Bergmann’s Dilemma: Exit Strategies for Internalists

25 It is worth noting here that this fact about versions of WAI allows them to avoid the “vicious regress” horn of Bergmann’s dilemma that we discuss in the next section of this paper. Since no such higher-order conceptualizations are required, there is no inevitable regress of conceptualizations in need of additional justification in the ways that Bergmann suggests (concerning versions of strong awareness internalism). Even when some conceptualizations are involved in the states of weak awareness—perhaps, e.g., when the object of weak awareness is a proposition or the noticed relationships between its conceptual constituents—WAI requires neither that such conceptualizations themselves be justified (since it is, for example, just the seeming that results from the noticing of conceptual relationships that is relevant to justification, and not the various conceptualizations that may themselves give rise to the seeming), nor that the justification for such conceptualizations, should it obtain, itself be conceptualized in some way (ad infinitum). (For more on these issues, see the next section of this paper, particularly our notes 33.)

26 Fumerton (2007). Fumerton's version of WAI can employ our defensive strategy with respect to direct acquaintance states by arguing that it is intuitively the case that the truth of the relevant belief is not “accidental” from the subject's perspective in the case where a subject is in a state of direct acquaintance with (i) that belief, (ii) the fact that makes the belief true, and (iii) the correspondence between these two things. Indeed, Fumerton appears actually to endorse this defensive strategy, in his (2007). We take our remarks to be elucidating that sort of response, identifying the general strategy underlying it, and showing how it applies to various versions of WAI.

27 We mentioned in our note 8 that such conceiving may or may not amount to belief. If it does not, we hereby register some uncertainty that it requires justification, though we will not press this concern here.

28 The complexity of the concepts being applied is ever increasing, since each new concept-application refers to the prior concept-application. To see the complexity most clearly, note that the ‘A1’ in A2 is simply a place-holder for the content of A1 itself. So, stated more explicitly, A2 is: S’s application to X2 of the concept being in some way related to the appropriateness of S’s application to X1 of the concept being in some way related to the appropriateness of B. The complexity increases for each step of the regress. See Bergmann (2006a, 18).

Some may think that we have already shown that the complexity regress does not go through by simply naming the prior levels of conceptualizations, rather than referring to them by their complex contents. After all, couldn’t any subject do this, and so avoid having to form increasingly complex conceptualizations? The answer appears to us to be ‘No’ (and so we hereby disagree with the direct reference solution to the complexity regress as it is formulated by Crisp (2009)). This is because each of these conceptualizations must be justified for the subject. In order for them to be justified for the subject, the subject will (at least typically) have to give explicit consideration to their complex contents in the ways that we go on to suggest.

29 Bergmann (2006a, 16).

30 For one contemporary example of the appeal to experiences as regress-stopping justifiers, and of the claim that experiences neither can be nor need to be justified, see Pryor (2005).

31 It is important to note that, as we will be understanding the term here and in what follows, “justifies” does not mean on-balance justifies. Rather, it simply means provides a reason for believing. We would use the latter expression, were it not so wordy.

32 As do Feldman and Conee (2008), for example.

33 We will not provide a further defense of the view that seemings can be foundational justification-contributors here. Our goal is just to indicate plausible strategies that internalists can employ in avoiding Bergmann’s dilemma. The idea that seemings are providers of justification is a prevalent and seemingly plausible view defended by Huemer (2001), Conee (2004), and others; Bergmann himself takes seemings to provide reasons in the way that we suggest (see our note 21). The idea that seemings can be foundational justifiers is also prevalent and plausible. Seemings, as we have characterized them, are experiences (they are experienced inclinations). Being experiences, they are not themselves subject to being justified or unjustified. As Conee and Feldman recently put it, “Its seeming to S that E supports P is a non-doxastic state, not a belief that E supports P . . . [A] seeming is not the sort of thing that needs justification. So, if a seeming justifies, then it is capable of stopping a regress of justifiers that are in need of justification” (2008, p. 96). Huemer (2001) endorses this idea, too.

As an aside, we note that the preceding remarks about seemings give some indication why the “seeming” version of WAI that we defended straightforwardly avoids the regresses facing versions of SAI.

34 We urge the reader to recall our cautionary note regarding the word “justifies” (see our note 31).

35 We again defer to Huemer (2001) and Conee (2004) for defenses of something near to this universal proposition.

36 Two notes are worthwhile here. First, it may seem to some readers that the parenthetical remark is ill-formed—what might it be for one to conceptualize that p? In our terminology, conceptualizing that p amounts to conceiving of the subject of p in the way that p suggests (e.g., conceiving of the subject of p as falling under the predicate-concept of p). Second, we simply note that, here and in what follows, we often forego explicitly stating the universal quantifier for the sake of ease of expression.

