CHAPTER XXIII.
CRITERIA OF ERROR IN REALISM.

§ 405. A realist, whether he explain human knowledge by means of a causal interaction between object and subject or not, must at least admit that somehow or other the object determines the content of thought, and that between percepts, concepts and judgments on the one hand, and real entities on the other, there is a correspondence or resemblance. So vital is this idea of correspondence to the theory of knowledge and truth that neither the absolutist nor the pragmatist is able quite to get rid of it. Kant, though rejecting correspondence in the realist's sense, retains it in another form. In the Critique of Pure Reason the syntheses of the imagination anticipate and correspond to those of conception. The structure of the rational mind, which is so minutely analysed by Kant, also corresponds to a type; and it is precisely because of this correspondence and its universality that 'objective' knowledge is possible. This idea of truth as conformity with a type or ideal is further developed by the Critical and Neo-Critical writers of to-day. In their view, Reality is a systematic whole, a coherent organised unity, constituted by the experience of the Absolute; and this Reality, considered as significant and intelligible, is Truth -- what the scholastic would call Ontological Truth. Ontological Truth, or Reality considered as intelligible, is the foundation of Logical Truth. Logical truth consists, therefore, in the conformity of the individual and finite mind with Reality itself; for it is essential to Truth "to manifest itself in the thinking of finite subjects." And precisely in proportion as this Ideal or Absolute Truth is realised in us, we possess knowledge or logical truth. Indeed, Dr. Joachim acknowledges that we can never rise above the level of knowledge which at the best attains to the truth of correspondence, since all human discursive knowledge must ever remain thought about another.^{1} The pragmatist, too, admits, as we have seen, at least provisionally, that truth consists in 'agreement with reality.' So that the correspondence-notion has, at any rate, an excellent claim to our attention, since it would appear that after all there must be some kind of correspondence between the mind of the knower and the object which he knows.

§ 406. Assuming, then, with the realist, that truth consists in an adaequatio intellectus et rei, and that this adaequatio or correspondence is due to the determination of the content of thought by the object, in order to ascertain the validity of any particular claim to truth, we need criteria by which to decide whether this correspondence exists, i.e., whether our thoughts have been objectively determined or not. This does not mean that per impossibile we must be able to compare our thoughts and our percepts with real things as with something outside of thought and perception, since clearly it is only by means of thoughts and percepts that things are known. The only comparison possible for us is a comparison between things as thought and as perceived. The realist asks for no 'miraculous second-sight' by means of which to detect the agreement or disagreement of copy and original, of idea and reality.^{2} Truth is correspondence of thought and reality, but this correspondence can be established without comparison, provided we can answer the question 'When has thought been determined by the object and when has it not?'

§ 407. Now there are two ways of attacking this problem. We may start from methodical doubt, or we may start from what I may call, perhaps, methodical and rational assurance. Granting that the function of thought is to give us knowledge of reality, we may begin by assuming that our thoughts as a rule go wrong, in which case we shall require criteria of truth by which to decide when they have not gone wrong. Or we may begin by assuming, not that perception and thought are psychological monstrosities, but that, in general, they perform well the function they are obviously intended by nature to perform; in which case we shall need, primarily, not criteria of truth, but criteria of error by which to determine when per accidens a function has gone wrong. This difference of standpoint is no mere verbal question such as whether we are to cut off a slice and keep the apple or to cut off the apple and throw away the slice; it is one of the utmost consequence in making up our minds in regard to the epistemological value of human cognition. If we begin by doubting everything and affirm that no proposition has more than a claim to truth until it is validated, we can never get at truth at all, for ex hypothesi our criteria of truth themselves have only a claim to truth, and so has the, validation-process; and so, also, for that matter, has our original statement that no proposition has more than a claim to truth until it is validated. Scepticism is just as much the inevitable outcome of a doctrine which begins by asserting that no truth is more than a claim till its validity has been tested, as it is of the doctrine which makes truth an absolute and impossible ideal. On the other hand, if we begin by assuming -- what in reality is not an assumption at all, but a self-evident truth -- that ordinarily and per se perception, conception and judgment, when functioning normally, are capable of giving us knowledge, then we have something positive and definite with which to start, and can proceed to establish criteria by means of which to eliminate abnormal functionings which are likely to lead to error.

