The word universal, used as a noun, has belonged to the vocabulary of English-writing philosophers since the sixteenth century, but the concept of universals, and the problems raised by it, has a far longer history.
Indeed, Plato may be taken to be the father of this perennial topic of philosophy, for it is in his dialogues that we find the first arguments for universals and the first discussion of the difficulties they raise.
Plato believed that the existence of universals was required not only ontologically, to explain the nature of the world that as sentient and reflective beings we experience, but also epistemologically, to explain the nature of our experience of it.
He proposed a solution to his problem, but he also recognized the objections to his particular solution. Ever since, except for intervals of neglect, philosophers have been worrying about the nature and status of universals.
No account has yet been propounded that has come near to receiving universal acceptance; this reflects not merely disagreement on the answers to be offered but also, and perhaps more importantly, disagreement on exactly what the questions are that we are, or should be, trying to answer.
That in some sense or other there are universals, and that in some sense or other they are abstract objects— that is, objects of thought rather than of sense perception—no philosopher would wish to dispute; the difficulties begin when we try to be more precise.
They may be indicated (although not defined) by the abstract nouns that we use when we think about, for example, beauty, justice, courage, and goodness and, again, by the adjectives, verbs, adverbs, and prepositions that we use in talking of individual objects, to refer to their qualities and to the relations between them.
In saying of two or more objects that each is a table, or square, or brown, or made of wood, we are saying that there is something common to the objects, which may be shared by many others and in virtue of which the objects may be classified into kinds.
Not merely is such classification possible, for scientific and other purposes; it is unavoidable: All experience is of things as belonging to kinds, however vague and inarticulate the classification may be.
Whatever we see (to take sight as an example) we see as a something—that is, as an object of a certain kind, as having certain qualities, and as standing in certain relations to other objects—and although every individual object is unique, in that it is numerically distinct from all others, its features are general, in that they are (or might be) repeated in other objects. Even if there were only one red object in the world, we would know what it would be like for there to be others, and we would be able to recognize another if we were to meet with it.
Generality is an essential feature of the objects of experience, recognition of generality is an essential feature of experience itself, and reflection of this generality is shown in the vocabulary of any language, all the words of which (with the exception of proper names) are general.
Universals are, by tradition, contrasted with particulars, the general contrasted with the numerically unique, and differing theories of universals are differing accounts of what is involved in this generality and in our experience of it. The leading theories of universals—realism, conceptualism, nominalism, and resemblance theories—can best be explained by an examination of the doctrines of the main exponents.
In following that sequence we shall be adhering approximately (although not precisely) to the chronological order in which the rival theories developed, and we shall be historically selective, in that we say almost nothing of the periods in the history of philosophy during which the controversies continued (for example, medieval philosophy) but of which a detailed knowledge is not necessary to a general understanding of the issues involved. The aim here is to present the different views that have been held, not to trace the fortunes of each view throughout the history of the subject.
Realist and conceptualist theories of universals are, by long tradition, regarded as opposed because according to realism universals are nonmental, or mind-independent, whereas according to conceptualism they are mental, or mind-dependent.
For the realist, universals exist in themselves and would exist even if there were no minds to be aware of them; if the world were exactly what it is now, with the one difference that it contained no minds at all, no consciousness of any kind, the existence of universals would be unaffected. They are public somethings with which we are somehow or other acquainted, and a mindless world would lack not universals but only the awareness of them: They would be available for discovery, even if there were nobody to discover them.
For the conceptualist, on the other hand, universals are in the mind in a private sense, such that if there were no minds, there could be no universals, in the same way as there could be no thoughts or imagery or memories or dreams. As will be seen, whatever may be said for or against realism, pure conceptualism cannot be a satisfactory theory, for it is essentially incomplete; it says something about our consciousness of universals but nothing at all about any basis for this consciousness.
Consequently, philosophers who have been conceptualists either have been so because they have been interested only in the epistemological question, in the conceptual structure of human thought and experience, or have combined their conceptualism with another theory designed to answer the ontological question—that is, the question what there is in the world corresponding to our mental concepts or ideas, what our concepts are concepts of. The antithesis between the two theories of realism and conceptualism is not, therefore, as clear-cut as it has often been presented to be.
The two main versions of realism are those of Plato and Aristotle. Plato’s came first, and the difficulties it raised, some raised by Plato himself, others added by Aristotle, were what led Aristotle to devise his own quite different, but still realist, account.
Plato and Aristotle were both realists in that they accorded to universals an existence independent of minds; where they differed was on the nature of the existence and the status that they believed universals to possess.
PLATO. Although it is possible to give, in some detail, a statement of what may be called Plato’s theory of universals, and to give it full documentary support by quotations from his writings, we would be mistaken to regard it as a final and fully worked out theory.
It was a theory toward which Plato can be seen working his way throughout his philosophical career, not so much by independent arguments as by intertwining strands of thought, all leading in the same general direction. There were a number of facts about the world and our experience of it by which he was impressed and puzzled. His theory evolved as an explanation of them, but he was never satisfied that he had solved his problem.
He was his own first critic, and a penetrating one, and to the end of his life he was torn, as is brought out in his dialogue Parmenides, between the conviction that his theory was fundamentally correct and the recognition that it posed problems that he found himself unable to solve. It should not be thought, therefore, that he ever produced a final account that he was prepared to rest content with and that needed an Aristotle to find fault with it.
Plato’s interest in questions about universals was first aroused by Socrates, by whom he was greatly influenced, whom he introduced as one of the speakers in all his dialogues (with the single exception of The Laws), and who in all but the later dialogues appears as the central character actually directing the conversation.
