Bertrand Russell

Bertrand Russell was the best-known British philosopher of the twientieth century. He was born in an aristocratic family in 1872 and died at the age of 98 in 1970, living most of his life in England. His fame is due in part to his achievements as a professional philosopher, but he would not be nearly so well-known had he not composed numerous popular essays and been a highly-visible political activist. For his literary efforts, he received the Nobel Prize in 1950. He was unsuccessful in his attempts to gain elected office, but he was an icon of the anti-war movement. He campaigned passionately for the abolition of nuclear weapons and against the war in Vietnam. He convened a "Vietnam War Crimes Tribunal" in 1966. His moral and political views resulted in his not being allowed to hold a visiting professorship at City College of New York in the 1930s.

Russell's Contributions

Russell's early work was primarily in the area of foundations of mathematics. He was the first to see the significance of the work of the German philosopher Gottlob Frege, who had invented a powerful form of symbolic logic and had tried to prove that all truths of arithmetic could be derived from that logic (which included set theory). Russell reformulated the symbolic logic to make it more easily comprehensible, and together with Alfred North Whitehead, he used it as the basis for a logicist system in the monumental Principia Mathematica (three volumes, 1910-13). Of particular significance in that book was a solution to a paradox Russell had discovered in Frege's system, what has come to be known as "Russell's paradox." Russell had discovered that Frege's system was inconsistent because it allowed the existence of the set of all sets which do not belong to themselves (the "Russell set"). This set belongs to itself if and only if it does not belong to itself. Russell labored for years to find a solution, which is known as the theory of types.

In 1905, Russell published one of the most influential articles in twentieth-century philosophy, "On Denoting." This article applied symbolic logic to a perennial philosophical problem: the problem of non-existents. In 1905, France was a republic and had no king. This being the case, what does the expression "the present king of France" mean? Does it refer to nothing? To a non-existent something? Russell's solution was to say that the expression, which is a "definite description" (see below), has no meaning except in the context of a sentence. So the sentence, "The present king of France is bald," means that there is one and only one king of France, and that individual is bald. The sentence is meaningful and false. This treatment of the problem of non-existents is a paradigm of what has come to be known as "analytic philosophy," the dominant philosophy in English-speaking countries today. Russell is credited, along with G.E. Moore and Ludwig Wittgenstein, with having founded the analytic movement in philosophy. Russell wrote many more books and articles on issues in metaphysics and epistemology, but none of them had the influence of the 1905 article. The Problems of Philosophy was written as a text for students.

Perceptual Relativity

Russell begins by describing our ordinary beliefs about the world, e.g., the belief that he is sitting at a black rectangular table. Most people would not find any reason to doubt the truth of their belief, and they might be inclined to say that the belief is certain. But philosophers, from the very beginning in ancient Greece, have subverted our confidence in what we ordinarily believe. Russell contends that beliefs such as the one just mentioned are very likely wrong. The basis of the belief is sense-perception. We see and feel the table, and on that basis we believe both that it exists and that it is black and rectangular. Moreover, we believe that if others were to apply their sense-perception in a similar way, they would have the same beliefs.

Perhaps this is so, if we describe the table in a rough way, as having a rectangular shape. But suppose we tried to say what its shape is exactly. Russell contends that it would look and feel slightly different to different individuals. This could be due to a number of factors. People looking at the table from different angles or distances see it differently. People with abnormal vision will see it differently from those with normal vision. This variation in what we sense, based on our point of view, has come to be known as perceptual relativity. Perceptual relativity is one of the most important reasons to embrace skepticism. Here is what the ancient Pyrrhonian skeptic Sextus Empricus had to say about it:

Therefore, since everything apparent is viewed in some location and in some position, each of which produces a great deal of variation in the phantasiai [images of them], . . . we shall be forced also by this mode to have recourse to suspension of judgment. (Outlines of Pyrrhonism, Book I, Chapter 14, "The Ten Modes")

Appearance and Reality

The quotation from Sextus concerns what is "apparent," or what appears to the human senses. In non-philosophical circles, the way things appear to people is primarily the domain of the visual artist or designer. More practical people do not concern themselves with appearances. Their interest lies instead in reality. If my car will take me where I want to go, I might not be at all concerned with the fact that the paint is faded and the metal is rusted. The philosopher steps in at this point and asks what exactly it is that merely appears and what is the reality behind the appearances. Appearance is what is relative to human perception, so if a feature of an object does not exist except in relation to perception, it is a mere appearance. Ordinarily, we believe that colors are real, that they exist whether or not they appear to some perceiver or other. But perceptual relativity shows that there are colors only insofar as colors are perceived. The reason is that there is no single color that can be called "the" color of an object, due to the fact that the object looks differently to people in different positions. Following George Berkeley, Russell contends that perceptual relativity extends to shapes, hardness, etc., as well as colors. In fact, all features detectible by the senses are subject to relativity. This leads to a skeptical conclusion, as Sextus had pointed out. We must say that the senses, at least by themselves or "immediately," do not give us knowledge of what the features of objects really are.