§ 408. The latter standpoint is that of the realist. The first principle of his theory of knowledge is that knowledge is possible, and that his cognitive faculties are capable of attaining to truth, which is the purpose and end of their existence. His aim, then, as a logician, is not to devise criteria of truth, but rather criteria of error. Error is a fact; but it is not a sufficient reason for us to distrust altogether the deliverances of our faculties. For, being a fact, it must have a cause, and we may be able to discover this cause and so to eliminate error by distinguishing from the rest those circumstances in which the deliverances of our faculties are not to be trusted.

We have said that truth consists in conformity between the content of thought and the nature of the object or system of objects to which it is referred. And conversely error consists in a positive disconformity between the two. In other words, truth is the reference of the content of thought to an object to which it corresponds; while error is the reference of the content of thought to an object to which it does not correspond. Now the correspondence in which truth consists arises when the content of thought has been determined by the object to which it is referred. Error, on the other hand, since it consists in positive disconformity, implies that thought has been determined by something other than the object to which it is referred. In order to decide, therefore, under what circumstances the deliverances of our cognitive faculties are not to be trusted, we must know what other causes may determine the content of thought besides the object to which it is referred.

There are several causes which may determine the content of thought. It may be determined (1) by the object of sense-perception through the mediation of the phantasm; (2) by habit or association of ideas, and (3) by the constructive activity of thought itself guided by some purpose which it seeks to realise. And each of these determining causes may per accidens, and under certain circumstances, lead to error. Error, however, will not occur until a false objective reference has been made, for in abstraction from objective reference neither sense-impressions nor ideas are either true or false. Hence the causes of error which we are about to consider are material rather than formal causes; they account for the disconformity between thought and thing, not for the false objective reference, which, as we shall see, is due to quite another cause.

§ 409. All that the senses give us as such (i.e., abstracting from any intellectual function that may co-operate in human sense-perception) are the external appearances of things; and appearances cannot, in themselves, be false, though they may be misleading; which is, I take it, what we mean when we speak of 'false appearances.' Now 'false appearances' may be due either to objective or to subjective conditions. The bent appearance of a straight stick when it is partly in and partly out of water is an instance of the first, for the optical illusion in this case is due to the refraction of the rays of light which have to pass through the water in order to affect the eye of the observer. Again. the Muller-Lyer illusion is due to objective conditions. A straight line appears shorter if it is terminated by arrow-heads which are turned inwards, and longer if the arrow-heads are turned outwards the reason being that in the first case the motion of the eye by which the length of the line is estimated, is checked, and in the second case increased, by the arrow-heads. The 'false' appearance in both these examples is due to the determination of the sense-impression by something else besides the object upon which attention is focussed. It is not merely the stick and the straight line which determine the 'form' which the percept takes but the stick and the straight line in certain circumstances. Neither case, however, is strictly normal, for ordinarily we do not perceive things through water, nor do we require to judge by the senses, alone and unaided, of the length of lines in more or less complicated figures. And though in neither case can we get rid of the optical illusion, it need not lead to false judgment, provided we take account of the circumstances. For in both cases the tendency to error can be counteracted, in the first by handling and in the second by measurement. Nevertheless, the examples given may serve to establish two useful criteria in regard to sense-perception. First we must be careful to take account of the circumstances under which perception takes place, and, if abnormal, must experiment in order to discover whether the special circumstances make any difference to what we perceive. And, secondly, if accuracy as to detail is required, we must make use of instruments which place the senses in conditions in which they are known to be reliable.

