Curt Ducasse, Philosophy as a Science, 1941


Philosophical Experience and Philosophical Theory

ADMITTEDLY, there has been no dearth of theorizing in philosophy, but too often the theories have been presented as if they were revelations from God to philosophers rather than speculations to be tested. In consequence, philosophical theorizing has been largely barren of the fruits that speculation, recognized as such and accordingly subjected to appropriate controls, has proved itself capable of yielding in other fields. This state of affairs, I believe, is chiefly owing to the fact that philosophical theorizing has in the vast majority of cases been carried on with little or no clear consciousness of the structure of explanatory theories in general, or of the empirical generalizations that philosophical theory has to explain, or of the nature of the primitive facts by reference to which the validity of philosophical theories is to be tested, or of the sorts of empirical tests the theories must meet.

Part II of the present work seeks to clarify these very questions, and it is for this reason that the attempt has been made (chap X) to point out the sort of facts that are primitive for philosophy, and (chap. XI) to describe the general nature of explanatory theories and of their data, and the criteria for choice between rival theories. In the present chapter, the attempt will be made to elucidate the questions remaining for consideration among these, and to make evident as a result that theorizing has in philosophy the same nature and functions as elsewhere, and that its validity is here also susceptible of being empirically tested.

1. Approach to the Problem of Theory through the Example of Ethics. -- For the sake of the lucidity that goes with concrete examples, the discussion will be worded first in terms of the field of ethics. It will not, however, present any particular ethical theory, but only a theory as to the nature of ethical theory in general -- that is, to use a terminology now becoming familiar, an ethical metatheory. Afterwards the considerations adduced will be shown to apply also, mutatis mutandis, to the other branches of philosophy and therefore to define a theory of the nature and method of philosophy in general -- that is, a philosophical metatheory.

2. What Facts Are Primitive for Ethics? -- The facts primitive for a given science, we have already seen, are the sort of facts about which (directly or indirectly) are all the questions asked by the science, and by appeal to which also the theories constructed by the science are finally tested empirically. The facts primitive in this sense for philosophy, it was contended, consist of certain appraisals. It is now submitted that the facts primitive in the same sense specifically for ethics consist of some of the appraisals that would be expressed by such statements as: "This act is wrong"; "This man is evil"; "This ought to be done"; "This state of affairs is good," etc. These are appraisals of particulars and are all, no matter what the particular form, appraisals of the kinds called approval or disapproval. Those among them which constitute the primitive facts for ethics are the spontaneous or empirical ones, in the sense of these terms already stated -- that is, the ones which are not deduced from some ethical theory which the judging person was taught or constructed for himself, and which are not simply an imitation of other persons' approvals and disapprovals.

That some approvals and disapprovals are spontaneous in this sense is shown by the obvious fact that, for instance, neither teaching nor the holding of any ethical theory is needed by a mother to disapprove of an attempt to harm her infant, or by a man to disapprove of his house being set on fire or to approve of being helped to put out the fire, or indeed by a dog to disapprove of another dog's attempt to take his bone away from him. Moreover, unless some approvals and disapprovals were spontaneous in the sense stared, no categorical but only hypothetical approvals or disapprovals could even be derivative.

3. Empirical Ethical Generalizations and Ethical Norms. -- By comparing the concretely given individual subjects of certain sets of spontaneous approvals or disapprovals, it is possible to arrive inductively at certain ethical generalizations, which are therefore also spontaneous or empirical in the sense of being neither derived from any ethical theory nor borrowed from other persons. For example, a number of acts, each of which we spontaneously disapprove, may resemble one another in being all of them cases of our having had the fruits of our labor taken from us without our consent. We may conclude inductively with the empirical ethical generalization that we disapprove of any act of this kind -- a kind commonly called, perhaps, stealing or robbery or exploitation. The verbal expression of this generalized disapproval would be some such statement as "stealing from us is wrong," or perhaps, more sweepingly, "stealing is wrong." Empirical generalizations such as this are the analogues, in the field of ethics, of such empirical generalizations in the field of natural science, as that heat causes wax to melt, or that glass is brittle, etc.

