Substance and Attribute
SUBSTANCE AND ATTRIBUTE
The concepts of "substance and attribute" are the focus of a group of philosophical problems that have their origins in Greek philosophy and in particular the philosophy of Aristotle. The concepts are, of course, familiar to prephilosophical common sense. Yet although we are acquainted with the distinction between things and their properties and are able to identify the same things among the changing appearances they manifest in time, these commonsense notions give rise to a group of philosophical problems when we come to scrutinize them. Thus we may wonder what it is that remains the same when, for example, we say that the car has new tires and lights and does not run as smoothly as it used to, but is still the same car; or when we say that although we could hardly recognize him, this man is the same one we went to school with thirty years ago.
It is interesting to note that the principal term for substance in the writings of Aristotle is ousia, a word that in earlier Greek writers means "property" in the legal sense of the word, that which is owned. (This sense is familiar in English in the old-fashioned expression "a man of substance.") The word ousia also occurs in philosophical writings before Aristotle as a synonym for the Greek word physis, a term that can mean either the origin of a thing, its natural constitution or structure, the stuff of which things are made, or a natural kind or species. The Latin word substantia, from which the English term is derived, is a literal translation of the Greek word hypostasis ("standing under"). This term acquired its philosophical connotations in later Greek and occurs principally in controversies among early Christian theologians about the real nature of Christ. A third philosophical term, hypokeimenon ("that which underlies something"), is used by both Plato and Aristotle to refer to that which presupposes something else.
There is, however, little of philosophical importance to be learned from the etymology of the terms in which problems are formulated and discussed. We shall first consider the questions to which the concepts of substance and attribute give rise in some of the philosophers for whom they have been important. We may then ask which of these questions remain as live philosophical issues at the present time and what answers can be given to these surviving questions.
Aristotle's account of substance has been the most influential in the history of philosophy. His account is, however, obscure and probably inconsistent. The difficulties of elucidating and reconciling the various parts of his doctrine have been part of the cause of its influence—it has offered a continuing challenge to commentators and critics from Aristotle's time to the present. "Substance in the truest and primary and most definite sense of the word is that which is neither predicable of a subject nor present in a subject; for instance, the individual man or horse" (Categories 2a11). The explanation is obscure, but the examples cited leave no doubt of what Aristotle means here: Substance in the most basic sense of the word is the concrete individual thing. However, he goes on at once to mention a second sense of the word: "Those things are called substances within which, as species, the primary substances are included; also those which, as genera, include the species. For instance the individual man is included in the species 'man' and the genus to which the species belongs is 'animal'; these, therefore,—the species 'man' and the genus 'animal'—are termed secondary substances." These secondary substances are predicable of a subject. "For instance, 'man' is predicated of the individual man" (Categories 2a21–22), as when we say "Socrates is a man." Aristotle seems to have the idea here that essences or natures are substances, and the more qualities they comprise, the more substantial they really are; he explains, "Of secondary substances, the species is more truly substance than the genus, being more nearly related to primary substance" (Categories 2b7). For example, the species Canis domesticus shares more qualities in common with the individual dog Tray than does the genus Canis.
This notion of essences as substances is treated at length by Aristotle in the Metaphysics and seems to be his preferred sense of the term. The intimation that the more qualities something has, the more substantial it is, has the advantage of suggesting that being a substance is a matter of degree and not an all-or-nothing matter. This hint, which Aristotle does not develop, contains an important idea, as will be seen later. But the doctrine of secondary substances has little else to recommend it and involves a serious logical confusion between the relations of class membership and class inclusion, as well as the notorious difficulties of the doctrine of essences.
