CHAP. I THE THEORY OF ASSOCIATION OF IDEAS 34!
arise, that is really all that, as psychologists, we can be con cerned with. Our principles are nothing but working hy potheses: we do not know and we do not care if they turn out to be fictions, when examined critically.
That is the way in which psychology surely might be studied. And if we studied it in this way we should escape some controversies. I, for instance, should lose all right and all desire to criticize the " Laws of Association " on the ground of their untruth, if they only ceased to proclaim themselves as statements as to the real movement of the mind. Within the same field of empirical psychology I should offer what I think is a more convenient hypothesis, and any objection to that which rested on metaphysics would be at once ruled out of court. We might perhaps thus advance the study of the subject in a way which now seems quite impossible. And if we did not make much advance in knowledge, we should save ourselves at least a good deal of bitterness.
§ 35. The suggestion is offered in great humility, since the obstacles it must meet with are overpowering. The first obstacle is the prejudice of a bad tradition. It is supposed that the psychologist must be a philosopher. He is used to think himself so, and he is not likely to accept a lower place. And this objection is in fact, I fear, unanswerable. I would give him the name of philosopher for his asking, but I could not admit him as a student of first principles. And the second obstacle is like the first. We get into what, I suppose, deserves the name of an antinomy. The psychologist is to confine him self within certain limits; he is not to cross over into meta physics. But unfortunately if he is not a metaphysician he will not know what those limits are. And it is the same to some extent with all the sciences. The physicist, for instance, is constantly tempted to think that his ruling ideas are ultimate facts. And this temptation is fatal to the mere specialist. It is only, on the one hand, a general culture and largeness of mind, or else some education in metaphysics, which saves him from this error. And it is much worse in psychology.12 The subject brings with it a special temptation; and, if all the truth must be told, the same great minds that devote themselves to physics, to chemistry, or to biology, do not take up psychology. And then again the psychologist is probably a dabbler in meta- 342 THE PRINCIPLES OF LOGIC BOOK II. PT. II physics. A little metaphysics is not enough to show that his so-called principles are fictions. And our leading English psychologists perhaps only know a very little metaphysics. And, having a limited acquaintance with the subject, they persuade themselves, and (what is worse) one persuades the other, that they have completely mastered it. It is to be feared that this evil must to some extent continue.
And there is a final obstacle. The student of metaphysics may form an opinion as to the real nature of psychical phe nomena. And knowing, as he thinks, the truth about these facts, he will be led to insist on a psychological interpretation which is strictly true. He will interfere with the empirical psychologist, and will himself contribute, by what he thinks good metaphysics, to the begetting of bad metaphysics in opposition. This is certainly an error, but it is an error, I fear, which will never quite vanish. When a man has once seen that every single science except metaphysics makes use of fictions, he is apt to conclude that the next step is for him to remove these fictions and to substitute the truth. But, if he looked closer, he would see that human beings can not get on without mythology.13 In science, in politics, in art, and re ligion it will always be found, and can never be driven out. And, if we confine our attention to science, we must say that there is only one science which can have no hypotheses, and which is forbidden to employ any fiction or mythology, and that this science with some reason is suspected of non- existence.
§ 36. We have approached a large subject which we can not deal with, and which might well occasion misgiving and doubt. We need give way to neither in our rejection of the principles of the school of Association. We reject them in the name alike of metaphysics, of psychology, and of logic. In behalf of metaphysics we protest against the basis of dog matic Atomism, and we protest against the superstructure of a barbarous mythology. It is not true that mental phenomena are mere particulars. It is not true that ghosts of impressions leave their graves. It is ridiculous to couple the existent and non-existent, or the present and the absent, by a relation which implies the presence of both. In defence of psychology we protest against an hypothesis which has to postulate phe- CHAP. I THE THEORY OF ASSOCIATION OF IDEAS 343 nomena which are clearly absent, and then to postulate their removal by a process which is not present. When a single hypothesis explains the facts, it is surely unscientific to employ a complication which works no better. And, in behalf of logic, we must protest once more. The essence of inference can hardly be a principle which later we recognize as a principle of error; and which, if the theory of Association were true, we should hardly get to perceive was false. It is an ill omen for Logic if it fails to show that what in the highest stage is accepted as a canon, was active from the first development of the soul as the guide of its conduct and ruler of its life.
NOTE TO CHAPTER I § I. Though I have no space, and perhaps no strict right, to deal with the subject here, I must yield to the temptation of making some very brief remarks on the doctrines noticed in §2 of this chapter. These go by the names of Indissoluble Association and the Chemistry of Ideas.