Unfortunately, we are presented with difficulties of interpretation, the details of which we shall not enter into here, because our knowledge of Socrates is derived entirely from descriptions given by other writers, one of whom was Plato. Hence arises the problem of deciding which of the doctrines ascribed to “Socrates” in the Platonic dialogues are those of the actual Socrates and which of them are extensions or even entirely new doctrines developed by Plato himself.
In general, it is accepted that the “Socrates” of the early dialogues does represent the views, and even more the methods of philosophical inquiry, of Socrates himself but that as time went on Plato more and more used him as the spokesman of Plato’s own views, the transitional stage being marked by such dialogues as Phaedo and the Republic.
We may conclude that while Socrates did not explicitly hold a theory of universals (and we have Aristotle’s word for it, in Metaphysics 1078b, that Socrates did not hold the view Plato put forward), his philosophical questions were such that Plato held they could not be answered except by such a theory; in other words, Plato, in putting a theory of universals into Socrates’ mouth, was not attributing it to Socrates as what he had actually expounded but was maintaining it as the logical consequence of Socrates’ own arguments: Socrates stopped short of propounding such a theory himself but was logically committed to it.
Socrates’main interest was in the human virtues, and his aim was to secure a satisfactory definition of the virtue under discussion. His questions were all of the form “What is X?,” whe re “X” stood for beauty, courage, piety, justice, and so on, in one case (Meno) even virtue itself.
The answers that he received he rejected because they were too narrow or too wide, but more commonly because instead of giving the essential definition of the virtue they gave instances of it or mentioned kinds of it.
Thus, it was no answer to the question “What is piety?” to reply that a man is acting piously if he prosecutes a murderer; again, it was no answer to the question “What is virtue?” to reply that the virtue of a man consists in managing a city’s affairs capably, that a woman’s virtue consists in managing her domestic affairs capably, that there are different virtues for an old man and a young man, for a free man and a slave, and so on.
Granted that there are many virtues, what is wanted is the one and the same form that they all have and by which they are virtues. The search, then, is for the single and essential form common to all things of the same kind, by virtue of which they are things of the same kind.
The “things” about which Socrates in fact asked his questions were limited because his philosophical interest was limited, but even he did not confine himself to human conduct. He acknowledged, for instance, that health or size or strength must be the same in all its instances, with the consequence that we answer the question “What is health?” only when we have given the essence of health—that is, what is common and peculiar to all instances of health.
Plato took this further and maintained (although not without hesitation) that there must be an essence common to all things of a given kind, whatever that kind was. It would apply not only to abstract virtues, such as justice and courage, but also to natural objects, such as trees, and to artifacts, such as tables.
An object would not be a table unless it had the same essence (of tablehood) as all other tables; despite the different shapes and sizes that individual tables may possess, there must be a single form or essence, common to them all, which constitutes their being tables and distinguishes them from other objects, such as chairs or beds. Plato summarized his position in the statement “We are in the habit of postulating one single form for each class of particulars to which we give the same name”.
And he held it to be true not only of objects designated by nouns (such as “bed” and “table”) but also of attributes or qualities indicated by predicates (such as “beautiful” and “greater than”). As there must be a form or essence of bedhood somehow common to all beds, so there must be a form or essence of beauty (or the beautiful) common to all things that are beautiful.
So far Plato had done nothing more than take over the Socratic contrast between the single general, essential form common to a class of particulars and the particulars themselves and extend it more widely than Socrates had done: He found the same contrast not only in the realms of ethics, aesthetics, and mathematics but also in the everyday world of sense experience.
But he went on to ask the questions that Socrates had never asked, namely what are we to say about the relationship between the universal form and its particular manifestations, and what are we to say about the nature and existence of the universal itself? His answer was to develop the theory known as the theory of Forms, according to which each universal is a single substance or Form, existing timelessly and independently of any of its particular manifestations and apprehended not by sense but by intellect.
His arguments can be distinguished, although not entirely separated, into two general kinds, metaphysical or ontological and epistemological. If knowledge is to be possible at all (and Plato did not doubt either that it was possible or that in certain spheres it was actual), it must be of what is stable and unchanging.
However, the familiar world of ordinary experience does not meet this requirement, for the one constant and striking feature of all objects (and their qualities) in this world is that they are subject to change and decay: Both natural objects and artifacts come, or are brought, into being, undergo changes throughout their existence, and sooner or later die or disintegrate and disappear.
This is the Heraclitean doctrine of flux, which Plato accepted and which he believed required as its counterpart a nonsensible realm of unchanging stability, without which there could be no knowledge. What can be known must be real, unitary, and unchanging: These are the Forms.
|change and decay|
Particulars are only semireal, real to the extent that in some way or other, or to some degree or other, they manifest the Forms, unreal to the extent that being material, they lack the perfection of pure Forms and are subject to the laws of material change and decay. Thus, Forms are required, to confer on particulars such reality as they do have, to constitute their being what they are and of what kinds they are.
A bed is a bed rather than a table because it somehow manifests the Form Bed. A Form is required not only to explain a particular object’s being what it is but also to cause its being what it is; the doctrine is thus not merely a logical but a metaphysical doctrine.
Plato emphasized this in the analogy of the sun (Republic VI), where he compared the chief Form of all, the Form of the Good, with the sun, which as the light-giving and life-giving agent in the physical world is the prime material cause of natural life as well as of our awareness, through our senses, of the material world.
Another consideration that led Plato to suppose the Forms as transcendent substances was the presence of what he thought to be contradictions in the material world: What is real cannot contain contradictions; therefore the material world cannot be more than an appearance of reality.
