SOURCE: "The Dialectic of Experience," in Mind, History, and Dialectic: The Philosophy of R. G. Collingwood, Indiana University Press, 1969, pp. 7-58.

[In the following excerpt, Mink discusses Speculum Mentis as a work that introduced and coordinated the major issues elaborated in Collingwood's later writings.]

1 COLLINGWOOD'S PENTATEUCH OF FORMS OF EXPERIENCE: A SUMMARY

Throughout his life Collingwood occupied himself with the relations and differences among art, religion, science, history, and philosophy, regarded sometimes as ways of life, sometimes as types of experience, and sometimes as modes of knowledge. At least one of his books is devoted to each of these, as their titles indicate: Religion and Philosophy (1916); Outlines of a Philosophy of Art (1925) and Principles of Art (1938); The Idea of Nature (1945); The Idea of History (1946); and, of course all of these, but also the Essay on Philosophical Method (1933), the Autobiography (1939), and the Essay on Metaphysics (1940) occupy themselves specifically with philosophy. Only The New Leviathan (1942) does not discuss explicitly and at length these special forms of experience and knowledge.

Again and again he returned to the task of drawing up constitutions for the kingdom of knowledge, and although he never argued that his Pentateuch of Art, Religion, Science, History, and Philosophy is an exhaustive list, in practice it remained canonical, although the importance of Religion as a separate category became attenuated after 1928, as he came to identify his categories less with "ways of life" and more with types and methods of knowledge. In an important way this series of constitutions provides a reduced and brilliant image of the changes (or development) of his thought, which sometimes are matters of nuance or emphasis, sometimes more fundamental alterations.

But although he undertook repeatedly to distinguish history from science, or philosophy from religion (always in such a way as to describe their connections as well as their differences) it is only in Speculum Mentis that he drafted a complete constitution, with equal attention to each of the five areas. While in his later books many of his ideas underwent modification and change, I am convinced (and it is consistent with his own later views about historical understanding) that it is only against the background of Speculum Mentis that his later views can be correctly interpreted.

Speculum Mentis deals not with a single problem but with an entire family of problems, all of which arise out of reflection on the entire range of experience. As stated in the Prologue, the progenitor of these problems is the disintegration in the modern world of the conception—or rather of the possibility, since the conception survives to plague us by its lack of exemplification—of a "complete and undivided life," of experience regarded as an integrated whole, which Collingwood took to be a characteristic virtue of the medieval outlook. This is an outlook which we could not recover if we would—an enterprise which the attempts of neo-medievalists have not made more attractive. Yet, lacking that integrity by which the diversity of human activity was hierarchically organized under the unifying principle of faith, the special problem of modern life is that an increasingly unsatisfied demand for beauty, faith, and knowledge coexists with unwanted overproduction of art, religion, and philosophy. Moreover, these are detached from each other, so that they are separately unsatisfying and collectively unrelated. One can dabble in each, but how to combine them in a single complex unity, a life, is not understood.

There do exist ways of life which are alien but fascinating in their single-minded exclusiveness, and which for this very reason provide rich themes for biography and fiction. It is the plunge into a single form of life which unites Thomas Merton and Gulley Jimson, but at the same time separates the religious life from the artist's life. To one who is curious about both forms of life but exemplifies neither, there are problems of understanding. What connection, if any, is there between aesthetic experience and religious experience? There are, of course, theories about this: the view that religious experience is a work of aesthetic imagination or the counterview that art is a form of religious expression—even when demonic. Both views are, as answers to such questions tend to be, reductive. Similar questions, in their most general form too well known to require listing, arise about the relation of religion to science, science to art, history to philosophy, and so on; again, answers are most often reductive, and are notable more for their number than for their cogency.

In Speculum Mentis, Collingwood sets out unabashedly to adjudicate the rival claims of all the major kinds of experience or "forms of consciousness.' Such an ambition might appear immodest, but it is in fact what philosophers have always attempted collectively and often individually. It is sanctioned by both traditional and contemporary views of the philosopher's job, and both views are represented in the division and classification which Collingwood makes of major kinds of experience. He has two criteria for recognizing an activity as a "form of consciousness": it must in fact be capable of being regarded (even though in the end this may not be consistently possible), as a "way of life," enlisting all of one's faculties and energies, and at the same time it must be a claim to knowledge about the world or to a method of achieving knowledge. The former criterion is the older meaning of "philosophy," the sense in which both slave and emperor could be Stoic philosophers. The latter is the modern meaning of philosophy as the critique of knowledge—the sense in which Hume could give up philosophy as a young man and Whitehead could take it up as an old one. That Collingwood unites the two indicates his belief that the elucidation of experience and the critical analysis of theory can and must be done together.

These two criteria also jointly limit the possible forms of consciousness. They rule out the political and the economic lives merely as such, because utilitarian activity does not make a claim to knowledge although it may be the practical side of a form of consciousness which does. The criteria do not admit the differences between social, monastic, and reclusive life as fundamental, nor do they distinguish between types of personality or between cosmic attitudes such as pessimism and optimism. In fact, Collingwood finds only five candidates: Art, Religion, Science, History, and Philosophy. It is well to capitalize the names, as he does, not as honorific designations nor in order subtly to reify them, but as a reminder that in Speculum Mentis they do not necessarily refer to the professional or institutional activities called by those names in common language and by Collingwood himself in his later books. One might think of them initially as what we comfortably, if not very clearly, refer to as the aesthetic attitude, the religious life, scientific inquiry, the historical consciousness, and the philosophical temper.

Art. By "Art," Collingwood means the activity and the products of imagination suspended from all questions and claims about the reality of the objects of imagination. Art (or imagination) asserts nothing, or rather, it can suppose everything without considering the question of the coherence of different acts or products of imagination. Whether a portrait "looks like" its model is not a question for Art (although this may be difficult to explain to a sitter who has paid for a satisfactory likeness). Nor is it a flaw in Sophocles' Antigone that the officious Creon of that play cannot be recognized as the wise and patient Creon of Oedipus Rex. Hence a primary characteristic of imagination is what Collingwood calls the "monadism of art," the reference being not to Leibniz's principle that "each monad mirrors the universe" but to his description of monads as "windowless," unrelated to and unaffected by all other monads.

Now "art" here clearly means not the artifacts viewed in galleries and the compositions reproduced in concert halls, but imaginative acts, whether of artist or spectator. Aesthetic experience as such does not distinguish between a limited class of things called "works of art" and the rest of the world; it is a possible attitude which may be taken up toward anything. It may be illustrated by what Collingwood elsewhere called the "principle of the picture-frame," referring to the conscious act of attention by which, imaginatively, a frame can be set around part of the extended visual or auditory fields and what lies within the frame perceived as pure spectacle. Such an experiment is not difficult, and its results can be extraordinary: a bleak industrial wasteland, grimy and depressing, can take on an entirely different visual character when a segment is "framed" and regarded, so to speak, two-dimensionally rather than three-dimensionally. (Learning to draw in perspective requires a similar cultivated act of attention; in Collingwood's view it is quite right to regard this as already an imaginative act rather than a technical skill in the service of imagination.) Looked at from this standpoint, it is possible to see (although the example is not Collingwood's) the sense in which photography can be one of the beauxarts. Many would deny photography the name of art on the grounds that a technical apparatus performs all of the functions usually associated with artistic creation. And so it does; but selecting a point of vantage, composing the picture within the frame, and changing the frame in the process of developing the prints are nevertheless all imaginative acts, in Collingwood's sense.

