Abstract
Stewart Cohen’s New Evil Demon argument raises familiar and widely discussed concerns for reliabilist accounts of epistemic justification. A now standard response to this argument, initiated by Alvin Goldman and Ernest Sosa, involves distinguishing different notions of justification. Juan Comesaña has recently and prominently claimed that his Indexical Reliabilism (IR) offers a novel solution in this tradition. We argue, however, that Comesaña’s proposal suffers serious difficulties from the perspective of the philosophy of language. More specifically, we show that the two readings of sentences involving the word ‘justified’ which are required for Comesaña’s solution to the problem are not recoverable within the two-dimensional framework of Robert Stalnaker to which he appeals. We then consider, and reject, an attempt to overcome this difficulty by appeal to a complication of the theory involving counterfactuals, and conclude the paper by sketching our own preferred solution to Cohen’s New Evil Demon.
This is a preview of subscription content, log in via an institution to check access.
Similar content being viewed by others
Notes
This way of introducing the name ‘Biv’ commits us to the Limit Assumption (Lewis 1973: 19–21) that there is exactly one closest world to actuality in which there is a brain in a vat. This assumption could be avoided at some technical cost. Equally we assume there is just one brain in a vat at the nearest world with a brain in a vat. These assumptions make no difference to the substance of our argument.
Comesaña (2002: 255) calls this claim “Demonic Reliability”.
Of course, one need not accept the conclusion of the New Evil Demon. Sceptics will reject premise (1), and ardent reliabilists will reject (NED)—as will we.
Comesaña talks about brains-in-vats (or victims of evil demons, to be precise) in the plural and does not refer to a particular brain-in-a-vat by means of a proper name, as we do. However, changing our version of the New Evil Demon to match his would merely complicate the formulation of the principles at issue slightly: instead of referring to Biv’s beliefs in w BIV we would, on his formulation, need to refer to the beliefs of brains-in-vats in their respective world’s w BIV1 … w BIVn.
Those who don’t will still have to blame (NED) or embrace scepticism.
See (BonJour 1980) for the clairvoyance cases. Sosa (2001: 390) recognizes that “we countenance possible higher beings who gain knowledge by properly and reliably forming (‘warranted’) beliefs, despite the fact that their epistemic ways, while successful in their world, would be miserably inadequate in ours.” He does not appear to acknowledge that this poses a threat to (R1).
The statement of (IR) is complicated slightly at (Comesaña 2010: 579); however, the additional clause added there is irrelevant to our concerns. One might also wonder whether (IR) ought to be formulated meta-linguistically, but we shall ignore these issues here.
The principle is due to Grice (1989: 47–49), who calls it “Modified Occam’s Razor”.
See Stalnaker (1978).
We assume here, as does Comesaña (in effect), that the context is one in which the conversational participants do not presuppose that there are no brains in vats. We also follow Comesaña in restricting our attention to just two worlds.
One complication is worth mentioning: on Stalnaker’s account it is not the case that both ‘readings’ are available in one and the same context. In this particular case, the horizontal cannot be asserted as it is true in every world in the context set. We discuss this issue in more detail below.
Comesaña (2002: 258) says “‘Ernie’s beliefs are justified’, then, is ambiguous; and so is [(3)], and for the same reasons.” And he qualifies his use of ‘ambiguous’ as meaning ‘expresses distinct diagonal and horizontal propositions’.
(Comesaña 2002: 256, emphasis in original).
The same assumptions are made here as were made in connection with [5I].
By ‘the horizontal’ we mean, of course, the top-most horizontal; for we are considering an utterance of (3I) in @.
Ironically, Comesaña introduces this kind of talk of truth and falsity simpliciter in a context in which it is inappropriate to assess a claim for truth simpliciter. He writes,
In the demoners’ world, taking experience at face value is highly unreliable: it usually yields false beliefs, and it would yield mostly false beliefs in counterfactual applications—in the demoners’ world. Here, in this world, taking experience at face value is highly reliable: it usually yields true beliefs, and this is not just a statistical correlation, for it would yield true beliefs if used in appropriate circumstances. So the crucial question is what we are referring to when we say that a justified belief must have been produced by a process most of whose outputs would be true. True where? The immediate answer is: true simpliciter, that is, true in the actual world. (Comesaña 2002: 256)
However, it is clear that the beliefs in question must be true, not in the actual world, but in their respective nearby variants of the actual world. Obviously Comesaña means that the belief forming process must be reliable in @, and not that the beliefs formed by this process must be true in @.
