Lecture Supplement on Epistemic Justification:

Foundationalism, Coherentism, Externalism and Internalism

I. Some General Points Regarding Justification:

One needs to note that there are a variety of types of justification: procedural justification (involving appeals to rules or regulations), legal justification (involving appeals to laws), moral justification (involving appeal to moral rules, laws, or considerations), epistemic justification (involving epistemic considerations), and pragmatic justification (involving appeals to goals).

It is also important that we recognize that amongst the “things” which can be (and may need to be) justified are: beliefs, actions (or behaviors), individuals, and groups.

Justification generally involves appeals to rules, goals, standards, or criteria. Moreover, it is often a matter of degree. While there is both an activity of justification and a state of being justified, it would seem that both must have a non-accidental analysis. That is, when speaking of warranting evidence, one needs to consider defeasibility! We must, however, distinguish prima facie justification and unqualified justification. Distinguish justification as an “absolute,” as a “relative,” and as a “threshold” (like a lethal dose) concept.

Epistemic justification isn’t simply concerned with belief, it is supposed to help us ensure that we go beyond belief toward knowledge. Thus it is essentially bound up with the “goals of epistemology.” Generally this means that it is concerned with helping us attain truth. There are a number of different ways this can be factored in however. For many philosophers (especially what we will call the “internalists” [see below], this is often framed in terms of the notion of epistemic responsibility. In his The Structure of Empirical Knowledge, for example, Laurence BonJour maintains that epistemic justification ensures the agent has reasons for her belief:

...to accept a belief in the absence of such a reason, however appealing or even mandatory such acceptance might be from some other standpoint, is to neglect the pursuit of truth; such acceptance is...epistemically irresponsible.[1]

-Note: the very notion of epistemic responsibility needs to be carefully considered. Among the things which need to be investigated are: the viability of epistemic voluntarism, the question of what one’s epistemic duties are (assuming that there are any), the question of how strong they are, and the question of how these duties are justified.

Other epistemologists want to want to talk about how the relationship between beliefs and truth can be non-accidentally strengthened without emphasizing the role of reasons, justifications, and responsibility. These “externalists” [see below] tend to focus on mechanisms, processes, or other relationships which have the propensity to engender true beliefs, in the hope that focusing our attention on these can help us identify, mechanisms for securing such beliefs.

Whatever the orientation of the epistemologist, however, it is clear that epistemic justification is concerned with facilitating the attainment of knowledge.

II. Foundationalism and Coherentism:[2]

The distinction is best approached by first discussing The Problem of Epistemic Regress. This ancient problem points out that the structure of epistemic justification seems to allow for only the following options:

There are some basic propositions (beliefs, etc.) which provide a valid foundation and justification for all other propositions (beliefs, etc). They are not themselves justified by appeal to any other propositions (beliefs, etc.). This option is called foundationalism.

Propositions (beliefs, etc.) are justified by appeal to other propositions (beliefs, etc.) in an unending and infinite chain of justificatory regress. This option is called infinite regress.

Propositions (beliefs, etc) are justified by appeal to other propositions (beliefs, etc.) which together form a finite collection of propositions in a manner which allows for a closed system of justification. This option is called coherentism.

Propositions (beliefs, etc.) are justified by appeal to other propositions (beliefs, etc.) but these propositions themselves are not themselves justified. These beliefs are either acknowledged to be arbitrary or are said to be accepted on faith.

The process of epistemic justification is ultimately unsustainable. This option is called skepticism.

Proponents of any one of these options, of course, has to argue against the other alternatives and argue for their own view of epistemic justification.

Foundationalists draw a distinction between mediate and immediate [epistemic] justification. They hold that while some propositions (beliefs, etc.) are justified by appeal to other propositions (beliefs, etc.), and, thus, are mediately justified, others must be immediately justified. The justifiability of these propositions (beliefs, etc.) does not depend upon or involve an appeal to any other propositions (beliefs, etc.). Another way to characterize their view is to speak of the thesis of epistemic priority— the view that some claims have a privileged epistemological status which allows them to play a unique role in epistemological justifications. Foundationalists hold that the process of epistemic justification requires that there be some propositions (beliefs, etc.) which have “priority” in the justificatory process—they are usually said to be either self-justified (self-evident, self-warranted) or justified by experience (or direct awareness, or some other non-doxastic element). Of course, other propositions (beliefs, etc.) are justified by appeal to these. In other words they provide the “foundation” upon which further propositions (beliefs, etc.) are justified (by mediation). As Louis Pojman notes, the priority here is asymmetrical: “…the basic beliefs transfer justification and knowledge to the derived belief but not vice versa.”[3]

