You are here

1: Justification, Internalism, and Deontology

My topic is warrant: that, whatever precisely it is, which together with truth makes the difference between knowledge and mere true belief. More specifically, my topic is contemporary views of warrant. I shall begin by looking briefly at the twentieth-century received tradition with respect to warrant; but first, how shall we initially pin down, or locate, or characterize this property or quantity I propose to discuss? It is that which distinguishes knowledge from mere true belief, of course; but note also that there is obviously something normative or evaluative about warrant. To attribute warrant to a belief is to appraise that belief, and to appraise it favorably; and we use such terms as ‘warranted’, ‘justification’, ‘justified’, and the like as “terms of epistemic appraisal.”1 To say that a belief is warranted or justified for a person is to evaluate it or him (or both) positively; his holding that belief in his circumstances is right, or proper, or acceptable, or approvable, or up to standard. We evaluate a person's beliefs (more exactly, her believings) as warranted, or justified, or rational, or reasonable, contrasting them with beliefs that are unwarranted, unjustified, irrational, unreasonable. The evidentialist objector to theistic belief,2 for example, claims that a theist who believes in God without evidence or argument is so far forth unwarranted and unjustified in that belief; he offers a negative appraisal of the belief or its holder. (Perhaps he claims that in believing in God in that way she is flouting some duty, or (more charitably) is suffering from a sort of cognitive dysfunction, or (still more modestly) that the module of our cognitive establishment that issues in theistic belief is not aimed at truth but at something else.)3

In the same way we may appraise the belief that all contemporary flora and fauna arose by way of random genetic mutation and natural selection from primitive forms of life, which in turn arose via similarly ateleological processes from inorganic material. And of course the less spectacular beliefs of everyday life are also subject to such evaluation and appraisal. We appraise a person's beliefs, but also her skepticisms or (to use another Chisholmian term) her withholdings, her refrainings from belief. An unduly credulous person may believe what she ought not; an unduly skeptical (or cynical) person may fail to believe what she ought. Further, we may hold a belief more or less strongly, more or less firmly; we appraise not only the belief itself, but also the degree to which it is accepted. If I believe that Homer was born before 800 B.C. and believe this with as much fervor as that New York City is larger than Cleveland, then (given what are in fact my epistemic circumstances) my degree of confidence in the former proposition is excessive and unwarranted.

Finally, warrant comes in degrees. Some of my beliefs have more by way of that quantity for me than others. Thus my belief that I live in Indiana has more by way of warrant, for me, than my belief that Shakespeare wrote the plays commonly attributed to him; my belief that 2 + 1 = 3 has more warrant than my belief that the Axiom of Choice is equivalent to the Hausdorff Maximal Principle. (This is not to say, of course, that I am not equally rational and equally justified in accepting these beliefs to the degrees to which I do in fact accept them; for I believe the latter member of each pair less firmly than the former.) But then we can distinguish degrees of positive epistemic status, at least for a given person.4 Initially, then, and to a first approximation, warrant is a normative, possibly complex quantity that comes in degrees, enough of which is what distinguishes knowledge from mere true belief.

Second, a problem that is less trivial than it initially seems: what shall we call this quantity? I propose to call it ‘warrant’; but those of us brought up in that benighted pre-Gettier era learned at our mother's knee that knowledge is justified true belief; and even in this enlightened post-Gettier age we still think of justification and knowledge as intimately related. So why not call this property ‘justification’? Because it would be both misleading and unfair. ‘Justification’ suggests duty, obligation, requirement; it is redolent of permission and rights; it brings to mind exoneration, not being properly subject to blame—it connotes, in a word (or two) the whole deontological stable. And the problem is that one of the main contending theories or pictures here (one with impressive historical credentials going back at least to Descartes and Locke) explicitly explains the quantity in question at least partly in terms of fulfilling one's epistemic duties, satisfying one's epistemic obligations, conforming to one's epistemic requirements. To use the term ‘justification’, then, as a name for that quantity would be to give this theory and its relatives a confusing and unwarranted (if merely verbal) initial edge over their rivals. So ‘justification’ is not the right choice. In earlier work5 I borrowed Roderick Chisholm's more neutral term “positive epistemic status” as my official name for the quantity in question. That locution, however, is too long; so I shall use the term ‘warrant’ in its place. Of course, ‘warrant’ has deontological associations of its own (even if they are not quite so insistent); perhaps (as Ernest Sosa suggested in conversation) ‘epistemic aptness’ is a better term. On balance, however, I prefer ‘warrant’—but we must be careful not to be misled by its residual deontological insinuations.

I. Internalism

The main story of twentieth-century epistemology is the story of three connected notions: justification, internalism, and deontology. I propose to begin my study of contemporary views of warrant by examining some internalist theories of warrant; but what is this ‘internalism’? What does it mean to call someone an internalist? The term is in considerable disarray. Different people use it differently; it expresses distinct ideas loosely related by analogies and family resemblance. How can we gain an understanding of internalism? What is the central notion here, the notion in terms of which we can see how the rest of those loosely related ideas hang together? What is the source of the attraction of internalism, and what makes it plausible? And how is it connected with the fundamental question of the nature of warrant?

The basic internalist idea, of course, is that what determines whether a belief is warranted for a person are factors or states in some sense internal to that person; warrant conferring properties are in some way internal to the subject or cognizer. But in what way? The pH level of my blood is a condition internal to me, as is the size of my heart; but clearly these are not internal in the relevant way. So what is the relevant way? The first thing to see, I think, is that this notion of internality is fundamentally epistemic. Warrant and the properties that confer it are internal in that they are states or conditions of which the cognizer is or can be aware; they are states of which he has or can easily have knowledge; they are states or properties to which he has cognitive or epistemic access. But not just any old epistemic access will do; I have epistemic access to the distance from the earth to the moon and to the depth of the Pacific Ocean (I own an encyclopedia), but that is not access of the relevant sort. What is required is some kind of special access. Perhaps (as Chisholm suggests) S can determine by reflection alone whether a belief has warrant for him; or perhaps he can determine with certainty whether a belief has the property that grounds and confers justification; or perhaps there is a certain kind of mistake—a mistake about warrant or the properties that confer it—that he cannot nonculpably make. So the relevant sense of ‘internal’ is strongly epistemic; the internalist holds that a person has some kind of special epistemic access to warrant and the properties that ground it.

The externalist, by contrast, holds that warrant need not depend upon factors relevantly internal to the cognizer; warrant depends or supervenes upon properties to some of which the cognizer may have no special access, or even no epistemic access at all. Take a paradigm externalist view: that of the early Alvin Goldman, for example, who holds “to a first approximation” that a belief has warrant if and to the degree that it is produced by a reliable belief-producing mechanism.6 What makes this view externalist? Why isn't it internalist? After all, my belief-producing mechanisms, unlike my house or my car, are surely internal to me. What makes the claim externalist, I suggest, is that the properties that on this view confer warrant are not such that I need have any special kind of epistemic access to them. On externalist views, warrant-making properties are such properties (of a belief) as being produced by a reliable belief-producing mechanism, or standing in a causal chain appropriately involving the subject of belief, or standing in probabilistic relation R to certain other relevant propositions; and none of these properties is one to which we have the relevant kind of special access.

