Knowledge is more than true belief. True beliefs that come from just lucky guesswork do not qualify as knowledge. Knowledge requires that the satisfaction of its belief condition be appropriately related to the satisfaction of its truth condition. In other words, knowledge requires a justification condition. A knower must have an adequate indication that a known proposition is true. A traditional view, suggested by Plato (c. 428–348 or 347 b.c.e.) and Immanuel Kant (1724–1804), proposes that this adequate indication consists of evidence indicating that a proposition is true. This view requires that true beliefs qualifying as knowledge be based on justifying evidence, or reasons.
When reasons are beliefs, we have inferential justification: one belief is justified on the basis of another belief. How is the latter, supporting belief itself justified? Is another supporting belief always needed? According to foundationalism, another supporting belief is not always needed.
Two-Tier Structure of Justification
Foundationalism acknowledges a two-tier structure in justification: Some justification is noninferential (that is, foundational), and all other justification is inferential (that is, nonfoundational) in that it derives ultimately from foundational justification. This view was proposed by Aristotle (384–322 b.c.e.), in Posterior Analytics, as a view about knowledge. It emerges too in René Descartes's Meditationes de prima philosophia (1641; Meditations on first philosophy), where it joins with the assumption that foundations of knowledge must be certain. Foundationalism is represented in varying forms in the writings of John Locke (1632–1704), Kant, Bertrand Russell (1872–1970), Clarence Irving Lewis (1883–1964), and Roderick M. Chisholm (1916–1999), among others. Foundationalism about evidence and justification explains a belief's (or a proposition's) having justification for a person; it does not explain one's showing that a belief has justification or is true.
Versions of foundationalism vary in explaining noninferential (foundational) justification, and in explaining how justification transmits from foundational beliefs to nonfoundational beliefs. Some philosophers have assumed that foundational beliefs must be certain (that is, indubitable or infallible). This leads to radical foundationalism, which requires that foundational beliefs be certain and that such beliefs guarantee the certainty or the truth of supported nonfoundational beliefs. Radical foundationalism faces two problems. First, very few (if any) empirical beliefs are certain. Second, beliefs that might be candidates for certainty (for instance, the belief that I am thinking) are not informative enough to guarantee the certainty or the truth of specific inferential beliefs about the external world (for example, beliefs of the sciences).
Modest foundationalism states that foundational beliefs need not possess or yield certainty and need not guarantee the truth of justified nonfoundational beliefs. A noninferentially justified, foundational belief has justification that does not derive from other beliefs, even if the causal basis of foundational beliefs includes other beliefs. Modest foundationalists have offered three influential approaches to noninferential (foundational) justification: (1) self-justification, (2) justification by nonbelief, nonpropositional experiences, and (3) justification by a reliable nonbelief origin of a belief.
Recent proponents of self-justification include Chisholm and C. J. Ducasse (1881–1969). They held that a foundational belief can justify itself, apart from evidential support from something else. Proponents of foundational justification by nonbelief experiences dissent. They hold, following Lewis, that foundational perceptual beliefs can be justified by nonbelief sensory or perceptual experiences (for example, my nonbelief experience involving seeming to see snow falling here) that either make true, are best explained by, or otherwise support those foundational beliefs (for example, the belief that snow is falling here). Advocates of foundational justification by reliable origins hold that noninferential justification depends on nonbelief belief-forming processes (such as perception, memory, and introspection) that are truth conducive, that is, they tend to produce true rather than false beliefs. Such reliabilism invokes as a justifier the reliability of a belief's nonbelief origin, whereas the previous view invokes the particular sensory experiences underlying a foundational belief. Proponents of modest foundationalism typically agree that noninferential justification for a belief can be defeated upon the acquisition of new justified beliefs, at least in most cases.
Wilfrid Sellars and Laurence BonJour have argued against non-inferential justification, on the ground that one is justified in holding a belief only if one has good reason to think that the belief is true. This ground, they suggest, implies that the justification of an alleged foundational belief will depend on an argument of this type:
- My foundational belief that P has feature F.
- Beliefs having feature F are likely to be true.
- Hence, my foundational belief that P is likely to be true.
If the justification of alleged foundational beliefs depends on such an argument, these beliefs will not actually be foundational. Their justification will depend on the justification of other beliefs, namely, the beliefs represented by premises (1) and (2).
