… that collapsed circus tent of epistemology—those acres of canvas under which many of our colleagues still thrash aimlessly about.
—Richard Rorty
The study of ignorance is an exercise in epistemology; but it is also an exercise with implications for epistemology as a field of study. It implies a critique and expansion of traditional, analytical epistemology. I have noted those implications throughout; they have formed a subtext to my discussion. For scholarly impact, they require a systematic and appropriately technical presentation that is, however, not the focal purpose of this book. What I can attempt here is to distill these implications in this epilogue for readers with special interest in meta-epistemology.
The present is an exciting time for epistemology. Happily, each of the elements of this critique has already been advanced by other philosophers in works of influential scholarship. I have referred to their projects throughout this text: social epistemology, “knowledge-first” epistemology, virtue epistemology, feminist epistemology, and resistance epistemology. But what is new, I believe, is the recognition that these developments, especially in tandem, open the conceptual space required for any adequate understanding of ignorance; the converse claim that ignorance can open new lines of inquiry in all these approaches; and the ultimate vision of an epistemology that is centered on the interaction among understanding, knowledge, and ignorance.
Epistemology, under standard definitions, has a capacious portfolio: it studies the nature, scope, and limits of human knowledge. In theorizing knowledge, it sustains a subsidiary interest in our cognitive faculties: perception, memory, imagination, emotion, reason, intuition, and the mind itself—especially insofar as these relate to the acquisition, possession, and retention of knowledge. This subsidiary work, however, is often shelved under philosophy of mind; and epistemology proper (as practiced by Anglophone epistemologists, at least) has held a much narrower focus.
It is fair to say that modern epistemology has been haunted by the insolent skepticism that is the legacy of Cartesian doubt. The highest priority has been the assurance of the possibility of knowledge. (No one has doubted the possibility of ignorance.) Knowledge has had to be characterized defensively, so as to ward off the demons of doubt. Thus, epistemologists have concentrated on explicating the sources, structure, and justification of genuine knowledge, especially in contrast to mere belief. Certainty has been the gold standard for “knowledge in the strong sense”; anything less is dross. Epistemic analysis is, therefore, pointedly normative or regulative; and the primary, likely the sole, epistemological value—is truth.
Although skepticism has been a philosophical stance since the ancient Greeks, the crucible in which modern epistemology took form was much later. The necessity of distinguishing genuine knowledge from belief was burned into Western thought by the conflicts among religious doctrines during the Reformation and Counter-Reformation, and especially by the conflict between religion and science. It gained traction with the elevation of mathematical science (formal systems) as the idealized model of knowledge. And in the early twentieth century, positivism and Reichenbach’s dictum that epistemology should concentrate exclusively on the context of justification (leaving the context of discovery to psychologists) secured the focus.
I do not denigrate this history. The public culture of ignorance and contemporary conflicts that arise from religious beliefs, conspiracy theories, pseudo-science, image manipulation, and disinformation campaigns show us that distinguishing genuine knowledge from belief is of high priority.
During the past century, vigorous debates on the problem of knowledge were conducted largely within the traditional schema: belief, truth, justification, and (since 1963) the absence of Gettier conditions—all directed toward and expressed in propositions. This schema has enabled an extensive research agenda; the vast literature it has generated displays technical refinement with increasingly sophisticated formulations and elaborations. It has also imposed a severely limited vision of epistemology. For some, it may still be a circus tent with a great show; but for others, it is collapsed—a dull preoccupation. As a comprehensive account of knowledge, I believe its potential for illumination has largely been exhausted. Although continued elaboration may yield surprises, I think its best value going forward is not as the commanding paradigm for the field, but rather as a finely honed heuristic instrument to be deployed within an expanded vision of epistemology.
In what ways does this traditional schema impose a “severely limited vision”? I have claimed that it is restrictive in that it:
One could insist, of course, on restricting the term epistemology to the narrower focus; but that would be arbitrary and clearly incomplete as a theory of knowledge. Let us begin the unpacking with the first two of these claims, and then pursue the remained in the order listed.
The interpretation of knowledge as a specially certified set of propositions is reflective of philosophy’s fixation with language. But we recognize there is a difference between a set of such propositions and all that is involved in knowledge—say, one’s knowledge of mathematics. The schema cannot bridge that gap. What more is required?
If we point out that knowledge also involves coherence among our multiple beliefs, theorists may reply that tests of propositional coherence are built into the warranting condition (which is likely to be reduced to known propositions as well). But knowledge clearly involves more than propositional coherence: it involves conceptual integration, conceptual frameworks, cognitive structures within which concepts can be grasped and propositions formed, for example.