The basic thrust of internalism in epistemology, therefore, is that the properties that confer warrant upon a belief are properties to which the believer has some sort of special epistemic access. But why think a thing like that? What is the source of internalism, and why is it attractive? To see why, we must turn to a different but connected idea: that of epistemic justification.

II. Justification

It would be colossal understatement to say that Anglo-American epistemology of this century has made much of the notion of epistemic justification. First, of course, there is the widely celebrated “justified true belief” (JTB) account or analysis of knowledge, an analysis we imbibed with our mothers’ milk. According to the inherited lore of the epistemological tribe, the JTB account enjoyed the status of epistemological orthodoxy until 1963, when it was shattered by Edmund Gettier with his three-page paper “Is Justified True Belief Knowledge?”7 After 1963 the justified true belief account of knowledge was seen to be defective and lost its exalted status; but even those convinced by Gettier that justification (along with truth) is not sufficient for knowledge still mostly think it necessary and nearly sufficient for knowledge: the basic shape or contour of the concept of knowledge is given by justified true belief, even if a quasi-technical fillip or addendum (“the fourth condition”) is needed to appease Gettier. Of course there is an interesting historical irony here: it isn't easy to find many really explicit statements of a JTB analysis of knowledge prior to Gettier. It is almost as if a distinguished critic created a tradition in the very act of destroying it.8 Still, there are some fairly clear statements of a JTB analysis of knowledge prior to Gettier. Thus, according to C. I. Lewis, “Knowledge is belief which not only is true but also is justified in its believing attitude.”9 A. J. Ayer, furthermore, speaks of knowledge as “the right to be sure”;10 for reasons that will be clearer a bit further along, I believe this is a statement of a JTB account of knowledge. And even if there are few explicit published statements, there is an extensive and impressive oral tradition.

So one element in the received epistemological tradition in the twentieth century is that justification is necessary and (with truth) nearly sufficient (sufficient up to Gettier problems) for knowledge. But what exactly is justification? Here we are offered a wide and indeed confusing assortment of alternatives. Consider the following: In the third edition of Theory of Knowledge (1989) Roderick Chisholm speaks of the question ‘what is knowledge?’ and suggests that

The traditional or classic answer—and the one proposed in Plato's dialogue, the Theaetetus—is that knowledge is justified true belief. (p. 90—see also the quotation from The Foundations of Knowing in n. 8)

According to Roderick Firth,

To decide whether Watson knows that the coachman did it we must decide whether or not Watson is justified in believing that the coachman did it. Thus if Watson believes that the coachman did it, we must decide whether his conclusion is based rationally on the evidence.11

Laurence BonJour12 holds that the traditional JTB account of knowledge is “at least approximately correct” (pp. 3–4); he continues:

We cannot, in most cases at least, bring it about directly that our beliefs are true, but we can presumably bring it about directly (though perhaps only in the long run) that they are epistemically justified. (p. 8)

It follows that one's cognitive endeavors are epistemically justified only if and to the extent that they are aimed at this goal, which means very roughly that one accepts all and only those beliefs which one has good reason to think are true. To accept a belief in the absence of such a reason… is to neglect the pursuit of truth; such acceptance is, one might say, epistemically irresponsible. My contention here is that the idea of avoiding such irresponsibility, of being epistemically responsible in one's believings, is the core of the notion of epistemic justification. (p. 8)

If a given putative knower is himself to be epistemically responsible in accepting beliefs in virtue of their meeting the standards of a given epistemological account, then it seems to follow that an appropriate metajustification for those principles must, in principle at least, be available to him. (p. 10)

Hilary Kornblith joins BonJour in linking justification with epistemically responsible action:

Justified belief is belief which is the product of epistemically responsible action; epistemically responsible action is action guided by a desire to have true beliefs. The epistemically responsible agent will thus desire to have true beliefs and thus desire to have his beliefs produced by reliable processes.13

Earl Conee and Richard Feldman claim that:

Doxastic attitude D toward proposition p is epistemically justified for S at t if and only if having D toward p fits the evidence S has at t.

Conee adds that:

A person has a justified belief only if the person has reflective access to evidence that the belief is true.… Such examples make it reasonable to conclude that there is epistemic justification for a belief only where the person has cognitive access to evidence that supports the truth of the belief. Justifying evidence must be internally available.14

William P. Alston considers and rejects an account of justification in terms of responsibility or duty fulfillment and proposes instead that:

S is Jeg [‘e’ for ‘evaluative’ and ‘g’ for ‘grounds’] justified in believing that p iff S's believing that p, as S did, was a good thing from the epistemic point of view, in that S's belief that p was based on adequate grounds and S lacked sufficient overriding reasons to the contrary.15

“Adequate grounds,” furthermore, “are those sufficiently indicative of the truth of p.”16 Alston also reports that he finds “widely shared and strong intuitions in favor of some kind of accessibility requirement for justification.”17

According to Ernest Sosa:

What does matter for justification is how the subject performs with respect to factors internal to him,… it does not matter for justification if external factors are abnormal and unfavorable so that despite his impeccable performance S does not know.18

The evil demon problem for reliabilism is not Descartes’ problem of course, but it is a relative. What if twins of ours in another possible world were given mental lives just like ours down to the most minute detail of experience or thought, etc., though they were also totally in error about the nature of their surroundings, and their perceptual and inferential processes of belief acquisition accomplished very little except to sink them more and more deeply and systematically into error? Shall we say that we are justified in our beliefs while our twins are not? They are quite wrong in their beliefs, of course, but it seems somehow very implausible to suppose that they are unjustified [his emphasis].19

In “Justification and Truth,” Stewart Cohen holds that the demon hypothesis entails that “our experience is just as it would be if our cognitive processes were reliable” and hence that we would be justified in believing as we do in fact, when our cognitive processes are reliable. So reliability, he argues, cannot be a necessary condition of justification. He also seems to join BonJour in thinking of justification as a matter of epistemic responsibility. Keith Lehrer joins Cohen in elaborating this view:

Imagine that, unknown to us, our cognitive processes, those involved in perception, memory and inference, are rendered unreliable by the actions of a powerful demon or malevolent scientist. It would follow on reliabilist views that under such conditions the beliefs generated by those processes would not be justified. This result is unacceptable. The truth of the demon hypothesis also entails that our experiences and our reasonings are just what they would be if our cognitive processes were reliable, and therefore, that we would be just as well justified in believing what we do if the demon hypothesis were true as if it were false. Contrary to reliabilism, we aver that under the conditions of the demon hypothesis our beliefs would be justified in an epistemic sense. Justification is a normative concept. It is an evaluation of how well one has pursued one's epistemic goals. Consequently, if we have reason to believe that perception, for example, is a reliable process, then the mere fact that it turns out not to be reliable, because of some improbable contingency, does not obliterate our justification for perceptual belief. This is especially clear when we have good reason to believe that the contingency, which, in fact, makes our cognitive processes unreliable, does not obtain.20

The early Alvin Goldman, on the other hand, takes a very different view:

The justificational status of a belief is a function of the reliability of the process or processes that cause it, where (as a first approximation) reliability consists in the tendency of a process to produce beliefs that are true rather than false.21

And according to the later Goldman of Epistemology and Cognition:22

(P1) A cognizer's belief in p at time t is justified if and only if it is the final member of a finite sequence of doxastic states of the cognizer such that some (single) right J-rule system licenses the transition of each member of the sequence from some earlier state(s) [where]… (ARI) A J-rule system R is right if and only if R permits certain (basic) psychological processes, and the instantiation of these processes would result in a truth ratio of belief that meets some specified high threshold (greater than .5) (pp. 83, 106).

