Epistemic Poverty, Internalism, and Justified Belief: A Response to Robert Lockie, Nikolaj Nottelmann

Author Information: Nikolaj Nottelmann, University of Southern Denmark, nottelmann@sdu.dk

Nottelmann, Nikolaj . “Epistemic Poverty, Internalism, and Justified Belief: A Response to Robert Lockie.”[1] Social Epistemology Review and Reply Collective 5, no. 1 (2016): 12-28.

The PDF of the article gives specific page numbers. Shortlink: http://wp.me/p1Bfg0-2y8

Please refer to:


Image credit: U.S. Army, via flickr


In his recent Social Epistemology article “Perspectivism, Deontologism and Epistemic Poverty” Robert Lockie aims to disarm the so-called “epistemic poverty objection” to the deontological conception of epistemic justification (DCEJ). I first offer a regimentation of that objection, inspired by Laurence BonJour. I then turn to examining Lockie’s counter-arguments. As it turns out, rather than addressing directly conceptual issues within epistemology, Lockie’s main efforts go into arguing that generally epistemic subjects from outside contemporary advanced communities are not as poverty-stricken, as some modern epistemologists may have thought. I review Lockie’s arguments to that conclusion as well as alternative ways of arguing for a similar point, and conclude that they do not decisively undermine the poverty objection. I then turn to Lockie’s argument that a suitable version of epistemic access-internalism may successfully counter the poverty objection. I here conclude that the version of access-internalism Lockie needs is non-standard as well as implausible. The upshot is that even if Lockie’s article has brought several interesting and original concerns to bear on the debate over DCEJ, he has not defeated the poverty objection. 

The Epistemic Poverty Objection

The epistemic poverty objection, in its recent form, was developed by Laurence BonJour (even if arguably the underlying idea is much older).[2] In 2002 BonJour wrote:

It is certainly possible that a person’s epistemic situation, the kinds of cognitive tools and methods of inquiry available to him or her, might be so dire and impoverished (as a result of either individual or cultural deficiencies) as to make it difficult or impossible to come up with strong evidence or good epistemic reasons for belief about many important matters.[3]

Shortly after this passage, BonJour passes the following diagnosis:

Cases of epistemic poverty are cases in which it seems possible to fulfill one’s epistemic duty without being epistemically justified.[4]

It follows that any conception of epistemic justification (EJ) committed to the falsity of that modal claim must itself be false. But according to BonJour, the deontological conception of epistemic justification (DCEJ) is thus committed: “The central claim of the deontological conception of [epistemic] justification is that satisfying one’s intellectual duty or responsibility in relation to the acceptance of a particular belief is both necessary and sufficient for that belief to be epistemically justified.” [5] So, in virtue of its sufficiency claim, DCEJ must be false. This, in short, is the poverty objection to DCEJ.

Before we may begin to assess BonJour’s argument in detail, clarifications seem needed. E.g. how much, in the relevant context, need we say about contentious notions such as “good epistemic reasons” and “intellectual duty”? Concerning the first term, I propose that at least initially we follow BonJour’s lead and unpack this notion in terms of truth conduciveness. In BonJour’s preferred words, good epistemic reasons offer a relevant subject “at least some chance of finding the truth”.[6] As we shall soon, this seems precise enough for our present purposes.

Relative to a proposition p and a subject S, I suggest we take an epistemic reason to be something (a perceptual episode etc.) indicating to S either that p is true or indicating to S that p is false. Following BonJour’s lead, relative to a situation, we may then take a good epistemic reason for S to believe/disbelieve p to be an epistemic reason, such that, should S in that situation come to believe what the epistemic reason indicates, S would then have a fairly high chance of believing the truth concerning p.   What exactly constitutes a sufficiently high chance or even “some” chance, luckily we may leave vague here. All that matters, as we shall soon see, is that at least sometimes, for some possible configurations of a subject, a proposition, and a situation, plausibly no good epistemic reasons are available, yet belief or disbelief is nevertheless irresistible.

This last claim is a psychological one. Yet, it is easily underpinned by philosophical considerations naturally falling out of the suggested framework: The above conception of a good epistemic reason allows us to enforce an appearance-reality distinction and separate epistemic reasons that are really good (relative to a subject and a situation) from those that are only apparently so. The only bit of proper empirical psychology we need now add to the mix is the seemingly plausible thesis that sometimes subjects cannot help forming beliefs based on what appears to them to be good epistemic reasons, even if those reasons are not really good ones.

If we allow the proponent of the poverty objection this thesis, it is clear that her commitments regarding the extent and nature of “intellectual duty” need not extend beyond the minimal. All she needs is the normative principle that (dutifully) ought implies (psychologically) can, together with the plausible empirical thesis that in some scenarios some subjects psychologically could not have brought themselves in a position to discover that certain epistemic reasons, which to them appear to be good ones, are in fact no good. No matter the full extent of our “intellectual duty”, since it is some kind of duty, it then follows that it cannot involve resisting belief under such circumstances.