If it were presupposed that there is no brain in a vat then w BIV would not be compatible with what was presupposed, and so would not belong to the context set. Similarly, if something false were presupposed, @ would not belong to the context set.
Note that we are not disputing Comesaña’s contention that (3I) could have expressed a truth; only that it in fact does so in some actual context with [3I] as its propositional concept.
Of course, speakers can choose which of two meanings of an ambiguous word or phrase to convey; but Comesaña does not regard the two readings of the relevant sentences as disambiguations.
In fact, we only get the second reading for (3IC) if we replace ‘is’ with ‘would be’—perhaps with the word ‘actually’ occurring between ‘would’ and ‘be’. Let this claim be (3IC*). Then we can grant for the sake of argument—something that Comesaña needs—that (3IC*), like (3IC), is equivalent to (3I), under the hypothesis that ‘actually’ has the two senses Lewis postulates.
See previous note.
The example is due to Stalnaker (1999, 156).
See Heim and Kratzer (1998: 239–259).
Compare, in this connection, the following proposed definition of ‘good son’: for all x, x is a good son iff x loves his mother. If it is then claimed that sentences involving ‘good son’ have multiple readings on the grounds that ‘his’ can refer to x or to some contextually salient individual, it would not, we think, be plausible to maintain in addition that ‘good son’ is not ambiguous.
In fact, in such a context it would be appropriate to utter the corresponding indicative conditional. But on Stalnaker’s view, indicative and subjunctive (counterfactual) conditionals express the same kind of conditional proposition; so this does not affect our point.
The effect is that the consequent of (3IC), on the reading we are now considering, where ‘actually’ picks up its reference from the derived context, is (3-2); accordingly, (3IC) itself is false (at both @ and w BIV). Of course, there remains the reading where ‘actually’ is simply interpreted as referring to @; the consequent of (3IC) is then equivalent to (3-1), and the conditional itself is true (at both @ and w BIV). (Notice, finally, that this vindicates our association of [3I] with (3I) in the previous section.).
In fact, note that if Comesaña’s were to admit that his approach is a version of the ambiguity theory, his approach would be significantly more complex than more straightforward ambiguity accounts of ‘justification’: on the type of ambiguity approach we are considering now, the ambiguity in ‘justified’ would be derived from the ambiguity of ‘actually’. Surely, it would be theoretically much simpler to claim directly that ‘justified’ is ambiguous between the two readings (R1) and (R2) than to take the explanatory detour of going via (IR).
If interpretations on which the referents of these terms are displaced away from the parametric value determined by the context of utterance are available, they are hard to come by. For instance, one can use ‘here’ to refer to a place one does not occupy if one is standing in front of a map and pointing at a representation of a place one is not in; but the circumstances which make this possible are quite particular. It does not seem that the relevant interpretations are systematically available in the way required by Comesaña’s strategy.
In effect, we are appealing here to the thought that, whatever lies behind Cohen’s original intuition that brains in vats have justified but unreliable beliefs, the intuition would be equally legitimate were we to consider as open (i.e. consider as actual) the possibility that we ourselves are brains in vats.
Note that there is a similarity between our view on this issue and Goldman’s (1988), with the crucial difference, however, that we do not consider blamelessness a type of epistemic justification (Goldman’s ‘weak justification’ is partly defined in terms of blamelessness). On our account, however, ‘justified’ is neither ambiguous, nor polysemous, and also not indexical in the way suggested by Comesaña.
(Cohen 1984: 283), for instance, claims that “‘rational’ is a “virtual synonym […] for ‘justified’”.
See, for instance, (Strawson 1952: 256), who claims: “It is an analytic proposition that it is reasonable to have a degree of belief in a statement which is proportional to the strength of the evidence in its favour.” See also (Locke 1690 [1975]: IV, xix, 1.; Hume 1777 [1902]: X, i, p. 87; Quine and Ullian 1978: 16; Williamson 2000: 192).
They are too high for some propositions, such as that she has hands, and too low for others, e.g., their negations.
Sutton (2005: 373) seems sympathetic to this view, claiming that there is “no reason to think that reasonableness can be identified with justification in the philosophers’ sense” and that “reasonable people are sometimes, through no fault of their own, in the wrong conditions, and then they form unjustified beliefs.” However, it should be mentioned here that Sutton (ibid.) also suspects that “the notion of reasonableness […] is far too vague” to conclusively settle the issue of whether or not Biv is reasonable in forming the beliefs that she forms.