Now: how shall we understand this blooming, buzzing confusion with respect to justification? There seem to be at least four central ideas in these quotations. First, of course, there is the pervasive connection between justification and warrant or knowledge. Second (BonJour, Cohen, Kornblith, the first Alstonian notion), justification is a matter of epistemic responsibility; a belief is justified if the person holding it is not guilty of epistemic irresponsibility in forming or maintaining it.23 Third (Alston, Conee, Lehrer and Cohen, Cohen, Sosa, Lehrer), we have the suggestion that there is an internalist component to justification (Goldman seems to demur): the believer must have cognitive access to something important lurking in the neighborhood—whether or not he is justified, for example, or to the grounds of his justification (that by virtue of which he is justified [Alston]), or to the connection between those grounds and the justified belief. Of course this must be some kind of special access; perhaps S can determine by reflection alone, for example, whether he is justified (Alston, Conee, Lehrer and Cohen, BonJour, Chisholm). Fourth, there is to be found in many of the quotations the idea that justification is a matter of having evidence, or at least depends upon evidence (Alston, Firth, Conee, Conee and Feldman, Chisholm).

Finally, there is another and broader (if vaguer) notion of justification also evident in some of the quotations (Lehrer and Cohen, Cohen, Sosa): one that is hard to put at all precisely but seems to be a generalization of the notion of justification taken deontologically. This is the idea that everything is going properly from the perspective of the knowing subject, or insofar as the knowing subject himself is concerned, as knowing subject. What is involved in the” perspective of the knowing subject,” or “the knowing subject as the knowing subject”? This isn't entirely easy to say. Part of what is involved, however, is that the way in which the subject's experience is connected with the world is excluded. Perhaps he is a brain in a vat, or perhaps his experience is induced by a Cartesian demon and is wholly misleading; nevertheless he is justified if everything goes right from there on, so to speak—if, in particular, he reasons aright and forms the appropriate beliefs given the course of his experience. Call this broad justification. There will be different ways of filling in what it is for things to go right from there on. It could be a matter of forming beliefs responsibly (Cohen), or in such a way as to properly pursue one's epistemic goals (Lehrer and Cohen); it could be a matter of a sort of intrinsic fittingness relating experience and belief (see chapter 3); and it could also be a matter of faculties downstream of experience (reasoning and belief formation, say) functioning properly, being subject to no dysfunction.

So we have several different suggestions as to what justification is: being formed responsibly, being reliably produced, being such that the believer has adequate evidence, being formed on the basis of an internally accessible and truth conducive ground, being an evaluation of how well the believer has pursued her epistemic goals, and so on. There is also the connection with knowledge, with internalism, and with evidence. How shall we understand this welter of views as to the nature of justification? And how does it happen that justification is associated, in this way, with evidence? And what is the source of the internalist requirement, and how does it fit in? And why is justification associated, in this way, with knowledge?

III. Classical Deontologism

Here what we need is history and hermeneutics: archaeology, as Foucault says (although, pace Foucault, there is no reason to think we will uncover a hidden political agenda). We must go back to the fountainheads of Western theory of knowledge, those twin towers of Western epistemology, Descartes and Locke. For some topics—the nature of proper names, perhaps, or the question of serious actualism (that is, the question whether objects can have properties in possible worlds in which they do not exist)—a grasp of the history of the topic is not obviously essential to a grasp of the topic. Not so for internalism and epistemic justification: to understand the contemporary situation of those notions we must look carefully at their history, in particular at some of the ideas of Descartes and, perhaps even more important, Locke. Commonly (and correctly) thought of as the fountainheads of the tradition of classical foundationalism, Descartes and Locke are equally and perhaps even more significantly the fountainheads of the tradition of classical internalism.24 Contemporary discussion has paid scant attention to the source and origin of the internalist tradition. This has led to confusion, and failure to see the main issues with real clarity, or even (more modestly) the degree of clarity we can attain. Here the ahistoricism of contemporary analytic philosophy has served us ill. To get a proper understanding of justification and internalism, to understand the basic internalist insight, we must trace that tradition back to its source in Descartes and (more crucially) Locke. I have neither the space nor the competence for a really proper historical investigation here; what follows, I fear, is little more than a gesture in that direction. It is an important direction, however, and it is there we must look in order to understand contemporary internalism.

The first thing to see is that for Descartes and Locke the notion of duty or obligation plays a central role in the whole doxastic enterprise. Rodericks Firth25 and Chisholm26 (and other contemporaries) point out that there is a strong normative component in such basic epistemological concepts as justification and warrant; Chisholm (as we shall see) goes on to claim that this normative component is really deontological,27 having to do with (moral) duties, obligations, requirements. In the contemporary context it required a real insight to see clearly the normative character of these epistemic concepts. For Descartes and Locke, however, deontological notions enter in a way that is explicit in excelsis. Their thought—that duty and warrant are closely related—is not, of course, inevitable; you might think that we have little or no control over our believings, so that we have few significant duties in connection with them. Or you might think that while indeed we have considerable control over what we believe, the deontological notions of obligation, duty, and permission do not apply; the most we can say, you think, is that some beliefs (or some habits of belief formation and retention, or some epistemic strategies) are more valuable in themselves or more useful to us than others, but no belief or way or strategy of forming and holding beliefs is such that it is obligatory either to accept it or to reject it.

For Descartes and Locke, however, deontological notions enter crucially. Following Augustine (De Libero Arbitrio) Descartes gives his classical account of the origin of error:

But if I abstain from giving my judgment on any thing when I do not perceive it with sufficient clearness and distinctness, it is plain that I act rightly.… But if I determine to deny or affirm, I no longer make use as I should of my free will, and if I affirm what is not true, it is evident that I deceive myself; even though I judge according to truth, this comes about only by chance, and I do not escape the blame of misusing my freedom; for the light of nature teaches us that the knowledge of the understanding should always precede the determination of the will. It is in the misuse of the free will that the privation which constitutes the characteristic nature of error is met with.28

According to Descartes, error is due to a misuse of free will, a misuse for which one is guilty and blameworthy (“and I do not escape the blame of misusing my freedom”). There is a duty or obligation not to affirm a proposition unless we perceive it with sufficient clarity and distinctness; that there is such a duty is something we are taught by “the light of nature.”29 According to Descartes, being justified is being within our rights, flouting no epistemic duties, doing no more than what is permitted. We are justified when we regulate or order our beliefs in such a way as to conform to the duty not to affirm a proposition unless we perceive it with sufficient clarity and distinctness.