Two Questions

At this point, one wonders how far this skepticism will carry us. If we do not know immediately through the senses what the features of objects are, can we know them in any other way? Or even worse, can we even know whether the objects exist at all? If you take away all the characteristics of a thing, what remains? Russell points out that these are some of the most difficult questions posed by philosophy. His task in what follows is to answer them.

Sense-data

In order to begin this task, Russell introduces a technical notion not found in ordinary language. When we sense something, the immediately sensed object, e.g., the color we see, is called a "sense-datum." Other philosophers have called these items "sensations," but Russell has a reason to invent a technical term. First, there is an ambiguity in the word "sensation," in that it can signify a "sense-datum" or a perceptual process. Secondly, sensations seem to be the product of physical sense-organs, such as the eye. But the very existence of our eyes, ears, and whole body is at issue here. So Russell takes sensation to be, not the product of the workings of the sense-organs, but the experience of being aware of sense-data such as colors, shapes, textures, etc. When we sense a color, there is a sensation of a sense-datum. Sense-data are private objects, subject ot perceptual relativity. The color black which is my sense-datum is not a feature of the table I suppose I am sitting at, or of any other object. Yet I say the table is black. What is the relation between the sense-datum and the table that could account for my belief?

Idealism

Our ordinary conception of a table is that of a physical object. Tables, chairs, the Earth, galaxies, are all physical objects. We can say that "matter" is the collection of all such objects. Using relativity arguments, among others, George Berkeley argued in the eighteenth century that matter does not exist at all. He called his doctrine "immaterialism," but Russell follows the normal practice and calls it "idealism." Russell does not think Berkeley's arguments were ultimately successful, but he recognizes that they at least undermine our certainty in the existence of matter.

As Russell interprets Berkeley, there are no physical objects. (Berkeley himself claimed that there are physical objects, only no matter, so he did not understand "matter" in the same way as did Russell. For Berkeley, matter is something unperceivable, while physical objects are collections of what Russell calls sense-data and hence are quite perceivable.) Nonetheless, sense-data do indicate the existence of something beyond them. This mental thing is a mind that causes them to exist, the mind of God. So Russell construes Berkeley as claiming that the real table is an idea in the mind of God. Thus the name "idealism."

Existential Doubt

Our doubts about the existence of matter, fueled by perceptual relativity, can be extended even further. How do I know that other people exist? It seems that it is on the basis of my perception of their bodies, the sounds that come from their mouths, and various other information garnered by the senses. But when the existence of physical objects is in doubt, the existence of other people is thereby thrown into doubt as well. It may be that I myself (along with my sense-data) am the only thing that exists. Such a view is called "solipsism."

David Hume extended this doubt to its limit (Treatise of Human Nature, Book I, Part IV, Section VI, "Of Personal Identity"). Supposing that I exist, which Descartes held cannot be denied when one thinks about it, what evidence do I have that I existed earlier? All that is present to me are my sense-data (as Russell would put it), including apparent memories. But how can I say that these memories are connected to a single being, I myself, which exists over a period of time?

So it seems that all that is certain is the fact that there is a perception of sense-data at a given time, though perhaps not by a unified self. This is a very thin starting-point for knowledge, but Russell nonetheless held that it is good enough. We can have knowledge about physical objects and other people on this basis, he will argue.

From Sense-Data to Matter

As we have seen, it is common-sensically believed that sense-perception gives us good evidence of the existence of physical objects. But how do sense-data make reference for anything beyond themselves? Moreover, why should we think they make reference to objects that are quite unlike them in a certain way? That is, sense-data are private objects that are relative to individual perceivers. But we suppose physical objects to be public objects, there for all to perceive. Their features are independent of any individual point of view. Again we pose the question, how can private sense-data give us good reason to believe in the existence of public objects?