That a single object should be both beautiful (in one respect) and ugly (in another), or large (in comparison with a second object) and at the same time small (in comparison with a third), was enough, in his view, to show that the Forms were more than immanent.
Therefore, not only must there be Forms in order to cause particulars to be what they are, but the Forms must be separate from the particulars because they must be free of the imperfection and defectiveness with which particulars are inevitably infected. The Forms are thus not only independent substances but perfect and ideal patterns, which particulars must fall short of.
This comes out especially in the consideration of mathematical (primarily geometrical) and value concepts, namely those of ethics and aesthetics. For a line to be straight or a figure to be circular, there must be the Forms of Straightness and Circularity.
But it is well known that no actual line is ever perfectly straight and no figure is ever perfectly circular; however carefully and precisely drawn, it possesses some curves or kinks that more minute scrutiny could disclose.
And what we are thinking about when we study or discuss a geometrical theorem is not the diagram of the circle drawn, freehand or mechanically, on the blackboard but the circle represented by the diagram.We thus have both the diagram of the circle, adequate as a diagram but imperfect as a circle, and the perfect Form of Circularity of which it is a diagram.
While this gives rise to the question, which cannot be pursued here, whether Plato distinguished between the Form of Circularity (of which there could not be more than one) and a Perfect Circle (of which, if there could be one, there could be more than one—as required by, for example, a theorem involving two intersecting circles), there is no doubt that he did think a Form not only was the perfect pattern, of which a particular was an imperfect manifestation, but also was what the particular would be if, per impossibile, it could be perfect.
Thus, to take an aesthetic example, Beauty (or the Beautiful) not only is the pattern that beautiful particulars inadequately manifest but also is itself perfectly beautiful; it is a substance possessing in perfection the essence that its derivative particulars possess only partially or in some degree.
As Plato came to realize later, and as Aristotle repeated, if a Form stands to its particulars as “one over many,” and if the Form is an ideal pattern of which the particulars are imperfect copies, then an infinite regress argument (known as the third-man argument) is generated: For the Form to be predicable of itself as well as of its particulars, it must share a character with them; but then there will be a Form of this character; this second Form will be predicable of itself, requiring a third Form of it, a fourth, and so on ad infinitum.
As was indicated above by the geometrical example, Plato believed that his theory of Forms accounted for the possibility of knowledge of universal truths, which was the only kind of knowledge strictly meriting the name.
When, by working out or following the proof, we learn that a square constructed on the diagonal of a given square has an area equal to double the area of the given square, we have learned a truth that is necessary and universal.
It is not something that happens, as a matter of fact, to be true of the squares in our diagram but might turn out not to be true of some other squares; that is, it is not an empirical generalization that subsequent experience might show to be false as a generalization.
We have a piece of a priori knowledge, which no possible experience could affect, namely that if a square has a given area, and if a second square has its sides equal in length to a diagonal of the original square, then the area of the second square must be double the area of the first.
Our knowledge is not knowledge of our diagram squares, or any others that we care to draw, for, as we have seen, they are not in fact squares. But it is knowledge, and the only thing, therefore, that it can be knowledge of is the Form Square (or the Square).
What defeated Plato in any attempt to give a complete account of his theory was the problem of describing the relation of Forms to particulars. In different places he spoke of the Forms “being in” their particulars, of particulars “participating in” their forms, and of particulars “copying” their forms.
Literal interpretation of any of these phrases gives rise to logical difficulties, and to take them metaphorically is to leave the statement of the theory imprecise and the problem unanswered. In Plato’s final writings (Epistle VII) on the subject there are signs that he was inclined to think that the fault lay with the inadequacy of language to describe what he wanted to describe, but the trouble is deeper than mere paucity of vocabulary.
We can form some kind of a picture of his two worlds if we think of the world of Forms as actually existing somewhere, populated by objects like the Standard Meter and the Standard Pound, and we can then think of actual particulars as being imperfect copies of the originals. But that picture, taken literally, is false, because Plato’s Forms do not exist in a place or at a time.
The mystery of their “existence” becomes impenetrable when we are asked to use the word exist in a way that we are incapable of conceiving. In his theory of Forms, with the Forms not immanent but transcendent, the problem of their relation to particulars becomes not almost impossibly difficult to solve but in principle insoluble.
ARISTOTLE. Aristotle, Plato’s pupil and successor, is often regarded as the careful scientific-minded thinker, anxious to restrain philosophy within the range of the observable and to avoid the imaginative speculations of Plato.
While this picture is in general correct and in particular fits Aristotle’s criticisms of Plato’s theory of Forms regarded as universals, his own theory of a Form as the object of a definition that describes a thing’s essential nature becomes in the end as obscure as Plato’s.
His criticism that Plato’s theory does nothing to provide a scientific explanation of the nature of things applies equally forcibly to his own theory of essences, and natural science, as we know it, began to progress only when, many centuries later, it liberated itself from this aspect of Aristotelianism.
But Aristotle’s theory of universals, which is nowhere fully elaborated and has to be pieced together from different passages, is important, both because it offered an alternative to Plato’s and because it is more obviously attractive to common sense. His objections to Plato are numerous and detailed but are not all of equal weight.
Basically, apart from the infinite regress argument, which he took over from Plato, they come to two: First, that Plato, by making the Forms perfect, separate substances, introduced an unnecessary and unhelpful duplication, and second, that Plato confused the categories of substance and property. Nothing is accounted for by making the Forms perfect patterns of particulars.