Maxims such as "stealing from me is wrong," "stealing from anyone is wrong," etc., are ethical norms. If they are arrived at in the way just described -- directly by induction from spontaneous disapprovals or approvals -- they are empirical ethical norms. Ethical norms, however, are in some cases arrived at deductively from some ethical theory the validity of which has already been well tested. They are then rational (i.e., theoretically grounded) ethical norms.

4. What an Ethical Theory Explains and Predicts. -- Approvals and disapprovals are always somebody's approvals and disapprovals, but what an ethical theory attempts is not to point out what past events in a person's life are responsible for the fact that he now approves and disapproves what he does, nor to state what sorts of training, if he were subjected to them, would cause him to approve or disapprove something else. An inquiry of this sort -- into the genesis of an individual's present approvals and disapprovals or into the laws of such genesis -- would belong not to ethics but to the psychology of learning.

Again, ethical theorizing, like physical, does not attempt to show that the generalizations it concerns are valid -- for in either case this is assumed already on the basis of inductions from particulars -- but only to show why they are valid and exactly how far their scope extends. This is done by discovering premises from which these empirically discovered generalizations could have been deduced, and from which others empirically confirmable can be deduced.{1}

What we must now specially notice, however -- for if we do not, hopeless confusion ensues -- is that the problem it is attempted to solve by a theory of, for instance, right and wrong, is never, "What do the words right and wrong, considered as it were in vacuo mean?" (for, simply by a verbal convention, they can be made to mean anything one chooses) but always, "What do these words mean, that is, imply, as applied to this, that, and that, or as applied to all cases of kinds K, L, M by person P or by persons P, Q, R?" For only then is the question as to what the words right and wrong mean not one of arbitrary definition but of real definition -- that is, of what the words mean as actually applied by person P, in contrast with what he may propose to mean or think he means by them. This point is of cardinal importance not only here but for philosophical method in general, and will be considered subsequently in greater detail (chap. XIV).

It is then obvious that the implication of a term of approval or disapproval as spontaneously applied, e.g., of the term wrong, can be investigated -- or a theory as to its implication tested empirically -- only if we specify some of the particular facts or classes of facts to which the term is spontaneously applied, for the validity or invalidity of a theory of the nature of wrongness is always strictly relative to the spontaneous applications of the term which are to be explained. But, as we shall see before long, we must specify not only which applications of the term the theory attempts to explain, but also whose spontaneous applications they are -- whether those of a given person, or those on which all the persons of a given group agree, or those on which everybody agrees. The need for this is usually overlooked, with confusion of issues as a result.

5. The Two Parts of an Ethical Theory and the Empirical Tests of Its Validity. -- Reduced to essentials, ethical theorizing, like physical or any other, consists in attempting, on the one hand, to invent a definition of the predicate of the generalizations to be explained, and on the other to specify a method for determining whether or not a given concrete case or class of cases is to be accepted as meeting the definition -- the definition and method together satisfying certain requirements soon to be described. In the physical example of wax, the generalization was that wax has the property of melting when heated, and the predicates to be defined were "heated" and "melting." In the case of an ethical theory, the generalizations concerned might be, for instance, that killing, adultery, and stealing are wrong. Let us note, however, that if, as we are assuming, these generalizations formulate inductions from a number of spontaneous disapprovals by a given person or group of persons, then these generalizations state properties of the person (or persons) P. That is, they are inductions that person P is such that whenever he perceives or imagines killing, adultery, or stealing, he disapproves it (the word wrong being used to state the disapproval).

The definition sought of the predicate wrong is then one satisfying, first, the requirement that if its definiens D is substituted for that predicate in the given empirical generalizations, the resulting statements will be true. That is, if the proposed definition of wrong is that to be wrong is to have character D, this definition must be such that killing, adultery, and stealing do have character D. To determine whether or not they do have it, we have to apply the method (accompanying the proposed definition) for determining whether or not character D is present in a given case or class of cases. The statement of the method will have the general form: A given case (or any case of a given kind) shall be admitted as having character D if and only if, when it is subjected to a test of kind T, the outcome of the test is perceived to be O. Specification of such a test is necessary because the cases in which theorizing (instead of direct observation) has to be resorted to are those in which the subjects of the several generalizations present themselves to perception as heterogeneous; that is, a respect of homogeneity D in such cases can perhaps be conceived but not directly perceived.

