Abstract
In this paper, I will discuss what I will call “skeptical pragmatic invariantism” (SPI) as a potential response to the intuitions we have about scenarios such as the so-called bank cases. SPI, very roughly, is a form of epistemic invariantism that says the following: The subject in the bank cases doesn’t know that the bank will be open. The knowledge ascription in the low standards case seems appropriate nevertheless because it has a true implicature. The goal of this paper is to show that SPI is mistaken. In particular, I will show that SPI is incompatible with reasonable assumptions about how we are aware of the presence of implicatures. Such objections are not new, but extant formulations are wanting for reasons I will point out below. One may worry that refuting SPI is not a worthwhile project given that this view is an implausible minority position anyway. To respond, I will argue that, contrary to common opinion, other familiar objections to SPI fail and, thus, that SPI is a promising position to begin with.
This is a preview of subscription content, log in via an institution to check access.
Similar content being viewed by others
Notes
SPI, so understood, is defended in Conee (2005a, 52 f.), Douven (2007) and Davis (2004, 2007, 2010, 2015). Bach (2010) “suggests, without endorsing,” it. Fantl and McGrath (2009, pp. 185–194) claim that the view is “not out of the question.” BonJour (2010, p. 78) defends what may be seen as a version of SPI, but presents his version of SPI only as a “tentative conclusion.” Schaffer (2004) defends a version of SPI, but has retracted the view by now. Note that Bach would couch his view in terms of implicitures rather than implicatures. The difference between implicatures and implicitures will mostly be irrelevant for my concerns, but see below for some further remarks.
To be precise, this thesis distinguishes epistemic invariantism only from epistemic contextualism. Epistemic invariantism is also supposed to be incompatible with all forms of epistemic relativism. Thus, epistemic invariantism also entails, first, that the proposition expressed by knowledge sentences doesn’t vary with the epistemic standard of the context of assessment and, second, that the truth-value of that proposition is neither relative to the epistemic standard of the context of use nor to the epistemic standard of the context of assessment. See e.g. MacFarlane (2005) for further discussion of the various forms of relativism.
Conveying that p is supposed to be roughly equivalent to intending to make one’s audience believe that p.
One may worry that even substitutional implicatures cannot make a literally false claim appear true (rather than appropriate). But this worry rests on shaky intuitions. First, it is unclear whether DeRose’s claim in the bank cases would indeed seem true rather than appropriate. See e.g. Davis (2010, p. 1154). Thus, it is unclear whether the implicature needs to make his claim appear true. Second, it is unclear whether the relevant implicatures (namely, loose use implicatures, see below) cannot make a literally false claim appear true. See e.g. DeRose (2012, p. 716) for the intuition that they can.
Schaffer (2004) is an exception. Note, though, that the worry I will level against SPI below does not depend on the above precisification of the position.
There are some controversies surrounding the correct analysis of loose use. Bach (1994), for example, treats loose use in terms of implicitures, Davis (2007) treats it in terms of implicatures. DeRose (2012, 714 ff.) suggests that loose use might be a semantic rather than pragmatic phenomenon. To sort out these issues, we would have to delve deeply into the discussion on the semantics-pragmatics distinction. This would go way beyond the scope of this paper. Since, however, the pragmatic account of loose use seems much more prominent than the semantic account, it should be fair to just assume that some such account is correct. Whether, in the end, the account involves implicitures or implicatures will mostly be irrelevant for the discussion to come. See below.
Bach (2010, 122f) “suggests without endorsing” this proposal. To be precise, Bach uses the phrase “provided things are normal” instead of the phrase “unless something unlikely happens.” Nothing should depend on this modification however. I have made this modification only for the sake of unity.
Note that not just any “unlikely” event is going to be relevant in the present context. I will take it to be tacitly understood that we are only concerned with those unlikely events that entail that p is not the case. As regards DeRose’s knowledge ascription in the bank cases, for example, a relevant unlikely event would be that the bank changes its hours. An irrelevant unlikely event would be, say, that an elephant sings the national anthem. The same goes for the second proposal to interpret the loose use implicature below.