A comprehensive epistemology would surely embrace more forms of knowing than propositional knowledge. Some stalwarts defend the intellectualist position, demanding that knowing how, knowing what something is like, and knowing an individual must be reduced to or cashed out as a set of propositions if they are to be regarded as having any genuine knowledge content. It is a view suitable for the Age of Assessment, but most philosophers accept all of these as distinct ways of knowing.1
Moreover, there are important epistemic concepts that are undertheorized by and alienated from the traditional schema, such as expertise, understanding, and wisdom. Richard Mason and Catherine Elgin have independently argued that our central epistemological focus should shift from knowledge or belief to understanding—or, at minimum, that epistemology should now attend to understanding understanding.2 (I have titled this book “Understanding Ignorance,” not “Knowing Things about Ignorance.”) As Mason explains, that project cannot be based on the traditional schema: understanding incorporates tacit or knowing as well as explicit propositional knowledge, and it also admits of degrees. Jonathan Kvanvig has argued that although knowledge has immense value, so does understanding; and a case can be made that it has even more value.3 He believes that understanding is not a form of knowledge, but that knowledge incorporates understanding. For me, understanding is both a broader and more basic concept than knowledge, not narrower, and it points to the concept of wisdom.
It is a nice question whether knowledge or understanding is more basic. Earlier I noted Williamson’s view that knowledge is primitive and logically prior to belief, and the thin interpretation of ignorance that results.4 Williamson does not discuss understanding in Knowledge and Its Limits, but presumably his analysis would require him to take understanding to be a derivative mental state, normative to truth. But I find it equally or more plausible that understanding is the primitive concept, logically prior to knowledge and belief.
The upshot of these claims is that epistemology properly includes more forms of cognition than propositional knowledge. As this purview becomes more inclusive, the role of ignorance may be further revealed and elaborated.
Even if one accepts the standard schema as the commanding analysis of knowledge, it follows that ignorance has a richer structure than a single, simple negation of knowledge. If genuine knowledge requires the meeting of four criteria, the failure of any one of the four criteria (belief, truth, warrant, and the absence of Gettier conditions) or any particular combination of failures represents a different way of not knowing, a situation in which S does not have knowledge of p. It is worth sorting out these epistemic failures.
My larger point in detailing these cases is to show that, although the schema may suggest a single, unified way of knowing p, it also allows for several varieties of not knowing p; and the differences are epistemically significant, in part because they point to the particular sources of ignorance. They foreshadow, at least partially, the structure of ignorance.
A second point: there is an asymmetry in the way the criteria implicitly relate to S. Both belief and justification seem to require some sort of “cognitive possession” by the believer. S must believe and have warrant. It is not simply that warrant must exist; S must “possess” it, if S is to have knowledge. (But it would open the trapdoor to circularity if that means S must know that she believes p and must know the facts of the warrant and that they are epistemically compelling.) By contrast, S cannot be the authority for affirming the truth of p; that would simply reassert S’s belief. And, by definition, the absence of Gettier conditions in affirming p can never be claimed by S; they are always unknown unknowns. Thus neither of these criteria is “cognitively possessed” by the believer; each states S-independent, objective facts. This asymmetry reflects the subjective and objective perspectives implicit in the schema.
In a recent series of articles, the Dutch philosopher Rik Peels has argued that “contrary to what one might expect and to what nearly all philosophers assume, being ignorant is not equivalent to failing to know,” at least on the four-criterion traditional interpretation.5 Rather, Peels claims that ignorance is a lack of true belief. (He does not claim that having true belief is sufficient for genuine knowledge; for him, knowledge and ignorance are asymmetrical in a way quite different from the asymmetry I just described.) Now there is no doubt that a lack of true belief is a failure to know; either the belief or truth criterion or both have failed. (In my account, these conditions create three different forms of ignorance.) The salient claim for Peels is that the failure of the warrant condition will not by itself constitute ignorance.