Similarity

A natural way of bridging the gap between private sense-data and public objects is by appeal to similarity. For example, if we both look at the table in normal light from a similar distance, we will both have black sense-data as a result. But this appeal to similarity is illegitimate unless we already know that other people exist. Given the danger of solipsism, this is not yet known to us. Simply to assume that other people exist would be begging the question, because if other people exist, then public objects exist, and whether they exist is the point at issue. We would have to assume that public objects exist in order to show that they do.

It might be objected that it certainly seems as if other people exist. And this is so. But this seeming consists entirely in the fact that we have sense-data of a particular sort. These sense-data might be the products of dreams, hallucinations, or some other process internal to my own mind. The question remains, how can our private experiences give rise to knowledge of public objects? They must, if such objects are to be known at all.

Simplicity

Russell grants that solipsism is capable of explaining the content of my private experiences. There is no contradiction in my supposing that there is nothing public corresponding to my sense-data. Even if we allow that public objects exist, we also acknowledge that we have dreams, and as Descartes pointed out, such dreams are so elaborate as to resemble strikingly the physical objects we take to exist. Yet, although solipsism is a way of explaining the data of sense, it is not the best way.

A striking feature of our sense-data is their order and regularity. On the hypothesis that public objects exist, if I look away from the table, then look back, I have reason to expect a great deal of similarity between the resulting sense-data. On the solipsistic hypothesis, I must account for this order by appeal to myself. But how is it that I (if there even is a unified I) provide the means to fill in the gaps in my sense-data so predictably? As David Hume pointed out, it is natural to associate the regularity of our sense-data with the continued existence of public objects. Russell adds that this is the simplest explanation. Not only does it explain the regularity in the appearance of inanimate physical objects, but it also explains the regularity in the appearance of animate physical objects. So if I have certain sense-data, it is simpler to account for them through the existence of a cat exhibiting signs of hunger than to account for it purely privately.

Human Behavior

The greatest advantage of the hypothesis of the existence of public objects is its ability to explain sense-data that are similar to the ones I attribute to my own body. It is simpler to suppose that they are the product of public objects, bodies, which are like mine in that they are the bodies of another human being.

The hypothesis of public objects also allows me to account for my dreams, which is difficult to do given solipsism. Dreams are produced when my body is asleep and my ordinary outlets of sense are temporarily not functioning (or not functioning normally). None of this should be news to anyone. The postulation of public objects has always been the preferred explanation of our private experience, and it explains all aspects of it. As Russell puts it, "Every principle of simplicity urges us to adopt the natural view."

Belief in Physical Objects

Of course, it is odd to talk about the "postulation" of physical objects or the "hypothesis" of their existence. Our original belief in the existence of public physical objects is instinctive, not the product of rational demonstration. In fact, as Hume recognized, it is natural to think that the appearance of the object just is the object itself. (On the other hand, people do not ordinarily identify "sense-data" with physical objects, since it never occurs to them to think about sensations or "sense-data" in the first place.) So the natural belief that physical objects exist should not be rejected despite the doubts induced by perceptual relativity. Their existence is the most simple explanatory hypothesis to account for the characteristics we find our sense-data to have.

Given that we acknowledge the existence of physical objects, the next question to be answered is about what kind of things they are. Philosophy's task is to give an account of these objects in a systematic way, conforming to our deepest instinctive beliefs about them. (But not the instinctive belief that they are just as they appear to be!) This will be an explanatory system, to be sure, and as such it is uncertain. But the uncertainty is diminished by the fact that its parts will be in harmony with one another. It is generally allowed by philosopher that coherence in a system is an indication of its correctness.

The Nature of Physical Objects

Science has come to explain the phenomena of physical nature through motion, a trend that began with the seventeenth-century rejection of Aristotelian teleology. It has been assumed by most scientists and philosophers that the motions in the physical world and the motions perceived by the sense are the same, but we have seen that this view is undermined by perceptual relativity. Berkeley emphasized the point that what is swift to one perceiver is slow to another, for example. He also noted that the space we feel is distinct from the space we touch. His conclusion was that space is only a relation of ideas (Russell's "sense-data"), but Russell wants to allow that space exists. So his own conclusion is that the space of perception is different from that perceived. Whether seen or felt, the shapes found in different people's sense-data will be different. But publically-existing space does not change at all.