For example, the definiens of wrong proposed might be "forbidden by God"; and it is not ascertainable by direct perceptual observation of cases of killing, adultery, or stealing, whether or not these modes of action are forbidden by God. The proposed definition of wrong therefore has to be coupled with specification of a method by which to ascertain this indirectly. Such a method might be, "consult the Bible." Then the first of the empirical tests of adequacy of the theory (of the meaning of wrong as spontaneously applied by P), which that definition and this method together constitute, is that killing, adultery, and stealing, which were spontaneously disapproved by person P, should be found among modes of action described in the Bible as forbidden by God. If they are so found, the theory to this extent fits the facts for which it attempts to provide a conceptual explanation -- the facts, namely, consisting of spontaneous disapproval by person P of the three kinds of action mentioned.

The second of the tests of adequacy of the theory is that the character D be not possessed by given modes of action that P spontaneously approves. That is, the facts for which the theory attempts to furnish a conceptual explanation are not only that P spontaneously disapproves modes of conduct A, B, C, but also that P spontaneously approves modes of conduct E, F, G.

But the theory must in addition, like explanatory theories anywhere else, meet the requirement of successful predictiveness, this comprising the third and fourth empirical tests of its validity. Meeting this requirement will mean that modes of conduct H, I, J, which have character D, and modes of conduct K, L, M, which do not have character D -- but approval or disapproval of which by P was nor taken into consideration in the devising of D -- will turn out to be also modes of conduct that P respectively disapproves and approves spontaneously.

As already stated, P may be a given person, or a given group of persons, or even possibly everybody. But unless specification is made as to who P is -- that is, whose spontaneous disapprovals and approvals the statements "A, B, C are wrong" and "E, F, G are right" respectively express -- the predictions implicit in any theory of these disapprovals and approvals remain a priori insusceptible of being either verified or confuted, because these predictions are of additional spontaneous approvals or disapprovals by the same persons. The theory therefore remains insusceptible also of comparison in respect to predictiveness with any supposedly rival theory. Two ethical theories are rivals only when each attempts to explain conceptually the same given set of approvals and disapprovals by the same person or persons.{2}

6. Grounds for Choice between Rival Ethical Theories. -- Failure to specify -- or indeed to realize the need of specifying -- which disapprovals and approvals a given ethical theory attempts to explain conceptually, and equally whose disapprovals and approvals they are, is responsible for much of the disagreement to be found in the field of ethical theory; and much of the disagreement in other fields of philosophy has an analogous source. On the other hand, if these points -- viz., what exactly the theories are about -- are specified, one knows immediately whether the theories really are rivals or not.

The theories, moreover, then become amenable to empirical confirmation or confutation, as in other fields. If they are rivals, choice between them is possible on the same sorts of empirical grounds (described in chap. XI) as in the case of any other explanatory theories. Such real joining of issues as there has been in ethical theorizing has been owing to the fact that, as Rueff points out, the parties to the disputes have been -- actually even if tacitly -- pretty well agreed to begin with that certain modes of conduct were wrong and certain others right. The rules which describe these modes of conduct are, he writes, "the practical morals of the moment in which we live and for our particular group . . . . . They are the product of life."{3}

7. Can Ethical Theory Resolve Conflicts between Ethical Judgments? -- The occasions that give rise to ethical theorizing are conflicts between two or more of the ethical judgments -- the approvals or disapprovals -- of one person; or between the ethical judgments of one person and those of another; or hesitation between approval and disapproval in a given case (this being describable as conflict between inclination to approve and to disapprove). When a rational, i.e., systematic solution of such conflicts is desired, theorizing is automatically resorted to. But since an ethical theory is a conceptual explanation of given spontaneous ethical judgments, and two really conflicting judgments cannot logically both be vindicated by the same theory, it seems impossible that theory construction should resolve the conflicts that give rise to it. Each party to the conflict will take his own ethical judgments as the ones to be fitted by the theory; and the result will be two theories instead of one, and two theories which will not be rivals in the sense described above. That is, they will not be rival explanations of the same judgments, but the rivalry will be between the judgments themselves.