We may now reconstruct a Bonjour-style poverty objection to the deontological conception of epistemic justification as follows:

1. In order for a subject S to be epistemically justified in believing a proposition p in some situation C, S’s belief that p must be supported by good epistemic reasons for S to believe p in C;

2. [BonJour-style understanding of epistemic reason-goodness] In order for a reason to be a good epistemic reason for S to believe p in C, in C S must have a fairly high chance of believing the truth concerning p, should she believe as the reason indicates;[7]

3.[Appearance-reality distinction] It is possible for an epistemic reason to appear very good to a subject in a situation, even if here it is not really a good epistemic reason;

4. There are possible situations and subjects, such that here a subject psychologically could not have brought herself not to believe what is indicated by reasons that appear to her to be good epistemic reasons, even if really believing by those reasons does not here offer the subject a fairly high chance of believing the truth concerning p;

5. [Ought implies can] If a subject S psychologically could not have brought it about that X, S has not violated her intellectual duty by failing to bring it about that X;

6. ERGO: There are possible situations and subjects, such that here a subject is not epistemically justified in believing a proposition p, even if she has not violated her intellectual duty by failing not to believe p (from 1-5);

7. ERGO: Any theory of epistemic justification implying that necessarily, if a subject has not violated her intellectual duty by believing p, she is thereby justified in believing p, is false (from 6).

Before proceeding to Lockie’s counter-arguments, let us briefly take stock of the dialectical situation:

The argument above is logically valid. It derives its conclusion (7) on the basis of assumptions (1)-(5). (1) seems impregnable in this context, even if it denied by certain theories of EJ, such as “mad dog” process reliabilism[8]. Still, it seems hard to find a theory of EJ hospitable to the negation of (7), which also clearly undermines (1). After all, it sounds highly odd to say that a particular belief does not breach any intellectual duties, is epistemically justified, but is nonetheless in no legitimate sense epistemically reasonable.[9] It arguably makes much more sense for the champion of DCEJ to fight over the proper understanding of epistemic reasonableness in the context of the poverty argument, thus targeting (2) and (3). We shall return to this matter later below.

(4) also appears to be fairly uncontroversial. This premise is supported by strong empirical evidence, e.g. concerning the effects of what Alston termed “cultural isolation”.[10] Alston here offers the stock example of a native, who, along with his peers, unhesitatingly adopts the cosmological beliefs central to the traditions of their tribe. Never has the native subject had any chance to encounter counter-evidence, and the relevant beliefs are passed on by the most respected local authorities with overwhelming force. Cartesian skeptical ideals notwithstanding, it seems highly plausible that in such situations a subject has no psychological possibility but to believe the traditional tribal wisdom, even if his reasons here do not offer him anywhere near a high chance of believing the truth about cosmological matters.

It is no conclusive objection here that many cosmological myths probably are not interpreted too literally by the communities to which they are central, or that typical affirmative attitudes towards them perhaps do not clearly fit into the extension of the English term “belief”.[11] Surely it is hard to deny that at least some members of traditional isolated communities have clearly and irresistibly believed select cosmological propositions on the basis of non-truth-conducive reasons; say the proposition that starlight derives from small holes in a semi-spherical canopy spanning the surface of their world.

The ought-implies can principle embedded in (5) also does not seem like the most obvious place to begin attacking the poverty objection. Surely, in an epistemological context an unrestricted principle of this sort has come under fire from various quarters[12]. Yet, it should be clear that really the argument could go through even on a much weaker version of the fifth premise: It suffices, if in at least some of the possible doxastic situations verifying (4), the subject is off the hook because she could not have believed otherwise. Denying this seems a very steep price to pay for the defender of DCEJ. After all, it seems highly plausible that at least sometimes the fact that a subject could not in any way have brought about a different doxastic outcome suffices to get her off the hook. Not least if we interpret the last status as her simply not being an appropriate target of blame[13] on account of the actual outcome.

Given the above examination of the poverty objection, it would seem that the defender of DCEJ is well advised to direct her fire against (2) and (3). But (3) seems highly natural, once (2) is in place. So the defender of DCEJ faces the principal task of arguing against (2). Yet, clearly (2) could be undermined by a direct attack on (3): If there is no legitimate distinction between epistemic reasons that are really good and epistemic reasons that (perhaps with certain qualifications) appear good to a subject, of course this undercuts any motivation for a view of epistemic reasonableness that irresistibly suggests such a distinction. But denying (3), on the face of it, amounts to a very radical epistemic subjectivism. The difficult project, now, is formulating this subjectivist position and supporting it adequately, without simply begging the question against the poverty objection. Of course, if (7) if false, DCEJ is true. This may perhaps lend support to some form of epistemic subjectivism (we shall return to this matter further below). Yet, supporting (3) by virtue of denying (7), would simply beg the question against the poverty objection.

Above, I have made an effort carefully to formulate the poverty objection to DCEJ in a form faithful to Lockie’s principal sources. Ensuingly, I have quickly diagnosed the strengths of that objection. This footwork I deem required for a fair assessment of Lockie’s arguments, since no similar regimentation occurs in his article and it would otherwise seem very hard to estimate the relevance and force of his specific attacks. At last, we may now devote our full attention to Lockie’s text.