The implications of identifying art with imagination rather than with a class of artifacts, with complex human activity rather than with physical objects, were not fully worked out by Collingwood until The Principles of Art. But the problem at hand is a different one: the claim of Art to be a possible way of life and a form of knowledge. One might think, regarding with a disenchanted eye the excesses of a century and a half of Romanticism, that neither claim has been made good or is likely to be. But it remains to understand why this should be so. In Collingwood's view, it is because the aesthetic life is inherently unstable and the aesthetic claim to immediate and intuitive knowledge is inherently inconsistent. The instability of a life of pure imagination might be thought due to the fact that a world of physical necessities exacts ultimate penalties from anyone who tries to convert it into pure spectacle. Standing in the middle of a busy road, one might manage to regard the approaching vehicles with disinterested imagination, but not for long. And imaginatively framing a bowl of slowly rotting fruit as model for a still-life will hardly protect a painter from the pangs or the effects of hunger. But although such considerations are obvious, Collingwood's point is that the world of fact does not merely constrain the aesthetic imagination but is necessary to it; and therefore in regarding itself as rejecting the world of fact the aesthetic consciousness is deceiving itself. But imagination, as such, is incapable of reflecting on itself, and therefore cannot recognize that it is necessarily dependent on what as a form of consciousness it excludes.

The instability of the "aesthetic life," in fact, is a consequence of the impossibility of maintaining the attitude of imagination without assertion. This attitude is like—and in Collingwood's view it is—an attitude of questioning or supposing which anticipates no answer or resulting assertion. But questioning, he argues, presupposes assertions which make the questions important or relevant; and it implies the possibility of an answer. Otherwise there would be no way of distinguishing one supposai from another.

This may seem a strained and over-intellectual analogy when applied to art. It is not without relevance to representational art, the epic and the novel, drama, and narrative dance; but in what sense are absolute music and non-objective art "supposais" or "questions"? They are, of course, suspensions of assertion; but a painting by Kandinsky would seem to be a suspension of assertion in a sense quite different from Dante's Inferno. The latter, despite its topographical detail, does not assert that one will find the entrance to Hell or the Mount of Purgatory if one explores the actual surface of the world we walk. There seems a lacuna here. In part, it was recognized and developed by Collingwood in The Principles of Art, where by careful argument he identified imagination with expression (an identification which, he remarked in Speculum Mentis, Croce had unsuccessfully attempted). But even so, it would be a mistake to suppose that the attitude of "questioning" means that a work of art asks a specific question or set of questions which could be given alternative verbal formulation. It refers rather to the act of imagination in general, not to this or that act of imagination; and in this sense, an imaginative act may be considered a general supposai that its object can be isolated from all other experience—without regard to what this other experience may be.

The instability of imagination as a claim to knowledge results from the fact that this claim cannot even be made without abandoning the standpoint of the immediacy of intuition. This may seem obvious, but there is also a special argument for it which is essential to Collingwood's project of relating Art to other forms of experience. The inadequacy of any general theory of intuitive knowledge is simply that no theory itself is intuitive. A claim to knowledge must at least be defended by an argument whose cogency as an argument is not incompatible with the truth of its conclusions. But this is just what occurs when the deliverances of intuitive immediacy are extended to include all possible knowledge: the possibility of an argument for this conclusion is ruled out together with all other arguments. It is not prima facie self-contradictory to believe that in art truths are revealed which are not accessible in other ways, and indeed this has been widely believed. But in any case one cannot defend this view without appealing to other standards of knowledge beyond the revelations of art. In his well-known study of Beethoven, Mr. J. W. N. Sullivan claimed that Beethoven had reached heights of understanding not available to more earth-bound spirits, and had expressed these insights in his late quartets and sonatas. But how could Sullivan discover this? If by the same process, presumably we should have another quartet—by Sullivan—rather than a book. But if by some other means, then Beethoven's late works, if they do in fact show forth a kind of knowledge, are checks for large amounts on a bank which is never open for business.

Collingwood's special argument for rejecting the claims of intuition both rests on and reveals his fundamental conception of experience, a conception whose most notable feature is the rejection of all dualisms: the dualism of data and interpretation (here he departs from empiricism), and the dualism of emotion and intellect (here he departs from rationalism in both its exigent and mitigated forms). In this case the dualism rejected is that of intuition—whether sensation or imagination—and thought. All experience, from the perception of a color to the logical analysis of abstract argument, Collingwood regards as both intuitive and conceptual; "it is all intuitive and all conceptual.' The merits of this thesis may be debated later; at the moment it is its consequences that we are after. And the immediate consequence is that Art as a claim to knowledge is a claim to be merely intuitive, whereas it is at the same time implicitly conceptual although it represses its dim awareness of this fact and cannot in fact make it explicit without ceasing to be Art as such.

But to say that Art as a form of life is implicitly conceptual does not mean (as it did for Schopenhauer) that a work of art fleshes out a concept in vivid and particular illustration, as if every narrative must have a moral and every bit of music a "meaning.' Collingwood uses the vocabulary of Hegel: intuition is "immediate," conceptual thought "mediates.' But what is meant is not difficult to see: the immediacy of intuition (which still refers to imagination, regarded as a kind of knowledge) refers to the surface qualities of experience as such: the blueness of a blue patch, the pitch and timbre of a note or the legato or staccato quality of a melody. It also refers to the ease and fluency of a logical inference or to the frustrating puzzlement of a paradox; many years later, Collingwood pointed out in The Principles of Art that as emotion is never independent of reason, so reason is never independent of emotion, and there are "emotions of reason.' "Mediation" is simply experience reflected on and become self-conscious. In Art such reflection might seem to result in the acceptance of "standards," such as correct perspective in drawing, or the Unities in drama; the aesthetics of classicism was such an attempt to subordinate imagination to abstract and even mathematical formulae. Collingwood's point is that, entirely apart from this, any conscious control of imagination has already imposed some standard, however implicit. So the dramatist introduces a character, the painter thumbs out a line ("A little more? There, that's it.') because the work in progress seems to demand it. It is useless to ask him why, or to try to explain why, in any theory of general application. But some standard of relevance, fittingness, or appropriateness is at work here. It is not applied to experience, because it is part of experience itself; but it is not part of intuition itself. There survives here the ancient Platonic doctrine that recognition and comparison are implicitly conceptual; but Collingwood denies that a purely intellectual concept is being applied to a purely sensuous experience.'This blue is brighter than that" is both intuitive (so far as the blues are concerned) and conceptual (so far as the relation "brighter than" is concerned), and neither can be isolated from the other. Without the relation, there are not two blues; and without the blues there is no relation. Intuition and conceptual thought are not two different kinds of experience or functions of mind. Rather, intuition is that aspect of any experience which is immediate and unreflective; thought is that aspect of any experience which reflects upon itself or is capable of doing so. The form of experience called Art is precisely that aesthetic attitude which claims to be wholly intuitive; in fact, however, it is implicitly conceptual. To articulate its structure through self-reflection does not destroy the immediacy of imagination but supersedes its tacit claim to stability and exclusiveness.

Religion. The life of Art, therefore, is an error when it claims to be an exclusive possibility. But this failure is felt before it can be understood. It is felt as the failure of the imagination to assert a world of fact which it necessarily presupposes, to commit itself to a claim which it suggests in failing to make it. Art imagines but does not assert, even though imagining is a kind of supposai and supposai is a transient stage in the process leading to assertion. What carries this process to its end Collingwood calls "Religion.' Religion is imagination which believes in the reality of its own products.