Aristotle's main purpose in the Categories is to contrast the independent way of existing proper to substances with the parasitic mode of being of qualities and relations. Substances can exist on their own; qualities and relations, only as the qualities of or relations between substances. The key to this distinction is given by the phrase "present in a subject." (The Greek word for "subject" here is hypokeimenon, literally "underlay.") Substances are never "present in a subject." This does not mean, as Aristotle explains, that a substance is never "present in" something else as a part of a whole. On the contrary, he cites heads and hands (Categories 8b15) as substances although they are parts of bodies. Rather, x is present in y when it is "incapable of existence apart from" y. This notion introduces a third sense of substance as that which is capable of independent existence. This sense is of considerable importance in later philosophy, but Aristotle does not develop it. He uses it chiefly to emphasize the distinction between substances on the one hand and their qualities and relations on the other. A quality—"red," "sweet," or "virtuous"—cannot exist apart from an x that has the quality. Relations such as "larger than" or "to the left of" cannot occur in the absence of the x and y that they relate.
It is true, of course, as Aristotle's critics have pointed out, that it is no more possible for a substance to exist without qualities than for qualities to exist without a substance. However, it is possible to point to prima facie examples of qualities existing without substances—the blue of the sky, for instance, or a red afterimage floating in my visual field. Surely the sky is not a substance, nor is my visual field. However, one cannot point to any instances of substances existing without qualities. Even if it makes sense to suppose that such a thing could occur, it is clearly incapable of being identified. Aristotle does not consider these problems. What he seems to have meant, although he does not express himself clearly, is that what is capable of independent existence is the concrete individual thing, a substance with its qualities and in its network of relations to other substances. But even here there is an obvious difficulty. Once we introduce the notion of relations involving other substances, we put a restriction on independent existence.
A fourth criterion of substance is that "while remaining numerically one and the same, it is capable of admitting contrary qualities" (Categories 4a10). This Aristotle calls "the most distinctive mark of substance." This notion is developed, more by later philosophers than by Aristotle himself, into the conception of a center of change and so of a substratum that underlies and supports its qualities. Finally, Aristotle emphasizes the notion of substance as a logical subject, "that which is not asserted of a subject, but of which everything else is asserted" (Metaphysics 1029a8), and he links this sense of the term with the concept of substratum. This logical criterion has been criticized as making the notion of substance dependent on the structure of Greek (and some other Indo-European languages), in which subject-predicate sentences are a standard mode of expression, and upon a restricted and now outmoded view of logic in which all statements canonically expressed are in a form in which a predicate is affirmed of a subject. It is not the case that sentences in all languages fall into a subject-predicate form or that this form of expression is adequate for a developed logic.
The various notions of substance as (1) the concrete individual, (2) a core of essential properties, (3) what is capable of independent existence, (4) a center of change, (5) a substratum, and (6) a logical subject are never thoroughly worked out and reconciled in Aristotle. He appears to emphasize now one and now another mark of substance as of paramount importance. The quotations cited above have been chiefly from the Categories ; the topic is taken up and discussed at length in the Metaphysics. The discussion is tentative and not finally conclusive, but Aristotle seems to favor alternative (2), substance as essence, as his preferred sense. But the whole treatment is important not for the answers that he gives but for the questions that he raises. Discussions of substance in later philosophers have tended, with few exceptions, to take over one or more of the six senses proposed by Aristotle as the clue to the problem.
atomists and medievals
Of the philosophical theories of antiquity, one other is of some consequence. Ancient atomism, founded by Leucippus and Democritus, developed by Epicurus, and expressed in its most attractive form in the De Rerum Natura of the Roman poet Lucretius, suggests that the truly real and substantial elements of nature are the atoms out of which everything is composed. It is these that are fundamental, unchangeable, and, in the last resort, capable of independent existence. The problem of substance and attribute was not much discussed by the ancient atomists, but their theories provide material for an answer to the question raised by Aristotle.
During the Middle Ages, discussion of this problem was very naturally centered upon the theological repercussions of rival theories. In particular, the doctrines of the Incarnation of Christ and of transubstantiation depended for their rational justification upon a plausible theory of substance. But these theological outworks produced no new basic insights that can be regarded as an improvement on the work of Aristotle. Indeed, they are just variations upon Aristotelian themes.