In his Essay Concerning Human Understanding, Locke is, if anything, even more explicit about this deontological component of the epistemic:

Faith is nothing but a firm assent of the mind: which if it be regulated, as is our duty, cannot be afforded to anything, but upon good reason; and so cannot be opposite to it. He that believes, without having any reason for believing, may be in love with his own fancies; but neither seeks truth as he ought, nor pays the obedience due his maker, who would have him use those discerning faculties he has given him, to keep him out of mistake and error. He that does not this to the best of his power, however he sometimes lights on truth, is in the right but by chance; and I know not whether the luckiness of the accident will excuse the irregularity of his proceeding. This at least is certain, that he must be accountable for whatever mistakes he runs into: whereas he that makes use of the light and faculties God has given him, and seeks sincerely to discover truth, by those helps and abilities he has, may have this satisfaction in doing his duty as a rational creature, that though he should miss truth, he will not miss the reward of it. For he governs his assent right, and places it as he should, who in any case or matter whatsoever, believes or disbelieves, according as reason directs him. He that does otherwise, transgresses against his own light, and misuses those faculties, which were given him.30

Here again there is the clear affirmation that we have an epistemic or doxastic duty: a duty, for example, not to afford a firm assent of the mind “to anything, but upon good reason.” To act in accord with these duties or obligations is to be within one's rights; it is to do only what is permissible; it is to be subject to no blame or disapprobation; it is to have flouted no duties; it is to be deontologically approvable; it is, in a word, to be justified. Indeed the whole notion of epistemic justification has its origin and home in this deontological territory of duty and permission, and it is only by way of analogical extension that the term ‘epistemic justification’ is applied in other ways. Originally and at bottom, epistemic justification is deontological justification: deontological justification with respect to the regulation of belief.

Now perhaps Descartes accepts a justified true belief account of knowledge; for he thinks that one is justified only in accepting just those propositions that are clear and distinct; and those propositions are just the ones he thinks we know. Locke, however, clearly does not; for him, knowledge and belief are two quite different states, and duty or obligation applies only to the latter. Your duty, he says, is to regulate your beliefs in such a way that you believe a proposition only if you have good reasons for it; those reasons would be propositions that are certain for you, and of which, accordingly, you have knowledge. But knowledge itself does not involve fulfillment of duty, epistemic or otherwise; indeed, here the dual concepts of obligation and permission do not really apply. Knowledge, he says, is a matter of noticing connections among ideas, and is only of what is certain. But if a proposition is certain for me, he holds, then there is no question of regulating my belief with respect to it. The reason is that I have no control with respect to such propositions, so that whether I believe is not up to me. Speaking of self-evident propositions, he says “all such affirmations, and negations, are made without any possibility of doubt, uncertainty or hesitation, and must necessarily be assented to, as soon as understood” (Essay, IV, vii, 4, p. 269). While Locke speaks here of just one of the several kinds of items of which we can have certainty, he clearly thinks the same thing about the others.

So Locke does not equate warrant—that quantity enough of which is sufficient, with truth, for knowledge—with epistemic justification, or, as we could call it to remind ourselves of the reference to duty and obligation, deontological epistemic justification. Nevertheless, deontological justification is of the very first importance for him as it is for Descartes. His central thought is that being justified in holding a belief is having fulfilled one's epistemic duties in forming or continuing to hold that belief. He adds (in agreement with Descartes) that if I go contrary to my epistemic duty, I lose a most important epistemic quality; and both hold that if I do so, then, although I may happen to “light on the truth,” it will be (no thanks to me) only by mere chance, by accident. (An important component of our idea of knowledge is that if a person just happens to “light on truth,” if he believes what is true by chance or accident, then the belief in question may be as true as you please but does not constitute knowledge.)31 This thought—the thought that being justified in holding a belief is having fulfilled one's epistemic duties in forming or continuing to hold that belief—is the fons et origo of the whole internalist tradition. It is this notion of deontological justification that is the source of internalism: deontology implies internalism. Let me explain.

IV. Deontology and Internalism

First on the Cartesian—Lockean deontological conception of justification, whether S's beliefs are justified, obviously, is up to S and within her control. He who has justification, says Locke, “may have this satisfaction in doing his duty as a rational creature, that though he should miss truth, he will not miss the reward of it [that is, of being justified, having done his duty].” The classical internalist thinks we need give no hostages to fortune when it comes to justification; here our destiny is entirely in our own hands. The fates may conspire to deceive me; I could be wrong about whether there is an external world, or a past, or other persons; for all I can be really sure of, I may be a brain in a vat or the victim of a malevolent Cartesian demon who delights in deception; I may be wholly and hopelessly deceived. Even so, I can still do my epistemic duty; I can still do my best; I can be above reproach. Justification (unlike, say, a strong constitution) is not something that happens to a person; it is instead a result of her own efforts. Perhaps I can't take credit for my good digestion or my charming disposition; I can take credit for being justified. As the classical deontologist sees things, justification is not by faith but by works; and whether we are justified in our beliefs is up to us.

It is really this deontological feature of the classical conception of justification that leads to the internalist result; but to see just how, we must make a brief detour through a steep and thorny area of ethics. Most of us will agree that a person is guilty, properly blamed, properly subject to censure and moral disapproval, if and only if she fails to do her duty (where among her duties might be that of refraining from doing certain things). So

(a) you are properly blamed for failing to do something A if and only if it is your duty to do A (and you fail to do it).

Of course we also think that someone who has done no more than what she nonculpably thinks duty permits or requires, is not culpable or guilty in doing what she does, even if we think that what she has done is wrong. You are the governor and it is up to you to decide whether a certain prisoner is to suffer the death penalty. You reflect as carefully and impartially as you can and make your decision; perhaps you believe that it is your duty in the circumstances not to commute the death sentence and let the law take its course. Then I shall not properly hold you blameworthy or guilty for doing what you do, even if I think you made the wrong decision. You cannot be faulted for doing what you think is the right thing to do—provided, of course, that you came to that judgment in a nonculpable way. (If you formed the judgment out of vengefulness, or pride, or lordly contempt for those whom you take to be your inferiors, then things are very different.) So we also have

(b) If a person nonculpably believes that doing A is morally required or permitted, then she is not guilty (not to be blamed) for doing A; and if she nonculpably believes that refraining from doing A is morally required or permitted, then she is not guilty (not to be blamed) for refraining from doing A.

It is plausible to add, still further, that if I believe that it is my duty, all things considered, to do A, then I am guilty, culpable, morally blameworthy if I do not do A.32

Sadly enough, however, these principles taken together appear to lead to trouble. For suppose I nonculpably think I am permitted to refrain from doing A. Then by (b) I am not guilty and not to be blamed for thus refraining; but then by (a) doing A is not my duty. So if I nonculpably think it is not my duty to do something A, then it is not my duty to do A; and if I nonculpably think it is not my duty to refrain from doing A, then it is not my duty to refrain from doing it. Furthermore (given the addition to (b)), we can argue similarly that if I think it is my duty to do A, then I am culpable if I do not do A, in which case it is my duty to do A. But isn't this consequence wrong? You and I might argue at considerable and heated length about what duty requires in a given set of circumstances. Perhaps I think you ought to commute that sentence; you think the right thing to do is to let it stand. You could not sensibly claim that since you do in fact believe that is your duty, and believe that nonculpably, you automatically win the argument. It isn't given in advance that I am always right about what my duty requires, so long as I am nonculpable in holding the opinion I hold. If that were so, why should I come to you, asking for advice as to what my duty really is, in a given situation? So (a) and (b) both seem correct; taken together, however, they seem to entail a proposition that is clearly false.