To attempt to explain the nature of one set of entities by postulating a second and better set does not solve a problem but merely repeats it at a different level: Whatever the question was that needed to be answered about particulars, it will need to be answered again about the Forms; mere multiplication answers nothing.
Second, Plato was guilty of a logical mistake in treating a Form both as an individual substance (which the “separation” thesis requires) and as a property (which it would have to be to be a universal). Substances are individuals and have properties, but they cannot be properties, yet Plato’s theory treats them as both.
For Aristotle the only true substances were single individual objects, such as Socrates or this table. (It is true that Aristotle introduced a difficulty by treating genus and species also as substances, for they are what it is the aim of science to know, but they are secondary substances, and the knowledge we may gain of them is knowledge about primary substances—that is, the individual objects met with in experience.)
Universals, therefore, are not substances existing independently of particulars. They exist only as common elements in particulars: The universal X is whatever is common to, or shared by, all x’s; it is what is predicated of the individual. Individual objects are to be classified into kinds according as they share the same property, and the kinds are to be subdivided into genus and species by the differences between more determinate properties.
Thus, all colored objects belong to the genus “color” because they all alike have the property of being colored, whereas red objects and green objects belong to different species of the genus, because the first have the property of being colored red and the second have the property of being colored green. One of the primary tasks of natural science is to divide and classify natural objects by genus and species into the real kinds to which, by nature, they belong.
Aristotle’s theory is more economical than Plato’s, requiring only one world of being instead of two, the contrast between the two theories being indicated by the labels that they later acquired in medieval scholastic philosophy: Plato’s was a theory of universalia ante rem (universals independent of particulars), and Aristotle’s of universalia in rebus (universals in things).
And with the possible exception of ideal concepts, such as those of geometry, which Plato had argued had no actual instances, Aristotle’s account seems better to fit a fact, or what we take to be a fact, of human experience, namely that a particular really is an instance of its universal.
Not only should we say that we get our idea of red, for example, from seeing red objects, such as fire engines or ripe tomatoes, but we should also say (except for philosophical theories of perception) that the object really was red, not that (as with Plato) the tomato tried unsuccessfully to be red but that (with Aristotle) it actually was red.
The properties that an object has, and that together constitute its nature, its being an object of that kind, whatever that kind may be (for example, whether it is a horse or a table), are really in the object, in some sense of “in.”
If objects do not and cannot possess any of the characteristics that according to experience and the scrutiny of observation they appear to have, then scientific knowledge becomes either altogether impossible or unrelated to the natural world.Aristotle’s view avoids the Platonic paradox that nothing in the observable world can ever be what it seems to be.
The contrast between the two views comes out again in their accounts of how we apprehend universals. They are agreed both that awareness of universals is implicit in ordinary sense experience (for it is this awareness that conditions our experience as being what it is) and that we are aware of universals not by sense itself but by intellect.
Plato could not say that we become aware of them by abstraction from particular instances, because they have separate existence and never are more than defectively instantiated: If our concept of X were only what we could abstract from imperfect instances, we never could apprehend X itself. The human soul has prenatal knowledge of universals and of their mutual relations, and postnatal experience of the ordinary world serves, or may serve, to revive this knowledge in suitable circumstances.
Thus, experience does not directly provide us with new apprehensions (of universals) or with new knowledge of necessary truths (connections between universals) but acts as a stimulus to remind us of what we already know but have hitherto in this life forgotten.
Plato’s argument here, if it is to be regarded as an argument, is a transcendental one (in Immanuel Kant’s sense of the word): Our knowledge is a priori, that is, of such a kind that we cannot get it from experience, although we do get it in experience; therefore it must be innate, that is, knowledge of what we originally knew prior to any experience.
As a transcendental argument it could be effective only if it could be shown that there was no other possible way of accounting for our apprehension of universals and our knowledge of universal truths. And Aristotle thought that there was another, less fanciful and less speculative way, derived from actual experiences and memories of previous experiences.
Apprehension of a universal, or formation of a concept, is not a sudden once-and-for-all business, given in a single experience, but a gradual process. Sense perception gives rise to memory, and memory conditions subsequent perceptions, so that they are not merely perceptions but recognitions of what is in some degree or other familiar from previous perceptions. Awareness of characteristics thus becomes clearer and more explicit with the growth and variety of experience.
By a process of induction, namely intuitive induction, the first primitive awareness of a universal (necessary to any perception) becomes stabilized in the mind, leading ultimately to a clear and articulate concept of it. Thus, for Aristotle, as for Plato, grasp of universals is by the intellect, but it is by the intellect gradually working on what it is at first dimly and indeterminately conscious of in the data of sense perception.
A simple example from arithmetic will illustrate his point. As children we learn to count. We get the idea of 2 from being faced with pairs of objects, and we learn that 2 + 2 = 4 from coming to “see,” for instance, that two apples plus two other apples are equal in number to four other apples.
But we also come, sooner or later, to “see” that the number 2 characterizes any pair of objects, and that 2 + 2 = 4 is a necessary truth, applicable to any two pairs compared with a quartet. We have the power, which becomes actualized in experience, of intuiting clearly the universal in the particular and of intuiting the necessary in the matter of fact; this, for Aristotle, is the beginning of scientific knowledge.
AUGUSTINE. Medieval philosophy was not primarily interested in questions about the nature of human knowledge. But its concern with metaphysics, especially in those aspects that carried theological implications, led to a continuation of the dispute between the two versions of realism and later to a nominalist rejection of both.
Platonic realism was championed by St. Augustine, for whom divine illumination performed much the same function as Plato’s Form of the Good, rendering intelligible by its light the necessity of eternal truths that the human intellect could grasp.
