Year 2000 Introductory Lecture Notes, Philosophy 102

The topic of this course is theory of knowledge or epistemology. The term ''epistemology" means "treatment of the subject of knowledge," the Greek word 'episteme' meaning 'knowledge.' Western philosophers have engaged in epistemology since at least the time of Plato in ancient Greece, but the practice did not have a distinguishing name until the nineteenth century, when some German philosophers coined the term 'Erkenntnistheorie,' which is translated as 'theory of knowledge.' Theory of knowledge is now one of the "core" areas of contemporary analytic philosophy. At the end of the twentieth century, there are specialists in epistemology in virtually all philosophy departments granting graduate degrees, and the literature has become vast. These lectures will cover the second edition of Theory of Knowledge by Keith Lehrer of the University of Arizona, one of the top American departments in epistemology. (One person, Brian Leiter, ranks it as first, while another, Keith DeRose, disagrees.)

Issues in Epistemology

Accounts of knowledge frequently do not take the form of a theory in the manner of studies of gravitation or of the evolution of species. The issues we will be studying include understanding what knowledge is, whether (and to what extent) we have any knowledge, the conditions under which we have any knowledge we do, and more. Some of the best-known epistemologists were skeptics who held that there is no phenomenon of knowledge about which there could be a true theory. (Lehrer says that they have a theory of ignorance or "agnoiology" (page 206).)

The most basic question in epistemology is about the nature of knowledge. Whether we have any knowledge, and if so, how much, cannot be settled unless we agree about what knowledge is. Since this is a rather broad question, it might best be approached by narrowing it down, asking first who or what might be capable of knowledge. Most of philosophers historically have taken adult human beings to be the paradigm case of the sort of being that can know. One reason we are investigating knowledge is that we, adult human beings, are vitally concerned in determining whether we are in possession of it. But are babies, non-human animals, or machines the sorts of beings that might have knowledge? Any answer we might give to this question about knowers has implications for the question about knowledge itself. If, for example, we conclude that to know, a being must have a certain capacity for understanding (perhaps lacked by a computer, which only responds blindly to instructions), then understanding is a component of knowledge.

Assuming that at least adult human beings are knowers, epistemologists from the very start made a distinction between knowledge and mere opinion. This distinction helps us to understand what knowledge is. It gives us a contrast with something that is not knowledge but has something in common with it. Both knowledge and opinion (in adult humans) require that someone thinks that something is the case. Two people may have opposite opinions about an issue. One thinks that something is the case and the other thinks it is not the case. They cannot both be right, and the one who is wrong lacks knowledge. Distinguishing knowledge from opinion in this way gives us a start toward understanding what knowledge might be. But as Plato argued, someone holding a correct opinion would still lack knowledge if the correctness of the opinion were a matter of luck. Traditionally, it has been held that a correct opinion needs to be "supported" or "secured" by evidence. And so the focus of most of Western epistemology has been on the study of evidence.

If knowledge is distinguished from opinion by the strength of evidence, a new question arises: how strong must the evidence be for a person to know something? My friend once told me that my neighbor was been stealing pears from the pear tree in my back yard. And in fact she was. Did my friend's testimony give me enough evidence so that I had knowledge rather than true opinion? Or did I need something stronger; for example, to see a person who looks just like my neighbor taking pears from my tree or to hear the neighbor confess?

A skeptic would say that even this stronger evidence is not strong enough for knowledge. Someone might be disguised as my neighbor in order to avoid detection, for example. I am capable of being fooled by such a disguise. Or, in a more extreme example, there might be space aliens with the ability to mimic human forms who fooled me completely. How strong our evidence must be for us to have knowledge will be a central concern of this course.

Ancient Greek Theories of Knowledge

Ancient Greek philosophy got its start with the idea that things are not always what they appear to be, that human beings are subject to deception. Starting from the common-sense view that we get information about the world through our senses, the early philosophers began to suspect that the information thus received is defective in certain ways. Most obviously, it is limited to those objects and events within the range of the senses. Even this information is often distorted, as when a square tower in the distance looks round. This distortion extends to our memories of what we have perceived through the senses. And we cannot sense the future (with normal sensory powers).