The revival of philosophy in the seventeenth century in a form that was relatively independent of the religious framework of medieval philosophy produced several systems for which the notion of substance is fundamental. In the work of René Descartes the concepts of substance and attribute become associated naturally with those of the conscious self and its states, and the problem of substance becomes associated with the problem of personal identity. Descartes had been thoroughly trained in the form of Aristotelian scholasticism current in his day, and his notions of substance are in part derived from this and in part inconsistent with it. He gives a formal definition of substance as follows: "Everything in which there resides immediately, as in a subject, or by means of which there exists anything that we perceive, i.e. any property, quality, or attribute of which we have a real idea is called a Substance ; neither do we have any other idea of substance itself, precisely taken, than that it is a thing in which this something that we perceive or which is present objectively in some of our ideas, exists formally or eminently. For by means of our natural light we know that a real attribute cannot be an attribute of nothing" (Philosophical Works, translated by Haldane and Ross, 2nd ed., Cambridge, U.K., 1931, Vol. II, p. 53). In other words, what we are directly aware of are attributes of things and not the things themselves. But it is a logically self-evident principle (known by "the natural light" of reason) that an attribute must be an attribute of something, and the something is a substance—known by this inference and not directly. So far Descartes does not depart from scholastic doctrine, but he goes on to affirm that substances have essential attributes. For example, thought is the essential attribute of mind, and extension is the essential attribute of matter. But he does not explain what a substance is apart from its essential property. What is the mind apart from thinking or matter apart from extension? Unless this question is answered, how can Descartes answer the later empiricist criticism that the concept of substance is meaningless because empty of content?
In another context (ibid., p. 101) he gives an alternative definition of substance. "Really the notion of substance is just this—that which can exist by itself, without the aid of any other substance." This second definition is a bad one, being circular in expression; but clearly Descartes has in mind both here and in the quotation above simply the Aristotelian criteria (3) and (5). On the basis of these definitions, Descartes postulates three types of substance: material bodies, minds, and God. But the first two, being in a certain sense dependent on God for their existence, clearly have a lower grade of substantiality. Descartes's conception of substance and attribute is made impossible to understand by the vagueness of the notion "attribute" by which he seeks to clarify the idea of substance. If "attribute" means "property or relation," it simply is not true that all attributes are attributes of substances. For example, a color may have properties that are not properties of the colored thing. It is true of the color red that it is produced by light of wavelength about 7000 angstrom units, but this is not true of red objects. In any case, it seems that Descartes has simply defined substance and attribute relative to each other so that his explanation is circular and thus uninformative: Attributes are what qualify substances and substances are what have attributes.
Descartes's second account of substance as that which is capable of independent existence was taken up and developed by Benedict de Spinoza in his Ethics. Spinoza was a student of Descartes and may be regarded as one who developed some of Descartes's ideas to consistent but surprising conclusions. Reflecting on Descartes's second account of substance, Spinoza showed that if by substance we mean, according to his definition, "that which is in itself and is conceived through itself," it is easy to show that there can be only one such being, the whole universe. Thus Spinoza equated substance with God and nature, the three terms being synonymous for him. This "hideous hypothesis," in David Hume's ironical phrase, has won for Spinoza the inconsistent titles of atheist and pantheist. What he did, in fact, was to demonstrate the alarming consequences for religious orthodoxy of Descartes's second definition and to indicate obliquely that substantiality in this sense is a matter of degree. Nothing in the universe is completely independent of its environment, although some things are more independent than others. A human being has a certain degree of independence of his environment but can exist only within a certain range of temperature, pressure, and humidity, and with access to air, food, and water. Other things may be more or less independent of their surroundings, and the extent of their freedom in each case is an empirical question. Spinoza did not draw this conclusion, but it is implicit in his development of Descartes.