I reject this interpretation. Note that if Martha, who lacks justification for her true belief that a sliced onion will reduce the pain of a bee sting, has her belief confirmed by medical authorities or discovers an explanation for the palliative effect, she will have removed her ignorance on that point. Similarly, a detective who correctly believes she has identified the killer but lacks sufficient evidence will continue her investigation to discover motive, means, and opportunity. If she succeeds, she both gains a justification for her belief and removes her ignorance. I believe that the lack of warrant for true belief does indicate ignorance. Moreover, Peels has reopened the door to epistemic luck: on his account, Gettier conditions do not create ignorance. According to Peels, Jim would not be ignorant of the outcome of today’s baseball game, because he holds a true belief. This seems counterintuitive. What term would one use to describe epistemic defects like Jim’s failure to know the outcome of today’s baseball game or our detective’s missing knowledge? To discount the relevance of warrant and luck for ignorance also disvalues the securing of justification and the achievement of knowing without epistemic luck.
A brief pause to note an odd case: can one be ignorant of something that is, in fact, false?
Putative examples are suspicious. Suppose Sherman is ignorant of the principles of phrenology. Suppose Miriam is unaware of a false report in the local paper. Suppose Ian is unaware that the Piltdown Man is thought to have supplied the “missing link.” Such cases seem to involve double targets of ignorance. Sherman is ignorant of the principles of phrenology, and did not know they had been debunked. Miriam did not know about the report, so did not know that it was false. Ian had never heard of the Piltdown hoax. More generally, one might be ignorant of certain impossibilities; but that is better understood as one’s being ignorant of the truth that such-and-such is impossible, not of the falsehood that such-and-such is possible. Ignorance is never a lack of ignorance.
The overarching claim in this section is that, even within the confines of the traditional schema, one finds evidence of the more complex structure of ignorance.
Because the aim of the traditional schema is to secure genuine knowledge against doubt, it adopts an absolutist model; that is, its logic treats key concepts as though they are bivalent (on/off) properties: either S knows that p or S does not. Yet two of its criteria center on concepts that admit of degrees. Belief is one such term: there is an epistemic difference between a belief that I do not realize I have until asked (“Why yes, I believe my desk lamp in the study is on now”) and a deep conviction (“I believe that life is more important than property”). Earlier, we encountered “degree of belief” as an interpretation of probability (chapter 10). I also noted the difference between not having a belief and refusing to believe. In short, belief can involve gradations of conviction and awareness.
There is a similar problem with the concept of warrant. Treating warrant or justification as bivalent concepts denies the scalar reality: a belief might be more or less justified, and the strength of its warrant might alter over time. In addition, it is misguided to imagine that the requirement for “sufficient warrant” is identical for all types of knowledge claims, or that we should believe only claims that have the strongest warrant or are certain. Such follies occur when epistemologists idolize mathematical knowledge, extending its model to empirical sciences, and thence to all forms of “genuine” human knowledge.
Are knowing and not knowing therefore also a matter of degree? If the two key components of belief and warrant have gradations, surely the target concept (knowledge) must also have gradations. In our consideration of ignorance as a boundary (chapter 5), I contrasted the “disjunctive” and “scalar” interpretations of knowing, presenting various “borderline” cases of knowing that involve degrees of recall, recognition, articulation, awareness, and so on. Admittedly, these matters, at least prima facie, are indicators of knowledge, not essential components of the logical structure of knowledge, but they do bear on the affirmation that “S knows p.” They indirectly reflect the gradualism of knowing.
Knowing how and knowing what it’s like do not have the same sort of traditional schema; the conditions for their certification are far less standard. But it is clear that there are degrees of skill—one can be more or less skilled at rope-jumping or organ-playing. It is not an either/or designation. It is controversial whether the same is true for knowing what it’s like, but I think even here, at least in many cases, the knowing can be gradual. I can learn what it is like to fly a plane somewhat by piloting a model airplane. But I learn more by sophisticated simulation programs. I learn still more by sitting with a pilot during flight. And I come to “really” know by piloting in various aircraft under varying conditions.
Any adequate epistemology should accommodate a gradualist understanding of key terms like belief, warrant, and knowing. The bivalent interpretations of the traditional scheme reflect the pull of certainty and its absolutist demands.
The traditional adherence to Reichenbach’s boundary between discovery and justification has had the negative consequences of disconnecting the theory of knowledge from the process of learning, and of divorcing normative epistemology from educational practice. Treating knowledge only as a fait accompli, a structure of certified propositions, has also severed knowledge from the capacities and virtues that are required for its acquisition and the vices that may thwart it.
During my exposition, I have turned to several approaches that cross or open Reichenbach’s boundary between discovery and justification. Making this boundary at least more permeable is necessary for a better understanding of ignorance.