Correspondence

So far, we have explained very little. The hypothesis that explains the gaps in our sense-data is that there are persistent public objects to which the sense-data correspond. We need to say more about what this correspondence is supposed to be. The phenomenon of regularity of sense-data is best explained, claims Russell, through the medium of the human body. Physical object interact causally with our bodies, and it is this interaction that gives rise to the sense-data. So changes in sense-data ought to reflect that changes in the positions of our bodies relative to the objects that indirectly cause the sense-data. This hypothesis is strengthened in various ways. First, there is a cohrence in our sense-data. The testimony of one sense-modality fits with that of another. My visual sense-data conform to my tactile sense-data, and the conformity is best explained by both being caused by the same physical object. Moreover, my sense-data conform to the sense-data of other people (whom we now postulate exist as well as us). Putting it all together, we can say that there is a physical space that corresponds to the space of our private sense-data.

Knowledge of Physical Space

How far does our knowledge of physical space extend? Only, Russell maintains, as far as is needed to explain the coherence of our sense-data. An example Russell gives is that of distance. When there is a solar eclipse, we can say that the moon is between the sun and the earth. The moon is nearer to the earth than is the sun, as it blocks the light of the sun. But what is the real, physical distance between the two? Perceptual relativity dictates that this is not something that is to be found in the sense-data. So though there is a correspondence between the space of sense-data and that of public physical objects, we can only take it so far. To some extent, the true nature of physical space is unknown to us, and so Russell embraces a limited skepticism.

Knowledge of Time

The same considerations hold for the relation between duration as it is felt privately and public duration. As noted above, Berkeley held that the same motion might appear swift (taking less time) to one person and slow (taking more time) to another. But as with space, there still must be correspondence enough to explain the coherence of sense-data (at least "so far as we can see"). As Kant noted, the order of succession of what Russell calls sense-data caused by public objects must correspond to the order of succession in the objects themselves, if we are to make any sense at all of the notion of a public object. But Russell notes that this correspondence is pretty rough. Thunder and lightning occur simultaneously in the public world, but the sense-datum of a lightning-bolt precedes by seconds that of a thunder clap. And the light we see as on the sun actually had left it eight minutes ago.

Knowledge of Physical Objects

Having arrived at a quasi-skeptical conclusion regarding our knowledge of space and time, Russell turns to other characteristics of physical objects. As in the previous cases, there must be some correspondence between the variation in sense-data and the variation in the public world. But again, all we can know of what goes on in physical objects concerns their relations to one another. A body is called hard when we have sense-data indicating that another body does not penetrate it. But we do not know hardness itself. Russell concludes that we do not know the non-relative ("intrinsic") properties of bodies through the senses. More generally, he thinks we are not justified in supposing that the properties of objects resemble what we find in our sense-data, except in the relative ways needed to explain the coherence of the sense-data themselves.

Idealism

Russell has maintained that the existence of physical objects is the simplest hypothesis to explain the coherence of sense-data. It does seem simpler than the solipsistic hypothesis. But is it the best hypothesis of all? Suppose that we grant that there must be public objects to explain the sense-data. Why must these objects be physical? Why might they not instead by mental? You might be inclinde to answer by saying that it would be strange to think of tables and chairs as mental entities. But Russell rightly notes that it is pretty strange or even absurd to think that we know of public objects only what is needed to explain the coherence of our sense-data. Common-sense in this case must take a back seat to philosophy.

Berkeley's Argument for Idealism

Russell now advances an argument for idealism that he attributes to Berkeley. We begin with the fact that sense-data are mental entities: their existence is to be perceived by our minds. Moreover, sense-data are what we know best. We have immediate knowledge of their existence and character. Whatever else we know, we would have to know through sense-data. So all we know immediately about tables, chairs, and the whole world of public objects is through them. There is, moreover, no reason to think that we know anything about sense-data beyond what is known immediately. So sense-data are really all there is to a thing. Since the being of sense-data is to be perceived, tbe being of public objects is to be perceived. Of course, it will be objected that sense-data are private, and that we must be able to account for their coherence. Berkeley does this by invoking God as an explanatory hypothesis. God has notions of those objects and causes sense-data in human being corresponding to those notions. Berkeley noted (though Russell did not) that this is a simpler hypothesis than one that postulates material objects, since God would be thinking of those objects anyway, so there is a gain in economy in just eliminating the middle-man, so to speak.

Fallacies

Berkeley's idealism has won practically no adherents since he presented his case for it. This might be explained by the strangeness of the hypothesis. But Russell cannot rely on this common-sense rejection of idealism. Instead, he needs to undermine the argument in its favor. He claims that Berkeley's reasoning is fallacious.

