Epistemology, History of

views updated


Epistemology, or the theory of knowledge, is that branch of philosophy which is concerned with the nature and scope of knowledge, its presuppositions and basis, and the general reliability of claims to knowledge. The pre-Socratic philosophers, the first philosophers in the Western tradition, did not give any fundamental attention to this branch of philosophy, for they were primarily concerned with the nature and possibility of change. They took it for granted that knowledge of nature was possible, although some of them suggested that knowledge of the structure of reality could better come from some sources than from others. Thus, Heraclitus emphasized the use of the senses, and Parmenides in effect stressed the role of reason. But none of them doubted that knowledge of reality was possible. It was not until the fifth century BCE that such doubts began to emerge, and the Sophists were chiefly responsible for them.

During the fifth century BCE human practices and institutions came under critical examination for the first time. Numerous things that had previously been thought to be part of nature were seen not to be. Thus, a general antithesis was drawn between nature and human convention or custom, and the question of where the line was to be drawn between them arose. The Sophists asked how much of what we think we know about nature is really an objective part of it and how much is contributed by the human mind. Indeed, do we have any knowledge of nature as it really is? Protagoras, for example, seems to have held, if Plato's report is to be believed, that everything is as it appears to a man, that appearances are the only reality. This is the meaning, or part of it, of his famous dictum "Man is the measure of all things." Gorgias was, if anything, more radical, claiming that there was no such thing as reality, that if there were, we could not know of it, and that even if we could know of it, we could not communicate our knowledge of it.

This general skepticism led to the beginning of epistemology as it has been traditionally knownthe attempt to justify the claim that knowledge is possible and to assess the part played by the senses and reason in the acquisition of knowledge. Before Plato, Democritus, the Greek atomist, had already drawn a distinction between those properties ordinarily attributed to things which, in his view, really belong to themfor example, size and shapeand those which, as he put it, are a matter of convention (nomos ), a function of the mindfor example, color. It was Plato, however, who can be said to be the real originator of epistemology, for he attempted to deal with the basic questions: What is knowledge? Where is knowledge generally found, and how much of what we ordinarily think we know is really knowledge? Do the senses provide knowledge? Can reason supply knowledge? What is the relation between knowledge and true belief?

The Nature of Epistemology

Epistemology differs from psychology in that it is not concerned with why people hold the beliefs that they do or with the ways in which they come to hold them. Psychologists can, in principle, give explanations of why people hold the beliefs they do, but they are not necessarily competent, nor is it their province, to say whether the beliefs are based on good grounds or whether they are sound. The answer to these questions must be sought from those who are experts within the branches of knowledge from which the beliefs are drawn. The mathematician can give the grounds for believing in the validity of Pythagoras's theorem, the physicist can give the grounds for believing in, say, the indeterminacy principle, and an ordinary but reliable witness can provide the grounds for believing in the occurrence of an accident. Normally, when the beliefs are true and the grounds sufficient, it is permissible to claim knowledge, and whether a particular truth can be said to be known may be determined by reference to the grounds that are appropriate to the field from which the truth is drawn. The epistemologist, however, is concerned not with whether or how we can be said to know some particular truth but with whether we are justified in claiming knowledge of some whole class of truths, or, indeed, whether knowledge is possible at all. The questions that he asks are therefore general in a way that questions asked within some one branch of knowledge are not.

role of skepticism

To characterize the questions asked by the epistemologist as extremely general is not, however, sufficient. Interest in very general questions of this sort and in the nature of general concepts is typical of philosophy as a whole. What distinguishes epistemology other than the fact that its interests center on the concept of knowledge? When a philosopher asks whether something is possible, the question must be set against the consideration that this thing may not be possible. It must be set against a general skepticism concerning the matter in question. To be called upon to justify the possibility of knowledge or of certain kinds of knowledge makes sense only on the supposition that it or they may not be possible. It is no coincidence that epistemology began in the context of a form of the Sophists' general skepticism about knowledge, for until such doubts had been raised, the possibility of knowledge was bound to be taken for granted. Once the doubts had been raised, they had to be answered. How they were to be answered depended on the nature of the doubts and on the degree to which any particular philosopher was susceptible to them.

Views on the nature of knowledge

Perhaps the most characteristic form of skepticism about knowledge has been based upon the premise that we ought not to claim knowledge about anything unless we are absolutely sure about it, unless there is no possibility of our being wrong. Once given this, it is possible to point out that it is at least logically possible to be wrong about most, if not all, the things that we ordinarily claim to know. Philosophers who have been impressed by this argument have generally tried to show that there are at least some things that we can claim to know, about which we cannot be wrong. Even so, most of the things that we normally think we know cannot, on this view, be said to be known at all. This consequence can be mitigated, although not removed, if it can be shown that the things accepted as known in the strict sense give reasons for believing the things that we normally take ourselves to know. Philosophers who have taken this course have differed both on what this "certain knowledge" is and on how it is connected with what we ordinarily suppose ourselves to know. Rationalists have generally attempted to show that the primary truths that constitute this certain knowledge are related to other truths somewhat as the axioms of a formal or geometrical system are related to the theorems.

Empiricists, on the other hand, have taken the view that the truths which constitute ordinary knowledge can be constructed out of the primary truths, as a building is built up from its foundations. They have differed again on the nature of the primary truths themselves. Rationalists have looked for them among the deliverances of reason, whereas empiricists have claimed that sense experience alone can provide such truths. Other philosophers have accepted part of the skeptical argument to the extent of denying the status of knowledge to some class of truths, reserving that status for some privileged class. Plato is a case in point in that for at least part of his life he maintained that sense experience never provides knowledge at all, this being reserved for a kind of awareness of or acquaintance with a world of quite distinct entities called Forms. In respect to the world of sense experience we have only opinion or supposition. This view of sense experience has not been uncommon among philosophers.

The concept of knowledge

A quite different way of dealing with the skeptical argument would be to question the initial premise that knowledge requires absolute certainty. One would not normally claim knowledge about something unless one were sure about it, but that is very different from asserting that a man could not be said to know something unless what he claimed to know was absolutely certain. Knowledge does not actually require this; it requires only that there be the best of grounds for what is claimed. To say this is to say something about what knowledge is, about the concept of knowledge itself. Hence, the skeptical arguments and the answers to them are not entirely independent of the conceptual question about the nature of knowledge. An understanding of the concept of knowledge is a prerequisite of embarking upon any attempt to answer other epistemological questions. Most philosophers have had something to say about the nature of knowledge, although many have taken its nature for granted. From this have stemmed a number of traditional epistemological difficulties.

Greek Philosophy


Plato (c. 428347 BCE) was influenced by several viewsthe moral teaching and philosophical practices of his master Socrates, the views of the Sophists already mentioned, and such pre-Socratic views about the nature of reality as the Heraclitean view that the sensible world is in a state of flux and the Eleatic view that reality is one and unchanging. He came to hold that reality cannot be changing or imperfect and that it must therefore consist of a world of Forms or Ideas independent of the sensible world. The exact reasons why Plato postulated a world of Forms are not altogether clear. But probably, as Aristotle says, he was influenced by Socrates' search for the essences of, for example, moral virtues. But because justice, for instance, is never found in this world in a proper and perfect form, he postulated its separate, ideal existence in order to provide the standard by which sensible instances of justice might be judged. The Forms might be known by reason, not by the senses. Whether there was a Form for every sensible particular is arguable, with respect to Plato's earlier philosophy. However, it is clear at any rate that by the time he came to write the Timaeus he believed this to be the case. Thus, in the first place the Forms were probably standards or exemplars of which sensible things were imperfect copies. At the same time, however, they functioned as universals, entities meant to explain how it is possible to think generally about things of one kind and how it becomes possible to attach a meaning to common names. The fact that the Forms had to fill both these roles gave rise to certain logical difficulties which Plato himself pointed out in the dialogue Parmenides and which he tried to deal with in the later dialogues. The Forms were always objects of knowledge and, in his earlier thought, the only objects of knowledge. Sensible things were, in his view, objects of opinion only.

Knowledge and true belief

The distinction between knowledge and true belief is first made by Plato in the Meno in the context of another Platonic epistemological doctrinethe theory of recollection (anamnesis). In this dialogue Socrates claims to elicit from a boy without instruction the answer to a geometrical problem. Since the problem is a geometrical one, it is one that cannot be answered by an appeal to the senses. Socrates therefore claims that he is enabling the boy to recollect something that he had known in a previous existence. He maintains that it is a doctrine well known to priests and poets that the soul has long ago experienced all things in its various existences. Hence, in a sense the soul knows all things, but because it has forgotten them, it has to be reminded of them. The example suggests that Plato may intend the doctrine, at least in part, as an explanation of our knowledge of a priori truths. Indeed, in the Phaedo he uses the doctrine to explain how we see things as instances of the Forms: Sensible things remind the soul of what it already knows and what it cannot know from sense experiencethe perfect Forms. The Meno does not claim so much. Indeed, Plato goes on to suggest that merely arriving at the right answer to a problem may not constitute knowledge but only true belief. Knowledge requires an ability to give the grounds (logoi) on which the answer rests. Nevertheless, Plato says, true belief may sometimes be, in its practical effects, as good as knowledge.

This distinction between knowledge and true belief is retained by Plato in later dialogues, although he is not always so charitable to belief as such. At the end of Book 5 of the Republic Plato begins a long argument, involving the famous similes of the sun, line, and cave to show how the soul may be drawn up by education to a true knowledge of the Forms, the final stage in the process that Plato calls dialectic. At the outset Plato makes a threefold distinction between knowledge, ignorance, and an intermediate state that he calls belief. Each of these states of mind has, he says, an object. The object of knowledge is what exists; the object of ignorance is, paradoxically, what does not exist; and the object of belief is that which is between existence and nonexistence. The last seems to be identified with the sensible world. Plato appears to think of these states of mind as forms of acquaintance with some kind of object, although the allocation of the objects in question is puzzling on any account. He rejects the identification of knowledge and belief on the grounds that belief is liable to error, whereas knowledge can never be. His conception of knowledge is thus a strict one.

Knowledge and sense perception

Plato's reasons for maintaining that we cannot have knowledge of the sensible world are that we should be in error if we attributed properties to sensible things absolutely. A thing is beautiful, heavy, or good only in relation to other things; hence, Plato concludes, nothing is really beautiful, heavy, or good except the standards of Beauty, heaviness, and Goodness themselves, and they cannot be sensible things. When we judge that sensible things are beautiful, heavy, or good, we are in error and cannot therefore be said to have knowledge.

There are two objections to this argument. First, if we realize that terms like "beautiful" are relative terms, we shall not necessarily be in error in saying that sensible things are beautiful; error will arise only if we are tempted to think that they are beautiful absolutely. Second, not all terms are relative in this way; "red," for example, is not. Perhaps Plato eventually took account of both of these points, but it seems clear that he tried to deal with the second by reference to the Heraclitean doctrine that the sensible world is in a state of flux. If this doctrine is true, it is impossible to attribute any property to sensible things unequivocally. This position is put forward in the Cratylus and most fully expressed in the Timaeus. It depends, of course, on whether the Heraclitean doctrine is true and free from unpalatable consequences. It is examined in Plato's most extensive consideration of knowledge, the Theaetetus, a dialogue probably written at about the same time as the Parmenides with its criticism of the traditional theory of Forms.

In the Theaetetus Socrates engages in a discussion with a young mathematician, Theaetetus, concerning the nature of knowledge. Theaetetus first answers the question "What is knowledge?" in a manner typical of the dialogues, by giving examples of knowledge, but is then prevailed upon to give the answer that knowledge is esthesis (perception or sensation; the Greek word from which it comes is ambiguous). This view is identified with that of Protagoras, the Sophist, to the effect that everything is what it seems to a man and that esthesis is of what is and must be infalliblethat is, what seems to a man is so and cannot be wrong. If knowledge is esthesis, it is thus an incorrigible awareness of something purely relative to the perceiver. Socrates enlarges on this view, indicating the extent to which our judgments about empirical things are relative in this way. Finally, the point of view is made absolutely general by the introduction of the Heraclitean doctrine of the flux. The joint effect of the doctrines is that all judgments about empirical things are relativethe classic Platonic point of view. This conclusion is reinforced by reference to various versions of what has become known as the argument from illusionan argument stating that we cannot be said to perceive the real properties of things because of the possibility of illusion or because of the causal aspects of perception itself.

Having expounded the view that all empirical judgments are relative, Socrates sets out to refute it. He refutes the pure Protagorean view that appearances are the only reality by pointing out that there are acknowledged experts in different fields of knowledge concerning objective phenomena; moreover, the Protagorean view is in essence self-refuting because no absolute truth can be claimed for it on its own terms. He refutes the Heraclitean doctrine of flux by claiming that if it were true, it would be impossible to say anything about anything, a consideration which he treats as a reductio ad absurdum. This section of the dialogue ends with the consideration that certain properties of thingsthe existence of things, their identity with themselves, and their difference from other thingsare ascribable to them only as a result of a judgment by the mind. Knowledge of the sensible world cannot therefore be a simple matter of having sense impressions, of esthesis in the sense specified; it must also involve judgment by the mind.

Possibility of false beliefs

In the next section Theaetetus suggests that knowledge consists of true judgment or belief. This suggestion is eventually refuted by the consideration that it is possible to believe something truly when one's grounds are insufficient. Most of the section, however, is given up to a discussion of false belief or judgment, for Socrates wonders whether this is possible. False belief cannot be a belief in what is not, for, as Parmenides showed, there is no such thing as what is not. On the other hand, if false belief consists in erroneously taking one thing for another, it is difficult to see how it is possible, for the believer must know one or both of the things in question if he is to be in the position of taking one for the other. However, if he knows at least one of the things, how can he mistakenly take it for the other? Socrates considers various possibilities, but the only cases in which he will allow the possibility of error are those in which a man fails to recognize something correctly because he has fitted the wrong sense impression to the wrong memory impression. (It is in the context of this discussion that Socrates introduces the analogy of the wax and the seal to illustrate the nature of the mind; the mind literally receives impressions from things outside it.) Since this account will not cope with judgments like 7 + 5 = 11, where there is no question of erroneous recognition of a sensible thing, Socrates introduces another analogy, likening the mind to an aviary, with pieces of knowledge represented by the birds. A man may know something in the sense that the idea of it is in his mind as a bird is in the aviary, but he may not have it at hand. That is, he may know it implicitly but not explicitly. Even here the original difficulties recur, however. How can he mistake an explicit piece of knowledge for something else?

The difficulties in this section of the dialogue depend upon construing errors of judgment as mistakes of identity and equating knowledge with direct awareness. The mistakes allowed by Plato are not strictly mistakes of identity but mistakes in matching one thing with another, the sense impression with the memory impression. Plato returns to the matter in the Sophist, where he tries to provide a new logical analysis of the nature of judgment. He distinguishes between judgments of identity and existential judgments and probably between both of these and subject-predicate judgments (as they were later called). Judgments generally assert that one thing participates in another (at least the latter being a Form), but judgments of identity and existence assert the participation of something in certain especially important Formsthose of sameness and existence. These Forms are two of a list of five to which Plato gave special attention, the others being difference, rest, and motion. To say that something is identical with another thing is to say that it participates in the Form of sameness in relation to that other thing. Mistakes of judgment can arise over whether something really participates in another thing, and to that extent the difficulties of the Theaetetus are resolved, although it may be questioned whether Plato has given a really adequate account of judgment.

True beliefs and logoi

In the last section of the Theaetetus, it is suggested that knowledge may consist of true belief together with the giving of a logos. It will be remembered that knowledge was associated with the ability to give a logos in the Meno. Here Socrates recounts a "dream" according to which the elements of reality are perceptible but unknowable and without a logos, whereas the compounds which are formed from them are knowable and have a logos. The notion of a logos is a vague and possibly ambiguous one, but its connection with knowledge seems to imply at least that knowledge must be expressible in propositions. Socrates rejects the "dream view" (which may possibly be attributable to Antisthenes) on the ground that knowledge of compounds would not be possible unless there was already knowledge of the elements. Propositional knowledge must depend upon a knowledge by acquaintance of something, in Plato's view presumably a knowledge of the Forms. But what is the logos which, when added to belief, may turn it into knowledge? Three suggestions are considered. First, that a logos is simply the manifestation of the judgment in speech (a possible meaning of the word) is clearly insufficient. Second, that it consists of the recounting of the elements of the thing known is insufficient in that this may not actually amount to knowledge of how the elements are put together. Third, that it consists in the identification of the thing in question is rejected on the ground of circularity, for if being said to know something requires that one know the distinguishing mark of the thing in question, the account is manifestly circular. But nothing less than this is sufficient. The dialogue therefore ends inconclusively.

It is clear that in this dialogue Plato was working toward a view of knowledge which is not too far from the ordinary one. His thought, however, was never entirely free from ambivalence in this respect, and perhaps he never entirely lost his distrust of the senses. The ideal of knowledge as a kind of apprehension of a system of Forms remained. It was the task of the philosopher to investigate this system by means of dialectic, the techniques of logical division and classification. For Plato knowledge was always a state of mind and had to be accounted for accordingly. This presupposition lies behind the inconclusiveness of the Theaetetus. Yet most of the traditional epistemological problems arise in the course of Plato's argument, and it is worth attention for this reason alone.


Aristotle (384322 BCE) was not so affected by skeptical arguments as Plato was. He does, it is true, try to answer Protagoras (Metaphysics Γ 5ff.), and he does so in a way very much like that of Plato in the Theaetetus, by pointing to the standard case in each class of judgment. Even in his early (and now fragmentary) work the Protrepticus, he had emphasized the necessity of an appeal to the expert in deciding issues in any particular art or science. This remained his approach throughout his life. Aristotle's preoccupations with epistemology appear in two provinces in particularin his theory of science and in his theory of the mind and its faculties. But his approach to epistemology was not so much the attempt to justify the claim that knowledge exists as the description of what knowledge and its presuppositions are.


Like Plato, Aristotle held that knowledge is always of the universal. Insofar as we can be said to know particular things, we know them as instances of a universal; we know the universal in the particular. But it must be emphasized that for Aristotle universals are inherent in particulars; he vehemently rejects the Platonic notion of a world of separate universals or Forms. (The only exceptions to the inherence of forms in matter and God and the most divine part of us, reason in the highest sense.) Knowledge therefore depends ultimately on the soul's or mind's reception of the forms of things. The soul itself, as he made clear in De Anima, is not a distinct, spiritual entity but the set of faculties possessed by the body insofar as it has organs to manifest them.

