What Entitlement Is

Abstract

The paper is an examination of Tyler Burge’s notion of epistemic entitlement. It begins with consideration of a recent attempt to understand entitlement, including the ways in which it differs from the more traditional notion of justification (Casullo in Acta Analytica, 22, 267–279, 2007). The paper argues that each of Casullo’s central contentions rests upon confusion. More generally, the paper shows that Casullo’s interpretation tries to force Burge’s work into a framework that is not suited for it; and that the interpretation also suffers from not being even minimally informed about the actual character of Burge’s epistemological views. The paper’s second half explains the distinction between entitlement and justification, including ways in which Burge’s understanding of the relevant notions has changed over time.

Tyler Burge (1993) introduced the term “epistemic entitlement” in the context of an extended discussion of the epistemology of interlocution. Entitlement was contrasted with justification, each of which is a subspecies of epistemic warrant. Roughly speaking, for Burge a justification consists in having a reason to hold a belief. An entitlement, by contrast, does not consist in possession of a reason. Yet each of justification and entitlement provides an epistemic right to hold a belief. Burge has argued that we are entitled to rely upon, not just interlocution, but also perception, inductive and deductive inference, reflection, memory, and certain claims to self-knowledge and self-understanding. The primary challenge, in each case, is to explain how possession of an epistemic right is possible in the absence of a warranting reason.

Burge’s discussions of the various entitlements we enjoy are often long and are invariably challenging. Partly for these reasons, Burge’s commentators and critics have very often misunderstood, not only his positions on these matters, but even the very distinction between justification and entitlement and, therewith, the notion of epistemic entitlement itself.

The present paper is intended to be a first step toward improving this situation. It begins with consideration of a recent attempt—Casullo (2007)—to come to terms with Burge’s notion of epistemic entitlement. I think that Casullo’s paper is worth discussing, not because it provides even an approximation to a correct understanding of epistemic entitlement, but because examination of the mistakes Casullo makes provides a natural and didactically useful platform for a proper discussion of the matter. This strategy is further supported by the fact that a number of writers have referred approvingly to Casullo’s discussion.¹ And it is supported, finally, by the related fact that his article is the only sustained published attempt, by a commentator, to understand Burge’s distinction.

The first half of the paper explains what is wrong with Casullo’s interpretation. The overarching goal of the paper, however, is not critical but constructive—it is to provide a better understanding of the nature of epistemic entitlement, and of the ways in which it differs from justification. The paper’s second half is accordingly devoted to contributing to such an understanding.

Casullo seeks to answer three questions concerning Burge’s notion of epistemic entitlement. How does entitlement differ from justification? How strongly externalist is entitlement? And is the notion of epistemic entitlement new, or has Burge merely coined a new term for a familiar epistemological concept? In the course of offering answers to these three questions, Casullo gives two principal arguments, together with certain inferences drawn from them. The first argument contends that, on Burge’s official explications of justification and entitlement, there is an inconsistency between his claim that our primary warrant to rely upon the word of others is an entitlement (rather than a justification) and his explanation of this entitlement. Sections 1 and 2 of the article discuss this argument. The second argument is to the effect that the key to understanding Burge’s distinction—and hence, the notion of epistemic entitlement itself—is recognizing that he employs a particularly strong conception of that in which possession of a reason consists. In sections 3 and 4, I elucidate this reasoning, showing that and why it is confused. Section 5 briefly considers Casullo’s claim that Burge’s notion of epistemic entitlement is not a conceptual or substantive—as opposed to a linguistic—innovation. The final two sections of the paper provide my own answers to the questions concerning epistemic entitlement. Section 6 explains changes that Burge’s understanding of the distinction between justification and entitlement has undergone over time. And section 7 answers the questions directly, from the perspective of Burge’s current thinking on the matters.

1 ᅟ

As noted, Burge maintains that there are two distinct forms of epistemic warrant, which is a super-species concept of epistemic good.² Epistemic warrants are either justifications or entitlements. To a first approximation, justifications involve reasons, which can be either self-sufficient premises or elements in an inference. Entitlements, by contrast, are epistemic rights to believe which do not rest upon reasons, or the ability to access such. Burge presents entitlement as an epistemologically externalist conception of epistemic warrant and justification as internalist.

Casullo confines his attention to what were then Burge’s two most extensive discussions of epistemic entitlement—those of the lately mentioned interlocution and of perception, respectively (Burge 1993, 2003a).³ The first question Casullo asks concerns the relationship between entitlement and the more familiar notion of justification. He quotes Burge as follows:

Although both have positive force in rationally supporting a propositional attitude or other cognitive practice, and in constituting an epistemic right to it, entitlements are epistemic rights or warrants that need not be understood by or even accessible to the subject (1993, 458).

Casullo distills from this passage a pair of theses concerning Burge’s notions of justification and entitlement: 

• (C1) Both have positive force in rationally supporting beliefs; and

• (C2) Both are epistemic rights or warrants.

Concerning differences between justification and entitlement, Casullo notes that Burge stresses two features peculiar to entitlements:

• (E1) Entitlements need not be understood by the subject; and

• (E2) Entitlements need not be accessible to the subject.

Subsequently, Casullo draws attention to an additional pair of features of entitlements, as well as to two new properties of justification:

• (E3) Being entitled to rely on a resource does not require being able to justify reliance on that resource; and

• (E4) Being entitled does not require being able to conceive such a justification.

• (J1) Justifications involve reasons people have and have access to; and

• (J2) Justifications involve reasons that must be available within the cognitive repertoire of the subject (268–9).⁴

Casullo maintains that taking these theses together allows us to isolate four prominent features of Burge’s understanding of our distinction. First, justifications do, and entitlements do not, involve reasons. Second, justification in its requirement of reasons requires as well understanding. Third, entitlement does not require access to reasons, while justification does so require. And fourth, being entitled does not require the ability, or even the conceptual wherewithal, to provide a second-order justification where warrant is attached to belief. I think these points are obviously correct.

Casullo begins with the access and second-order justification requirements, noting that they are closely related. We can distinguish having access to the ground of a belief from having access to the ground’s adequacy in conferring warrant upon the belief. Simplifying from Casullo’s presentation a bit, we get two possible senses in which one might hold that warrant requires epistemic access:

• (A1) S has access to the ground of the belief that p iff S’s belief that p is based on ground G, and S has access to G; and

• (A2) S has access to the adequacy of the ground of the belief that p iff S’s belief that p is based on ground G, and S has access to justification for the belief that G is an adequate ground for S’s belief that p.

Casullo proposes to examine the application of these principles to Burge’s account of specifically interlocutionary entitlement.⁵ Burge encapsulates his account of this entitlement in the following principle, which he calls the Acceptance Principle:

• (AP) A person is entitled to accept as true something that is presented as true and that is intelligible to him, unless there are stronger reasons not to do so (1993, 467).

Burge holds that this principle indicates the most fundamental source of our epistemic warrant for accepting the word of others. The explanation of the entitlement makes essential reference to a propositional content presented as true and as comprehended by the recipient of the testimony. Casullo attempts to understand Burge’s views here by applying to them Alston’s theses (A1) and (A2). From this perspective, having access to the ground of the warrant would appear to require having access to the presentation-as-true (or proposition asserted). Having access to the adequacy of the ground, in contrast, would require more—it would require, in effect, having some reason to believe that (AP) is itself true.