Douven (2007, 333 n.) endorses this view. See also BonJour (2010, 73 f.). Interestingly enough, Douven explicitly considers the first interpretation of the loose view discussed above but dismisses it without further argument. Bach, on the other hand (who, remember, suggested the first interpretation), also considers the just mentioned second interpretation but dismisses this interpretation without further argument. (He just claims that “it might seem extreme to suppose that most of what ordinarily passes for categorical knowledge is really conditional in content” (2010, p. 123).)
See Davis (2007, p. 420) for considerations that go into a similar direction.
See Dinges (2015b) for some discussion of the former issue.
See Davis (2007, p. 413) for related considerations. A common worry for pragmatic accounts of the bank case intuitions is that the relevant implicatures aren’t properly calculable. The above considerations are too vague to fully address this worry. Even so, they should suffice to shift the burden of proof to those who think that SPI falls prey to some such worry.
See Baumann (2011, p. 160) for the resulting case.
Blome-Tillmann endorses the above meaning determination principle only for expressions that are not “composed out of simpler expressions.” I ignore this restriction because we will be dealing only with non-composed expressions anyway.
I will later argue that SPI must be rejected because the knowledge claims in the bank cases fail relevant tests for non-literality. Note, however, that this objection is entirely independent from the general considerations on meaning determination above: If knowledge claims involve loose use, they should pass the tests for loose use whatever we think about the determinants of the meaning of an expression.
Recent empirical studies on the bank cases corroborate this assessment. It is striking that when people are asked about the truth-value of the knowledge claims in the bank cases, their responses generally range somewhere in the middle between “true” and “false.” See e.g. Hansen and Chemla (2013). These results strongly suggest that DeRose is at least not a paradigm case of an ordinary “knower.”
This move should also suffice to alleviate Hawthorne’s worry that, if SPI is true, “no one—not even the philosophically unsophisticated—ever believes that he knows such ordinary proposition as that Manchester United beat Coventry City 2–1, or that the plane from Detroit is late, or that the towels are in the dryer” (2004, p. 119). Assuming that, according to SPI, knowledge claims like the ones mentioned by Hawthorne normally do convey what they literally express, SPI is perfectly compatible with the view that ordinary speakers have the pertaining beliefs. See Douven (2007, 344 f.) for further discussion of this issue.
On this basis, we can also respond to the related worry that, unless we go for full-blown skepticism, there is no non-arbitrary way to set the boundary between knowledge and non-knowledge. See e.g. DeRose (2012, 712 f.). We should set the boundary between knowledge and non-knowledge such that the knowledge ascriptions that we would retract under slight skeptical pressure do not count as true and those that we wouldn’t so retract do count as true. (BonJour (2010, p. 58) worries that, if we set the boundary between knowledge and non-knowledge in this not fully skeptical way, we cannot explain why knowledge is “a supremely valuable and desirable cognitive state, one whose possession marks the difference between full cognitive success and at least some degree of cognitive failure”. This may be so. But I, at least, fail to see why knowledge should be such a state to begin with. Knowing is to be in a very good epistemic position. But this does not mean that one’s epistemic position could not be better.)
See Sosa (1999, p. 148) for a fallacy that might be involved in skeptical arguments. He argues that knowledge requires “safe” not “sensitive” beliefs. On that basis, he goes on to suggest that skeptical arguments may sound compelling because “[s]afety and sensitivity, being mutual contrapositives, are easily confused.”
Proponents of the present meaning determination objection may never have intended to argue against versions of SPI that don’t entail skepticism by definition. If that is so, the present response only shows that their objection doesn’t carry over to the form of SPI we are concerned with here (but see the first response above).