The ancient atomist Democritus, a contemporary of Socrates, proposed that the senses yield no knowledge, or at best, "obscure" knowledge. "The truth is that what we meet with perceptually is nothing reliable, for it shifts its character according to the body's dispositions, influences, and confrontations" (Wheelwright, The Presocratics, 182). The most telling limitation of the senses is their inability to see objects smaller than a certain size. "Whenever [an investigation reachers the point where] obscure knowledge can no longer see the objects because of their smallness, and also cannot hear or smell or taste them nor perceive them by touch, [the investigator must then have recourse to] a finer means of knowing" (Wheelwright, 183). By this, Democritus meant rational investigation, the kind of reasoning that produced a theory of reality, a "metaphysical" theory, according to which what is real is moving mass of atoms in a void.

There is a deeper failure of sense-perception than its mere grossness: it does not satisfy our thirst for explanation. We perceive myriad changes going on around us all the time. The senses might be able to detect, with greater or lesser accuracy, that they have taken place, but they are silent on the question of why they have. The ancient Greeks in general were attracted to the idea that the changes we sense in the world can be understood through an explanatory principle ("arche") that is not accessible to the senses themselves.

For Democritus, what explains what we sense is the action of tiny indivisible particles of matter, atoms, which cannot be detected by the senses. What appears to the senses as a smooth, continuous surface, say a slab of polished marble, is in reality a conglomeration of atoms with a great deal of empty space between them. For the atomists, things in reality are not what much like what they appear to be.

According to atomist metaphysics, some of the properties which we attribute to the marble on the basis of sense-perception are properties the marble does not possess. The slab of marble is not smooth or continuous. It does not have the green color we see or the coolness to the touch that we feel. On the other hand, the size and shape of the marble are more or less the way we perceive them to be. This bifurcation of properties into real and apparent came to be known in the seventeenth century as the distinction between "primary" and "secondary" qualities, respectively.

It is incumbent upon an atomist theory of knowledge to describe the "finer means of knowing" and to show how it yields knowledge. There are many questions to be answered. What are the standards governing rational theorizing? How can one metaphysical theory be shown to be better than another? Why should rational theorizing be thought to be any more reliable than simply using the senses? Do we really have a good enough reason to give up our common-sense beliefs about the objects around us? These are questions that are even today central to epistemology. (To some extent they were addressed by the atomist Epicurus, who lived about a century after Democritus. For more information, follow this link.)


Plato provided the first detailed theory of knowledge, though most of his philosophy is concerned with metaphysics. According to Plato, what is real is "forms" such as "justice itself" or "triangle itself." These forms exist separately from the objects of the senses, which in turn are described as pale copies of the forms. (For more on Plato's theory of forms, see my lecture notes for Philosophy 1). Plato seems to have arrived at the theory of forms as a solution to his main philosophical problem: to understand what a thing really is. Nothing in the unstable, transitory realm of the senses is suitable as the nature of the thing, so Plato was forced to have recourse to another realm of reality (sometimes referred to as "Plato's heaven"), populated with eternal, unchanging objects. It is justice itself, rather than any circumstances surrounding an act, which makes a just act just. As with Democritus, Plato's theory of reality generated the question as to how one could know the theory is true when the reality it posits cannot come before the senses?

The metaphysical theory of forms was developed through a through a long process of give-and-take ("dialectic"). Available alternatives were carefully examined and then rejected. For Plato, the theory of the forms provides the best account of the nature of things. He could, like the atomists, have claimed that there simply must be forms. In modern terminology, he could have treated them as "theoretical postulates" that must exist because things could not be what they are without them. But Plato went farther and claimed we know the forms themselves directly: they are known through "recollection" of an acquaintance we have all had with the forms before birth. This looks like an ad hoc appendage to the theory of forms: the theory of recollection is pretty implausible in itself, so was it a desperate attempt to describe knowledge of a reality that appears to be unknowable?

The theory of recollection was invoked in Plato's dialogue Meno to explain a fact about human knowledge, not to buttress a theory of reality. Plato tried to show that mathematical knowledge is possible a priori, that is, without recourse to sense experience. In the dialogue, Socrates talks to a slave-boy who had no mathematical training, and through a series of questions evokes from him the Pythagorean Theorem. The only way to explain how this is possible, Plato held, is by postulating that the boy had always had the information, though it had never come to consciousness. The idea that human beings are capable of knowledge a priori, or independent of sense experience, became very widespread in subsequent theories of knowledge.

