S. Joel Garver


The Problem of Epistemic Obligations

Section I

Sometimes people believe things that they ought not to believe. Sometimes they merely fail to believe what they should believe or to know what they ought to know. Consider the following scenario:

Patrick hopped out his just parked car and hurried up the driveway, noting the time with some alarm. The back door was open and as he entered his parents came into the kitchen, their faces mixed with expressions of worry, relief, and anger.

"Patrick! Where have you been? You said you would be home around's nearly midnight! Your father and I were so worried."

"I lost track of the time," Patrick said, though that wasn't exactly true. He hadn't checked his watch until he started for home, but he had known it was getting late. He didn't think it was that important. His father motioned him to a chair.

"Why didn't you call?"

"It didn't occur to me. I was..."

"It 'didn't occur' to you? You should have realized we were worried."

"I was busy."

"That's no excuse. You should have known to call...At any rate, you're grounded. No car for a week. Now get to bed."

Patrick's father makes several claims of interest: first, Patrick ought to have realized that his parents were worried, and second, Patrick ought to have known to call them. Moreover, Patrick falsely thought the lateness of the hour to be unimportant and should have known better. It is of some interest that such prescriptive language in this context does not attach simply to certain actions (like calling home), but rather to doxastic and epistemic phenomena: knowing, believing, and realizing. As the case of Patrick exemplifies, and as we shall further see, cases of this kind regularly center around beliefs and knowledge that are closely connected with evaluation and the accomplishment of practical ends, though they are not limited to such cases. By realizing his parents' worry Patrick is expected not only to gain the factual knowledge that his parents are worried and that he is the cause, but also the evaluative knowledge that it is his responsibility to do something to alleviate that worry. Similarly, knowing to call is knowledge that he ought to call and knowing the time should have been coupled with a realization of its practical importance.

Examples of this sort can be easily multiplied. The man who beats his wife ought to realize that this is vicious and he may wrongly believe that violence can facilitate problem-solving. According to many psychologists and social workers the beliefs of such a man are fostered by numerous social factors--cultural stereotypes, economic and legal inequities, and so on--as well as his failure to deal adequately with his own insecurity, inability to communicate, dominating personality, and dependency (e.g. cf., Hoeffler 1983; Martin 1981:25-71).

Another example. The person who does not believe in God really ought to believe according to Christian understanding and, moreover, she is responsible for her unbelief. In terms of the teachings of St. Paul in Romans 1 and 1 Corinthians 2, her unbelief might be explained as a possession of knowledge of God's existence and an awareness of what He is like coupled with a simultaneous suppression of same knowledge, expressed (at least in the case we are considering) as unbelief--or even false belief. According to St. Paul such unbelief is a form of affectively charged yet culpable blindness to the clear revelation of God in the created world, his image in her own personhood, and her conscience which reveals His will (Ro 1:18-2:16; 1Co 2:6-16; for some exegesis of these passages see Gaffin 1995; or on your own see Ro 1:19, 21 and 28 and 1Co 2:8, 11, and 14; note the use of words in the Greek text: "apokalupto," "gnosis," "ginosko," "oida," and "dexomai").

Such responsibility for belief is not, however, limited simply to cases in which a subject has believed wrongly, and thus may be to blame, but also extends to right beliefs for which a subject may be praised. Consider those, who through study and training, are able to form various sort of beliefs or to acquire certain kinds of knowledge: that T.S. Eliot is using sacramental imagery, that the shadow in the X-ray is not cancerous, that there's a fatal flaw in the theory that quantum fields interact through virtual particles, and the like. We regularly praise people who master areas of knowledge for their insight, their creativity, their research, their vision--assuming, therefore, their responsibility for that knowledge.

