The "standard analysis" of knowledge in recent philosophy has been of knowledge as "justified true belief." In a "belief," someone mentally assents to some proposition; if this belief is "true," then there is some fact about reality that makes the proposition true; and then if the belief is "justified," it means that the believer has some evidence or good reason for the belief.

Such a definition of knowledge goes back to Plato, where it is extensively discussed in his dialogue, the Theaetetus [201c-210b]. Whenever "justified true belief" is mentioned, credit is usually given to Plato for it. However, what is rare in that respect is acknowledgement that Plato rejected this definition of knowledge [note]. We see the moment this occurs in the Theaetetus:

SOCRATES:  So, apparently, to the question, What is knowledge [epistémê]? our definition will reply: 'Correct belief [dóxa orthé] together with knowledge [epistémê] of a differentness'; for, according to it, 'adding an account [lógos]' will come to that.

THEAETETUS:  So it seems.

SOCRATES:  Yes; and when we are inquiring after the nature of knowledge, nothing could be sillier [euéthes] than to say that it is correct belief together with a knowledge of differentness or of anything whatever.

So, Theaetetus, neither perception, nor true belief, nor the addition of an 'account' to true belief can be knowledge.

THEAETETUS:  Apparently not. [210a-b, from Plato's Theory of Knowledge, Francis M. Cornford, Library of the Liberal Arts, 1957, p.161 -- Greek terms added]

What happens in Plato's discussion is that for the "account" (or for us, the justification) to be strong enough to qualify the true belief as knowledge, the "account" must itself already be knowledge, producing a circular definition. No modern account of "justified true belief" is immune to this problem, and disregarding Plato's discussion does not help in the least. Indeed, anyone familiar with Plato's epistemology knows that he cannot accept "justified true belief," since he believes that we have knowledge that we are born with and are unaware of, i.e. it exists first without the presence of any corresponding belief [note].

One complication with "justified true belief" is that justification can mean importantly different things. I have discussed elsewhere the difference between "reasons for belief" (rationes credentis) and "reasons for truth" (rationes veritatis). Reasons for truth involve actual evidence that connects the believer to the matters of fact that determine truth. On the other hand, in much of life, few people have the time or expertise to know the actual evidence in all matters. We may have good reasons for believing much of what we do, even though we may be unaware of most of the logically relevant evidence. Thus, someone may read an entire encyclopaedia and come away with a great deal of knowledge, but in most of the matters covered by the encyclopaedia the articles have little chance to go over the grounds or background of the subjects very thoroughly. Nevertheless, we say that the reader ends up with a great deal of knowledge. If we ask him why he believes such and such a thing, all he can do is cite the authority of the encyclopaedia. That would be a logically fallacy if either of us believed that this proved the truth of his belief, but it is no fallacy if we understand the limitations of the appeal. We trust that the encyclopaedia has solicited experts in their fields, and that these authorities are themselves familiar with the reasons for truth of the matters they assert. On a humbler level, we credit the authority of honest persons who are eyewitnesses to certain events. Witnesses can be mistaken, but there may be no better evidence, and no other way to recover the truth, than their testimony. As Will Rogers (1879-1935) said, "All I know is what I read in the papers." A terrifying prospect for reliable knowledge.

In most matters, we have little choice but to rely on authority. This also means, in turn, that we actually have no way of knowing if our knowledge is really knowledge or not. We could have lived for two thousand years believing that the universe revolved around the earth, and been justified, but then end up told that we were wrong. Indeed, there is now little awareness of what the rationes veritatis were for a stationary earth. Even most astronomers probably believe that it was hybris, the ego-centrism of humanity, or the stupidity of religious superstition that dictated the centrality of the earth. It is now pretty much forgotten that the matter was determined, neither by astronomy nor by arrogance, but by physics:  ancient and mediaeval physics did not have the principle of intertia, and it was thought that a constant force was needed to keep at object moving at a constant velocity. Since a force could be felt, and we felt nothing of the sort on the earth, the earth consequently was not moving. Only Galileo reconceived the whole matter and introduced the really very counter-intuitive notion that an object in motion would tend to continue in motion with an unchanging velocity. Nothing like this is seen in ordinary life -- it is postulated in defiance of the evidence of the senses -- when my car breaks down (and reaches the bottom of the hill), it is going to come to a stop. In this case, where people's knowledge becomes deficient is not in current reasons for modern theories in astronomy or physics, but in what the reasons used to be for the old theories, i.e. we have a failure of historical knowledge about those old reasons.