Lockie Against Epistemic Poverty

Much of Lockie’s text is devoted to attacking ascriptions of massive epistemic poverty, not least such as are arguably informed by cultural bias: Contrary to the opinions of certain bigots, according to Lockie

… we should see rationality (and irrationality) as a species-specific human capability, potentially attainable by all—or nearly all. I hold that rationality is not situated in one arbitrarily specific way-station located between a humanly unbounded God’s-eye reification and the perspective of our ancestors living in caves: a specific way-station to be found currently only in a micro-thin stratum of the statistically/logically educated élite.[14]

Lockie offers at least two basic lines of argument to support the above claim (I hope not to have missed further important arguments in his text). His positive argument, very briefly unfolded in the text, is empirical and appeals to the fact that even cultures steeped in irrationality to the bigot Western eye have in fact prospered and survived; together with the implicit premise that epistemic rationality is a precondition of such survival and prospering:

Highly contextualized humans … faced and face enormously complex, open-ended challenges across a great range of startlingly different situations: challenges involving frequent life-and-death situations in the harshest of environments. The idea that after surviving such challenges to adulthood, they were and are not, as a population, rational; or at least not as rational as us in contemplating our supermarket and smart-phone decisions, is on the face of it, not an attractive position to have to defend.[15]

I deem Lockie’s empirical-historical line of attack here to be somewhat at odds with his over-all purposes: Surely some cultures have impressively survived under very harsh conditions. But if we take this as solid evidence of their being “as a population” at least as rational as us modern scientifically literate Westerners, what should we say about those cultures, which like The Roman Empire, arguably have managed to fall into decline and perish even under conditions of relative plenty? For the present purposes, however, let us ignore this point and simply grant Lockie that even people entirely unable to appreciate modern statistics and logic (such as emblematic “illiterate agrarian peasants”[16]) do in fact base very many of their important decisions on epistemically rational assumptions.

We may also grant him that this is important to their collective survival. It remains hard to see how this offers support to the deontological conception of epistemic justification (DCEJ). Rather the relevant line of thought seems congenial to the appearance-reality distinction, which the poverty objection needs for its premise (3): Surely it cannot have much survival value for people living under harsh conditions merely to base their decisions on reasons, which to them seem truth-conducive, but really are not: A community, which in frequent cases of drought invariably sacrifice their best live stock to the Earth Goddess, because this seems to them a reliable way of securing relief from famine, is threatened on its existence, unless this procedure is somehow really a reliable way of promoting its survival.

So far the poverty objection is only further bolstered. But does Lockie’s above observation undermine other of its key assumptions, then? I think not. Of course progress is made with bereaving premise (4) of some of the evidence apparently supporting it: At least this premise is not immediately verified by observations of the ways of actual “illiterate peasants” as some bigots would perhaps think. Yet, as was already brought out above, still the premise seems overwhelmingly plausible. As Lockie concedes earlier in his article, “people are, after all, quite often irrational”.[17] And this clearly holds, even if rationality is more or less evenly distributed among literate and illiterate communities, as Lockie (perhaps too quickly) seems to conclude. As long as we accept the appearance-reality distinction for epistemic rationality, it seems safe to maintain that—even in the best of communities—some people do not believe as rationally as they think they do. And for such cases the poverty objection remains in full force.

Lockie’s positive attack on epistemic poverty, then, offered little ammunition against the poverty objection. However, he also launches a negative defense, claiming that some methods for evaluating the epistemic rationality of people in non-Western societies are in fact deeply unfair. Lockie here reviews Alexander Luria and Lev Vygotsky’s famous 1930’s investigations into the logical acumen of illiterate Uzbek peasants. Here test subjects were orally presented with the following puzzle:

In the far north, where it snows, the bears are white;

Nova Zemblya is in the far north, and it is always snowy there;

What color are the bears in Nova Zemblya?[18]

Many test subjects failed to deliver the response “white”. Rather they refused to answer the final question, quoting insufficient information or their lack of authority on the matter.

I entirely share Lockie’s central verdict that such test results are very poor evidence of widespread Uzbek peasant irrationality. I also think that the case is relatively clear: A crucial reasonable (or at least not clearly unreasonable) belief attributable to the typical Uzbek peasant here, is the belief that still they were not in a position appropriately to pronounce on the color of the Nova Zemblya bears to a distinguished audience like the official Soviet academics (or something like it).[19] A belief in this vicinity would seem to account very well for their verbal behavior. Unfamiliar with academic logic puzzles and their associated language games, they could hardly be faulted for their failure to appreciate that the test question was really supposed to be heard as a disguised conditional of the form: If the above dual assumptions are true, what are you then in a position to conclude about the Nova Zemblya bears?

But also I see no consequences of this for our conception of epistemic rationality. Marking an illiterate Uzbek as irrational on the basis of his refusal to perform in the Soviet logic test seems very much like faulting a gorilla for physical weakness, because it shows unwillingness to enter an Olympic weightlifting competition. Even after the test, the possibility remained live that the Uzbeks were rational in their beliefs in exactly the same sense as typical Western intellectuals often credit themselves with being. And general empirical reasons would suggest that probably the typical Uzbek peasant would prove fairly competent with basic logical deduction in settings closer to home, such as planning his harvest or negotiating trading prices for his crops at the market. Just as even after a gorilla’s failure to enter a weightlifting competition, observations of its physiology and living conditions suggest that it is very physically strong in exactly the same sense as are elite human weightlifters, including the ability to lift heavy objects over the top of its head.

