9 Misconceptions about Classification (Rule 3)The problem of the relation of classifications to the situation is a difficult one, and thinkers have often gone wrong on it. In fact, the theory of "sets" or classes is still a very busy field in mathematics and logic. What we say here about classification and constructions in general must perforce be elementary and partial. We are interested in fallacy: in this connection two main misconceptions of the function of classification have led to broad errors in argument. These errors are related to the philosophical systems of Platonism and relativism, but we shall try to characterize them without a long digression into the history of philosophy.
Classifications are expounded in general terms, the so-called universals. This means only that when speakers use general terms in the language, such as "man," "woman," "child," they are classifying objects in the world according to certain traits or referring to the class of such objects itself.
1. Peter is a child.
2. The child is the father of the man.In (1) the speaker says that Peter belongs to the class of the child (children). In (2) the speaker says something in general (hence "general terms") about children, as well as about fathers and men. General terms like "child" mean certain constructions, and the constructions in turn obviously stand in some way for the situation. All children have a great number of traits in common. The situation in the world of human beings seems to impose a construction for "child," so that a language which lacked a term for the child-construction (that is, one can imagine, had only a word for person, which provided no guide to age or maturity) would certainly seem to lack a much needed term. Perhaps it would be correct to assume that the construction child is so much needed that every language has one or more names to mean it. In this common-sense way of talking about general terms, it will be seen that there are two notions at least that will stand examination. What can it mean to say that the situation in the world "imposes" a construction? What can it mean to say that a term is "needed" for a construction? Reification and relativism give very different answers to these questions, and each can lead to serious error.
In connection with Rule 3 (that classifications must be adequate for the purpose for which they are designed), we will argue that these errors arise from a misunderstanding of the role of purpose. Reification holds that the structures are imposed by a suprasensible world, and that the only purpose that the classifier should have is to learn to know them as they are. Relativism holds that there are no structures in the world or anywhere else; there are just totally discrete particulars. Hence in relativism no limits are set to the activity of the classifier, which is anyhow a mere exercise in fiction. We shall take up these misconceptions in turn in the form in which they result, for argument, as fallacies.
In the Phaedrus Plato says that classifications should not proceed, as a bad carver might, by cutting up the situation just anyhow, but should break it at the joints." The assumption is that there are natural structures in the situation which it is one's duty to discover. These structures are closer to reality than other characteristics, which are mere accidents due to the recalcitrance of the material in taking shape in imitation of the idea. The idea is in a heaven, and it is "ideal," that is, perfect, changeless, timeless. The things of this palpable world try to imitate the ideas, with indifferent success. The ideas, the eternal structures, are after all free from the gross embodiment of sense experience, in their pristine perfection. What you have got to do, unless you want to hack away as a bad carver, is to discover the natural structures, those closest to the ideas. Then your classification will produce neat joints and chops.
This view has the common-sense merit of recognizing that constructions stand for characteristics of the situation. It has the disadvantage of a great unprovable assumption, namely, that the characteristics referred to by the "natural" constructions are somehow more real than the other characteristics present. The preferred status of some structures over others is that the Platonists prefer certain purposes over others in classifying. These purposes turn out, on analysis, to be precisely those purposes imbedded in a culture and so traditional in their view of the world that they are felt to be natural and get built into the language. They have names in the common speech, such obvious names as "man," "redness," "goodness." Language seems to render this Platonism a natural thing to believe, and it is not surprising that it makes a strong appeal. This is not only a view that the division between, say, child and man is a natural division; it goes deeper. The term "redness," for instance, since it exists in the language, would seem to be a term for something, to name something, just as "the New York Central Railway" names a very complicated and ramified system. It does not seem farfetched to assume that, since redness has a name, it must exist. "Red is a primary color," is a perfectly legitimate expression in the language but one implicitly making an astonishing claim. Does this not seem to say that there exists a class of colors, rather like the class of children, and that red is a member of this class? Not this book or this country schoolhouse, of course, but redness itself, the redness that makes red the surfaces of this cheek or this rose. The inventor of language was a Platonist. Modern Platonists adhere to what they call the "perennial philosophy," and they acknowledge this philosophy has a "natural piety toward language."