The existence of mere "reasons for belief" highlights the general difficulty with justification. If we say that knowledge requires that our "justified beliefs" be true, it is quite reasonable to ask, "How do we know what is true?" -- if not Pilate's "What is Truth?" After all, "belief" and "justification" are all about mental contents, but objective truth is not about a mental content, but about a matter of fact in reality. To get at the truth, it seems like somehow we must get around things which are merely our mental contents and catch of glimpse of reality. Of course, catching a glimpse simply means that we get a new mental content. Indeed, the connection between justification and truth raises the old Cartesian Problem of Knowledge. Its form in Descartes himself was a question about perception, but the relationship between justification and truth raises the issue quite generally for all knowledge. If the "justification" of knowledge is not to be hopelessly disconnected from truth, and so from knowledge, it must indeed be connected to truth.

It is the nature of that connection that creates the real Problem. The only direct connection that Descartes could imagine between us and external objects was a causal one. Because this involves a genuine physical link between belief, perception, and the physical things they concern, it is still a popular possibility. The difficulty, already evident in Descartes, was that causes are only logically sufficient to their effects, i.e. a physical cause is not necessarily the only thing that can produce the effect. Thus, a perception may be caused, not by its represented object, but by dreaming, by fever -- as a hallucination -- by insanity, drugs, sensory deprivation, brain damage, subliminal messages from corporate sponsors, extra-terrestrial mind control, or -- the favorite of Descartes -- the Deceiving Demon. This has often struck people as funny, but we know that hallucinations are possible, and the principle is always the same. As Hume would argue independently, an effect need not resemble its cause in the least; and this makes it very hard to understand how causality can convey into the mind some veridical representation of the world. Indeed, it is a pretty hopeless proposition, which probably accounts for the recently popularity of theories like Wittgenstein's, that justification conforms to an internal "language game" rather than to an external reality. That languages are then dignified with an objective and factual existence denied to anything else, however, prevents this kind of thing from being a coherent theory.

A disconnect between justification and the external facts seems to be behind the counterexamples by which Edmund Gettier impeached "justified true belief" in 1963. Some of the counterexamples involve inferences that turn out to be true by accident. Thus, in one of them:

Smith has applied for a job, but has a justified belief that "Jones will get the job." He also has a justified belief that "Jones has 10 coins in his pocket". Smith therefore (justifiably) concludes that "the man who will get the job has 10 coins in his pocket."

In fact, Jones does not get the job. Instead, Smith does. However, as it happens, Smith also had 10 coins in his pocket. So his belief that "the man who will get the job has 10 coins in his pocket" was justified and true. But it does not appear to be knowledge.

Here the man who gets the job and has 10 coins in his pocket is actually not the man to whom the belief originally referred. One thing we can say is that the inference was "unsound" because, even if logically valid, it turns out to have a false premise. But that does not appear to be the real defect. If a proper justification connects a representation to its objects, then this kind of inferential justification has abstracted out the identity of the object, i.e. that the belief was about Jones, not Smith. It is an accident, based on an unsound argument and a shift between individual subjects, that the inference turned out to be true.

Another example goes like this:

Smith walks into a room and seems to see Jones in it; he immediately forms the justified belief, "Jones is in the room." But in fact, it is not Jones that Smith saw; it was a life-size replica propped in Jones's chair. Nevertheless, Jones is in the room; he is just hiding under the desk reading comic books while his replica makes it seem as though he is in. So Smith's belief is not only justified but also true.

Again (in what sounds like a Dilbert cartoon), there is a defect in the relation between the belief and its object, since the replica is mistaken for Jones, whose nearby presence nevertheless makes the belief accidentally true. As with the inferential example, the justification begins with something false, namely the misidentification of the replica as the real person. This severs the link to truth, leaving it as an accidental matter that the belief ends up being truth nevertheless [note].