The justification of one's belief that P, however, does not require justified belief that premises (1) and (2) are true. Given such a requirement, I will be justified in believing that P only if I am justified in believing that my belief that P has feature F. Given these requirements, I will be justified in believing that (1) my belief that P has F only if I am justified in believing an additional proposition: that (4) my belief that (1) has F. In that case, we have no plausible way to preclude that similar requirements apply to the latter proposition—namely, (4)—and to each of the ensuing infinity of required justified beliefs. We do not have, however, the required infinity of increasingly complex beliefs. The conclusion is that if the justification for a belief must be accessible to a person holding the belief, that accessibility should not itself be regarded as requiring further justified belief. Some critics of foundationalism have missed this conclusion.
Current debates over internalism and externalism regarding justification concern what sort of access, if any, one must have to the justification for one's justified beliefs. Internalism incorporates an accessibility requirement of some sort on what provides justification, whereas externalism does not. Debates about internalism and externalism are prominent in contemporary epistemology and show no sign of easy resolution.
Proponents of foundationalism must specify the exact conditions for noninferential justification. They must also specify the exact conditions for the transmission of justification from foundational beliefs to inferentially justified (nonfoundational) beliefs. Modest foundationalism, as noted above, allows for nondeductive, merely probabilistic connections that transfer justification from foundational to nonfoundational belief. Proponents of modest foundationalism have not, however, reached consensus on the exact nature of such connections. Some proponents of modest foundationalism hold that "inference to a best explanation" underlies the transmission of justification from foundational to nonfoundational beliefs in many cases. The belief, for example, that snow is falling here can, in certain circumstances, provide a best explanation of various foundational beliefs about one's perceptual experiences (including one's seeming to see snow falling here). This, however, is a controversial matter among contemporary proponents of foundationalism.
Versions of foundationalism that restrict noninferential justification to subjective beliefs about what one seems to see, hear, feel, smell, and taste confront a special problem. These versions must explain how such subjective beliefs can yield justification for beliefs about conceiving independent physical objects (for instance, beliefs about household objects). Such subjective beliefs do not logically entail beliefs about physical objects. Given that extensive hallucination is always possible, it is possible that one's subjective beliefs are true while the pertinent beliefs about physical objects are false. This consideration challenges any version of foundationalism that includes the view that statements about physical objects can be translated, without loss of meaning, into logically equivalent statements merely about subjective states characterized by subjective beliefs. Perhaps a proponent of foundationalism will identify some nondeductive relations that explain how subjective beliefs can confer justification on beliefs about physical objects. Currently, however, no set of such relations has attracted consensus acceptance from supporters of foundationalism. Some versions of foundationalism allow for the noninferential justification of beliefs about physical objects, thereby avoiding the problem at hand.
In sum, then, foundationalism offers an influential response to questions about the nature of inferential justification. It is, however, a work in progress given its controversial features.
See also Knowledge ; Logic .
Alston, William P. Epistemic Justification. Ithaca, N.Y.: Cornell University Press, 1989.
Audi, Robert. The Structure of Justification. Cambridge, U.K.: Cambridge University Press, 1993.
BonJour, Laurence. The Structure of Empirical Knowledge. Cambridge, Mass.: Harvard University Press, 1985.
Chisholm, Roderick M. The Foundations of Knowing. Minneapolis: University of Minnesota Press, 1982.
Lewis, Clarence Irving. An Analysis of Knowledge and Valuation. La Salle, Ill.: Open Court, 1946.
Moser, Paul K. Knowledge and Evidence. Cambridge, U.K.: Cambridge University Press, 1989.
Paul K. Moser
Foundationalism seeks to discover whether there exist ultimate bases and foundations of human knowledge, and if so, to discover what these are. Though the term is newly minted, it designates an ancient and honorable concern among nearly all of the major philosophers of the Western tradition. The earliest example is Aristotle's compelling logical argument in the Posterior Analytics that, insofar as knowledge is based on evidence, and that evidence in turn is articulated in premises, and those premises rest on still other premises—eventually we will need to reach premises that are not just "prior and better known" than the conclusions drawn from them, but that are themselves not dependent upon any prior knowledge, being instead those that are self-evident or evident simply in terms of themselves alone.
