Skip to main content
Log in

How to Forget that “Know” is Factive

Acta Analytica Aims and scope Submit manuscript

Abstract

This paper examines, and rejects, a recent argument to the effect that knowledge is not truth-entailing, i.e. that “know” is not factive.

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

Similar content being viewed by others

Notes

  1. Hazlett (2010). Unless otherwise indicated, all parenthetical page references are to that paper.

  2. Asterisks will be used to mark unacceptability, rather than ungrammaticality, throughout.

  3. I am assuming here that if, for example, A claims that he knows that a certain thing is not self-identical, and B claims that no one could know such a thing, the burden of proof would evidently fall on A rather than on B.

  4. Note that the title’s point is re-affirmed elsewhere in Hazlett’s text, for example, in his statement, “I’ve called factive verbs a myth, and I mean something by that”(500).

  5. For some independent problems facing Hazlett’s purported derivation of p as a conversational implicature of “S knows that p,” see the present paper’s Appendix.

  6. Hazlett obliquely admits the relative rarity of acceptable conjunctions of “S knows that p” with the negation of p, when he claims, speaking of uses of “S knows that p” whose conjunction with the negation of p would be unacceptable, that they are “the most common uses of “knows” we have”(514).

  7. See Stjernberg (2009) for some discussion of this issue.

  8. Notice that, since there would be little point to the notion of conversational implicature if conversational implicatures could not be effectively distinguished from entailments, and since the primary evidence held to ensure the possibility of such a distinction is that conversational implicatures are cancellable whereas entailments are non-cancellable, commitment to the diagnostic value of cancellability is not optional for those employing the notion of conversational implicature: if one’s reliance on cancellability turned out to be problematic, so would be one’s reliance on the notion of conversational implicature. [Weiner’s (2006) recent discussion of cancelability, in addition to other limitations pointed out by Blome-Tillmann (2008) and by Borge (2009), does not clearly acknowledge that point.]

  9. I would like to thank Allan Hazlett for his comments on the present paper’s penultimate version.

References

  • Blome-Tillmann, M. (2008). Conversational implicature and the cancellability test. Analysis, 68, 156–160.

    Google Scholar 

  • Borge, S. (2009). Conversational implicatures and cancellability. Acta Analytica, 24, 149–154.

    Article  Google Scholar 

  • Grice, H. P. (1989). Studies in the way of words. Cambridge: Harvard University Press.

    Google Scholar 

  • Hazlett, A. (2010). The myth of factive verbs. Philosophy and Phenomenological Research, LXXX, 497–522.

    Article  Google Scholar 

  • Kvanvig, J. T. (2006). The paradox of knowability. Oxford: Oxford University Press.

    Book  Google Scholar 

  • Millikan, R. G. (2005). Language: A biological model. Oxford: Oxford University Press.

    Book  Google Scholar 

  • Stjernberg, F. (2009). Restricting factiveness. Philosophical Studies, 146, 29–48.

    Article  Google Scholar 

  • Weiner, M. (2006). Are all conversational implicatures cancellable? Analysis, 66, 127–30.

    Article  Google Scholar 

Download references

Author information

Authors and Affiliations

Authors

Corresponding author

Correspondence to Savas L. Tsohatzidis.

Appendix

Appendix

As explained in the text, Hazlett’s hypothesis that utterances of “S knows that p” conversationally implicate, rather than entail, p, is precluded by the cancellability requirementFootnote 8; it may be interesting to note, however, that that hypothesis suffers from important further problems as well.

As is well known, the conversational implicatures of an utterance of a sentence are calculated, according to Grice, on the basis of what, in uttering that sentence, a speaker literally says. Hazlett, however, claims that when, in uttering a sentence of the form “S knows that p,” a speaker conversationally implicates p, the alleged implicature is calculated simply on the basis a certain thing that, in Hazlett’s view, these sentences entail. Specifically, Hazlett hypothesizes that sentences of the form “S knows that p” entail corresponding sentences of the form “S has the epistemically warranted belief that p;” he then claims that, since “it is a conceptual truth that epistemically warranted beliefs tend to be true” (508), and since, given this allegedly conceptually guaranteed ‘tendency,’ a cooperative speaker who ascribes an epistemically warranted belief to a subject S will be taken to conversationally implicate, unless he provides explicit indications to the contrary, that the belief he ascribes to S as epistemically warranted is true, utterances of “S knows that p” come to conversationally implicate p.