There is more to Platonism than advice to eschew new-fangled terminology and modernistic or otherwise reprehensible ways of looking at things, that is, of making constructions. The senses can never give knowledge of redness, only of this red object or of that -- and even the best example is not perfect. The senses, likewise, cannot give knowledge of child, only of Peter or Paul or Hazel. But these particulars can grow old and die, and knowledge must be an ideal ageless perfection. There is no point in a feeble copy. However clear the photograph, one would prefer to see the person. However typical the person, one should prefer to know the Idea. Peter or Paul or Hazel are imperfect copies of the ideal "child." Thus, in Platonism the ideal child and redness are conceived of as existing purely, of floating about, as it were, ready to enter into particular bodies or surfaces.
This reification, thing-ization, of preferred constructions is a fallacy when it leads to mistakes in argument. Philosophers may do as they choose in deciding what things are "real" and what are not. But when they carry over their reifications into other fields, then they are asking for unnecessary trouble. The history of science, of affairs, even of mathematics has been troubled for generations uncounted by demands for "real definitions," assertions that, something is because it must be, denegations of whole areas of inquiry as unreal and wicked.
Reification, then, is the "hypostatizing" of entities, that is, the making of abstractions into substances. It is the assumption that child exists (in a Platonic heaven) over and above Peter, Hazel, Ethel. It is not necessary to make this assumption even if language suggests it. Universals, such as "redness" and "child," can be interpreted in a way that makes no philosophical claims as to whether or not the constructions they name refer to entities in the real world. This way of handling them is called Reduction.
Reduction is a rule asserting that any sentence containing a general term, if the general term designates an intelligible construction, can be reduced to a sum of sentences containing mention only of particulars. Abstractions ("redness," "justice") are included under general terms. Thus, A reduces to a1 or a2 or a3 . . . or a1, where A is in a sentence containing the general term, and a2, a3sub>, etc., are sentences referring only to particular instances. Thus the sentence, "A child needs security," as far as the child-construction goes, reduces to "Peter needs security, or Hazel needs security, or Ethel needs security," and so for all children. "Red is a color," similarly reduces to a mention of red books, red balloons, and other objects. The word "color" in "Red is a color" points out that whenever an object is red it is colored, that is, belongs to the larger class of all red, green, blue books and all other colored objects. This may seem a strange kind of "reduction," involving, as it does, expansion to unlimited numbers of sentences -- in the case of "child" one, in fact, for every individual young human being that now is, ever was, or ever will be. The "reduction" consists in the boiling down of general terms to particulars. The difficulty presented by the great number of sentences is only practical. Theoretically, it is soon clear how the process works out for the sentences in question, and since the series are fairly orderly, a sort of induction can be applied to show the reduced content of the general terms. (In point of fact, it seems that children actually learn the use of general terms in much this way.)
Consider "Peter is a child." This begins immediately to give the extension of the class by naming one member. Reduction of "child" in this case begins by simply designating a class member by its proper name. When the extension of the class is evident, the general term can be eliminated another way: it is now possible to single out essential characteristics which Peter shares with Hazel and Ethel and other children and then to say that Peter has these characteristics, thus giving the necessary and sufficient conditions for verifying the sentence "Peter is a child": Peter is human and aged y years, where y = a number under 12.
What reifiers do is resist the rule of Reduction. For them the question "What is a child?" is not a question of how "child" is used in the language, or rather, it is a question that asks for more information, namely, what a child really is. In addition to Hazel, Ethel, Peter, and all the children, for the reifier there is another entity in the world, child which Peter, et al., in their different and imperfect ways partake of, share in, embody. Similarly, with colors, there is this red object, and that, and moreover redness, the idea which enters into teacher's pencil, Mary's blush, the Russian flag. Redness and child exist together in heaven, along with Justice and bed (Plato's most famous examples).
In these simple cases reification is not apt to lead to confusions of the sort "Is redness red?" or "Is childness childish?" At least in the arguments about ordinary affairs, it more often leads to pseudo-philosophical meanderings about whether justice is an absolute or relative good, whether art can be for art's sake, whether a savage dog which is not man's pal is really a dog after all.