37 The brackets here and elsewhere are offered as a reader’s aid. The first bracketed clause is the content of E*. The second bracketed clause is the content of C2, which contains two other bracketed clauses that are, respectively, the contents of E2 and C1, i.e., the relata mentioned in C2.

38 The reader may worry that additional experiences will still be needed, since S will have to justifiably conceive of E* as relevant to C3, and so on. We address this worry in what follows.

39 For an example of this complexity, and of Bergmann’s understanding of the complexity regress in this way, the reader is directed to our note 28.

40 It is not terribly important that the reader perfectly grasp these contents, for reasons that will become apparent in what follows. The most important thing for the reader to note is that, as we go on to say, C4 and C5 involve repeated iterations of the same partial contents.

41 In effect, the reiterations involved here are analogous to reiterations of the “it is true that” clause of “it is true that p” to get (e.g.) “it is true that it is true that it is true that it is true that p,” in at least the respect that the reiterations in either case add nothing that requires additional substantive justification. Essentially the same reasoning and evidence that justifies C4’, for example, will justify C5’, and so on, just as essentially the same reasoning and evidence will justify “it is true that it is true that p” and “it is true that it is true that it is true that p.” This is true even if the resulting propositions are in some ways more difficult to comprehend. Any individual who goes through the reasoning that we have just exhibited will see that the more complex propositions are readily justified in the same manner as the less complex ones, thereby eliminating his need to explicitly grasp the particular contents of each in order to be justified in (at least de re) believing them to be true. We explain this further in the body of the paper. For further defense of these ideas, see our note 43.

As an additional dialectical point here, we note that Bergmann himself, in addressing the view of Evan Fales (1996), considers the “it is true that” iterations that we have just mentioned and concedes that they do not involve “any harmful increase in complexity” (2006a, 41; emphasis added). We take it that the conceptualizations mentioned in this paper also ought not to be regarded as “harmfully” complex, for the reasons just given. Thus, though we think that Bergmann is right that Fales’ view—which is similar to the one discussed here, in that it also involves the claim that internalism will require allegedly harmless conceptual iterations—generates unbearably complex propositions (via iteration of the clause, “it is transparent that…”), which themselves require additional substantive justification, we think that the view that we defend here avoids this difficulty entirely.

42 We note that our characterization of strong awareness (following Bergmann’s own characterization) makes no requirement that any belief or conceptualization be de dicto rather than de re, nor do we see why it should. Justified de re beliefs such as the one described here seem clearly to make it the case that the subject takes the object of awareness to be justificatorily relevant to his belief in a way that avoids the SPO. The subject’s object-level belief clearly does not remain “accidental’ from his perspective when his perspective includes such justified beliefs.

43 The case is in this respect analogous to a case where S knows that p and that the Double Negation replacement rule is truth-preserving.  Suppose that S contemplates whether [it is not the case that it is not the case that p] is true.  Knowing that p itself is true, S concludes that the more complex proposition is true.  Suppose that S then continues this pursuit, thinking about the propositions created by continuing to add two negations to the propositions at earlier levels.  At some point in the not too distant future, these propositions will become too complex for S to believe de dicto, but it still seems that S can justifiably believe things about these complex propositions.  For instance, since S knows that at each stage two negations are being added to the preceding proposition, and since S knows that such a maneuver is truth-preserving (and that p is true), it seems that S can justifiably believe de re that the increasingly complex propositions are true as well.  S can justifiably believe this about these propositions even though S's mind cannot grasp these propositions due to their complex contents. Likewise, S can justifiably believe de re, as a result of undergoing the reasoning that we present, that the increasingly complex propositions we mention in this paper are true, and this without needing to outright grasp their complex contents.

44 We reiterate here that the regress mentioned in that premise is, according to Bergmann, the complexity regress, and is not a regress having to do with the (potentially infinite) number of beliefs or conceptualizations. We have thus refuted that premise. Still, for the sake of completeness, we briefly consider the issue of the number of requisite beliefs or conceptualizations in what follows.

45 Bergmann suggests something like this response in his (2006a, 16).

46 It is at least plausible to think, for example, that many of us occurrently but tacitly believe that 10 is greater than 9, that 11 is greater than 9, that 12 is greater than 9, etc. For a defense of the claim that tacit beliefs are actual beliefs, different from mere dispositions to believe, see Crimmins (1992, 58-74).

47 Though we have just mentioned conceptualizations or beliefs, we will go on to speak only of beliefs, noting here that a parallel account could be given replacing “beliefs” with “conceptualizations.”