Means of knowledge

Sense perception is the receiving by the sense organ, the faculty of which is the respective sense, of the sensible form of a thing, as he puts it, without its matter. He also describes sense perception as an actualizing of the potentiality that the sense organ possesses as its faculty. It is not easy to see how this account of the matter can be worked out in detail. Granted that the hand becomes hot when it touches a hot object, what happens to the eye when we perceive color? Aristotle thinks that each sense is affected in a way peculiar to it, so that each sense has its own special sense object. The eye has color, the ear, sound; and so on. In addition to the special senses, there is a common sense, which has no sense organ peculiar to it. It is a faculty of all the sense organs or at least of those which are capable of perceiving the same qualities of objects; for example, size and shape are perceptible by both sight and touch. Aristotle speaks of both the special sensibles, such as color, and the common sensibles, such as shape, as essential to the respective senses. Apart from these there are the incidental sensibles, which are the things that possess these properties. Aristotle speaks of them as incidental sensibles because if we use our eyes, we are bound to see color; it is not essential that we see a particular objectto use Aristotle's example, the son of Diares. At any given time he may be the object of our vision, but he does not have to be. Some interpreters have spoken of these incidental sensibles as perceived indirectly, but there is no warrant for this interpretation in De Anima. As Aristotle points out, it is possible to see indirectly that sugar is sweet if we know that what we see directly is sweet. But this is quite different from perception of the incidental sensibles.

With this rather passive account of perception Aristotle gives a more active account when he stresses the role judgment plays in perception. Indeed, in the Posterior Analytics he speaks of the senses themselves as discriminative capacities. It is through such judgment that perceptual errors, such as mistaking the identity of a thing, can arise. Aristotle tends to say that when a sense is concerned simply with its own special objectfor instance, sight with something whitethere is the least chance of error. On occasion, however, he seems to imply that here error is impossible because the reception of the form in this case is something purely passive, so that there is no question of judgment. His position is not altogether clear, and there may be some confusion in his mind on this matter.

The same combination of passive impression and active judgment can be found in Aristotle's account of such faculties as imagination and memory, and to some extent there is a parallel account of the intellect itself. The persistence of the exercise of the sense faculty after its actualization by a sense object leads to the setting up of images. But imagination cannot exist in the full sense without the exercise of some form of judgment. Likewise, memory must depend not only on having images but also on a referring to the past.

It is in the account of the intellect that the parallel with sense perception comes out most clearly. There is, first, a reception of form, in this case not sensible form but intelligible form. This corresponds to having concepts. Second, there is the combination of concepts in judgment, and it is only here that the possibility of error arises. Because the higher faculties depend for their existence on the lower, the exercise of the intellect, which is in itself nothing but a mere faculty, depends on the prior exercise of sense perception. Hence, Aristotle says, the soul never thinks without an image. Owing to the influence of St. Thomas Aquinas, this has often been interpreted as the basis of empiricism. If the issue had been raised by Aristotle, the outcome might have been this doctrine, but it is not clear that he did raise it. To say that the activities of the intellect are always dependent on the workings of the lower faculties is not in itself to say that the only ideas we can have are provided by sense experience. Finally, Aristotle distinguishes between an active and a passive intellect. The intellect thus far described is the passive intellect; the active intellect, something purely actual and without potentiality, is necessary in order to make possible the actualization of the faculties of the soul. In Aristotle's thought it is given no other function.

What knowledge is

Since knowledge is concerned with the universalwith formany knowledge which is expressible in judgment must consist of an apprehension of an essential connection between forms. To know something about a thing is to be able to subsume it under species and genus and thus to know what is essential to it. It is matter which is responsible for what is accidental, and matter is in the last analysisas prime matterunknowable. To a large extent Aristotle's conception of knowledge in the full sensethat is, scientific knowledgeis coincident with Plato's. For Aristotle knowledge implies order; sense experience without this is something less than knowledge. This notion of order or organization is akin to Plato's notion of logos. Similarly, the idea that knowledge consists of classification in terms of genus and species and thus of a charting of the essential connections between forms is akin to Plato's conception of dialectic as concerned with the structure of the world of Forms.

Knowledge and definition

The connection in Aristotle's thought between knowledge and classification in terms of genus and species also entails a connection between knowledge and definition, for definition itself is in terms of genus and species. Aristotle distinguishes between nominal and real definition, the first being designed to give knowledge of terms only, the second giving knowledge of the essence of the thing itself. The difference turns largely on the fact that giving the essence of the thing involves the explanation of its cause. Thus, Aristotle says that we think that we have knowledge in the primary sense when we can give the cause of the thing. To give the cause of a thing involves the demonstration of its essence from first principles, and this is the function of science. The first principles themselves can be known only by a form of intuition; one sees their truth in their instances. It is possible to explain the principles of one science in terms of another science, but this process must come to a stop somewhere. It is the mark of a foolish man, Aristotle says, to think that everything can be proved. Principles such as the law of contradiction, which are implied in all demonstration, can be proved only dialectically. A dialectical proof is one that starts not from necessary first principles but from what is commonly accepted. In this case the proof consists in getting the man who denies the law of contradiction to say something and then to show him that what he says implies the law; he is thus convicted by his own testimony.

Aristotle thus presents us with a concept of an ideal of scientific knowledge and gives some account of what is presupposed by it. The difference between knowledge and true belief is, in his account, presumably dependent on whether what is claimed is essentially and necessarily true, a part of science as he sees it. But Aristotle gives little in the way of a justification of the claim that knowledge is possible at all, for he clearly feels no need to do so. To that extent he is out of the main stream of epistemological thinkers.

hellenistic philosophers

If anyone in the ancient world was an empiricist, it was Epicurus, the leading Greek atomist.


In the view of Epicurus (341270 BCE) all knowledge resulted from contact with the atoms of which the soul is composed by atoms from outside. It is true that he said atoms could sometimes stimulate the soul directly without affecting the senses, providing humans with visions of the gods, but in general the senses had to be involved. Atoms affecting the sense organs produced sensations; mass stimulation of the sense organs resulted in a presentation or appearance (phantasia ) to the soul. Sense experience in the more general sense occurs when an incoming presentation is fitted to a general conception or abstract idea, which itself results from repetition of sensations. This is the nearest thing to judgment in Epicurus's system, and this is the most usual source of error. Epicurus insists that all sensations are true and that they are the ultimate standard to which we must refer all our judgments; they admit of no other check. Since they are the ultimate standard of judgment, there is no other source for a metaphysical theory about the world. This, it has often been pointed out, does not fit in very well with the claims that Epicurus makes in order to give an atomistic picture of the world.

It is not altogether clear what Epicurus meant in saying that all sensations are true, especially since he also maintained that phantasmata, dreams or the delusions of the mad, are true. In the context of the atomist conception of the physical basis of perception, however, the view seems to imply a doctrine, common in the history of the subject, that anything in the mind which is caused cannot be liable to error. In reality questions of truth and falsity do not arise in such circumstances, a consideration implicit in Epicurus's statement that sensations are without a logos (not the sort of thing to involve judgment), but the conclusion that error is impossible has frequently been drawn.


The rival Stoic school was founded by Zeno (fl. c. 300 BCE), but the main figure was, perhaps, Chrysippus (c. 280c. 204 BCE). The Stoics were also empiricists to a large extent, although there is doubt whether at least some of the school did not admit innate ideas. The central notion of Stoic epistemology was intuition or apprehension (katalepsis ). This, as is put by the Skeptic critic Sextus Empiricus (c. 200 CE), was their standard of truth. Like the atomists the Stoics thought that things make impressions on the soul, although they differed from the atomists over which physical processes were involved. They made no suggestion, however, that these impressions were necessarily veridical. This was true only of those impressions that were clear and distinct (enarges ). Whenever an impression is received in the soul, the soul has to register it by a process known as assent, but there cannot be said to be knowledge until there is apprehension, until the soul is gripped by the impression (katalepsis literally means "gripping"). When this apprehension can properly be said to exist is clearly open to question, and this was pointed out by the Skeptics. Hence, later Stoics were forced to say that apprehensive impressions were a guarantee of truth as long as there was no objection.


Meanwhile, the Skeptics were making attacks upon the dogmatic schools, as they called them. The general tendency of this school was to accept the doctrine of impressions and phantasiae, but to maintain that there was no ground for going beyond them. Thus, it was necessary to be content with appearances and not to seek for the hidden truth about reality. The arguments against dogmatism were probably unsystematized initially, but they were gradually put into order. Probably under Aenesidemus (first century BCE) a list of ten (or eight) arguments (tropes) was drawn up. Some of these were forms of the argument from illusion, stressing the possibility of illusion and error in order to suggest that there was no reason to think that we ever gain knowledge of the real truth about things. Perceptions, they said, are always relative to the circumstances, the perceiver, and so on. Hence, no phantasia is a criterion of truth, and nothing else can be. A later Skeptic, perhaps Agrippa (first century CE), systematized the arguments even further by constructing general forms of skeptical argument. The final form stated that because of the earlier arguments nothing could be known in itself but only in relation to other things; however, something could be known relatively to other things only if these other things could be known absolutely. Because this had already been shown to be impossible, nothing could be known.

Some Skeptics came to see that this conclusion, put so baldly, was too dogmatic. When Arcesilaus of the Academy (the New Academy of the third century BCE) was sufficiently influenced by Skepticism to claim that knowledge was impossible, perhaps claiming Socratic practice as a precedent, the Skeptics still thought that this view was a species of dogmatism. Carneades (214129 BCE), a later academic who tried to meet the arguments of Chrysippus of the Stoa, not only maintained that absolute knowledge was impossible but also tried to substitute a theory of probability for it. He distinguished three grades of probability in respect to perceptions: (1) the simply probable, (2) the probable and confirmed by its consistency with its concomitants, and (3) the probable, confirmed and tested for inconsistency with the system to which it belongs. The last grade is science as we ordinarily know it. But even this would have been too much for the pure Skeptic. Skepticism as a system received its fullest expression in the works of Sextus Empiricus in the second century CE.


In the third century CE Platonism was revived by Plotinus, the founder of the Neoplatonic school. This was Platonism in its more mystical aspects, although Plotinus often uses Aristotelian notions, sometimes with a Platonic twist. The soul, as opposed to the body, is given preeminence, so that perception and knowledge are made a function of the soul. The soul has its own activities, which are manifested in perception and memory; the body and its impressions are merely instruments for the soul to use. The main function of the soul qua intellect is to contemplate the Forms, above which is the supreme principle or entity, the One. Unity with the One is the soul's goal.

Medieval Thought

It was Neoplatonism which, according to St. Augustine (354430), brought about his salvation from Manichaeism in theology and skepticism in philosophy. Neoplatonism offered a supposedly positive doctrine in both metaphysics and epistemology and one which St. Augustine could largely accept, thus ignoring the other, heterodox views. St. Augustine's thought is therefore Neoplatonic in its essentials. As a result he took it for granted that knowledgeand, most importantly, knowledge of Godwas possible, and he felt no further need to question this assumption. The same is true of most other medieval thinkers. Since philosophy was closely linked with theology, it was axiomatic that knowledge of God was possible in some sense, and skepticism about knowledge in general was rejected by an appeal to whatever philosophical system was thought best able to explain that knowledge. Insofar as there was argument, it was about the presuppositions and sources of knowledge, not about whether it existed.

Knowledge of a thing involves, it is commonly thought, knowledge of its general characteristics and, therefore, its subsumption under a universal. Medieval thinkers differed according to their philosophical tradition, according to whether they were Platonists or, after the Aristotelian revival in the thirteenth century, Aristotelians. But the main dispute was over theories of universals. Since the dispute had theological implication, it was heated. The argument had its source in certain questions put by Porphyry, a disciple of Plotinus, about the exact status of species and universals. These questions, the answers to which Porphyry thought were obscure, were discussed by Boethius in a commentary. The main schools of thought on the subject were the realists, conceptualists, and nominalists. Realists thought that universals had an objective existence, although their view of this existence depended on whether they were Platonists or Aristotelians. Conceptualists held that universals existed only as concepts in the mind; nominalists held that the only universal things were words. Such theories, however, were rarely found in their pure form.

realist theories of universals

The division among realists is best seen in the differences between St. Augustine and St. Thomas Aquinas (12251274).


St. Augustine gave preeminence to the soul, in Neoplatonist fashion. In his view the soul has its own functions and is not directly influenced by the body. Perception is based on the impressions produced by the soul when the body is stimulated. Experience, however, involves inference, as the soul subsumes its impressions under concepts. To have such concepts is, for St. Augustine, to be aware of Forms in the Platonic sense, the one difference being that in his view the Forms are thoughts in the mind of God. Thus, universals have a real existence in the mind of God, and all knowledge, even sense knowledge, involves some awareness of God. There is an ascent from lower forms of knowledge, like perception, to higher forms, with knowledge of God at the peak.

Thomas Aquinas

The Aristotelian revival in the thirteenth century led to St. Thomas Aquinas's acceptance of a more Aristotelian point of view than Augustine's. Like Aristotle, Thomas rejected self-subsistent universals, maintaining that universality is a function of the mind. Nevertheless, there are real similarities between things because of their common form. Hence, species have more than a mere mental existence.

The Thomist theory of knowledge consists largely of an explanation of how knowledge of Forms is possible. When the senses are stimulated, the soul's potentiality is actualized; a sensory image, or phantasma, is set up, corresponding to the object of perception. But since the universal aspects of such objects can be apprehended only by the intellect, they must be transferred from the phantasma to the intellect. Indeed, phantasmata as such are not objects of awareness on our part. The mind is aware only of the universal aspects of things, not their particularity, which is available only to the senses. Something has to illuminate the phantasmata in order to make clear their universal characteristics. Thomas employs Aristotle's distinction between the active and passive intellects here. The active intellect abstracts the universal or species from the phantasma, and this is imposed upon the passive intellect as a concept, which is then verbalized. Concepts thus exist only as the result of an abstraction of the universal aspects of things, and the essence of Thomas's empiricism is that all knowledge depends on sense experience in this way. Even knowledge of self-evident truths, which Thomas admits, as well as knowledge of the essential nature of things, is in the last resort dependent on sense experience, and all our thoughts must be based on experience. Thomas can be looked on as the founder of empiricism in the sense that he held that all the materials for knowledge come ultimately from experience and from nowhere else. Unlike the later philosopher John Locke, however, he did not set out to justify the doctrine in detail.

conceptualist theories of universals

Although Thomas may be classed as a realist in his theory of universals because he maintained that there are objective similarities between things by virtue of their common form. He could not do so without the notion of concepts in the mind.


Peter Abelard (10791142) had previously held a conceptualist theory of universals, emphasizing the extent to which universality is a function of the mind. He held that universals are really concepts (sermones ) involved in judgments that particular things have something in common. They are arrived at by abstraction from particular things, by attending to features of things considered in themselves. Abelard even used the notion of a generic imagethat is, an image not of any particular thing but, supposedly, of what is common to a whole class of thingsin order to account for our ability to think of things generally. He was anxious to reject both realism and the contemporary nominalism held by Roscelin of Compiègne (d. c. 1125), who maintained that universals were just words or even names. He did not, however, deny that concepts had a basis in things. Hence, in a sense Thomist realism and Abelardian conceptualism are very much two sides of the same coin.

nominalist theory of universals

Similar considerations apply to the great nominalist thinker of the fourteenth century, William of Ockham.


Even Ockham (c. 12851349) did not quite maintain that the only universal things are words, for he held that words are conventional signs corresponding to concepts that are natural signs of things. Universality lies in the sign-significate relation, in the fact that both words and concepts can be signs of a class of things. To the question "What are universals?" Ockham initially replied that they had only a logical existence; they were meanings, the contents of the mind when thinking generally. For Ockham the term universal was what he called a term of second intention. A first intention is the state of mind involved in the apprehension of particular things; a second intention is that involved in the apprehension of first intentions. The term universal thus picks out the content of our apprehension of our first-order apprehensions of a class of things. "Redness" is not the name of an entity but of the content of the relation which exists between the sign or concept "red" and particular red things. Ockham later took another step toward conceptualism by holding that universals had a mental, not just a logical, existence. He then held that universals are the concepts that the mind has and which are the natural substitutes for things themselves. Ockham probably took this step for reasons of economy, for in the earlier account universals were a sort of intermediary between the mind and particular things. The place given to generality remained the same, however. There was nothing general in the world; generality depended on the relationship between the mind and particular things.

The real novelty of Ockham's approach lay in his holding the view, new to medieval thought, that the mind itself could have apprehension of particular things. Thomas, for example, had denied this, holding that the mind could be concerned only with universal characteristics abstracted from phantasmata. The consequent gap between the mind and the senses inevitably involved a representative theory of perception; the mind was confronted only with the representatives of things. Ockham denied all this. He held that the mind could be concerned directly with the particular by means of intuitions. Intuitive knowledge is a direct knowledge of a thing or its existence. The senses provide an intuition of a thing's existence, and the intellect provides an intuition of its nature. John Duns Scotus (c. 12661308) had held that intuition of a particular was a necessary condition of the abstraction of the universal from it, but he had also held that this intuition must, in this life at any rate, be confused. Ockham denied this. In his view intuitions may be perfect or imperfect respectively, according to whether they are dependent merely on present experience or whether they also involve past experience. The possibility of imperfect intuitions, however, depends on the possibility of direct, perfect intuitions.

Although Ockham thought that this kind of direct knowledge does exist, he did not claim that all intuition was equally clear; clarity, moreover, was not always enough to guarantee truth. In the first respect, he claimed, as St. Augustine had done earlier, that we have clearer knowledge of our own mind than of other things. In the second he maintained that God can give us an intuition of something that does not in fact exist (a consideration which looked forward to René Descartes's suggestion that God might be a deceiver). This is not the natural course of things, however.

Much of Ockham's thought is in the Stoic tradition, and to some extent this can be said about Descartes, the first of the rationalist thinkers of the seventeenth century. By this time, however, the questioning of the accepted points of view of the Middle Ages had led to increased skepticism. Descartes's theory of knowledge is therefore in the fullest sense the beginning of that "search for certainty" whose elements were found in Plato but had not been prominent after him.

Seventeenth-Century Rationalism

descartes and cartesianism

The emergence of science during the Renaissance and the disputes that it produced led to a certain skepticism about claims to knowledge and to the search for a method, like that of science, which would determine the truth once and for all. Descartes (15961650) was the pioneer in this new tradition, and although his roots were in the Middle Ages, he was to a large extent an innovator. Being a mathematician of distinction, he saw the solution to problems of epistemology in the systematization of knowledge in geometrical form, although he did not carry out the full program himself. This involved starting from axioms whose truth was clear and distinct. He describes the ideal method in the second chapter of the Discourse on Method as (1) not to accept as true anything of which we have not a clear and distinct idea, (2) to analyze the problem, (3) to start from simple and certain thoughts and proceed from them to the more complex, and (4) to review the field so thoroughly that no considerations are omitted. Of what do we have clear and distinct ideas? To deal with this problem, Descartes employs the method of doubta form of skepticism. This method involves setting on one side anything that can be supposed false until one arrives at something that cannot be supposed false.