Plato had another motivation in claiming that the forms are the objects of our knowledge. Just as with other conceptions such as justice, he sought an understanding of what knowledge is, and how it is to be distinguished from opinion. In so doing, he set the ground rules for the investigation of knowledge in Western thought. The notion that knowledge involves more than true opinion, the added element being an account of why the belief is true (which is implicit in the preceding discussion) originated with Plato. It is sometimes called the traditional analysis of knowledge. The essential difference between the two is that knowledge is stable and opinion is fleeting. What makes knowledge stable is that it is "tied down." Plato believed that knowledge cannot be stable in the requisite way if its objects are the transitory things which are detected by the senses. Given these considerations, we can see that Plato's theory of knowledge is an attempt to understand what knowledge is, and then to answer the question as to how it can be obtained. Theory of knowledge and metaphysics go hand-in-hand.


Aristotle, Plato's pupil, made several important contributions to epistemology. Like his teacher, he used a dialectical method to build a theory of reality. The resulting theory was rather unlike that of either the atomists or Plato, in that it was not skeptical about the properties of things detected by the senses. Indeed, the "categories" by which he classified the forms of real things included "quality," which includes the perceptual property color. Further, with few exceptions, Aristotle did not postulate the reality of objects inaccessible to the senses. Thus his theory of knowledge fits pretty well with the point of view of common sense.

It is possible, according ot Aristotle, to find order in a changing world. What is required is an adequate stock of experience. When presented with objects which are different but contain something in common, the mind is able to form general concepts which apply to all of those things. (Nowadays, this is known as "concept empiricism.") The process of abstracting concepts from concrete objects does not require any access to a non-sensible world. Rather, it is based on the mind's ability to grasp the forms inherent in the sensible objects themselves. The mind becomes "informed," as it were, upon repeated exposure to objects embodying the form.

Knowledge of what a thing is by grasping its form could be called basic knowledge. It is direct insight into the nature of a thing. Aristotle believed that without basic knowledge, there could be no knowledge at all. He was, in modern terminology, a "foundationalist." Aristotle's argument for foundationalism has been highly influential in epistemology, especially in the twentieth century. (Our text devotes two chapters to it.) If there were no basic knowledge, then whatever we know is based on some other knowledge we have. But that knowledge would have to be derivative as well, requiring that it be based again on some other knowledge. The basing relation either continues infinitely or else loops back on itself. And neither of these alternatives is acceptable. (We will discuss this argument in great detail later on in the course.)

The Stoics

A philosophical movement beginning in the generation after Aristotle, Stoicism, was concerned with an issue that had a profound influence on the subsequent course of epistemology. As has been mentioned already, much of the tendency in Greek philosophy was to be skeptical of the testimony of the senses. Aristotle's theory of knowledge was friendlier to the senses, in that he held that through repeated sensory experience we can grasp the forms of things. But what of a specific judgment made on the occasion of a sensory experience? Let us say that Socrates perceives someone who looks like his student Plato. Does he know that it is Plato? Given the fact that we are frequently deceived in the judgments we make based on what we perceive, we need to be able to lay down conditions which will distinguish cases in which Socrates knows that he sees Plato from those where he does not.

The Stoics supplied such conditions. They held that several requirements must be met in order for one to have knowledge. Let us apply them to our example. First, the perception Socrates has must have come from Plato himself, not from someone who merely resembles Plato. Second, it must be a perception which bears the characteristics of Plato. If Plato were to disguise himself beyound recognition, his presence before Socrates would produce the wrong kind of perception. Plato's presence and Socrates's having an accurate perception of him is not enough for his knowledge that it is Plato. A third condition is that Socrates must also be able to distinguish a peception of Plato from a perception of someone (or something) else greatly resembling Plato. This condition of knowledge is much-discussed in the contemporary literature of epistemology. While the condition is plausible, it holds a certain danger of skepticism. It might be that Socrates could not distinguish Plato from someone in Elea very closely resembling him. If so, it might be that Socrates does not know that Plato is before him, even though the man is Plato, looks like Plato, and would not be mistaken by Socrates for anyone else in Athens. (For more on the Stoic epistemology, follow this link.)