48 Bergmann (2006a, 16).

49 One might worry here that even potential awareness versions of SAI face a version of the “infinite conceptualizations” regress that we have not yet considered. For it seems that such versions of SAI may still require for justification that a subject be able to hold an infinite number of justified conceptualizations at some time, even if they do not require that he actually do so. And perhaps, the thought goes, no (finite, human) subject is even able to hold an infinite number of justified conceptualizations at some time, so that no such subject is able meet the justification requirements of even potential SAI. (Thanks to Earl Conee here.)

It seems to us, however, that versions of potential SAI do not actually have this requirement. They do not require that a subject be able to hold an infinite number of justified conceptualizations at a time. Rather, what such views require is just that, for every individual conceptualization that is a member of this infinite set of conceptualizations, the subject is able to (justifiably) hold it at the relevant time. This is the difference between such views requiring, on the one hand, that it is possible that, at t, the subject form every conceptualization and, on the other hand, that for every conceptualization, it is possible that the subject form it at t. The latter is true, does not require potentially holding an infinite number of beliefs at a time, and is all that appears to be required by versions of potential SAI.

50 On this account, the analogue to E* would be DA*:

DA*: S’s having the thought that direct acquaintance justifies while being directly acquainted with the fact that direct acquaintance justifies and the correspondence between this thought and this fact.

51 Though we have noted this on occasion throughout Section III of this paper, it is worth explicitly reiterating here that the versions of SAI defended so far also do not fall prey to the SPO (since they always require that the subject have a perspective on his evidence that relates it to the truth of the relevant conceptualization(s) or belief(s)), and so avoid Bergmann’s dilemma entirely.

52 We take ourselves to have addressed the seemingly pressing objection to (or, really, concern with) our defense of WAI—namely, the concern that the view defended there is not really a version of internalism or is inconsistent with the internalist “spirit”—in Section III of this paper.

53 We would like to put particular emphasis on the word “potentially” here.

54 Again, such views appear to be considered live options by Bergmann. See Bergmann (2006a, 14-19; p. 14 in particular). We note also that the view under discussion here differs from what was previously called “potential SAI” in that it allows for both potential conceptualization and potential awareness; “potential SAI” allowed only for potential conceptualization/belief.

55 Even if such reasoning is difficult or not immediately obvious, it still appears to be such that it is able to be accomplished, on careful reflection, by a good number of people.

56 One might worry that any version of SAI that appeals to the mere potential awareness of the justification contributor is now subject to the SPO, and so to the dilemma’s other horn. (Thanks again to an anonymous referee here.) Indeed, we have this worry ourselves. Nevertheless, we mention this possibility to leave open a further possible response to the dilemma that may avoid some of the skeptical worries potentially induced by the versions of SAI so far defended. One particular possibility that we find promising in this vein is a view that requires actual strong awareness of the justification-contributor at the object-level, but merely potential awareness of the justification-contributors at all meta-levels. It seems to us that such a view will avoid the SPO, at least concerning the first-order belief in question, and will also avoid entailing skepticism. But we leave this only as a last suggestion, and are unsure whether the view could avoid charges of arbitrariness.

57 In his defense of premise V, Bergmann explains why its resulting in radical skepticism has the result that we should not accept internalism. He says, “It is one thing to think that justification clearly requires something one naturally expects it to require, such as a good reason, and then to acknowledge that most of our beliefs lack such a reason. An internalist (open to skepticism) who thought this would insist that if we have no good reasons for any of our beliefs, so much the worse for our beliefs; she wouldn’t be inclined to think that we don’t need good reasons for them after all. But it’s another thing to think that, in order for any belief to be justified, the person holding it must have (or have the ability to form) an infinite number of beliefs of ever-increasing complexity. For in that sort of case, the problem isn’t merely that the requirement has skeptical implications. It’s also that the requirement just seems excessive, especially when stated so bluntly. I think this is why the internalist, in response to my discussion of strong awareness, will not (and should not) be inclined to adopt the radical skeptical results in question” (2006a, 22).

58 See Bergmann (2006a, Ch. 7). As one particular example, Bergmann explicitly claims that “it’s possible to depend on X1 to justifiedly believe that X1 is reliable” (2006a, 196).

59 Fumerton (2006, 180) makes a similar point. He claims that if a view like the one we are endorsing here (about seemings as reasons) is true, then it would simply be unreasonable not to let a proponent of the view defend his view by appealing to the very things it says are reasons. Similar claims are defended in Conee (2004).

60 For remarks on the sorts of circularity that are problematic, see again Bergmann (2006a, Ch. 7).

61 The sort of circularity endorsed here, if any circularity is endorsed, seems to be an instance of Van Cleve’s (epistemically harmless) ‘rule circularity’. See Van Cleve (1984).

62 Special thanks to Earl Conee and to an anonymous referee at Philosophical Studies for helpful comments on earlier drafts of this paper. We have also benefited from conversation with John Shoemaker.