Furthermore, part of leading a virtuous life requires pursuing and deepening certain sorts of knowledge, not in the accomplishment of a set number of epistemic tasks, but in being characterized by traits or by shaping one's epistemic life in a particular way. A healthy marriage is characterized by a deepening understanding and knowledge of each partner by the other; and while some of this knowledge may be practical or personal, it is not without significant propositional content. In Christianity, the believer is urged to grow in her knowledge of God the Father and of Christ, to continually "put on" Christ who is the wisdom of God, and to renew her mind (Ro 12:2; 13:14; Eph 4:22-24; Col 1:9-10; 3:10; 2Pe 3:18). It is also apparent, then, that people who are characterized by their ability to correctly assess situations and arrive at the proper evaluations are possessed of a virtue that includes certain virtues of the intellect. The opposite sort of people are not merely insensitive, but possess an intellectual vice.

The first set of examples--involving epistemic blame--betray a common phenomenon which can be seen as culpable epistemic blindness and, using terminology drawn from the Christian tradition, manifests sin which requires nothing short of a "change of heart" or conversion (metanoia). The second set of examples--involving epistemic praise and achievement--also suggests that epistemic virtue requires certain "attitudes of the heart" or deep-seated affections, desires, and identifications.

Christianity provides an important insight here into the nature of our epistemic lives: that there are deep and important connections between the way we believe, what we believe and know, and who we are, that is, the state of our "hearts"--the depths of our selves from which arise various aspects of our lives: emotional, volitional, imaginative, dispositional, rational, psychological, creative, or spiritual (Pr 27:19; Is 51:7; 29:13; Jer 17:9; Lk 8:12, 15; Acts 2:37; 5:33; this emphasis was largely the suggestion of Mark Talbot). Though I address this issue in more detail in later chapters, I will note here that this approach may provide, I think, a necessary corrective to epistemological models which take "There is a book on the table before me" or "There is a tree outside my window" as paradigmatic for discussions of justification and knowledge (for a different critical approach to these paradigms see Lorraine Code's essay in Alcoff and Potter 1993:15-48). Whether any part of our epistemic lives are unaffected by issues of the heart will concern us throughout the following chapters.

From our everyday speech and from the Christian tradition it is apparent, then, that people are held responsible for much of what they believe and know. Furthermore, there are, at least in a wide range of cases, obligations associated with believing, knowing, realizing, being aware, and so on. There are several important questions that must be addressed at this point. First, though we have spoken indiscriminately concerning belief and knowledge (and being aware, realizing, etc.), is there a difference between these categories that is significant in respect to the prescriptive and normative language that we have been using? Second, is this language of obligation in respect to these epistemic categories truly prescriptive (or deontological) or is it merely normative in some weaker sense? Third, if our language is truly prescriptive, are these requirements, prohibitions, and the like best thought of in moral terms or in some nonmoral manner? Allow me to address these concerns in turn, since they are important for clarity in the following chapters.

Let us consider the first question. It would seem that the differences between belief and knowledge are quite important for a discussion of what obligations might attach to them. It is more natural, in most contexts, to say "You ought to have known" or "You should realize" than it is to say "You ought not to have known" and "You shouldn't realize." On the other hand, it is perfectly natural to say "You ought not to believe that p" or "You shouldn't believe that p." It is probably less common to say "You ought to believe." But there is nothing odd about saying "You ought to realize" or "You ought to have known." Thus the connection of obligation and responsibility with belief is asymmetrical to their connection with knowledge. Furthermore, it makes some sense to consider a subject believing something at will, but it makes far less sense to consider a subject knowing something at will. Thus direct voluntary control and responsibility is related to belief in a different way than to knowledge. (These points were suggested to me by Bill Alston in correspondence.)

In response to this line of thought let us begin by considering some of the differences between belief and knowledge. Sometimes "belief" and "knowledge" are used to contrast two mutually exclusive categories: "What do you mean you believe she came home? Don't you know?" This use of the terms, however, is limited to certain special contexts. In philosophical contexts, at least, knowledge is not taken to exclude belief (or vice versa) and, in fact, the view that knowledge essentially involves belief (e.g., knowledge is justified true belief) is still common. I am thinking of knowledge and belief in this philosophical way--such that propositional knowledge (though perhaps not practical, intuitional, or acquaintance knowledge) essentially involves belief. Belief, however, is a wider category than knowledge, simply identifying a kind of propositional attitude distinct from, say, doubt, desire, or hope. This attitude or psychological stance can obtain regardless of whether the proposition in question is in fact true or whether the attitude is justified.