Perhaps Michael Polanyi is the twentieth century thinker who has rejected this bifurcation most directly. I cited him (first in chapter 3) for his advocacy of tacit (proposition-resistant) knowledge. But it is a different aspect of his work that I have in mind here. The thesis of his classic 1958 book, Personal Knowledge,6 is that “impersonal knowledge” is a fiction, that scientists necessarily “participate in” the knowledge they possess and create, and that such knowledge requires a personal commitment and results in personal transformation. Scientific knowledge bears the marks of the human enterprise that produced it.
Gettier, in showing that epistemic luck can undermine apparently situations of knowing, is addressing the context of discovery—the context of coming-to-know—as well. Luck is a matter not of the structure of knowledge, but of how one came to believe it was true.
The most direct and sweeping challenging to Reichenbach’s barrier, however, is the rise of virtue epistemology. Taking knowledge to be an achievement opens the connection between the ways in which one pursues knowledge and the quality of the knowledge one obtains.
My point here is that these assaults on the Reichenbach distinction move us in a positive direction. It is not that I believe epistemology should become or replace cognitive psychology; the two fields properly have different aims and methods. But if there are epistemically relevant issues that arise within the process of coming to know, any adequate epistemology should be inclusive of them and attend to them. For me, all these approaches are useful in part because they also reveal the interaction with ignorance (though they are not normally presented in that way). Thus, Polanyi’s notion of tacit knowledge is one of unknown knowns. And Gettier describes epistemically erosive conditions of which the believer is ignorant. Virtue epistemology opens up space to consider the role of ignorance as motivator and as a factor in such virtues as intellectual humility, epistemic restraint, and discretion.
The “S” of “S knows that p” might as well stand for Solo. Knowing is understood to be a mental state of individuals, and knowledge is possessed by individuals. As a corollary, epistemic autonomy is taken as ideal, while epistemic dependence is regarded as deficient. Thus, first-hand knowledge—seeing with one’s own eyes, checking the proof for oneself, directly verifying purported facts—is preferred if one’s knowledge is to be sterling. The best knowledge is literally self-evident—evident to oneself. Reliance on the authority or testimony of others for information and knowledge is déclassé. Such dependence risks crediting hearsay, even gossip, and other mediated information, even if it is common practice and the basis of didactic teaching.
This framing of thought is, of course, traceable to Plato: the model of knowing-as-seeing, culminating in noesis, the moment of intellectual insight experienced only by a well-prepared individual. But this model can have both authoritarian and democratic forms. It suited well the Reformation and the rise of liberalism. Descartes’s method of doubting all inherited knowledge to believe only ideas he himself found indubitably clear and distinct is a radical version. The impulse to epistemological foundationalism betrays an atomistic individualism and a certain self-importance; it seeks to shed the messy structure others have made so as to build a superior version oneself. W. K. Clifford’s dictum that our beliefs not exceed our direct evidence is a stringent rule for the autonomous, individual knower.7 In the democratized form, every rational human is equally capable of discerning the truth, of possessing and being a source of knowledge; and each individual carries the epistemic responsibility of verification, of knowing first-hand.
Yet in life as we live it, knowledge is constructed within an epistemic community; it is possessed and justified or rejected in the epistemic interactions of a community. We all hold membership in different, plural, and overlapping epistemic communities, and most of anyone’s personal “storehouse of knowledge” is derived from other sources. It is a distortion to lump all testimony into a tainted category: it not only ignores practice, it also obscures the rich and varied epistemic relationships of trust, authority, and expertise that are exhibited in the giving and receiving of testimony. Epistemic dependence is not only frequently necessary, but in some situations, such as the deferral to experts, it may be epistemically virtuous or even obligatory.8
The appropriate response is now coalescing: a shift from the individual knower to the epistemic community, with a correlative shift from epistemic autonomy to forms of epistemic dependence. One place to begin is the reassessment of the epistemic value of testimony: C. A. J. Coady produced a provocative analysis that branded “autonomous knowledge” an “illusory ideal,” and bolstered testimony as a common and legitimate source of knowledge.9 In the groundbreaking work of Alan I. Goldman, the value of testimony is placed within the larger context of other practices, such as argumentation and a set of traditions that represent epistemic communities such as law and science.10
Social epistemology and especially feminist epistemology have inspired reflection on the ethical dimensions of epistemic communities. I earlier cited Miranda Fricker, who has identified practices of “epistemic injustice” (chapter 7), such as the systematic devaluation of the testimony of certain groups, which creates an epistemic oppression usually tied to other forms of prejudice and marginalization. Social epistemologists of varying stripe have brought attention to sources and forms of socially constructed ignorance, to the privilege and power that permit certain types of willful ignorance, and to the need for an “epistemology of resistance” that reveals and disrupts structures of epistemic oppression.