That there is a goal to this skepticism is, it might be objected, prejudged, for Descartes points to the fact that he has often been deceived to suggest that he may always be deceived. This conclusion is not, however, admissible, since to establish his premises, he must at least know that he has sometimes been deceived. The truth is that Descartes has a definite conception of what knowledge must be, and most of what we ordinarily call knowledge does not fit that conception. His approach is therefore not strictly that of the general skeptic. It consists in setting on one side anything that does not possess the mark of genuine knowledge, this mark being that we should have clear and distinct ideas of the thing in question. We have a clear idea of a thing when it is open to the mind, when we are clearly aware of it; an idea is also distinct when we have a full knowledge of the nature of its object and of the means whereby that object can be distinguished from other things. Many philosophers have believed that we do not have certain knowledge of empirical truths but that we do have it of mathematical truths. Descartes agrees to the extent that he maintains that we can have clear and distinct ideas of the objects of mathematics, but he also maintains that if God were a deceiver, he might have caused us to have false beliefs even here. Hence, it is at least a possible hypothesis that there is an archdeceiver who brings it about that I am mistaken in all my beliefs. Is there anything which is free from this contingency?

"Cogito ergo sum"

The result of Descartes's inquiry into this matter is that there is at least one proposition which is indubitable in the sense that I cannot be wrong in maintaining it. This is the proposition "I think, therefore I exist" (Cogito ergo sum ). Descartes is definite that this is not to be treated as an argument despite its form; it is an indubitable proposition. (In a sense Descartes had been anticipated in this by St. Augustine's "If I am mistaken, I exist" [Si fallor sum ], but St. Augustine had not used the proposition for exactly the same purposes.) It is reasonably clear that I cannot deny either "I think" or "I exist" without absurdity, although this is not enough to show that the cogito is in any way a logical truth. Yet for Descartes it must have the kind of necessity that is generally attributed to logical truths; it must be logically impossible for the proposition to be false. Moreover, it must have content such that its truth entails the existence of something with a specific naturenamely, a spiritual or thinking substance which has certain ideas, particularly those of God and material objects. Only then can Descartes go on to justify belief in such objects.

In effect, therefore, Descartes says that I can doubt everything except that I doubt. Since doubting is a form of thinking, I cannot doubt that I think, and since thought requires a thinker, I cannot doubt the existence of myself as the thinker. It might be objected that even if there is no reason to reject this position, it has not been shown to be necessary. If I cannot doubt that I doubt, this may be a contingent matter, not a logical necessity. Once given the cogito, however, Descartes can go on to use it as the premise of an argument whose outcome will be the justification of our belief in a world independent of ourselves.

The status of perception

Granted that we have ideas of a world of material things, what prevents those ideas being mere figments of the imagination? Ideas in themselves are purely mental entities (although Descartes is never very clear about their exact nature); they may or may not represent the things they purport to represent. Ideas can be innate, adventitious, or fictitious. To say that they are adventitious is to say that they come from things outside us; to say that they are fictitious is to say that they are produced by the mind itself. Innate ideas are a priori, inborn. Which ideas these are, if any, may be disputed. But, at all events, our ideas of material things are clearly not innate. Why, however, are they not merely fictitious?

To say that an idea is fictitious is not to say that it is impossible for it to be an idea of something objective. To some of our ideas perhaps nothing could possibly correspond; these would be logical impossibilities and would have no objective validity in Cartesian terms. For an idea to have objective validity, the reality in it must be caused by something that has the same reality, either formally or eminently, in itself. A machine, to use Descartes's example, may be formally the cause of a man's idea of it; his idea may be a copy of the machine, the two having the same form. But if the man conceives of the machine himself, then he or his mind is eminently the cause of his idea; the idea is not a mere copy of its cause, for the source of the idea is something higher. If, then, our ideas of material things are to be objectively valid or have objective reality, they must either be copies of actual material things or be produced by something higher. If they are produced by something higher, they were produced either by our minds or by God. To show that our ideas of material things do correspond to those material things, Descartes has to show that they are not produced in this way either by our own minds alone or by God.

Now, ideas in themselves, Descartes maintains, cannot be strictly true or false; it is the use we make of them that can be true or false. Hence, truth and error are functions of judgment. Nevertheless, we have a natural disposition to believe that our ideas are veridical. In Meditation III, Descartes points to this natural disposition and to the fact that our perceptions do not depend on the will as reason for believing in the veridicality of our perceptions, although he rejects these considerations as insufficient. In Meditation VI, however, he has recourse to the same considerations, although they must now be viewed in the context of the view that perception is a faculty of the mind plus body and does not express the essence of the mind alone, which is concerned solely with thinking. The passivity of our perceptual ideas thus seems to be invoked in order to reject the notion that our ideas of material objects could be a product of our own minds. Ideas of material objects, however, might still be caused by God. Yet if this were true and if nothing answered to those ideas, God would be a deceiver, for he would be giving us a natural disposition to believe in the existence of things which do not exist in fact. God, Descartes maintains, is not a deceiver. This point is taken as axiomatic and provides the ultimate guarantee of our belief that we do perceive an objective world.

Existence of God

The existence of God is therefore a cardinal point in the chain of argument. Descartes produces two sets of arguments designed to demonstrate his existence as a necessary truth. The first argument, in Meditation III, is based on the same considerations about the causes of our ideas as those adduced in connection with ideas of material things. The idea of God, which Descartes takes to be objectively valid, could be produced only by something having the same reality formally or eminently in it. We, being inferior creatures, could not produce it, and it could not come from any other source except God himself. Hence, there must be a God. This argument is a version of the so-called Cosmological Argument.

The other argument, to be found in Meditation V, is a version of the Ontological Argument first invoked by St. Anselm. God must through his essence possess all positive attributes in perfection. Existence is a positive attribute; hence, God must exist. It is now generally recognized that neither argument is sufficient to demonstrate the necessary existence of God. However, the necessary existence of God must be demonstrated if the argument concerning the existence of material things is to have any cogency.

If Descartes's argument for the existence of God had been sound, he would have shown that all the reality in our ideas must be in their causes. More is required if it is to be shown that our ideas are, at least in some cases, copies of their causes. This is a problem for any representative or causal theory of perception, any theory that holds that our ideas and perceptions are mental entities which are, at best, only representatives of things outside them. From Descartes's point of view there is the general consideration that God is not a deceiver; any errors or illusions to which we are subject are the results of judgments we make because of the ideas we receive. This makes it incumbent on us, if we are to avoid error, to take due account of the clarity and distinctness of our ideas. We are right in judging that our ideas correspond to their causes only if those ideas are clear and distinct.

Primary and secondary qualities

In Descartes's opinion there is a big difference between primary qualities, such as figure, magnitude, and motion, and secondary qualities, such as color. Primary qualities can be known by an intuition of the intellect (inspectio mentis ). Our ideas of them are clear and distinct because of the part they play in mathematics, and in mathematics, therefore, the intellect can be regarded as having a proper knowledge of reality. This is not to say that we cannot make mistakes concerning the primary qualities of objects; judgment can be as liable to error here as elsewhere. However, since the ideas of them are clear and distinct, we have the assurance that in general objects do have such qualities. We have no such assurance in the case of secondary qualities. The intellect is not involved here, but since the senses were, Descartes maintains, provided only for the conservation of life, it does not matter very much whether our ideas of secondary qualities correspond to the actual qualities of objects.

It would not be generally admitted today that mathematics does provide the kind of knowledge of reality that Descartes requires. The question of the exact connection between mathematics and the world is a complicated one. Granted, however, that mathematical ideas have a precision not possessed by other ideas, it does not follow that we have a precise knowledge of any qualities of physical objects. For it remains an open question to what extent such ideas are applicable to physical objects. There is a genuine distinction between primary and secondary qualities in that the first are measurable in a way that the second are not. This, however, is not sufficient to justify the claim that knowledge of primary qualities is knowledge of reality in a way that knowledge of secondary qualities is not. In some placesfor instance, Principles, Part IV, Section XI, and Dioptric, Section VIDescartes tries to reinforce this view by arguing that since the effects in the brain caused by the stimulation of our senses possess only the properties of motion, figure, and extension, there is no means whereby we could come to apprehend any other properties of objects. There is a circularity in this argument, since its premise is that we do know of the primary qualities possessed by brain processes, whereas all ideas, being the effects of physical processes, should be in the same position.

In sum, Descartes's theory of knowledge is essentially one of a representative kind. It is based on the idea that the mind or soul, being something very different and distinct from the body, can have as its contents only ideas, which are, at best, representatives of physical things. The mind has its own activities, and its nature is to be active. Through these activities it can come to have knowledge of abstract mathematical truths. But all sense knowledge, as opposed to intellectual knowledge, can come only through the medium of ideas, and that these ideas correspond in any way to the physical objects presented to the senses is inevitably open to question. The justification of our belief that they do depends, in the long run, on the affirmation that there is a God and that he is not a deceiver. Descartes thinks that he can demonstrate that there is a God, taking as true by definition that he is no deceiver and that our natural disposition to trust our senses is therefore justified.


Since Descartes conceived of the soul and the body as distinct substances with distinct essencesthat of the soul being thought and that of the body extensionhe was inevitably faced with the problem of how one could act on the other. He was never very clear on this point, although he came to insist that there must be some quasi-substantial union between the two. In some placesfor example, the Dioptric he speaks of brain processes giving "occasion" to the soul to have sensations or ideas. Some of his followers, who thereby became known as occasionalists, took up this notion and tried to explain the apparent link between soul and body by saying that God puts ideas into the soul on the occasion of the bodily processes. Arnold Geulincx (16241669) said that God puts the ideas there by means of the bodily processes; Nicolas Malebranche (16381715) said that God acts directly on the mind on the occasion of the bodily processes. Since Malebranche had leanings toward the views of St. Augustine, he interpreted this occasionalism in terms of the Augustinian doctrine that we know all things in God.


In other respects Malebranche tends to follow Descartes, although often with greater emphasis. He, too, insists that we have clear and distinct ideas of figure, extension, and movement, since these qualities, being conceivable in mathematical terms, are open to the intellect. He also lays great emphasis on our liability to error in anything connected with the senses, especially if we think that the senses provide us with knowledge of things as they actually are. In one respect, however, he adds a certain sophistication to the Cartesian distinction between the ideas or sensations that arise in the mind without any intervention on our part and the judgments that we make and which depend on our will. Sometimes what we see differs from what would be expected on the basis of the sensations resulting from the stimulation of our senses. We may see a thing in its right size, for example, although the actual pattern of stimulation received in the eyes is different; alternatively, we can be subject to illusion when the conditions of stimulation are not abnormal. In such circumstances we are not generally aware of making any judgment in order to correct the sensations received. Malebranche therefore distinguishes between natural judgments, or judgments of sense, and free judgments. Free judgments depend on our will, but natural ones do not; they are, he says, a kind of complex sensation in that they do not depend on us. They are, he explains, made by God in us, in consequence of the laws of the union between soul and body. As judgments they can be true or false, but as sensations they may occur against our will and are certainly not due to our will. Malebranche is in an ambivalent position here, but his difficulties show a certain honesty.


It has often been remarked that what makes the thought of Benedict (Baruch) de Spinoza (16321677) especially interesting is that it combines the quite different and, as generally conceived, quite disparate traditions of nominalism and extreme rationalism. In his nominalism he belongs to the tradition of William of Ockham and, more particularly, of Thomas Hobbes (15881679). Hobbes maintained that only names were universal. Although names were signs of images of things that constitute our conception of them, there was nothing universal in the conceptions themselves. Only the use to which names are put was universal, for they are taken as signs of many things. Hobbes used this to mount an aggressive attack on the paraphernalia of Scholasticismessences, substance, and the like. He was a tough-minded mechanist who thought that reality consisted solely of corporeal bodies in motion.

Although he did not have the same motive, Spinoza was similarly opposed to the apparatus of Scholasticism. Indeed, he may have been influenced by Hobbes. In his Ethics, Part II, Spinoza held that as a result of the stimulation of our bodily senses by many things, confused, composite images arise, and it is these images that general words represent. There is no place for universals existing in things. Since these images may be set up differently in different men, "universal notions," as Spinoza calls them, may differ from man to man. The knowledge that is expressed by their means can only be confused. Spinoza is not content to leave knowledge there, however; he has the conception of knowledge of a much higher kind, and his working out of this conception is in effect the systematizing of Cartesianism. To make Cartesianism consistent, however, he had to change much in it.


Although according to Descartes's view the clarity and distinctness of an idea was a necessary condition of its truth, it was not a sufficient condition of its truth. It was always possible to raise the question of whether any particular idea did correspond to reality and, in particular, to its cause. This was a consequence of the dualism between the mind and its ideas and the physical world, a dualism inherent in Cartesianism. Spinoza replaced this dualism by a monism according to which the mental and the physical were two aspects of one thingultimately, God or Nature. In adopting this view, he was again carrying out the implications of Cartesianism, for Descartes had asserted that in the proper sense the concept of substance belongs only to God, for only God is self-subsistent, or causa sui. Hence, in Spinoza's view everything is a modification or mode of the one true substance, depending on God for its existence. The Cartesian distinction between mental and material substance becomes a distinction between the two infinite attributes of God in Spinoza's theory. Bodies are modifications of God qua extended, and minds are modifications of God qua thinking. They are not distinct things; they are merely parallel aspects of the one true substance. The order of ideas in the mind is the same as the order of things. Hence, the objects, or ideata, of ideas, insofar as these have reference to God, are always truly represented by them. No ideas can be absolutely false, and insofar as they refer to God, they must be true. Ideas can be considered false only from the point of view of what we ordinarily call a single human mind, not from the point of view of God.

Truth and falsity

Because everything can be deduced from the essence of God, everything is subject to necessity, and this applies to ideas as much as to everything else. There is no room in Spinoza's theory for the Cartesian distinction between ideas and the will; for him the will and the intellect are the same. Falsity cannot therefore arise from the exercise of our will in the employment of ideas. For Spinoza the exercise of the will in judgment is not something additional to having ideas, and he emphasizes that it is wrong to think of ideas as simple pictures that may or may not correspond to their objects. To have an idea of something and to make a judgment about it are one and the same, and it is by virtue of this composite conception that ideas, not only judgments, can be true or false. Considered as part of God's thinking, ideas cannot, of course, be false, for in that case they must always correspond, qua modifications of God as thinking, to the modifications of God as extended. But considered as ideas in a single human mind, they can represent their objects confusedly or inadequately. In having a confused idea of an object, we fail to see it as following necessarily from the nature of things. Such ideas, Spinoza says, are like consequences without premises.

In the Ethics, Part II, Definition 4, Spinoza defines an adequate idea as "an idea which, insofar as it is considered in itself without relation to an object, has all the properties or intrinsic marks of a true idea." If an idea is confused, it cannot be adequate. Therefore, truth must have an intrinsic criterion, not just the extrinsic criterion of the correspondence of an idea to its object. In other words, the clarity and distinctness that Descartes had looked to for the foundations of knowledge is the mark of every true idea. Truth and adequacy thus merge. For full truth an idea must be seen to be true, and this is possible only insofar as it is seen to follow from the nature of things. Truth, Spinoza says, is its own criterion. This is connected with another of Spinoza's epistemological viewsthat knowledge must ultimately be reflexive. Anyone who really knows something must know that he knows. If a man knows that something is necessarily so, he must know that he knows this, since the truth of what he knows must be manifest.

The doctrine of truth that Spinoza presents is commonly referred to as the coherence theory of truth, and it is normally associated with the doctrine of degrees of truth, knowledge, and reality. The distinctions between grades of reality that exist between the one true substance and its various modifications is paralleled by distinctions between kinds of knowledge. True knowledge, the having of adequate ideas, entails seeing things as following from the essence of God. Knowledge can be more or less inadequate or confused to the extent that a thing is not seen as following necessarily from that essence. Absolute truth consists in having adequate ideas, although every idea has some degree of truth since it must have a counterpart in the order of things. In other words, an idea, although necessarily true in some respect, has greater truth to the degree that it is adequate and to the degree that its object is seen as fitting in with the order of things. Like most versions of the coherence theory of truth, this is really not a theory of what truth is or what is meant in calling an idea or judgment true, but, rather, a theory of when or under what conditions an idea or judgment can be seen to be true. A judgment can be seen to be true if it coheres with the system of judgments that characterize reality. But coherence theorists tend to say that any judgment which can be seen to cohere with other judgments in this way is thereby "more true" than those which do not cohere. In effect, they tend to use the word true so that it is more or less equivalent to verified. A judgment that has a higher degree of verification by virtue of its coherence with other judgments is said to be ipso facto more true. The coherence theory of truth is thus a genuine epistemological theory, a theory of the conditions under which we can be said to know a proposition as true.

Kinds of knowledge

Spinoza distinguishes three grades of knowledge. Full knowledge Spinoza refers to as the third kind of knowledge and characterizes it as intuition. This kind of knowledge, he says in the Ethics, Part II, "proceeds from an adequate idea of the formal essence of certain attributes of God to an adequate knowledge of the essence of things." To have this knowledge is the goal of philosophyto see things sub specie aeternitatis, as conforming to a kind of necessity. The right method in philosophy as set out in the Treatise on the Correction of the Understanding is to rid the mind of confused and inadequate ideas and to lead it to ideas which are adequate. It is significant that Spinoza calls this kind of knowledge intuition, because in its essence it consists of seeing the world as a coherent whole bound by necessary connections. Most rationalists have ended up with some such conception, and for them reason is inevitably second best. So it is for Spinoza.

What Spinoza calls reason is the second kind of knowledge, below intuition. This is best described by distinguishing it from the first kind of knowledge, which is knowledge derived from vague experience and is also called opinion or imagination. This corresponds roughly to sense experience. (Knowledge from hearsay, the fourth kind of knowledge, added in the Treatise to the bottom of the list, is of little importance for present purposes.) From sense experience we gain confused ideas of things without respect to their place in the general order of things. We may obtain knowledge of a similar status from signsthat is, from reading or hearing words which allow us to form ideas similar to those of the imagination. Both of these sources may lead to the setting up of the universal notions referred to earlier, notions that vary from person to person and cannot provide genuine knowledge. In the course of our experience, however, we may light upon notions that are common to all people, such notions as those of extension and other general attributes of reality. These notions Spinoza calls common notions, to be sharply distinguished from the universal notions already discussed. These common notions correspond to the ideas of primary qualities that Descartes had allowed to be clear and distinct because they were the objects of intellectual intuitions. For Spinoza, too, they provide the starting point of the sciences, and as such our ideas of them are adequate. They can be seen to be true of reality inasmuch as they reflect all-pervasive features of reality. It is for this reason that they are common to all humans; they are not subjective like the universal notions.

Reason or science thus consists in elaborating the essential features of the attributes of which we have common notions. Like Descartes, Spinoza conceived of science as based on the model of mathematics in general and geometry in particular. His conception of the right method in philosophy itself is modeled on the geometrical method. Indeed, Spinoza had tried to set out the Cartesian philosophy in a geometrical fashion according to the rules that Descartes had preached but had not practiced as well. Thus, although science, the systematization of knowledge, is ultimately derived from sense experience, it reflects the actual order of things more truly than experience does. Nevertheless, the goal of all knowledge is not just this systematization but the seeing of things as a whole, sub specie aeternitatis. For this reason intuition stands above reason.