Suppose someone to argue, "The true friend has never yet been found. Damon was not a real friend of Pythias, David of Jonathan, Johnson of Boswell. Real friendship implies absolute equality, an equal sharing -- a having of 'all things in common.' Otherwise, one will have the advantage of the other. And since there can be no absolute equality in all things, even among identical twins, there can be no real friendship in the world. Perhaps there can be friends among the angels, but certainly not among men."
How can such an argument be met? It is set up in such a fashion that no evidence can possibly refute it. The speaker hypostatizes an entity, apparently in heaven among the angels, of true friendship, which involves (a saying of the ancient Greeks) "having all things in common." Then he demonstrates that this kind of sharing is impossible in the sense of an equal sharing, which is certainly one sense of sharing. By what is apparently a heavenly intuition, the speaker knows that true friendship implies equal sharing. Since such equality is impossible in the nature of things, therefore, true friendship is illusory.
The only possible answer is to point out to the speaker that the way one learns how to use the term "friendship" is precisely by observing how it is applied in the cases of Damon and Pythias, and other friends. This is obviously a high-level abstraction, involving references to many kinds of behavior -- speaking well of each other, being helpful, liking to be together, having mutual trust and confidence. But it can in principle be reduced to particular events under each kind. What would the speaker say if one pointed this out to him? Presumably that he knew all this sort of thing, that his point was precisely that all this was not enough to constitute friendship. There is no way to answer such an argument. As long as the speaker persists in the habit of making a thing out of an abstraction and of saying that some other thing is clearly not that thing, he is at once so right and so wrong that argument is impossible. Obviously nothing on earth is identical with anything else on earth, to say nothing of something in heaven which is at once immutable, atemporal, aspacial, perfect.
The firm way to deal with suspicious abstractions in argument is to demand their reduction to particulars. If the opponent cannot or will not show that such a reduction is possible in principle, then further discussion is pointless, since no evidence or counterargument is relevant. One finds oneself involved in an unruly discussion. But if reduction is admitted, at least in theory, then it is possible in principle to test the argument. It may be possible to find a counter instance, to make experimental application, to verify or falsify.
The examples are designed to show some chief varieties of reification.
EXAMPLE COMMENT Case history of patient J.S. Young man, 28, complained to his pastor that his wife, in-laws, and associates at work were all "persecuting" him. Pastor sought the help of clinic, reported to clinic that there seemed to be the following grounds for the young man's complaint: Investigation showed J.S. aggressive, accusing those around him of plotting against him. These persons retaliated by avoiding his company, talking about him behind his back, and generally behaving badly towards him. His wife, especially, resented his jealousy and finally left him. What caused the pastor to seek the help of the clinic was the fact that J.S. refused to see how he was responsible for the way others behaved towards him. He maintained that everybody he knew had been "got to by a secret power." His wife and his friends were in fact, he claimed, the agents of that power. At his last interview with his pastor, J.S. threatened the pastor with a knife, claiming that the pastor was also an agent, trying to win his confidence in order to betray him. Diagnosis. Paranoia. Patient's social relationships are disturbed by his self-destructive tendencies. These he projects into an outside independent agency. Recommendation. Confinement and extended therapy. J.S. actually was being persecuted in a way. His own aggressive behavior and distrust had alienated his wife, his relatives, and his associates. They avoided him and "talked about him." But notice that J.S. does not mean by "persecution" a construction that reduces to the sum of any particular occurrences in the behavior of others towards him. Instead, he sees all this as being inspired by a "secret power." He feels persecuted by a power, a kind of Platonic entity which is the cause of all the particular occurrences of persecution. The point is that there is a class of events having the common property of acts of persecution; the reification consists in making this property into an independent entity in the world, lying behind the acts and mysteriously causing them. It can be easily seen in this case that such reification is insane. It may be harder to detect the insanity in more ordinary cases. Commissar Petrov writes in Pravda, "Loans to smaller nations, subsidies, bilateral trade agreements, mutual defense pacts with the inevitable concomitant of air bases and arms supplies -- all of these have an obvious policy behind them of forming aggressive alliances against the Peoples' Democracies. This is not even in effect denied, for the United States admits that its