Another rationalist philosopher, Gottfried Wilhelm Leibniz, makes the concept of substance fundamental to his philosophical system. He uses two of the Aristotelian criteria of substance, substance as a center of change and substance as a logical subject, but adds the concept of simplicity. The basic elements of Leibniz's metaphysical system were what he called monads. In his Monadology he defines monad as "nothing but a simple substance.… By 'simple' is meant 'without parts.'" That there are such simple substances follows, for Leibniz, from the admitted fact that there are compound things, which can be nothing but collections of simple things. Leibniz seems here to have been influenced by the arguments of the ancient materialists for the existence of atoms. His monads, however, were supposed to be immaterial substances, centers of change and thus subjects of predicates. Unfortunately, by describing his substances in this way, he deprives the term of meaning just as Descartes had done. He does indeed affirm that his monads are centers of activity, but this activity is manifested only in their tendency to move from one state to another. But if the essence of something is to be the x that undergoes changes and of which predicates can be affirmed, it can have no positive character of its own. In Bertrand Russell's words, "substance remains, apart from its predicates, wholly destitute of meaning" (The Philosophy of Leibniz, p. 50).
locke, berkeley, hume
Leibniz had criticized the British empiricist philosopher John Locke for professing to find substance an empty concept. The weakness of Locke's criticisms of the concept was that he concentrated his attack on the notion of a substratum of qualities. This is not the most important of the Aristotelian senses of the term. But if "substratum" can be shown to be an empty notion, it is easy to raise skeptical doubts about some of the associated senses, particularly those of substance as a center of change, as the concrete individual, and as a logical subject. Locke points out that we find in experience groups of qualities that occur together in time and place. We therefore presume these qualities to belong to one thing and come to use one word, "gold," "apple," or "water" (whatever it may be) to refer to the collection of properties "which indeed is a complication of many ideas together." Further, "not imagining how these simple ideas can subsist by themselves, we accustom ourselves to suppose some substratum wherein they do subsist, and from which they do result, which we therefore call substance " (Essay concerning Human Understanding, Book II, Ch. 23).
Substance, then, is not a positive concept but merely an "obscure and relative" notion of "the supposed but unknown support of those qualities we find existing, which we imagine cannot exist sine re substante without something to support them." Since Locke has already tried to show that all our meaningful concepts originate in experience, substance is an awkward counterexample to his theory of knowledge. Indeed, he would probably have rejected it altogether but for certain associated moral and theological doctrines that his cautious and conformist temperament made him forbear to reject outright. Moreover, he seems to have been unable to reject Descartes's principle that attributes must inhere in a substance, although he does not submit this supposed logical truth to any rigorous examination.
However, Locke's empiricist successors, George Berkeley and Hume, were fully aware of the importance of Locke's criticism and his reduction of the notion to "an uncertain supposition of we know not what." Berkeley's attack on the concept of material substance owes much to Locke, and Hume was content to write off the whole idea as an "unintelligible chimaera." Moreover, Hume extended the skepticism of Locke and Berkeley in respect of material substance to question, on analogous grounds, the existence of spiritual substances or selves. It is clear that a mind whose function is merely to be the bearer of states of consciousness is as vacuous a notion as Locke's material "we know not what."
Immanuel Kant's Critique of Pure Reason (1781) transformed the notion of substance, as it did so many other philosophical concepts. In Kant's view, "substance" does not refer to a feature of the objective world independent of human thinking. On the contrary, the unity and permanence of substances are features contributed by the human understanding to the world of phenomena. This represents a very radical revision of the concept of substance. Substance shrinks from being a fundamental feature of the objective world to an aspect under which men cannot help classifying their experience—and they cannot help themselves not because of the nature of external reality but because of the structure of their own cognitive apparatus.
Since Kant's day the permanent and valuable features of philosophy have been those that have grown out of the immense development of the formal and natural sciences from the end of the eighteenth century to the present, a development that has shown the falsity of the scientific assumptions on which the Kantian revolution was built. For example, Kant believed that Newtonian physics, Euclidean geometry, and Aristotelian logic were finally and beyond all question true of the world, and some features of his system depend on these assumptions. This development has presented the problem of substance as a problem soluble, if at all, in the light of empirical evidence drawn from the relevant sciences. It has, moreover, made clear that there is no one problem of substance but a number of subproblems that can be treated independently.