Because the second and third kinds of knowledge involve adequate ideas, they cannot give rise to falsity. Sense experience alone can be the source of falsity. Through sense experience we can have only confused ideas, since ideas reflect particular modifications of reality in some finite respect, not in relation to the infinite attributes of God. Sense experience is ordinarily thought of as a passive form of knowledge, as opposed to forms of knowledge that demand the use of reason. This passivity, Spinoza thinks, is only a sign of the inadequacy of our ideas in this case. Activity on the part of the mind is, conversely, adequacy in its ideas. Spinoza points out that the human mind is part of the infinite intellect of God. Hence, "when we say that the human mind perceives this or that, we say nothing else than that God, not in so far as he is infinite, but in so far as he is explained through the nature of the human mind, or in so far as he constitutes the essence of the human mind, has this or that idea" (Ethics, Part II). Just as the ideas of sense experience are confused and inadequate only in relation to our mind, although when considered as God's ideas they are true, so the ideas of sense experience are passive in relation to our mind but are nevertheless part of God's active thoughts.

In sum, for Spinoza the goal of all knowledge is seeing the world as a single whole. On the way to this lies reason or science, which attempts to reveal things as subject to necessity by means of self-evident, necessary truths about things. All else, although not absolutely false, is the source of illusion. But, as in everything else in Spinoza, an adequate understanding of his theory of knowledge also involves an adequate understanding of his complete metaphysics or theory of reality. This is true of philosophers in general but never more so than in Spinoza's case.


In many ways Gottfried Wilhelm Leibniz (16461716) is Spinoza with a strong dash of common sense. Spinoza's monism, especially its assertion of the necessity of things and the apparent consequence that free will is impossible, was anathema to Leibniz. In these respects Leibniz revolted against Spinoza, but in other respects he was very much like Spinoza. He, too, drew no distinction between the will and the intellect and made activity and passivity in the mind a function of the clarity and distinctness or otherwise of our ideas. Furthermore, although common sense told him that there must be a plurality of things, not just one, he had to conceive of each ultimate thing as a simple substance possessing all the properties of Spinoza's one substance.

Leibniz simple substances had to have a unity in plurality in that, although simple, each one had to be capable of reflecting the whole universe from its point of view. Since Leibniz took as axiomatic that in every true proposition the predicate is contained in the subject, everything that can be said about a substance is so because of the nature of that substance, and all the relations which it has to other things must arise from the nature of that substance and be internal to it. It is for this reason that every true substance must reflect the universe from its point of view and in this way be a microcosm of the macrocosm. The only thing, Leibniz thought, which could be both simple and capable of reflecting the universe in this way was something like the soul. In consequence, he postulated the existence of a plurality of simple substances, spiritual in nature, which he called monads. But for the monads, he said, Spinoza would be right.

Necessary and contingent truths

Since the properties of each monad were internal to it, it might be thought that Leibniz, like Spinoza, would have maintained that everything was a matter of necessity. However, although Leibniz maintained that all the properties of a substance are internal to it and thus follow from the nature of the substance, for other reasons he maintained a clear distinction between truths that are necessary in the logical sense and truths that are dependent on the facts. He was thus perhaps the first to draw a clear distinction between necessary or logical and contingent or factual truths. The first he called truths of reason, the second truths of fact. These truths had different principles as their basis. Truths of reason were dependent on the principle of contradiction, since their necessity turned on the fact that their denial would result in a contradiction. Leibniz thought that such truths, when their terms are defined, could be reduced to identical propositions of the form "A is A." The reduction to such identical propositions would therefore proceed by means of chains of definitions. Mathematical truths are of this kind, and Leibniz was one of the first to seek a basis for mathematics in logic.

Truths of fact, on the other hand, could not be justified by reduction to identical propositions; their basis had to be found in a separate principle, the principle of sufficient reason. This principle received different formulations at different stages of Leibniz's thought. Insofar as it was meant to allow for the contingency of matters of fact while providing a rationale for them, Leibniz tended to formulate the principle by reference to the choice of God. In creating this universe, God could choose from a number of possible worlds each having a different order or structure. Since, as Leibniz thought, for various different reasons, the number of monads is infinite, the number of such possible worlds is also infinite. Any contingent truth about this world has for its justification the fact that in choosing this world, God chose it as the best of all possible worlds. The truth remains contingent because it is dependent on God's choice, but a sufficient reason for its truth is that God chose it as part of the best of all possible worlds. At other times, however, Leibniz's appeal to God's choice has fewer theological implications. For example, in his correspondence with Samuel Clarke, he tries to refute the idea of an infinite absolute space by saying that in such a space God would have no sufficient reason for putting the universe here rather than there. This means that there would be no way of telling where the universe was and that it would, in consequence, make no sense to speak of it as being in one place rather than another. This use of the principle of sufficient reason amounts to something like the use of the verification principle by logical positiviststhe meaning of a proposition lies in the method of verification.

When Leibniz maintains, however, that in every true proposition the predicate is included in the subject, he seems to undermine the distinction between truths of reason and truths of fact. For this doctrine would make all propositions into what Immanuel Kant was later to call analytic propositions or judgments, propositions that are logically necessary. In consequence, it has been maintained (for example, by Louis Couturat) that in some of Leibniz's writings the principle of sufficient reason merely states that all true propositions are analytic, whereas the principle of contradiction states that all analytic propositions are true. In fact, this is probably a consequence of Leibniz's main views rather than a statement of it. If every proposition about a substance attributes to it a property that is part of its essence, then all such propositions must be analytic even if they are to be characterized as truths of fact on other grounds.

Leibniz accepted this conclusion but tried to evade the contradiction implicit in characterizing a truth of fact as analytic by explaining that the number of properties possessed by any substance must be infinitely great, as the points of view from which a thing can be regarded are endless. We, being finite creatures, cannot complete the analysis of any given substance. Hence, we cannot know for certain whether any given attribute really does belong to it; we cannot, without completing the analysis, even know whether it is possible for this substance to possess the property; it may be a contradiction to suppose it. God, being infinite, can complete the analysis, and so for him all propositions about things are analytic or logically necessary. We, on the other hand, can know only that if a proposition is true, it is necessarily true, but we cannot know for certain whether any given proposition is true. In our judgments about the truth of propositions, we have to depend on probabilitiesthat is, we have to estimate what reasons are sufficient for our conclusions. Thus, for us any judgment of fact is contingent. For God contingency enters only in that he has chosen what substances there should be, which of all possible worlds should exist. For him everything thereafter is necessary. Hence, the principle of sufficient reason comes into consideration in two related waysfirst, in that it guides, without determining, God's choice of a world and, second, in that it guides our decision concerning which world God has chosen.

Perception and appetition of monads

So much for the logical consequences of Leibniz's point of view. Given the metaphysical system according to which there exist, an infinite number of spiritual entities or monads, other consequences also follow. According to Leibniz, every monad has perception and appetitionthe apparent passive and active features of mental life. Since everything about a monad is internal to it, these features can indeed be only apparent. Appetition is that aspect of a monad responsible for internal change and development. No monad can affect or be affected by any other monad. A perception is any property of a monad that results from its development but that may reflect changes in other monads. (This use of the term perception is, although influenced by Leibniz's metaphysics, clearly very general, but its use was very general throughout the seventeenth century.) What may seem to be activity on the part of one monad with respect to another is really having distinct perceptions, whereas passivity is having confused perceptions. Here Leibniz sides with Spinoza.

Self-consciousness of monads

Leibniz's criterion of a distinct idea of a thing is the ability to list the characteristics which distinguish the thing from other things. This clearly involves a degree of self-consciousness, and this is possessed only by the monads constituting the human soul. Although all monads have perceptions in that other things are represented in them, not all have apperception. To have apperception, the monad must be conscious of what is involved in its perceptions, and those perceptions must therefore be distinct. The distinction between perception and apperception means that perceptions can sometimes be unconscious. Leibniz brings forward a number of arguments in support of this view, ranging from the argument that reflection upon perceptions must come to a stop with perceptions that are not self-conscious to the argument that there must be what he calls petites perceptions. When we hear the roar of the sea, he argues, we are not aware of hearing each little ripple despite the fact that the waves are made up of ripples. Since the overall perception must correspond in complexity to its object, he concludes that there must be perceptions corresponding to the ripples, and these little perceptions are therefore unconscious. This is not a psychological discovery but a philosophical analysis the premises of which are open to dispute. Like Descartes, Leibniz accepts the representative theory of perception in thinking that perception consists in having ideas which are, or may be, representative of objects. If this theory is rejected, Leibniz's argument about petites perceptions loses much of its force.

Error in perception

Just as Leibniz sides with Spinoza in maintaining that activity and passivity are to be explained in terms of the distinctness of our ideas, so he agrees with him, against Descartes, over the explanation of error. There is no room for the individual will in Leibniz's system. Appetition is only the impulse that provides the development of the monad's perceptions; it in no way corresponds to the will. Error is merely a matter of having confused ideas, and since these are correlated with passivity, the passive aspects of mental lifesense perception and the likeare the source of error.

Innate ideas

Yet although Leibniz can distinguish between ideas of perception and ideas of reason, or the understanding, it remains true that according to his view in a sense all ideas are innate. None is literally produced by things affecting our sense organs. Yet the distinction between ideas in terms of their clarity and distinctness does mean that it is possible to say that some ideas are what Kant called a prioriin no way derived from the senses. These are ideas such as those of mathematics, and Leibniz criticized his empiricist adversary Locke for failing to take sufficient account of these ideas. Indeed, to the empiricist principle that there is nothing in the intellect that was not first in the senses Leibniz replied, "Except the intellect itself."

Rationalism generally tends to emphasize the part played by the intellect in contradistinction to that played by the senses. It holds that real knowledge is that provided by the intellect, for only there is the certainty which knowledge requires. Moreover, it is by means of the disciplines that are peculiarly the province of the intellect that knowledge is to be obtained and preserved.

British Empiricism

In general, empiricism stands in opposition to rationalism both in its views about the main source of our ideas and in its views concerning the source of true knowledge. Thus, it is often, historically speaking, a reaction against rationalism. The so-called British empiricists of the seventeenth and eighteenth centuries, however, were empiricists only in tendency. The first, Locke, was a complete empiricist concerning the source of our ideas, but he was often a rationalist in allowing other than empirical knowledge. Locke's new way of ideas, as it was called, was an attempt to show that all the materials for knowledge are derived from sense experience. Locke did not claim, however, that all knowledge was founded on experience in any other sense. George Berkeley, who carried on Locke's new way of ideas and even sharpened some of Locke's claims, especially on the subject of abstract ideas, was fundamentally a metaphysician with a special way of looking at the world. David Hume, the last of the trio, claimed to introduce the experimental method into philosophy, following in the steps of Newton, and of the three he had by far the best right to be counted an empiricist. Indeed, his empiricism led him extremely close to skepticism concerning a number of claims to knowledge; such skepticism, he believed, could be avoided only by "inattention" to philosophical issues. But all three of these philosophers were united in their opposition to any doctrine of innate ideas.


Book I of Locke's Essay concerning Human Understanding is devoted to an attack on the doctrine of innate ideas, and the positive doctrine begins only in Book II. At the outset Locke (16321704) had claimed that he was following the "historical plain method," the object of which was to "set down any measures of the certainty of our knowledge." This historical plain method consists in classifying our different ideas and plotting their source as a prelude to an assessment of claims to different kinds of knowledge. Despite appearances this is not a psychological method; it is a method of philosophical analysis designed to discover the logical character of different ideas. In this way Locke distinguishes between ideas of sense and ideas of reflection. Ideas of reflection result from the operation of the mind itself upon ideas of the sense. There is no other source of ideas.

Locke also distinguishes between simple and complex ideas of both kinds, complex ideas being formed by the mind in compounding simple ideas. He seems to think that what it is to have a simple idea of sense is fairly obvious; it is to be aware of a particular quality of an object mediated by a single sense. The criterion of simplicity was, however, a problem for all British empiricists. In having simple ideas the mind is passive, but some activity is allowable in the forming of complex ideas.

Primary and secondary qualities

Among simple ideas of sense Locke makes an important distinctionalready implicit in Descartes and othersbetween ideas of primary and ideas of secondary qualities. Primary qualities, such as bulk, number, figure, and motion, are, Locke thinks, inseparable from the bodies in which they are found. Bodies could not exist without them. Secondary qualities, such as color, sound, and taste, are, on the other hand, "nothing in the objects themselves but powers to produce the various sensations in us by their primary qualities." In other words, the primary qualities of things produce sensations in our minds that are ideas of secondary qualities, but "secondary quality" is a misnomer to the extent that there is really no such quality in things. Thus, our ideas of primary qualities actually correspond to the things that produce them, whereas our ideas of secondary qualities, although produced by things, resemble nothing in those things, being purely subjective. Locke brings forward a number of arguments for this conclusion, arguments based mainly on the assimilation of our perception of secondary qualities to sensations of pain. That is, he takes the perception of, for example, warmth or color to be the same kind of thing as feeling pain.

Account of perception

He thinks of perception in general as identical with merely having sensations, and he thus fully embraces the causal theory of perception according to which perceiving is having sensations caused by things. He goes further than this in respect to primary qualities, for here he also accepts the representative theory of perception according to which our ideas represent the things that cause them. This theory, as we have seen, was the stock in trade of seventeenth-century philosophy. Like most theorists of this pattern, Locke can give no good reason for the view that any of our ideas resemble their causes, and he cannot take the rationalist course of appealing to an intellectual intuition of some properties of things. It is a fair guess that he, like Descartes and others, was influenced by the success of physical science in maintaining that physical properties like extensionproperties which are measurableare the properties of things. There is also the connected fact that these properties are perceptible by more than one sense, as Aristotle had noted in his theory of common sensibles.

Modes, substances, and relations

Complex ideas may be exhaustively subdivided into ideas of modes, substances, and relations. Ideas of modes are ideas of things that depend on substances for their existencefor example, the idea of a triangle or a murder. Ideas of substances are ideas of particular things taken as existing by themselvesthe complex idea of substance, he goes on to say, consists mostly of powers. Ideas of relation, finally, result from a comparison of one idea with another. Locke came to have some dissatisfaction with this classification of complex ideas, and in the fourth edition of the Essay he introduces a fourfold classification of ideassimple ideas, complex ideas, ideas of relation, and general ideas. However this may beand there is room for dissatisfaction with Locke's second classification, tooall ideas other than simple ones are in some way formed by the mind out of simple ideas.

Locke classified ideas of space, time, and number as ideas of modes. That is to say that they are ideas of entities which depend for their existence on particular things. We build up our ideas of these entities out of our ideas of particular things when seen in the appropriate relations. Kant later showed that such a view of the source of our ideas of space and time was untenable; Leibniz commented on the fact that Locke failed to take account of the a priori nature of the ideas of space and time and attributed the failure to Locke's inexperience with mathematics.

Locke maintained that our ideas of physical substances are mostly ideas of powers and that the idea of power is an idea of another mode. Since what we know of physical substances is largely due to their effects on us or on other substances, ideas of physical substances all mainly ideas of power. The effects, Locke thinks, are due to the motions of the invisible parts of things, but owing to the weakness of our senses, we are unable to perceive the nature of these causes. We have little or no knowledge of the "real essences" of things. What we do know of things is of their "nominal essences"their nature merely in respect of the classifications into which we fit them. Such classifications may not correspond to the real nature of things. Here Locke clearly shows how much he was influenced by the physical sciences. He thought that classifications show the way to the nature of things, but that owing to the weakness of our senses, we are unable to do more than gain a general impression of the nature of those physical processes. Therefore, we have to be content with an ordering of things according to their effects rather than as they are themselves.

Theory of general words

Locke adds to the account of ideas a discussion of language and of the words corresponding to the various ideas. It is in the context of this discussion that he puts forward his theory of the meaning of general words, a theory that was to come under attack from Berkeley. This theorythat the meaning of general words is given by the general or abstract ideas to which they correspondis in effect Locke's theory of universals. He expresses the problem by asking, "Since all things that exist are only particulars, how come we by general terms; or where find we those general natures they are supposed to stand for?" His answer is that words are general by being signs of general ideas, and we form general ideas by abstraction, "separating from them the circumstances of time and place, and any other ideas that may determine them to this or that particular existence." Thus, words become capable of representing a number of individuals by standing for an abstract idea. Locke's view is therefore a form of conceptualism in that the universal or general element lies in our ideas or concepts, not in anything nonmental. Given a liberal enough interpretation of the word idea, there is perhaps no great difficulty in understanding what Locke is getting at, although the implication that the meaning of words must always consist in their standing for something (in this case an abstract idea) is certainly wrong. The idea terminology is a vague one, common though it was in the seventeenth and eighteenth centuries, but most of those who employed it would have denied that ideas were simple images of things. Moreover, such an interpretation is far from consistent with much that Locke says about ideas. Nevertheless, the use of the term "abstract idea" is not without its difficulties, especially since Locke says that such ideas must represent things.

Kinds of knowledge

In his account of ideas and their classification Locke is the strict empiricist, maintaining that all ideas must be ultimately derived from simple ideas of sense, either directly or as a result of the operations of the mind upon these. His account of knowledge, given in the last book of the Essay, is less empiricist in character; indeed, it owes an obvious debt to Cartesianism. He begins with the claim that knowledge is nothing but "the perception of the connexion of and agreement or disagreement and repugnancy of any of our ideas." This agreement or disagreement can be classified into four kinds: (1) identity or diversity, (2) relation, (3) coexistence or necessary connection, and (4) real existence. It is the fourth kind which presents the difficulties. How can the knowledge of the existence of a thing be a matter of the perception of the agreement or disagreement between our ideas alone? This could be so only if our knowledge of the existence of things could be a priori. Locke thinks that some knowledge of this sort can be shown to be a priori, but it is knowledge of the existence of sensible things that presents the greatest difficulties.

Locke distinguishes between three degrees of knowledge in a manner which is reminiscent of Spinoza's distinction between the three kinds of knowledge. There is, first, intuitive knowledge; second, demonstrative knowledge; and, third, "sensitive" knowledge of particular finite existences. The last Locke adds almost as an afterthought on the ground that it has by no means the certainty that belongs to the first two, although it goes beyond mere probability and is commonly thought of as knowledge. (Locke's conception of the standard to which knowledge must attain is noteworthy here.) Apart from the different degrees of certainty that are to be attached to these kinds of knowledge, they also differ in that intuitive and demonstrative knowledge can be concerned with relations between ideas (we can see that white is not black, and we can reason from one idea to another) but sensitive knowledge is concerned only with the existence of the objects of ideas. This is not to say that there cannot be intuitive and demonstrative knowledge of existence, too. Locke thinks that we have intuitive knowledge of our own existence (compare Descartes's cogito ) and demonstrative knowledge of God's existence (by means of a version of the Cosmological Argument; Locke distrusts the Ontological Argument). But how can the existence of anything be known from ideas alone? Locke sometimes appears to say that such knowledge consists in the perception of the agreement of certain of our ideas with the idea of existence, but in general he acknowledges that more is required than thisreal existence and not merely conceived existence. The difficulties here are especially evident in connection with sensible knowledge.