There are crucial elements of this account that one might wish to question. Thus, one might argue that, contrary to what Hazlett assumes, sentences of the form “S knows that p” do not entail corresponding sentences of the form “S has the epistemically warranted belief that p,” since, for example, (a) and (b) could be true even though (a') and (b') could be false (assuming that they would not be nonsensical):

  1. (a)

    Mary knows that she is hungry/thirsty/angry.

  2. (a')

    Mary has the epistemically warranted belief that she is hungry/thirsty/angry.

  3. (b)

    Mary knows that she is attracted to the man she just saw.

  4. (b')

    Mary has the epistemically warranted belief that she is attracted to the man she just saw.

But even assuming that Hazlett’s hypothesized entailment relation obtains, his account faces the problem that it requires acceptance of a clearly false thesis regarding conversational implicature. To see this, notice that Hazlett professes not to know anything about what “S knows that p” means except that it does not entail p and that it does entail “S has the epistemically warranted belief that p”—in particular, he is explicitly agnostic as to whether his epistemic warrant condition is “sufficient for the truth of ‘S knows p’” (509), and he is equally agnostic as to “whether there are additional necessary conditions on the truth of ‘S knows p’” (509). But anyone who is to that extent non-committal about what the truth conditions of “S knows that p” are—and so, about what, in the Gricean sense, utterers of sentences of that form literally say—would hardly seem to be entitled to make hypotheses as to what such utterers conversationally implicate (after all, implicature-generating inferences are supposed to be inferences that derive, on the basis of a full specification of what speakers literally say, what these speakers mean without literally saying). To justify his procedure, then, Hazlett would have to specify a special principle whose truth would entitle him to make hypotheses as to what utterers of “S knows that p” conversationally implicate, even in the absence of full information as to what they literally say. And, as far as I can see, the only principle that, if true, would provide him with the requisite entitlement, would be the following:

(H) If a sentence S1 entails a sentence S2, then whatever can be conversationally implicated by utterances of S2 can also be conversationally implicated by utterances of S1.

Principle (H), however, is certainly false. For example, an utterance of (i) can conversationally implicate (ii), but no utterance of (iii) can conversationally implicate (ii), even though (iii) entails (i):

  1. (i)

    It’s cold in here.

  2. (ii)

    I want you to close the windows.

  3. (iii)

    It’s cold in here, and all the windows are closed.

Similarly, an utterance of (iv) can conversationally implicate (v), but no utterance of (vi) can conversationally implicate (v), even though (vi) entails (iv):

  1. (iv)

    John is nervous sometimes.

  2. (v)

    It is not the case that John is always nervous.

  3. (vi)

    John is always nervous.

If, however, principle (H) is false, then the supposed fact that “S knows that p” entails “S has the epistemically warranted belief that p,” together with the supposed fact that utterances of “S has the epistemically warranted belief that p” conversationally implicate p, does not authorize the conclusion, which Hazlett was interested in drawing, that utterances of “S knows that p” themselves conversationally implicate p. In short, Hazlett’s purported derivation of the truth of p as a pragmatic implicature, rather than as a semantic entailment, of utterances of “S knows that p” would be unsuccessful even if it were not already disqualified by the cancellability requirement.Footnote 9

Rights and permissions

Reprints and permissions

About this article

Cite this article

Tsohatzidis, S.L. How to Forget that “Know” is Factive. Acta Anal 27, 449–459 (2012). https://doi.org/10.1007/s12136-012-0150-8

Download citation

  • Received:

  • Accepted:

  • Published:

  • Issue Date:

  • DOI: https://doi.org/10.1007/s12136-012-0150-8

Keywords

Navigation