Man is above the beasts, not only because he can acquire, by the mind alone, knowledge of eternal truths, but also because even in sensation he judges of material objects by incorporeal standards: In judging a physical object to be beautiful he implies the objective existence of Beauty, both as a universal and as a standard.
Again, the intelligible structure of the temporal world, which the reason of man (but not the senses of the beasts) can grasp, is itself nontemporal; for example, the concepts and truths of mathematics, although empirically applicable, are timeless necessities.
Ideas as objective essences are exemplars contained “in the divine intelligence.” Thus, Plato’s theory of Forms enters theology, and the question arises whether Augustine in his theory of ideas supposed that men were in direct contact with the mind of God. It is fairly clear that he did not but much less clear how he could avoid it.
THOMAS AQUINAS. The leading exponent of Aristotelian realism was Thomas Aquinas, who, although professing the greatest reverence for Augustine, departed widely from Augustine’s views.
Thomas’s metaphysics is, like Aristotle’s, teleological, maintaining that the nature of things and events is to be explained in terms of the ends that they serve, and he extended Aristotle’s contrasts between potentiality and act, between form and matter, and between essence and existence. Essences are universals, which have no being apart from existence but which are intelligible without the supposition of existence.
The existence of things does not follow from their essence—otherwise existence could not be, as it clearly is, contingent. Universals are apprehended directly by the mind, but only in the material things the nature of which they comprise; they are not to be found in themselves, although by the processes of abstraction and comparison the mind can approximate to thinking of them in themselves.
The chief follower in the Thomist tradition was John Duns Scotus, who nevertheless rejected much in Thomas, such as the distinction between essence and existence, and followed Avicenna in differentiating between the “thisness” of an individual object (which distinguishes it from other objects of the same kind) and the nature of an individual object (which distinguishes it from objects of other kinds).
CRITICISM OF REALISM. Although each of the two versions of realism received vigorous support in the long disputes of medieval philosophy, and although Augustinianism for a time prevailed, Aristotle’s version has had the longer-lasting influence, especially on philosophers brought up in the British tradition of empiricism. That things do have common characters and that the characters are objectively real seems hardly deniable, and this is part of what Aristotle’s theory asserts.
But although it is more hardheaded than Plato’s, it does raise its own difficulties, two of which may be mentioned. First, how much does it in fact explain of what it purports to explain? We do not account for two tables’ being tables better by saying that they have a single characteristic (or set of characteristics) in common than by saying that they are both imitations of a single Form.
And if what is to be accounted for is rather our ground for saying that they are tables, which is a question not about their being tables but about our justification for believing or claiming to know that they are, then admittedly we are perceptually aware of the characteristics of each, and of their similarity.
But is saying that some (or all) of the characteristics of the one table are like (even exactly like) the characteristics of the other what the Aristotelian means to do when he maintains that there is a universal common to them (and any other tables)? This may be doubted, for the Aristotelian asserts that a single universal is present in each of the objects, or that each is an instance of it, all the objects of a given kind sharing in the universal of that kind.
But this is metaphorical talk, and to explain by metaphor is not to explain at all. As a descriptive statement “These two tables are the same shape” is unobjectionable; as an explanatory statement it is less obviously illuminating. Second, Aristotle’s supposition that objects belong to real kinds, which are there for us to discover, ignores the fact that distinctions between kinds or classes are not found but made by us, as was later emphasized by John Locke.
This difficulty is not fatal to the Aristotelian theory, which could accommodate it by emphasizing different levels of determinacy in a universal or class characteristic, but it leads to the question, pursued by Ludwig Wittgenstein in the twentieth century, whether it is necessary that any single characteristic at all be common to all members of a single class.
If it is not necessary, our recognition of objects as belonging to a certain class does not have to depend on the apprehension of a universal shared by all its members, for it may be that nothing, even in the metaphorical sense, is shared. Aristotle’s theory, which prima facie has the merits of being simple and realistic, is perhaps both too simple and not realistic enough.
As has already been indicated, conceptualism should not be regarded strictly as a rival theory to realism, even if some of its exponents have mistakenly so regarded it.
Starting from an extreme Aristotelian position, that everything which exists is particular, conceptualism concentrates on the fact that generality is an essential feature of both experience and language, and it seeks to answer the question how mental concepts are formed, how they can be general if the data of experience from which they are formed are particular, and how words are general in their significance. Nominalism carries the process further by maintaining that only words are general.
Both theories, even if they answered their own question satisfactorily, would have to face the question what basis in reality there is for the generalization inherent in experience, thought, and language. Some versions ignore this question altogether; others answer it in terms of the similarities and differences to be found between particulars.
The essential difference between the theories of conceptualism and nominalism is that while both profess to answer a question about language—how words are general, or how words have meaning—nominalism does it more economically, without interposing concepts between words and what words stand for.
The conceptualist says that a word is general or meaningful because in the mind there is a corresponding general concept; he then has to explain what a general concept is. The nominalist thinks that the meaningfulness of a word can be accounted for without postulating a separate mental entity called a concept.
Conceptualism is primarily associated with the three classical British empiricists, Locke, George Berkeley, and David Hume, all of whom propounded views about what, in the terminology of the time, were called general ideas.
They were all empiricists in that they agreed that all ideas, or the elements that ideas are composed of, come from, and can come only from, experience: The mind can work on what is given to it by sense experience but can neither have ideas prior to any experience (a denial of the doctrine of innate ideas and, by implication, of Plato’s suggestion of prenatal acquaintance with the Forms) nor create ideas de novo.