Existence of external world

In Book IV of the Essay Locke tries to justify the claim that we have knowledge of the existence of particular sensible things by showing that our ideas do correspond to the things that cause them. Whereas complex ideas may not always correspond to things because of the part played by the mind in forming them, simple ideas receive no contribution from the mind; they are entirely passive. Unfortunately, it does not follow from this that they are necessarily veridical. He adduces further considerations, stressing the passivity of the mind in receiving ideas and the way in which the senses may cohere in their reports. None of these considerations is really sufficient, and Locke admits that they do not amount to proof. In fact, by simultaneously embracing a general empiricist approach and a representative theory of perception, Locke cannot provide a guarantee of, or even any general argument for, the veridicality of the senses. He cannot provide any independent access to the objects of perception other than that provided by the senses themselves. Like most empiricists, Locke accepts the correspondence theory of truth in that the truth of a proposition consists in its correspondence to the facts. (Truth, he says, signifies "the joining or separating of signs, as the things signified by them do agree or disagree one with another.") But he has no general warrant for the belief in the correspondence of ideas to things.


The main aim of Bishop Berkeley (16851753), as he conceived it, was to defend common sense and religion against skepticism and atheism. But both his metaphysics and his theory of knowledge, connected as they are, can be partially regarded as attempts to rid Locke's theory of impurities. Locke's view of the world involved, besides minds and their contents (ideas), material substances, for the most part unknowable. Berkeley wished to get rid of material substances precisely because they were unknowable. In his view the existence of material objects consists only in their being perceived; their esse is percipi (their existence is to be perceived). In consequence, they must be regarded as complexes of ideas whose cause cannot be any substance underlying them but must be a spirit, the only active thing. Some of our ideas may be caused by ourselves qua spirits, but insofar as ideas have what we normally think of as an objective order, they must be caused by the supreme spirit, God. The laws of nature according to which ideas are ordered are guaranteed by God, and our ordinary way of looking at and talking about the world obscures this fact. Berkeley therefore thought that it was necessary to rid Locke's views of those elements which prevented this insight. The main issues concerned the notion of substance as the cause of our ideas, with the connected doctrine of the distinction between primary and secondary qualities. Berkeley also found fault with Locke's doctrine of abstract ideas, partly for its own sake but also because he thought it one of the main supports for the doctrine of substance. We might, that is, have an idea of substance by abstraction; it had to be shown that this was impossible.

Knowledge of the external world

At first, in the New Theory of Vision, a work on both optics and philosophy, Berkeley maintained that physical objects are primarily objects of touch. Vision, he asserted, could provide us with no direct perception of the distance of things, for the retina of the eye is only a two-dimensional surface. Our sensations of sight (and, like others of his time, Berkeley thought that perception fundamentally consists in having sensations) can only be of expanses of color. When we perceive things as at a distance from us, what really happens is that the visual sensations which we have suggest to us certain ideas derived originally from touch and connected with the visual sensations by experience. The New Theory of Vision consists largely in the working out of this theory in detail. Berkeley came to see, however, that there was no reason for making this distinction between sight and touch. All senses should be alike in these respects. Insofar as we have knowledge of what we take to be physical things, it is because we have, as the result of experience, so connected ideas that having certain sensations or ideas suggests other sensations or ideas. These ideas make up a collection that we identify as an object. Thus, for Berkeley objects are, in some sense, identical with collections of ideas.

In the beginning of the main part of his Principles of Human Knowledge, Berkeley flirts with the view, now known as phenomenalism, that all we mean when we say, for example, that there is a table in our study when we are not there is that if we were there, we should perceive it. But he adds as an alternative the suggestion that what we mean is that "some other spirit actually does perceive it," and this is his main view. What we ordinarily take to be things are really bundles of ideas in some spirit's mind; they have a certain stability even when we are not perceiving them because God still is. Indeed, it is God's having ideas according to a certain order that guarantees the order of our ideas.

Sensations and ideas generally (for although sensations are, strictly speaking, one species of idea, Berkeley often uses the terms interchangeably) are entirely passive. Their esse is percipi. On the other hand, spiritsGod or ourselvescan be active. We cannot have ideas of spirits, although we have a notion of them, since we can understand the word spirit and at least know that we are the source of some of our ideas. There is no room in this for any substance's underlying our ideas, since we have no idea of such a thing. Furthermore, the special place that Locke had given to primary qualitiesthat of being the properties essential to material substanceis untenable, and if what Locke said about secondary qualities is right, there are no grounds for making any distinction between them and primary qualities. They are equally dependent on the mind, so that if secondary qualities are subjective (and Berkeley accepts and adds to Locke's arguments for this conclusion), they must all be. All qualities are ideas in the mind.

Theory of universals

Berkeley's fiercest attack upon Locke was directed against his doctrine of abstract ideas. Berkeley interpreted Locke as asserting the existence of ideas or images that possess contradictory properties. The abstract idea of a triangle must be simultaneously scalene, isosceles, and equilateral. This is clearly impossible. It is doubtful whether Berkeley is right in this interpretation, but he clearly thought that if such ideas were admitted, there could be little objection to the admission of the idea of substance, toothat is, the idea of a physical but in principle imperceptible object.

In the place of abstract ideas, Berkeley introduced a theory of universals that was nominalist in character. Universals are merely particular ideas that are representative of other ideas in the same class in the way in which a particular man may be representative of other men; hence, their universality lies only in their power of representation. There is no need to assume the existence of general ideas since general words need not correspond to general ideas in order to have meaning. In other words, Berkeley challenged the theory of meaning that asserts that all words are names and refer to somethingunum nomen, unum nominatum (one name, one thing named), as the Scholastics put it. In his view general words stand for a number of particular ideas belonging to the same class. General words are different from names in that general words represent a number of things indifferently. It must be confessed, however, that it is difficult to see clearly what, according to Berkeley's view, is involved in understanding a general word. Certainly, it involves having an idea which indifferently represents a whole class of things, but what is it to see that it does so?

Refutation of skepticism

Berkeley's general view has certain consequences. It means, for example, that important sections of mathematics have to be abandoned. There must, Berkeley believed, be a least perceptible size. Since all our ideas are ultimately derived from sensations, there can be no ideas of infinitesimals or points. For the most part, however, Berkeley considers himself to be defending common sense against the attacks of the metaphysicians. The vulgar, he maintains, believe that "those things they immediately perceive are the real things" (that is, not imperceptible substance), but philosophers believe that "the things immediately perceived are ideas which exist only in the mind." Berkeley characterizes his own view as the joining of these two notions in that he equated real things and ideas. He thinks that given his view that ideas, which are the objects of immediate perception, are the real things, there is no room for doubt concerning the real nature of thingsa doubt which Locke had expressed. Moreover, since what is immediately perceived is by definition free from error, only the wrong use of ideas in judgment can give rise to error. Error is thus a product of the imagination. Insofar as we rely upon our sense perceptions as directly given, we must be free from error. Thus, Berkeley claims, his view prevents skepticism and "gives certainty to knowledge."

Concept of knowledge

Apart from the reference to God and spirits, Berkeley is a strict empiricist not only in the sense that he believes that all the materials for knowledge are derived from sense perception (as Locke, too, believed) but also in the sense that knowledge is itself founded on sense perception. Locke was not such a complete empiricist; he thought that knowledge in the strict sense is founded on intuition and demonstration, and he made skepticism possible to a certain extent over sense perception because he thought that its veridicality could not be completely shown. According to Berkeley, knowledge derived from reasoning must ultimately be founded on knowledge based on sense perception. Sense perception, in turn, is no longer conceived of as having ideas that are produced by objects and may not always represent their causes.

Berkeley has given up the representative theory of perception with its assumption that something so underlies our ideas that they may be representative of it. His rejection of the representative theory of perception is the basis of his claim to combat skepticism. Yet, as Hume asserted, it has often seemed a claim that fails to produce conviction, because the claim that what is directly perceived is free from error is true by definition. The question of how we know when we have direct perception still remains, however. Not all ideas are objects of immediate or direct perception; some are ideas of the imagination. According to Berkeley, these are less regular, vivid, and constant than ideas of perception, and they are more "dependent on the spirit"; they can be distinguished from ideas of sense by these criteria. But are all ideas of perceptible things ideas of things immediately perceived, and if not, how do we tell which are?

In the first of the three Dialogues between Hylas and Philonous, Berkeley argues that by sight we immediately perceive only light, colors, and figures; by hearing, only sounds; by taste, only tastes; by smell, only odors; and by touch, only tangible qualities. Here he appears to be arguing from the premise that these things are the special or proper objects of the senses. Although it is difficult to know what, if anything, is special to sight and touch, it is easy to see what is meant in the case of the other senses. Even if we grant that we hear only sounds, taste only tastes, and smell only odors, it does not follow, however, that we cannot be mistaken about the characteristics of these objects in particular instances. Are we necessarily free from error in hearing when we hear that a sound is loud or soft?

Nor is our attribution of colors necessarily free from inference as it should be if the perception of color is immediate. What, then, really counts as an object of immediate perception? In answering this question, Berkeley is subject to the same difficulties that have beset more modern philosophers when they have sought to base the philosophy of perception on the notion of sense data. If the foundations of knowledge are found in the deliverances of the senses, there must be certain perceptions that are incorrigible in the sense that they cannot logically be subject to doubt. But what counts as incorrigible perception? Berkeley tries to answer this question by assimilating perception to having bare sensations. Sensations, however, are not the sort of thing that can be right or wrong. The mere passivity of sensation, as opposed to the will, does not show that error arises from the will. If this criticism is valid, Berkeley's theory does not satisfactorily prevent skepticism in the way that he supposes.


Locke thought that his inquiry had revealed the limitations of the understanding by showing that there are parts of nature that our senses cannot discern. Berkeley, on the other hand, thought that there was nothing which our understanding could not grasp. Sense perception gives us complete knowledge of reality, and we have in addition notions of spirits, including God. Indeed, we could regard our ideas as a sort of divine language by means of which God speaks to us, so that our senses, if viewed correctly, continually reveal the glories of God. Hume (17111776) agreed with Berkeley in thinking that there is nothing in nature that lies beyond the reach of our senses, but, contrary to Berkeley, Hume reached the conclusion that our understanding is very limited and that skepticism is the only reasonable attitude toward knowledge. That Hume was intentionally a skeptic has been disputed, but there is no doubt that this is the logical outcome of his views. He thought that whatever Berkeley's claims, his arguments were in fact skeptical: "They admit of no answer and produce no conviction." In effect, therefore, Hume's position is that of following the principles of empiricism to their conclusion without any ancillary claim to knowledge of the inner workings of nature or of God. His conclusions are also something of a reductio ad absurdum of empiricism.

Nature of ideas

Hume begins by drawing a sharp distinction between impressions and ideas, impressions being the perceptions of sense and ideas the perceptions of the imagination or memory. In this he claims to be restoring the term idea to its original use. Every simple idea must have a corresponding impressionthe idea of red, for example, resembling the impression of redand complex ideas may be formed out of simple ideas. As with Locke, both impressions and ideas may be divided into those of sense and those of reflection, impressions and ideas of reflection being impressions and ideas of the mind's reflection on impressions or ideas of sense.

The criteria of the simplicity of an impression or idea are as much a problem here as with Locke. To have a simple impression is to have an elementary perception that cannot be further broken down into other perceptions, and this will function as a building block out of which the rest of knowledge may be constructed. Hume takes very strictly the principle that to every simple idea or perception of imagination or memory there must correspond an impression, or perception of sense, although at the very outset he admits a possible exception in the idea of a color in a series. We may have the idea of such a color from the principle of a series without ever having seen it. This possible exception, however, Hume refuses to take as important.

The principle that every simple idea must correspond to an impression is vital for a delimitation of the understanding and as a weapon against rationalism. Impressions and ideas can, however, be distinguished only by the superior force and vivacity of impressions; they cannot be distinguished in terms of their relations to physical objects or minds, for our knowledge of physical objects is derived solely from impressions, if at all. Likewise, among ideas, ideas of memory have a superior liveliness to ideas of the imagination. It is extremely doubtful whether this is always true, and this, in turn, casts doubt on any attempt to characterize remembering and imagining on empiricist lines by reference to the contents of the mind alone.

Theory of universals

Hume follows Berkeley in his theory of abstract ideas or universals. In his view there are no abstract ideas, strictly speaking; however, ideas can be particular in their nature and general in their representation. Hume's only addition to Berkeley's account is his attempt to indicate how this can happen through the association of ideas. The occurrence of one idea may dispose the mind to call up all other ideas associated with it. Hence, the understanding of a general word lies in the disposition of the mind to have the ideas of those things to which it may be applied. This is not a very plausible account in itself since the notion of understanding cannot be analyzed in terms of habits or dispositions of minds, but it is at least an attempt to tackle the problem. The solution is in accord with Hume's general approach; his account of belief is similar.

Space and time are difficult notions for an empiricist to deal with, for, as Kant pointed out, particular phenomena seem to presuppose space and time rather than vice versa. Hence, it is difficult to see how our ideas of space and time can be derived from our ideas of particular phenomena. Locke had nevertheless classified our ideas of space and time as ideas of modes. Hume attempts to deal with our perception of spatial extension and temporal duration in terms of the order in which impressions or ideas appear. But in consequence he has to admit that ultimately the impressions that are ordered in this way cannot themselves be extended or of extended objects, nor can they take time. In general, Hume's treatment of space and time is one of the more puzzling parts of his work.


Hume's greatest reputation derives from his treatment of causality, although his approach to this subject is similar to his approach to the problem of our knowledge of the external world or of ourselves. His approach is founded upon a distinction between different kinds of relation. There are "relations of ideas," which depend completely on the ideas related, and factual relations, which can be changed without changing the ideas. This is a distinction between logical and matter-of-fact relations, and it leads to a distinction between logical truths and factual truths that parallels Leibniz's. Hume is interested in the causal relation because he believes that it is the only matter-of-fact relation that can lead us from one idea to another. Causality is not a logical or a priori connection, but it is a connection. This assertion is of the utmost importance. Why, however, do we think that there is some necessity in causal connections? It cannot be a logical necessity; also, it cannot be derived from a more general necessity such as might be provided by a principle of universal causality, for Hume believes that such a principle must be contingent and that the evidence for it must be derived from our knowledge of particular causal connections.

He therefore proposes to "beat about the neighbouring fields." He notes that we generally take a cause to be antecedent to its effect and contiguous to it in space. More important, in experience there is a constant conjunction between cause and effect. In a sense these factors provide the basis for our belief in the necessity of the causal connection. Hume takes belief to be a lively idea associated with a present impression, and here the principles of the association of ideas again play a part. What makes an idea a belief is the feeling of being determined by habit or custom to pass from the impression to the associated idea. This feeling is an impression of reflection. It is in such an impression that Hume finds the source of our idea of necessary connection between cause and effect, for the "experimental method"the resort to experienceshould show us that there is no impression of power as such. It is due to habit or custom that we pass from cause to effect, and our belief in the necessity of doing so arises from the impression of a reflection of being determined to do it.

It is important to note just what Hume has achieved here. He has not in any way justified our belief in the necessity of the causal connection; he has merely attempted to explain the origin of the belief by giving a psychological explanation, not a philosophical justification, of the belief. But he has rejected all theories of occult powers in things, so that in one sense he may be considered to have said that what we mean by calling one thing the cause of another is that it is a uniform and contiguous antecedent of another event. To this extent his account is a reductive analysis; he analyzes our notion of cause by reducing it to notions that we understand. Yet Hume can find no justification for inferring the occurrence of one event from that of another; certainly, one event does not logically imply the occurrence of the other, but what other justification is there? Hume's conception of justification fails just because he recognizes no other kind of justification except that one thing logically implies another. Although Hume is commonly said to have raised the problem of induction, he made no real attempt to solve it himself, nor could he within his framework.

Knowledge of the external world

In Hume's account of knowledge of the external world the skepticism already implicit in his account of causality comes to the fore. Like Berkeley, Hume distinguishes between the beliefs of the "vulgar" and of the "philosophical system." The vulgar believe that we are aware of perceptions only, but they also believe that some of themour perceptions of primary and secondary qualitieshave permanent existence. The philosophical system holds that there is a distinction between objects and perceptions and that only objects are permanent. Hume claims to side with the vulgar, but he sees no reason to distinguish any perception from any other. The mind is like a theater in which scenes come and go. Yet he does admit that it is natural to believe in a world of permanent objects. Reason can provide no justification of this belief, but we can give a psychological explanation of it like the account of our belief in the necessity of causality.

Our impressions have a certain coherence and constancythat is, they fit together and recur in the same order after intervals. As a result, the imagination tends to carry on by custom or habit, and it attributes more regularity to objects of perception than they actually possess. Thus, we come to believe in a world of permanent objects, and we tend to reconcile what reason tells us of the interrupted nature of perceptions and what our imagination suggests about their regularity by a "philosophical" (as opposed to a commonsense) belief in a world of permanently existing objects. Nevertheless, a "direct and total opposition betwixt our reason and our senses" remains. Hume often speaks as if objects were just bundles of perceptions, but he has to deal with the belief that they are more than this. For such a belief he can give no justification, although he offers an explanation of its origin. In the last resort he can only recommend inattention to both our senses and our understanding. This is nothing if not skepticism.

Personal identity

Very much the same account is given of our knowledge of ourselves, a fact which may seem even more paradoxical. Once again, Hume uses the appeal to experience to indicate that we have no impression of the self. He rejects once and for all Berkeley's suggestion that we have a notion of the self. Belief in the self must therefore be parallel to belief in an external world, and Hume proceeds similarly. Belief in our identity through time must result once again from the coherence and constancy that exists between our impressions and ideas, as a result of which the imagination takes them to be impressions and ideas of a single self. Once again, however, no reason can be given for this belief, a fact that worried Hume more than his other tendencies toward skepticism. He returned to the topic in an appendix to the Treatise of Human Nature, but in the end he could find no way of ridding himself of his worry except a game of backgammon and a good dinner.