§ 410. False appearances are sometimes due to subjective conditions, to the inaccuracy of the senses or to their abnormal state. The sense-impression does not always correspond exactly with the objective stimulus. For instance, a straight edge of paper may, as has already been pointed out, be perceived as bent if applied to a line of nerve-terminals which itself is bent. And though the circumstances in this case, being those of a psychological experiment, are clearly abnormal, in no case can the senses be relied on for more than a certain degree of accuracy. In respect of accuracy, however, the senses vary considerably, so that the inaccuracies of one sense acting under certain conditions may be counteracted by employing the same or another sense under different conditions, or again, and more especially, by the use of instruments.

Other errors, v.g., illusions and hallucinations, may be due to the abnormal conditions of the senses or of the brain. Giddiness may cause an apparent motion of the room in which one stands; paralysis in one of the muscles of the eye may cause the stone-mason to strike his hand instead of his chisel while abnormal sensibility in any organ may lead, not only to exaggerated perception, but to the perception of what is not real at all, but merely subjective. It is clear, then, that the senses are trustworthy only so long as they are in their normal condition ; otherwise they are likely to lead us astray.

§ 411. The 'relativity' of sensation seems to present a more serious difficulty. But this is not really so; for in all cases the illusion due to the relativity of sensation may be counteracted by varying its circumstances or its antecedents. If we want to judge of colour accurately, we must judge of it in various lights and with various backgrounds though in ordinary life we do not need this exceptional accuracy, for colour is only one among many qualities by means of which we distinguish objects. Moreover, contrast seldom effects a total change in the colour of an object, and since it is by differences of colour for the most part that we distinguish objects, and since differences are not toned down but exaggerated by contrast, it is probable that the latter assists rather than hinders the true perception of real things. The 'relativity' of sensation is also quantitative. Two objects of the same weight but of different size are perceived as of different weight if one is lifted immediately after the other. But here, again, if we want accuracy we do not trust the unaided senses, but have recourse to instruments. Fechner's law, perhaps, might be cited as an instance of quantitative illusions due to relativity; but to me it seems that to perceive differences proportionately is a more accurate way of perceiving them than to perceive them by addition and subtraction.

One more illusion illustrating a relativity of a different kind. When seated in a railway train the fields and hedgerows seem to rush past us as we proceed; whereas, in reality, they are motionless. The cause of this illusion is that when we are unconscious of our own motion, movement, if any, appears to belong to the object. The alleged 'apparent' motion of the sun round the earth is due to a similar cause, though here there does not seem to be an illusion strictly so called. For we do not perceive an actual motion in the sun ; all we perceive is that its position relative to our point of observation changes from hour to hour. There is no 'false' appearance; but the error of the older astronomers was due to a gratuitous assumption leading to a false inference. Illusions due to the relativity of motion, however, rarely lead to error, and need never occur provided we are careful to ascertain whether the object or system of objects relative to which another object is said to move, is itself really stationary.

Ordinarily, then, appearances are to be trusted; but (1) we must exclude those cases in which the condition of the senses is abnormal; (2) when the circumstances are abnormal we must ascertain whether they make any difference to the percept and (3) if accuracy is required, we must employ objective methods of measurement which place the senses in such circumstances that they are reliable even in regard to details.

§ 412. Habit also determines the content of thought; and again per se it is not a cause of error. For a habit is but a tendency to repeat an act already elicited, and assuming that the previous act corresponded with its object, so also will its repetition. Habit, however, depends largely upon physiological conditions, particularly upon cerebral connections; and the same set of neurones may function on many different occasions not only in the experiencing of fact, but also in reading and conversation. Hence the course which the nervous impulse takes when free from intellectual control is somewhat haphazard, so to speak, and thus arise dreams and the play of fancy. It is clear, however, that neither dreams nor the play of fancy lead to error under normal conditions; for, although both have 'false' appearance of objectivity, there is, strictly speaking, no objective reference. Hallucinations and 'fixed' ideas, too, are clearly due to abnormal and pathological conditions, and, if continued, lead eventually to madness. With these cases, therefore, we are not concerned; but there are two cognitive processes in which habit is operative, both of which are of considerable importance from our point of view. I refer to memory and to the subsumption of particular cases under general ideas and laws.