I should briefly mention some potential worries for SPI that, I think, have been answered elsewhere already: One may want to argue against SPI on the grounds that the alleged implicature should be felicitously cancelable, but isn’t. Worries along these lines seem properly addressed e.g. in Davis (2007, pp. 411, 431): The implicature cannot be (explicitly) canceled because it is entailed by what is literally said (as is the implicature that it is close enough to three o’clock of an utterance of “It is three o’clock”). One may want to argue against SPI on the grounds that, by appealing to loose use, the proponent of SPI can no longer explain disagreement and retraction data. Worries along these lines seem properly addressed e.g. in Davis (2007, 406 f.): It is perfectly appropriate to retract claims once it is pointed out that they are strictly speaking false. (See below for some further remarks on this issue.) Finally, one may want to argue against SPI on the grounds that loose use cannot explain why it would seem inappropriate for DeRose to deny knowledge in the low standards case. Worries along these lines seem properly addressed e.g. in Davis (2007, 408 f.): Negative statements like “It is not three o’clock” can be used loosely to convey that it is not close enough to three o’clock. The same goes for “I don’t know the bank will be open.” This sentence may also convey that DeRose is not close enough to knowing that the bank will be open, which is false in the low standards case.
The same goes for (substitutional) implicitures: According to Bach (1994), a typical utterance of “I have nothing to wear” expresses the falsity that the speaker has nothing to wear whatsoever but truly implicites that she has nothing appropriate to wear for a given occasion. Correspondingly, the following dialogue seems fine:
A: I have nothing to wear.
B: Sure you do. There are socks in the drawer, shirts on the shelf ...
A: Oh, c’mon. You know what I meant!
Similar dialogues could, I think, be devised for all relevant cases of implicitures. Thus, the present objection applies to SPI even if the proponent of SPI appeals to implicitures rather than implicatures.
My presentation of the awareness principle closely follows the presentation in Blome-Tillmann (2013, Sect. 3). Major points of divergence will be spelled out in the main text. Here I will briefly outline three minor points of divergences that should be mentioned nevertheless. First, Blome-Tillmann does not show that the awareness principle applies to implicitures. This makes the range of targeted positions unnecessarily narrow. Second, he does not identify the category of a substitutional implicature (or impliciture) as the category of implicatures to which the awareness principle applies. Rather, he seems to think that the relevant class of implicatures involves violations of Quality\(_{1}\) (the conversational maxim enjoining you to be sincere). See his principle (DP) and FN 39 on p. 4311. This is wrong, I think, because not all substitutional implicatures involve Quality\(_{1}\) violations, and even so, they all allow for the construction of dialogues of the above sort. We only have to gear B’s objection to whatever maxim is violated at the level of what is said. For example, suppose I say, “Jane will come or she won’t” in response to the question of what to do about Jane’s absence. I do not violate Quality\(_{1}\) because I do believe what I literally say. Still, I substitutionally implicate, say, that nothing can be done about Jane’s absence because what I literally say is entirely uninformative (it’s a mere logical truth that we all believe already anyway). Correspondingly, a stickler may object, “Well, these are all the options, aren’t they? I wanted to know what to do about Jane’s absence.” And, again, it would be fine to respond, “Oh, c’mon. You know what I meant! (There’s nothing we can do.)” Third, I have argued that B’s objections can be understood either in terms of the idea that B fails to see the implicature or in terms of the idea that she is being a stickler. Blome-Tillmann (2013, 4311 n.) envisages only the first kind of understanding. The stickler reading, however, often is the only plausible reading available (particularly when we consider implicitures).