Account of perception

Hume rejected Lockean substance, with the result that perceptionsimpressions and ideasbecome the substantial entities in his ontology (as he in effect admits in Treatise, Book I, Part 4, Ch. 5). In retaining the terminology of perceptions, especially that of impressions, Hume clung to the skeleton of the causal or representative theory of perception. But the skeleton no longer had flesh, despite the suggestiveness of the terminology. Thus, Hume is forced to take his starting point from perceptions that are logically independent of any owner and any object. In one place he says that there is nothing objectionable in the idea of an unperceived perceptiona very odd notion. From this he has to build a world that fits the common supposition that there are physical objects and persons. The premises from which Hume derives his position are unacceptable, but given them, he can provide no reason whatsoever for belief in such a world and has to say that the belief is just a product of the imagination. This is skepticism with a vengeance, but it is the logical outcome of his approach.

reid's criticism

Hume's contemporary Thomas Reid (17101796) thought, rightly enough, that Hume's conclusions were manifestly absurd. Finding nothing wrong with the arguments presented by Hume, he concluded that the fault must lie in the premises and proceeded to attack the whole "way of ideas" which was the source of these premises. Reid maintained that it was necessary to make a strict distinction between sensation and perception, a distinction that the doctrine of impressions and ideas blurred. Reid was quite right about this, and his account of the nature of sensation and perception is interesting for its own sake. Sensation, he said, is an act of the mind that has no object distinct from that act, its prototype being pain. Perception is a much more complicated affair, involving a conception of an object and a belief in its existence. In many cases we fail to note the distinction because most of our perceptions are accompanied by sensations. Sensations provide in themselves no basis for inference about the nature of perceived qualities of things, although things cause the sensations. Reid expresses the relation between perception and sensations by saying that the sensations suggest the perceptions; sensations are natural signs of perceived qualities. This suggestion and the sign relationship involved are not a matter of experience, for we do not experience sensations in the same way that we perceive things (to have a sensation is not to perceive anything in itself). The relationship is a natural one like, he says, that between expressions of emotions and the emotions themselves.

Whatever the worth of this account, it is certainly a completely different analysis of perception from that of other philosophers of the period. In claiming to defend our ordinary beliefs by this account against the skepticism introduced by Hume, Reid set himself up as a philosopher of common sense, a position adopted by G. E. Moore in the twentieth century. Reid did retain some of the features of British empiricism, however, especially in a distinction between original and derived perceptions, a distinction that is in some respects very similar to that between simple and complex ideas or impressions. Moore similarly employed much of the apparatus of sense data invoked by modern empiricists.


Kant (17241804) represents the juncture of seventeenth-century rationalism and British empiricism. Brought up in the tradition of post-Leibnizian rationalism, he was, as he put it, awakened from his dogmatic slumber as a result of a reading of Hume. His critical philosophy, as expressed in the Critique of Pure Reason, can be characterized as an attempt to draw the boundaries between the proper use of the understanding and the improper use of reason in making assertions of speculative metaphysics and as an attempt to show how the understanding can provide objectively valid knowledge of those things which Hume left to the imagination.

classification of judgments

Kant bases his approach upon a twofold distinction between types of judgment. There is, first, a distinction between a priori and a posteriori judgments, the first being judgments whose truth can be known independently of experience, the second being judgments whose truth can be known only through experience. A priori judgments are pure when they involve only concepts that are themselves independent of experience. On the other hand, it is not a necessary condition of a judgment's being a priori that it should involve only such concepts. The concepts that are involved in a posteriori judgments, however, must be derived from experience, must be empirical.

The second distinction, that between analytic and synthetic judgments, is different. Analytic judgments are judgments about a thing that give no information about the thing, although they may serve to analyze or explain the concepts involved. This is because the concept of the predicate is contained, albeit covertly or obscurely, in the concept of the subject. The denial of these judgments involves a contradiction; hence, they correspond to what Leibniz called truths of reason. Synthetic judgments, on the other hand, do give information about a thing; in them the concept of the predicate is not included in that of the subject, and their denial does not involve a contradiction.

Kant now combines the two distinctions. Analytic a posteriori judgments are clearly impossible, but there is no difficulty, Kant thinks, about analytic a priori judgments or about synthetic a posteriori judgments. There remains the class of synthetic a priori judgments. It is on these, Kant thinks, that the claims for metaphysics rest, and it is these in particular that empiricists refuse to admit. Kant's program is to show whether and to what extent such judgments exist. The outcome of the program is that although metaphysics in the traditional sense is impossible, synthetic a priori judgments are admissiblefirst, in mathematics and, second, in the form of the presuppositions of objective experience or science. This program, which constitutes the critical philosophy, is Kant's substitute for traditional metaphysics.

synthetic a priori knowledge

The possibility of synthetic a priori knowledge means that not all knowledge about things can be derived from experience. Nevertheless, Kant thinks that all such knowledge is based on experience. It starts from what he calls intuitions, but since knowledge involves the possibility of making judgments about things, it cannot consist of intuitions alone; it must also involve concepts. To have a sensible intuition is to have a simple awareness of something by means of the senses. This awareness Kant analyzes in a way that derives much from the British empiricists. A sensible intuition consists, first, of a sensation as the content of the intuition. Its form consists of spatial or temporal extension. Hume had to admit that impressions have an order, but he drew the consequence that the impressions themselves were unextended and nontemporal when at their simplest. Kant generally argues that sensations have only intensive qualities, qualities that can vary only in degree. Nevertheless, since the intuition consists of the sensation plus the formthat is, its relations to other sensationsspatiotemporal form is something which is a necessary part of our experience. One cannot, as Locke seemed to suppose, build up ideas of extension from first impressions. Spatiotemporal form is a necessary, a priori characteristic of experience.

Since Kant, however, has assumed a theory of perception similar to the representative theory, this a priori spatiotemporal form applies only to things as they appear to usto phenomena. It does not apply to whatever may be thought to lie behind our experiences (things-in-themselves). This fact Kant expresses by saying that spatial and temporal characteristics (and primary qualities in general) are empirically real but transcendentally ideal. The characteristics in question are not merely subjective; they are objectivevalid for all menbut only in relation to phenomena, not to things-in-themselves. (Throughout, Kant's criterion of objectivity is the criterion of intersubjectivityvalidity for all men; he is pointing out that from the point of view of the critical philosophy something may be objective in this sense without being a feature of something independent of the mind.)

Pure a priori intuitions

Kant goes on to argue that we have not only a sensible intuition involving a priori spatiotemporal features but also a pure a priori intuition of space and time themselves. It is by virtue of this that the science of mathematics is possible. In order to do geometry, for example, it must be possible to make constructions in space, an idea that presupposes that we have an intuition of space. (Kant insists that this is an intuition, not a concept, but his reasons for this are complex and difficult to understand.) Arithmetic similarly presupposes an intuition of time. It is for this reason that mathematical judgments are both synthetic a priori and possible. Kant took it for granted that Euclidean geometry was the geometry of space, so that his account provided the justification of that geometry. It has often been suggested that the discovery of other geometries has undermined his case, and to some extent it has. But Kant would still have insisted that some intuition of space is a necessary condition of the possibility of any geometry, and there is something to be said for this position. Our concepts of space and time are not concepts that can be simply abstracted from experience.

categories of understanding

Space and time, then, provide the form of all experience, just as sensation provides the content. What is given in this way must be subsumed under concepts in judgment if knowledge is to result. "Thoughts without content," Kant says, "are empty, intuitions without concepts are blind." But in itself the formation of judgments is not enough for knowledge. The judgments that we make might be just the work of the imagination, as Hume in effect supposed in considering our knowledge of the external world. What criteria, then, have to be observed in the case of objectively valid judgments? Kant's answer is that such judgments have to conform to certain principles of the understanding and that these principles are derived from the pure or formal concepts, which Kant calls categories, of the understanding. Only insofar as our judgments conform to these principles can the judgments that we make about appearances be intersubjective, true for all men. Objectivity can be a question of this intersubjectivity alone because no valid judgments can be made about things-in-themselves. What, then, are these principles, and what are the categories?

In the section of the Critique known as the "Transcendental Analytic" Kant puts forward two arguments for categories. The first, the "metaphysical deduction," tries to argue for the existence of the categories directly, by finding the key to the list of categories in the traditional table of judgments provided by formal logic. To each of the traditional headings under which judgments can be classified logically, there corresponds, Kant believes, a concept that provides the principle of construction of an objectively valid judgment. The second argument the "transcendental deduction," attempts to show that the existence of categories of the understanding is a necessary condition of possible experience. The two arguments are complementary in that the transcendental deduction depends upon the metaphysical deduction for the actual list of categories, while the metaphysical deduction does not really show that categories are necessary to objective experience.

It is true that later, in discussing the principles derived from the categories, Kant brings forward specific arguments in each case, so that it might be said that the case for accepting each of these principles does not depend entirely on an acceptance of the metaphysical deduction. Yet, it is the metaphysical deduction alone that provides the guide to which categories and which principles we should seek. Today, the metaphysical deduction is almost universally rejected. There can be no validity in the attempt to derive a table of categories for objectively valid judgments from a table of judgments classified according to purely logical principles. It remains, therefore, that if the transcendental deduction is valid, it is possible to accept some categories as necessary, but apart from arguments for specific cases one cannot determine which categories are necessary by reference to any general rule.

Transcendental deduction

The argument of the transcendental deduction is very complex, and only the most general outline will be given here. First, the senses provide us with a manifold of sensations set out in space and time. Second, in order that they may form a unity, the understanding, with the aid of the imagination, has to synthesize them. The imagination helps us to see the manifold as a manifold in space, and in the form of memory it ensures that we also see it as a unity over a temporal period. Kant calls these two forms of synthesis the synthesis of apprehension and the synthesis of reproduction. Third, in the synthesis of recognition the manifold has to be given a principle of unity by subsumption under a concept, so that we see the manifold as a such-and-such. The results of this, however, could still be only subjective. In order to attain objective validity, the understanding must enable us to conceive of the manifold as united in an object. What Kant calls the transcendental unity of apperception is the awareness of experiences as part of one consciousness and as having an object, although neither the owner nor the object of those experiences can be found in the experiences as such. Objective experience presupposes these features; otherwise, the situation would be, as Hume in effect supposed, a mass of experiences whose connection with a person or objective world is merely contingent.

Fourth, the judgments we make about the manifold of experience thus unified must themselves conform to certain principles of unity. It must be possible, for example, to see certain connections within experience as that of ground to consequent, and our judgments must presuppose such connections in the things joined in them. Thus, we arrive at the idea that if objective experience is to be possible, it must conform to such categories as ground and consequent. The categories are concepts of the principles of connection of things in judgment, if that connection is to be more than a mere subjective one. They are categories because they are applicable to anything.

Transcendental judgment

The categories derived from the logical table of judgments according to the metaphysical deduction are purely formal. For example, Kant believes that the category of ground and consequent is derivable from the logical notion of a hypothetical judgment. This purely formal category of ground and consequent can be given content only by being applied to phenomena in such a way that it emerges in more material form in terms of the particular relation of ground and consequent that is applicable in the case of phenomenathat is, in terms of the relation of cause to effect. Kant formulated a doctrine of schematism to explain how we can apply the purely formal categories to experience. A schema is a kind of principle for the construction in the imagination of anything that falls under a given concept. It is that which enables us to identify a given object as an instance conforming to the concept. Thus, the schema for each of the categories can be thought of as the principle for the application of the pure category to phenomena in time. The notion of ground and consequent applied to phenomena in time emerges, as we have seen, as the notion of cause and effect (an essentially temporal notion). It is only from these schematized categories that it is possible to derive the principles of the understanding according to which all objectively valid judgments must be viewed. These principles Kant discusses and argues for separately.

In all this Kant believes he has explained how judgments about mere phenomena can be objectively valid although they are confined to these phenomena. The judgments in question are by no means applicable to things-in-themselves, to whatever lies behind phenomena. Such notions as those of a world lying behind phenomena and of a real self that is aware of them are noumena, and as such they must be thought of as limiting concepts. To treat such concepts as if they were concepts of an ordinary kind and to use them in a systematization of knowledge, as is done in speculative metaphysics, is wrong and is liable to produce fallacies. A noumenon is merely a "nonphenomenon," and the concept of a noumenon is essentially a negative one, but reason tends to treat such concepts as positive, and from this the illusions of speculative metaphysics stem.

critique of metaphysics

The judgments that the understanding allows us to make are conditional in that they are relative only to possible experience. Pure reason tends to assume an absolute, something unconditional, which provides the basis of the unity of all judgments of the understanding. It thus provides us with ideas whose proper use is only regulative, in that they are ideas of the goals or limits toward which the understanding may strive without being able to apply them directly to experience. To use these ideas, as speculative metaphysics does, as ideas under which we can directly subsume realitysuch ideas as those of the absolute unity of the thinking subject, the absolute unity of the world in space and time, and the absolute unity of the conditions under which anything can be thought at all, the entity of entities or Godis the source of antinomies and other forms of contradiction or fallacy. Kant sets out to expound these contradictions and their resolutions in detail, but it is impossible to follow the argument here. The section of the Critique known as the "Transcendental Dialectic" is a critique of rational psychology, speculative cosmology, and metaphysical theology with its attempts at a demonstration of the existence of God. Although there is much other material in the Critique, this section essentially completes the critical philosophy in its attempt to show that the understanding can provide objectively valid knowledge of phenomena and to reveal the limitations of the proper use of the ideas of reason.

An assessment of Kant's work in the theory of knowledge is difficult to give. It contains many extraordinary insights, although their detailed development often leaves much to be desired. Above all, perhaps, it takes as its starting point the analysis of experience provided by the British empiricists, and this undoubtedly limits it.

Post-Kantian Idealist Philosophy


German philosophy after Kant is in many ways a commentary upon Kant's philosophy, either as further development or opposition. The idealism which was so characteristic of nineteenth-century philosophy was begun when Johann Gottlieb Fichte (17621814) found fault with the Kantian view of things-in-themselves that are beyond the reach of knowledge and proceeded to reject the notion on grounds similar to those which are commonly used against any causal or representative theory of perceptionthere can be no good reason for believing in such things. With the rejection of things-in-themselves, even as a limiting concept, we are left merely with experiences or phenomena, and it is of these that, in the idealist view, reality must consist. The general problem of idealism that Fichte thus introduced was how it was possible to distinguish among experiences those which are purely subjective and those which are really objective. The problem is how we can distinguish between what is contributed by the mind and what is not, between the self and not-self, as Fichte put it. In Kant's view objectivity was equivalent to validity for all people, but that it was at all possible to distinguish between what was due to the mind and what was not seemed guaranteed only by the existence of things-in-themselves. With the rejection of the latter, experiences and experiencer became only two sides of the same coin. For this reason the general trend of idealism was toward the coherence theory of truththe view that experiences and judgments are true to the extent that they cohere with one another, forming a coherent system. This view was naturally associated with the doctrine of degrees of truththat judgments have varying degrees of truth to the extent that they cohere with each other. This more or less intelligible view was, however, complicated by being involved with the view that judgments about the empirical world have a very low degree of truth because they bring with them paradox and contradiction. The sensible world is therefore only appearance, and reality must be something else.


The belief that the sensible world is only appearance is perhaps less marked in G. W. F. Hegel (17701831) than in some of his idealist successorsfor example, F. H. Bradley. Hegel was influenced not only by Kant but also by Greek thought, especially by Platonist and Neoplatonist conceptions of an intelligible world of Forms with a structure of its own. Nevertheless, Hegel's relation to Kant may be roughly characterized by saying that he attempted to restore the functions of reason that Kant had forbidden. Whereas Kant had tried to justify the processes of the understanding while underlining the contradictions involved in an improper use of reason, Hegel tries to show that the understanding involves its own paradoxes, which can be resolved only by the use of reason; this, in Hegel's view, is by no means improper.

Dialectical method

Contradictions arise during the application of philosophical categories like those of the One and the many, so that the philosopher finds himself asserting both a thesis and its antithesis, in a manner similar to that expounded in Kant's antinomies. There is, Hegel thinks, a method which reason can pursue in order to resolve any such contradiction. Reason has to find a synthesis, some category that will reconcile those which produce the apparent contradiction. But the resolution may, in turn, find itself opposed to a further antithesis which demands another synthesis and so on. This method Hegel calls dialectic. According to him, it provides the key to understanding how the ideas of reason may be charted. In the end they will be seen to be dependent on the ultimate, absolute idea that provides the ground for everything else. Thus, the idea of something absolute and unconditional which Kant had rejected is restored.

It must be confessed, however, that as a method in the strict sense, Hegelian dialectic is sadly defective in that there appear to be no rules for its use. Hegel presents a series of insights, sometimes real, sometimes imaginary, into the relationships between very general and abstract philosophical ideas, like those of being and essence or consciousness, self-consciousness, and reason. Dialectic provides the architectonic according to which these relationships may be charted, and Hegel is excessively thoroughgoing in its use. The resultthe Hegelian systemis a complete map of all forms of knowledge and of all philosophical ideas, constructed on a single plan. The attempt is ambitious; the ground for its validity, slender in the extreme. It would be foolish, however, to deny the incidental insights.

Theory of knowledge

Hegel's theory of knowledge may be found partly in his Science of Logic and partly in his Phenomenology of Mind. In the Science of Logic he explains his view of Kant, criticizing Kant's trust of the understanding and Kant's, to Hegel's own mind, undue restriction of the functions of reason. Then, through the dialectic he charts the notions most central to reason, beginning with the opposition between the categories of Being and Nothing, the synthesis of which he finds in Becoming. These are notions which reason finds indispensable for any account of the world and upon which logic must depend.

In the Phenomenology Hegel sets forth his view of perception most clearly. There Hegel begins by pointing out that consciousness appears to be an apprehension of what is immediate, of what is, which is, it appears, a confrontation of the ego with something else (as Fichte also supposed). But sense knowledge proper must involve a subsumption of this immediate consciousness under universals or concepts, and, moreover, there is no way of grasping the particular that is thus subsumed under concepts except by reference to other concepts. Proper names and even words like "this" are, in Hegel's view, general words, since they apply to a multitude of different things (Hegel here ignores or fails to appreciate the way in which they so apply); hence, they furnish us with no means of identifying a particular independently of universals. Sense knowledge thus turns out to be a mediated knowledge, a knowledge which is possible only through the medium of universals and which is not a direct knowledge of reality.

There is, however, Hegel argues, a contradiction between the fact that we take ourselves to perceive things which are unitary entities and the fact that our knowledge of them can exist only through a plurality of universals which are themselves unconnected. This contradiction is resolved only because the intellect provides us with a higher universal that constitutes the basis or condition for applying the lower-order universals in sense perception. This higher universal is force, the idea of "lawlikeness." The unity of the objects of perception is due to the lawlike connections that exist between the universals under which they are subsumed. This is something that can be discerned only by the intellect, which thus produces the synthesis of the contradictions apparent within consciousness. This, of course, does not end the matter for Hegel, as the phenomena of consciousness are equally phenomena of self-consciousness. The opposition between consciousness and self-consciousness requires a synthesis by reason.


The kind of general argument that Hegel used can also be found in the English idealist Bradley (18461924), although he was far less attached to Hegel's method, referring to the dialectic as "the bloodless ballet of the categories." (In spite of being chronologically out of order, emphasis is placed on Bradley here because Bradley, although often difficult to understand, is generally easier than Hegel for English-speaking readers. The slight differences between Hegel and Bradley matter little in light of their essential similarity of purpose.) In his Principles of Logic Bradley argues that all judgments are only conditionally true in that the identification of the portion of reality which is their subject involves subsumption under universals. The judgment as a whole therefore says that some universal can be ascribed to reality only on condition that some other universal or universals may be ascribed to reality. All judgments, although categorical in that they are concerned with reality (reality being the subject of all judgments), are also hypothetical in this sense. Hypotheticals must likewise rest upon a ground; their truth is dependent upon connections within reality (Hegel's force).