basic policy in the Cold War is 'defense' against communism. But what is denied is that all these activities reveal another policy, a deeper and more underlying strategy. This is the indispensable policy, in fact, of all monopolistic capitalistic nations: colonialism. All these activities of the United States carry out the capitalist necessity of imperialistic colonizing of so-called backward areas. The frank colonialism of the last century is no longer possible: the empires of Great Britain, France, the United States can not stand anywhere against the political aspirations of downtrodden peoples. So, as the powers are forced to yield political empire, with the other hand they impose an economic-military empire on the same peoples. Thus the activities of the United States serve two purposes: they attempt an aggressive encirclement of the Peoples' Democracies, and at the same time they enslave the backward areas in a new-style colonialism." This argument, like most of the arguments of dogmatic Marx-Leninism, bases its proof on a reified entity. Starting with the common characteristic underlying the diverse activities of U.S. foreign policy, defense against communism (which it interprets as aggressive alliances against the "Peoples' Democracies"), the proof then alleges another common characteristic, colonialism. It is certainly possible that the United States could be doing two things at once. But what is the evidence for this second purpose? Pravda cites only contradictory evidence, the break-down of empire (apparently referring to the new-found independence of India, the Philippines, etc.), and alleges that this has led to a "new-style colonialism." There is a curious sentence about the "indispensable policy" of "all monopolistic capitalistic nations" being colonialism. And this is then referred to as a "necessity." The reader of this article is now in the presence of an Entity. Like the secret power of the previous example, this construction of an indispensable policy cannot be reduced to any supporting evidence, nor refuted by any contradictory evidence. The writer, as a Marx-Leninist, is speaking of a "truth" more basic than any mere accidents embodied in historical events. If pressed to explain, he would presumably cite other and still more fundamental entities -- the "logic" or determinism of historical process, the dialectic of the class struggle. A rule attributed to the nominalist William of Occam warns that "entities are not to be multiplied beyond necessity." This means that you do not keep on appealing to entities, such as colonialism or the secret power, to explain what is already sufficiently understood. As such entities go beyond all possible evidence, they are unarguable. What, if not evidence, can possibly support or refute them? King Aroo overhears one character reproving another for throwing away his future. Now King Aroo had just learned that his own future was nothing but a smear of tea leaves, so he begins looking through ash cans for the future that got thrown away. (This is from Jack Kent's famous comic strip for March 6, 1957.) King Aroo's search is funny because the reification is so concrete that it might be something found iu the ash can. Superstitious people seem to use the term "Future" as if it named a sort of country where events are already taking place which in time will be reproduced on this earth. These events can be dimly seen in crystal balls, tea leaves, cards, palms of hands. All you have to do is wait for them to arrive, like a letter already in the mail. "The State is the divine idea as it exists today . . . It is the absolute power on earth; it is its own end and object. It is the ultimate end which has the highest right against the individual." Hegel. King Aroo deserves a smile. Not so Hegel. His is a reification which, unfortunately, has millions of followers. The "State" its own end? What mystic thing is a state that can be an end apart from the welfare of individual people? "No, one thousand times, No! The supreme end is the individual, and collective institutions should have no more hold over him than is needed for his own individual development." Salvador de Madariaga. (Quotations from Hegel and de Madariaga are taken from Sontag and Beddie, Nazi-Soviet Relations. New York: Didier, 1948, p. VIII.)
Personification (see #29) is a mark of reifying. Amateur psychologists often speak of the Ego or the Id as alternative souls presiding by turns inside the head (like the "ghost in the machine" derided by the philosopher Gilbert Ryle. Amateur strategists say things such as, "Strategy dictated the policy of island hopping." If this means only that strategical considerations led to the plan of island hopping, there can be no objection -- economy "dictates" the use of personification in numerous more or less literary sentences (including some in this book). But if the speaker resists such a reduction of strategy, then a new thing has been born into the world. If one wants to waive aside such military considerations as subordinating operational tactics to over-all designs -- there are these considerations, of course, but also something else much bigger, namely, Strategy, or even Grand Strategy, which not only controls the plans but also, somehow, the planners, who are its instruments -- then shake hands with Strategy and introduce him to Id and Ego.