The actual situation, however, is seldom as hopeless as these reflections would suggest. Beside the judgment concerning which the parties disagree, there will usually be many other ethical judgments in which they agree; and it will be possible to construct a theory adequate to explain given ones among these and to predict successfully the remainder of them. To resolve the remaining disagreement rationally will then consist in deciding it on the same theory already admitted by both parties as adequate to explain and predict all the ethical judgments in which both agree. That is, a solution is rational if the judgment it proposes is one logically coherent with -- deducible from the same theory as -- the remainder of one's own judgments; and the possibility of resolving conflicts of ethical judgments by means of appeal to ethical theory thus postulates desire in each party for such integration or coherence among his own judgments. If this is not present, no rational but only an arbitrary solution is possible. An example will make clear that rational solution of conflicts between ethical judgments, through ethical theorizing, is a familiar and working process.

Suppose for instance that a number of persons all spontaneously disapprove of killing, adultery, and stealing, but that some of them countenance lying and some disapprove it. Suppose this disagreement leads them to seek a theory adequate to explain the ethical judgments mentioned as to which they all agree. And suppose they eventually settle upon, as adequate to do so, a theory whose constituent "definition" and "method" are:

The definition that any act of a given kind is spontaneously disapproved (the disapproval being formulated by calling the act wrong) means that acts of this kind generally bring about more human unhappiness than happiness; and, the method for deciding whether or nor a given kind of act is "wrong" according to this definition, that the disagreeing parties shall inquire together into the probable effects of acts of the given kind, and then take a vote as to whether the unhappiness probably involved in these effects is or is not greater than the happiness.{4}

Suppose further that according to this theory, gambling and drunkenness also would be wrong; and that it turns out that gambling and drunkenness actually are disapproved spontaneously by all the persons in our example. The theory is thereby shown to be capable also of true predictions. So far as they have tested it, the theory then fits all the spontaneous ethical judgments in which they all agree. It is natural for them to ask next whether, on this same theory, lying would be also wrong or on the contrary right. Let us suppose that they agree that it is a mode of conduct generally likely to cause more unhappiness than happiness, i.e., that, according to this same theory, lying would be wrong. Then obviously to approve it would be irrational, that is, theoretically incongruous with their disapproval of killing, adultery, stealing, gambling, drunkenness. And "invalidity" -- "unwarrantedness," "unjustifiability" -- of a given one of one's own spontaneous ethical judgments can mean nothing whatever except theoretical incongruity, in the sense just described, of the given judgment with the rest, or the majority, or the most confident, of one's own spontaneous ethical judgments.

A claim that a judgment thus theoretically incongruous is nevertheless rational and valid could be based only on the proffering of a rival theory, equally capable of explaining and predicting the other judgments but implying that lying is right. If such another theory were offered, then additional predictions implicit in each theory would have to be examined, and a rational choice between them would be a matter of which one predicts truly the greater number or the most confident of the spontaneous approvals and disapprovals in which the persons agree. If both theories predicted them equally, though one still implied that lying is wrong and the other that it is right, then no rational but only an arbitrary choice would be possible between the two theories. This, however, would be a very extraordinary situation, for it is usually very difficult to devise a theory that will imply everything another theory is found to imply, except for one specified thing as to which it has the contrary implication.

8. One Ethical Theory, or Several? -- The capacity of a given ethical theory to explain or predict the empirical generalizations with which it concerns itself can be tested by a person other than the one whose ethical judgments these generalizations represent, provided only that these judgments be made known to him. Indeed, if they are known to him, he is in just as good a position to construct a theory of them as is the person who made the judgments. For as we have seen, the validity of a theory of these judgments is not a matter of whether the person (or group) spontaneously making them believes that, by the given value-predicate, he or they mean what the theory says this predicate means. It is a matter of whether the account of its meaning offered by the theory stands the tests of validity described above (sec. 6) better than does any rival account. Scientific status for a theory requires that the theory be susceptible of confirmation or invalidation by outside critics, and ethical theories are thus susceptible of this.