Here, as Aquinas says, we must make a distinction. An attractive way out of this quandary is offered by the distinction between objective and subjective duty or rightness. You are guilty or blameworthy if you fail to do your subjective duty, but not necessarily guilty for failing to do your objective duty. Guilt, being properly blamed, being properly subject to censure—these things go with violation of subjective duty. Perhaps my objective duties are constituted by virtue of their being, of the options open to me, the ones that contribute most to the greatest good; or perhaps they are constituted by God's commands; or perhaps they are the ones that bear a certain particular relation of fittingness to the circumstances. Then a person might well not know or be able to see that a given action was the right one, the dutiful one, in the circumstances. Perhaps I suffer from a certain sort of moral blindness; I simply cannot see that I have an obligation to care for my aging parents. Then I am not blameworthy for failing to care for them, unless my moral blindness itself somehow arises from dereliction of duty. Assume, just for purposes of argument, that the ground of the obligation not to steal is the divine command “Thou shalt not steal.” I could hardly be blamed for stealing if I (nonculpably) didn't know that stealing is wrong or didn't know, of a given act of stealing I am performing, that it is wrong, or didn't know, of a given act of taking something, that it is indeed an act of stealing. You are guilty, or to blame, or properly subject to censure only if as we say, you knowingly flout your duty. Ignorance may be no excuse in the law; but nonculpable ignorance is an excusing condition in morality. Indeed, it is sometimes also an excusing condition in the law; according to the M'Naghten Rule you aren't legally culpable if you cannot tell right from wrong.

Now how, exactly, does this help with respect to the above quandary? Well, the problem was that (a) and (b) seemed to entail that I couldn't make a nonculpable mistake about what my duty was; but that seemed wrong, since it is perfectly sensible for you to challenge my belief as to what duty requires, even if you don't for a moment believe that I arrived at that belief culpably. And the resolution is that while I can't make a nonculpable error about my subjective duty, the same does not hold for my objective duty; but what we dispute about, when we dispute about what my duty, in a given circumstance, is, is not my subjective but my objective duty. It is easy enough, in the right circumstances, to make a mistake about that.33

Given that no one is guilty for doing what she nonculpably believes is right, you might expect that we would ordinarily be receptive to the claim of ignorance as an excusing condition. The fact is, however, that in many circumstances we are extremely reluctant to accept such a claim. I take part in a racist lynching: you will not be impressed by my claim that, after careful reflection, I considered that the right thing to do. We are deeply suspicious of such claims. We are not ordinarily receptive to the claim, on the part of murderer or thief, that, after due consideration, she thought the course she took most appropriate, morally speaking, of those open to her.

The reason, I think, is that there are many moral beliefs we don't think a properly functioning human being can (in ordinary circumstances) nonculpably acquire. We do not think a well-formed, properly functioning human being could honestly arrive at the view that it does not matter how one treats his fellows, that if inflicting suffering on someone else affords me a certain mild pleasure, then there can be no serious objection to my so doing. We do not think a person could honestly come to the view that all that matters is his own welfare and pleasure, other persons being of value only insofar as they contribute to that end. It is not, of course, that we think it logically impossible (in the broad sense) that there be persons who honestly arrive at such views; it is rather that we think it simply would not, more exactly, could not happen, given ordinary circumstances and what is in fact the nature of human beings. A theist will be likely to view this as a matter of God's having created us in such a way that we can simply see that heinous actions are indeed heinous; nontheists will account for the same fact in some other way.

In either case, however, we are likely to think that if a cognitively nondefective person comes to believe that such actions are perfectly right and proper, it must be because of some fault in him. Perhaps at some time in the past he decided to accept these views, and the pressure of that commitment has brought it about that these beliefs are now second nature to him. A part of what is involved in our blaming people for holding corrupt beliefs, I think, is our supposing that a human being whose faculties are functioning properly, and who is blameless in forming and holding her beliefs, will reject these beliefs, just as we think a human being whose faculties are functioning properly will accept modus ponens as valid but fail to pay the same compliment to Affirming the Consequent. We think a properly functioning human being will find injustice—the sort depicted, for example, in the story the prophet Nathan told King David34—despicable and odious. Only a cognitively defective person could conscientiously come to think that the behavior of the rich man in that story is anything but morally abhorrent; any normal adult who gives the matter a moment's thought can see that injustice of that sort is wicked and reprehensible. (Indeed, we need not limit ourselves to adults: small children often exhibit a very well developed sense of justice and fairness.) In the face of this natural tendency or prompting, to accept such behavior as perfectly proper requires something like a special act of will—a special act of ill will. Such a person, we think, knows better, or at any rate should have known better; she chooses what in some sense she knows to be wrong. And if, on the other hand, we think a person really does lack this inclination to see these actions as morally wrong, then we think he is in some way defective, that some of his cognitive faculties are not functioning properly.

We therefore object, from a moral point of view, to certain kinds of actions that (in the short term anyway) are entirely conscientious; we hold that a person may be doing what is wrong or wicked or impermissible even if he thinks that way of acting is quite in accord with his obligations—even if, indeed, he thinks the action in question is part of his duty. Our objection here is that he ought not to think that action permissible or obligatory; and the fact that he does think so shows (if his cognitive faculties are functioning properly) that at some point he has culpably done something that has clouded his own moral vision. We think those whom we thus hold responsible really know better. They have rejected what is plain to anyone of goodwill. They have ignored or suppressed the promptings and leadings of nature—the natural tendency to find unjust behavior reprehensible, for example—and have instead chosen a different route, perhaps one that legitimates a desire for self-aggrandizement, one that gives free rein to that perverse and aboriginal sin, pride. So there is a link, here, between objective and subjective duty, a link we think provided by our nature. This link is constituted by the fact that—so, at any rate, we are inclined to think—a certain kind of mistake is not possible for a well-formed human being who is acting (and has acted) conscientiously.

So for a large and important class of cases we think objective and subjective duty coincide, and do so because of our cognitive constitution; there is a large class of cases in which a properly functioning human being can just see (all else being equal) that a certain course of action is wrong. Now it is this same thought—the thought that in a large class of cases objective and subjective duty coincide—that underlies classical internalism. This coincidence of objective and subjective duty is the driving force behind the classical internalism of Descartes and Locke. We can see this in more detail as follows.

A. The First Internalist Motif

According to Locke and Descartes, epistemic justification is deontological justification. And here they are clearly thinking of subjective duty or obligation; they are thinking of guilt and innocence, blame and blamelessness. If I do not have certainty but believe anyway, says Descartes, “I do not escape the blame of misusing my freedom.” Locke, clearly enough, is also thinking of subjective duty (“This at least is certain, that he must be accountable for whatever mistakes he runs into”). But then the first internalist motif follows immediately:

M1. Epistemic justification (that is, subjective epistemic justification, being such that I am not blameworthy) is entirely up to me and within my power.

All that is required is that I do my subjective duty, act in such a way that I am blameless. All I have to do is my duty; and, given that ought implies can, I am guaranteed to be able to do that. So justification is entirely within my power; whether or not my beliefs are justified is up to me, within my control. My system of beliefs may be wildly skewed and laughably far from the truth; I may be a brain in a vat or a victim of a malicious Cartesian demon; but whether my beliefs have justification is still up to me.