Recall our first thesis concerning justification:

• (J1) Justifications involve reasons people have and have access to.

This principle suggests a promising line of demarcation between entitlements and justifications. Justifications consist in having reasons, and entitlements do not. Therefore,

• (P1) Justified belief that p involves having an accessible reason to believe that p; entitled belief does not.

However, Casullo claims that there is a problem with (P1). The problem is that it appears to be in conflict with Burge’s view of interlocutionary entitlement. We saw that according to (AP), one is entitled to rely upon the word of others just in case a content p is intelligible and presented-as-true. Casullo writes: “Both understanding the content of a communication and understanding its assertive force, however, are paradigmatic examples of states that are propositional” (273). Since (AP) bases testimonial entitlement on propositional states, it bases them on reasons; indeed, reasons to which the believer has access. Casullo suggests that Burge therefore has a problem, insofar as (P1) is taken to be a correct and fundamental statement of the latter’s view.⁶

Here is the context for the passage just quoted, which contains Casullo’s reasoning for the first major conclusion of his article:

According to (AP), S is entitled to accept p as true provided that p is intelligible to S and p is presented as true to S … Both understanding the content of a communication and understanding its assertive force, however, are paradigmatic examples of states that are propositional. Since the entitlement to accept the testimony of another is based on propositional states, it is based on reasons … (273).

Casullo concludes:

Therefore, according to (J1), the warrant conferred by (AP) is justification rather than entitlement. But, according to (AP), the warrant that it confers is entitlement. So, (P1) appears to be incompatible with Burge’s account of testimonial entitlement even on the narrow construal of reasons (273).

As I say, the first major conclusion of Casullo’s paper is stated here. So long as we restrict attention to the notions of access and reasons, in attempting to understand what separates entitlement from justification, we seem to arrive at the result that Burge’s explicit statement about the differences between the two forms of warrant [(P1)] is incompatible with his own characterization of the nature of our specific entitlement to rely upon interlocution [(AP)]. I turn now to critical examination of Casullo’s argument for this claim.

2 ᅟ

Casullo has argued that absent some alternative way of understanding the distinction between justification and entitlement, Burge’s explanation of our entitlement for relying upon the word of others [(AP)] is inconsistent with his stated claim that entitled belief necessarily obtains in the absence of reasons [(P1)]. In the present section, I will argue that Casullo’s reasoning for this claim rests upon confusion about what it is to have a reason for belief. His argument is also vitiated by the attempt to force Burge’s distinctions into an epistemological framework that does not befit them. I begin with the second point.

(A1) and (A2) above concern the notion of the “ground” of a belief, as well as the notion of “basing” a belief upon that ground. These are familiar terms in the epistemological literature, but it is important to notice that Burge does not use them. More importantly, the way in which he views epistemic warrant—including especially entitlement—precludes simple talk of the “ground” of a warranted belief. Take for example perceptual entitlement. Being entitled to a perceptual belief involves the obtaining of at least the following conditions: (a) enjoying a perceptual experience the content of which is partly individuated by reference to causal-formative interactions between related content-instances and subject matter; (b) the reliability of the process which issues in the perceptual experience, in the thinker’s normal environment; and (c) the correct negotiation of the transduction of perceptual experience content into the propositional form appropriate to belief (see Burge 2003a, sections VI and VII). Burge’s arguments concerning perceptual entitlement are obviously complex. My point in bringing them up here is simply to point out that each of (a)–(c) is part of what warrants a perceptual belief to which we are entitled. And none of the elements is appropriately thought of as the “ground” of entitled perceptual belief.

Part of the difficulty with talk of grounds here is that it encourages a certain analogy between reasons and justification, on the one hand, and perceptions and perceptual entitlement, on the other. It is tempting to suppose that just as reasons are the ground operative in justification, perceptions are the ground at issue regarding perceptual entitlement. But the analogy is not a good one. The reason is the warrant, in the justification case. We have just seen, however, that it is incorrect to suppose that in the case of perceptual entitlement perception is the warrant. Perceptions are only part of what constitutes, or undergirds, perceptual entitlement. Moreover, perceptual states are not on Burge’s view necessarily conscious, and are not therefore necessarily accessible, as reasons were conceived by Burge at the time to be.

A further problem with this approach derives from the fact that all epistemic warrants for Burge require reliability. And so in order for a full warrant—whether a justification or an entitlement—to be accessible to a thinker, the reliability of the warrant must in some sense be accessible. Now in certain cases of justification, the reliability associated with the warrant is constitutively accessible to the believer. (See section 7) But no form of epistemic entitlement requires that the believer have access to the relevant sort of reliability. Since, again, reliability is always part of a warrant, and this part of the warrant is never required to be accessible in the case of epistemic entitlement, it is particularly inapposite to think of entitlement in terms of “grounds” and “basing.”

I believe that insistence upon thinking in terms of grounds and basing is largely a product of trying to assimilate all epistemic warrant to the model of inferential reasoning. Because Burge explicitly and repeatedly points out that not all warranted belief fits this model, it is a definite mistake to insist upon using the model in attempting to understand his views.⁷ Having said that, I will try to engage with Casullo’s argument, which does essentially employ the notion of ground. But I will point out places where this insistence causes problems.⁸

Let us set out Casullo’s argument explicitly for ease of reference. We begin with Burge’s Acceptance Principle, stated in the previous section:

• (AP) A person is entitled to accept as true something that is presented as true and that is intelligible to him, unless there are stronger reasons not to do so.

Casullo thinks it follows from Burge’s commitment to this principle that interlocutionary warrant is reason-based, contrary to the characterization of entitlement in (P1). This is supposed to follow because, according to (AP), interlocutionary entitlement is based upon propositional understanding—understanding both of the conceptual content of the presentation-as-true, and of its assertive force.

There are two large problems with this argument. The first is that (AP) is not intended to be a full explanation of our entitlement to rely upon others. When one looks at the full explanation, it becomes clear that there are warranting elements, in relevant cases of interlocution, other than the propositional state presented-as-true. Partly for this reason, it is not appropriate to view this state as the “ground” of interlocutionary entitlement. This is of course a specific version of a more general problem. Entitlements always involve more than one warranting element. One cannot know what the elements are, in the case of a given entitlement, without looking at the details of the explanation that Burge provides. Casullo ignores the details of Burge’s actual views. And this is a principal reason that his interpretation fails.

Indeed, Casullo effectively attributes to Burge the very view that the latter opposes. Burge opposes—explicitly, clearly, and repeatedly—the view that basic interlocutionary warrant is a matter of making an inductive inference from, inter alia, the propositional content presented-as-true. This view of interlocution came to be called “reductionism.” Burge is obviously an opponent of reductionism. Yet Casullo interprets Burge as holding a reductionist position, according to which interlocutionary warrant is a matter of making an inference from the content presented-as-true.