In light of this we can make sense of why talking of direct control over belief is not necessarily problematic while the idea of direct control over knowledge is very odd. It is because knowledge essentially involves non-accidentally true belief. We could have direct control over knowledge only if our believing something could make it true or if our believing something would make the belief non-accidentally related to the truth in the proper way (i.e., justified, warranted, reliably, etc.). That is obviously not something that is generally under our control.

This is not, however, due to the fact that no form of voluntary control can have a role in knowledge. After all, if knowledge involves belief and we can have some responsibility for that belief (either by direct control or by means of certain acquired habits of belief-formation), then we have some responsibility for what we know. This is the case even if we had no responsibility for the truth of the belief and only indirect influence over the belief's non-accidental relation to the truth.

Some further points. Although both belief and knowledge are dispositional in character, belief will admit of degrees while knowledge is all or nothing: a subject may believe with more or less confidence or conviction, but a subject either knows something or she doesn't. Moreover, to call something "knowledge" is to ascribe positive epistemic status to it. That is to say, knowledge is a normative notion (and it is the kind of normative notion that does not admit of degrees). To call something a "belief," however, is simply to indicate what kind of attitude a subject holds towards a certain proposition (i.e., "S believes that p" as opposed to "S hopes that p" or "S doubts that p"). It does not ascribe a normative or positive epistemic status to that propositional attitude.

With these points in hand, we are now prepared to address the relationship of prescriptive language to belief and knowledge. The reason it is odd to say "You ought not to have known" is due to the fact that an ascription of knowledge to a subject carries with it a normative judgment that the belief involved in the knowledge is somehow permissible, normative, or has positive epistemic status. The statement, "You ought not to have known," is odd for the same reason that the statements "You were wrong to do what was right" or "You shouldn't have said what you were permitted to say" are odd. It is, however, natural to say, "You ought not to have believed." This is due to the fact that, though knowledge does not exclude belief, belief is a broader category than knowledge in that it can include items that lack positive epistemic status or do not count as knowledge. It is, then, typically those items that are thought to lack positive epistemic status (or are not taken to count as knowledge) that we say one ought not to believe.

Similarly, "You ought to believe that p" is somewhat odd since that statement carries with it more than just the normative judgment that p is the sort of proposition towards which the subject ought to take an attitude of belief. It also implies the judgment that such belief could and most likely would have positive epistemic status or count as knowledge for the subject in that situation (i.e., the belief concerns something obvious or at least something that should have been obvious to that particular subject given his background, experience, desires, non-epistemic obligations, etc.). Though it is not senseless to say "You ought to have believed" (since knowledge includes belief), it is usually more natural to say, "You ought to have known." This is due to the fact that, though knowledge does not exclude belief, knowledge is a narrower category than belief in that it includes only items that have positive epistemic status. It is, then, those beliefs that are thought to have potential positive epistemic status and are thought to be obvious to the subject that we say one ought to have known.

The asymmetry between belief and knowledge, then, is important. It does not, however, prevent us from speaking of obligations or normative concerns regarding both belief and knowledge. The same basic concepts seem to govern our use of normative language in respect to both belief and knowledge. Thus I suggest the following rough outline of the way in which normativity attaches to belief and knowledge. Prohibitions attach to what we take as mere belief (belief that, for whatever reason, cannot also be taken to be knowable). Permissions attach to beliefs concerning what we take to be knowable. Requirements attach to beliefs we take to be knowable and (within certain parameters) to be obvious to the subject. Epistemic virtue is a habit for the permissible and the fulfillment of requirements, while epistemic vice leads away from requirement and indulges prohibition.

For the time being I am taking prohibitions, permissions, and the like in terms of our actual applications of their terminology, which, I am supposing, reflects--though perhaps only darkly--the actual structure of epistemic normativity. This discussion will be "cleaned up" as we progress, for instance, we will see that normative language does not attach to belief quite so directly but only derivatively.