The first of these doctrines is supposed to have a very great metaphysical importance. Mere chance conjunction, if often repeated, will beget, we are told, an union of ideas which is irresistible. This shows that what seems to be a necessary connection may be no more than an accidental adherence. From this we conclude that a necessary connection is no canon of truth. And this proves that our trust must be placed elsewhere. The Logic of Experience tells us, of course, what it is we are to trust to.
For myself, in the first place, I never could get any information from that Logic which seemed intelligible, and so I will confine myself to the former part of the preceding statement.
§ 2. The first fault I have to find is that it does not go far enough. We need not have a repeated conjunction. One single instance is enough to give rise to a necessary connection. For, as we should say, what is once true is true always.
§3. I have to complain, in the second place, that all kinds oi combination are called association. But association surely implies that the elements which are joined might not have been joined. And should be proved, or at all events made probable, before co-existei is assumed to be mere association.
84 It may be replied, "Even if the things are connected, yet, as we perceive them, their union for us must be chance conjunction, and therefore association." But this again should in no case be asserted without some ground. It is not always self-evident that the mind co^ have had one element without the other. And where you fail show that this is the case, you cannot talk of association.
344 THE PRINCIPLES OF LOGIC BOOK II. Px. II § 5. I shall be answered, " What we prove is that in certain cases mere chance association has produced necessary connection; and we argue from this that it may be fairly suspected of doing so in all. The possibility is proved and the possibility is enough." I can not enter here into the merits of this argument which I shall hereafter show is logically vicious (vid. Bk. III. II. Chap. III. §22)14; but sup pose that for the present we admit it. What conclusion follows? That we are fallible men? We knew that before. That we are to trust to anything else? Then what else? Admit for argument's sake the possibility that all our beliefs are baseless, what then? Why nothing. If we mean to go on living and thinking, we dismiss this possibility as idle. Suppose we all are victimized by chance con junction, are we not right to be so victimized?
§6. Association implies other conditions. It implies contingent circumstances. When a chance conjunction is taken by an error for a necessary connection, the mistake really consists in defective analysis. The remedy is found in the progress of analysis, assisted probably by fresh fact. Where this remedy is impracticable, no remedy can be applied. For no other is possible.
§ 7. Apart from mental chemistry, which we shall consider pre sently, a connection of ideas could not continue to be necessary when it demonstrably has arisen from association. And this is quite ob vious. For the connection of ideas supposes a content which is ideally inseparable, and the knowledge of the association involves this ideal separation. The experience which shows the fact of the association, is at the same time the analysis which loosens its bond.
§8. This however is a minor point. To the objection that possibly all truth may be nothing but chance association, we reply (as above) that, supposing this for argument's sake to be true, we can not trouble ourselves with idle possibilities. But if you wish to go beyond this idle possibility, you must show cases where unreasoning chance conjunction has produced false belief -without confusion. You must show, that is, that the belief in the connection was wholly false; that it was not a true belief in a real fact made false simply by a confusion between the relevant and irrelevant elements in the connec tion. But this, I think, has never been shown.
§9. If association rests on conjunction in perception, then that is a valid ground for belief. It is deceptive merely so far as it is unanalyzed, and confuses the irrelevant with the relevant. Otherwise it is a proof of necessary connection. But then this latter is not mere association. For it is not every conjunction in presentation which can be called an association, but only those conjunctions which result from chance. And chance disappears before analysis.
§ 10. I will now turn to the doctrine of mental chemistry. Ele ments by virtue of repeated chance conjunction are said to cohere in such a way that they form a third product which has the qualities of neither. But this in the first place would not be association, since that term implies that the individuals continue. In a chemical union l5 CHAP. I THE THEORY OF ASSOCIATION OF IDEAS 345 the molecules of the substances cease to be molecules of either sub stance. It is therefore nonsense to say that they are associated.
§11. This of course may be said to be a question of words. But the fact of such union in the case of ideas has, at least in my knowledge, never yet been shown. It can not be called impossible, nor should I at least have said that it was even improbable, but I have never seen any certain instance of it.