Lockie’s empirical arguments against widespread epistemic poverty, even if successful, arguably should not trouble proponents of the poverty objection much. But perhaps it is too early to give up on Lockie’s general strategy here. It still could well be that more principled reasons against widespread epistemic poverty offer more comfort to DCEJ. One strategy could be an argument based on fundamental hermeneutical principles. Such an argument could gain momentum from Donald Davidson’s insistence that “we have the idea of belief only from the role of belief in the interpretation of [natural] language”[20] together with his famous holistic theory of the conditions for such an interpretation. The argument could go like this:

8. In order to evaluate the justification for subject S’s belief, we need to attribute to her (fairly) definite belief contents;

9. [Davidsonian Principle] We cannot attribute to S (fairly) definite belief contents, unless we are in possession of an interpretation of her language;

10. [Principle of Charity] We cannot acquire an interpretation of S’s language, unless we attribute to most of her utterances truth and sincerity; [21]

11. We cannot attribute truth and sincerity to most of S’s utterances, unless we assume that most of her beliefs are true;

12. We cannot assume that most of S’s beliefs are true, unless we assume that most of S’s beliefs are epistemically reasonable;

13. ERGO: Before we may even begin to evaluate the epistemic justification of S’s beliefs, we must assume that most of her beliefs are epistemically reasonable. Any assumption involving S’s whole-sale epistemic poverty is simply incoherent (from 8-12).

Several controversial twists and turns were taken in this argument, which I have no space to discuss here. However, by endorsing the argument, the opponent of the poverty objection is offered some interesting moves. As was the case with Lockie’s outspoken empirical argument, the most natural reading of “epistemically reasonable” here is not a subjectivist one. Rather the assumption of massive truth in S’s belief-base seems to enforce an assumption of objective reliability in her belief-forming ways. Still the argument offers the deontologist a promising way to at least cull the appearance-reality distinction (3), and hence limit the impact of the poverty argument, without going fully subjectivist.

The above Davidsonian argument in some sense targets (3), because it is supposed to generalize modally. Thus it even covers all possible reflexive epistemic evaluations. This fits well with Davidson’s insistence that all interpretation (linguistic as well as psychological) is at bottom radical linguistic interpretation.[22] Thus, for any epistemic subject the possible gap between what seems to her epistemically reasonable, and what must be seen as in fact epistemically reasonable, significantly narrows, even if the concept of epistemic reasonableness involved is not for that reason a subjectivist one: If the basic line of thought behind (4) above is right, very often we cannot help believing what presently seems to us epistemically reasonable. But the Davidsonian argument now assures us that it would be incoherent for us to presume that most of those “forced” (hence blameless?) beliefs are not also in fact epistemically reasonable. The possibility that such a belief is merely apparently epistemically reasonable is not excluded, but should now be regarded as an exotic one, reserved for what is—in Davidson’s words—“new, surprising, or disputed”.[23] Naturally the same line of reasoning applies to all possible interactions between diverse cultures and language communities.

Where does this leave us? At the very best, the poverty objection has been somewhat mollified. Due to the weakening of (3) necessary to accommodate the points above, it can now only saddle DCEJ with problems in the region of the “new, surprising, or disputed”, i.e. outside the province of beliefs constituting our “vast amount of agreement on plain matters”[24]. Only here is it possible that a believer irresistibly and blamelessly believes on apparently good epistemic reasons, which coherently may simultaneously be regarded as really being bad reasons. So only in such cases will an advocate of DCEJ have to concede that some unreasonable beliefs are nevertheless justified, because they are blameless. Perhaps this a softer bullet to bite.

But surely proponents of the poverty objection like BonJour will not be satisfied: To say that DCEJ has epistemic justification (EJ) right, as long as we stick to a certain province of beliefs, invites the problem that the extension of that province here seems fenced off by factors entirely contingent and alien to the dispute over our conception of EJ. Whether or not the content of a certain belief falls within a domain “new, surprising, or disputed” or is part of “our vast agreement on plain matters” is determined by historical (psychological and sociological) contingencies, whose relation to epistemic goals like truth and knowledge is at best a loose and contentious one. So Davidson seems quite a dangerous ally for the deontologist. Also, the basic conception of epistemic rationality underlying the argument’s premise (12) is an objectivist one.

So the possibility remains very live that a significant portion of a particular subject’s beliefs are, due to no fault of her own, simply epistemically irrational (because unreliably formed). Again, it seems strange simply to insist that such a very unreasonable belief is epistemically justified, just because the believer could not help being unreasonable here. So by basically endorsing (2), while only restricting (3) to a minority class of beliefs rather than going fully subjectivist, arguably a would-be “Davidsonian deontologist” has not really mounted a convincing defense against the poverty objection.

Of course there might be other ways to argue from assumptions concerning the limited extent of epistemic poverty to the unsoundness of a BonJour-style poverty argument apart from those reviewed above. But at present I fail to see any strategy in this terrain, which would serve Lockie’s purposes well. This concludes my examination of this part of Lockie’s defense.