This definition of religion is not novel; it resembles Feuerbach's and is subject to the criticisms which have been brought against Feuerbach. It is easy to object that the definition seems designed to apply only to religions with an elaborated mythology and rules out purely ethical religions such as Confucianism. Yet no definition of religion can be made as a satisfactory inductive generalization which will simultaneously satisfy the demand of including everything ever referred to by the term and yet avoid the sterility of utter vagueness. The utility of definitions in such cases is to call attention to certain features commonly overlooked or underestimated; and in Collingwood's case it is intended to find a way of relating some, if not all, important features of religion to some, if not all, important features of art and science. From this standpoint, the relation, if arguable, is clear and memorable: Religion is Art asserting and worshipping its own object. But since this object is asserted as real, many consequences follow: where Art is monadic, hence pluralistic and tolerant, Religion claims its assertions to be true, hence all incompatible assertions to be false. Where the products of Art are not related at all to each other (although we may compare them), Religion is cosmologica!, and for the first time conceives the world as a single ordered whole; hence it is also social, both in the sense that it defines a community of believers and in the sense that this community must have some attitude toward unbelievers, for example, that "all those who are not for us are against us.' Finally, it is credal, its creed preserving, defining and encouraging the central imaginative act.

But Religion, too, is subject to an inner development of the sort which revealed the instability of Art both as a way of life and as a claim to knowledge. This comes about, again, because Religion, although a legitimate and necessary kind of experience, cannot preserve its characteristics unaltered once it becomes conscious of them. The special problem of religion is that the direct objects of religious consciousness are symbolic, pointing to meanings which they do not contain. Yet their efficacy as symbols depends on the fact that religious consciousness ignores this distinction, which, in Collingwood's language, is only "implicit.' Hence the literal acceptance of ritual and creed is natural and inevitable, but unstable. In one of his apter metaphors, Collingwood says that Religion is "thought growing up in the husk of language, and as yet unconscious that language and thought are different things. The distinction between what we say and what we mean, between a symbol or word and its meaning, is a distinction in the light of which alone it is possible to understand religion; but it is a distinction hidden from religion itself.'

Now this observation is in principle the legacy of the nineteenth century study of comparative religion and of the development of the "higher criticism" of the Old and New Testaments, which ended forever the literalist interpretation of Biblical mythology and chronicle except for sectarian groups determined to remain in a state of intellectual arrest. But Collingwood puts uncritical Religion in a novel framework in showing its development out of Art and its issue in Science. The conflict between "science and religion" can be seen in a new light: its basis is no longer a conflict between the superstitious literalism of religion and an enlightened scientific method, but a conflict between religion not yet conscious and religion become conscious of the distinction which defines it. (The historical implication is that the intellectual method of the sciences was itself a product of this religious instability; and in fact Collingwood maintained this historical thesis in The Idea of Nature, the Essay on Metaphysics, and finally in The New Leviathan.) Once the religious attitude has become aware that what it says is not what it means, it is a legitimate object for rational criticism, but the enlightened rationalism which points out that the concrete imagery of religion is neither historically true nor scientifically possible is using a weapon which has been forged and put into its hands by the development of Religion itself. And yet there is no return; as criticism owes its possibility to Religion itself, so in turn it produces a theology which reinterprets religious assertions to accommodate the meanings which have been uncovered. But as Collingwood observes, this reinterpretation is not a defense of Religion, but its negation. Theology cannot recover the lost innocence of the encompassing religious life, because it has recognized distinctions and accepted critical standards the unawareness of which is an essential characteristic of what it purports to defend.

It has, in being explicitly a mode of thinking, forfeited its chance to return to the innocence of implicit thought.

The transition from immediacy to self-consciousness within Religion is a revolution sharper than the transition from Art to Religion. That replaced supposal by assertion, this replaces language by thought, or rather by a fusion of thought and language in which the two cannot be sharply distinguished. For the first time, therefore, a logic of propositions is possible, the proposition being the meaning expressed in a statement—e.g., the meaning, unstatable in itself, of equivalent sentences. Religion, like Art, therefore, turns out to be a "philosophical error": not as worship, which is the natural state of the religious consciousness, but as theology, which introduces standards of explication of meaning and criticism of inference which are alien to the primary religious consciousness. In fact, "theology is a manifestation not of the religious spirit but of the scientific spirit," and therefore it is not a mode of experience but a transition to a different mode of experience.

Science. By "Religion" Collingwood obviously means something more limited than the extended contemporary senses of this term. On the other hand, by "Science" he means something much broader than the contemporary restriction of the term to the natural sciences. In neither case is he attempting a general description of all the things which happen to be called by a single name, but rather is identifying something like an ideal type to which a variety of instances more or less closely approximate. The first characteristic of "Science" is that is a kind of thinking aware of itself as such or "explicit"; in both Art and Religion, thought is present, but only as "implicit." The primary objects of this self-conscious kind of thinking are not concrete objects of imagination, as in both Art and Religion, but concepts, or abstract universals.'Object of imagination" is itself, for example, a concept, as is "object of thought." And so are all the terms—"imagination," "relation," "supposal," "assertion"—in which Art and Religion are described, but which do not appear in their own vocabularies.'Classification is the key-note of the scientific spirit; but classification is nothing but the abstractness of the scientific concept."

Collingwood's discussion of Science yields nothing in abstractness to Science itself. He is not generous with examples, but it is clear that his paradigm of science is not experimental inquiry but mathematics; it is not Faraday whom he has in mind but Plato and the Platonic insistence on the unintelligibility of phenomena, like triangles scratched in the sand, apart from the intelligibility of concepts, like the concept of triangle. In effect, he imputes Platonism to science (and not unrea* sonably so) when he describes Science as the "affirmation of the abstract or classificatory concept as real." (Art, it will be recalled, makes no assertions at all; Religion asserts the reality of an object, but one which is imaginative rather than conceptual.) Such a description resembles Whitehead's "fallacy of misplaced concreteness" (which it antedated by a year), but it is easier to bring out Collingwood's point if one remembers that he was not, like Whitehead, referring directly to contemporary physics. Otherwise some of Collingwood's statements seem simply perverse: "Sensuous experience is … unnecessary to the scientist, and all he has to do is to think." This is true of a mathematician if of anyone; and Collingwood emphasizes that mathematics is an exact or a priori science, the only one which is an unalloyed instance of his description. He claims, however, that the description has other applications, because it accounts for the mechanism (determinism) and materialism of empirical natural science: mechanism because physical events are regarded as indifferently comparable instances of abstractly formulated laws, materialism because the abstract universal is "indifferent to its own particulars." This is, I think, exactly the meaning of "materialism" which Russell had in mind when he said that "matter is whatever satisfies the equations of physics"; and in this sense Collingwood is quite right in saying that "mathematics, mechanism, and materialism are the three marks of all science"; but the description is less interesting than it seems because "mechanism" and "materialism" are so broadly conceived that they include what ordinarily have been regarded as alternatives to them (e.g., vitalism, since the concept of "enas an abstract universal to its instances).

Collingwood knows perfectly well that modern natural science is empirical and that scientific hypotheses are related in complex ways to statements of experimental and observational evidence. He recognizes the heuristic necessity of factual experience in suggesting hypotheses, even in mathematics; but, strangely, he does not discuss what contemporary philosophy of science has so exhaustively explored, the confirmation and disconfirmation of hypotheses by the facts of the case. In part, this omission is dictated by his scheme: he wishes to reserve the "realm of fact" as the proper object of "historical" rather than of "scientific" thinking. But it is not merely gratuitous, if one distinguishes between an abstract and concrete sense of "fact": what counts as a fact in science is nothing more man the givenness, the datum-character of properties relevant to a hypothesis: "Scientific fact is a fact purged of its crude and scientifically scandalous concreteness, isolated from its historical setting and reduced to the status of a mere instance of a rule." And, of course, it goes without saying that "concrete" fact is not just an aggregate or congeries of abstract "facts."