These problems can still be stated in something like their original Aristotelian form, but we may find ourselves looking in different areas of knowledge for their answers. There is no one unitary science, such as metaphysics or ontology, that can be looked to for a solution. For example, the notion of substance as a logical subject of predicates (as when we say of a piece of gold, "It is heavy," "It is yellow," "It is malleable," "It melts at 1063° C," and so on) is now seen to be a problem of interest to formal logic and to linguistics. It is a technical question of logic whether all sentences about individual things can be (or must be) expressed in subject-predicate form. And it is a technical question of linguistics whether all languages use such a form to express these notions, or indeed have a subject-predicate syntax at all. (The answer in both cases seems to be "No.")
The question "What, if anything, is capable of independent existence?" can be seen, insofar as it relates to material things, to be a question to which physics, chemistry, and biology give us the answers. (If the question is asked about the existence of nonmaterial things such as numbers or propositions, we have first to make clear what is meant by "existence" in such contexts.) We see that independent is not a term with a clear meaning but, rather, is an elliptical expression. "X is capable of independent existence" means "X is capable of existing without regard to features y 1, y 2, · · ·, yn of its environment." Since these conditions are so numerous, it is easier to express the concept negatively: "X is not independent" means "X is incapable of existing apart from conditions z 1, z 2, · · ·, zn " or "z 1, · · ·, zn are necessary conditions for the existence of X. " On this interpretation, a substance in the sense of something that is capable of completely independent existence is something for whose existence there are no necessary conditions. The specific values of the variable z will vary with the value of X. For example, if X is a piece of ice or a lump of metal, one of the z 's will be temperature; if X is a green plant, the z 's will include light and oxygen; and so on. It may well be that nothing in the universe is independent of all conditions, but whether this is so is an empirical question.
Aristotle's favorite, but least satisfactory, account of substance was that of substance as essence, an essence being a set of qualities that conjointly embody the nature of the thing they qualify, are grasped by intellectual intuition, and are expressed in the definition of the thing. But developments in the sciences (especially in biology) and in the philosophy of science over the past century have shown that this notion is illusory. Definitions, in the contemporary view, are either descriptions of current linguistic usage or recommendations for linguistic conventions. They cannot seek to explicate the essential nature of the definiendum because naturally occurring objects have no such invariable natures. Definitions in formal sciences like mathematics and logic do delineate the invariant properties of the definienda precisely because they are proposals for conventions.
There remains for consideration substance in the senses of (a ) a center of change, (b ) a substratum of qualities, and (c ) the concrete individual thing. Senses (a ) and (b ) are closely akin and are both vulnerable to the empiricist line of criticism made famous by Locke. We may regard a particular thing as qualified by different properties at different times (for example, when an insect changes from egg to caterpillar to pupa to moth), or as qualified by a group of qualities at the same time (for example, when we say that a lump of sugar is white and sweet and soluble). Both of these ways of looking at substance lead to the unanswerable question "What is it that is the bearer of the qualities in each case?" But the answer to this cannot even be as satisfactory as Locke's "something we know not what," for by thus separating the subject (or hypothetical bearer of the qualities) from its predicates, we effectively prevent ourselves from saying anything about it. For to say anything about it is merely to assign to it one more predicate. This way of explaining substance makes it an empty concept.
Yet the obvious alternative to this blind alley seems no more promising. Suppose that when we say "Some apples are red" we do not mean what contemporary logic teaches us to mean: There is an x that has both the property of being an apple and the property of being red. Suppose that instead we mean: That set of particular properties which we call "apple" includes the further property of being red. Then the relation "being predicated of" turns out to be nothing more than the familiar relation of being a member of a group. This conclusion looks innocuous until we realize that this interpretation would make all subject-predicate affirmations either necessarily true or logically false. For the proposition "The set of properties Q 1, Q 2, · · ·, Qn contains the property Qn " is a logically true statement. And if we amend it to make it informative thus: "The set of properties Q 1, · · ·, Qn contains the property Qn +1" we do not have an informative proposition but, rather, a logically false one.