The Skeptics

In fact, this line of attack against the Stoics' condition of knowledge was made by some members of Plato's academy who have come to be known as "Academic skeptics." It is symptomatic of their more general attitude that the reasons in favor of one opinion can always be countered by reasons against. This attitude is quite popular in the present day. It is commonly believed that, at least with respect to many subjects, nothing can be definitely established, as every argument pro and contra is made from a certain point of view, and no point of view can be objectively established. (This view was made especially forcefully by Nietzsche at the end of the nineteenth century.)

As the skeptics realized, the attitude "no more this than that" is quite powerful, in that it takes away our reasons to act. If I have no better reason to think that food will nourish me than to believe that it will kill me, why eat? Some skeptics tried to handle this objection by drawing a distinction between theory and practice. Knowledge belongs only to the realm of theory, while action can be based on opinion. Even if we have no knowledge of what the "right" action is, we can simply observe the behavior of others and imitate it. Another approach, which has been attributed to Carneades, was to allow that although we lack knowledge generally, we can still have good reasons for belief. Carneades's theory of plausible belief was adapted by Roderick Chisholm in the 1960s in the context of a theory of knowledge.

Later skeptics (ostensibly following Socrates's contemporary Pyrrho, and hence called "Pyrrhonian,") produced what is one of the most compelling attacks against the possibility of knowledge: the "problem of the criterion." As was alluded to earlier, one way into skepticism is to claim that it is impossible to settle some disputes, specifically when each side brings its own battery of reasons which are not shared by the other. Now we would like to be able to adjudicate the dispute by calling up some further reason that will favor the reasons given by one side or the other. But the skeptics (stealing a page from Aristotle) posed this dilemma. If it is a reason held by just one side, then the question has been begged against the other side. There would be a circle of reasons. So perhaps we could bring in a reason that is neutral between the two. It may settle the dispute, but it may not. For either side could question the propriety of using that reason. If a further reason is given to back up the reason newly in dispute, the process can repeat itself. This problem (also known as the "diallelus" or "wheel") is discussed by a number of prominent epistemologist these days.


In the long period from the fall of the Roman empire to the Renaissance, little emphasis was placed on epistemology. Most of the discussion in philosophy was in metaphysics, which became particularly important because of philosophy's role as a buttress to Catholicism. But in the sixteenth century, the arguments of the ancient skeptics became known again, precipitating a kind of intellectual crisis, particularly because the Protestant Reformation had brought into vivid relief the problems of reconciling divergent points of view. (See Richard H. Popkin, The History of Skepticism: From Erasmus to Spinoza.)

Descartes, writing in the first half of the seventeenth century, tried to squelch the new skeptical tendencies. To do this, he employed a radical method, the "method of doubt." His idea was to find the most compelling skeptical objections to the possibility of knowledge and to overcome them. The most important objections were extensions of those made by the Academic skeptics against the Stoics. In order to know, one must be able to distinguish situations in which an object is present from one in which there is instead something very much like it.

Descartes stretched this objection to the limit. He noted that he could not infallibly tell whether he was awake or asleep, as he had often been deceived on the matter. Moreover, he might be fooled by an evil demon so powerful that even his beliefs about simple truths of mathematics could be false, the result of delusion. Descartes believed that he could show that these possibilities can be neutralized. Unfortunately, philosophers have come to recognize that it is very difficult to overcome these skeptical possibilities once they are taken seriously, as Lehrer notes in Chapter 1. One might argue that Descartes has set the bar too high: in his quest to avoid deception, he allows as knowledge ("scientia") only what admits of irrefutable proof. A noteable feature of Descartes's epistemology was his view that "scientia" is the product of reason, operating independently of the senses or "a priori". It is quite plausible to suppose that mathematical knowledge is formed in this way. Moreover, it is plausible to think that the existence of God could only be known through reason, as Descartes believed is the case. But Descartes went further and like Plato held that the fundamental characteristics of the physical world are known indendently of the senses. This position has come to be known as "rationalism."


In the second half of the seventeenth century, philosophers began to develop a number of different approaches to knowledge. Leibniz and Spinoza put forward broadly rationalist theories. Much more influential on current epistemology was John Locke, who sought to base our knowledge of the physical world on the senses, an approach now called "empiricism." A key thesis of Locke's empiricism is that the only items available to the mind as a basis for knowledge are "ideas" ultimately derived from the senses, and hence that knowledge is a relation of "agreement or disagreement of ideas."