Lockie on Epistemic Subjectivism and Internalism.

In the opening pages of his article, Lockie focuses more narrowly on what I deemed the most promising strategy earlier above, namely a direct attack on (2). Lockie submits:

The deontic understanding of justification, plus the ought-implies-can principle, leads to the threat of being never able to convict an agent of being unjustified, even when he is radically awry, should he or she merely be unable to apprehend the fact that he or she is mistaken. The solution to this problem for the internalist is to mark a distinction between “absolute” and “practical” justification—between fulfillment of “objective” and “subjective” duties. To escape the epistemic poverty objection, deontic, oughts-based justification must be restricted in its application to the “subjective”, “practical” realm.[25]

Soon after, Lockie writes of epistemic rules or standards that are “subjectively right”, and takes the “internalist” position to be that only compliance with such rules is relevant to “genuine” EJ.[26] But what does Lockie mean by “the subjective realm”? Cutting exegetical concerns short, we have already seen above what Lockie needs it to mean: A realm of reasons, which does not spawn an appearance-reality distinction like (3). For as soon as this distinction is in place, the deontological conception of epistemic justification (DCEJ) is in trouble: A reliabilist conception of epistemic rationality as was embedded in (2) does the trick, but so, as we shall see below, would in fact any other conception, which leaves the advocate of the poverty objection room to argue that a believer could irresistibly believe on reasons, whose epistemic goodness is merely apparent.

Here it is somewhat unfortunate that Lockie runs together epistemic internalism, subjectivism and deontologism. First, by his insistence, evident from the quotation above, that the required subjectivism amounts to “internalism” in a standard sense. Second, by his claim that such “internalism” has a “deontic core”, for which reason he also uses the term “deontogically internalist theories”.[27] Neither of those claims, I shall now hope to bring out, is felicitous for Lockie’s purposes.

As for the second claim, the problem is this: If (3) could only be undermined by a theory of epistemic rationality with “a deontic core”, this would beg the question against the poverty objection, whose aim precisely is to dissuade us from accepting any such theory. If internalism concerning EJ is to play a part in a convincing attack on the poverty objection, surely it must be motivated by reasons independent of DCEJ. And perhaps it can be motivated thus. Prominent standard defenses of internalism concerning EJ like Conee and Feldman (2008) do not directly appeal to deontic concerns. Rather they appeal to intuitive equivalences between allegedly representative epistemic subjects. The merits of such defenses have been the subject of a large and continually blooming literature, which I cannot hope to evaluate or review here. Rather, I shall simply grant Lockie that there is indeed a non-circular way of underscoring an internalist position concerning EJ in the context of facing the poverty objection.

This leaves us with the first problem touched upon above; the relation between internalism and the radical stripe of subjectivism Lockie needs in order to combat (3). Unfortunately for Lockie it is far from clear that standard internalist positions are in fact anywhere near as subjectivist as he needs them to be. Since he seems to favor so-called access-internalism[28], we shall consider that version first. An authoritative statement is provided by Robert Audi:

Some of our examples suggest that what justifies a belief—the ground of its justification—is something internal to the subject. The internal, in the relevant sense, is what we might call the (internally) accessible: that to which one has access by introspection or reflection … (…) … To have (internal) access to something is either to have it in consciousness or to be able, through self-consciousness or at least by reflection, whether introspective or directed “outward” toward an abstract subject matter, to become aware of it, in the (phenomenal) sense that it is in one’s consciousness.[29]

Shortly put, Audi says here that in order to count towards the epistemic justification of a belief, a reason must be internal to the believer’s consciousness, at least given a suitable process of reflection. This position is subjectivist in the sense that the contents of a believer’s consciousness are subjective phenomena if anything is. But the position is not subjectivist in the radical sense Lockie needs: Of course “seemings of epistemic reasonableness” are, in the relevant sense, in the believer’s consciousness. So, if it seems to a subject that one of her beliefs is epistemically reasonable, an access-internalist can allow this “state of seeming”[30] positively to count towards the epistemic justification of her relevant belief. Still, the access-internalist is in no way committed to let such seemings epistemically trump all other internally accessible evidence. So even within an access-internalist framework the possibility remains very live that a belief is simply unjustified, even if it seems highly reasonable to the believer. And there is no immediate reason why the access-internalist should block this possibility: Surely a believer recklessly following her immediate impressions of her own reasonableness without consulting the totality of her accessible reasons through careful reflection should hardly count as the paradigm of a justified believer.