The second problem is related to these considerations. It simply does not follow from the fact that a propositional state is implicated in a warrant that the warrant is reason-based. Consider again the sentence from the passage from Casullo quoted at the end of section 1: “Since the entitlement to accept the testimony of another is based on propositional states, it is based on reasons …” (273). This is a clear non sequitur. A belief which is warranted by reasons is, paradigmatically, an inferentially obtained belief. And despite the role that propositional content must play in any interlocutionary warrant, the warrant is—in the fundamental cases with which Burge is concerned—not inferential. It is not based on reasons.

It is true that all reasons are propositional in form. But it certainly does not follow from this that all propositional states function as reasons. To suppose, with Casullo, that this does follow is to commit the elementary logical fallacy of affirming the consequent. It is to conflate necessary and sufficient conditions.

In the case of entitlements generally, that which provides the warrant for a belief is never a single “thing.” The point has already been illustrated for the case of perception. I will now show that it holds as well with respect to interlocution.

Burge’s substantive explanation for our entitlement to rely upon the word of others draws conceptual connections between three elements: propositional representational content, rational sources, and truth. A prima facie intelligible purported presentation-as-true is a prima facie indication of the presence of a rational source. Rational sources can be relied upon, prima facie, to present truth. And truth is the fundamental representational norm for propositional competencies, states and events, which so to speak sets the most fundamental epistemic norm, epistemic warrant. This, in barest outline, is why on Burge’s view we are (most fundamentally) epistemically warranted in accepting the word of others.

An explanation of an entitlement must make constitutive mention of the various elements which conspire to warrant non-reason-based belief. To understand an entitlement one must therefore look at the details of the explanation. In the interlocution case, we have just seen, there are three elements to the explanation—the ostensibly presented content, the presumed source, and truth. Apparent propositional content is a defeasible indication of the presence of a rational source as its author. Rational sources are necessarily, other things equal, reliable guides to truth. It is these two relations which provide the warrant for relying upon interlocution.⁹ Notice that we have made no mention of any inference from the propositional content in question. There is no inference involved in basic cases of interlocutionary warrant. Such warrant is not reason-based. It is an entitlement rather than a justification.

These points illustrate not only the inadvisability of applying the framework of grounds and basing to Burge’s epistemological views. They illustrate as well why it is that entitlement cannot be understood in abstraction from the details of the actual explanations that Burge provides. One can identify the warranting elements involved in a given entitlement only by attending to the actual, substantive explanation of the entitlement.

It was mentioned earlier that there is never a single state or content which underlies a case of epistemic entitlement. But we can say something even stronger than this. In some cases of entitlement that is no state or content whatsoever, which provides warranting force for the entitled belief. An unsophisticated thinker’s reliance upon deductively valid inferential patterns is one to which she can be epistemically entitled. She need have no understanding of the relevant logical notions as such. What constitutes this sort of entitlement is the thinker’s by-and-large correct execution of inferences in accordance with the relevant pattern. Again, there is in this case no state or content which could possibly be identified with the “ground” of the entitled belief or reliance. Such cases clearly exhibit the inappropriateness of the ground/basing model for understanding epistemic entitlement.¹⁰

Now obviously, one could simply define “ground” to mean warrant. Then it would not be amiss to ask after Burge’s view of the ground for any given warranted belief or set of beliefs. However, in so doing, one would lose much of the usefulness of the notion epistemic ground. One would no longer be able to require that there be a specific “thing” which functions as epistemic ground. One could not then insist that the thinker do something to base her belief upon its ground. And one would no longer be able with propriety to rely upon the already stretched analogy, between warrant and warranted belief, on the one hand, and the set of premises of an inference and the inference’s conclusion, on the other.

Casullo is correct that propositional representational content plays a necessary role in one’s warrant to rely upon the word of others. He is even correct that it plays a specifically epistemic role—as opposed to a merely causal-enabling one. But Casullo is certainly mistaken in thinking that it follows from this fact that such warrant is reason-based and inferential in character.

To summarize the results of this section: Casullo has not shown that there is any inconsistency, nor even any tension, between (AP) and (P1). It does not follow from the undisputed fact that a propositional state plays a role in the explanation of an entitlement that the relevant warranted belief is inferentially based upon that state. And it definitely does not follow from the fact that all reasons are propositional that all propositional states function as reasons.

3 ᅟ

Perhaps Casullo recognizes that a line of interpretation which yields the result that two of a thinker’s proximate theses are inconsistent has little to be said in its favor. So even though he clearly believes his first argument to be sound, given Burge’s own characterizations of the forms of warrant, he moves on to a new strategy. Casullo considers some of Burge’s more recent formulations of the distinction between justification and entitlement. He quotes the following passage from Burge:

Entitlement is epistemically externalist inasmuch as it is warrant that need not be fully conceptually accessible, even on reflection, to the warranted individual. The individual need not have the concepts necessary to think the propositional content that formulates the warrant … Justification is warrant by reason that is conceptually accessible on reflection to the warranted individual (2003a, 273).

The notion of epistemic externalism will occupy Casullo and us below. For now, we focus upon the notion of accessibility at work in the passage. Burge seems to be claiming that appeal to it is at least one central way of understanding his distinction. Therefore, we have the following:

• (E5) Entitlement is warrant that need not be conceptually accessible to the subject.

  And,

• (J3) Justification is warrant by reason that is conceptually accessible to the subject (273).

A suggestion concerning the appropriate understanding of these theses is provided by Burge’s discussion of the relationship between his epistemological framework and Wilfred Sellars’s rejection of the “myth of the given.”¹¹ Sellars upheld the Kantian view that reasons must be propositional. We have seen that Burge agrees with Kant on this point. But Sellars maintained, in addition, that only reasons or other propositional states can warrant belief. Burge’s doctrine of entitlement is precisely a denial of this thesis. Burge holds that this second Sellarsian thesis is true of justifications, but not warrants generally (not entitlements).

Casullo claims that Burge’s view has two key differences with Sellars’s, the first of which we have lately noted. “First, Burge denies that only reasons can warrant beliefs; entitlement is also a form of warrant. Second, he introduces a conception of reasons that is richer than the Sellarsian conception” (274). This stronger conception of reasons is central, Casullo maintains, to Burge’s contention that perceptual representations entitle rather than justify perceptual beliefs. We are entitled to the latter, rather than being justified in holding them, precisely because the transition from perceptual states to perceptual belief is not a piece of reasoning which requires access to the ultimate ground of the entitlement. Recall that justification for Burge essentially involves reasons. Casullo judges that justification, in contradistinction to entitlement, does require such access to the ground of the warrant.

Casullo appears to notice that, on Burge’s view, higher non-human animals and young human children possess some warranted perceptual beliefs and yet lack reasons for those beliefs. Casullo believes that Burge holds that such thinkers lack concepts which would be necessary even to think such reasons. They lack concepts such as reliable and defeating condition. Casullo concludes: “Therefore, the conjunction of Burge’s contention that justification requires having reasons with his contention that having reasons requires possessing at least some epistemic concepts suggests that having a reason to believe that p requires that the epistemic principle governing the adequacy of that reason be conceptually accessible to the cognizer” (275).