The fourth and final book of Locke's An Essay Concerning Human Understanding, entitled "Of Knowledge and Probablity," is a prototype of epistemology today. Locke understood the task of epistemology to be fourfold:

Locke also discussed the crucial notion of probability. Unlike most recent epistemologists, he believed that when the strength of our evidence makes our belief only probable, we lack knowledge. Finally, Locke discussed some touchy issues of religious belief, declaring that we are never justified in believing what is "contrary to reason" but allowing for faith in what is "above reason."


George Berkeley found Locke's account of knowledge to be deficient. If our knowledge consists in the relation of agreement or disagreement among ideas, how can we claim knowledge of the physical world, which is supposed to lie beyond ideas, as their cause? Rather than abandon Locke's account of knowledge because of its skeptical tendencies, Berkeley embraced and extreme "innovation," which was to identify physical objects with collections of ideas. More recent incarnations of this strategy are known as "phenomenalism," and are discussed in Chapter 3.


David Hume, in the second half of the eighteenth century, finally drew the conclusion that almost nothing can be known. Hume advanced a number of powerful skeptical arguments. The most famous of these was to the effect that we can never have knowledge of general truths about matters of fact that go beyond what we have experienced. No matter how many cases we have observed of A being accompanied by B, we cannot justifiably infer that all As are Bs. Hume also advanced arguments against the use of probabilisitic reasoning, knowledge of the existence of external objects, and even of knowledge of our own identity over time.

This threw him into fits of despair, so much so that he wrote that at times he considered himself a monster unfit for human company. Hume's arguments led to a low point for epistemology in the modern era. Some philosophers simply threw up their hands and gave up the quest to validate our standards of evidence, opting instead to appeal to common sense. (Especially Hume's contemprary Thomas Reid, for whom Lehrer expresses admiration.) Others resorted to elaborate philosophical schemes designed at least in part to avoid skepticism. Still others tried to meet Hume's arguments head-on.

The Twentieth Century

There were some important developments in the late eighteenth and the nineteenth centuries, but we will pass over them. Much work in the field of epistemology in the first half of the twentieth century was centered around the refutation of skepticism from an epmiricist standpoint. Of particular note are Bertrand Russell and G. E. Moore. Russell proposed that we can know that the physical world exists because its existence is the best explanation for the occurrence of ideas in our minds. The topic of explanation as evidence will be discussed in Chapter 5. Moore attacked the plausibility of skepticism. He held that it was more reasonable for him to believe that he had two hands than to believe that he had no hands at all (as would be the case if physical objects do not exist). If any skeptical argument has as its conclusion that he does not know that he has two hands, then there must be something wrong with the argument. Recent epistemologists have struggled to understand how it is that the skeptical argument can be so compelling when its conclusion is so implausible. (For a representative paper, by Yale Professor Keith DeRose, click here.)

By the twentieth century, most empistemologists had abandoned the strictures of Descartes and Locke, allowing that we can know what is merely probable. So there were various attempts to adapt probability theory to epistemology. A very detailed adaptation can be found in Book II of C. I. Lewis's 1946 Carus Lectures, published as An Analysis of Knowledge and Valuation. This was to become a prototype of a modern theory of knowledge. A Harvard graduate student, who went on to teach at Brown University until his recent death, was Roderick Chisholm, of Brown University. Chisholm tried to combine some features of Aristotlelian foundationalism with some of the proposals of the more moderate ancient skeptics (as well as some other elements) into a positive account of knowledge. His first book, Perceiving: A Philosophical Study, was published in 1957.

In 1963, Edmund Gettier published a very short paper, "Is Justified True Belief Knowledge?" which produced a massive response. (Gettier's question will be considered in the notes for Chapter 1.) Many of the leading epistemologists of the present day cut their teeth on the "Gettier problem." One of them was Chisholm's student, Keith Lehrer, the author of our text. Aside from proposing some of the myriad solutions to the "Gettier problem," Lehrer rejected foundationalism and was even more inclined than Chisholm to borrow from the skeptics. The result is a theory of knowledge with many attractive features, which we will examine in the sequel.

[ Lecture Notes | Philosophy 102 Home Page | G. J. Mattey's Home Page ]