§ 12. In the case of emotions this " chemical union " does seem to take place. But even here there might be doubt if the emotion should properly be considered as an " union." It might rather be a new reaction on a fresh compound material.16 But, however that be, it is true that the emotional product often can not be analyzed.17 It can only be reconstructed perhaps in part hypothetically. And again if we take intellectual functions, there is no doubt that in the process of mental development " faculties " are produced which are different in kind from what went before.18 But then again these functions are hardly unions of ideas. When you strictly keep to mental objects, I think you must say that no instance of what looks like chemical combination has yet been found.
§ 13. It is of course mere waste of time to bring forward as evi dence cases where the fact of the association is not admitted. It is for example a mere circle to instance the idea of visual extension, since visual sensations without extension are the merest hypothesis. Not only can this alleged fact not be observed, but there are very strong reasons for rejecting it wholly.* It is not less idle to bring forward a product, such as the sensation of white, and then roundly assert that it is the fusion of different sensations. Perhaps it is, but you would have to show the existence of these sensations in the particular case, and give some reason for your belief that they were transformed. It is finally ridiculous to adduce, as a chemical product, an idea which can be separated at once and with ease into its component parts. J. S. Mill when hard pressed seems to play as his trump card the idea of infinity (Hamilton, Note to Chapter XV.). But infinity, as he understands it, hardly calls for analysis. < it falls apart into its elements, for it is a mere mechanical union.
The conclusion must be that the chemistry of ideas is no more than a hypothesis. I do not think in any case it would be the right way to state the fact. But the fact itself has not been clearly shown to exist.19 In the second place, were we convinced that mere chance conjunction was able to lead to it, then nothing would follow except what we know, viz. that there is some general antecedent probability that any conviction is false. This result makes no difference either to theory or to practice.
*Vid. Stumpf, Raumvorstellung.
346 THE PRINCIPLES OF LOGIC BOOK II. PT. II ADDITIONAL NOTES 1The length of this Chapter was, I think, justified by the general state of psychology in England in 1883. The reader is asked to bear in mind that perhaps the greater part of it was written at a particular time, and for a special purpose, and hence should now be superfluous. This remark does not, however, apply to the doctrine that Association holds only between universals. This doctrine, as I have long ago mentioned (in Mind N. S., No. 20, p. 472, and, I think, elsewhere) — I owe to Hegel (Encyk. §§452 foil.). For my indebtedness to him in psychology otherwise see on Bk. III. I. VII. I may add here that, so far as I remember, in 1883 I had not yet made the acquaintance of Herbart, with Drobisch and Volkmann, or even of Waitz or of Lotze's Medicinische Psychologic. I had, however, read most of Steinthal's Sprachwissenschaft, I.
2 "If the condition is known." It would be better to say "is taken to fall within the subject," and for "where the condition is unknown" to write " where you can not specify the condition."
3 " Analytical." Cf. Bk. III. I. VI. § 10.
4 " When they are dead, etc." I should certainly have used other language here with reference to "survival," if I had been better ac quainted with the excellent work done in psychology by Herbartian writers. The doctrine that every mental state still survives and is active below the conscious level, was, and is, as a working hypothesis, not to be treated with contempt.
5 "What is left behind is a mental result." Cf. §26. We have here the problem of Dispositions. This, I should say, is in the end insoluble, if you ask for anything beyond an empirical Law or Laws. Cf. Mind (O. S.), No. 47, p. 363, No. 13 (N. S.), p. 25, and No. 33, p. 9.
6 " Logical." Cf. Bk. III. I. III. §§ 20 and 23. Association becomes logical by its use for, and subordination to, a logical end; where, that is, it is controlled, for the purpose of truth, by the identity and indi viduality of an object. Cf. Mind (O. S.), No. 47, pp. 381-2, and Essays, pp. 362 foil.
7 " They all have content." Cf. §§ 30 and 31. The point here is that you do not anywhere have a psychical fact which is purely simple and in this sense unique. Everything given, we must remember, is always in some sense itself qualified by its context. See T. E. V.
8 "This is not merely an assertion I have chosen to make." Prof Sully on the other hand, in The Human Mind, I, 331 (1892), says (I understand) that it is so. While directing the reader to this criticism I may add that I neither saw nor see any need to reply to it. The reader who can not deal with it for himself will, I think, have read this Chapter in vain.
9 " Primitive Credulity." Cf. Bk. III. I. VI. § 32, and Essays, p. 377.
10 " This is a result." See Note 5.
CHAP. I THE THEORY OF ASSOCIATION OF IDEAS 347 11 " In Professor Bain's account, etc." Though the temptation was irresistible, I am sorry that I treated Bain, here and elsewhere, with so little respect. Let me say now that, so far as I know, he was the only writer of his time and school who made an original con tribution to psychology. Though one may find him at times to be absurd, one seldom finds that he has not at least made an instructive effort to see the facts for himself.