Russell agrees with Berkeley on many points: sense-data are mental, they are the only things immediately perceived, and public objects, if known at all, would have to be known through them. The disagreement has to be over the claim that our knowledge of sense-data exhausts our knowledge of objects. As we have seen, Russell, thinks that we can know of the existence of public objects corresponding to our sense-data. For Russell's Berkeley, there is something corresponding to our sense-data and explaining their coherence, but this is something mental, i.e., something in the mind of God. But why claim that this something has to be mental?

This is where Russell thinks he has detected fallacious reasoning. For Berkeley, anything besides the mental is unthinkable. If I am to hypothesize something public corresponding to my sense-data, it must be "in my mind" just because it is thought of. If I think of a tree, that tree is in my mind as the object of my thought. But this does not mean, Russell goes on, that the tree itself is something mental just because it is thought of. An object may be "before the mind" without being mental, just as my thought of my wife is before my mind, while she is not a purely mental object. Similarly, we can think of a public cause of the coherence of our sense-data without thereby being committed to that cause being a mental object.

Acquaintance

An argument for idealism related to that given by Russell's Berkeley centers around the notion of acquaintance. Recall the claim that we have no knowledge of anything outside our sense-data (or their mental cause). The reason might be that we are not directly acquainted with any such thing. But if we are not so acquainted with an object, then it is of no relevance to us, and we might as well say that it does not exist.

Russell grants that we do not know matter through acquaintance. We are acquainted only with our private sense-data. But matter is at the same time relevant to us. We had better take it seriously, because it is what orders our sense-data, including our feelings of pleasure and pain. So can we know things with which we are not acquainted? The answer obviously (for Russell) is yes. I am not acquainted with the first man who walked on the moon, but I know him under that description. If we distinguish between knowledge by acquaintance and knowledge by description, it is at least possible for us to have knowledge of matter.

Knowledge of Things

Russell does not deny that knowledge by acquaintance is the simplest and first knowledge we have of things. In particular, we know other things through the sense-data with which we are acquainted. To know a thing by description requires an additional element, knowledge of general principles. Some philosophers have held that it is a general principle that an object cannot be red all over and green all over at the same time. I cannot know this truth by being acquainted with the color red or with the color green. In general, we need to be careful to distinguish knowledge of things, which can be by acquaintance and description, and knowledge of truths (such as general principles).

Knowledge by Description

An example of knowledge by description might be this: I know Neil Armstrong as the first man to walk on the moon. A more theoretical example is that I know that this table is "the physical object which causes such-and-such sense-data." In both cases, we attach to some object the description "the so-and-so." To know that the description applies to the table, I have to know some general truths about causality. But I also must know, by acquaintance, the sense-data which are supposed to be caused by the table itself.

Objects of Acquaintance

Sense-data are the starting-point of our knowledge, but we must be acquainted with more things in order to extend our knowledge through description. So we are acquainted with our memories. More exotically, we have an acquaintance with our being acquainted, an awareness of awareness which we call self-consciousness. Acquantance with my seeing the sun is acquaintence with the fact "self-acquainted-with-sense-datum." So I have knowledge by acquaintance of my acquaintance with the sense-datum.

Definite Descriptions

Other objects of acquaintance, which are required for knowledge of general truths, are universals such as whiteness, diversity and brotherhood. We must operate with knowledge of universals in order to be able to use language. Once we have the universals, we can combine them in expressions of the form "the so-and-so," to form definite descriptions. It is through definite descriptions that we know objects by description. So I know Neil Armstrong under the description "the first man to walk on the moon." A definite description must satisfy two conditions in order to yield knowledge. There must actually be an individual fitting the description, and there may be no more than one. So I cannot know anything about "the first grizzly bear to walk on the moon" (no bears have walked on the moon) or about "the man who walked on the moon" (many men have).

Acquaintance and Description

Sometimes definite descriptions designate objects with which we are acquainted. A common-sense example is one in which I am acquainted with my car, and I know it under the description, "the white car in my garage." I might describe objects with which I am not acquainted but with which I easily could become acquainted, as "the man just outside the door." But to know such an object, one must at least be acquainted with components of the description. I need to be acquainted with the door, or at least acquainted with something that gives me knowledge by description of the door. Clearly, the description takes us beyond that with which we are acquainted. This is what allows us to have knowledge of the cause of sense-data. We are only acquainted, strictly speaking, with the private objects of our experience, but we can describe their causes and know the causes under the description.

[ Slide Presentation | Philosophy 1 Home Page]