Epistemic Obligations

Just an observation: deontic talk, in much of ordinary discourse, is exaggerated. This point is important for epistemology as well as for theories of the rationality of action.

Paradigm example: my wife gets a “free gift” card from retailer X. She already had 2 of them, and went to X to collect on Saturday. Monday another shows up in the mail, and she says, “dang I should have waited–I could have redeemed all 3 together!”

It is obvious that what she says is false–it isn’t the case that she should have waited. The same should be noted, I think, when people say that you should believe the truth. No, that’s false, and the explanation of the mistake is that it is very common to confuse the theory of obligation with the theory of value: it is, in some sense, the best one can do to believe the truth; but in that sense, one has no obligation to conform to whatever is best for one to do. (This post is prelude to another on the connections between obligation and value, one with more theoretical bite–the point of the present post is to keep two theoretical points distinct, the point about the attraction of exaggeration concerning obligation talk and the point about the implications of the best explanation of why it counts as exaggeration.)


Epistemic Obligations — 11 Comments

  1. Hi Jon,
    I’m not sure it is obvious that she speaks falsely when she says “I should have waited”. I’m not even sure she *does* speaks falsely in that scenario. Why think as much? I don’t mean the question rhetorically…I wonder if you have an argument up your sleeve for thinking those kinds of assertions are false and/or are exaggerations.

    For whatever it is worth, I make those kinds of assertions all the time, and I think they are both true and unexaggerated. Maybe that just says I’m weird though… 🙂

  2. Hi Dennis, first, I think there are a number of different notions of obligation, and so there may be one on which the claim in question is true. But I meant to isolate something like practical rationality for the example, and here I think it is obviously false. Unless, of course, you are willing to endorse wholesale obligation violation for all of us because we aren’t multibillionaires: buy the stocks that go up and we’d all be rich. Better to say that there is a failure on the value side than on the obligation side, since there one gets the rather trivial claim that there is something we each could have done that would have further our own interests beyond what we actually did (holding fixed, of course, the worries that being rich has its own costs…).

  3. I’m with Dennis; I see no pressure towards describing such utterances as false. They’re perfectly natural and appropriate, and plausible semantics are available to render them true. (Deonic modals — indeed, modals generally — are flexibly context-sensitive.) So what’s the problem?

    Of course there are *particular interpretations* of such utterances that are obviously false. But they’re not the correct ones. Indeed, that they are obviously false counts strongly against their being correct interpretations.

  4. Jonathan, sounds a bit far-fetched to think that those who don’t invest in the stock market in a way that makes them rich should have behaved differently. And the connection you cite between what is natural and appropriate to say and how the semantics should treat the claims in question ignores the pragmatics/semantics interplay.

    I put the example in question in terms of the truth or falsity of what she said, but I really don’t want the focus to be on ordinary usage or anything metalinguistic. I want it to be about obligations we have and don’t have, and I have no idea how my wife had an obligation of any sort not to go to the store on Saturday, though I can see pretty clearly how it would have turned out better for her had she waited.

  5. Jonathan and Dennis, it may be worth noting as well that my reaction that what she said was false isn’t meant to imply that there isn’t a perfectly straightforward and natural interpretation of her remark that I’d endorse and in fact do endorse in responding to her. It’s the value interpretation, in terms of what would have been a more efficient use of time. But, as I’ve been pushing, one can endorse all of that and still maintain, as I do, that she violated no obligations to herself or anyone else by going to the store on Saturday.

  6. My initial impulse is to side with Jon K. on the point under discussion, though I find myself troubled by three points/questions:

    1) Right now, this discussion would benefit from sharpening our intutions. Specifically, what might be some of the independent criteria for testing those intuitions? For example, I take it that one mark of contexts where deontic talk is rightly applied concerns whether or not it’s appropriate to sanction someone for non-compliance, as would be the case if one failed to fulfill an obligation. But quite a bit here hinges on how broadly we construe those sanctions. For example, Jon’s wife’s regret might be seen as a weak form of self-imposed sanctioning. In which case, the deontic talk might turn out to be correct after all, with the modesty of the sanctions explaining why the deontic talk appears to be exaggerated.