The upshot is that the poverty objection could easily be mounted on assumptions acceptable to a standard version of access-internalism. Since this is a central point, I believe it worthwhile to spell this internalist poverty objection out in detail:

14. In order for a subject S to be epistemically justified in believing a proposition p in some situation C, S’s belief that p must be supported by good epistemic reasons for S to believe p in C;

15. [Access-internalist understanding of epistemic reason-goodness] In order for a reason to be a good and sufficient epistemic reason for S to believe p in C, in C S’s belief that p must be supported by the totality of S’s internally accessible evidence;

16. [Appearance-reality distinction] It is possible for an epistemic reason to appear good and sufficient to a subject in a situation, even if here it is not really a good epistemic reason;

17. There are possible situations and subjects, such that here a subject psychologically could not have brought herself not to believe what is indicated by reasons that appear to her to be good epistemic reasons, even if really believing by those reasons does not here offer the subject belief supported by the totality of her internally accessible evidence;

18. [Ought implies can] If a subject S psychologically could not have brought it about that X, S has not violated her intellectual duty by failing to bring it about that X;

19. ERGO: There are possible situations and subjects, such that here a subject is not epistemically justified in believing a proposition p, even if she has not violated her intellectual duty by failing not to believe p (from 1-5);

20. ERGO: Any theory of epistemic justification implying that necessarily, if a subject has not violated her intellectual duty by believing p, she is thereby justified in believing p, is false (from 6).

The only revision of the original poverty objection, apart from the replacement of (2) with (15) enforced by access-internalism, is the replacement of (5) with (17). But (17) seems like something a typical access-internalist should accept due to the possibility of epistemic poverty in the form of irreparable cognitive deficiency[31]: Plausibly some believers are simply, due to no fault of their own, simply too dull not often to believe by their immediate impressions of their own epistemic reasonableness, ignoring contrary reasons easily accessible to them through reflection. In that case, by plausible access-internalist standards, they come out epistemically unreasonable and unjustified, even if they are blameless.

We may also notice that not even a traditional deontological defense of access-internalism would be of much help to Lockie (disregarding the problem that arguably such a defense would beg the question against the poverty objection). This defense goes roughly as follows:[32]

21. [DCEJ] EJ belief equals blameless belief;

22. [Ought implies can] But a subject cannot appropriate be blamed for failing to consider evidence, which she cannot access;

23. Only evidence accessible to a subject can count towards her justificatory status (from 21 and 22);

24. But no evidence is accessible to a subject unless it is internally accessible, i.e. internal to the subject’s consciousness (perhaps conditional on suitable reflection);

25. ERGO: [Standard access-internalism] Only evidence internally accessible to a subject can count towards her justificatory status (from 23 and 24).

We may observe here that the deontological “core” assumption (21) only serves to limit the domain of what counts towards a subject’s EJ. In no way does it privilege specific types of internally accessible evidence to the point of collapsing the appearance-reality distinction concerning epistemic rationality.

Lockie does not explicitly consider a shift of allegiance to that other main camp within contemporary epistemic internalism; mentalism. However—as should be clear by now—such a shift would offer him no improvement. An authoritative short statement of mentalist internalism reads:

The justificatory status of a person’s doxastic attitudes strongly supervenes on the person’s occurrent and dispositional mental states, events, and conditions.[33]

Again this position is in no way forced to privilege, among the relevant mental states constituting the supervenience base of a person’s justificatory status, her seemings concerning epistemic reasonableness. And neither does awarding such privileges seem like a very good idea: Plausibly some subject could have among her mental states a strong impression of being epistemically justified in believing p, even if the totality of her mental states do not constitute an adequate basis for that belief.

I conclude that no standard form of internalism offers Lockie the kind of subjectivism he needs in order to undermine the poverty objection. However, his preference for access-internalism is hardly a coincidence. A modified version of that position could do the trick for him: Above I appealed to a possible subject believing on the mere immediate appearance of rationality, too impoverished to take into doxastic account the totality of her internally accessible reasons.

In order to make access-internalism sufficiently subjectivist for Lockie’s purposes, we need to insist that such agents are impossible. This could be done by significantly strengthening the access-condition far beyond what is usually seen in the internalist literature: We could insist that even if a reason is accessible, in the standard sense that a subject could access it through reflection, it is not accessible in the sense relevant to EJ, unless at the relevant time the subject is also able doxastically to respond to it.

Given this modification, the reasons (accessible in the standard sense), to which a strongly impoverished agent is unable to respond, are at once no longer deemed accessible to her in the sense relevant to EJ. This revised access-internalism would give Lockie what he needs: In cases where a subject believes on what irresistibly appear to her as being good epistemic reasons, all counter-acting reasons are now (in the revised sense) inaccessible to her, hence cannot count against her epistemic reasonableness. Lockie now needs the further assumption that in the entire absence of contrary reasons, a strong appearance of epistemic reasonableness constitutes a sufficient ground for belief. Then, in cases of irresistible belief on apparently good reasons, such beliefs are now also really reasonable by definition, since due to their inaccessibility no relevant contrary reasons counting against the belief’s EJ remain. Finally, then, the appearance-reality distinction is collapsed and (3)/(16) is defeated.

However, this radically subjectivist modification of access-internalism also makes the position highly suspect. It now more or less offers a carte blanche for spurious excessive epistemic self-confidence. It even serves to underwrite an important intuition fueling the poverty objection in the first place, namely that it ought not be possible to improve a subject’s epistemic status (more epistemic justification) simply by making her epistemically poorer (more block-headed and over-confident, thus putting further reasons out of her reach).[34]

I conclude that even if Lockie is right that a version of access-internalism is the right place to look for support against the poverty objection, the specific access-internalism he needs is implausible and radically non-standard. I see no strong reason to endorse such a position, even for access-internalists. In fact the dialectic could well work the other way: If access-internalism does not hold resources credibly to stave off the poverty objection to DCEJ, perhaps it ought better rid itself of any remaining “deontic core” and part with DCEJ altogether.