The way out of this dilemma is not to ask such misleadingly general questions as "What is the locus of change?" or "What is the bearer of properties?" We can ask for the detailed history of a particular thing, an insect, a plant, a man or what not, and the answer will be given to us by the relevant sciences—embryology, anatomy, physiology. We can ask for the detailed structure of a particular thing, a piece of gold, a moth, a man, or what not; again the relevant science—physics, chemistry, anatomy—will give us the answer if the answer is known. But we cannot ask for the history or structure of things in general, for there is no science of things in general.
A similar criticism awaits the last of the Aristotelian answers to the question about substance: A substance is a concrete individual thing. We cannot sensibly ask what makes things-in-general concrete individuals. The notion of a concrete individual thing is clear in its standard cases, like men, tables, mice, or stones. But it is unclear in its nonstandard applications. Is a cloud a concrete individual or is it just the particles that make it up that can be so called? Is a rainbow? Or a dream table? Can electrons be called individual things when it is impossible in principle to identify them and trace their continuous histories? Examples such as these show the futility of trying to find a general formula that will clarify the notion of a concrete individual thing. We can, of course, ask the psychologists what perceptual characteristics of things lead us to class them as individuals. That a set of jointly occurring properties stands out in our perceptual field, that it moves as one, that it persists through time—all these and other characteristics will lead us to regard a thing as a thing. But there is no decisive test which will enable us to decide, if we are doubtful, whether a certain x is really a concrete individual or not. In borderline cases this must be a matter for decision, not diagnosis.
Aristotle. Metaphysics and Categories (Oxford translations). Oxford, 1908 and 1928.
Blanshard, Brand. The Nature of Thought. London: Allen and Unwin, 1939. Ch. 3.
Broad, C. D. An Examination of McTaggart's Philosophy. Cambridge, U.K.: Cambridge University Press, 1933. Vol. I, Ch. 7.
Leibniz, G. W. Monadology, edited by Robert Latta. Oxford, 1898.
Locke, John. An Essay concerning Human Understanding, edited by J. Yolton. London, 1962. Book II, Ch. 23.
Russell, Bertrand. The Analysis of Matter. London: Kegan Paul 1927. Ch. 23.
Russell, Bertrand. Human Knowledge. London: Allen and Unwin, 1948.
Russell, Bertrand. An Inquiry into Meaning and Truth. London: Allen and Unwin, 1940.
Russell, Bertrand. The Philosophy of Leibniz. Cambridge, U.K., 1900.
Shoemaker, Sydney. Self-Knowledge and Self-Identity. Ithaca, NY: Cornell University Press, 1963.
Strawson, P. F. Individuals. London: Methuen, 1959.
Anscombe, G. E. M. "Aristotle. " In Three Philosophers. Oxford, 1963.
Anscombe, G. E. M., and S. Körner. "Substance." PAS supp. 38 (1964).
Basson, A. H. "The Problem of Substance." PAS (1948–1949).
Bennett, Jonathan. "Substance, Reality, and Primary Qualities." American Philosophical Quarterly (1965).
Cousin, D. R. "Aristotle's Doctrine of Substance." Mind (1933).
Kneale, William. "The Notion of a Substance." PAS (1939–1940).
Mei, Tsu-Lin. "Chinese Grammar and the Linguistic Movement in Philosophy." Review of Metaphysics 14 (1961). An original and penetrating critique of the notions of subject and predicate.
Sellars, Wilfrid. "Substance and Form in Aristotle." Journal of Philosophy (1957).
Stebbing, L. Susan. "On Substance." PAS (1929–1930).
D. J. O'Connor (1967)