The priority is often characterized as a form of “immunity:”

immunity from error (infallibility)

immunity from refutation (incorrigibility) or

immunity from doubt (indubitability)

-psychological indubitability

-logical indubitability

-indubitability because of lack of grounds for doubt

Whatever form of immunity is chosen however the foundational propositions (beliefs, etc.) must have some doxastic legitimacy which they can transmit to other propositions (beliefs, etc.) if foundationalism is to be possible [this is why the response of those who would rely upon “arbitrariness” does not seem viable to epistemologists]! Now whether they speak of prima facie warrant, self-evidence, certainty, infallibility, or incorrigibility, the foundationalists have their work cut out for them.

self-evidence in logic and the principle of non-contradiction

self-evidence and Descartes

self-evidence and experience

-In their Contemporary Theories of Knowledge, John Pollock and Joseph Cruz maintain that:

we claim that in determining whether a belief is justified, importance must be attached to perceptual states, but this cannot be accomplished by looking at beliefs about perceptual states. This suggests that justification must be partly a function of the perceptual states themselves and not just a function of our beliefs about the perceptual states. This sort of view is called direct realism, and a version of direct realism will be defended in this book.[4]

It is clear that the role of the foundational elements must be extensively discussed. William Alston makes this clear in his “An Internalist Externalism:”

...I have no temptation to restrict the topic of epistemic justification to the activity of justifying. Surely epistemology is concerned with the epistemic status of beliefs with respect to which no activity of justifying has been carried on. We want to know whether people are justified in holding normal perceptual beliefs, normal memory beliefs, beliefs in generalizations concerning how things generally go in the physical world, beliefs about the attitudes of other people, religious beliefs, and so on, even where, as is usually the case, such beliefs have not been subjected to an attempt to justify. It is quite arbitrary to ban such concerns from epistemology.

But though the activity of responding to challenges is not the whole of the story, I do believe that in a way it is fundamental to the concept of being justified. Why is it that we have this concept of being justified in holding a belief and why is it important to us? I suggest that the concept was developed, and got its hold on us, because of the practice of critical reflection on our beliefs—in short the practice of attempting to justify beliefs. Suppose there were no such practice; suppose that no one ever challenges the credentials of anyone’s beliefs; suppose that no one ever critically reflects on the grounds or basis of one’s own beliefs. In that case would we be interested in determining whether one or another belief is justified? I think not. It is only because we participate in such activities, only because we are alive to their importance, that the question of whether someone is in a state of being justified in holding a belief is of live interest to us.[5]

Different versions of foundationalism arise either through differences regarding the foundations or differences regarding the process which allows as to builds upon the foundation (the processes of mediate justification).

The different orientations due to the “building factor” are generally traced to the differences between appeals to:

deduction

induction

probability or

inference by best explanation

I started with foundationalism because it is generally the preferred response to the problem of epistemic regress, and because while it has a number of significant challenges which are well-recounted, many epistemologists nonetheless believe that ultimately they can be met and a viable foundationalism will be developed.

Coherentism is harder to characterize. In his Problems of Knowledge: A Critical Introduction to Epistemology, Michael Williams maintains that:

one might think of a system of beliefs as like a space station, held together by its internal structure and wheeling along through space without resting on anything. Or one might use the analogy of an ecosystem, in which organisms, playing their distinctive roles, keep each other going. Where the foundationalist sees an architecture of knowledge, the coherence theorist sees something like an ecology with beliefs occupying interdependent niches.[6]

In his Pyrrhonian Reflections On Knowledge and Justification, Robert Fogelin points out that one should not think of systematic or holistic justification in “linear” terms:

nor does the coherentist permit what might be called a circular form of linearity, that is, a structure of reasons that simply loops back on itself. For the standard coherentist, linear circularity is a bad form of circularity. In place of such linear conceptions of justification, the coherentist pictures justification using such metaphors as a network, a mesh, a system, or an organic totality of beliefs. The fundamental idea is that the items in coherent systems of beliefs must stand in relationships of mutual support.[7]

The core challenges coherence theorists confront are: (a) characterizing what they mean by ‘coherence’, clarifying why coherence might be taken as indicative of truth, (c) clarifying how one can assess coherence, and (d) clarifying why one set of coherent propositions (beliefs, etc.) should be preferred to other sets of the same.