Above, I spent some effort regimenting a standard poverty objection to the deontological conception of epistemic justification (DCEJ), before evaluating Robert Lockie’s recent attacks on that objection. My investigations revealed that neither Lockie’s empirical arguments against widespread and asymmetrically distributed epistemic poverty, nor his harnessing of access-internalism, did ultimately succeed in significantly weakening the poverty objection. I also attempted to bolster and augment Lockie’s preferred arguments.

First, I tried to mount a hermeneutical argument against massive epistemic poverty. I argued that even if this argument seemed more relevant than Lockie’s empirical arguments, ultimately it achieved nothing beyond somewhat mollifying the implications of the poverty objection. Second, I argued that a radically non-standard access-internalism might in fact yield the radical subjectivism Lockie needs. However, not only is it hard to see how such a position could ever be defended without simply begging the question against the poverty objection. Also, it seemed implausible in its own right.

Where does this leave us? First, we should hardly rule out entirely that the poverty objection to DCEJ may somehow be undermined. But even with this objection down, DCEJ is not off the hook. As I have argued elsewhere[35] it faces a number of other pressing objections. One tempting way out at this stage is to give up on the claim central to traditional DCEJ that blamelessness suffices for EJ. After all, this is the claim targeted by the poverty objection.

If we retreat to the idea that blamelessness is an insufficient, but necessary requirement for EJ, that concept would retain a distinctive deontic flavor, while dodging the poverty objection altogether. Perhaps surprisingly, Lockie hints at such a solution, when he writes that “we are rational … (…) … if we use what we have got well, diligently, and to the limits of our ability.”[36] Now, the poverty objection is precisely that using “what we have got” to the limits of our ability does not always amount to using it well enough for justification. But literally, Lockie could be read as admitting here that those are two different aspects of epistemic rationality anyway.

Still, even this concession perhaps is not enough: In my opinion (see my 2013), in order clearly to stand apart from rival conceptions, DCEJ is forced to understand EJ in terms of blamelessness. But unlike EJ, blamelessness is not obviously a graded phenomenon. One can be more or less justified in believing a proposition, but either one is justly reproachable for believing it, or one isn’t. Also blamelessness is not obviously an epistemic notion, neither is it clear how it could be turned into one.[37]

Arguably, talk of “epistemic blamelessness” makes sense only in so far as epistemic concerns may, pro tanto, count towards an agent’s blameworthiness. It makes dubious sense to say of a blameless agent that she nevertheless remains blameworthy in some specific epistemic sense.[38] Rather, if, all things considered, she is not blameworthy for holding a belief, this means that she is beyond just reproach for holding it in any legitimate sense. If this is right, the relation between doxastic ethics and epistemology is at best a contingent one. We should then stop trying to make sense of basic epistemological concepts in terms of ethical ones.


Alston, William P. “The Deontological Conception of Epistemic Justification.” In Epistemic Justification: Essays in the Theory of Knowledge, 115-152. Ithaca: Cornell University Press, 1989

Audi, Robert. Epistemology: A Contemporary Introduction to the Theory of Knowledge, 3rd ed. London: Routledge, 2011.

BonJour, Laurence. “Internalism and Externalism.” In The Oxford Handbook of Epistemology, edited by Paul K. Moser, 234-263. Oxford: Oxford University Press, 2005.

Davidson, Donald. Inquiries into Truth and Interpretation, 2nd ed. Oxford: Clarendon Press, 2001.

Feldman, Richard, and Earl Conee. “Internalism Defended”. In Epistemology: An Anthology, 2nd Edition, edited by Ernest Sosa et al., 407-21. Oxford: Blackwell, 2008.

Firth, Roderick. “Chisholm and the Ethics of Belief.” The Philosophical Review 68, no. 4 (1959): 493–506.

Goldman, Alvin I. “What is Justified Belief?” In Justification and Knowledge, edited by G.S. Pappas, 1-23. Dordrecht: D. Reidel, 1979.

Goldman, Alvin I. “Internalism, Externalism, and The Architecture of Justification.” Journal of Philosophy 106, no. 6 (2009): 309-338

Lockie, Robert. “Perspectivism, Deontologism and Epistemic Poverty.” Social Epistemology, (2015): 1-17. doi: 10.1080/02691728.2014.990281

Needham, Rodney. Belief, Language, and Experience. Chicago: The University of Chicago Press, 1972.

Nottelmann, Nikolaj. Blameworthy Belief: A Study in Epistemic Deontologism. Dordrecht: Springer, 2007.

Nottelmann, Nikolaj. “The Deontological Conception of Epistemic Justification: A Reassessment.” Synthese 190 (2013): 2219-2241.

Owens, David. Reason Without Freedom: The Problem of Epistemic Normativity. London: Routledge, 2000.

Peels, Rik. “Let’s Bite the Bullet on Deontological Epistemic Justification: A Response to Robert Lockie.” Social Epistemology Review and Reply Collective 4, no. 12 (2015): 42-50.