This is to say that Casullo believes that Burge is committed to a certain strong sense of “reason”—in the latter’s claim that justification requires reasons, whereas entitlement does not. Casullo formulates the thesis to which he believes Burge is committed, thus:

• (P1*) Justified belief that p requires conceptual access to the epistemic principle governing the ground of the belief that p; entitled belief that p does not.

Because he takes Burge to be committed to this strong notion of that in which possession of a reason consists, Casullo concludes that (AP) is not inconsistent with (P1) after all. For despite the fact that (AP) requires that testimonial entitlement epistemically involves propositional states that the cognizer has access to, they do not yet form reasons for the belief which she forms on the basis of comprehending the content and assertive force of the relevant utterance or inscription. They do not form reasons for her because the cognizer need not—and often will not—possess the concepts necessary even to think the explanatory ground of the warrant, viz. (AP) itself.

This is the second major conclusion of Casullo’s paper. He argues that it is Burge’s strong conception of reasons, embodied in (P1*), which provides the key to understanding the distinction between entitlement and justification. The next section comments critically upon this argument.

4 ᅟ

It was argued in section 2 that Casullo fails to show any sort of incompatibility between (AP) and (P1). That is to say, he does not demonstrate that there is an inconsistency or even a tension between the view that propositional states must be implicated in certain entitlements, on the one hand, and the view that entitlements of their nature are precisely warrants which do not involve reasons for belief, on the other. So there is no particular interpretive impetus for the search for an alternative understanding of the distinction between justification and entitlement, different from that which Burge himself provides. Nothing has been shown to be amiss with the characterization of the distinction between justification and entitlement in terms of reasons as ordinarily conceived.

Nevertheless, it will be instructive to examine the particular flaws in this, Casullo’s second major argument. Such examination is in any case necessary for understanding Casullo’s answers to the three questions concerning both the nature of entitlement and its significance.

Casullo states as a premise of his second main argument that Burge has a very strong conception of reasons. In reality, so far from having a strong conception of reasons—one which, indeed, requires the reasoner to have access to the epistemic principle governing her warrant—Burge has a relatively weak conception of that which is required for possession of a reason. And it is both certain and obvious that he does not hold (P1*).

As I have noted, Casullo confines his attention to only two of Burge’s articles, one of which is Burge (2003a). It is surprising therefore that he takes Burge to be committed to a very strong conception of reasons. For Burge several times in this paper states unequivocally that he does not have such a conception of reasons. At the very beginning of the paper he claims:

Inferentially supported beliefs are justified, or are at least backed by beliefs that are capable of justifying them. Higher non-linguistic animals have reasons or justifications for some of their beliefs in this sense. My notion of having a reason or a justification does not require reflection or understanding. That is a further matter (2003a, 505, fn. 1).

He later says: “Animals need not be able to understand or be guided by the norms that in fact apply to their formation of beliefs” (2003a, 513). This claim applies both to entitlements and justifications. It therefore applies as well to reasons. And at the end of the paper, Burge makes the point for a third time:

Functioning as reasons does not require deliberative self-conscious reflection on reasons, or even understanding the notion of reason (2003a, 547, fn. 37).

In these passages, Burge clearly states that he does not hold a strong conception of reason, such as that embodied in (P1*). He also attributes reasons to creatures that cannot possibly have reasons for their beliefs in Casullo’s strong sense. Burge thinks it is plausible that what he calls “higher animals” have reasons for some of their beliefs (examples of higher animals are primates). From this claim together with (P1*), it follows that some at least of these animals have conceptual access to the epistemic principle governing the ground of their warranted beliefs. But this is evidently absurd. Higher animals lack the conceptual wherewithal to think even about their perceptions, much less epistemic matters. Since they can have reasons for their perceptual beliefs, yet cannot do that which is required by (P1*), the principle cannot be a correct statement concerning the nature of reasons.

There is no reason to think that Burge endorses anything like (P1*). Indeed he explicitly and repeatedly rejects strong conceptions of reasons simpliciter, in the very paper that Casullo discusses and quotes from. So Casullo’s second major argument fails along with the first.

Casullo may wish to lodge an objection at this point. He quotes (275) in support of his interpretation the following passage from Burge:

Understanding the warranting force of being in a relevant-seeming perceptual state and having some grip on the general shape of conditions that undermine such force are part of having a reason (2003a, 529).

So Casullo appears to want to argue as follows. Burge claims in this passage that possession of a reason requires having a grip on the “general shape” of conditions that undermine warrant. But then he is committed to (P1*) after all, since having a grip on this shape precisely requires having conceptual access to the epistemic ground supporting any given justified belief. And this is all that (P1*) claims.

However, in this passage Burge is obviously not describing reasons as such. He is describing reasons for perceptual belief. Because perceptual representational states are non-conceptual and therefore non-propositional, having a reason for a perceptual belief must go beyond the relatively simple inferential model which we have seen underlies one central aspect of Burge’s conception of justification. Perceptual experience itself cannot constitute or provide reason for belief. Only a propositional state could do that. This is why having some meta-representational concepts such as perceptual state is required for the possession of reasons for perceptual belief. For perceptual experience to be called upon to participate in an inferential warrant would require a kind of meta-justification. Such a justification is not required in other cases. In other cases, ordinary first-level inferences provide justification. A simple example is inferring p from p&q. Such an inference can provide justification for one’s belief that p. No meta-representational concepts are necessary for executing such an inference.

Perception is the only competence, or resource for reason, which constitutively involves non-conceptual representational contents.¹² Only such contents can figure in perceptual states strictly so-called. A reason for a perceptual belief must indicate why it is a good thing to hold, and it must constitutively involve the content of the perceptual belief. The content of the belief, insofar as the belief is warranted by perception, must be involved in the explanation. This content derives from the correlative non-conceptual perceptual state content. But this last content cannot directly participate in a reason, since reasons are propositional. A perceptual state itself cannot be, or figure directly in, a reason for belief. And the perceptual belief is not a reason for itself. It follows that insofar as the perceptual state is to figure in an explanation of why the perceptual belief which derives from it is a good thing to hold, it must figure indirectly, in the form of a proposition about the perception. This is why reasons for perceptual beliefs require meta-representational abilities. And it explains why reasons for perceptual belief are perforce different from reasons for other sorts of belief, which do not require meta-abilities.

Now it is perhaps tempting to think that Burge does hold an analog of (P1*), one which is restricted to reasons for perceptual belief. But this too is a mistake. Not even reasons for specifically perceptual belief require epistemic concepts or access to epistemic principles. Let us examine the evidence for these matters contained in Burge (2003a). Three quotations are particularly relevant. First, in discussing the applicability of epistemic norms to creatures who lack the ability to conceptualize the norms, Burge writes:

In critical reasoners, these norms become objects of reflection. But to be applicable to mental states of non-critical reasoners, many such norms need not be conceptually accessible to reflection. In young children and higher subhuman animals, they are not conceptualized. Such children and animals cannot think about them. They lack the concepts epistemic, warrant, entitlement, reason, reliable, competence, entails, perception, and perceptual state (2003a, 521).

This passage is not discussing necessary conditions upon either reasons or reasons for perceptual belief. It is discussing necessary conditions for thinking about epistemic norms. And having a reason—even a reason for a perceptual belief—does not require the ability to think about such norms.