Poets can get very different effects from that of King Aroo by the use of similar devices. Jean Cocteau depicts a character in a diligent search, not for his own future, but for his own death, as if there were a special (not just a different) death for everybody. In his Orpheus the same poet personifies Death as a woman surgeon. Other writers speak of a character's life as a mystical Gestalt -- a sort of space in which the drama of his fate is worked out: the world is a stage and human beings are puppets moved by destiny. The Greeks personified the fates as the Eumenides, and T. S. Eliot depicts the Eumenides howling outside the window in Family Reunion. Such devices are of the utmost importance in literature, where they effect a dramatic economy. It is philistinism to complain of reification in poetry -- there is no harm in the device, provided it is understood as a play of imagination.
Relativism stands at the opposite extreme from reification. If you want to render an account of a situation, the Platonists say that you must select for your classifications only the essential characteristics, the essential being those that make the situation what it is because they reflect the eternal reality. For the relativists, no characteristics are essential or even intrinsic. None are more "real" than any others. The relativists are impressed with the obvious fact that situations can be classified in any number of ways -- thus the criticism of classifications must center on the purpose of the classifier. Since his freedom in making constructions is without limit, but since he does in fact choose one set of constructions over the others, why does he choose as he does? Well, he chooses to suit his purpose. The relativists add that since there is nothing in the situation to fix his purpose, it must be given by something extrinsic. This extrinsic, purpose will be found in the preoccupations of the classifier, his preconceptions, the peculiar concerns of his times, the unique interests of his class.
Relativists are thus poles apart from Platonists: there are no real characteristics in the situation, as the reifiers think, to "determine the constructions. The model of the nominalist or relativist explanation is the famous case of the ass that was placed exactly equidistant between two bundles of hay. Only the relativist ass does not starve to death in indecision, but makes a choice. This choice is not determined by anything in one bundle of hay to make him prefer it to another -- that would be an intrinsic consideration. No, his choice is some extrinsic matter, such as the habit of always turning left to his dinner. Of course, there are more than two bundles of hay, for there is an indefinite number of possible choices in classifying situations.
Relativism bids us consider the case of the historian who sits down to write an account of, say, a battle. Thousands of objects are involved -- men, guns, weapons. Thousands of events transpire -- melees, charges, cries and deaths. The historian cannot know all of these. Moreover, he already knows more than he can use, in most cases. What does he do? He talks generally about the movement of wings and centers, about the breaking and holdings of fronts, about advances and retreats. He selects and arranges the materials and in so doing imposes a structure on what is essentially structureless., In short, as he cannot know all the facts and cannot use even all the facts he knows, the historian's structure is an invention of his own. It will be determined, not by the data, but by his own values, prejudices, needs, social or political commitments. If this is true of the account of a contained story, such as of a single battle, how much more so must it be of large "wholes," such as the history of the Punic Wars or, for that matter, of the Roman Empire.
The historian does not tell the truth about the past; he does not even tell a lie. (This assumes he doesn't fabricate evidence or conceal facts.) For relativism there is no truth beyond the particulars, that is, the actual documents and other data. There is only fiction. One account will be as "true" as another, for all accounts are relative (hence "relativism") to the narrow intellectual and moral concerns of the writer. As these differ characteristically from age to age, it will be necessary for each generation to rewrite history for itself. .Charles Beard speaks of the views of the historian as being "arbitrarily established." F. J. Teggert says that an historian may seem to be presenting a picture of some distant time but that he is perforce speaking with the "voice of his own generation," and giving tongue "to the ideas and aspirations of his own community."
Relativistic theories have serious implications beyond the technical problems of writing and evaluating history. The Roman Empire is obviously a very large, multiform "situation," but other persons besides historians often must try to understand situations that are quite manifold and complex enough. It may well be that no constructural machinery can be devised that will adequately classify the events of the Roman Empire. But it does not follow from this that all situations are impervious to understanding. Even historical situations are not always vast and unwieldy, and relativists could hardly maintain that there is a difference in essential kind between historical and other events.