But confirmation or disproof of the capacity of an ethical theory to integrate conceptually the spontaneous ethical judgments of a given person or group is after all only historical or maieutic criticism. It is internal criticism -- criticism ex concessis; and ordinarily the sort of criticism we make of ethical theories is not as detached as this. For example, if a consequence of a given ethical theory were that matricide is right, we should not ordinarily limit our critical examination of the theory to inquiring whether or nor the judgment that matricide is right turns out in fact to be an additional one of the spontaneous judgments of the person or group whose judgments the theory attempts to integrate conceptually. What we should ordinarily say is rather that since the theory would entail that matricide is right, whereas in fact it is wrong, the theory is invalid.

But the fact which is the basis of our own assertion, that in fact (vs. in theory) matricide is wrong, can be only that both we ourselves and the persons to whose decision we submit the justness of our criticism -- that is to say, all persons "who count" -- spontaneously disapprove of matricide. And criticism on this basis is still criticism ex concessis although the group whose spontaneous judgments we require the theory to fit is now a different and presumably a more inclusive or in some sense more "authoritative" one. Criticism of an ethical theory -- or in general, of a philosophical theory -- is always unavoidably thus ex concessis and therefore ad hominem because the facts with which the theory ultimately concerns itself are facts intrinsically ab homine, viz,, are appraisals by men. Ad hominem is here also ad rem.

Accordingly, there would seem to be only two theoretically possible ways to eliminate the plurality of theories (nonrival) which arises from the intrinsic relativity of the facts concerned to a variety of individuals or groups. One way would be to take, as the facts which ethical theory is really (vs. has historically been) required to fit, and with which the incongruous facts are to be aligned, the facts consisting of only the spontaneous ethical judgments in which everybody agrees. But it can be confidently asserted that there are no such facts.

The only remaining way to unity of theory -- which is a scientific desideratum -- is then to postulate that all persons would spontaneously appraise the same thing alike if they were "in the same position,"{5} and then (in the statement of the facts which the unitary theory is to fit), to specify the appraising person or group concerned not denotatively and therefore absolutely, but descriptively and therefore relationally, by an account of his or its "position" in the sense stated.

That is, the facts which the theory will have to account for will then be stated not in the form, "person or group makes appraisal A of entity E," but in the form, "any person or group in position P makes appraisal A of entity E." This (by postulation equivalent) way of formulating the same facts eliminates the conceptual irreconcilability which would appear between such facts as that "person P appraises entity E as bad" and "person Q appraises it as good," if the "positions" of P and of Q were assumed to be the same.

A theory capable of integrating conceptually all spontaneous appraisals will thus necessarily be a relativistic theory in a sense analogous to that in which the "relativity theory" in physics is relativist. It too, that is to say, will eliminate or at least abstract from the individual observer or judge by conceiving the judgment he renders as a function of his "position" as well as of the entity judged. But a unitary theory, precisely because of its necessarily relativistic character, will not imply that it is necessary for some persons to alter their appraisals for the sake of congruousness with the appraisals of other persons. For a given appraisal can be either really congruous or really incongruous with given others only if all of them are appraisals from the same "position"; and persons other than a given individual or group are always ex hypothesi at other "positions" at a given time and can only imperfectly imagine his or its "position." A unitary theory will not even imply that it is necessary for a person to alter some of his appraisals for the sake of congruity with the rest of his own appraisals, if they are appraisals made by him from significantly different "positions," or if the appraisals concerned are not functions solely of his "position" but in part also of an additional variable such as "taste" seems to be; for then the various appraisals cannot be said to be really either congruous or incongruous with one another.

9. Are Ethical Theories Prescriptive? -- Theory is in itself no more prescriptive in ethics than anywhere else, but only predictive. In the natural sciences, the predictions ultimately concern certain perceptually observable facts in nature; in ethics, certain spontaneous ethical judgments, viz., spontaneous approvals and disapprovals. But although there are spontaneous ethical judgments -- and the validity of any ethical theory is ultimately tested by reference to some of them -- a given spontaneous judgment may come to be repudiated as incongruous with the theory which adequately fits the rest of them, and replaced by its opposite. In such a case a fact which the theory does not fit is being altered to fit the theory. This is genuinely possible here, because the facts with which the theory concerns itself consist of ethical judgments -- approvals or disapprovals -- and there is such a thing as changing one's judgment. Nothing similar is possible in natural science, where the facts theorized about neither consist of nor are dependent upon human appraisal. To alter a specimen which conflicts with a given biological theory does not save the theory, for the theory in this case predicts not a judgment by man, which he can alter if he wills, but something that nature under specified circumstances produces, no matter what man wishes may be, and no matter what man may later do to the thing so produced.