What Collingwood is aiming at here should now be clear: he wishes to show that what he chooses to call History emerges from Science as its development and fulfillment, as Science in its turn emerged from Religion and Religion emerged from Art. He thinks that the history of science since the Renaissance suggests this because of the introduction into sciences such as astronomy, geology, and biology of the temporal dimension: astronomy becomes the history of the universe, biology the history of species, and so on. This can hardly be taken seriously as more than an academic pun, because "history" in this case means no more than "change over time," and has nothing significant in common with what Collingwood himself means by "History.' But he adverts to a second argument for the necessary passage from Science to History, namely that the framing and exploration of hypotheses itself presupposes the possession of an ordered body of facts, and "these facts, as actually ascertained by observation and experiment, are matter of history". That this argument is not lightly advanced is indicated by the fact that Collingwood restated it without significant change, years afterward, in the conclusion to The Idea of Nature.

Now this argument inverts the commonly accepted opinion that the mere observation and recording of facts belongs to an elementary "natural history" stage of science and is to a more advanced theoretical stage as butterfly-collecting, say, is to genetics. In the accepted view, it is the ordinary activities of historians which are proto-scientific, rather than the ordinary activities of scientists which are proto-historical. Collingwood surely cannot mean that what we know as science will, as it becomes more sophisticated, look more and more like what we know as history. (It has been widely held, and one can at least imagine the possibility, that future historiography will be more like present natural science—like a science of society, that is—whenever it is not merely chronology or antiquarianism. None of this has anything to do, of course, with the use of scientific methods, like chemical analysis or carbon-14 dating of artifacts, by historians.)

Yet Collingwood is not just playing Paris to History's Aphrodite. The clue to his apparent inversion of the order of thought from the unintelligible aggregation of particular facts to the intelligible unity of an explanatory theory lies, I think, in the notion, as yet not clarified, of self-consciousness. Science as science need not be consciously aware of what it presupposes; yet once it has become so aware, it has reached a coign of vantage beyond itself. The analogy here is not to what has earlier been said of the relation between Art and Religion but to the different relation between Religion and theology: theology (which is implicit Science) is Religion become critical of its own meanings and claims. And something like this occurs when Science becomes aware of its own activity and simultaneously aware that, insofar as it attempts to describe, explain, or justify itself, this activity is not itself scientific inquiry, but something else.

Now when Science reaches the stage of recognizing this, Collingwood claims, it is able to see how much it has taken for granted. Beginning in medias res, as it were, it frames hypotheses about whole classes of events which exist for it as data, although the events belong to the past and as such are objects of historical knowledge (even though this may be individual memory). Perhaps this can be accepted as a truism; but it does not seem to justify the startling conclusion that "natural science as a form of thought exists and always has existed in a context of history, and depends on historical thought for its existence.' For one thing, the form of thought which we call specifically history "did not exist before the eighteenth century.' What Collingwood himself refers to in The Idea of Nature as science is at least as old as the Renaissance; how can it "depend for its existence on a form of thought" which it antedates by centuries?

Misunderstanding, at this crucial point, of the relation of Science and History is easy and almost inevitable, and Collingwood himself does not avoid misstating his point or obscuring it by special pleading. His argument in summary seems to be: Science as abstract, theoretical, and hypothetical presupposes a world of concrete facts to which its statements refer but which they do not exhaust. What does deal with them, at least in principle, is History. So Science as a form of thought is less adequate than and is fulfilled in History.

But as Collingwood sometimes fails to distinguish between nature and science, so here he fails to distinguish between history and the theory of history. Hence the shimmering ambiguity of a statement like this: "I venture to infer that no one can understand natural science unless he understands history: and that no one can answer the question what nature is unless he knows what history is.' The word "history," in the most natural reading of this, should mean "historical thought" in the first statement and "historical reality" in the second. Yet if we interchange these, we get two entirely different statements, namely that one can understand natural science only in terms of its own (real) history, and nature itself only as the object of a science historically understood.

To bring out the same point in a different way: one should not interpret Collingwood (despite his own encouragement) to mean merely that scientists use records of past observations, and so on, and therefore in order to be complete scientists should be better historians. Rather he means: if one reflects on the process of scientific inquiry (and a scientist may well not do so) he becomes aware that scientific facts are classes of occasions on which certain observations have been made, and these observations are, for us, historical facts. The problem is not, except in rare cases, whether they are indeed facts; it lies in the concept of historical fact, and this is a matter of the theory of historical knowledge. Thus it is not that a particular piece of scientific research leads on to a specific problem of historical research, but that the scientific attitude becomes aware of itself as an historical phenomenon and raises questions which call for a theory of history. Why, for example, has organized science been a phenomenon of Western civilization, and cumulative science as we know it today a phenomenon of the post-Renaissance period? Can one give a satisfactory account of what natural science is without taking these questions into consideration? It is not surprising that when Collingwood came to write that part of his systematic philosophy which would deal with Science, he produced not a philosophy of science but a history of the idea of nature, and found that different concepts of nature have been the presuppositions of different (Greek, medieval, modern) ideals of scientific inquiry.

History. As Science is the assertion of abstractions (concepts, theories), History is the affirmation of fact. As a form of thought, its object is the past and the present, but it is only incidentally the "study of the past," and that because there are no future facts. But, in any case, History is not the affirmation of facts as a disjunct plurality but of the "world of fact" as a concrete unity; the former is the province of Science, because its "facts" are abstractions from the latter. It is the concreteness of fact, not its pastness, which distinguishes it as the object of History. And the emphasis on "concreteness" rather than on pastness is underlined by Collingwood when he takes perception to be the origin and a specific instance of historical thought. There is a faint echo in this somewhat odd use of the term "history" of Locke's "historical, plain method," a non-temporal meaning which otherwise survives today only in the term "natural history.' It is clearly not Collingwood's intention merely to insist on a niggling etymological propriety; yet what is historical consciousness, apart from the fact that it is not Science, and what is the "concrete" apart from the fact that it is not the abstract?

As usual, Collingwood is less than helpful in his habit of conflating several different meanings in a single term. In this case there seem to be two main ones: with respect to knowledge, the concrete is the complete rather than the partial; with respect to the objects of knowledge, it is the independence of those objects from our ways of understanding them. Moreover, for Collingwood, Science deals only with particulars, artificially marked off or abstracted from the web of relations in which they actually stand; only History can deal with individuality, for "individuality is concreteness.' (And like Spinoza and Hegel, Collingwood's final judgment is that there can be only one true "individual," that which appears to the incomplete perspective of History as "the world of fact.')

