also been taken to challenge the claim that inquiry is always directed at a question.Instead, the thought goes, some inquiries may be directed at a proposition.17In the above case, Ann isn't curious or wondering whether the door is locked.Rather, she wants to confirm, or wants to be sure, that the door is locked.Ann's inquiry, on this view, is to be understood as directed at the proposition that the door is locked, and the desire to confirm or the desire to be sure are to be understood as propositional inquiring attitudes.A possible reply here, however, is that Ann's double-checking is still directed at a question: whether the door is locked, rather than the proposition that the door is locked.If Ann finds out that the door isn't locked, for example, her inquiry seems thereby settled, even though what she finds out is different from the proposition that the door is locked.
After finding out, she will proceed to lock the door, but in doing so she no longer seems to be inquiring.18

Alternatives
There are a number of alternatives to Ignorance Norm.One has belief in place of knowledge: No-Belief Norm One ought not: believe p Q and inquire into Q. 19iven that knowledge entails belief, No-Belief Norm is a strictly stronger norm than Ignorance Norm, and there are more cases of inquiry that it forbids, including cases in which the inquirer believes but doesn't know p Q .As such, it is subject to the same objections as Ignorance Norm. 20n response to cases of double-checking and confirmation as counterexamples to No-Belief Norm, Millson (2020) has argued that these cases involve credences, while No-Belief Norm concerns outright belief.This helps explain the sense in which, as Friedman (2019b) argues, outright belief is a kind of settled opinion-belief is a question settling attitude, whereas credences are not.Whether such a response is also available in the case of double-checking where agents know, rather than just believe, the answer to their question, depends on how knowledge, credence, and outright belief are all related.
Other alternatives to Ignorance Norm have been proposed.Goodman and Holguín (2022) propose a norm about being sure: Unsurety Norm One ought not: be sure that p Q and inquire into Q.
Unsurety Norm allows knowing and inquiring at the same time, as long as knowledge doesn't entail being sure.It can thus accommodate some cases of doublechecking and confirmation.Ann knows her door is locked, but she isn't sure of it, so she double-checks.Goodman and Holguín (2022) understand being sure as a state that doesn't require absolute certainty or credence 1.But if so, Unsurety Norm seems nevertheless susceptible to double-checking objections.Suppose Ann is sure her door is locked, but she isn't absolutely certain of it.Since locking her door is very important to her, she wants to have even greater confidence in her door's being locked, and so she double-checks.If double-checking while knowing (but without being sure) is permissible, double-checking while being sure (but not maximally sure) seems like it should be permissible as well.

The aim of inquiry
Discussions of the norms of inquiry often involve claims about the aim of inquiry.
Many, for instance, take Ignorance Norm to go hand in hand with the claim that the aim of inquiry is knowledge.It's not entirely clear what it means to say that X is the aim of inquiry, and likely not everyone means the same thing when they make such a claim.Saying that X is the aim of inquiry can be understood in various ways.
It can be constitutive of inquiry: inquiring just is trying to bring about X, and one doesn't inquire without such an aim. 21It can also be understood as a normative claim: rational inquirers aim at X. 22 however, I will take for granted that there is such a thing as the aim of inquiry, in order to discuss the relation between it and norms of inquiry such as Ignorance Norm.
One prominent view is that the aim of inquiry is knowledge: 23 K-Aim aim of inquiry is knowledge.
By and large, arguments against Ignorance Norm have been taken to be arguments against K-Aim, on the assumption that Ignorance Norm follows from K-Aim. 24This assumption is most likely based on some implicit bridge principle like the following: Bridge If φ-ing aims at A, then one ought not: φ while A.
An inquiry-and-knowledge instance of Bridge makes K-Aim entail Ignorance Norm.
However, it's not at all clear that Bridge generally true.Suppose, for example, that Ann is a professor who announces an important message in class.Her goal is that all students in class hear the message.To be on the safe side, she repeats the announcement.Suppose, however, that everyone already heard the message the first time.Then Ann violated a norm entailed by Bridge: one ought not try to get an audience to hear a message when the audience already heard it.However, it doesn't look like Ann did anything wrong.She wasn't sure whether everyone heard, and in her situation repeating the message seems the rational thing to 21 Kelp (2021b), for instance, takes a knowledge aim to be constitutive of inquiry. 22Falbo (2023a), for instance, understands the aim of inquiry in such a normative way.For a more skeptical discussion of the aim of inquiry, see Friedman (forthcoming).
24 Two exceptions are Carter and Hawthorne (forthcoming) and Haziza (ms) who offer ways of accepting K-Aim while rejecting Ignorance Norm.
In response to such a case, we can either reject Bridge, or appeal to some kind of secondary norm.If we reject Bridge, we might say that an A aim of φ-ing instead entails that one ought not to φ while being sure (or similar) of A. If we don't, we might appeal to a secondary norm that an agent who φs satisfies as long as they aren't sure (or similar) of A. However, a response that doesn't look plausible is to say that Ann's announcement didn't, after all, aim at getting everyone to hear, but rather aim at being sure (or similar) that everyone has heard.Ann stops repeating the message once she's sure everyone has heard it.But her aim was only that everyone hear it.
A proponent of K-Aim may similarly respond to counterexamples to Ignorance Norm.They might reject Bridge and thus the move from K-Aim to Ignorance Norm, and instead opt for some weaker version of Ignorance Norm, for example: don't inquire while being certain that you know.Or they might appeal to secondary norms as already discussed above.In either case, the idea would be that your aim is still knowledge even if you're improving your epistemic state beyond knowledge, as long as you're not sure that you know, don't know that you know, or the like. 25  One alternative to K-Aim in light of cases of inaccessible knowledge and doublechecking is the following:

I-Aim
The aim of inquiry is epistemic improvement. 26  A central motivation for I-Aim comes from counterexamples to Ignorance Norm: cases where inquiry seems permissible despite the possession of knowledge.Proponents of I-Aim argue that we can make sense of those cases by positing that the aim of inquiry is to improve one's epistemic position.Epistemic improvement can come in many forms: stronger justification, higher accuracy, higher-order knowledge, and so on.In cases of double-checking or inaccessible knowledge, for instance, the agent can improve her epistemic position beyond knowledge by inquiring. 27,28  25 See Carter and Hawthorne (forthcoming) and Haziza (ms) for defenses of K-Aim along these lines.
27 Archer (2021) and Falbo (2023a) make the additional argument that inquiry is permissible in cases where knowledge cannot be gained, as long as epistemic position can be improved.
28 Additional alternatives to K-Aim take inquiry to aim at true belief, justified belief, or understanding.See e.g.Kelp (2021a, ch. 1) for discussion.

Positive norms
Ignorance Norm and the like place negative requirements on inquirers: one ought not to know, not believe, not be sure of the answer to one's inquiry.Are there also positive requirements that inquirers must satisfy?Is there, for instance, anything that inquirers must know to properly inquire?Willard-Kyle (forthcoming) has argued for the following norm: Knowledge Norm is motivated by the idea that there's something wrong about inquiring into a question when you know that a correct answer exists.Asking 'Who ate the cake?', for example, seems improper if you don't know that the cake was eaten.Indeed, utterances like the following are clearly incoherent: (5) #I don't know if anyone at the cake, but who ate it?
A positive requirement on inquiry is partly inspired by Plato's Meno.On one way of understanding the Meno paradox, it poses the following dilemma: If you know the answer to the question you are asking, then nothing can be learned by asking.If you do not know the answer, then you cannot recognize a correct answer even if it is given to you.Therefore, one cannot learn anything by asking questions. 30  The first horn of the dilemma can be taken to advocate Ignorance Norm: inquire only if you don't know the answer.The second horn should be rejected: even if you don't already know the answer, you may still be able to recognize a correct answer if you have some partial knowledge. 31On this solution of the Meno paradox, inquiry is possible as long as the inquirer has some but not too much knowledge.The idea that knowing something is required for inquiry motivates a positive requirement on inquiry. 32Knowledge Norm captures one such requirement.But it's not the only one.A norm even more directly related to the Meno paradox is the following: 29 See also Deigan (ms) for a related norm. 30 Sorensen (2018, §6.1). 31Sorensen (2018).I won't argue here that this is the correct interpretation or response to the Meno paradox.For a more in-depth discussion, see Fine (2014).
32 See Willard-Kyle (forthcoming) for such a connection to the Meno paradox.
Meno Norm One ought: inquire into Q only if one can recognize a correct answer to Q.
The idea is that there's something wrong in inquiring without the ability to recognize a correct answer.There is some plausibility to such a norm.Suppose that Ben tries to find out, on his own, which numbers are the prime factors of 105.Suppose, however, that Ben cannot multiply.Thus, when Ben considers the numbers 3,5,7, he doesn't recognize that they are the prime factors of 105, because he cannot multiply them.Ben may know that there is an answer to his question, but his inquiry may still be irrational if he lacks the ability to recognize the correct answer.Of course, what such an ability amounts to exactly remains to be clarified.33