12 " And it is much worse in psychology." The reader here must not forget that I was pointing to psychology as it was in England, still at least in the main, in 1883. To use such language now would be absurd and even monstrous.
13 " Mythology." Cf. my Appearance and Essays, and, in particular on psychology, see Mind, Vol. IX, N. S., No. 33. It is hardly neces sary, I hope, for me to inform the reader that this view of truth, as being more or less mythological everywhere outside metaphysics, came to me from Plato and again still more from Hegel.
14 Cf. here Appearance, p. 620.
15 " Chemical union." I am of course here not endorsing but merely adopting the view offered me as to the real nature of chemical union.
This, I presume, might be the place to discuss the doctrine of Fusion— whether of one sensation with another or with ideas and Dispositions— or again of these two last, each with themselves or with one another. But I can not venture upon such a difficult subject here.
is "It might rather...material." This seems certainly 1 better view. No psychical state, as a unity, can be wholly resolved into the mere compounding of units— even " chemically."
" " Often can not be analyzed." The meaning here is that there are some emotions, to."analyze" which, except quite inadequately, you must have recourse to what you believe is their origin.
is " Functions " and " faculties." This statement of course will apply to much of what are called "instincts."
19 "But the fact itself...exist." By "the fact itself I mean the appearance of a complex psychical state, in which the elements, as such, have become indistinguishable, and which can nevertheless be taken as the mere product of their union. There are but few psychologists, I should think, who would now accept such a fact However that may be, we can and must, I should say, maintain that new products have been developed, but certainly not by chemical union of ideas. Every new mental product is rather a fresh reaction from the individual totality; and, to explain, we.must seek.to find the Laws by which Dispositions are formed, and by which the result former experience meets and more or less transforms the incoming stimulus.
CHAPTER II CHAPTER II THE ARGUMENT FROM PARTICULARS TO PARTICULARS1 § i. At the point which we have reached a discussion of this subject may seem inexcusable. If we have shown that no association is possible except between universals, and that in the very lowest stages of mind universals are used, we may fairly be reproached by the reader who is anxious to learn something new, if we linger over errors the root of which has long since been torn up. For supposing that the results we have attained to are sound, the question is settled. To reason directly from particulars to particulars is wholly im possible. It must be at most a desire of the mind which this world can not gratify, a postulate a priori given by an intuition, that disappears before analysis and is rejected by experience.
§ 2. But since it is possible that the reader of this Chapter has not accepted the conclusions we obtained, since it is not unlikely that he has passed them over, let us try once again if we can not do something to turn the light into this refuge of darkness. We must not expect to persuade the disciple of the Experience Philosophy. It is not for anything we are likely to offer, that he will desert the fashionable and easy creed in which he has been reared. But at least we shall have tried not to leave him an excuse. He must not say that we have been afraid to look his idol in the face.
There is however one thing we will not do for his sake. We decline to supply a direct examination of the well-known chapter in J. S. Mill's Logic. It would require much more space to set out the ambiguities inherent in that chapter,2 than we can give to the discussion of the question itself; a dis cussion to which, I may remind the reader, I consider that at this stage he has no right.
§3. Why should we not reason from mere particulars? Do our reasonings never rest upon fact? And what are facts if they are not particulars? Either then we never, starting CHAP. II THE ARGUMENT FROM PARTICULARS 349 from fact, conclude to fact, or else we infer particulars from particulars. This result may so be deduced from first prin ciples. And common experience supports the result. From cases we have known we go to fresh cases without an appeal to any general principle. We have seen something happen and, given a new instance, we argue at once that it will hap pen again. But we have no reason other than this fact to give for our conclusion. We thus in the second place have proved our thesis a posteriori, as before we proved it a priori. And now we add an indirect proof. If for reasoning were wanted major premises, then the lower animals could not reason. But they do reason, and therefore the thesis is proved. § 4. How shall we escape from this array of proofs? Are they not unanswerable? To me they seem unanswerable, and I have not the smallest wish to escape them. I admit them and embrace them; but I ask a question, What is it that they prove?