What is required of the "justification" then is that it contain some kind of cognitive connection between belief and truth that is not accidental or based on some falsehood or mistake. This works even with a ratio credentis, whose force consists in its reference to something or someone that contains the relevant ratio veritatis. The Gettier counterexamples illuminate two things:  (1) Since knowledge is fallible and corrigible, a justification may seem good enough at some point, and we may be in possession of something that, as far as we can tell, qualifies as knowledge at the time. We accept it as knowledge, as we did with the Ptolematic universe, until the justification falls through and better evidence appears. (2) We would never accept anything as knowledge unless we believed that the justification had some connection to the fact of the matter that determines truth. Since we can never peek around our representations and grasp reality in itself, the correction of our knowledge means that what we believe connects us to the fact of the matter changes. The Problem of Knowledge, then, is that we ask how a justification can connect us to the fact of the matter at all. This remains the essence of the issue, and Modern Philosophy squirms around uncomfortably trying to make a connection between justification and something that will determine and ground it.

Theories that try to accomplish this fall broadly into two categories:  (1) the Externalist, which, supposing that reality is "out there," figure that there must be a cognitive connection between internal states, of belief and justification, and external reality. These can be causal connections, which founder on the Cartesian paradoxes, or they can be more broadly conceived, as with Aristotelians hoping to use something other than the Aristotelian concept of efficient causation, which is usually what is meant by "causation" now. All Externalist theories of cognition have the difficulty that, by definition, external reality is not as such accessible to us. It is outside, and whatever is conveyed in to us from it, we have no way of comparing this with the original and ascertaining that our representations are conformable to their source. Aristotelians have no difficulty with this, since they are comfortable with using their metaphysics as the basis of their epistemology ("ontological priority" in philosophy) and are not troubled by modern, critical questions (from Descartes and Kant) about how we can know the metaphysics ("epistemological priority" in philosophy) -- though in spirit the questions about knowledge go back to Socrates and Plato and were simply neglected by Aristotle. Externalist theories tend to go with a "correspondence" theory of truth, that representation must correspond to the nature of things or the fact of the matter.

The Externalist theories may be contrasted with (2) the Internalist, where our beliefs or justifications must simply connect up to something that is already present in our minds. In response to Descartes, his Rationalist successors, Spinoza and Leibniz, both used kinds of Internalism, with Spinoza that there is nothing external to God, which, being the only thing that exists, is us, and with Leibniz that all our knowledge is the internal unfolding of the representation contained in a monad, our own individual substance, which has possessed all its knowledge from eternity and has no real causal relationship to anything external (except God). As bizarre as these theories get, Descartes and then the first Empiricist, John Locke, look like blessed common sense in comparison. However, Locke, as a kind of Externalist, falls to the critique of Berkeley and Hume, who cannot see, given an Empiricist epistemology, how our possession of "ideas" of perception and experience give us any evidence of the existence of external metaphysical substances. Where Berkeley rejects the existence of matter, Hume drifts into a thorough ontological Skepticism, with matter, souls, and God all rejected as insufficiently justified by experience. Hume can then follow something like a "coherence" theory of truth, that our "ideas" need merely be made consistently with our "impressions," our original experiences, to maintain truth. Note that for both Rationalists and Empiricists, something that would count as adequate justification of a belief, as in Plato, must already be an item of knowledge.

The problem with the sort of Internalist theories of Rationalists and Empiricists is that they seem bizarre and autistic. The metaphysics of Spinoza or Leibniz is a lot to swallow, and Leibniz still needs God to ensure the "pre-established harmony" in the representations of different monads. Hume's Skepticism leaves us ignorant of reality. As Hume says of the real workings of nature:  "These ultimate springs and principles are totally shut up from human curiosity and enquiry." Yet approaches like this are now quite popular in philosophy, where truth is commonly thought to be determined, not by any external reality, but by the structure of language (Wittgenstein), by "paradigms" (deconstruction), or by power (either in a Marxist or Nietzschean form -- or in popular combinations of the two, now grandly called "Theory" by people otherwise unfamiliar with philosophy). Most of these versions end up being relativistic or just plain nihilistic. It is said that Michel Foucault, dying of AIDS, did not believe that AIDS even existed. It was just an ideology created by some hostile power structure. If the story is true, it is an extreme case of an Internalist epistemology refusing to believe that an external reality even existed. This did not prevent the reality from exercising its own irresistible power [note].