B. The Second Internalist Motif

Descartes and Locke, I say, are speaking there of subjective duty. But of course they are also speaking of objective duty. Descartes claims that it is clear to us that we must not give assent to what is uncertain: “the light of nature,” he says, “teaches us that the knowledge of the understanding should always precede the determination of the will.” And Locke holds that it is my duty to regulate my belief in such a way that I believe only what I have good reasons for, that is, only what is epistemically probable with respect to my total evidence.35 One who does otherwise, he says, “transgresses against his own light, and misuses those faculties, which were given him.” Such a person, he says, “neither seeks truth as he ought, nor pays the obedience due his maker, who would have him use those discerning faculties he has given him, to keep him out of mistake and error.” To regulate my belief in this way is my objective duty; what makes an act of believing permissible or right is its being appropriately supported by the believer's total evidence. But Locke also holds that this is my subjective duty; if I do not regulate my belief in this way I am blameworthy, guilty of dereliction of epistemic duty. (Merely trying to regulate it thus is not sufficient; I must succeed in so doing if I am not to be blameworthy.) Objective and subjective duty thus coincide. Similarly for Descartes: if you give assent to what is not certain then (ceteris paribus) you are blameworthy, have flouted subjective duty as well as objective duty. So the second internalist motif:

M2. For a large, important, and basic class of objective epistemic duties, objective and subjective duty coincide; what you objectively ought to do matches that which is such that if you don't do it, you are guilty and blameworthy.

And the link is provided by our nature: in a large and important class of cases, a properly functioning human being can simply see whether a given belief is or is not (objectively) justified for him. (In the same way, in the more general moral case, certain heinous acts, so we think, are such that a properly functioning human being cannot make a nonculpable mistake as to whether those acts are morally acceptable.)

The second internalist motif has three corollaries.

First: if it is your subjective duty to regulate your belief in this way, then you must be able to see or tell that regulating belief this way is indeed your duty. Locke and Descartes clearly hold that a dutiful, conscientious person whose cognitive faculties are functioning properly will not make a mistake as to what is the right method or practice for regulating belief. Descartes claims that it is clear to us that we must not give assent to what is uncertain: “the light of nature,” he says, “teaches us that the knowledge of the understanding should always precede the determination of the will.” And Locke says that the person who does not regulate his belief according to the evidence “transgresses against his own light, and misuses those faculties, which were given him” (my emphasis). So the first corollary:

C1 In a large and important set of cases, a properly functioning human being can simply see (cannot make a nonculpable mistake about) what objective epistemic duty requires.

To grasp the second corollary, we must note first that (according to both Descartes and Locke) I do not determine directly, so to speak, what it is that I am obliged to believe and withhold. According to Locke, I determine whether a given belief is acceptable for me or justified for me by determining something else: whether it is supported by what is certain for me—whether, that is, it is probable with respect to what I know. Similarly for Descartes: I do not directly determine whether a proposition is acceptable or justified for me; I do it by determining whether or not it is clear and distinct for me. So I have a way of determining when a belief is justified for me; to use a medieval expression, I have a ratio cognoscendi for whether a belief is justified for me. As we have seen, Descartes and Locke think that a well-formed human being cannot (in those basic cases) make a conscientious error as to whether a given belief is justified for her; but then, in those cases, she will also be unable to make a conscientious mistake about whether a given belief has the property by which she determines whether that belief is justified for her. Locke and Descartes therefore believe that a well-formed, conscientious human being will (at least in that large and important basic class of cases) be able to tell whether a given belief has the property that forms the ratio cognoscendi for justification. So the second corollary:

C2 In a large and important class of cases a properly functioning human being can simply see (cannot make a nonculpable mistake about) whether a proposition has the property by means of which she tells whether a proposition is justified for her.

As we have just seen, Locke and Descartes hold that I have a means of telling whether a given proposition is justified for me; I do it by determining whether it is supported by my total evidence (Locke) or whether it is certain for me (Descartes). But note that what confers justification on a belief for me, the ground of its justification, is, as they see it, the very same property as that by which I determine whether it is justified for me. According to Locke, the ratio essendi (to invoke the other half of that medieval contrast) of justification is the property of being supported by the believer's total evidence, whereas according to Descartes it is the property of being certain for the believer. But then the ground of justification (the justification-making property) is identical with the property by which we determine whether a belief has justification: ratio cognoscendi coincides with ratio essendi.36 (This is not, of course, inevitable; in the case of measles, velocity, blood pressure, weight, and serum cholesterol our ratio cognoscendi does not coincide with the ratio essendi.)

If so, however, then there is another kind of error a properly functioning dutiful human being cannot make; such a person is so constructed that (in that class of basic cases) she cannot conscientiously come to believe, of the justification-making property, that a given belief has it when in fact it does not. According to Locke, a properly functioning human being could not both be appropriately dutiful in forming her beliefs (in these cases), and also mistakenly believe, of some proposition, that it was supported by her total evidence; according to Descartes, such a person in such a case could not mistakenly come to think that a belief was certain for her when in fact it was not. We have a certain guaranteed access to the ratio cognoscendi of justification; but if ratio cognoscendi and ratio essendi coincide, then we also have guaranteed access to the latter. So the third corollary:

C3 In a large, important and basic class of epistemic cases a properly functioning human being can simply see (cannot make a nonculpable mistake about) whether a proposition has the property that confers justification upon it for her.

Now the fact of the matter seems to be, contra Locke, that cases in which it is obvious what my total evidence supports are, after all, relatively few and far between. It is easy enough to make a nonculpable mistake about what my total evidence supports; it is often difficult indeed to tell whether a belief has (what Locke sees as) the ratio cognoscendi of justification. Perhaps Locke sometimes saw this; significantly enough, he sometimes retreats to the weaker view that what confers justification is the belief's being such that upon reflection I think it is supported by my evidence. Here it seems clear that I do have the requisite special access.

C. The Third Internalist Motif

There is still another and somewhat less well defined internalist motif here. According to Locke and Descartes, I have a sort of guaranteed access to whether a belief is justified for me and also to what makes it justified for me: I cannot (if I suffer from no cognitive deficiency) nonculpably but mistakenly believe that a belief is justified or has the justification-making property. This is the source of another internalist motif; for it is only certain of my states and properties to which it is at all plausible to think that I have that sort of access. Clearly you don't have this sort of access to the pH level of your blood, or the size of your liver, or whether your pancreas is now functioning properly. The sorts of things about which it is plausible to hold that you can't make a mistake, will be, for example, whether you believe that Albuquerque is in New Mexico, whether you are now being appeared to redly, whether you are trying to get to Boston on time, or whether you are trying to bring it about that, for every proposition you consider, you believe it if and only if it is true. So the justification-making property will have to attach to such states as my believing thus and so, my being appeared to in such and such a fashion, my aiming at a given state of affairs, my trying to do something or other, and the like. These states are the ones such that it is plausible to hold of them that I cannot make a nonculpable mistake as to whether I exhibit them. But they are also, in some recognizable, if hard to define sense, internal to me—internal to me as a knower or a cognizing being. Thinking of justification in the deontological way characteristic of classical internalism induces epistemic internalism: and that in turn induces internalism of this different but related sort. It is not easy to think of a name for internalism of this sort, but perhaps the name ‘personal internalism’ (calling attention to the way in which my beliefs, desires, experience, and aims are crucial to me as a person) is no worse than some others.

Of course it is not necessary that the things to which a person has this special access are internal in this sense; there could be a being who had guaranteed and indeed logically incorrigible access to properties that were not in this way internal to him. If the bulk of the theistic tradition is right, God is essentially omniscient: but then it is impossible (impossible in the broadly logical sense) that he err on any topic whatever, internal to him or not. Not so for us.