One further objection to our view is also of interest. In particular, it might be held that a belief is justified if it would be irrational to suspend judgement on the issue which it concerns. But then, since Biv can’t discriminate her situation from ours, it would be irrational for her to suspend judgement on the question of whether she has hands in the hope that new evidence might come to light which would settle the issue. Accordingly, her belief that she has hands is justified; and similarly, mutatis mutandis, for her other beliefs. Fascinating as the issues raised by this objection are, we cannot address them adequately here; for doing so would take us too far afield. For instance, in order to respond to them fully, we would have to decide whether suspending judgement is simply a matter of having partial, rather than full, outright belief, or whether it is its own, sui generis attitude—a question which deserves a full paper length treatment of its own. A second issue is whether we should think of the rationality at issue as theoretical or practical; and the relationship between these is also a complex and involved matter. Accordingly, we do no more here than note this objection, and leave readers to make of it what they will, perhaps returning to address the issues it raises elsewhere.
Cohen (1984: 282) insists on this point, too, but draws different conclusions from it than we do.
Thanks to Jane Friedman on this point.
Note that since the above concern is that something different is meant when ‘justified’ is used on different occasions, it doesn’t so much matter what the semantic mechanism is which brings about this result. Thus, the above concern about running the risk of divorcing normativity from epistemic justification concerns both Comesaña and ambiguity theorists such as Sosa equally.
References
BonJour, L. (1980). Externalist theories of empirical knowledge. Midwest Studies in Philosophy, 5, 53–73.
Cohen, S. (1984). Justification and truth. Philosophical Studies, 46(3), 279–295.
Comesaña, J. (2002). The diagonal and the demon. Philosophical Studies, 110(3), 249–266.
Comesaña, J. (2010). Evidentialist reliabilism. Noûs, 44(4), 571–600.
Conee, E., & Feldman, R. (1985). Evidentialism. Philosophical Studies, 48, 15–34.
Feldman, R., & Conee, E. (2001). Internalism defended. American Philosophical Quarterly, 38(1), 1–18.
Fumerton, R. (1988). The internalism/externalism controversy. Philosophical Perspectives, 2, 443–459.
Goldman, A. I. (1988). Strong and weak justification. Philosophical Perspectives, 2, 51–69.
Grice, H. P. (1989). Studies in the way of words. Cambridge, MA: Harvard University Press.
Hawthorne, J., & Magidor, O. (2009). Assertion, context, and epistemic accessibility. Mind, 118(470), 377–397.
Heim, I., & Kratzer, A. (1998). Semantics in generative grammar. Oxford: Blackwell.
Hume, D. (1777 [1902]). An enquiry concerning human understanding and concerning the principles of morals. Oxford: Oxford University Press, L. A. Selby-Bigge.
Lewis, D. (1970). Anselm and actuality. Nous, 4(2), 175–188.
Lewis, D. (1973). Counterfactuals. Oxford: Blackwell.
Locke, J. (1690 [1975]). An essay concerning human understanding. Oxford: Clarendon.
Pryor, J. (2001). Highlights of recent epistemology. British Journal for the Philosophy of Science, 52, 95–124.
Quine, W. V., & Ullian, J. S. (1978). The web of belief. New York: McGraw-Hill.
Soames, S. (2005). Reference and description: the case against two-dimensionalism. Princeton, NJ: Oxford, Princeton University Press.
Sosa, E. (1993). Review: Proper functionalism and virtue epistemology. Noûs, 27(1), 51–65.
Sosa, E. (2001). Goldman’s reliabilism and virtue epistemology. Philosophical Topics, 29(1/2), 383–400.
Stalnaker, R. (1978). Assertion. Syntax and Semantics, 9, 315–332. Reprinted in: Stalnaker, R. (1999). Context and content (pp. 78–95). Oxford: OUP (page references are to reprint edition).
Stalnaker, R. (1999). Context and content: Essays on intentionality in speech and thought. Oxford: Oxford University Press.
Strawson, P. F. (1952). Introduction to logical theory. London: Methuen.
Sutton, J. (2005). Stick to what you know. Noûs, 39(3), 359–396.
Sutton, J. (2007). Without Justification. Cambridge, MA: MIT Press.
Williamson, T. (1997). Knowledge as evidence. Mind, 106(424), 717–741.
Williamson, T. (2000). Knowledge and its limits. Oxford: Oxford University Press.
Author information
Authors and Affiliations
Corresponding author
Additional information
Thanks to two anonymous referees for this journal. The paper is fully collaborative; authors are listed alphabetically.
Rights and permissions
About this article
Cite this article
Ball, B., Blome-Tillmann, M. Indexical Reliabilism and the New Evil Demon. Erkenn 78, 1317–1336 (2013). https://doi.org/10.1007/s10670-012-9422-3
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s10670-012-9422-3