But it is essential to note that a theory rationally explaining and predicting the majority of our spontaneous ethical judgments, or the most confident of such judgments, does not prescribe that we modify others of our ethical judgments which clash with the theory. The theory, as such, only informs us of the modifications to be made in these others if we wish them to become rationally congruous with the rest, or the firmest, of our spontaneous ethical judgments. And it is this wish for theoretical integration, if we have it, which prescribes that we make those modifications.

However, the erroneous belief that ethical theory is prescriptive may have another source beside confusion between information as to the changes needed for theoretical integration, and the wish for such integration. The belief may be due instead to the fact that approval or disapproval of some thing is an implicit or incipient prescription that it be done or not done. Accordingly, ethical theory, being about approvals and disapprovals, is about prescriptions (implicit ones) -- not in the sense that it makes any, but only in the sense that it explains and predicts certain ones, and informs us of certain modes of conduct or states of affairs which our wish for theoretical coherence of all our prescriptions (if we have it) prescribes that we prescribe.

10. Ethical Theories vs. Ethical Principles.- -- We have described above a species of conceptual apparatus consisting of a definition of some ethical predicate and of a method for deciding whether any given empirical subject does or does nor exemplify the definiens. When such a conceptual apparatus is constructed by a person to explain certain of his spontaneous ethical judgments, it constitutes a theory of these judgments; and if it meets the tests of validity for explanatory theories in general, it provides the judging person or group with the information they need to make their judgments, in cases where they are in doubt, theoretically coherent with the spontaneous ones the theory explains and predicts.

But the manner in which such a conceptual apparatus comes to be adopted may be very different from this. Instead of being constructed by a person or group for himself or itself, it may be presented ready-made to their minds by somebody else, and be adopted not at all because of any explanatory power it may have, but simply because the psychological manner of its presentation -- be this suggestion, sugarcoating, pressure, or anything else -- is psychologically adequate for getting it adopted by them. In such a case that conceptual apparatus can no longer be described as for them a theory explanatory of given spontaneous judgments. It is rather a principle of ethical judgment. It is a general rule for determining what sorts of things to approve and disapprove -- a general rule for arriving at specific ethical norms.

Adoption of it insures unity of principle in ethical judgments derived from it, but question arises as to what other sort of merit, if any, it may have. Evidently it does not have that of explaining or predicting the spontaneous judgments of the person who has adopted it, since it was not desired to do this and rather censors or supersedes them. Such a rule of judgment can only have the merit of being a means to an end desired by whoever urged or forced its adoption. And it is difficult to see what sort of end this could be other than to bring about or to prevent -- through the ethical judgments dictated by adoption of the rule -- situations respectively approved or disapproved spontaneously by the person who urges adoption of the rule.

When a child adopts a rule from his parents, the situations which the parents aim to bring about or prevent are ones that they spontaneously approve or disapprove, and believe the child also will later respectively approve and disapprove spontaneously. Where adoption of a rule is psychologically enforced by a tyrant upon his subjects, the situations intended are ones that he himself spontaneously approves, or disapproves notwithstanding that he knows that his subjects appraise them differently. But in any case, adoption of a rule or enforcement of it on someone is rationally defensible only by reference to expected fruits consisting of situations spontaneously approved or disapproved by some specified person or persons. If no spontaneous approvals and disapprovals existed, no set of maxims for conduct, be they ever so coherent in principle, would have anything whatever to recommend it over any different but equally coherent set. Even when a philosophical theory or principle is free from formal defects, an appeal -- initial or terminal -- to philosophical experience . is thus indispensable for validating or invalidating it. On the other hand, philosophical experience without philosophical theory would remain as precarious, as poor in content, and as limited in applicability as would physics if deprived of physical theory and confined to empirical generalizations.