One final point. It is clear that Descartes and Locke each embrace a version of doxastic voluntarism: the view that we have at least some voluntary control over our beliefs. Of course both hold that I cannot withhold assent from what is genuinely self-evident; thus, according to Locke (as we saw), “all such affirmations, and negations, are made without any possibility of doubt, uncertainty, or hesitation, and must necessarily be assented to, as soon as understood” (Essay IV, vii, 4, p. 269); and Descartes would concur. But Descartes thought there were many propositions—propositions that are not initially self-evident for me—such that it is within my power to accept them and within my power to reject them; and it is my duty to accept them only if they are or become certain for me. (Of course, he recognized that it might take a certain discipline to be able to reject all beliefs that are less than certain; perhaps one must reflect on the fact that the senses often deceive us, or on the dream argument, or on the hypothesis that we might be in the doxastic control of a malicious demon. In fact Descartes suggests somewhere that it might be a good idea to reread the first couple of his Meditations when tempted to believe what is less than indubitable; in the same spirit Victorians encouraged youths tempted to idleness to read inspirational poetry extolling industry.)37 Similarly for Locke: he held that it is up to me whether I deliberate, although once I have gone through this process and see that a given belief is strongly supported by what is certain for me, then it is no longer up to me whether I accept it. But many beliefs, he thought, are such that prior to reflection I have a strong inclination to accept them, which inclination is resistible.38 It is possible for me to accept them and possible for me to withhold them; and he held that the latter is what duty requires. It is my duty to withhold assent unless the candidate for belief is supported by what is certain for me. So should we add doxastic voluntarism as a fourth internalist motif?

I don't think so. First, note that in the typical case where a belief suggests itself—even a belief that is not certain—it is not the case that it is within my power to accept it and within my power to reject it. Driving down the road I am confronted with what appears to be an approaching automobile; it is ordinarily not, in such a case, up to me whether I believe that there is an automobile approaching. (And, pace Descartes, it doesn't help much to reread the first couple of his Meditations.) When I see a tree, or the sky, it is ordinarily not within my power to withhold such propositions as there's a tree there or today the sky is blue. You offer me a million dollars to believe that the population of the United States exceeds that of China; I can try my hardest, strain to the uttermost; it will be in vain. But does epistemic deontologism as such imply the denials of these facts? Not at all. Perhaps Descartes’ and Locke's views do; and perhaps epistemic deontologism implies a doxastic voluntarism of some sort, a sort of weak doxastic voluntarism; but it does not as such imply that there are any beliefs at all such that merely by an act of will I can either acquire or lose or withhold them. One who construes justification deontologically need not believe that it is ever my duty, in given circumstances, to believe or withhold a given proposition. Perhaps instead my duty is to follow Locke's advice and reflect on the question whether the belief in question is supported by my evidence; perhaps it is my duty to adopt or strive to adopt policies of a certain sort. It is within my power to adopt policies that influence and modify my propensities to believe; I can adopt such policies as paying careful attention to the evidence, avoiding wishful thinking, being aware of such sources of belief as jealousy, lust, contrariety, excessive optimism, loyalty, and the like. Perhaps it is my duty to cultivate the epistemic or doxastic virtue Ernest Sosa speaks of; and perhaps a belief of mine is justified in case it is epistemically virtuous, the sort of belief that an epistemically virtuous person would hold in that context. Of course, even this implies at least a certain degree of indirect control over my beliefs. Suppose I am at present epistemically vicious and it is my duty to achieve intellectual virtue. Then it is within my power to achieve this state; but the beliefs I would hold were I epistemically virtuous would presumably be different from the beliefs I would hold if I persisted in my epistemically vicious condition; so I have at any rate a certain indirect control over what I believe. But this control may be as indirect and tenuous as you please.

Internalism, therefore, is a congeries of ideas loosely connected and analogically related. If we go back to the source of the internalist tradition, however, we can see that internalism arises out of deontology; a deontological conception of warrant (or, as in Locke's case, of an aspect of warrant) leads directly to internalism. Seeing internalism as thus arising out of deontology permits us to see why it includes just the elements it does, and how those elements are related.

V. Back to the Present

Suppose we return to the twentieth century; we are now in a better position, I think to understand the swirling diversity it presents with respect to justification. According to the twentieth-century received tradition, as we saw above, (1) justification is necessary and (along with truth) nearly sufficient for knowledge, (2) there is a strong connection between justification and evidence, and (3) justification involves internalism of two kinds (epistemic and personal internalism). Further, justification itself is taken as a matter of epistemic responsibility or aptness for epistemic duty fulfillment (Firth, Lehrer, Cohen, Kornblith, Chisholm), as an “evaluation” of how well you have fulfilled your epistemic goals (Lehrer and Cohen), as being believed or accepted on the basis of an adequate truth conducive ground (Alston), as being produced by a reliable belief-producing mechanism (Goldman), as being supported by or fitting the evidence (Conee, Conee and Feldman, Firth, many others), and as a matter of things going right with respect to the knowing subject qua subject. The project was to try to understand this diversity, and to see what underlies the close connection of justification with knowledge, the internalist requirement laid upon epistemic justification, and the stress upon evidence in connection with justification. I think it is now easier to understand this diversity.

By way of pursuing this project, note first that the basic Cartesian—Lockean idea of justification as fulfillment of epistemic duty or obligation is, of course, directly reflected in the work of all of those (for example, BonJour, Cohen, Kornblith, and, preeminently, Chisholm) who see justification as epistemic responsibility or aptness for epistemic duty fulfillment. This deontological conception of epistemic justification is the basic and fundamental notion of epistemic justification; other notions of epistemic justification arise from this one by way of analogical extension. Deontological justification is justification most properly so-called.39 (In the next chapter I shall examine Chisholm's detailed attempt to explain warrant in terms of justification thus construed.)

Turn now to the second notion of the nature of justification: that it is or essentially involves having adequate evidence for the belief in question. We often say that a belief is justified when the believer has what we think of as sufficient evidence or reason for the belief, or (perhaps more exactly) that under those conditions the believer is justified in holding that belief. According to the ‘evidentialism’ of Conee and Feldman, you are justified in believing B just if you have sufficient evidence for it, or (as they put it) just if it fits your evidence. (Thus Conee: “Such examples make it reasonable to conclude that there is epistemic justification of a belief only where the person has cognitive access to evidence that supports the belief.”)40 Indeed, this equation of being justified with having evidence is so pervasive that the justified true belief analysis of knowledge has often been put as the idea that you know if and only if your belief is true and you have adequate evidence for it.41 How does this association of justification with evidence come about? The answer is easy, given the central historical position of Descartes and, more particularly, Locke. To be justified is to be without blame, to be within your rights, to have done no more than what is permitted, to have violated no duty or obligation, to warrant no blame or censure. Now suppose you agree with Locke that among your duties is that of not giving “firm assent” to any proposition (any proposition that is not certain for you, that is, self-evident or appropriately about your own mental life) without having good reasons, that is, evidence for it. Then, naturally enough, you will think that no one is justified in accepting such a belief without evidence or reason. Indeed, you may eventually go so far as to use the term ‘justified belief’ just to mean ‘belief for which one has good reasons’.