In Bradley, however, there is greater emphasis on the idealism. For Bradley universals correspond to ideas, so that in judgment we are attaching an idea to reality. However, since what is known can be connections between ideas only, reality as we know it is a system of ideas joined by what he calls internal relations. A judgment is true to the extent that the ideas which it ascribes to reality cohere with the whole system of ideas. By an internal relation, as opposed to an external relation, Bradley means a relation that is more than a contingent one. The relations that form the system that constitutes what we know of reality are more than mere contingent relations, although they are not so close as to be logical entailments. All that we can know of reality apart from the ideas under which we subsume it is that it is experience; it is the bare fact of consciousness from which Hegel starts. Bradley is indeed Hegel made more palatable for English tastes.

It is only fair to add that in his most explicitly metaphysical work, Appearance and Reality, Bradley finds paradoxes in the notion of relations in general. The idea that two things may stand in a relation gives rise, he claims, to an infinite regress. What is relational must be set down as appearance only. It follows that judgment can never amount to absolute truth, for all judgment involves the setting of ideas in relation; all judgment, in other words, involves inference. No judgment can therefore furnish more than a limited degree of truth or be about something which has more than a limited degree of reality.

Judgments about the Absolute, the sum total of reality, may have a certain intellectual incorrigibility, since they will in effect ascribe to the absolute reality what is merely part of itself, but they cannot add up to truth itself. This is inevitable, since Bradley takes all judgment to be asserting the identity of subject and predicate, of reality and idea, while maintaining an unbridgeable gap between thembetween the "that" and the "what," as he puts it. The notion of judgment therefore involves a contradiction in itself, and for this reason the understanding is condemned. Intuition or immediate awareness gives us the bare fact of experience as constituting reality. What it is unconditionally and absolutely only reason can tell us. Whereas Kant had maintained that reason can tell us nothing of what is absolute and unconditional and that only paradox can result from the attempt to make it do so, the Hegelian doctrine espoused by Bradley is that the limitations of the understanding can be seen only by going beyond its limits to what is not finite and not conditioned but absolute. The claims for reason had never been pushed so far before, nor have they been since.


With Arthur Schopenhauer (17881860) there was a partial return to Kant. Schopenhauer thought that Hegel's dialectical method was barren because it ignored Kant's insights into the nature of reason and the understanding. Yet he retained Hegel's idealist approach. Kant's phenomena became presentations or ideas in a sense similar to Berkeley's; that is, they became subjective experiences. Schopenhauer thought that the world consists of ideas or presentations and that necessary connections judged to exist between them are, as Kant thought, merely conditional; however, Schopenhauer does not accept the paraphernalia of Kant's categories for justifying such judgments. In his view all justifications for claiming objective experience rest on the principle of sufficient reason, which takes various forms according to the form of knowledge involved. It acts as a logical ground (the ratio cognoscendi ) in connection with logical truths, as a ground connected with the features of space and time (the ratio essendi ) in connection with mathematical truths, and as causality (the ratio fiendi ) in connection with ordinary empirical phenomena. Thus, the notion of causality is made to play the role of all Kant's categories in relation to empirical phenomena; causality is their only ground for necessity in this sphere. Schopenhauer finally looks to the will as the only ground for action, for moral necessity. As one class of phenomenon, action finds its explanation only in the will.

Will as the thing-in-itself

The world as idea or presentation is only one half of Schopenhauer's philosophy (his main work is titled The World as Will and Idea ). Although he accepted the idealist framework, he thought that the demand for a thing-in-itself as the basis of all our ideas was inescapable. He finds the nature of the thing-in-itself in the will. Reality consists of the manifestations of one force, the will, which uses consciousness as an instrument for its own self-promotion. Only in art is there anything like freedom from it, for only there does the mind achieve a state akin to the contemplation of Platonic Ideas, a sort of permanency which is foreign to the general manifestations of the will.

Although the last part of Schopenhauer's thought had some influence upon the romantic movement in nineteenth-century German thought, it has not received much welcome from the mainstream of philosophers. However, Schopenhauer's theory of knowledge, contained in the part of his philosophy devoted to the world as idea, contains much of interest along lines which are, in origin, Kantian.

Late Nineteenth-Century Philosophy

Philosophical thought in Germany during the nineteenth century tended to be either romantic or neo-Kantian. Neo-Kantian philosophy came under empiricist influences from Britain, and at the end of the century under Franz Brentano and Alexius Meinong this finally led to a return to realism, a movement that not only produced phenomenology (perhaps the dominant philosophy in Europe today) but also to some extent influenced Bertrand Russell and other realist philosophers in the English-speaking world.


Brentano (18381917) held that the objects of psychology were mental acts. Each mental act had an immanent objectwhat Brentano called an intentional objectthus reviving scholastic terminology. These objects were a kind of internal accusative to the relevant act, as a judgment is to the act of judgment. They provide the content of the act. But what is the status of these objects? Clearly, the question is especially pertinent because it is possible to think of or make judgments about things that do not exist. How can a real act have an unreal object?


This was the problem that Meinong (18531920) took up. He postulated nonexistent objects to explain the possibility of our thinking, for example, of things that do not or cannot exist. Similarly, false judgments were said to correspond to what he called objectivesnonexistent states of affairs which would be facts if only the corresponding judgments were true. Objectives could not be said to exist, for they were not things, but they might subsist. From a linguistic point of view, this doctrine implied a realist theory of meaning, according to which the meaning of any expression was given by a corresponding entity. The fact that these entities were not themselves mental entities (although they gave content to what is mental) implied a return to realism in a more general sense. Objects could be real, according to Meinong, without being actual.


Edmund Husserl (18591938), another disciple of Brentano, started from very much the same point of view as Meinong, maintaining that the proper philosophical task was to investigate the essence of mental acts and their objects. Philosophy consisted, in his view, in an inquiry into the essence of different manifestations of consciousness and the essences with which they are concerned. To study this, it was necessary to strip off all presuppositions, metaphysical or otherwise. Husserl later emphasized this aspect increasingly. He adopted the method of epoch the bracketing of presuppositionsin a manner akin, as he pointed out, to the Cartesian method of doubt. This would lead to pure consciousness as the one absolute, the one firm thing, and from this the philosopher may turn back to investigate the essence of different phenomena as they appear to consciousness. Thus, in effect the initial realist point of view led back to one that was more like idealism. But this belongs, properly speaking, to the twentieth century.

j. s. mill

Meanwhile, in Britain the predominant philosophy at the beginning of the nineteenth century was sensationalism with its attendant associationism. James Mill (17731836) took a radically empiricist point of view, trying to reduce perception to merely having sensations and other mental phenomena to sensations plus the ideas associated with them. His son J. S. Mill (18061873) brought greater sophistication to this point of view and, in so doing, led to its downfall. Like his father, J. S. Mill wished to reduce all knowledge to experience, to the association of certain ideas with basic sensations. He expressed a great admiration for Berkeley's New Theory of Vision and its explanation of how we come to see things as at a distance. He thought it possible to explain in a similar way how we come to think of ourselves as perceiving a permanent world of things. We have expectations that take us beyond the immediate sensations because of the associations built up in experience between our immediate sensations and ideas of "permanent possibilities of sensation." Our ideas of material things are simply ideas of these permanent possibilities of sensation. Like Hume, Mill approaches this problem psychologically; he seeks to explain why we believe in an external world. To the extent, however, that he is inclined to add that things are simply these permanent possibilities of sensation, his view is the extreme empiricist doctrine of phenomenalism, the doctrine that all we mean by "material object" is something about our experiences.

Mill's main contributions to philosophy perhaps lie in logic, ethics, and politics. His general approach, however, is psychological, based on a conception of experience as atomistic sensations which could be linked to derivative ideas by the processes of association. Mill's general point of view is perhaps nowhere more obvious than in his account of such necessary propositions as those of mathematics. These are, in his view, simply very highly confirmed generalizations. The only necessity is psychological necessity.

Mill's view of knowledge came under attack in the latter half of the nineteenth century from many sources in Britain and elsewhere. In Britain perhaps the main attack came from the returning idealism, particularly as represented by F. H. Bradley. His main line of thought, as already discussed, was Hegelian with less emphasis on the dialectic and greater emphasis on the idealist point of view, according to which reality consists of experience organized in thought by attaching ideas to it in judgment. In Appearance and Reality Bradley sought to show that all features of the empirical world are only appearance and that reality must consist of a form of experience which is absolute and unitary and transcends all the contradictions of appearance. His criticism of Mill is that the pure sensory given is a myth; all the content of our knowledge must come by way of ideasthat is, through thoughtand association "marries only universals." Experience in itself is nothing.


Criticisms of a different kind came from France and America. In France there were few philosophical developments of interest during the nineteenth century until Henri Bergson (18591941). Bergson was an anti-intellectualist who emphasized life against thought. Much of his approach was therefore biological. The space and time of which we are conscious, Bergson thought, are continuous; the division of it into things and processes is due to the intellect, which carries out the division according to our biological needs. Perception involves an awareness of the possible moves that our body can make in relation to an object; in contrast, sensation corresponds to a simple response to a stimulus. Like Bradley, Bergson thought that the atomistic sensations of the sensationalist were a product of intellectual analysis. There are actually no basic experiences of this sort. We perceive things as our biological needs cause us to do so. Similarly with memory; our body acts like a sieve. Without the body our mind might remember everything, and this would be both useless and even disastrous biologically. Our body saves us from this, causing us to select only that which is biologically useful.

Because of this emphasis on biological utility, there is a relativism inherent in Bergson's point of view, and it has much in common with American pragmatism as instituted by William James. Bergson went further than James in his emphasis on life, however. His starting point was a thesis about time that is really outside the scope of this article. Roughly, his view was that the time of consciousness (la durée ) is continuous; the ideas which thought presents to us are or seem discontinuous. The continuity of consciousness must be due to an interpenetration of those ideas, and as a result, they form a developing series in which, given that each member developed from what has gone before, each member must be unique. The development itself is due to a vital spirit, and the same is true of the universe at large.


Bergson's emphasis upon the continuity of consciousness has its counterpart in James's thesis of the stream of thought. In his Principles of Psychology William James (18421910) insisted, in opposition to the sensationalists, that there were no atomic sensations or ideas; distinct ideas are selected phases of one stream of consciousness, and these phases make up a continuity because each idea has a fringe which overlaps that of its neighbors. Thus far, however, James was content to argue that our ideas are determined by what things there are. We must, he maintained, distinguish between knowledge by acquaintance and knowledge about something (not to be confused with a somewhat similar but really different distinction made by Russell). The baby is acquainted with the universe but he has not yet selected anything from the mass of sensation with which he is confronted. Thus, he knows nothing about anything. James was later to go further in Bergson's direction. In Essays in Radical Empiricism he rejected the distinction between thought and things, embracing a thesis that is known as neutral monism, the thesis that reality consists of one stuff (in this case experience) out of which both the mental and the physical are to be constructed according to the principles which govern each. From this continuous experience a plurality of thoughts and objects can be developed, a plurality of things related by concatenation only, as experience tells us they are. In this James set himself in opposition to his other bête noire, idealism, with its emphasis on internal relations and its denial of plurality.

peirce and dewey

James's special claim to fame (although some would say notoriety) is perhaps his status as the founder of pragmatism, although this, too, has Bergsonian affinities. The original source for the pragmatist point of view was C. S. Peirce (18391914), a rather isolated figure. A man with a great wealth of ideas, Peirce, as a nonprofessional philosopher, was to some extent outside the main stream of philosophical thought. He came to philosophy as a mathematician and scientist. He was opposed to all intuitions of the Cartesian type, largely because of his belief in the power of hypothesis and his disbelief in ultimate inexplicables. Perhaps his greatest contribution, however, lies in his theory of signs and meaning. In this connection Peirce maintained that our conception of anything is determined by our conception of the practical bearings of that thing. In sum, meaningfulness is a question of practical utility, and the meaning of any given concept or expression is given by its precise utility.

James turned this theory of meaning into a theory of truth, much to Peirce's disgust. He maintained that the test of the truth of any belief was its fruitfulness. To say that a belief is true is to say that it is good in this sense. Such is the pragmatic theory of truth.

For John Dewey (18591952) it is knowledge that is successful practice; propositions are merely instruments which may take us to the goal toward which experimental inquiry is directed. There is no final truth; instead, there is "warranted assertability" when the judgments which we make lead us to the abstract goal of science in accordance with scientific method.

Twentieth-Century Realism

The course of twentieth-century philosophy was not smooth, and it is therefore not easy to chart. No more than a sketch will be attempted here. Undoubtedly, however, the main philosophical event at the start of the century was a swing from idealism to realism. In America there were the neorealists, such as E. B. Holt, W. B. Montague, and R. B. Perry, influenced by James and his theory of neutral monism; in England there was Samuel Alexander, and in Germany there was Meinong. The most important figures in this revolution were G. E. Moore and Bertrand Russell at Cambridge, and of these Moore was the originator in this respect.


G. E. Moore (18731958) began with a criticism of Bradley. Moore thought that Bradley had not taken far enough his rejection of the view that we abstract our ideas from experience. Insisting that concepts or ideas should be regarded as the objects, the meanings, of our thoughts, Moore went on to argue that there must be propositions as the objects of beliefs. Things are merely collections of concepts and as such enter into propositions as their constituents. Propositions are what we believe when we hold any belief, true or false. This amounts to an insistence upon a distinction between any mental activity or form of awareness and its object. Moore's article "Refutation of Idealism" in his Philosophical Studies is founded on just this point, for he finds the refutation of the doctrine that esse est percipi in a distinction, somewhat after the manner of Brentano, between the act of awareness in perception and the object of awareness, between consciousness and its object. Idealists, he maintained, failed to notice the distinction.

Moore was later to give up his doctrine of propositions, for in considering the problem of false belief, he said that "there do not seem to be propositions at all, in the sense in which the theory demands them." If there were, there would have to exist something corresponding to false beliefs, and the fact of its existence would make the beliefs true, not false. Belief cannot therefore consist in a relation between ourselves and an object. The rejection of propositions in this objectively existing sense may look like the abandonment of the very foundations of his realism, and so in a sense it is. But Moore did not give up the view that we do know of a reality independent of our minds, and when he abandoned the account of false belief which implied the existence of objectively existing propositions, he nevertheless maintained categorically that we must not give up the view that truth somehow consists in correspondence with reality.

Existence of the external world

Moore sometimes said that he never doubted that we do know things about reality that we ordinarily think we know. Therefore, he was not influenced by the usual skeptical arguments against this position. In his view the real philosophical problem was to analyze what we mean when we say that we have this knowledge. Moore has generally been one who raises the difficulties about this problem rather than one who gives the answers. He has definitely maintained, however, that we do have knowledge of many different kinds of thing, and in his notorious "Proof of an External World" (Proceedings of the British Academy, 1939) he gave as a good argument for the existence of such a world the fact that we can point to objects in it. Thus, he held up his hands, saying, "Here is one hand, and here is another," to prove the existence of an external world. Moore's thought moved toward the view that what requires defending is common sense, ordinary beliefs such as the belief that there are objective things, like his hand, in the world. When metaphysicians say such things as "Time is unreal," this is an affront to common sense and demands explanation. Ludwig Wittgenstein later said that the view defended by Moore was not strictly common sense, since it was a philosophical point of view. This seems correct; what Moore meant by common sense was a general realist point of view.

Account of perception

In his analyses of what we mean when we claim knowledge, however, Moore's discussion follows the lines of those which have been more influenced by skepticism. Thus, in his account of perception he brings in sense data as what we actually see or directly apprehend when we look at something. Characteristically, he distinguishes between the sense datum (the object that we actually see) and the sensation which we might be said to have of it. But in using the words "direct apprehension" and "actually seeing," he suggests that he wants that indubitability which other philosophers have sought as an answer to the skeptic. Direct apprehension cannot be of the physical object, for, he argues, when an envelope is held up, it cannot be this that we actually see, since some people may fail to identify it as such. There is room for error about the identity of the object but not, perhaps, about its color or even shape as seen, so it must be these that we actually see. Moore's main worry is about the exact relation that exists between sense data and physical objects, both of which, he thinks, certainly exist. The answer that he would like to give is that sense data are parts of the surfaces of physical objects, but the fact that different people may have conflicting views of these objects prevents him from giving this answer. The right answer and, for the same reason, the precise nature of sense data were always a puzzle to him.

It may be wondered why Moore felt it necessary to bring in sense data at all. The answer must be that for Moore there had to be things which we just know, and although we may know of the existence of physical objects, we may not be so sure of their exact qualities. Hence, the possibility of error plays a part in Moore's thought just as it has done in that of others, even if the influence of skepticism is not so explicit. Direct apprehension fills the same role in his philosophy of perception as intuition in his ethics. The notion has its parallel in the conception of knowledge employed by his Oxford contemporary John Cook Wilson and his disciples, such as H. A. Prichard. They thought the trouble with idealism lay in its failure to see that there was a distinct state of mind, knowledge, in which there was no possibility of error.


Bertrand Russell (18721970) was first an idealist but was converted to realism by Moore. From his early study of Leibniz, Russell took the view that philosophy consists in the analysis of propositions, and his interest in logic also brought him to a concern with language. During the early 1900s he became interested in Meinong, whose realism seemed to confirm Moore in Russell's line of thought. However, he came to think that Meinong's supposition that there had to exist objects to explain our ability to think of things which do not exist in fact, such as round squares, showed, as he put it, an insufficiently robust sense of reality. Partly in response to this and partly in response to a more complex theory of meaning put forward by the mathematical logician Gottlob Frege, Russell set forth the theory of descriptions.

According to the theory of descriptions, phrases of the form "the so-and-so" are incomplete in meaning. They have no meaning (in the form of an object of reference) in themselves; to give their meaning, it is necessary to analyze the meaning of the whole sentence in which they occur. Sentences of the form "The so-and-so is F " are really tantamount to composite sentences including as a part the sentence asserting that something exists corresponding to the description "the so-and-so." Where nothing of the kind exists, the whole proposition is false ("proposition" here being equivalent to "statement," not the objectively existing object of beliefs). The theory of descriptions became a tool of analysis, and Russell used it in many connections.

Theory of knowledge

The main importance of the theory in epistemology is connected with Russell's distinction between knowledge by acquaintance and knowledge by description. Knowledge by acquaintance is Moore's direct apprehension, but Russell has always been more concerned with the justification of claims to knowledge than was Moore. For Russell it was important that all knowledge be founded on knowledge by acquaintance, if it was to be possible at all, for only in knowledge by acquaintance is error absolutely impossible. In Problems of Philosophy Russell gave a list of entities of which we have knowledge by acquaintancesense data, memory data, the self, and universals. Of physical objects we have only knowledge by description because here error is possible.