Second, Burge writes at one point:

Children and higher non-human animals do not have reasons for their perceptual beliefs. They lack concepts like reliable, normal condition, perceptual state, individuation, defeating condition, that are necessary for having such reasons (2003a, 528).

This passage comes slightly closer to supporting (P1*), as a claim restricted to reasons for perceptual beliefs. For the passage concerns not access to epistemic norms but necessary conditions upon possession of reasons for perceptual beliefs. However, the passage does not claim that all the listed concepts are necessary for having such a reason. Burge simply presents items on the list as examples of concepts necessary for having reasons for perceptual beliefs which are not available to young children and non-human animals.

Our third passage makes it clear that, for Burge, some meta-representational concepts—but no specifically epistemic concepts—are required of reasons for perceptual beliefs:

in certain contexts, one can justify a perceptual belief by claiming I saw such and such. Such a claim can count as a reason (2003a, 529).

Notice that the claim featured here involves no obviously epistemic concepts. Indeed, the only meta-representational concept involved is see. That the claim can nevertheless count for Burge as a reason for a perceptual belief shows that he is not committed to the necessity of epistemic concepts for having reasons or justification for perceptual belief. One could of course have more sophisticated reasons for a perceptual belief, which do feature epistemic concepts. But epistemic concepts are not necessary for having such reasons, on Burge’s view. And because (P1*) requires access, for reason possession, to an epistemic principle, the fact that Burge holds that no epistemic concepts are necessary to having reasons for perceptual belief shows clearly that (P1*) is not true, even when restricted to reasons for perceptual belief.

In summary, Casullo has provided no evidence that Burge holds a strong conception of reason possession. And indeed Burge explicitly and repeatedly states that he does not, in one of the very articles that Casullo is supposed to be discussing.

5 ᅟ

Three matters remain. First, to evaluate Casullo’s claim that Burge’s notion of entitlement involves a merely terminological—as opposed to a conceptual or substantive—innovation. Second, briefly to trace the developments in Burge’s understanding of justification and entitlement which have occurred since his earliest published discussion of the matter. And finally, to explain precisely what entitlement is, how it differs from justification, and how it differs from previous conceptions of epistemic warrant.

Casullo notes that Burge maintains that entitlement is an epistemologically externalist form of warrant, and that justification is internalist. Casullo argues that, in fact, epistemic entitlement is only “moderately externalist.” And justification is only “moderately internalist.” This is a large part of Casullo’s broader campaign to show that Burge’s notion of epistemic entitlement is not new in any significant sense. Given the merely moderate natures of justification and entitlement, as Burge conceives of them, Casullo thinks it follows that Burge’s views are not significantly different from views held by others.

I think it is true that for Burge justification, taken as a whole, is only moderately internalist in character. In the final section of the paper, I will argue that there are different subspecies of justification. Some of these are only moderately internalist, but some are fully internalist—in the sense that they essentially involve access to the full warrant backing the belief in question, including its reliability. For now, I will simply grant to Casullo that there is a sense in which justification for Burge is only moderately internalist. For the central error that Casullo makes in this part of his article concerns rather entitlement.

Casullo argues that Burge’s notion of entitlement is at least moderately externalist, because entitlements do not satisfy either of two classical internalist conditions: (a) The requirement that the adequacy of a thinker’s ground for her warranted beliefs be accessible; and (b) The requirement that the epistemic principle governing the ground of a thinker’s belief that p be accessible. Because these conditions are phrased in terms of grounds, they do not clearly apply to Burge’s view. However, we can certainly allow that entitlement is not an internalist notion of epistemic warrant. But how strongly externalist is it?

In order to be strongly externalist in Casullo’s sense, a conception of epistemic warrant must deny that the ground of a thinker’s warranted belief that p is accessible. Does Burge deny this requirement? Here is what Casullo has to say on the matter:

Burge is not explicit on this point, but there is some evidence that suggests that he rejects strong externalism. Burge’s accounts of testimonial and perceptual entitlement suggest that an entitlement to believe that p requires that the subject have the concepts necessary to think the ground of that entitlement … [and some of] Burge’s remarks suggest that the conceptualization of perceptual representations is an essential feature of perceptual entitlement and that the ground of such an entitlement is conceptually accessible … An entitlement to believe that p is moderately internal: It involves access to the ground of the belief that p … (277–8).

Casullo proceeds to infer from this reasoning that Burge’s conception of entitlement is not strongly externalist. As noted, Casullo uses this result in arguing for the view that epistemic entitlement is not a new epistemological concept. His view is that Burge has simply coined a new term for a familiar and antecedently recognized epistemological notion.

The reasoning in this passage is confused. The claim concerning testimonial entitlement recalls Casullo’ earlier discussed argument. I explained in section 2 why this argument fails, and I will not rehearse the treatment here.

The second point made in the passage from Casullo is new. He claims that some of what Burge writes about perceptual entitlement suggests that the ground of such an entitlement is conceptually accessible. Casullo seems to think that the fact that the conceptualization of perceptual state content is allowed by Burge to be necessary for perceptual entitlement shows that the ground of the entitlement must be conceptually accessible. And if this were so, then perceptual entitlement would not be strongly externalist.

Casullo makes two claims here. First, he says that some of Burge’s remarks suggest that the conceptualization of perceptual representations is an essential feature of perceptual entitlement. There is more than mere “suggestion” of this thesis in Burge (2003a). It is an obvious commitment of the paper. Perceptual experiences are non-conceptual and therefore non-propositional. Entitlement is a form of epistemic warrant. Epistemic warrant attaches only to propositional states, attitudes, and events. It attaches paradigmatically to beliefs. One can have perceptual beliefs only insofar as perceptual experience content is transduced into conceptual (propositional) content. One can be entitled to perceptual belief only if one has perceptual belief. Therefore, the conceptualization of perceptual experience content is very clearly necessary to perceptual entitlement.

We can say something even stronger about the relevant sort of conceptualization. Thus far, we have not seen it to be anything other than a necessary condition for perceptual entitlement. And bare necessary conditions upon warrant do not automatically enjoy positive epistemic significance. However, the conceptualization of perceptual experience content which is necessary to the formation of perceptual belief actually plays a positive role in the explanation of perceptual entitlement. It is one of the things that must go well, in the exercise of the competence which produces the relevant belief, in order for one to come to enjoy perceptual entitlement—as opposed merely to something whose failure can undermine such warrant (Burge 2003a, section VII). It is part of the perceptual competence itself. We can therefore grant to Casullo that the relevant sort of conceptualization is indeed part of that in which perceptual entitlement consists.

What we cannot do, however, is grant Casullo’s second claim. Casullo’s inference from the fact that conceptualization is necessary to perceptual warrant to his conclusion that the ground of the entitlement must be conceptually accessible to the agent is clearly mistaken. The mistake is akin to one of the errors Casullo makes concerning interlocutionary warrant. We saw in section 2 that Casullo infers from the fact that there is a propositional state involved in interlocutionary entitlement that the entitlement is reason-based. But of course, being propositional in form is merely a necessary condition upon functioning as a reason—it is not also sufficient, as a moment’s reflection will show. More broadly, Casullo’s mistake is to expect to find one single thing that is the “ground” of interlocutionary entitlement. Casullo correctly finds comprehension of the content and force of the presentation-as-true as part of the warrant for testimonial belief. But he errs in thinking that this is the whole warrant.