They prove first that, when we go from experience of facts, this experience is the foundation of our inference. They prove again that we do not always go from an explicit major premise, and that therefore another way of reasoning is pos sible. And in defence of these results I am as zealous as any of my readers can be. If he likes to say beside that a syllogism in extension is a petitio principii and no argument at all, he will urge what long ago I have endorsed. But let us come to the conclusion. If you mean to argue to no more than this, that experience of particulars is a basis of inference, and that no explicit major is required, I am ready to support you. But if you mean to conclude, Therefore we reason from particulars as such direct to particulars, I object at once. The conclusion does not follow from the premises, and it also is wholly contrary to experience.
§ 5. We have in fact to do here with a common-place logical blunder. The thesis to be proved is that an inference is made direct from particulars, as such, to other particulars. The conclusion which is proved is that from experience of particulars we somehow get a particular conclusion. Not to see the enormous difference of these assertions is to fall into a gross ignoratio elenchi. To prove the thesis in dispute it is necessary to assume that either we go direct from particulars 35O THE PRINCIPLES OF LOGIC BOOK II. PT. II to particulars, or else advance through an explicit syllogism (perhaps even an explicit syllogism in extension). No sort of evidence is offered to show that this alternative exhausts the possibilities; and it disappears the moment we confront it with facts.
§ 6. In reply to the assertion that we are able to argue from particulars to particulars, I would ask in the first place what particulars are meant. Am I to understand that the past experiences in their particularity are the premises used in this supposed inference? If I am told this is so, of course I reply that we have here a mere psychological fiction. Par ticular images of past occurrences, which retain the special marks of the originals, are not available. The doctrine that each perished perception leaves an unblurred unabridged counterpart of itself, is a preposterous invention (cf. pp. 35-7, and Book II. II. Chap. I.).
§ 7. It is again a mere error which sees in the lowest form of inference the presence of one or more images of the past, together with a fact which they are used to qualify. When a present perception is modified by the suggestions of past ex perience, these suggestions do not come from particular images of perished events. This theory is a second pure invention (cf. ibid.).
§ 8. In the third place when, at a higher stage of develop ment, the past event is as such called to mind, and when we do argue from a particular image, yet even then we do not argue from its particularity, from its psychological environ ment and temporary colouring. We argue from the content, the idea which can exist in different times and under diverse psychological conditions. And once more, and in the fourth place, this idea itself need not be used as a whole, but we may argue from one part of it.
§9. A child has come to know that, when the dog is pleased, he wags his tail. On this he argues that, when the cat wags its tail, it must be pleased. What is it he proceeds from? The error we are considering actually supposes that one or more images of foregone occasions, presenting the dog pleased and with his tail in motion, come before the mind, and that, on this, the perception of the cat now moving its tail directly gives rise to the conclusion, The cat is pleased. But the CHAP. II THE ARGUMENT FROM PARTICULARS 351 question arises, How is it that one attribute is taken from the dog-images and given to the cat, without the rest going with it? Does not this use of one part of the dog-images, and the neglect of the rest, show that something happens to the images in question, and that, however it has come about, the inference is not drawn from the whole of any one of them? Suppose again that they differ among themselves, do we argue direct from the whole of all of them? But if not, from what else?
§ 10. The facts, I should have thought, would have left little doubt that the result of experience is a connection of attributes, where the differences of their particular subjects are blurred — a confused universal, which may appear to the mind in a particular imagery, but is used without any regard to that. I confess I should have thought that it was very clear that, in the special cases where we argue from recollection, we use the past event as a type or instance. And since both this past event and the present perception come to us as instances, we neglect some of the differences that exist between them. We do not know the principle, but we feel " it is the same thing " in both cases. But, if so, the premise from which the conclusion directly comes, is not the particular. It is an universal extract, what we call a " general impression."
§ ii. Reasoning from a particular to a particular is obvi ously an argument from analogy. In this we all know that we do not use the whole of that particular from which we argue. It was an inference by analogy which deceived the child (§9). He took from the dog a relation of qualities and transferred it to the cat. What he argued from was this general relation, and it was a false analogy, just because it was a bad generalization. Again, why do we object to false analogies? Is it not because in them we treat some fact as another instance of a rule, when there is no common rule and the facts are not instances? And is not this a hint that in true analogy we use a principle though we can not state it?
§ 12. This leads us to put another question. Suppose that per impossible we did have before our minds a number of particular images, and did argue from them directly; would not this inference be a very bad one? If I say " A, B, and C are a, and there is no difference between D and A, B, and^C, therefore D is a"— is not this a circle— a frivolous prtitiof 352 THE PRINCIPLES OF LOGIC BOOK II. PT. II