One may worry that DeRose’s wife is addressing what DeRose means to convey in the dialogue above because she implicates that the possibility that the bank changes its hours not only exists but is likely enough to be relevant even for being close enough to knowledge. This would explain why the “You know what I meant” response is unacceptable. Even though I find it very plausible that DeRose’s wife does implicate something along these lines, I think the proponent of SPI cannot plausibly appeal to this implicature. First, this would threaten to make her view superfluous: Consider DeRose’s wife in the high standards case. As in (D1), she mentions the error-possibility that the bank has changed its hours. Given that in (D1), she implicates that this possibility is likely enough to be relevant for the epistemic position at issue in (D1) (that is, for being close enough to knowledge), it is very hard to deny that there should be a similar implicature in the high standards case. In particular, DeRose’s wife in this case should implicate that the mentioned error-possibility is likely enough to be relevant for the epistemic position at issue in the high standards case (that is, I take it, for knowledge strictly speaking). Once we grant this implicature, however, we can argue that DeRose’s wife changes DeRose’s epistemic position: In the high standards case, but not in the low standards case, he gains testimonial evidence that the bank is relevantly likely to change its hours. Correspondingly, we can say that our intuitions about the bank cases vary simply because DeRose’s epistemic position varies; SPI is no longer required. (See Dinges (2015a) for further elaboration on this kind of proposal and its application to so-called “third-person” bank cases.) Second, it doesn’t seem that the “You know what I meant” response would become any more appropriate if we let DeRose’s wife cancel the alleged implicature, for example, if we let her respond with “I doubt that. Not that this is likely to any relevant degree, but the bank could have changed its hours.” Hence, the awareness objection could easily be reinstated. Thanks to an anonymous referee for pushing the worry discussed in this footnote.
To be clear, I am happy to grant that, if the error-possibility DeRose’s wife mentions becomes ever more far-fetched, the “You know what I meant” response may become plausible at some point. For example, the response may seem fine when she inveighs, “I doubt that, we might be brains in a vat after all.” But this observation is of no avail to the proponent of SPI. So long as SPI is supposed to provide an answer to the bank case puzzle, it must be capable of dealing with less far-fetched alternatives such as the one in (D1). After all, these alternatives figure in the bank cases. (And they do so for a good reason: It is very much unclear whether more far-fetched alternatives would trigger the intuitions constituting the bank case puzzle. See e.g. Gerken (2012, 141 f.).)
Davis (2007, 410 f.) similarly distinguishes loose use from what he calls “sloppy” or “careless” or “unenlightened” use.
To be clear, note that the phenomenon of enlightenment in high standards contexts, though closely related, must not be confused with the phenomenon of retraction: that we tend to retract knowledge claims when someone mentions an error-possibility. As indicated already, loose use may explain retraction. Still, when we retract claims because it is pointed out that they are strictly speaking false, we do so with what MacFarlane dubs an “exasperated grumble” (2011, p. 541). We precisely don’t consider ourselves enlightened. So, loose use cannot explain enlightenment even though it can explain retraction.
References
Bach, K. (1994). Conversational impliciture. Mind & Language, 9(2), 124–162.
Bach, K. (2010). Knowledge in and out of context. In J. K. Campbell, M. O’Rourke, & H. Silverstein (Eds.), Knowledge and skepticism. Topics in contemporary philosophy (pp. 105–136). Cambridge, MA: MIT Press.
Baumann, P. (2011). WAMs: Why worry? Philosophical Papers, 40(2), 155–177.
Blome-Tillmann, M. (2013). Knowledge and implicatures. Synthese, 190(18), 4293–4319.
BonJour, L. (2010). The myth of knowledge. Philosophical Perspectives, 24(1), 57–83.
Cohen, S. (2005). Contextualism defended. In M. Steup & E. Sosa (Eds.), Contemporary debates in epistemology. Contemporary debates in philosophy (Vol. 3, pp. 56–62). Malden, MA: Blackwell.
Conee, E. (2005a). Contextualism contested. In M. Steup & E. Sosa (Eds.), Contemporary debates in epistemology. Contemporary debates in philosophy (Vol. 3, pp. 47–56). Malden, MA: Blackwell.