If Externalist and Internalist views of justification both have their own hopeless problems and paradoxes, what is there left? What is left, as it happens, is the approach of Kant:  The phenomenal world is empirically real. Here the world that we know is both internal and external. It is internal because it is a content of consciousness. It is external because the ways in which we ordinarily judge things to be external, as being independent substances and existing in space, are forms that exist internally to phenomena. At one level, the theory is that those forms are there because we put them there ourselves; but then at another level, where the mental activity of "synthesis" generates consciousness itself, we are not consciously aware or in control of the process. We see the result. Nevertheless, it is common for people to see Kant as offering some version of Berkeley's subjectivism or of a Cartesian externalism, where phenomena are not really external objects, the things-in-themselves. However, the key to this is that things-in-themselves are actually neither internal nor external. Although Kant does not say so, these must be regarded as "categories" also, answering the objections from Spinoza to Hume that the concept of the external cannot actually be conveyed from something absolutely external. Indeed.

The characteristic of Kantian theory that overcomes the duality of internal and external I have reformulated as "ontological undecidability," that there is not sufficient reason to regard the world as either merely phenomenal or absolutely separate from our representation. As it happens, we get something like this today in Quantum Mechanics, where the presence of an observer contributes to objective characteristics of what is observed -- for which I have formulated a Kantian Quantum Mechanics. What is then behind the dualism of internal and external is the dualism of negative and positive transcendence. While that is important for the ultimate view of things-in-themselves, it is less important for a theory of knowledge. That is settled by the Friesian theory of immediate knowledge. What we get there is the view that ordinary perceptual knowledge is justified because in our perceptions we are directly acquainted with phenomenal objects. The objective ground of truth is present in consciousness, as the phenomenal objects themselves.

A Kantian theory has always been paradoxical and is likely to remain so. Nevertheless, it benefits strongly from restoring an element of common sense. We are no longer isolated in a Cartesian theater, windowless monad, or an isolated epistemic autism. The world and its objects are directly accessible to us, without actually contradicting the understanding of Descartes or Hume that perceptions are artifacts of consciousness. At the same time, since phenomena are more than physical objects, the reality of abstract objects, like those of mathematics, is not compromised by mere material existence. By the same token, this allows, as Hume did not, that matters of value, as themselves abstract objects, have equal standing with empirical matters of fact. Indeed, these virtues are already evident in Schopenhauer, for whom abstract objects and matters of value are phenomenally real, lacking only the application of Schopenhauer's version of the "principle of sufficient reason," especially space and time. This is all possible because, unlike in materialism, the world is as much part of consciousness as it is separate from it. Characteristics of conscious existence, like abstraction or value, are not merely subjective or internal but part of the complex of phenomenal objectivity. Indeed, we get to restore another element of common sense. Beauty is something that we perceive in objects, even though everyone from philosophers to the man on the street regard it as something subjective and "in the eye of the beholder."

The dictionaries define beauty as something intrinsic to the object (its color, form, and other qualities) or simply as the pleasure an object evokes in the beholder (The philosopher Santayana called beauty "pleasure objectified.") [Nancy Etcoff, Survival of the Prettiest, Doubleday, 1999, p.8]

This paradox draws us back again to Plato:

Now beauty, as we said, shone bright among those visions, and in this world below we apprehend it through the clearest of our senses, clear and resplendent. For sight is the keenest of the physical senses, though wisdom is not seen by it -- how passionate would be our desire for it, if such a clear image of wisdom were granted as would come through sight -- and the same is true of the other beloved objects; but beauty alone has this privilege, to be most clearly seen and most lovely of them all. [Plato, Phaedrus, 250d, after R. Hackford, Plato's Phaedrus, Library of the Liberal Arts, 1952, p. 93, and the Loeb Classical Library, Euthryphro Apology Crito Phaedo Phaedrus, Harvard University Press, 1914-1966, p. 485]