Thus, the essence of empiricism is the Epicurean doctrine, given fresh impetus in the seventeenth century by Pierre Gassendi, that nihil est in intellectu nisi prius fuerit in sensu (“Nothing is in the mind which is not first in the senses”).
Unfortunately, the passages in the Essay in which he discussed general ideas, or, as he more commonly and perhaps misleadingly called them, “abstract ideas,” are neither so clearly thought out and expressed nor perhaps even so consistent as to save him from varying interpretations. The initial difficulty concerns the word idea itself, which is the key word of his philosophy, but which he neither defined nor used so as to escape ambiguity.
Sometimes when he spoke of ideas in the mind he appears to have meant mental images such as occur in remembering, imagining, and dreaming; in this view thinking is done in images, which are particular in their occurrence and existence but somehow become general in their use.
At other times he meant, or at least has been taken to have meant, that abstract ideas are mental entities different from images. At still other times he showed signs of using the word idea not as the name for any mental occurrence at all but as shorthand for the meaning of a word.
Thus, the idea of red would be not an image of something red but what we mean by the word red or what we think an object to be when we think it is red; to have the idea of red is to be able to use the word red correctly and to be able to discriminate correctly between those objects that are red and those that are not.
Attention here will be paid mainly to the first view, of ideas as images, for it is a conceptualist view; so would be the second, that general ideas are mental occurrences different from images, but this all empiricists in that they agreed that all ideas, or the elements that ideas are composed of, come from, appears to be a view that Berkeley fathered on Locke rather than one Locke actually held and can come only from, experience: The mind can work on what is given to it by sense experience but can neither have ideas prior to any experience (a denial of the doctrine of innate ideas and, by implication, of Plato’s suggestion of prenatal acquaintance with the Forms) nor create ideas de novo.
Thus, the essence of empiricism is the Epicurean doctrine, given fresh impetus in the seventeenth century by Pierre Gassendi, that nihil est in intellectu nisi prius fuerit in sensu (“Nothing is in the mind which is not first in the senses”).
According to Locke we form general ideas by a process of abstraction from particular ideas. In two different places he gave what appear to be two different accounts of abstraction. In the Essay on Human Understanding he said that a general idea—for example, of man—is formed by leaving out of the particular ideas of various individual men all features that are not common to them all and retaining only what is common to them all.
The general idea of animal is arrived at by still further leaving out, “retaining only a body, with life, sense and spontaneous motion, comprehended under the name ‘animal.’” If this passage were taken in isolation, regardless of what else Locke said on the matter, there would be something to be said for the Berkeleian interpretation.
For Locke appears to have been saying that we start with a number of particular images, each, for example, of a different individual man of our acquaintance, and end with something that is still an image but is now a ghostly general image, characterized not by any of the features that are peculiar to any of the individual men but only by all those that all men share. It was not difficult for Berkeley to ridicule as logically absurd the suggestion of a mental image, all the features of which are (as, in this view, they would be) determinables.
|laws of contradiction|
In his polemic Berkeley did not consider the possibility that Locke might have been getting at something different, namely that mental images may be indeterminate, so that the logical laws of contradiction and excluded middle do not apply to them; for instance, a mental image of a cloudless night sky is an image of a number of stars but of no precise number.
Just how far Locke regarded himself as committed to ideas as images and how far he would have regarded his account as being philosophical rather than psychological (if he could have been induced to accept the distinction) is hard to say.
But it is fairly clear that his account is not philosophically satisfactory. He showed himself to be well aware that the real problem is one concerning the applicability and use of general words or terms. But as must have been obvious to him, significant use of words in speech or writing is not in fact paralleled by a corresponding string of introspectable images.
Therefore, at best, his claim that a general word is meaningful because it stands for a general idea would have to involve “stand for” in a dispositional sense; that is, a word is meaningful if a corresponding idea can be found for it. Even then he would be open to the nominalist criticism that nothing is explained simply by duplicating a general word with a general idea.
Furthermore, he stressed that almost all thought is verbal: The use of nonverbal imagery in thinking is restricted to a very narrow and primitive level. And, in fact, in the latter part of the Essay he showed signs of interpreting ideas not as pictures corresponding to words but as meanings of words, particularly when he was discussing modes—that is, concepts not necessarily used with existential reference.
To have an idea, for example, of murder or of gratitude is to understand and use the words murder and gratitude in a certain way, and to have a correct idea is to understand and use the words in the same way others do. The question whether A has shown gratitude in his conduct to B is a question not only what A’s conduct has been but also whether it sufficiently fits the accepted sense of gratitude.
Finally, Locke extended this to all general ideas and rejected the Aristotelian thesis that apprehending universals is apprehending real kinds, or real principles of classification. In maintaining this he was making a move toward a kind of nominalism, for he was emphasizing the fact that concepts, other than those determined by technical or arbitrary definition, are open-ended.
We do not find objects and their features divided by nature or God into real and objectively delimited classes; we observe objects and their features, but the distinction between one class and another is something we ourselves make by criteria of convenience and utility.
Similarities and differences are there for us to observe; whether the similarities are sufficiently close so that we can place the objects in the same or in different classes is for us to decide. A modern example would be the question whether a machine can think, or whether a computer can remember.
Such a question, Locke would insist, is to be answered only by seeing what operations the machine performs and then deciding whether they are sufficiently close to what we mean by thinking or remembering when we talk of our own activities to make it reasonable, rather than misleading, to describe them in these terms.
A consequence of this kind of conceptualism will be that concepts are not permanently fixed, as on a simple realist theory they would be; a concept is liable to development and change, as fresh experience or changes of view show the need or utility of it.