I find the concentration on epistemic communities particularly fertile when it is linked with virtue epistemology. In that juncture, intellectual virtues are not merely characteristics of the learner as an autodidact and individual knower; they are intellectual qualities and capacities that sustain the functioning of an epistemic community. Just as moral virtues require communal supports, arenas, and forms of engagement for their display and refinement, epistemic virtues also require the support, arena, and forms of participation of an epistemic community.
Understanding a culture of ignorance requires the insights of social, virtue, and feminist epistemologies. These approaches may be integrated; only their integration allows us fully to understand and the dynamics and the multiple forms of ignorance in our culture.
Traditional epistemology is normative and regulative in its placement of genuine knowledge as the sole intrinsic, epistemic value (dependent, of course, on the value of truth). Knowing—discovering and possessing knowledge—is the epistemic good: the more knowledge, the better. This is affirmed whether knowledge is valued intrinsically or instrumentally.
Interestingly, none of the new approaches cited to this point has challenged that presumption. Nonetheless, I argue that this presumption is problematic; it may be challenged within their more expansive vision of the field. Of course, knowledge is of enormous value. And although true beliefs are useful, we are all confident that genuine knowledge is more precious (though spelling out the rationale for that judgment may be difficult).11 The question is whether knowledge is always of value and is the only epistemic value, or whether it is responsive to other values.
How might the assumption of monistic epistemic value be challenged? One response is that it precludes the possibility that it might sometimes be better not to know. It therefore treats conditions of privacy or confidentiality, matters of dangerous or harmful knowledge, along with decisions not to know, as of ethical concern only, not of epistemological interest. Surprisingly, even virtue epistemology has usually characterized virtues and vices only in relation to the acquisition of knowledge, a valued goal. Virtues like discretion and intellectual humility, and vices like nosiness, tattling, and betrayal—qualities that reflect the value of restraining or withholding knowledge—are largely ignored. Yet these traits become important when we theorize a normative epistemic community.
A second, related response is that it fails to give proper attention to ignorance. As we have seen, ignorance may have value, as in rational and strategic ignorance (chapter 6). In recent literature, the strongest case for ignorance as a positive epistemic value has been made by Cynthia Townley, who argues both for its value for individual knowers and commends its role in optimal epistemic communities.12 In her account, the individual knower becomes an epistemic agent, acting within an epistemic community. A well-functioning community has pluralistic epistemic values, of which knowledge is only the primary one; ignorance under certain conditions is valued as well.
The view that epistemology should value both knowledge and ignorance (polar opposites) can be made coherent only if there are deeper, governing values at work—values that determine when the valence properly switches from one to the other. I quite agree with Townley that when we shift from the individual knower to the role of responsible agents in epistemic communities, we find that epistemic values other than knowledge are in play. Epistemic agents are responsible as believers, as testifiers who pass on knowledge, as receivers of information, as confidantes and authorities, and so on.
I would go a bit further and claim that we learn a great deal by thinking of epistemology as an arena of value theory and ethical concern. Just as we consider justice in the distribution of rights, privileges, and rewards, so we may consider justice in the distribution of epistemic roles and values like knowledge, ignorance, expertise, and sharing. The concern is broader than the ethics of belief, though that is important. It includes the ethics of curiosity, secrecy, confidentiality, transparency, accuracy of testimony, and other practices. Once we make this transition in perspective, new norms of epistemic flourishing may be articulated and familiar ones may take new forms.13
If Anglophone epistemology has too often been locked in Cartesian concerns, Enlightenment affirmations, and structural models of knowledge, it nonetheless has generated analytical progress. The critique summarized here does not aim to overthrow that progress, nor does it jettison the formal schema (though I believe that its continuing creative possibilities seem limited). Rather, it argues for a grander, more inclusive framing of epistemology within which the traditional analysis would occupy a particular place. In the innovative approaches I have cited, that reformative task has begun.
The excitement I find in epistemology is the vision of a study that embraces various forms of knowing and gives proper place to understanding; of an inquiry that links the ways in which we come to know and the structure of knowledge; of an epistemology that in all these dimensions addresses the interactions between knowledge and ignorance in community and the values that govern them; of a field that illuminates our epistemic predicament as we move among the known, the knowable, the unknown, and the forgotten, within the horizon of the unknowable.