[1] I am grateful to Jim Collier for the opportunity to write this response. Warm thanks extend to Rik Peels for his insightful comments on an earlier draft of this article.

[2] E.g. already Roderick Firth’s incisive objections against Chisholm’s version of DCEJ came at least very close to the core of of BonJour’s poverty objection, e.g when Firth insists that “it is, in short, an open question- not to be closed by definition—whether it is reasonable, in this ethical sense, to accept a proposition which is improbable”. Firth 1959, 498. Firth’s objection against DCEJ here in a sense is the inverse of the poverty objection: It is that immediately (“by definition”) DCEJ renders all epistemically reasonable (“probable”) beliefs blameless (“ethically reasonable”). But it is only a small step to see a problem of similar magnitude in the fact that immediately DCEJ renders all blameless beliefs epistemically reasonable without taking into account their objective epistemic quality. And implicitly, Firth seems to take this step: At least he talks of epistemic reasons in objectivist terms and explicitly regards it a problem that Chisholm cannot informatively say that some subject is blameless for holding a belief, because she has adequate evidence for it, since to him the purported explanans here is simply the explanandum stated once over (1959, 496).

[3] BonJour 2002, 236.

[4] BonJour 2002, 237.

[5] BonJour 2002, 236.

[6] BonJour 2002, 237.

[7] Remark that (2) need only spell out a necessary requirement, not a sufficient one. Hence it does not matter whether a sufficient condition for epistemic reason-goodness should involve the further requirement that the relevant belief is also psychologically based on such reasons.

[8] See here not least Goldman 1979, according to which production of a belief by a reliable process suffices for its EJ (given the satisfaction of some further counterfactual conditions). Of course even this position is at least consistent with the thesis that no process is reliable unless it involves support by good epistemic reasons.

[9] Anyway, Lockie seems to concede (1): He uses the terms “epistemically justified” and “epistemically rational/reasonable” more or less interchangeably. And traditional champions of DCEJ like Chisholm did not shy away from construing justification in terms of reasonableness. See e.g. Firth 1959.

[10] Alston 1989, 145.

[11] Famously, Needham 1972 makes much of such claims.

[12] See not least Owens 2000.

[13] The relations between notions such as ”duty”, ”blame”, and ”permission” in a doxastic context arguably are complex. I have no space for a detailed treatment of such matters here. A reader interested in my views may consult Nottelmann 2007. See also Nottelmann 2013.

[14] Lockie 2015,12.

[15] Lockie 2015,12.

[16] This is Lockie’s term (2015,12).

[17] Lockie 2015, 9.

[18] Quoted from Lockie 2015, 5.

[19] Here, on Lockie’s behalf, I try to meet Rik Peels’ complaint that Lockie does not sufficiently articulate the Uzbek beliefs, which are not proved irrational by the test. Peels 2015, 45. Peels offer an insightful analysis of Luria’s and Vygotsky’s specific findings. I shall not engage with those here, since arguably even an interpretation most congenial to Lockie’s purposes does not seem to offer him the desired ammunition against the poverty objection.

[20] Davidson 2001, 170.

[21] In a famous passage Davidson submits: ”Making sense of the utterances and behavior of others, even their most aberrant behavior, requires us to find a great deal of truth and reason in them. To see too much unreason on the part of others is simply to undermine our ability to understand what they are so unreasonable about. If the vast amount of agreement on plain matters that is assumed in communication escapes notice, it’s because the shared truths are too many and too dull to bear mentioning.” Davidson 2001, 153. I here take it that “agreement on plain matters assumed in communication” implies that sincerity as well as truth must be assumed.

[22] Davidson 2001, 125.

[23] Davidson 2001, 153.

[24] C.f. Davidson 2001, 153.

[25] Lockie 2015,3.

[26] Lockie 2015,3.

[27] Lockie 2015,4.

[28] Lockie in fact seems to equate ”deontologically internalist theories” with access-internalism. Lockie 2015,4.

[29] Audi 2011, 272-3.

[30] Not to be confused with a seeming state, of course!

[31] This is Alston’s preferred term for what many ordinary folk would call plain block-headed stupidity. See Alston 1989, 146.

[32] See e.g. Audi 2011, 276.

[33] Conee and Feldman 2008, 408.

[34] Similar concerns seem to trouble access-internalism even in its more palatable varieties. See e.g. Goldman 2009, 323-5.

[35] Nottelmann 2013.

[36] Lockie 2015, 9.

[37] This line of attack was already prominent in Firth 1959.

[38] Although, of course, we could say that counter-factually she would be blameworthy in a world, where only epistemic concerns mattered. But, I take it, this observation does not help out DCEJ, which conceives of EJ as some kind of (epistemic) blamelessness in the actual world. I thank Rik Peels for pushing me to clarify this concern.

Categories: Critical Replies

Tags: , , , , , ,

2 replies


  1. Deontological Perspectivism: A Reply to Lockie, Hamid Vahid « Social Epistemology Review and Reply Collective
  2. Response to Elqayam, Nottelmann, Peels and Vahid on My Paper “Perspectivism, Deontologism and Epistemic Poverty”, Robert Lockie « Social Epistemology Review and Reply Collective

Leave a Reply