III. The “Contextualistic Alternative” to Foundationalism and Coherentism:

As both Pojman and Alston note in their discussions of foundationalism and coherentism, there is at least one fairly new option: contextualism. Contextualists contend that since justificatory questions arise in particular contexts (which carry their own presumptions and assumptions), they can be resolved by appeal to these presumptions without assigning these some degree of “epistemic priority” or needing to fit them into an overall system of propositions (beliefs, etc.). In trying to quickly introduce the “contextualist alternative” viewpoint, let me use two citations. First, in her “Relocating Aesthetics: Goodman’s Epistemic Turn,” Catherine Elgin maintains that:

contemporary realists are prone to think that literal language at its best partitions its domain into natural kinds, or divides nature at the joints, or discloses the true and ultimate structure of reality. Somehow, the world is supposed to dictate its proper description. [Nelson] Goodman denies this. He believes that any order we find is an order we impose. Systems of categories are contrived to impose order. They divide a domain into individuals and group these individuals into kinds. They thereby equip us to describe, predict, explain, and complain about the entities thus recognized. But the success of one category scheme does not preclude the success of others. There is no unique way the world is, hence no privileged way the world is to be described. A single domain may be organized in multiple ways; and for different purposes, different classifications may be best.[8]

Second, in his “The Pragmatic Dimension of Knowledge,” Fred Dretske suggests that we:

...think of knowledge as an evidential state in which all relevant alternatives (to what are known) are eliminated. This makes knowledge an absolute concept but the restriction to relevant alternatives makes it, like empty and flat, applicable to this epistemically bumpy world we live in.[9]

The social or pragmatic dimension to knowledge, if it exists at all, had to do with what counts as a relevant alternative, a possibility that must be evidentially excluded, in order to have knowledge. It does not change the fact that to know one must be in a position to exclude all such possibilities. It does not alter the fact that one must have, in this sense, an optimal justification—one that eliminates every relevant) possibility of being mistaken.[10]

In his “What Perception Teaches,” Fred Dretske maintains that:

maybe the impossibility of walking on water means one can’t walk to New York from Paris, but that doesn’t mean a person (even a Parisian) cannot walk to New York City. They can walk there from Hackensack. No trick at all.

Exactly the same is true of perceptual reports. Whether I can see that my wife is on the sofa depends on where I came from in reaching that conclusion. If all I am claiming to have done (and there are many conversational contexts in which this is all I am claiming to have done) is to distinguish, on visual grounds, my wife being on the sofa from your wife being on the sofa or (depending on contrastive intent) my wife being on the sofa from my wife being somewhere else, that is no trick at all. The possibility of hallucinations, dreams and illusions—the usual bag of skeptical tricks—is totally irrelevant to these modest perceptual claims. What the possibility of hallucination and deception demonstrates is that one cannot arrive where one says one arrived—knowledge that one’s wife is on the sofa—from a state of total ignorance. It shows that one cannot distinguish, at least not on purely visual grounds, veridical from illusory visual experiences. It shows that one cannot, as it were, walk on epistemological water. In saying that one sees that one’s wife is on the sofa, though, one is not claiming to walk on water. All one is saying is that one can successfully tell (visually) the difference between one’s wife being on the sofa and her being wherever else she might be if she were not on the sofa. Epistemically speaking that discrimination is a piece of cake—the epistemological equivalent of an easy stroll across a bridge.[11]

Knowledge is what you get when you reduce some pre-existing, contextually understood, set of possibilities—the relevant alternatives—to one. You can do that, and thus have knowledge, without supposing that this pre-existing set can itself be reached from some universal set (all possibilities) by the same, or, indeed, by any, method. That would be like inferring that there must be a finite number of (whole) numbers because every (whole) number is finite.

Not only are all our ways of knowing open, the knowledge enabling relations we all depend on are open. They, too, stop short of limiting propositions. Consider indication. We come to know that there are deer in the woods by seeing the tracks that indicate this. Yet, tracks in the show indicate there are deer in the woods without indicating there is a physical world. Tracks in the snow are part of the physical world, not something that indicates there is a physical world. Tree rings indicate the age of the tree and we use this fact to find out how old the tree is by counting the rings. Yet, the rings indicate the age of the tree without indicating what we know to be a necessary consequence of this—that the past is real. We cannot answer Russell’s skeptical query about the past by looking for tree stumps with at least one ring in them. If there is a past, if the limiting proposition is true, tree rings will indicate how much of a past the tree has had, but they will not indicate that the tree has had a past.