Epistemicity
An important question concerns how norms of inquiry-zetetic norms-fit within normative epistemology.In particular, are zetetic norms epistemic?If they are, how do they interact with other epistemic norms?
One thesis here is the following: Z→E All zetetic norms are epistemic norms.
Z→E is defended by Friedman (2020).Friedman's argument is roughly as follows.
Inquiry has an epistemic end, and so zetetic norms are norms that guide us towards an epistemic end.Such norms are epistemic norms.
Even if not all zetetic norms are epistemic, perhaps some specific ones are.Hall and Johnson (1998) and Flores and Woodard (2023) Friedman (2020) formulates this principle without the 'ought' in the antecedent, resulting in a narrow-scope norm.However, there are well-known issues with norms of this kind.(See e.g.Broome (1999), and see Steglich-Petersen (2021) for discussion in the context of ZIP.)I take the present formulation of ZIP to be more plausible.There are a number of ways of responding to Friedman's argument.Perhaps the most natural is to reject the claim that ZIP is epistemic, and thus to reject Z→E.
For one thing, assuming ZIP to be epistemic leads to what seems an implausible conclusion.For another, there are independent grounds for thinking that ZIP is not epistemic.37Moreover, it seems plausible that what kind of requirements ZIP issues depends on the kind of requirement that features in its antecedent.For instance, if I practically ought to find out how many windows a building has, then presumably the requirement to take the necessary means to finding this out will also be practical, rather than epistemic.
A different sort of response to Friedman's argument accepts that ZIP is epistemic, but rejects the move to E→Z.One such view is proposed by Thorstad (2021), who argues that norms of inquiry such as ZIP and norms of belief such as Permitted Belief and Required Belief apply at different evaluative focal points in the epistemic domain. 38Tensions between norms at different focal points are to be expected, on this view, and they also arise in the practical domain.
Haziza (2022b) also grants the assumption that ZIP is epistemic, but argues that there are plausible epistemic norms of belief that are compatible with ZIP.For instance: Conditional Required Belief If your evidence supports p, you ought to believe p if you take a doxastic attitude to p.
The idea behind Conditional Required Belief is that it can be okay not to take any doxastic attitude toward a proposition for which you have evidence; but if you do take a doxastic attitude, it should be belief. 39Conditional Required Belief is compatible with ZIP: you can satisfy both by inquiring and not forming a doxastic attitude in propositions unrelated to your inquiry.If so, we can accept that zetetic norms are only a proper subset of epistemic norms. 40 issue conflicting requirements.ZIP conflicts not only with Required Belief, but even with the following much weaker norm: Permitted Belief If your evidence supports p, you may believe p. Permitted Belief permits forming beliefs on the basis of evidence unrelated to your counting task, but ZIP forbids it.So Permitted Belief and ZIP are inconsistent.If even such a weak norm conflicts with ZIP, perhaps no non-zetetic norm is compatible with it.If so, E→Z follows.
Friedman (2020) argue that there are epistemic norms on evidence gathering, and thus epistemic norms that tell us, at least in part, how to inquire.Such norms are epistemic, Flores and Woodard argue, because agents who fail to gather sufficient evidence are subject to distinctively epistemic sanction, in the form of reduction in trust.34Amoreradicalthesis is the following: E→Z Only zetetic norms are epistemic norms.Friedman (2020)argues for E→Z as well.She thus ends up with the view that the epistemic and the zetetic are one and the same.E→Z requires the rejection of some widely accepted norms.For instance, many accept that one ought to believe what the evidence supports.This is not a zetetic norm: it doesn't say anything about inquiry, and it applies even when one doesn't inquire.Such a norm must be rejected on Friedman's view.Friedman's argument for E→Z is based on the idea that zetetic norms can conflict with non-zetetic norms.So if zetetic norms are epistemic, and there are no genuine normative dilemmas in epistemology, 35 non-zetetic norms cannot be epistemic.To see the conflict, suppose that you ought to find out how many windows there are in a building.You must count the windows yourself, and you must get an answer soon.While counting, you'll get a lot of evidence unrelated to your task: evidence about the people in the street, about the weather, about yourself, and so on.Required Belief requires that you form beliefs on the basis of all such evidence.However, doing so will distract you from counting.So ZIP requires that you don't form those beliefs and instead focus on counting.ZIP and Required Belief, therefore, Here's Friedman's argument in more detail.First, inquiry is governed by the following instrumental norm: Zetetic Instrumental Principle (ZIP) If you ought to settle Q, then you ought to take the necessary means to settling Q.36Second, ZIP is epistemic.This is supported by the same reasoning that supports Z→E.Finally, ZIP conflicts with non-zetetic epistemic norms.Consider the following version of the evidentialist norm considered above:Required Belief If your evidence supports p, you ought to believe p.