To call History "concrete" thus refers at least to our inescapable sense of the expanding web of real relations in the environing world; it is the feeling of "More!" demanded by the real world of our every attempt to capture it in description or explanation, or rather it is this feeling elevated to conscious recognition. But "concrete" may also refer to the world of fact itself, standing resolutely over against any claim to knowledge of it. Collingwood is unlikely to have committed the simple fallacy of confusing knowledge of something concrete with knowledge as itself concrete, even though he does often verge on the philosopher's bad habit of letting nature solve his problems for him. (Speaking of "dialectical opposites" at one point, he says, "Hold up a stick and distinguish its top and bottom; there you have a concrete synthesis of opposites in an individual whole.' But one cannot explain how "opposites" are "synthesized" by pointing to a natural object, as if to say, "It has solved the problem, so surely we can.' Sticks do not have problems, nor syntheses either.) He not only recognizes the difference between historical knowledge and its object, but attributes to this difference the "breakdown of history.' The object of history is the infinite world of facts, an indefinitely complex totality of events and relationships; historical knowledge, on the other hand, is inescapably fragmentary and specialized. Hence its "superiority" to Science lies not in its practice but in its ideal conception of itself. The theoretical concepts of Science do not exhaust the world, but scientific thought can encompass them because they are its own constructions. The ideal object of History is the concrete world itself; it explains an event not by exhibiting it as an instance of a theoretical law but by tracing out in detail its real connections with the web of real events—it seeks the genealogy of events, as it were, rather than their genetics. But this means that every piece of history is necessarily regarded as part of a universal history of which there is no historian. Collingwood's point might be put this way: Science's concept of reality is attenuated but its performance is excellent; History's concept of reality is adequate but its performance is correspondingly unsatisfactory. Science dips a wide-meshed net which brings up a fraction of the ocean's teeming life; History dips a narrow-meshed net which brings up more than anyone can enumerate and describe; it faces the terrifying plenitude of was eigentlich gewesen ist.

So the sense of the concrete is in jarring conflict with the limitations of inquiry: "History is an unstable attitude which leads either back into science or forward into philosophy, according as the intellectual vigour of the historian is exhausted or stimulated by his attempt to get rid of the abstractions of science.' It is natural, of course, for professional historians (whose own form of experience may well be that of Science) to attempt to justify the specialization of historical inquiry by convincing themselves that historical events and processes can be isolated from one another as particular objects of inquiry; this is exemplified, for example, in periodization, although every historian will allow that there are no cut-off dates, the events earlier or later than which are of no relevance to, say, the Renaissance historian. But specialization cuts the Gordian knot only by relapsing into a prehistorical mode of experience: either it atomizes history into events which can be classified and regarded as instances of general laws, or it decomposes history into wholly unique and individual histories (e. g., biographies). The former is the atomism of Science, the latter the "Monadism" of imagination, or Art. So the historical consciousness is as unstable in its own way as the modes of experience which have preceded it. It is, in fact, the "reductio ad absurdum of all knowledge considered as knowledge of an objective reality independent of the knowing mind.' History, as it were, makes explicit an insoluble problem which was already implicit in the religious consciousness: it is the assertion of a completely objective and independent reality which in the end turns out to permit no knowledge of such a reality sufficient even to support the initial affirmation.

Probably no one would deny that this problem—or puzzle, or impasse—is the armature on which the history of philosophy has bit by bit sculpted its own body. Such a metaphor, if a bit strained, is a not inapt reference to Collingwood's view of mind, however, which he regards as having, like the history of philosophy itself, continuously recreated itself out of its own substance. So the Gordian knot of History's inability to realize its own ideal can be untied only by achieving the self-conscious recognition that the problem is one of its own creation; but in attaining this recognition it passes over into Philosophy. In technical terms, Collingwood regards History as taking for granted an epistemological distinction between subject (the activity of knowing) and object (the object of knowledge). But this distinction is itself an abstraction, a survival, as it were, of Science in the mode of History. How could it be avoided? By recognizing that "the world of fact which is explicitly studied in history is … implicitly nothing but the knowing mind as such.'

This may seem like solving a puzzle with a paradox; and Collingwood barely pauses to justify the identification of subject and object by the a priori argument that a mind's error about its own nature distorts its activity in a way appropriate to the error, from which it follows that no object of mental activity can be independent of the way in which it is known.

It would, I think, be a mistake to try to distill from Speculum Mentis alone what Collingwood might have meant by his cavalier identification of the historical object as the "knowing mind.' The record of his succeeding books, especially The Idea of History and the Essay on Metaphysics is, as we shall see, clearly the history of his continuing attempt to explain it to himself. In Speculum Mentis he was bemused by the use to which he could put the formula without fully elucidating or justifying it: he could use it to good effect in bringing about the transition to Philosophy as the "self-consciousness of mind" which History achieves, at the expense of ceasing to be History once recognition is achieved; and, more importantly, he could use it as a stick with which to beat the contemporary philosophy of epistemological realism which in his Autobiography, with bitterness unsoftened by time, he described as a "futile parlour game" for "minute philosophers.' But in exploiting the latter use he defeated the former purpose. It is not at the level of History but at the level of Philosophy that the self-knowledge of mind is relevant to the issue of realism; and he might better have left History at the stage of recognizing that it itself belongs to the world of fact which is its object, so that there is a history of histories, not as another field of specialization, like the history of plumbing, but as a constitutive element in the historical consciousness as such. For Gibbon the history of the Roman Empire was a series of purely Roman events. But the history of the Roman Empire now includes Gibbon and others, at least in the sense that a contemporary historian is aware that his concept of history is ingredient in his work and itself is part of the history of histories of the Roman Empire. The logic of the concept of history, it is evident, is very intricate. For Collingwood it is enough to show that History passes through phases leading to an awareness of itself as part of its own object. The elucidation of this state is then no longer History but Philosophy.

Philosophy. In the series beginning with Art and continuing through Religion, Science, and History, each stage has proved to be the explicit formulation of something implicit in the earlier stage; each is an achievement but at the same time an error regarded from the standpoint of the next stage. But the error of each is not like that of a self-sealing delusory system but rather like the unstable error of self-inconsistency, which itself throws up the criterion of consistency by which it may criticize and transcend itself. At each stage, the activity of thought (which is a generic term comprising imagination, conceptualization, affirmation and denial, etc.) assumes itself to be distinct from its object. Thought which reaches the stage of explicit self-consciousness and has itself as its own object is Philosophy. And its object is not merely itself as a diaphanous activity, of course, but itself together with its objects. To choose an example which is not Collingwood's, it is a debatable question (at least for mathematical intuitionists, such as Brouwer and Heyting), whether the kind of inference involved in reductio ad absurdum proofs is a satisfactory proof, indeed whether the Principle of Excluded Middle holds at all. Now this and similar arguments about the foundations of mathematics are arguments about mathematics rather than mathematical arguments; insofar as they are about the standards of mathematical reasoning, they are also about its objects: no one ignorant of mathematics could think about mathematical thinking. So the philosophy of mathematics is not a branch of mathematics, like abstract algebra, but meta-mathematics (it will be remembered that mathematics is an instance of Science, in Collingwood's terms); and similarly historical thinking, conscious and critical of itself, becomes metahistory or Philosophy.

Each direct form of consciousness gives rise to its reflective form, and this reflective form may either pass over into the next form of consciousness or remain (no doubt by what C. S. Peirce called the "method of tenacity") in the dogged and repetitive affirmation of itself. As the latter, each form generates an erroneous and dogmatic philosophy. Dogmatism, in fact, is "simply the resistance which a given form of experience presents to its own destruction by an inner dialectic.' Aesthetic philosophy (Collingwood obviously has Bergson in mind although he does not mention him by name) exalts feeling and "intuition" as a substitute for intellect. Religious philosophy, whether as theism or atheism, interprets the metaphorical statements of religion as literal, and undertakes to prove or disprove what the religious consciousness itself is unable even to define. Scientific philosophy is the justification, in logic and metaphysics, of the description of the world in abstract concepts. Finally, Historical philosophy is represented by epistemological realism, the view that knowledge makes no difference to its objects.