Conee, E. (2005b). Contextualism contested some more. In M. Steup & E. Sosa (Eds.), Contemporary debates in epistemology. Contemporary debates in philosophy (Vol. 3, pp. 62–66). Malden, MA: Blackwell.
Davis, W. A. (2004). Are knowledge claims indexical? Erkenntnis, 61(2/3), 257–281.
Davis, W. A. (2007). Knowledge claims and context: Loose use. Philosophical Studies, 132(3), 395–438.
Davis, W. A. (2010). The case for contextualism: Knowledge, skepticism, and context, vol. 1, by Keith DeRose. Mind, 119(476), 1152–1157.
Davis, W. A. (2015). Knowledge claims and context: Belief. Philosophical Studies, 172(2), 399–432.
DeRose, K. (2009). The case for contextualism: Knowledge, skepticism, and context (Vol. 1). Oxford: Oxford University Press.
DeRose, K. (2012). Replies to Nagel, Ludlow, and Fantl and McGrath. Philosophy and Phenomenological Research, 84(3), 703–721.
Dimmock, P., & Huvenes, T. (2014). Knowledge, conservatism, and pragmatics. Synthese, 191, 3239–3269.
Dinges, A. (2015a). Epistemic invariantism and contextualist intuitions. Episteme, FirstView, 1–14.
Dinges, A. (2015b). Innocent implicatures. Journal of Pragmatics, 87, 54–63.
Douven, I. (2007). A pragmatic dissolution of Harman’s paradox. Philosophy and Phenomenological Research, 74(2), 326–345.
Fantl, J., & McGrath, M. (2009). Knowledge in an uncertain world. Oxford: Oxford University Press.
Gerken, M. (2012). On the cognitive basis of knowledge ascriptions. In J. Brown & M. Gerken (Eds.), Knowledge ascriptions (pp. 140–170). Oxford: Oxford University Press.
Grice, P. (1989). Studies in the way of words. Cambridge, MA: Harvard University Press.
Hansen, N., & Chemla, E. (2013). Experimenting on contextualism. Mind & Language, 28(3), 286–321.
Hawthorne, J. (2004). Knowledge and lotteries. Oxford: Oxford University Press.
MacFarlane, J. (2005). The assessment sensitivity of knowledge attributions. In T. S. Gendler & J. Hawthorne (Eds.), Oxford studies in epistemology. Oxford studies in epistemology (Vol. 1, pp. 197–234). Oxford: Oxford University Press.
MacFarlane, J. (2011). Relativism and knowledge attributions. In S. Bernecker & D. Pritchard (Eds.), The Routledge companion to epistemology (pp. 536–544). New York: Routledge.
Meibauer, J. (2009). Implicature. In J. Mey (Ed.), Concise encyclopedia of pragmatics (pp. 365–378). Amsterdam: Elsevier.
Schaffer, J. (2004). Skepticism, contextualism, and discrimination. Philosophy and Phenomenological Research, 69(1), 138–155.
Sosa, E. (1999). How to defeat opposition to Moore. Philosophical Perspectives, 33(13), 141–153.
Unger, P. (1971). A defense of skepticism. The Philosophical Review, 80(2), 198–219.
Acknowledgments
I am grateful to Michael Blome-Tillmann, Elke Brendel, Aurélien Darbellay, Mikkel Gerken, Beate Krickel, David Lanius, Dan López de Sa, David Löwenstein, Erik Stei, Emanuel Viebahn, Julia Zakkou and the participants of the Fifth Annual Graduate Epistemology Conference (Edinburgh, 2015) for very helpful comments on earlier drafts of the paper.
Author information
Authors and Affiliations
Corresponding author
Rights and permissions
About this article
Cite this article
Dinges, A. Skeptical pragmatic invariantism: good, but not good enough. Synthese 193, 2577–2593 (2016). https://doi.org/10.1007/s11229-015-0867-1
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s11229-015-0867-1