It is not so much the size and complexity of a situation that makes relativists despondent. It is the predicament of theorizing in gen-era],. They believe, in principle, that since the account of any situation involves classifying, that is, selection and arrangement, the account is not of the situation but of the bias of the classifier. What you get from an account of the battle is not an understanding of the battle, but an insight into the historian. So expressed, relativism seems a form of the fallacy of origin (see #19); that is, it invites you to pay no attention to the truth of what is said but only to the source of the remarks. But more is involved. It is a theory that, since the selection and arrangement are never in the situation but only in the mind, and since an account is simply the presentation of the selection and arrangement, there can be no understanding of the situation. All that there is to understand is a construction of the mind.
Of all the possible selections and arrangements, the man making the account chooses that particular way which best implements his purpose. In this sense the account is strictly relative to the purpose, and it is good to call attention to this relativism, for an understanding of the account must involve the concerns of the man making it. Our way of saying this is given by Rule 3. The classification is, we say, in fact determined by the purposes of thn classifier. But does it follow that his account is not of the situation? The account is in fact limited by the situation. There is no reason to believe that the account of large situations (the Punic Wars) is different in anything except complexity from the small situation (say, the hundred marbles). Perhaps despondency over very ambitious history-writing is justified. But even Gibbon and Toynbee are bound by limits, and these limits are clear in the data. Aside from the obvious negative limits (Hannibal lacked H-bombs in his attacks on Rome), there are recalcitrant facts that also must be taken into all accounts (Hannibal failed to break Roman resistance). Historical and other vast situations present the characteristics they have and no others, just as do all situations. These characteristics are in principle classifiable, and the account can then be evaluated according to the well understood rules of classification. Evaluation of the purpose behind the account is an important matter, but it is a different matter.
EXAMPLE COMMENT Peter says to his wife, "Why are you reading those book reviews? Criticism is a farce. One critic likes what another critic cannot stand. Jones in the Tribune says that a book illustrates courage under adversity and that the characters are finely drawn and true to life. Smith in the Gazette says that the theme of the same book is that virtue is its own punishment and that the characters are puppets manipulated by a clumsy writer. Which is right? Neither! Each critic finds in the novel what he is looking for, and everything in the book proves anything you care to say about it. Why? Because there is no novel except in the reader's head, and what the reader likes and doesn't like depends entirely on his taste, This is why the old saying 'There is no disputing about taste' is so wise. What would two critics be disputing about? Two separate novels, one in the head of each. If it is silly to dispute about 'fictions,' it is silly to write or read criticism." In a sense, there is a novel in each reader's mind, but it hardly follows that these cannot be very similar in salient respects. The events in a novel can certainly be classified in various ways, and the choice of ways will be determined by the purpose of the critic. This purpose, in turn, may be causally related to his taste. Similarly, the characters in fiction do and say things that are subject to various interpretations, just as people behave in life. But we do in fact understand each other. This is to say, our account of each other's character makes it possible to predict behavior, to get along together, to cooperate. In a similar sense, good critics give insight into the characters in a novel and elucidate the action. There is nothing mysterious about this. If the purpose is to find hidden motives, then psycho-analvtical constructions throw light on the characters of a novel as of life. If the purpose is to show how the behavior is socially or economically directed, then the speech and actions will be ordered in a different way. A good novel is a rich and complex texture, and it will support many interpretations. Student report: "I know a lot of historians think that President Jackson's quarrel with the Bank was a misfortune for the country. This merely shows how they feel about banks and about Jackson. As for me, I happen to feel different about things. I happen to admire Jackson. The historians have a right to their opinion, and I have a right to mine. After all, no one can say which of us is objectively right, since there's no way to rim the nineteenth century over again without Jackson." "Objectively" here suggests the model of a laboratory where experiments can be "run over again." There are clearly other senses of "objectively," such as "in conformity with the evidence." Jurors cannot run a murder over again to see if the accused did in fact kill the victim, but they can nevertheless bring in an objective verdict,. In this model, unlike that of the laboratory, objectivity in a verdict ("opinion") implies a verdict in line with the evidence and free from the "subjective" elements of personal passion or bias. This is all rather obvious. What is important to note about the example is the widespread supposition that "everybody is entitled to his opinion." This is simply absurd in the kind of case under consideration. The student is not entitled to an opinion counter to informed historical judgment, just as a juror is not entitled to an opinion counter to the evidence. If the present book can be said to have a general thesis, it is that nobody is entitled to an irresponsible opinion on anything.