Russell also declares that "every proposition which we can understand must be composed wholly of constituents with which we are acquainted." (By "proposition" he meant the objectively existing entity in the early Moorean sense.) This is possible only if anything of which we have knowledge by description is reducible to things of which we have knowledge by acquaintance. Physical objects, for example, must be reduced to sense data or, since they are not always being perceived, to a combination of sense data and sensibiliaactual and possible sense data. They must, as Russell puts it, be considered as logical constructions from sense data; they are simply bundles of sense data and sensibilia.

From the logical point of view, names of physical objects are disguised descriptions, and the theory of descriptions shows that it is not necessary to suppose the existence of a special class of entities called physical objects in order to give propositions about them a meaning. What we are acquainted with when we perceive a physical object is a number of sense data; the physical object we know only by description, and any statement expressing a fact about it is a statement about a description. This statement is analyzable, so that it contains an existential proposition about something answering to that description, according to the theory of descriptions; it is, that is, about something falling under a set of universals, and these are objects of acquaintance as much as sense data are. The notion of an object of acquaintance is closely connected with that of a logically proper name, an expression that cannot fail in its reference. Descriptions can, of course, fail in that there may be nothing answering to them. With the reduction of physical objects to sense data, knowledge of physics is preserved, and this has always been a cardinal point in Russell's program. On the actual status of sense data, Russell's opinion varied, but when most impressed by physics, he made them, paradoxically enough, entities in the brain.

Logical atomism

At the beginning of World War I, when Russell had come under the influence of Wittgenstein, he held that the ultimate constituents of the universe are atomic particulars, which are terms of relations in atomic facts. All other facts were to be built up from atomic facts by the processes of logic. The atomic particulars are sense data, and the relations supply the universal element in the fact.

The theory of logical atomism is a metaphysical theory rather than an epistemological one, and this is even more obviously true of Wittgenstein's version in his Tractatus Logico-philosophicus. Russell differs from Wittgenstein in his more explicit interest in the nature of the constituents of atomic facts, and it is clear that his choice of these constituents was determined by his desire to found all knowledge on knowledge by acquaintance. The one exception to the program that Russell found at this stage was that of mental states like belief, for he could not see how statements about beliefs could be reduced to more elementary statements. In order to deal with the problem, he has flirted with behaviorism, since if belief is analyzable only in terms of behavior, this, in turn, may conceivably be explained in terms of sense data.

Russell changed his mind over the details of his logical atomism, especially when he was influenced by the logical positivists, but the framework remained the same. His account of memory is similar to that of perception; it is founded on memory experiences only contingently related to the past, although they may have a feeling of familiarity. The memory data are objects of acquaintance, but the past itself is not. In one general respect, however, Russell acknowledges that empiricism failsin our knowledge of the postulates on which, in his opinion, inductive inference and, therefore, science rest. Induction, he thinks, is founded on habit, and the principles implicit in such habits cannot themselves be derived from experience. Despite his belief in the limitations of empiricism, Russell never wavered in his defense of realism. He always embraced and keenly defended the correspondence theory of truth.

Nature of mathematics

Nothing has been said thus far of Russell's work on the foundations of mathematics, especially his great contribution, written with A. N. Whitehead, in Principia Mathematica, although this is in a sense part of his epistemology. In Principles of Mathematics Russell held that mathematical propositions were synthetic, but when, influenced by Frege, he embarked on the attempt to reduce mathematics to logic by deriving it from a small number of axioms containing only logical notions, he held mathematical propositions to be analytic. This is too complicated a matter for discussion here. The exact nature of mathematical propositions is still a matter of dispute, but the attempt to reduce them to logic may now be seen to be a great and splendid failure.


It may be noted briefly that Russell's partner in Principia Mathematica, A. N. Whitehead (18611947), took a very different road in epistemology. He tried to explain the properties of things in terms of their relations to one another. In perception the mind tries to graspor in Whitehead's term to "prehend"a part of the system of nature around it; it is reacting to the environment in biological fashion. There are no atomic sense data; to suppose that there are is to be liable to the "fallacy of misplaced concreteness"the view that because science has a concept (such as that of an instant), there must be entities of this sort in experience. To some extent Whitehead's thought is in the idealist tradition, but it also contains a certain Platonism. The particulars of which nature is composed, he thought, are events; the permanent characteristics that we recognize in them are objects. The objects are, as he puts it, ingredient "into events," not into just one event but through an indefinite neighborhood of events. We do not, that is, see single things with isolated characteristics; we view them as part of a system. Whitehead's thought is, however, difficult, and little can be done to make it intelligible in a short space.

Logical Positivism

Wittgenstein's Tractatus influenced a group of philosophers in Vienna who were mainly interested in the philosophy of science after the empiricist fashion of Ernst Mach. Wittgenstein had said that to understand a proposition is to understand what it would be like for it to be true. The Vienna circle, as this group became known, wrongly interpreted this as a general criterion of significance, and so the verification theory of meaning was born. According to this theory, meaningful propositions must be either analytic or empirically verifiable. The propositions of mathematics and logic were thought to belong to the first class, and the propositions of science to the second. Metaphysical propositions, belonging to neither group, were declared meaningless.

The members of the group differed over the details of this scheme, and a progressive relaxation in its rigor gradually took place. One of the biggest problems was the status of the verification principle itself, for on the face of it it is neither analytic nor empirically verifiable. The eventual outcome for some members of the group was to view it as a recommendation only. There were also other problems. The initial aim of the movement was, above all, to lay the foundations of science. Scientific propositions had to be preserved and metaphysics excluded. It became apparent that it was difficult, if not impossible, to provide a formulation of the verification principle which fulfilled both goals.

Moritz Schlick (18821936), the original leader of the group, felt compelled to interpret scientific laws as rules rather than statements. Another problem lay in the meaning of the phrase "empirically verifiable." Schlick held that ultimately there had to be a direct confrontation with experience. Other members of the movementfor example, A. J. Ayer (19101989)held that there had to be basic propositions which were directly and strictly verifiable (and thus absolutely incorrigible), although others could be indirectly verified by reference to these. This led to a distinction between strong and weak verification; propositions about physical objects, for example, might be only weakly verifiable, since an indefinite number of propositions about immediate experience would have to be invoked in order to verify them conclusively. In this, positivism was associated with the thesis of phenomenalism that statements about physical objects are analyzable into a collection of statements about sense experiences. The fact that such an analysis must be indefinitely long has resulted in a progressive modification of the thesis on the part of its main proponentsfor instance, Ayer.


Schlick's view brought with it the correspondence theory of truth. Otto Neurath (18821945), on the other hand, held that this involved an attempt to go outside the web of language and ran the risk of a lapse into metaphysics. Neurath maintained that there had to be "protocol propositions"observation reports on which science might be foundedbut that these need not be reports of immediate experience. The truth of a protocol proposition was determined by its coherence with other propositions making up the language of science. Thus, Neurath embraced a coherence theory of truth. This general line was developed by Rudolf Carnap (18911970) into a form of conventionalism. He put forward a "principle of tolerance," maintaining that logic had no morals; verificationism thus became a proposal for the best way of developing the language of science. Whereas Schlick and the earlier Carnap had maintained that basic propositions must be about immediate experience, Neurath had maintained the thesis of physicalism that protocol propositions must be in the language of physics. In line with his relaxation of the criteria, Carnap came to accept what he called the "thing-language"the language of the commonsense worldas the basis for scientific language. Today, positivism in its strict form is more or less a thing of the past.

Nature of science

It is noteworthy that Karl Popper (19021994), who was not a member of the movement but who was influenced by it and influenced it, held that the key to an understanding of science lay not in verifiability but falsifiability. He put this forward, however, not as a theory of meaning but as a criterion for the demarcation of science from metaphysics. The generalizations of science are, because of their very form, unverifiable, but they are falsifiable, whereas the propositions of metaphysics are not. Popper developed these views into a thesis about science as based on the hypothetico-deductive method. The aim of science is to put forward bold hypotheses, the deductive consequences of which must be subject to rigorous testing and criticism. This view is associated with a form of skepticism, for Popper sometimes maintains that we can never know the truth. The best that we can do is to put forward hypotheses and subject them to rigorous tests, for this is the way in which science progresses. Truth itself is just an illusion.

Contemporary Movements

Contemporary philosophy is in an untidy state of nonuniformity. In Europe perhaps the most prevalent philosophy is a phenomenology deriving from Husserl. This movement is also associated with existentialism, originally a reaction against the superrationalism of Hegel and, therefore, to some extent a form of irrationalism. Existentialists have added little to epistemology; they tend to take for granted the existence of an objective world, aiming only to present a picture of it and of man's place in it. Those existentialists who derive something from Husserlfor example, Jean-Paul Sartre and Maurice Merleau-Pontyare concerned mainly with descriptions of forms of consciousness, with phenomenology as descriptive psychology.

ordinary-language philosophy

Perhaps the most significant movement, apart from latter-day positivism and pragmatism, is the so-called ordinary-language philosophy, today most closely associated with Oxford. The leading spirit of the movement is, however, the Cambridge philosopher Ludwig Wittgenstein (18891951), who has had an immense influence. His work is not easy to summarize; it is in part a series of comments upon his earlier logical atomist views and the theory of meaning that it espoused. Only a small part of his work can be mentioned here. He has criticized the attempts implicit in much sense-datum philosophy to construct a private language by arguing that the results of such attempts would lack the essential conditions of a language. There would be no way of distinguishing between the occasions on which one was following a rule in applying an expression and those on which one was making a new decision so to apply it. He has also stressed the importance of bringing back terms to the language game (as he calls possible languages) that is their original homeordinary language. This, he maintains, is perfectly in order as it is; the important thing is to examine the uses to which expressions are put, with the recognition that language is a form of life and must be treated accordingly. Among other things this has led to the recognition of truths which are necessary but not analytic, truths which he calls "grammatical." These are truths which express nonanalytic connections between concepts. The emphasis upon such truths and the arguments which lead to them on the part of followers of Wittgenstein is in a sense a partial return to Kant. (The distinction between analytic and synthetic truths in general has in any case come under fire from several quarters, especially from the American logician W. V. Quine [19082000]. But his emphasis on the necessity of assessing the status of propositions only within a system is something more of a move toward the idealist point of view.)

The appeal to ordinary language has been used for many purposes. Gilbert Ryle (19001976), for example, has used it in order to plot, as he puts it, the logical geography of mental concepts such as mind, belief, or will. He, too, has attacked the notion of sense data, and he has made and emphasized an important distinction between knowing how and knowing that. For present purposes the main importance of the appeal to ordinary language lies in its confrontation of the skeptic. The most stringent appeals to usage (rather than merely to the functions of language, as in Wittgenstein's case) have been made by J. L. Austin (19111960), who has emphasized the extent to which many philosophers including skeptics, have departed from our ordinary use of words. It is clear, however, as Austin in effect admitted, that an appeal to what we ordinarily say cannot settle these issues, however much it may be a good first move. Arguments are first required in favor of our ordinary way of speaking.

It has been argued, for example, that anyone who says that we never know anything but only believe or suppose it robs the concept of belief of an essential contrast with knowledge, without which it would be meaningless. This argumentthe so-called argument from polar conceptsis invalid, because a philosopher can use the concept of belief as long as he has the concept of knowledge, as long as he knows what it would be like to know; he does not have to admit that anyone knows anything as a matter of fact.

Another argument is that we can never deny that we have knowledge altogether, because this would be denying the existence of the paradigm case by reference to which we have learned the meaning of the word knowledge. This argumentthe so-called paradigm-case argumentfails, in the opinion of the present writer, because it assumes that meaning is given by the applications of a term. Whereas it might be difficult to see how we could have come to attach meaning to a term unless we had learned some of its applications, it is not logically impossible that we should have done so. Hence, more complicated arguments are required.

This is the situation. Most philosophers would agree that if we are to be said to know a proposition p, we must believe p, p must be true, and we must have good reasons for believing in p. There is perhaps little argument over the first two conditions, although there might be some hesitation over the details. The problem is what counts as good reasons for believing in p. In the ordinary way we recognize different reasons according to the nature of the proposition involved. The skeptic denies that any of these are sufficient, and it is impossible to produce any knockdown argument which will dispose of the skeptic's claim. Each application of this claim must be assessed on its own merits, and the answer to the skeptic must therefore be a dialectical one in the Socratic sense. But the very existence of recognized forms of knowledge presupposes that there must be such knowledge. This is, however, only a presumption, not a proof.

Traditionally, and for good reason, skepticism has had biggest sway in connection with claims to knowledge of objects of perception, knowledge derived from memory, knowledge of other minds, and inductive knowledge. In each of these cases the skeptic may present too high a standard of knowledge, which cannot, in the ordinary way, be attained. But the temptation to accept such a standard may be increased by adopting certain views about the nature offor example, perception or memory. If perception is thought of as merely having sensations or memory as merely having images or ideas in the mind, there is necessarily a gap to be crossed from our own minds if there is to be objective knowledge in these fields. Hence, the talk of knowledge of an external world, external to our own minds. With a different conception of memory or perception, in which an essential connection is recognized between memory and the past or perception and an objective world, this gap is removed. This is not to get rid of the skeptical problem at one fell swoop, since it remains a question whether and under what conditions these concepts of perception and memory can be applied. This, however, is merely the general problem of how knowledge is possible in different fields; it can be dealt with dialectically. In other words, an incorrect conceptual analysis can worsen the skeptical problem, but a correct one cannot solve it. But the presumption remains that objectivity is possible in these fields, even if it is incapable of proof. Each doubt can be alleviated only by argument; there is no overall answer. Finally, the supposition that an answer can be provided by showing that there are forms of knowledge in which error is logically excluded and which are therefore absolutely indubitable is an illusion. First, there is no indubitable knowledge; second, it is not necessary for the general possibility of knowledge that there should be. Much of the history of epistemology has depended on this illusion.

See also Abelard, Peter; Alexander, Samuel; Antisthenes; A Priori and A Posteriori; Arcesilaus; Aristotelianism; Aristotle; Augustine, St.; Austin, John Langshaw; Ayer, Alfred Jules; Bergson, Henri; Berkeley, George; Boethius, Anicius Manlius Severinus; Bradley, Francis Herbert; Brentano, Franz; Carnap, Rudolf; Carneades; Cartesianism; Chrysippus; Cosmological Argument for the Existence of God; Couturat, Louis; Descartes, René; Dewey, John; Dialectic; Duns Scotus, John; Empiricism; Epicurus; Fichte, Johann Gottlieb; Frege, Gottlob; Geulincx, Arnold; Gorgias of Leontini; Hegel, Georg Wilhelm Friedrich; Hellenistic Thought; Heraclitus of Ephesus; Hobbes, Thomas; Holt, Edwin Bissell; Hume, David; Husserl, Edmund; Innate Ideas; Intuition; James, William; Kant, Immanuel; Knowledge, A Priori; Leibniz, Gottfried Wilhelm; Leucippus and Democritus; Locke, John; Logical Positivism; Mach, Ernst; Malebranche, Nicolas; Meinong, Alexius; Merleau-Ponty, Maurice; Mill, James; Mill, John Stuart; Montague, William Pepperell; Moore, George Edward; Neo-Kantianism; Neoplatonism; Neurath, Otto; New Realism; Ontological Argument for the Existence of God; Paradigm-Case Argument; Parmenides of Elea; Peirce, Charles Sanders; Perry, Ralph Barton; Plato; Platonism and the Platonic Tradition; Plotinus; Popper, Karl Raimund; Porphyry; Pre-Socratic Philosophy; Protagoras of Abdera; Quine, Willard Van Orman; Rationalism; Realism; Reid, Thomas; Roscelin; Ryle, Gilbert; Sartre, Jean-Paul; Schlick, Moritz; Schopenhauer, Arthur; Sextus Empiricus; Skepticism, History of; Socrates; Spinoza, Benedict (Baruch) de; Stoicism; Thomas Aquinas, St.; Universals, A Historical Survey; Whitehead, Alfred North; William of Ockham; Wittgenstein, Ludwig Josef Johann.


For the texts of individual philosophers see the separate articles devoted to them.

introductory books

Ayer, A. J. The Problem of Knowledge. London: Macmillan, 1956.

Russell, Bertrand. The Problems of Philosophy. London: Williams and Norgate, 1912.

Woozley, A. D. Theory of Knowledge. London, 1949.

general histories

Copleston, Frederick. A History of Philosophy. 7 vols. London: Burns, Oates and Washbourne, 19471975.

Hamlyn, D. W. Sensation and Perception. London: Routledge and K. Paul, 1961. A history of the philosophy of perception; contains bibliography.

O'Connor, D. J. A Critical History of Western Philosophy. New York: Free Press of Glencoe, 1964.

other recommended titles

Ayers, Michael. Locke: Epistemology and Ontology. London: Routledge, 1991.

Buryeat, M. F., ed. The Skeptical Tradition. Berkeley: University of California Press, 1983.

Fine, G. "Knowledge and Belief in Republic V." Archiv fur Geschichte der Philosophie 60 (1978): 121139.

Fogelin, Robert. Hume's Scepticism in the Treatise of Human Nature. London: Routledge and Kegan Paul, 1985.

Hankinson, R. J. The Sceptics. London: Routledge, 1995.

Lehrer, Keith. Thomas Reid. London: Routledge, 1989.

Lovejoy, Arthur O. The Revolt against Dualism. La Salle, IL: Open Court, 1930.

Moser, Paul K. "Epistemology (1900Present)." In Routledge History of Philosophy. Vol. 10: Philosophy of the English Speaking World in the 20th Century, edited by John Canfield. London: Routledge, 1996.

Moser, Paul K., and Arnold VanderNat, eds. Human Knowledge: Classical and Contemporary Approaches. Oxford: Oxford University Press, 1987; 2nd ed., 1995; 3rd ed., 2002.

Pears, David. Hume's System. Oxford: Oxford University Press, 1990.

Popkin, R. The History of Scepticism from Erasmus to Spinoza. Berkeley: University of California Press, 1979.

Shope, R. The Analysis of Knowing. Princeton, NJ: Princeton University Press, 1983.

Sosa, Ernest. "How to Resolve the Pyrrhonian Problematic: A Lesson from Descartes." Philosophical Studies 85 (1997): 229249.

Stroud, Barry. Hume. London: Routledge and Kegan Paul, 1977.

Stroud, Barry. The Significance of Philosophical Scepticism. Oxford: Oxford University Press, 1984.

White, Nicholas. Plato on Knowledge and Reality. Indianapolis: Hackett, 1976.

Williams, Bernard. Descartes: The Project of Pure Enquiry. Atlantic Highlands, NJ: Humanities Press, 1978.

D. W. Hamlyn (1967)

Bibliography updated by Benjamin Fiedor (2005)