If each of these is an "error," philosophy as such (or "Philosophical philosophy") must be the vantage point from which this can be recognized and argued. But what is this vantage point? Insofar as each is or claims to be a kind of knowledge, it would be nothing more than special pleading to claim that "Philosophy" somehow knows as they are the objects which are in some distorted way grasped by Art, Religion, and the rest. Yet in some sense this is nevertheless the case, although only in the sense that "object" has for Philosophy alone a critical meaning. Collingwood does not mean that Philosophy is like an experimental psychologist's comparing with what he sees to be a circle the reports that it is elliptical by subjects wearing distorting goggles. Rather it is like a psychologist who knows that what he himself see as elliptical is to be accounted for by the particular circumstances under which he himself sees it. Philosophy does not have a "truer" account of nature than does Science; but it has the explicit recognition, which is not essentially part of Science itself, that the idea of nature itself is partly constituted by the characteristics of Scientific thought. Nor does Philosophy have privileged access to the world of concrete fact which is the unattainable object of History. It is rather the self-consciousness of the genuinely historical attitude which recognizes that it is itself a determining part of the object of inquiry. There is no implication that Philosophy is free error, but at least it is free from the necessary errors consequent for each of the other forms of experience on its failure to recognize the extent to which it has determined what it supposes itself to have found.

Such a failure of self-consciousness is what Collingwood refers to in distinguishing the "concrete thinking" of Philosophy from the error of "abstraction" which is imputed to all the other modes of experience; and it is, I think, all that he means. In the vocabulary of Hegel, from whom the terms have been borrowed, "abstract" means "partial" and "concrete" means "whole.' For Hegel no theory is entirely false and no theory is entirely true, except that theory which shows the relation to each other within a single ("concrete") system of each partial or "abstract" theory. Collingwood has borrowed the terms but not all of their meanings or implications; he rejects, for instance, as "mere mythology" Hegel's notion of a "world-spirit" whose development appears in the evolution of human institutions and of nature itself. For Collingwood, the common characteristic of all forms of "abstract thinking" is the gratuitous assumption of a distinction between subject and object in which the latter is taken to be wholly different from and other than the former.'Concrete thinking" is the recognition that subject and object "can only be distinctions which fall within one and the same whole, and … this whole can only be the infinite fact which is the absolute mind.' And what holds for the subject-object distinction holds mutatis mutandis for other distinctions as well: condition and conditioned, ground and consequence, particular and universal, individual and society, determinism and indeterminism, "and in general every form of the two complementary abstractions one of which denies the whole to assert the part, while the other denies the part to assert the whole.'

It would be too mild to say that philosophers today are suspicious of references to "absolutes" (not to mention those who equally suspect references to "minds"). And it cannot be denied that Collingwood is more prophetic than analytical in his section on Philosophy. But if the language is Hegel's, the voice sounds strangely like the accents of Locke; and if Collingwood permits himself a poetic license which in his later work he did not disavow and even undertook to defend, he clearly does not intend by "absolute mind" any frivolous hypostatization. Nor is it merely an "ideal concept," i. e., what mind could be if it were not so unfortunately what it is, nor yet is it a set of characteristics shared by all minds (such a description would obviously betray the abstractness of Science playing the psychologist). Collingwood never abandoned the view that mind "is what it does"; and "absolute mind" is therefore the historical record, with all its richness of difference, of human activity into which thought enters in any of its forms. So the record of "absolute mind" does not include knee-jerks, breathing, belches, and blinks (although it may include blushes and winks); but it does include, as for Hegel, works of art, religions, sciences, political and legal institutions, systems of philosophy, and so forth, and it is only through the construction of such external worlds that "mind can possibly come to that self-knowledge which is its end.' It is in the history of these human worlds that mind discovers the speculum mentis, the mirror of the mind.'The absolute mind is an historical whole of which mind is a part," but a part, it is emphasized, which at the level of self-consciousness becomes a different part and therefore alters to a corresponding degree the whole itself.

2 THE DOCTRINE OF THE "CONCRETE UNIVERSAL," AND ITS CONSEQUENCES

Yet, "the absolute mind is not one stupendous whole! It lives in its entirety in every individual and every act of every individual, yet not indifferently, as triangularity is indifferently present in every triangle, but expressing itself in every individual uniquely and irreplaceably.' This comes near to being an Orphic saying, yet its meaning is not open to arbitrary interpretation and can, I think, be illustrated from a major aspect, omitted in the summary so far, of the five forms of experience.

Several times Collingwood describes the object of History or Philosophy (but not of Science) as a whole whose parts repeat in their structure the plan of the whole. This is, in fact, a summary description of the difference between Science, History, and Philosophy: the objects of Science, which are abstract universals, are not wholes of which their instances are parts, nor do they stand to their instances as something whose structure is repeated in the instances. Presumably Collingwood means by this the well-known fact that the concept of circle (or the class of circles) is not circular—as the concept of man does not have a backbone. But whereas the object of Science is the abstract universal, the object of History is said to be the concrete universal, which is by definition a kind of whole whose essential characteristics are also characteristics of its parts. The difference between History and Philosophy, in brief, is that History cannot grasp this object and Philosophy can, because History does not recognize, and Philosophy does, that nothing can meet this specification except mind itself, embodied in or projected into its activities.

Now the objection to Collingwood at this point will be either that the notion of a concrete universal is too unclear to admit of reasonable dispute, or that it is clear enough but refers to nothing real. As we shall see, there is something to be said for the first objection; and after Speculum Mentis Collingwood dropped the Hegelian terminology of "organic whole" and "concrete universal.' But he did not drop the nucleus of ideas which he was trying to express with this borrowed terminology, and they reappear in the Essay on Philosophical Method as the "overlap of classes" and the "scale of forms" and the principles associated with these ideas. One should at least not suppose that by "concrete universal" Collingwood (or anyone else) ever meant the abstract universal plus some conceptually undefinable and empirically unobservable added entity.'Abstract" and "concrete" are not two species of the common genus "universal," although the assumption that they are has vitiated almost all discussion of the issue and no doubt accounts for the fact that the issue has fallen into desuetude. But Collingwood never made this mistake. Speculum Mentis shows that he was aware from the beginning that the "concrete universal" is not a program for a new logic but is the leading idea of the historical consciousness; and as such it is the most important single strand which connects his earliest and last work. And it does so as one instance of its own meaning.

In Religion and Philosophy, Collingwood discussed and applied the idea of a whole whose structure is reduplicated in its parts, an idea which he regarded as identical with the so-called doctrine of "internal relations," i. e., the view that at least some entities are constituted entirely by their relations to other entities, as contrasted with the doctrine of "external relations" which holds that entities have some characteristics essentially, and independently of all other entities and their characteristics. So one finds Collingwood saying, "Every characteristic of the thing turns out to consist in a relation in which it stands to something else," and in a whole consisting of three parts, x, y, and z, "the inner nature of the part, x, then is entirely constituted by its relations to y and z.' Therefore, in such a whole, a part is simply one perspective of the whole: "the part is not added to other parts in order to make the whole, it is already in itself the whole, and the whole has other parts only in the sense that it can be looked at from other points of view, seen in other aspects. But in each aspect the whole is entirely present.' A number of examples are adduced to support this apparently paradoxical view: a musical duet, which is not (conceivably) the addition of two independent parts but a single entity constituted by the relation between the parts; or a dramatic scene as the interplay of two or more characters; or the identity of a single personality throughout its thoughts and acts.