To Plato, seeing beauty in the world draws us to the understanding and contemplation of the World of Forms, where beauty (and wisdom) originates. While the mature Kant did not have an aesthetic realism, the younger Kant of the Observations on the Feeling of the Beautiful and Sublime (1764) did, and this realism was restored by Schopenhauer and Fries. What it means is that, just as all matters of value are real parts of the world (which the mature Kant believed for morality), the beauty that we see in objects, subjective as it must often seem, nevertheless is really there. It bespeaks something, transcendence, that both Plato and Kant understood quite well but that is missing from most recent philosophy and from the opinion of most modern intellectuals.

The value of "justified true belief" is the search that it initiates into what it is that is indeed going to justify belief, and what this has to do with what is true. It is not too surprising that such a definition found little favor with major historical philosophers, from Plato to Hume. That the very philosopher who suggested it, Plato, already rejected it should have been telling people something during the popularity of the formula in recent philosophy. The difficulty, indeed, was always with that little bridge that needed to connect the justification to the ground of truth. That is what earlier philosophers instinctively worried about; that is the sharp nettle that Descartes grasped to bloody effect; and that is the rock upon which most post-Kantian philosophers have foundered, either ignoring the depth of the problem (as in analytic philosophy) or veering off into strange solipsistic, anti-realist exercises (as in linguistic and "post-modern" developments). The Kantian approach, indeed, is a bit strange itself -- as, truth be told, Plato's was already -- but it repays the effort.


Home Page

Copyright (c) 2007, 2008 Kelley L. Ross, Ph.D. All Rights Reserved

Knowledge, Note 1

Thus, as recently as 2006, Paul Boghossian of New York University says in his generally sensible book:

Suppose, then, that as visual observations of the earth from space seem decisively to confirm, this planet we live on is in fact round. Then our belief that it is round is both justified and true; according to the standard, widely accepted Platonic definition of knowledge, then, our belief counts as knowledge. [Fear of Knowledge, Against Relativism and Constructivism, Clarendon Press, Oxford, p.15 -- boldface added]

One would never know from such a discussion that the real "Platonic" definition of knowledge, according to Plato that is, rejected this account.

Return to text

Knowledge, Note 2;
Knowledge in Plato's Meno

Plato introduces something similar to "justified true belief" of the Theaetetus in an earlier dialogue, the Meno. He appears there to endorse the idea, but we also find the seed of the later problem. Thus, Meno has asked why knowledge [epistémê] is considered to be better than true opinion or belief [orthè dóxa]. The answer is:

SOCRATES:  For true opinions, as long as they remain, are a fine thing and all they do is good, but they are not willing to remain long, and they escape from a man's mind, so that they are not worth much until one ties them down by (giving) an account of the reason why [aitías logismós -- an account of explanations or causes]. And that, Meno my friend, is recollection [anámnêsis], as we previously agreed. After they are tied down, in the first place they become knowledge, and then they remain in place. [98a, G.M.A. Grube, "Meno," Five Dialogues, Hackett Publishing, 1981, 1986, p.86]

Thus, true belief becomes knowledge by the addition of Recollection, which is sufficient as a reason or account of the belief. Earlier, however, we got Recollection introduced as Plato's theory of knowledge:  "for searching [zêteîn -- inquiring, investigating] and learning [manthánein] are, as a whole, recollection" [ibid., 81d, p.70]. If the "account of the reason why" is itself knowedge, then certainly true belief becomes knowledge with its addition, but if we actually wanted this to be a definition, it would be circular.

Here, however, Plato does not offer this as a definition, simply as an explanation of why knowledge is prized higher than true opinion. In those terms, it is also rather odd. My experience is that opinions, true or otherwise, do not "escape from a man's mind." They usually are held rather tenaciously, even in the face of evidence and reason, usually just because they are one's own opinions rather than someone else's. We even have a special expression for those with stubborn opinions, "pig headed."