It also is interesting to note that epistemology, as a philosophical discipline, has traditionally focussed on what I have suggested as permitted belief, that is, beliefs concerning the knowable. Whether anything is knowable, what it is to believe knowledgeably (e.g., to have justification), and what objects are knowable are the traditional fare of epistemology. Our present project is somewhat broader than those traditional concerns (including, e.g., discussion of what we ought to know) as well as narrower (largely bypassing any discussion of skepticism and the limits of knowledge). Furthermore, the majority of discussion will focus on the categories of prohibition and permission, leaving requirement to the side until those categories have be clarified. Remember also that this rough outline of epistemic normativity is just that: a first sketch. As we progress we will approach these categories again and again, but from varying perspectives (e.g., from the perspective of everyday ascriptions justification).

This brings us to the second question: whether this normative language is truly prescriptive or deontological in force or whether it is merely normative in some weaker sense. We understand, I think, what it means for an action or state of affairs to be good (excellent, perfect, well-done) or bad (defective, imperfect, substandard) in a normative, though not prescriptive, sense. For instance, we may say that a soccer player's unsuccessful kick towards the goal was a bad kick, but we are not necessarily saying that the athlete violated any sort of prohibition or was irresponsible. Such prescriptive language is used only under certain additional conditions: if the soccer player had a great deal of control over his kick, if he failed to have been--for instance--as attentive and well-positioned as he ought to have been, if he had been regularly successful with similar kicks in the past, and if he was able to have done better. Only then would we say that it was a bad kick due to the athlete's irresponsibility and he ought to have done better.

Similarly, there is a normative aspect to belief-formation. For instance, upon meeting Ignatius and Catherine, Teresa immediately jumped to the conclusion that they were married. This belief was based upon shoddy evidence: they seemed to have known each other a long time and "looked good together." Normatively, then, Teresa's belief is defective, based upon slim evidence by a belief-forming habit that is insufficiently restrained and discriminating. But, unlike the bad soccer kick, we do not take Teresa's belief to be only normatively defective. We would, I think, go further and say that she shouldn't have supposed that Catherine and Ignatius were married. Prescriptive language enters the scene quite naturally.

Granted, there are similar cases in which we would not assess blame. For instance, if Teresa went to the refrigerator before shopping to check the date on the milk, she might, in the process, form a number of other beliefs peripherally and promiscuously: that the butter's almost gone, that there are only four cans of soda left, that there's too much condensation dripping inside the refrigerator.

Let us suppose, however, that Teresa was mistaken regarding the soda: there were in reality only three cans left, not four, as she supposed. There may or may not be something normatively defective in the formation of her belief, but, regardless, we wouldn't say to her, "You shouldn't have supposed there to be four sodas left" (the exception would be if she were somehow otherwise obligated to know the number of sodas, e.g., for her party tonight). This, however, does not disprove the idea that epistemic life is prescriptively governed. We would, I think, consider ourselves to be excusing Teresa's mistaken supposition in not blaming her (due, perhaps, to a recognition that Teresa was not properly responsible for these peripheral and promiscuous beliefs). But excuse presupposes typical blameworthiness and, therefore, prescription and responsibility as the rule.

The ways in which these prescriptive concepts govern prohibited cases of believing (for example, believing upon insufficient evidence), as well as those cases in which a belief is permissible, are relatively straightforward. On the other hand, cases in which there are requirements to know are more complex. They involve degrees of obviousness, other non-epistemic obligations, the teleology of belief-formation, and the like. As I have already noted, we will set such cases aside until we have better accounted for the simpler cases of prohibition and permission.

This brings us to the third question: whether this prescriptive language (prohibition, permission, requirement, ought, should, etc.) is ethical in its force or simply a matter of some kind of non-ethical prescription. There are a number of reasons to take such language as non-ethical. First, we can conceive of a case which makes little epistemic sense, but fine ethical or moral sense: young Calvin snoops into his mother's documents and finds out he had, at one time, been put up for adoption. Calvin asks his mother why she had done this and her first reply is, "You shouldn't know that!" Clearly she does not mean that the child's belief is epistemically defective, but rather that he pried into something into which it was morally wrong for him to pry. There is a distinction, then, between moral prohibition and epistemic prohibition even in respect to belief-formation.