It is worth noting that ethical theory can serve only to make future conduct theoretically coherent with past or present ethical experience. It might be that the remote consequences of something we spontaneously disapprove are of a sort we would approve strongly enough to make us approve instead of disapprove what brought it about if we knew that it has these remote consequences. If we do not know this, theorizing will never inform us of it. The only thing that would inform us of it would be observation that the remote consequences are of that sort. If such information were sought experimentally, the experiments would, let it be noted, consist in deliberately doing something which both spontaneous judgment and theory based on this pronounce wrong, in order to find out whether consequences of it more remote than theory has yet taken into account may not be of a sort that would ultimately lead us to pronounce the act or situation right instead of wrong. Since this does sometimes turn out to be the case, there is in the field of ethics justification for a certain amount of experimentation flying in the face of past experience and of theory based on it -- for a certain amount of adventure or exploratory conduct prima facie unethical instead of prudent or ethical.

11. Experience and Theory in Branches of Philosophy Other than Ethics. -- What has been said concerning the relation in ethics between theory and given general facts (whether directly obtained from primitive or from derivative facts) applies equally, mutatis mutandis, to branches of philosophy other than ethics. This may be shown by considering briefly one or two examples relating to the theory of knowledge. It is well to bear in mind, however, something already emphasized, viz., that theories, in philosophy as elsewhere, and in the theory of knowledge as in ethics, are more often directly about derivative facts of the given field than directly about its primitives, and that therefore, although appraisals are the primitive facts of philosophy, most of the terms as to the meaning of which philosophy is called upon to frame explanatory hypotheses are not themselves terms of appraisal but rather stand for things implicit in appraisals and brought out by prior study of the latter. This prior study, moreover, may to some extent or in some cases have been automatic rather than deliberate, and made by generations past rather than by ourselves, its results then being matters of "commonsense knowledge" rather than of esoteric and technical knowledge. We shall here, however, as in the case of ethics, take terms of appraisal as examples, because to do so will stress the fact that doubt of given appraisals is what originally generates philosophical theorizing.

As an example of a spontaneous judgment of appraisal in the field of the theory of knowledge, we may take this: "The inference that, since all men are mortal and all vertebrates are mortal, therefore all men are vertebrates, is fallacious." This appraisal can be regarded as spontaneous in the sense already specified. That is, it is an appraisal that would be made by many persons who, although logically sensitive, are wholly ignorant of syllogistic theory. If this appraisal were disputed, the question would immediately arise as to what exactly is being predicated of that inference when it is called "fallacious." Any answer constituting an explanation of the appraisal (whether tenable or untenable), will be one from which it would be logically possible to deduce the fallaciousness not only of the given inference, but also of some other spontaneous inferences resembling it more or less closely. That is, it will be an explanation not exclusively of the particular fact that the given inference is fallacious, but essentially of an empirically ascertainable general fact, that is, an empirical law -- for instance, of the law that all inferences of the form "all S is M, all P is M, therefore all S is P" are fallacious. The theory consisting of the definition, "A fallacious inference is an inference in which the middle term is undistributed," and of the method of identifications, "The middle term is the term which occurs in both premises, and a term is undistributed if it is predicate of an affirmative proposition or subject of a particular proposition," would explain that empirical law and demonstrate, as a case under it, the fallaciousness of the given inference.

But if the appraisal of that inference as fallacious were defended in this way, the objection would very likely be made that although the theory of fallaciousness stated does rationally justify that appraisal and some others, there are many other inferences appraised also as fallacious, the fallaciousness of which is not accounted for by that theory. To integrate these other appraisals with the given one, a more comprehensive theory would have to be produced. It might , be the traditional syllogistic theory; or, since not all inferences are syllogistic, some still more comprehensive one might be needed to account for all the spontaneous appraisals of fallaciousness taken as data. A theory adequate to do this would doubtless also reveal the fallaciousness of some inferences in cases where one found one's self unable to make a confident spontaneous appraisal.