The suggestion that Meno made earlier seems better, that "the man who has knowledge will always succeed, whereas he who has true opinion will only succeed at times" [97c]. Plato has Socrates disagree with this, saying "Will he who has the right opinion not always succeed, as long as his opinion is right?" Well, yes, as long as his opinion is right. The problem is with how one acquires a true opinion. As Socrates says in the Apology and Plato affirms in the Meno, someone with a true opinion has no understanding of why their opinion is true. In this, they are like poets, soothsayers, and prophets, who say many true things by inspiration but cannot explain the truth of what they are given. In the Meno, this is used to explain how those who have virtue, as true opinion, are unable to teach their own virtue. In the end, they are said to have virtue "as a gift from the gods" [100b].

In these terms, Meno's point can be well taken. Those who have true opinion will not always succeed because when they move out of their familiar circumstances and need to form an opinion about some unfamiliar objects, they will be lacking what Socrates in the Euthyphro calls the "model" [paradígma] that is provided by the idéa which is possessed by all things of one kind (in that case, piety, cf. Euthyphro 6e) and makes them what they are. They will have no guide to what they need and will likely err in judging the case. A striking example of this is given towards the end of the Republic. In the Myth of Er, we have a story of souls, having obtained rewards or punishment for their previous lives, choosing their next lives. One soul, returning from rewards in heaven, who had been virtuous "by habit without philosophy," carelessly chooses the life of a tyrant, only to realize, too late, that he will endure many horrors and an evil fate [Republic 619b-d]. He was lacking the "model" provided by knowledge to judge good or bad in such unfamiliar circumstances.

This passage could stand as a sort of pre-critique of Aristotle's later thesis that morality consists entirely of good habits. In the Platonic context, however, it gives us an excellent case of the limitations of true opinion. The person who is virtuous out of habit has true opinions and can judge correctly of those objects about which his opinions have been habitually concerned -- both where he acquired the habits and where they are customarily applied. Remove him to where his habits are no guide, and he is in trouble. This would be the right answer to Meno's question about why knowledge is prized more highly than true opinion. In the context of the Meno, we instead may get the strange statement that opinions "escape from a man's mind" just because Plato wants a device to return to the concept of Recollection, which had been introduced in the Meno but then neglected in the discussion after the famous example of the geometry construction using Socratic Method.

However this works, it remains the case that "justified true belief" is not offered as a definition of knowledge in the Meno, and if it were, it would clearly be circular, just as Plato concludes in regard to the definition in the Theaetetus.

Return to "Knowledge" text


Home Page

Copyright (c) 2007, 2010 Kelley L. Ross, Ph.D. All Rights Reserved

Knowledge, Note 3

The interest and controversy generated by the celebrated "Gettier Counterexamples" raises questions about the seriousness and even competence of modern academic philosophy. When the "standard analysis" of knowledge had been based on a theory advanced by Plato, but which had already been refuted and rejected by Plato himself, for cogent reasons, how in the world could it have become the "standard analysis" among our contemporary professional philosophers? How could anyone have been surprised when the counterexamples exposed its problems? Had these people actually read the Theaetetus? Whitehead's observation that Western philosophy consists of a series of footnotes to Plato is nowhere better confirmed in this case, with the disturbing addendum that the footnoters may not even have understood Plato, much less risen to anything like his level as a philosopher.

Return to text

Knowledge, Note 4

Boghossian [op.cit.] cites the French sociologist Bruno Latour, who refused to believe the recent revelation that Ramesses II had died of tuberculosis:  "How could he pass away due to a bacillus discovered by Robert Koch in 1882?" ["Ramses II est-il mort de la tuberculose?" in La Recherche, 307, March 1998, pp.84-85]. Latour has decided that to discover something is to create it. This is so nonsensical that I doubt he governed his ordinary life according to such a principle. If his car turned up missing, I expect he would believe that it had been stolen, and that it had been stolen before his discovery that it had disappeared. The discovery itself didn't cause anything, except his own concerns and inferences. All too often, this seems to be about the level of epistemological sophistication we can expect from sociology.

Return to text