Now duets and dramas—aesthetic objects in general—are often adduced as examples of complex wholes which cannot be successfully analyzed as aggregates of independent parts. And Collingwood seems to have confused this kind of whole with the more special kind of whole (for which duets and dramas do not serve as examples) whose plan is reduplicated in each of its proper parts. It is significant that there is no mention anywhere in Religion and Philosophy of the "concrete universal," nor any mention of the latter after Speculum Mentis. Collingwood's route of thinking, I suggest, must have been something like the following: in Religion and Philosophy he regarded himself as an adversary of a contemporary philosophy of science which claimed the merits of science as its own. This philosophy of science was mechanistic and materialistic and its consequence was the denial to the human being of any authentic freedom or independent self-existence. In the course of defending these characteristics of human personality, Collingwood found himself distinguishing between the notion of abstract identity (e. g., the sense in which we say that all men are equal before the law) and that of concrete unity (e. g., the sense in which men may form a Gemeinschaft in virtue of having common although not identical interests). There can be a concrete unity of two things which are part of the same whole, but ordinarily one would not say that such parts are identical with each other. Yet by a priori argument Collingwood concluded that if a whole is to be "strictly" knowable, its parts must be not simply added to one another but inter-connected in such a way that "each part is the whole.' Hence in any "genuine" whole unity and identity are the same thing. And this is stated as a general doctrine or philosophical principle, although in fact it seems to be true only of the specific case of personality. Formally, it is an ad hoc argument; yet it is not unilluminating. What we mean by "personality" is clearly not some indwelling entity but a pattern of behavior such that no matter how we analyze or divide human behavior—whether into simple response as on psychological tests or into complex events such as the carrying through of an expedition or a business venture—there are distinguishable characteristics (e. g., aggressiveness, caution, imaginativeness, reflectiveness) which describe not only momentary responses but complex actions over time and in fact the person himself. It is still no doubt much too strong to say that "each part is the whole," but at the same time it clearly will not do to regard a personality as merely the aggregate or average of a person's responses over time. There is an identifiable patterning, and to say that each part is the whole is a mildly misleading way of calling attention to the fact that there is no natural way of dividing human action into units or further unanalyzable parts. No matter how we divide a human career, we must divide it into parts which have the characteristics observable in the whole or in other parts. And it is not then too strange to say that "each part [is] also in a sense the others," remembering what that sense is.

But by the time he found himself adapting this argument to his purposes in Speculum Mentis, Collingwood had found a new way of generalizing it. The idea of a whole whose parts are "identical" with it and with each other now appears, not as a logical or metaphysical principle, generalized to refer to "any really organic whole," but as the special object of historical consciousness.

It is as if, in the interim, Collingwood had read Kant's "Idea of a Universal History" and had adapted to his own use Kant's notion of a "universal history" which would "connect into something like systematic unity the great abstract of human actions that else seem a chaotic and incoherent aggregate.' The idea of a whole "identical" with its proper parts is reasonably applicable to the unity of personality simply because the unity of personality is an historical unity. So by the time of Speculum Mentis Collingwood could, as it were, provide a gloss to Kant's reference to the "systematic unity" of history by suggesting that what makes it systematic is not anything like the logical relations of a mathematical or conceptual system nor the theoretical relations of natural science but the dialectical relation, already explored, of whole and part.

To call the relation "dialectical" does not, of course, either explain or justify it. Negatively, it is an indication that it is not to be understood either by analogy to mechanical models, which Collingwood never accepted, or by analogy to the biological model of "organic unity," which in Religion and Philosophy he did accept. The conception of the self in Speculum Mentis is not (as a mechanist would interpret it) that it is stratified like the many Troys, each layer built upon the one below but isolated from it so that the activities of each city are supported by the rubbler and shards of forgotten pasts. Nor is it (as an organicist would interpret it) a set of functions mutually related as ends and means, illustrated by the way in which the circulatory system returns to the digestive system in usable form the energy which the latter has converted. Rather the self is a reflexive process which takes into itself and retains, while it transforms, its own past experiences and activities. If a metaphor is wanted, its appropriate field is neither archaeology nor physiology but history, and its modality is not space but time. But the dialectic of experience is not really to be understood by metaphors, since metaphors are themselves extensions of its own meaning. It is not like anything more fundamental than itself; other things, less fundamental, are like it. The identification of any change as a process rather than as a sequence of states reflects the awareness by the self of itself as a process which incapsulates and transmutes its past in a way quite unlike an organism assimilating food or an avalanche taking into itself the objects in its path.

Whatever one thinks of the merits of the dialectic of whole and part, there can be no doubt that it comes very close to being the central idea the elucidation of which is the strand of continuity in the record of Collingwood's philosophical career.… [In] its various disguises, it solved some problems and started others. But since so far it has been stated very summarily and generally, it may not be out of place to observe that the application to history has an initial richness and plausibility which the generalized metaphysical version lacks. We cannot today think about history without some sort of essential, rather than merely chronological periodization.'Renaissance," "Enlightenment," "Feudalism," "Capitalism," even "Baroque," "Romantic," and other such designations are as indispensable as they are impossible to define precisely or to date sharply. An historical period is clearly neither merely the sum of its parts nor an entity which can be identified apart from them. It is, like national character or style in art, a partly but not completely analyzable pattern of complex form discernible in each of a wide range of instances. To say that "each part is the whole" means only that each represents the whole, that the Renaissance style is not Renaissance architecture plus Renaissance politics, and so on, but is wholly represented in any one of these (although we can see what is relevant in one only by comparing it with others), as all the elements of an artist's style can be identified in a single composition or painting (but only with other instances in mind).

3 THE DIALECTIC OF ETHICAL SYSTEMS

Now the little system of Speculum Mentis is itself self-exemplifying in this way. We have summarized Collingwood's discussion of Art, Religion, Science, History, and Philosophy as claims to knowledge or types of cognitive experience. But this is only part of what Collingwood means by "forms of consciousness"; they are also modes of action, or at least of the normative principles of action. So a practical ethics corresponds to each form of experience, and the connection between, say, Science as a mode of thought and Science as a mode of action is that each of these is a "part" in which the "whole" is fully exemplified. Thus they can be distinguished but not separated. And as each mode of thought is unstable and gives way to the next, or is true with respect to the mode it succeeds but is an error from the standpoint of its successor, so the corresponding systems of ethics go through a corresponding dialectical development.

To review this development from the beginning: Art is pure imagination, which makes no distinction between real and unreal and raises no questions about any relations among its self-contained objects. Now what sort of activity is analogous to this? It is hardly a guess: play, insofar as play is understood to be activity without ulterior ends and one which cannot be described as expedient or inexpedient, right or wrong. The question, of course, is not one of tactics or of rules. A move in chess may be expedient or inexpedient as a step toward check-mate, or legal or illegal according to the rules. But it is the activity of playing the game, including winning or losing, which is engaged in for its own sake. Chess as a livelihood, or as a way of making acquaintances, and so on, is of course not play at all. Nor should it now be any more difficult to identify the practical morality associated with Religion, remembering that this refers to the assertiveness of religious consciousness wholly without self-criticism: it could be nothing other than conventional morality, performances of all sorts which are chosen or avoided solely because they are or are not "done.' Such a morality of propriety is, like the play of Art, capricious: it neither has nor claims justification. But it is a development from play, as assertion is from supposai, and differs from it by being social; conventional morality is to individual caprice as the creeds of Religion are to the free imagination of Art. Collingwood wisely remarks that even though conventional morality can only assert and not defend itself, mere rebellion against it is not an advance but a lapse into individual capriciousness.