Second, we can distinguish between cases in which a statement of the form "You shouldn't have believed that p" is merely epistemic in force from those in which it is more clearly ethical as well. Consider the following case: Cyril believes that Lucy called him a demeaning name because Gregory told him Lucy had done so. Gregory, Cyril knows, likes to try to get people in trouble and, therefore, is not always trustworthy in such matters. Nevertheless, Cyril believes him. We would say that Cyril shouldn't believe Gregory and that his doing so is an epistemic failure on his part. Such belief-forming behavior among these young children, however, is probably not censurable in a strongly ethical sense. If, however, Cyril is not another child, but Gregory and Lucy's teacher, we would take Cyril to have ethically violated the standards of fairness and professional distance that is required of him (as well as having violated epistemic standards).

There are, after all, many kinds of prescriptive standards: epistemic, aesthetic, political, professional, or institutional, for instance. These standards, it seems, are not properly ethical. It seems further that we are, in this study, interested in epistemic standards for the most part. That at least is a case that can be made. I must, however, note some hesistancy about that perspective.

While I recognize a distinction between mere normativity and prescription, I do not know what a non-moral prescriptive standard is or what a non-ethical obligation is. Political, professional, and institutional standards, it seems to me, are those standards that one is morally obliged to meet by virtue of good character, contract, general moral principles, or obedience to authority. Can there be immoral laws? Of course, but such laws create no obligations. Prescriptive aesthetic standards (and not mere technical normativity) are the standards which artists are morally obliged to meet by virtue of general moral principles, their dedication to aesthetic normativity, their artistic integrity, and their duty not to squander their talents.

Similarly, I believe prescriptive epistemic standards are the standards which rational agents are morally obliged to meet by virtue of the goodness of truth, its value within human ethical life, and the contribute it makes to human flourishing. Furthermore, epistemic standards embody or stem from general moral principles, evaluations, and behavior-guiding norms, especially those concerned with the virtues of honesty, humility in the face of experience, and receptivity to the other. In this way even epistemic prescription is thoroughly ethical. There may be cases in which all ethical or prescriptive standards have been met in respect to some belief and that belief is still normatively defective (e.g., hallucinatory beliefs due to strong pain medication). That case, however, turns on the distinction between the normative and prescriptive, not between the merely prescriptive and the ethical.

Furthermore, those standards that are grouped as properly ethical may also be categorized in less obviously ethical terms: as parental, neighborly, environmental, theological, self-directed, or other sets of obligations. How does one draw the line between the ethical and the merely prescriptive? There are no sharp lines between ethical and non-ethical prescriptive standards, and I believe this is due to the fact that to be prescriptive is to be ethical. Epistemic prescription, therefore, can be seen to be ethical.

The strongest argument against such a viewpoint, it seems to me, is that put forward by G.E.M. Anscombe in her classic essay, "Modern Moral Philosophy" (1958). In that essay she argues that the entire category of "moral" is suspect, not in the sense that there are no obligations, but in the sense that it is unnecessary to speak of a "moral" or "ethical" realm over and above what we are obliged to do in order to meet particular ends. That is to say, imperatives are hypothetical and relative to a particular notion of what constitutes a "good life" for a human person or, to put it another way, a well-ordered life (analogous to Aristotle's teleological ethics of virtue). I am actually quite sympathetic to such a viewpoint, but fill in the notion of a "good life" in terms of a teleology of human persons as directed towards life together in God. The form of life which tends to promote that end gives rise to obligations that I would term "moral" and "prescriptive."

I recognize, however, that not everyone will agree with me on these last points and that the point is not crucial to my basic argument. I also recognize, in light of reasons already given, that some prescriptive standards are usefully grouped as "epistemic" in distinction from other non-epistemic ethical standards. Therefore, I will stick to the term "epistemic obligation" (permission, etc.) in the following chapters, unless I am dealing with the connections between epistemic standards and what I shall call "more centrally ethical" (or just "ethical") standards. Nevertheless, what I shall argue for in the following chapters is, in large part, a virtue-based approach to epistemology.

Such are my claims thus far. These claims demand further elucidation and support.

[ home | la salle university | connelly library | yahoo! | the bbc | about such things ]