Other terms of epistemic appraisal, such as "sound" and "erroneous," "correct" and "incorrect," etc., could be taken as examples, and the relation of philosophical theory to primitive philosophical experience illustrated in terms of them in a similar manner. But perhaps the most comprehensive terms of epistemic appraisal we could consider would be "knowledge" and "ignorance." It is true that, like other terms illustrated in the discussion of the relation of appraisal to description in chapter X, they can be and sometimes are applied without import of appraisal; but in the large majority of cases, when we declare an opinion to have the status of knowledge, a part at least of what we mean by this is that it is better than -- superior to -- certain others (which we might appraise more specifically as erroneous, or as possibly sound but unproved, etc.) and better also than absence of opinion, or of opinion having the status of knowledge, on the same subject. Doubt or dispute as to whether or not a given opinion does have the status of knowledge would -- in cases where additional examination of the data of the opinion does nor settle the matter -- automatically bring up the question as to what exactly we are predicating of an opinion when we appraise it as being knowledge; and the whole of the theory of knowledge develops, in the general manner already repeatedly illustrated, out of the attempt to answer this question.

That most of its familiar problems are derivative from this question about its primitives is obvious if we consider any plausible answer to it. For example, if the answer is proposed that knowledge is the status an opinion has if and only if it is based upon evidence sufficient to prove its soundness, then the need to specify a method by which to determine whether or not any given opinion is so based forces us to consider what exactly an opinion is, how opinions are related to propositions, and how the soundness or erroneousness of opinions relates to the truth or falsity of propositions; what propositions themselves are, and what their truth or falsity consists in; what is meant by saying that an opinion is based upon "evidence"; what sort of thing is capable of serving as "evidence"; what is the test of the "sufficiency" of given evidence to prove the soundness of a given opinion; how proof and probability are related, etc. All these questions, and many others that are admittedly philosophical but seem to have no connection with appraisals, are thus implicit in the initial question as to what is predicated of an opinion when it is appraised as being knowledge. Any attempt to answer the initial question exhaustively brings out all these, and thereby demonstrates that they are derivative from a question about appraisal.

[Table of Contents] [Chapter 13]


{1} That this is the task of ethical theory has been pointed out by J. Rueff in his book, From the Physical to the Social Sciences (1929), which deserves to be better known than it seems to be. The task of ethical theory, he contends is to "enunciate a system of initial propositions, axioms, and definitions which, when fed into the reasoning machine, will produce theorems coinciding with the rules of practical morals" (p. 79). [Back]

{2} The theory that "wrong," when applied to the subjects to which a person (or group) P spontaneously applies it, means D is supported or invalidated by observing whether the predictions implicit in D turn out true or false, and not by the testimony of P as to what he means by wrong. For his testimony as to this could represent only what he thinks he means by D, that is, only what he believes to be the theory of wrongness which would accurately predict his spontaneous applications of the term wrong. [Back]

{3} Op. cit., 75, 77. [Back]

{4} This hedonistic theory of right and wrong is introduced here only for purposes of illustration, and its intrinsic merits are therefore a side issue. It may not be amiss to remark, however, that the "hedonic calculus" usually proposed as method for deciding whether a given sort of act is wrong (on the hedonistic definition of wrong) is a process not performable even in principle, to say nothing of at will. No way is known for measuring pleasures and displeasures either in the "fundamental" or the "derived" manner (see chapter on measurement in Campbell's What Is Science?). Even if we knew how to measure them, no psychological or physical summation and subtraction operations are known, the results of which would be predicted by arithmetical summation and subtraction of the figures obtained by measurements. (Cf. "Symbols, Signs, and Signals," loc. cit., pp. 48-50). It should be noted that as "calculus of self-realizations" is for similar reasons, likewise flatus vocis. On the other hand, the method specified in the text above, even if not as neat as one could wish, is practicable, and is approximately the one actually used when decisions have to be reached as to whether less, or more, happiness than unhappiness would result from a given course of action. [Back]

{5} That is, if the information of any given person, his habits, and his relations to the entity appraised were the same as those of a given other person, he would appraise it identically. (This postulate is plausible in matters other than those described as "matters of individual taste"). [Back]

[Table of Contents] [Chapter 13]