Two further points here. (a) Conee and Feldman do not make the deontological connection: they do not say that the ratio essendi of justification is duty fulfillment, with the chief duty being that of believing (or, more plausibly, trying to bring it about that you believe) only that which fits your evidence. But there are plenty of contemporaries and near contemporaries who do; W. K. Clifford (that “delicious enfant terrible,” as William James calls him) trumpets that “it is wrong, always, everywhere, and for anyone to believe anything upon insufficient evidence”;42 his is only the most strident in a vast chorus of voices insisting that the or a primary intellectual duty is that of believing only on the basis of evidence. (A few others in the choir: Sigmund Freud,43 Brand Blanshard,44 H. H. Price,45 Bertrand Russell,46 and Michael Scriven.47 And (b) there are two quite different possibilities for the evidentialist; she might be holding, on the one hand, that the very nature of justification is believing (or trying to bring it about that you believe) on the basis of evidence (that justification just is believing or trying to believe in that way) or she might hold, more plausibly, that the nature of justification is fulfillment of epistemic duty, the chief among those duties being that of believing or trying to believe only on the basis of evidence. (Since Conee and Feldman do not mention epistemic obligation, it seems likely that they are to be taken the first way.)

Turn next to broad justification, the idea that things are going right from the perspective of the epistemic agent, or with respect to the epistemic subject as subject, or with respect to everything downstream of experience. This is an easy analogical extension of the deontological notion of justification. On that deontological notion, your experience might be wholly deceptive (you might be a brain in a vat, or the victim of a Cartesian demon); nevertheless, you are justified if you do your epistemic duty and form the beliefs your circumstances (and duties) require. Here the notion of things going right with respect to the epistemic subject as subject is just the notion of doing one's epistemic duty, being epistemically responsible. But it is an easy analogical step to the idea of things going aright with respect to other conditions relevantly internal to the knower and hence to the idea of everything's going right from the perspective of the epistemic subject or with respect to the epistemic subject as subject.

Lehrer and Cohen speak of epistemic justification as an evaluation of how well you have accomplished your epistemic goals. Here the idea is not that you have duties or obligations; it is rather that you have or may have epistemic goals: and you are justified to the degree that your epistemic behavior is a good way of attaining those goals. And here the word ‘rationality’ might be more appropriate than ‘justification’. What is really at issue is Zweckrationalität, means—end rationality, appropriateness of your means to your goals. This notion is similar to Richard Foley's conception of epistemic rationality, powerfully expounded in The Theory of Epistemic Rationality.48 Lehrer and Cohen's notion is not directly connected with the classical deontological conception of justification; however, it does have a sort of indirect connection. If you become doubtful that there are any specifically epistemic duties, or perhaps think there are some, but doubt that fulfilling them can play a large role in the formation and governance of belief, then this notion of means—ends rationality may seem an attractive substitute. Perhaps there is no such thing as epistemic duty; even so, however, there is such a thing as pursuing your epistemic goals well or badly, and it is not unnatural, given your doubts about the former, to think of justification in terms of the latter.

Finally, there is the conception of justification to be found in both the old and the new Goldman. According to the old Goldman (to a first approximation), a belief is justified if and only if it is produced by a reliable belief producing process or mechanism (see pp. 197–199). According to the new, a belief is justified if it is the last item in a cognitive process that is licensed by a right set of J rules; and a set of J rules is right in case it has a high truth ratio in nearby possible worlds.49 Here I think there is little connection with the classical notion of justification as involving fulfillment of epistemic duty. True, in the later Goldman there is the notion of a rule, and of a process permitted by a rule. But rules of this sort have nothing to do with duty or obligation; there is nothing deontological about them. Goldman's use of the term, I think, is to be understood another way: suppose you use the term ‘justification’ as no more than a name for what is necessary for knowledge and (together with truth) sufficient for it up to Gettier problems; and suppose you also think, with Goldman, that fulfillment of epistemic duty, no matter how fervent and conscientious, is substantially irrelevant to knowledge. Then you might find yourself using the term in just the way he uses it. Here there is only a fairly distant analogical connection with the classical conception.50

So much for the main contemporary conceptions of justification; they can all be understood, I think, in terms of their relation to the classical deontological conception. But the same can be said for the contemporary connection between justification and internalism. According to Conee, “Justifying evidence must be internally available”; his idea is that the evidence in question cannot be evidence you could get from your world almanac, for example, but must rather be evidence you can come up with just by reflection. Alston “find[s] widely shared and strong intuitions in favor of some kind of accessibility requirement for justification.” Here there is a reflection of the classical connection between deontological justification and epistemic internalism; it is the influence of classical deontologism that is responsible for those widely shared intuitions. Internalism in the personal sense is also widespread (just as we should expect, given the relation between internalism in the two senses). Thus Lehrer and Cohen argue that reliabilism must be wrong about justification: “The truth of the demon hypothesis (where my beliefs are mostly false) also entails that our experiences and our reasonings are just what they would be if our cognitive processes were reliable, and therefore, that we would be just as well justified in believing what we do if the demon hypothesis were true as if it were false” (see p. 9). Here the idea, clearly, is that only what is internal to me as a knower in the personal sense, in the way in which my beliefs and experiences are, is relevant to justification. This may be understood, I think, as a reflection of the connection between deontological justification and internalism.

Now classical deontological internalism has a certain deep integrity. The central notion is that we have epistemic duties or obligations; this induces internalism of both the epistemic and the nonepistemic sorts; and the central duty, Locke thinks, is to believe a proposition that is not certain only on the basis of evidence. Chisholm's carefully crafted internalism, as we shall see, exhibits all of these features, except that according to Chisholm the central epistemic duty is to try to achieve epistemic excellence. Other contemporary accounts, however, sometimes seize on one or another of the elements of the classical package, often in such a way that the integrity of the original package is lost, or at least no longer clearly visible. (I argue this point with respect to Alston in chapter 9 below.) For another example, consider Conee and Feldman, who see justification as a matter of having adequate evidence, and hold that this evidence must be internally available to the believer. This makes sense if combined, as in Locke, with the idea that justification is fundamentally a deontological matter of duty fulfillment. They say nothing about the latter, however, which leaves the internalism unmotivated and the connection between it and evidentialism obscure.

Lehrer and Cohen, to take still another example, speak of justification as “an evaluation of how well you have pursued your epistemic goals.” The internalism they display fits at best dubiously with this conception of justification. Suppose justification is an evaluation of how well you are pursuing your epistemic goals; it is then presumably an evaluation of the appropriateness of the means you use to the goals you choose. Suppose your doxastic goal is, for example, believing truth, or attaining salvation, or achieving fame and fortune: why would there be any necessity that you be able to tell, just by reflection, let's say, how well suited your means are for achieving those goals? And why think that only what pertains in a direct way to your experiences and beliefs is relevant to this question of how well those means fit those goals? What reason is there to think that an evaluation of how well you were pursuing your epistemic goals would have to measure something such that only your beliefs and your experiences would be relevant to it? The internalism of the classical conception lingers, but its root and foundation are no longer present.

By way of conclusion then: justification, internalism, and epistemic deontology are properly seen as a closely related triumvirate: internalism flows from deontology and is unmotivated without it, and justification is at bottom and originally a deontological notion. Given a proper understanding of these three, furthermore, we can gain some understanding of the kaleidoscopic variety of contemporary thought about justification and its connection with warrant. (As a bonus, we are also able to understand the connection between justification and evidence.) The most interesting question, of course, is whether the twentieth-century received tradition is correct here; can warrant (apart, perhaps, from a fillip to mollify Gettier) be explained in terms of justification? It is time to consider that question, and there is no better way to do so than by turning to the work of Roderick Chisholm, the subject of the next chapter.