The reasoning that Casullo employs in the perceptual case is also fallacious, and for largely the same reason. He thinks that since a part of that which bestows perceptual entitlement upon a thinker must be conceptual in form, the entitlement as a whole must derive from this conceptual state. And this does not follow. The relevant sort of conceptualization is necessary for perceptual entitlement, but it is not also sufficient.

Notice that even that proper part of perceptual entitlement that does involve conceptualization is not something to which a thinker has access. The conceptualization of perceptual state content occurs at the subpersonal level and is not something that any thinker has access to, even in principle.

As we have already seen, perceptual entitlement accrues to a cognizer in virtue of the presence of three warranting elements. First, the anti-individualistic individuation of perceptual state contents. Second, the reliability of the competence in its normal environment—the normal environment is the sort of environment relative to which the natures of the perceptual representational contents in which the competence issues are to be explained. Finally, the transition from non-conceptual perceptual experience to propositional perceptual belief must be negotiated in a proper way.

To what extent must a thinker have access to these three elements which combine to yield perceptual entitlement? The answer is to no extent. Being perceptually entitled requires only that one have a well-functioning perceptual and propositional competence. Young human children, and perhaps some subhuman animals, satisfy this condition. Yet it is plausible that none of them can think about any of the three elements. And it is certain that they need not do so, in order to enjoy warranted perceptual beliefs. Therefore, perceptual entitlement satisfies Casullo’s condition for being a strongly externalist conception of epistemic warrant. The fact that perceptual experience content must be conceptualized in order for there to be perceptual beliefs, and therefore perceptual beliefs to which we are entitled, has no tendency whatever to show that a relevant thinker must have conceptual access to a propositional formulation of the entitlement, or even to that part of it which explicitly concerns conceptualization.

We have been given no reason to doubt Burge’s claim that entitlement is a fully externalist notion of epistemic warrant. Because Casullo bases his negative verdict concerning the originality of Burge’s position principally upon his denial of this claim, his verdict has no force.

From this point, I will approach the questions concerning entitlement directly. The next section briefly explains changes that Burge’s views on the forms of epistemic warrant have undergone over time. The final section of the paper explains in some detail the distinction between justification and entitlement and, thereby, the nature of epistemic entitlement. It discusses how Burge’s views map onto the internalism/externalism distinction in epistemology. And it remarks briefly on the question of the originality of the notion of entitlement.

6 ᅟ

There are three primary phases discernible in Burge’s published work on the forms of epistemic warrant. The first is the article in which the distinction between justification and entitlement was first formulated (Burge 1993). The second is contained in Burge’s discussion of perceptual entitlement (Burge 2003a). The third is Burge (2013). While there is no space here for a complete discussion of these matters, I want to say enough to provide the reader with at least a general idea of how Burge’s thinking on entitlement and justification has evolved. I emphasize that some of what follows, especially concerning the transition from the first to the second phase (which Burge has not himself commented upon) is speculative. He has in his (2013) described in detail the second transition, so there we are on much firmer interpretive ground.

There are no substantive differences between the first and the second phases. But there are some differences of emphasis. I will highlight two of these. First, it was claimed in the earliest of these papers that “(j)ustification or entitlement aims at truth, since it rationally supports belief” (1993, 458), [and later in the paper Burge uses the expression “rational entitlement,” in a way that seems intended to cover all entitlement (1993, 470)]. As claims about warrant generally, these are true only in a very broad sense of “rationally.” Perceptually entitled belief, for example, is not strictly speaking rationally supported or supported by reason. As we have seen, the central warranting elements where entitled perceptual belief is concerned correspond to the individuation condition; the reliability condition; and the transition condition. None of these elements is of such a nature that it alone, or even in combination with the others, could provide reason for belief, in the strict sense. And because of this, I think it is slightly misleading to characterize perceptual entitlement as rationally supporting belief.

In a broader sense, of course, propositional formulations of entitlements do provide or indicate reasons for belief, including perceptual belief. In his later discussions, Burge avoids claiming that warrant generally rationally supports belief. I think that this shift of emphasis was borne of an increasing awareness of differences between, for example, perceptual entitlement and other forms of entitlement and warrant. Perceptual belief is not strictly supported by reason. Burge notes that it is, however, protected by reason—in the sense that reason certainly plays, for example, a constitutive corrective role for perceptual belief, when we consider the larger cognitive economy in which perceptual belief takes its position.

Second, Burge made in the early period the following claim: “The unsophisticated are entitled to rely on their perceptual beliefs” (1993, 458). This is clearly true. And it does not follow from this claim that only the unsophisticated are entitled to such reliance. But the claim may nevertheless suggest that perceptual entitlement is primarily the provenance of the unsophisticated, which is not true of Burge’s considered view.

All higher animals—human and subhuman, sophisticated and unsophisticated—are entitled to reliance upon their perceptual beliefs. Moreover, for cognizers generally entitlement is the basic and central form of perceptual warrant. It is relatively rare, even among those few who possess the requisite abilities, for one to hold a perceptual belief on the basis of reasons. Non-human animals and young human children, lacking meta-representational concepts, cannot have such warrants (justifications) for their perceptual beliefs. And, again, even where such conceptual equipment is available, it is typically utilized in forming perceptual warrants only in unusual circumstances. The point I am making here is quite minor. It is only that I doubt that Burge would later have made the claim with which the last paragraph began in the same unqualified manner.

The transition from the second to the third phase is marked by a slight change in understanding of entitlement and a more substantive change concerning justification. There are in Burge’s formulations of the nature of entitlement, in the first two phases especially, two sorts of characterization: (a) an access characterization; and (b) a functional-structural characterization. An example of the former is the claim that entitlements need not be understood by, or even conceptually accessible to, the warranted subject. The latter sort of characterization is exemplified by the claim that entitlement is warrant that does not consist in possession or use of a reason. The change concerning entitlement that marks the transition from the second to the third phase is a preference for the functional-structural characterization. Burge says that he “came to think that a functional-structural conception is more basic than a conception in terms of accessibility to consciousness” (2013, 28). I think that the chief reason for this change is the fact that on Burge’s later view even justifications need not be accessible to consciousness. Therefore, such accessibility cannot be the essence of the distinction between justification and entitlement.

In both of his (1993, 2003a) Burge held that justifications involve reasons that thinkers have and have access to. By “access,” Burge meant conscious access. The major change in his thinking about the nature of justification is rejection of this requirement upon justification. Burge came to believe that there is some empirical evidence for the existence of modular reasoning (he also came to doubt whether there is any genuine apriori, conceptual proscription of the notion of subpersonal-level reasoning). Such reasoning is not an act by the individual; it occurs in her cognitive subsystems and is not, even in principle, directly available to consciousness or reflection. Qua reasoning, such processing seems to be a form of justification. But it clearly fails to meet the earlier access requirement on justification; which, as I have already suggested, is one reason Burge came to prefer the functional-structural characterization of the difference between justification and entitlement. In fact, Burge thinks that warrant issued by modular reasoning is close to the borderline between justification and entitlement (2013, 33 n. 6).