In a sense, then, indication is an epistemological stronger relation than logical implication. If I know that A indicates B, then, by learning that A is true, I can come to know that B is true. But if I know that A implies B, it does not follow that I can come to know that B is true by learning that A is true. If B is one of our limiting propositions—that the past is real, that there is an external world, that there are other minds—then A, something I now to be true can imply it without indicating it. So I cannot learn that these propositions are true by discovering that they are logically implied by what I already know. To know they are true I would need something to indicate that they are. Skepticism gets its foothold on the fact that nothing we are aware of indicates that these propositions are true although their truth is assumed by the things we know.[12]

You can see that Dretske is, partially, building on some of the ideas suggested by Lehrer and Paxson, and that he is giving this view a decided twist. This is as suggestive as I can be here about how the contextualists would treat the topic of epistemic justification—you can read David Annis’ “A Contextualistic Theory of Epistemic Justification” in our text for further information on this orientation.[13]

IV. Externalism and Internalism:

In his Unnatural Doubts: Epistemological Realism and the Basis of Scepticism, Michael Williams maintains that:

the essence of externalism...is to allow knowledge when a person in fact meets certain conditions, whether or not he knows he meets them. These conditions may be “external,” not just in not being represented in the person’s knowledge or beliefs, but in having to do with his actual situation. The capacity for knowledge is thus like any other capacity: it depends partly on the powers of the individual and partly on the circumstances in which he is required to exercise them.[14]

The individuals we have been studying have all either advanced or presumed the internalist perspective thus far (though Goldman has a decided leaning toward externalism). Traditionally epistemologists have had this orientation. Externalism, while relatively new on the scene, is important. As Robert Fogelin notes in his Pyrrhonian Reflections On Knowledge and Justification:

...justificatory performances come in a variety of forms ranging at the extremes from those that involve complex ratiocination to those that rely upon the unreflective use of a perceptual power or capacity. Both can be carried out in a responsible or an irresponsible manner; both can establish or fail to establish the truth of some belief. Both are sources of knowledge. Until relatively recently, philosophers have often tended to think of knowledge as solely the product of intellectual activity. The externalists have made an important contribution to epistemology by breaking the spell of this intellectualist prejudice.[15]

In his Contemporary Epistemology, Ralph Baergen maintains that:

...internalists perform their epistemic evaluations from the first-person point of view, taking into account only that which was available to the subject when the belief being evaluated was formed (or maintained, or revised), while externalists evaluate from the point of view of a fully informed observer—what one might call the third person point of view.[16]

In turning to a more expansive treatment of epistemic justification (centered upon the case of perceptual knowledge), this distinction will become clearer. Thus by examining Alston’s The Reliability of Sense Perception, we will come to a yet clearer understanding of internalism and externalism, as well as of foundationalism and coherentism.

Alston’s “take” on epistemic justification will be complex, and to introduce it I will use a citation from Catherine Elgin’s Considered Judgment:

in “Two Concepts of Rules,” Rawls distinguishes between justification within a practice and justification of a practice. Justification within a practice conforms to the pure procedural model....Justification of a practice, however, appeals to considerations of utility. Although corroboration of antecedently acceptable conclusions and validation of antecedently acceptable methods contribute nothing to justification within a cognitive practice, they are not idle wheels....They provide reason to believe that the practice’s standards, rules, and methods are adequate, that the practice as constituted promotes its epistemic ends.[17]

Our prospects may look bleak. The failure of foundationalism deprives us of the resources for identifying a unique goal toward which cognition naturally aims. So practices cannot be justified by their promoting activities or producing results that realize such a goal. If we attempt to justify our practices by embedding them in a metapractice, the problem simply recurs. Unless the metapractice is itself justified, it is powerless to justify the practices it authorizes.

There is a third alternative: to construct our epistemic ends out of our actual interests and goals in theorizing. To be sure, we cannot identify the objectives we happen to pursue with the legitimate ends of inquiry. That would be to confuse the valued with the valuable....Still, by adjudicating among them, revising and amending as required, we may bring them into reflective equilibrium. And a system of tenable commitments in reflective equilibrium defines, I suggest, a worthy epistemic goal.[18]

Justification on this account is a twofold process. Cognitive practices are justified by their epistemic utility; they are answerable to standards they need not themselves acknowledge and may be faulted for failing to achieve ends to which they do not aspire. Individual methods, inferences, conclusions, and the like, are justified by the standards of the practices to which they belong, by rules they recognize as binding on them—but only if those practices in turn are justified.[19]

To see her point, however, we must turn to Alston’s book.