For example, a central question of twentieth- and twenty-first-century sociology, which concerns not only moral outlooks but also legal decisions and the development of law and penal policy, is the question under what conditions a man is to be held not responsible for his physical actions.
But the answer to the question is not to be reached simply by determining whether the physical, psychological, and medical facts of a particular case place it inside or outside the accepted scope of responsibility; it also leads to examining the notion of responsibility itself, which in the slow process of time undergoes modification. Experience being ineluctably conceptual, not only are concepts derived from experience, but concepts shape experience itself, as indeed Aristotle had hinted.
If there were nothing else valuable in conceptualism, it would be of importance as a corrective to the naïveté of extreme realism, which suggests that all the material of human experience falls into a scheme of pigeonholes or a fixed mold and that the task of inquiry is simply to find out what the scheme or mold is.
BERKELEY. George Berkeley, Locke’s immediate successor and fiercest critic, devoted the whole introduction of his main philosophical work, The Principles of Human Knowledge (1710), to a violent attack on Locke’s theory of abstract ideas, for reasons perhaps not primarily concerned with universals at all.
However, it is extremely doubtful whether he had, in fact, either studied Locke carefully enough or interpreted him correctly. Berkeley’s own theory of general ideas as particular ideas that become “general by being made to represent or stand for all other particular ideas of the same sort” is expressed in a way that might be a verbatim quotation from Locke himself.
And Hume’s enthusiastic comment that Berkeley’s view of general ideas as particular ideas used generally is “one of the greatest and most valuable discoveries that has been made of late years in the republic of letters” does Hume little credit; his examination of Locke was clearly no more thorough than Berkeley’s had been.
If Berkeley had done nothing but propound his account of general ideas, his contribution would have been nil. But, in fact, he did much more. Aware that a central strand in the supposed problem of universals was the fact of language and appreciating the question how sounds made by the human larynx or marks made on paper could be used to convey a meaning (this too had been stressed by Locke), he protested against the simple view of unum nomen unum nominatum, that every time the same word is used it is accompanied in the mind by the same idea.
First, this is empirically false, as anybody could find out by noticing the many different ideas (images) he might have on the different occasions he used the word; for example, red might be accompanied sometimes by an image of a red dress, sometimes by an image of a red apple, a red flower, and so on, which might in any case all be different shades of red.
Furthermore, it is not even true that every time a man uses a word that can be accompanied by an image, it is accompanied by one. The actual occurrence of an image, if not necessary, could not help to explain the meaningfulness of a word. Sometimes Berkeley wrote as if an image were necessary in a dispositional sense; a word is significant if a suitable image can be had or produced to correspond to it.
Thus, he compared a use of language—for instance, in conversation—to the use of algebraic symbols in a calculation: We can represent a given quantity by the symbol x, and we carry out the calculation without all the time thinking of the quantity represented by x; what matters is that we can, at any time we want to, replace x by the quantity. Similarly, words for the most part, as actually employed, function as cashable counters.
But Berkeley went on to emancipate himself even from this tenuous servitude to ideas as images. He hinted at it when he said that the important thing is the definition of a word, not the occurrence or recurrence of an idea: “It is one thing for to keep a name constantly to the same definition, and another to make it stand every where for the same idea: The one is necessary, the other useless and impracticable.”
But later he went even further and suggested what can be described as an operational theory of meaning. This is nowhere fully developed, chiefly because he abandoned serious philosophical inquiry while still a young man, but unmistakable indications of it persist throughout his writings.
In the Principles they appear in two ways: (a) the reminder of the diversity of function of language; and (b) the doctrine of “notions.” The tendency among philosophers to try to explain the significance of words in terms of corresponding ideas was due to a simple and entirely false view of language, namely that its sole function was informing, or “the communication of ideas”; this made it easier to think of ideas as pictures translated into words by the speaker and retranslated into pictures by the hearer. (The modern television analogy of visual pictures translated into radio signals by the transmitter and retranslated into visual pictures by the receiving set would not be entirely inapt.) But as Berkeley rightly emphasized, to inform is not the function of language, only one of its functions.
It has others, “the raising of some passion, the exciting to or deterring from an action, the putting the mind in some particular disposition”—to which we could add still others, such as asking questions, praying, vowing, swearing, making promises, declaring intentions, and expressing wishes or fears.
It is not entirely clear exactly what Berkeley intended the doctrine of “notions” to be. He acknowledged that his own principles did not allow him to say that we have (or can have) ideas of everything we may significantly talk of, because they did not allow him to say that we have ideas of mind or spirit (ideas being passive and mind or spirit being active); yet a man who uses the words mind and spirit (to which Berkeley added all words denoting relations) is not uttering meaningless gibberish. Therefore, it must be true of at least some words that we “know or understand what is meant” by them although we can have no corresponding ideas.
In these cases we have notions. Notions, as they appear in the Principles, do not solve any problem (if one exists) regarding how words that cannot be paralleled by ideas can be significant—they merely occur as a label for the fact that there are such words. They are not the answer but appear to be Berkeley’s name for the question.
If by “having a notion of x” he meant “knowing or understanding the meaning of the word x, although not being able to have an idea of x,” then the question how one can know or understand the meaning of an idealess word is not answered by saying that he has a notion, and there is no reason to think that Berkeley deluded himself into supposing that his doctrine of notions actually gave an answer to anything.
The Principles takes the matter no further than the negative conclusion not only that a word need not be accompanied by an idea but also that some words cannot be. This is the beginning of an admission that the intelligibility of language neither requires nor is illuminated by suppositions about mental imagery.