We may now formulate somewhat more precisely and accurately Burge’s distinction between justification and entitlement. Justification is warrant by reason. Justification accrues to an individual either through inference—whether personal or subpersonal—or through understanding of epistemically self-sufficient propositional contents. Entitlement is warrant that does not consist, even partly, in reasons.

7 ᅟ

We began with three questions concerning entitlement. We already have the outlines of an answer to the first question. We saw in the previous section how Burge views the distinction between justification and entitlement, including ways that his understanding of the distinction has changed over time. There are, however, some important details concerning the respective natures of justification and entitlement which have not yet been discussed. Discussing them allows us to provide answers concerning the remaining two questions: (I) how strongly externalist is entitlement? (and how strongly internalist is justification?) and (II) how novel is Burge’s notion of epistemic entitlement?

A few distinctions must be must be made at the outset. I believe that they are worth making in themselves, but they are necessary in any case in order to understand how justification and entitlement fit into the division between epistemic internalism and externalism.

Take first entitlement. We must distinguish between general and specific entitlements. On Burge’s view, all of us (including any subhuman animals that reason) are entitled to rely upon perception, memory, and inference. Reasoners who are also critical reasoners are in addition entitled to rely upon certain claims to self-knowledge, self-understanding, and reflection.¹³ These are general entitlements in the sense that they are (in most cases defeasible) entitlements to rely upon the deliverances of certain representational competencies, whatever those deliverances may be. Specific entitlements, by contrast, derive from characteristics of the individual representational and cognitive situations at issue. Thus, small human children, for example, are entitled to their perceptual beliefs generally. And such a child might enjoy a specific entitlement to her belief that there is a red sphere before her.

General entitlements are explained by showing that the representational competency in question is representationally successful, given pertinent limitations on perspective and information. Thus, children’s entitlement to rely upon perception is explained by pointing out that their perceptual competencies are reliable in the latter’s normal environment. By contrast, a child’s particular perceptual belief concerning the red sphere is explained, quite roughly, by noting that her perceptual representation as of a red sphere is the product of a perceptual competence that is reliably veridical in its normal environment, that this characteristic of the content as of a red sphere is in a certain sense preserved across the transition from perceptual state- to conceptual-content, and that the child’s resulting belief that there is a red sphere before her is formed in a normal way from these materials.

Notice that it follows from this that any belief which a thinker is justified in holding is also one to which she is entitled, in both the general and specific senses. Justified beliefs are products either of understanding or inference. Each is a competence that we are entitled to rely upon, both in general and in relevant specific cases. The converse does not hold, however—there are many situations in which a thinker is entitled to a belief but lacks a justification for it. I will call warrants to which correspond entitlements but no justifications mere entitlements.

Now, there are on Burge’s view some justifications which essentially involve elements to which the thinker is merely entitled. I will call these cases of impure justification. Impure justifications are always justifications that involve inference (as we will see, those justifications that are rather products solely of understanding—cases in which the propositional content is itself a reason for believing it; that is to say, where the content is self-evident—are such that the distinction between “pure” and “impure” cannot arise). They are cases in which the thinker reasons from certain premises to a conclusion, where the thinker is warranted in believing the premises, in which the inferential pattern employed is relevantly truth-preserving or truth-tending, but in which the reasoner lacks sufficient understanding of the inferential pattern for it to be a reason (or even part of a reason) for her belief.

Of course, if the individual is warranted in believing the conclusion of the inference, on the basis of the inference, she must also be warranted in employing the inferential pattern. But in cases of impure justification, this warrant is a mere entitlement. Impure justification therefore involves justification in the sense that the premise-steps of the inference are not only themselves warranted but also contribute toward providing reason, for the individual, for the conclusion-step. But it also essentially involves mere entitlement, in the sense that the thinker is entitled but not justified in her reliance upon the relevant inferential pattern.

In cases of pure justification, by contrast, the thinker relies on no mere entitlements (though we have seen that in every such case there are entitlements, both general and specific, which attach to the relevant belief). Justification that involves only understanding is always pure. This is because all that is necessary for such warrant is understanding of the representational content in question. Such understanding itself is a reason for the correlative belief. In these cases, the understanding is, or provides, the full warrant. So no mere entitlements are at issue.

Justification that involves inference is sometimes pure. It is pure when the thinker has enough understanding of the inferential pattern in question to appreciate, in effect, that (and perhaps why) it is truth-preserving or truth-tending. Then, the understanding and use of the inferential pattern itself provides (partial) reason for the conclusion-step of the inference, for the individual. In Burge’s view, many cases of deductive inferential justification are pure in this sense. Inductive mathematical justification can also be pure. The case of empirical inductive inference is complicated by the fact that much inductive reasoning is not well understood enough—either by the individual or perhaps even by philosophers and scientists—that use of relevant inductive inferential patterns can itself provide reason or justification for belief. All empirical justifications may therefore be impure. Nevertheless, we are often entitled to rely upon empirical inference, and therefore often (impurely) justified in doing so.

The importance of these distinctions will now be made apparent. The first of our two remaining questions concerns the relations of the forms of warrant, as Burge views them, to the internalism/externalism distinction in epistemology.¹⁴ More specifically, the issues are as follows: how internalist is Burge’s conception of justification? And, how externalist is his notion of entitlement? And of course, the answers to these questions are directly relevant to the final question, concerning the extent to which Burge’s notion of epistemic entitlement is original or innovative.

We have seen that justification is not a completely univocal notion for Burge. Justification can be either understanding- or inference-based. Understanding-based justification is fully internalist, in this sense: a thinker who warrantedly believes a content, on the basis solely of her understanding of it, has full conceptual access to the warrant. This is because her understanding is the warrant. She also has access, in the relevant sense, to the reliability of the warranting state. She can understand, that is to say, not only that the content is true but also—at least to a minimal extent—why it must be true (or why it must be true when judged). One class of justifications is therefore fully internalist, for Burge.

Among inference-based justifications, we must distinguish the pure from the impure. Pure justifications involve no mere entitlements. Where such justifications are concerned, every aspect of the warrant functions as a reason for the subject. A thinker who competently executes a deductive inference, for example; who is warranted in her beliefs in the premise-steps of the argument; and who furthermore has sufficient understanding of the relevant inferential pattern, enjoys pure justification for her belief in the conclusion-step of the argument. In this case, too, the warrant is fully internalist. The subject’s beliefs in the premise-steps of the argument are conceptually accessible to her. This follows from the fact that she has sufficient understanding—competence understanding—of the contents of the premise-steps of the argument. And part of having the relevant understanding of the inferential pattern in question is possessing some appreciation of the fact that it is truth-preserving. Thus, not only the warrant but also its reliability is accessible to the cognizer. This is what it means, as we are here understanding the relevant notions, for a warrant to be fully internalist.