Problematic status. Despite the tradition of persistent and perennial search for foundations that has seemingly characterized all previous philosophy, the inventors of the new term "foundationalism" want the designation to be taken as a form of criticism and even opprobrium. What could be worse, they ask in effect, than that spectacle of futility and irrelevance that has been exhibited by Western philosophers in their obsessive preoccupation with the foundations of human knowledge. Instead, philosophers should open themselves to the dawning of that new day, ushered in by the likes of Richard Rorty and his associates among the so-called Post-Analysts, in which we shall no longer worry over the question of whether our knowledge be with or without adequate foundations.
It must not be thought, however, that this apparent bland dismissal of concern with the foundations of knowledge is philosophically frivolous. Richard Bernstein, for example, has made the very telling point that even on its face any such enterprise as that of trying to discover the absolute foundations of any and all human knowledge is bound to be a futile attempt to find some fancied Archimedean point, from which all knowledge might be levered and suspended. Still more to the point have been the sorts of arguments developed by Alvin Plantinga and Nicholas Wolterstorff. Oversimplifying their arguments somewhat, what they are apparently concerned to point out is that any ultimate or absolute foundational truths upon which all the rest of our human knowledge would need to be based presumably would have to be either of two kinds. Either they would have to be truths of a purely formal kind, such as we are familiar with from logic and mathematics, and which do indeed seem to require no other evidence of their truth than just those very truths themselves. Or, as an alternative, they might be truths of simple observation or perception. For what other evidence can one have of such simple truths as "I am wearing shoes," or "There is a tree outside my window," other than that I just do perceive these things to be the case?
Accordingly, having established this much, Plantinga and Wolterstorff immediately proceed to give the coup de grace to any remaining and still struggling Foundationalists. For so far from the purely formal truths of logic and language being able to provide us with a foundational knowledge upon which we might be able to base our remaining knowledge of the world, it turns out that any and all purely formal truths, being no more than mere logical or linguistic truths, are in principle totally incapable of providing us with the slightest knowledge of the world or the way things are. Hence they are quite irrelevant for purposes of any foundationalism. And no less embarrassing is the case regarding ordinary observational or perceptual truths. Not only are such truths unable to provide us with unshakable foundations for our human knowledge, but it would appear, particularly from recent researches in the logic of science, that any and all perceptions and supposed data of observation are entirely relative to the conceptual schemes or frameworks in terms of which our perceptions and observations take place. Hence we have only to change our operative frameworks, and what we perceive will no longer be the same at all. And with that, the hope that experience and observation might provide us with an ultimate foundation for our knowledge of the world and of reality simply goes out the window.
What, then, is the consequence of such a demise and destruction of any and all forms of foundationalism, so far as human knowledge is concerned? Presumably, the answer that the Post-Analysts would give is that henceforth it shall be pragmatism, and not realism, that must be the order of the day, so far as human knowledge is concerned. Catholic thinkers have surely been on the defensive with regards to this sort of epistemological nihilism. T. Russman's Prospectus for the Triumph of Realism, however, is a skillful rehabilitation of foundationalism along Aristotelian and Thomistic lines.
See Also: realism.
Bibliography: r. j. bernstein, Beyond Objectivism and Relativism, (Philadelphia, Pa. 1983). a. plantinga and n. wolterstorff eds., Faith and Rationality (Notre Dame, Ind./London 1983). t. a. russman, A Prospectus for the Triumph of Realism (Macon, Ga. 1987).
[h. b. veatch]
The term foundationalism usually refers to theories about the structure of belief formation or belief justification. Beliefs may be formed or justified in one of two ways: non-inferentially (immediately) or inferentially (mediately). The division between "basic" and "nonbasic" beliefs is asymmetrical; nonbasic beliefs are formed or justified by appealing to basic beliefs, which are foundational. For classical foundationalists, a basic belief must be self-evident or incorrigible. Modest foundationalists argue that meeting other criteria, such as "evident to the senses," may also qualify a belief as basic; further, basic beliefs can be defeasible. Foundationalists in the science-religion dialogue often focus on defending the propriety of basic beliefs and inferences from them.
See also Nonfoundationalism; Postfoundationalism
f. leron shults