    2) Suppose that we grant that often deontic talk should be traded in for “could-have-done-better” talk. Clearly there are still some genuine, legitimate uses for deontic talk. So going back to Jon K’s original post, why does the obligation to believe the truth not count as one of these legitimate uses of deontic talk? I find my intuitions are much weaker in this context than in the paradigm example involving coupons and shopping.

    3) Finally, a proposal: the kinds of ‘exaggerated’ cases that Jon raises can preserve their deontic character if we take exaggerated utterances expressing that “I should’ve done A” to be elliptical for “Had I known then what I know now, then I should’ve done A.” As Jon indicates, he’s not interested in wrangling over ordinary language claims, but I wonder how one discerns whether Jon’s proposal or the one just sketched makes better sense of exaggerated deontic claims.

  7. Kareem, I think your 3rd point is central here, and I’m happy with it. It encodes the kind of epistemic encroachment that I am touting: the relevant kind of obligations are constrained by epistemic factors, in a very strong and central way. That’s because the kind of obligation I’m thinking about is the kind I take to be connected with moral responsibility, which is well-known for being subject to epistemic encroachment.

    This is all quite vague at this point, I realize. There are so many different kinds of deontic phenomena that it is easy for all of us to talk past each other. What I want is a notion of fundamental normativity regarding (the egocentric predicament concerning) what to do and what to think, and I hope that this type of normativity is closely linked to moral responsibility. If not, then I need a different account of why fundamental normativity involved epistemic encroachment.

  8. Hi there,

    I think that there are different kinds of obligations (or reasons) which trigger different responses to the gift card case in the first entry. (i) Of course, there is no “categorical” reason or obligation for your wife, independent of her desires, to wait till monday. In this sense an utterance of “I should have waited till monday” is clearly false. (ii) But if your wife has the desire to collect more of these gift cards to redeem them all at once or to get a better gift, then obviously there exists a “hypothetical” reason or obligation for her to wait until monday. In this sense an utterance of “I should have waited until monday” is true!
    Does this make sense to you? Or am I missing something?

    (I am no native speaker. Sorry for my poor English.)

  9. Hi Martin, Agree on (i), certainly. On (ii), she has no such desire, though there are hypotheticals that are true here. For one, she wants, in general, to maximize benefits with minimal effort. In that sense, she should have waited, because it fulfills that want.

    The issue I want to focus on, though, isn’t about whether there are some kind of obligations that are present, but upon the obligations or lack of them that answer to the central concerns of the egocentric predicament concerning what to do and what to think. Regarding that concern, we seek guidance, and a theory that tells my wife that she should have waited is mistaken. It isn’t sensitive enough to the perspectival character of the relevant kind of obligation.

  10. This raises interesting issues, which I think can be illustrated by strongly moral cases. Take for example recent scandals, in Iraq and Afghanistan, regarding the killing of innocent civilians (pregnant women, journalists, etc.). A common defense in these cases is lack of knowledge before the cases.

    Let’s take this defense at face value. Judging by your post and comments, it seems that you have to say it’s technically false to say “The soldier should not have killed the pregnant mothers.”

    One thing that’s interesting here is that the blameworthiness of the soldier is mitigated (again, accepting the defense), but other normative truths remain, e.g. “Pregnant mothers ought not to be shot.” I suspect that truths like this are trying to be preserved with uses of ‘should,’ as in “The soldiers should not have shot the pregnant mothers.”

    I pick this case not in the way people pick Hitler examples, but just to clarify intuitions. So in the shopping cases, it might be that ‘should’ preserves something like, “Gift cards should be combined.”

  11. Hi Jon,

    I know your focus is ultimately on epistemic obligations. And that you are concerned, as far as I can extract from your posts, to show that a theory which implies that there are any epistemic obligations simpliciter in similiar cases like these is false (though it is valuable to believe the truth, of course). Your point, if I get it right, is that there are no epistemic obligations independet of the perspective of the agent. Or, to put it a different way, that a theory which implies or postulates such obligations is false, since there aren’t in this sense. But if this is your idea, why is it not possible to relativize epistemic obligations to the egocentric perspective? According to this position, what an agent would be obligated to do would be a function or be determined from his egocentric perspective (his desires, beliefs and other contentful states, probably). Is this not a possible move?

Leave a Reply

Your email address will not be published. Required fields are marked *