In a much later work, Alciphron (1732), Berkeley returned to the topic and showed how (with the examples of force from physics and grace from theology) although frontal questions such as “What is force?” and “What is grace?” could produce no answer, yet these were genuine concepts, because it was true that the use of them (or of the words force and grace) could lead to fruitful results.
Or again, “the algebraic mark, which denotes the root of a negative square, hath its use in logistic operations, although it be impossible to form an idea of any such quantity.” In allowing that a concept could be fertile even though it could not be cashed, Berkeley was at once breaching the walls of strict empiricism and anticipating the theory construction of modern science, particularly of modern physics.
HUME. Immediately after Berkeley came David Hume, the third of the great British empiricists and the one who has had the most lasting influence on subsequent developments in the philosophy of that school. He devoted an early section of his Treatise of Human Nature (1739) to the subject of abstract ideas, professing to accept Berkeley’s doctrine of general ideas and producing arguments to confirm it. But in fact he was not merely repeating Berkeley’s views.
He took one step backward in maintaining that the use of every general word must be accompanied by a particular mental idea: “’Tis certain that we form the idea of individuals, whenever we use any general term.” But he took several steps forward in suggesting how a given idea can represent others of the same kind—that is, how the idea can become general.
Hume’s emphasis on the role of the word was even stronger than Berkeley’s had been. Whereas Berkeley had supposed that a word becomes general by its relation to a particular but representative idea, Hume put it the other way round, that a particular idea becomes general by being “annexed to a certain term.” “All abstract ideas are really nothing but particular ones ... but, being annexed to general terms, they are able to represent a vast variety.”
Where Berkeley had contented himself with maintaining that an idea became general by representing all ideas of that kind, Hume offered an account of how a particular idea could represent others that were not at the time present to the mind. It did this through custom or habit, by the association of ideas and the association of words.
At any given moment a man has only one individual idea before his mind, but because of the resemblances that he has found in his experience, the one individual idea is associated with others of the same kind, which are not actually present to the mind at the time but which would be called up by the stimulus of a suitable experience or a suitable word.
Thus, the possession of a general idea or a concept becomes a mental disposition, the readiness, engendered by custom, to have some idea belonging to a given kind, when the appropriate stimulus occurs, and the acquisition of a concept will be the gradual process of (1) learning by experience and habituation to recognize instances and to discriminate between them and instances of a different concept; and (2) having the appropriate associations and dispositions set up in one’s mind. To have a concept actually in mind at any given time is to have in mind an individual idea plus the appropriate associative dispositions.
Hume assigned words a key role in his doctrine of association of ideas, supposing that particular ideas, which resemble one another somewhat but not exactly or in all respects, tend to be associated with one another because each is associated with the same general word.
The differences between a ripe tomato and a scarlet-painted automobile are more numerous and conspicuous than their similarities, but the idea of the one can readily be associated with that of the other by the fact that the word red is used of each, and thus the idea of either could serve as representative of the class of red objects, whatever the variety of objects and the differences between the many shades of red displayed. “A particular idea becomes general by being annex’d to a general term; that is, to a term, which from a customary conjunction has a relation to many other particular ideas, and readily recalls them in the imagination.”
One could say that according to Hume we learn to think by learning to talk, not the other way round, and that in learning to talk the chief influence is that of custom and association. Here Hume failed, as nominalism also failed, to see that the attempt to account for the generality of an idea in terms of the generality of a word will not do, if taken only as far as he took it.
In the sense in which he insisted that every idea is particular, so is every word. Whatever reasons there are for denying the existence of general ideas as distinct from particular ideas will also be reasons for denying the existence of general words as distinct from particular words.
Paradoxical though it may seem, the sense in which the word red may be said to be general is such that the word red cannot occur in any sentences at all, for what occurs in a particular sentence is a particular word red.
The fourth word in the sentence “Some automobiles are red”may be very like the first word in the sentence “Red tomatoes are ripe,” but they are different individual words, occupying different positions in space (as printed).
Even in this case they are not exactly alike (for the first does not, and the second does, start with a capital letter), and other “reds” could be even more unlike—for instance, if they were printed in different fonts of type or were written down by different people.
Consideration of this point would have required Hume to say about a word’s being general what he (like Locke and Berkeley) said about an idea’s being general, namely that it was based on (or constituted by) the resemblance between particulars. (Difficulties in making out somebody’s handwriting stem precisely from its deviating more than usual from the familiar resemblances.) Conceptualism therefore comes down, in the persons of these three authors, on the side of resemblance as being the ontological basis of general ideas.
All that actually exists is individual; generalization, or concept formation, is possible only to the extent that individual objects and occurrences, their features, and the relations between them display perceptible resemblances to a greater or lesser extent. But Hume offered, or at least hinted at, a more sophisticated version of resemblance.
According to Locke, two objects would resemble each other if they possessed certain features in common—that is, if certain features of the one were identical (in an Aristotelian sense) with certain features of the other. Thus, one object possessing features abcd would resemble another possessing features adef, but less closely than it resembled one possessing features acdf.
But Hume saw that this raised difficulties for simple (or unanalyzable) ideas or qualities—for example, that “blue and green are different simple ideas, but are more resembling than blue and scarlet; tho’ their perfect simplicity excludes all possibility of separation or distinction.”
They may resemble each other “without having any common circumstance the same.” The notion of resemblance as an ultimate relation, without requiring that the respect in which two objects resemble each other should be a quality identical in each, propounded here by Hume, has been taken further in later developments of his theory.