What does represent the development beyond conventional morality (and is in fact the step taken by conventional moralists when conventions deteriorate and must be shored up by argument, viz., "Honesty is the best policy") is the utilitarian ethics appropriate to Science, which abstracts from the class of intentional actions their common element of purposiveness and calls it "utility.' That utility is an abstraction can be seen if one remembers that the notion of a calculus is essential to utilitarianism; John Stuart Mill effectively abandoned utilitarianism in his defense of it by introducing the notion of qualitative, or incommensurable, differences among pleasures.

In general, as Science abstracts from the untidy concreteness of physical objects the measurable properties of mass, velocity, etc., so in the consideration of action it abstracts the concept of an action as such as well as the measurable properties of the effects of actions. This point is best brought out by the contrast between the abstract ethics of utility and the concrete ethics of duty appropriate to History; the latter ethics regards action not as a purely instrumental means to an end, with no value other than that derivative from the value of the end, but as something required in a specific situation, and needing no inherited justification from anything lying outside itself. But Collingwood argues that this account, like the others, will not suffice. An action is regarded as obligatory because it is demanded by the facts of the situation. Action may, however, alter the facts; and insofar as it does so, it is no answer to the question of how they should be altered that they are indeed thus and so. So the ethics of duty ("Historical ethics") is ambivalent; it says on the one hand that the will is autonomous, and on the other that the world of concrete fact lays upon it an obligation which is its duty. Hence the peculiarity of law, which is the embodiment of concrete ethics. Law achieves, unlike utilitarian ethics, the notion of responsibility, but enforces it from without; one obeys the law not because it is the law but because disobedience has predictable and painful consequences, and this attitude, needless to say, is a relapse into utilitarianism. Or again, the instability of "concrete ethics" is signified by the inescapable conflict between the claim of law and the claim of individual conscience. But this again (Collingwood claims) reveals a distinction between individual and society, which as an abstraction indicates that History has not yet fully emancipated itself from Science.

So as History is fulfilled only in Philosophy, the ethics of duty is fulfilled only in "absolute ethics," in which the distinction between individual and society disappears, and "the agent acts with full responsibility," the sense of compulsion by external law having disappeared. At this point Collingwood is at best programmatic and at worst filling out his a priori scheme. (The section on "Absolute Ethics" is the shortest in Speculum Mentis, less than two pages long.) He returned again and again to his classification of types of ethical theories, but in every later recapitulation the scheme culminates not with "Absolute Ethics" but with Duty. One must assume that in Speculum Mentis Collingwood felt that Hegel had spoken the last word; later he decided that Hegel had had one word too many.

4 THE UNRESOLVED PROBLEM OF SPECULUM MENTIS

The eloquence of Collingwood's eulogies to "absolute mind" and "absolute ethics" cannot conceal the fact that the stage of Philosophy, far from being the developed and comprehensive fulfillment of Art, Religion, Science, and History, seems, unlike these others, to have no positive content of its own; the great climax of the drama of development turns out to be nothing but the playwright stepping in front of the closed curtain to remind the audience of what it has already seen. But no transformation scene was advertised, and it is just Collingwood's point that Philosophy stands apart in having no special object or method of its own. It is in fact part of his objection to "aesthetic philosophy," "religious philosophy," "scientific philosophy," and "historical philosophy" that they commit the "error of conceiving philosophy as one specialized form of experience, instead of realizing that it is merely the self-consciousness of experience in general.'

The record of the development from Art through History is itself the dialectical growth in thought's consciousness of itself; so one might say that it is Philosophy which fills up all the interstices of the series—or, since it is a dynamic series, Philosophy is the uneasiness with which each moment of thought backs into the next with its eye fixed unwaveringly on the last. Speculum Mentis, that is to say, is a philosophical book, and as such it must exemplify throughout what it may or may not illuminatingly describe in the final section as its own procedure. And this procedure, all things considered, is clearer and more cogent than the concluding description of Philosophy as the last term of the five-part series. It cannot be correct to say that Philosophy has no object: the whole of Speculum Mentis is a demonstration that the forms of consciousness, including Philosophy itself, and their relations to each other constitute the object of Philosophy. Nor can it be correct to say that Philosophy has no method: how are the features of a mode of experience identified and analyzed, for example, or the relation between one mode of experience and another discerned? The method of Philosophy at least must be the kind of thinking exemplified in the recapitulation of the series of modes of experience. Collingwood began Speculum Mentis with the "suspicion that a philosophy of this k i n d … is the only philosophy that can exist and that all other philosophies are included in it.' But by the end of the argument there is generated an uneasy and unresolved tension between two different ways of viewing philosophy: one is that philosophy has its own province, which includes those of other forms of consciousness but is distinguishable from them, and the other is that philosophy is only the reflective self-awareness of any mode of experience, the bringing to explicit (i. e., self-) consciousness of principles, criteria or presuppositions normally implicit in the thought and action of that mode of experience.

Now if one recognizes this unsolved problem as the major issue of Speculum Mentis, it helps to account for many of the otherwise extraordinary passages in that book. They are neither uncontrolled flights of speculative imagination nor unabashed special pleading for the philosophical party of absolute idealism but attempts to elucidate the peculiar logic of a theory which is an instance of itself, to achieve a systematic comprehension of philosophy as it is described and philosophy as it is exemplified. But, more significantly, the whole series of Collingwood's books then falls into place as a continuing attempt to answer the unresolved question of Speculum Mentis.

To put this development in the briefest way, the Essay on Philosophical Method ignores the other forms of experience to elucidate a method claimed to be the special province of philosophy; it fulfills, or attempts to fulfill, what in Speculum Mentis is left as programatic, and provides a conceptual analysis of the concept of philosophy which in the earlier book is poetically expressed. The succeeding books, The Idea of Nature, The Idea of History, and The Principles of Art, are analyses of other forms of experience from the standpoint of philosophical reflection on them. In the Essay on Philosophical Method, Collingwood discusses philosophy but does not permit himself to think about the way in which his own discussion exemplifies it; in the later applications, he exemplifies it but does not directly discuss it. Yet while attacking either half of the problem, the awareness of the other half persists. It was not until the Essay on Metaphysics that Collingwood attempted a synthesis which would illustrate that coin-cidentia oppositorum which, according to Speculum Mentis, is the living nature of thought itself.

The uneasy tension in Speculum Mentis between the claim of philosophy to be an independent (and the highest) form of experience and the conception of philosophy as the reflective self-consciousness of other forms of experience is itself a dialectical tension. The development of the former in abstraction from the latter results in the dialectic of concepts which is the subject of the Essay on Philosophical Method; the development of the latter in abstraction from the former results in the dialectic of mind which Collingwood had partially worked out even before the Essay on Philosophical Method but did not publish in explicit form (in The New Leviathan) until shortly before his death. The clue to understanding each is to see that it replicates the other, just as, in Collingwood's view, the "concrete universal" is a whole whose parts replicate the form of the whole and of each other; the later developments are intelligible only when they are seen to have a common dialectical form. The rather traditional defense in Speculum Mentis of the "concrete universal" is transformed by Collingwood into his theory of concepts as comprising a scale of forms … and into his theory of mind as constituted by levels of consciousness related to each other like the basic types of experience and concepts on a scale of forms.… Collingwood's ultimate dialectic of mind grows out of the dialectic of concepts, as that grows out of the dialectic of experience.