What of impure justification? We saw earlier that many cases of deductive inference, and most cases of inductive inference, feature or involve warrants which only partly consist in reasons and, therefore, are only impurely justifications. Such cases involve mere entitlements. That is to say, they feature certain entitlements to which no justification corresponds. Impure justifications are only partly internalist. They involve premise-steps of inferences which do function as reasons for the conclusion-step. This aspect of the warrant is internalist, inasmuch as the thinker must have access to (be able to think) the contents of the premise-steps. However, the subject’s warrant, in such cases, to rely upon the relevant inferential pattern is a mere entitlement. There is not enough understanding of the pattern for it to count as reason, for the individual, for belief in the conclusion-step of the argument. Thus, one aspect of the warrant is internalist. This is the understanding of the premise-steps of the inference. The remainder of the warrant, however, is externalist in character. The reliability of the inferential-pattern is not something to which the subject has access. Instances of impure justification are therefore are only partly internalist.

We know that Burge characterizes entitlement as an externalist form of epistemic warrant. But how strongly externalist is entitlement? The answer to this question has already been prefigured. Entitlements were characterized by Burge at one point as epistemic warrants which need not be accessible to the entitled subject, even upon reflection. Later, he came to prefer the functional-structural characterization, which had been also been employed in his early work. On the second conception, an entitlement is an epistemic warrant that does not consist in possession or use of a reason. So far as the internalism/externalism issue is concerned, these alternative characterizations are close to being equivalent.¹⁵ There are always elements to the warranting force of an entitlement which are not accessible to the warranted individual. Examples are the necessary, constitutive connections between propositional content and rational sources, on one hand, and between rational sources and truth, on the other—in the case of interlocution; and the anti-individualistic individuation of perceptual-state content (together with the reliability of the perceptual competence), in the case of perception. None of these warranting elements is commonly accessible to warranted believers who employ the relevant competencies. None of them needs to be accessible in order for entitlement to obtain. The proper functioning of the relevant representational competence, which is what gives rise to relevant general entitlements, does not require any sort of access.

Similarly, none of these elements need function as a reason, for individuals to enjoy the entitlement. Indeed, in some cases, the elements themselves—as opposed to a propositional characterization or formulation of them—cannot function as a reason. Reasons are propositional. But in any event, whether the elements of an entitlement are propositional or not, they do not function as reasons. And no access to them is necessary for the obtaining of the warrant. Therefore, entitlement is in one sense fully externalist.

In another, extreme sense, entitlement is not a fully externalist notion of warrant, inasmuch as no entitlement consists purely in reliability. All epistemic warrants, for Burge, necessarily involve representational competencies which are part of the psychological makeup of the cognizer. I think this concession is not terribly significant, however, inasmuch as any view according to which pure reliability—or, indeed, any sort of reliability which is completely divorced from psychological capabilities of the believer—is sufficient for warrant is not a serious contender in the debate.

What is important is obviously not whether we label entitlement (or justification) “internalist” or “externalist.” There are many conceptions of the designations of these terms. What is important here is that we understand the overall structure, as well as the details, of Burge’s views on the forms of epistemic warrant.

This brings us to our final topic. Casullo is clearly motivated by a desire to minimize the significance of the notion of epistemic entitlement. His efforts in this direction are not successful, however. The attempt rests crucially upon failure to understand even the rudiments of Burge’s views.

I see no particular need to defend the originality of Burge’s epistemological views. His work speaks for itself. However, because the issue has been raised, I will make three points in closing.

The first is that Burge has, in articulating his notion of epistemic entitlement and distinguishing it from justification, arguably gone much further than any previous theorist in accommodating the insights of both epistemic internalism and epistemic externalism. We have seen that Burge has constructed an epistemic model which does justice both to the rationalist insistence upon the reality and significance of cases of self-evident states and attitudes, on one hand, and to the empiricist insight that there are certain other belief-forming competencies—including most prominently perception—which differ significantly from the rationalist paradigm of warranted belief and knowledge, on the other. His view allows for, and indeed insists upon, the reality of certain types of warrant which are fully internalist (it bears recollecting that not all warrants which are fully internalist, on Burge’s view, involve self-evident contents or states). But Burge nevertheless maintains that many epistemic warrants—epistemic entitlements—are strongly externalist, in the sense that the warranted individual need not have any sort of access to the warrant.

Insofar as one is working with a univocal conception of epistemic warrant, there is pressure to view warrant as amounting to the same thing in all areas in which the notion applies. I think this is one principal reason that traditional accounts of warrant or justification were not apt to acknowledge significant differences between, for example, inferential warrant and perceptual warrant, respectively. They tended, indeed, to assimilate the latter to the former.

Of course, distinctions had been made previously, even intratheoretically, between different forms of warrant. I believe, however, that Burge’s distinction between entitlement and justification is the first such distinction which is principled, independently motivated, and thoroughly developed.

A related point is that no one working in these areas has ever provided, as Burge has, an effective and principled restriction upon the reliability requirement for perceptual warrant. Epistemic warrant requires reliability in some sense. But New Evil Demon cases show that circumscribed instances of unreliability are compatible with warrant. One of the most significant challenges in contemporary epistemology is to reconcile these two facts in a principled and otherwise satisfactory way. I believe that Burge has done this.

Finally, and perhaps most fundamentally, I think that Burge has gone much further than any previous philosopher in providing a conceptual underpinning for the widely accepted necessary, constitutive connection between epistemic warrant and truth. Most epistemologists have acknowledged, albeit sometimes only implicitly or even reluctantly, that considerations which do not in some way bear positively upon the truth of a belief cannot contribute to an epistemic warrant for the belief. Yet no one has ever explained, in a detailed and reflective way, why this is. Burge provides such an explanation. I will end by providing a rough summary of this explanation.

Representation-yielding systems—such as perceptual- and propositional attitude-producing competencies—necessarily have the function of producing veridical or correct representations. The claim is not that this is the only function that they have. Many such systems have biological and practical functions as well. But it is apriori necessary that, for any representational competence, it has a function in producing veridical representations. This is a necessary, constitutive part of what such a competence is for.

Functions inevitably yield standards for meeting them, some of which are norms. Contributions toward meeting norms are goods for the system, its deliverances, or the cognizers whose psychological systems include the relevant competence—relative to the function in question. The primary representational norm is veridicality or correctness. A relevant entity meets the primary representational norm insofar as it is veridical or yields veridical representations. A secondary representational norm is to be reliably veridical. A tertiary one is to be reliably veridical, given any pertinent limitations upon information and perspective.

Reflection upon the relevant norms—including reflection upon cases—yields the result that the third representational norm mentioned sets belief’s primary epistemic norm, epistemic warrant. Being epistemically warranted is being produced by a natural representational competence which is reliably veridical in the competence’s normal environment (Burge 2003a).

The account toward which I have gestured here explains the constitutive connection between epistemic warrant and truth. It also explains the fact that reliability in every possible circumstance is not necessary to every kind of epistemic justification or entitlement. It is not necessary, in particular, for perceptual entitlement. Burge’s account not only yields the correct verdict in the case of perceptual warrant, concerning New Evil Demon cases, but it explains the correctness of the verdict. I believe that the explanation which I have summarized is one of Burge’s most original and enduring contributions to philosophy.