In memory we definitely set ourselves to recall some past event, something which we have seen, or heard, or read; so that the cerebral processes which in part at least condition memory are not free to function according to the law of least resistance, but are subject to intellectual control. In what this control consists we do not know, for in spite of much modern research on the subject memory is still a mystery. Nevertheless, it is clear that the course which the nervous impulse takes is controlled in memory, and that in consequence ideas tend to revive in the order in which they occurred on the occasion which we are trying to recall. To me, indeed, this seems to imply not merely intellectual control, but an intellectual memory, the activity of which is conditioned physiologically, but as a habit is independent. At any rate, this would account for the fact that the intellect accepts or rejects ideas suggested by physiological associations according as the events we are trying to remember do or do not seem to us to be therein correctly recalled -- a fact which is not fully accounted for by internal consistency or inconsistency in what we remember. In any case, we do remember, and in so far as we have certain remembrance, our memory is ordinarily trustworthy, and if we wish for greater certainty we may appeal to written records and other objective criteria.

§ 413. The functioning of habit in perception occurs both at the sensory and at the intellectual level. The integration of the percept itself is due partly to habit, and the subsumption of the percept under a general idea or category is also due partly to habit. In both cases, however, the functioning of habit is controlled by objective conditions, and as the habit itself has been built up by previous perceptual and conceptual activities, themselves objectively determined, the correspondence of percept and concept is not destroyed, but rather facilitated and made more accurate. Thus, the perception of distance and of the third dimension is largely due to habit; but, though in the perception of long distances error may occur, and illusions of 'reversible perspective' are familiar to the experimental psychologist, here as before the senses are reliable except in special circumstances, and here as before their deficiencies may be easily remedied. Similarly in regard to subsumption. A few characteristics suffice whereby to recognise an object and to cause a certain concept to function in the mind. A few strokes of the pen are sufficient to suggest a soldier, a wheel-barrow, a bird, or any of the common objects with which we are familiar. And in the same way real objects are instantaneously subsumed, and usually correctly subsumed, under a general idea by force of habit, and without any careful examination of their characteristics. But sometimes the idea suggested by habitual association may be wrong, and the predicates which are contained within that idea may turn out to be incompatible with this particular concrete instance, as in the case of Madame Tussaud's policeman who is found to be not a policeman at all, but a wax figure incapable of movement. Seldom, however, when formal and deliberate assent accompanies subsumption, does error occur; for unless the object is familiar and the subsumption obvious, we hesitate, and only subsume after a more or less careful examination of the characteristics of the object in question. In most cases certainty is possible if we require it; and if we do not require it, we are content with probability, and so do not give a full and complete assent.

The influence of one's 'point of view' and of expectancy in perception and subsumption are particular cases of the influence of habit. The plain man and the artist admire the sturdy grandeur of a venerable oak; the botanist examines its peculiar structure and estimates its age; the builder laconically remarks that it is a fine piece of timber and would cut into a number of valuable planks. Each has considered the oak from a different point of view, yet neither has been thereby led to make a false judgment. Again, expectancy is seldom the cause of error. It may lead to a momentary illusion, but this is usually corrected spontaneously and immediately. Only in abnormal cases does expectancy lead to a false judgment, as in pathological cases, and in pre-perception, where the conditions are those placed by the experimental psychologist.

§ 414. Habit, then, though it is more or less concerned in almost all cognitive activity, seldom, if ever, leads to error, provided due care is exercised in any particular case, whether it be one of memory, of perception, or of subsumption under general ideas. There remains, then, what I have called -- for want of a better name, which should be equally comprehensive -- the constructive activity of thought. Under this head I include any kind of purposive enquiry, inference and postulation, since all these cognitive processes may be regarded as constructive. That purpose functions in almost all cognition can hardly be questioned but since, as I have already shown in a previous chapter,^{3} purposes do not affect the content of thought, but only its intent, i.e., the questions which we ask, the influence of purpose is not ordinarily, nor need it ever be, a cause of error. That de facto purpose does sometimes lead to error I grant; and as to why it does so I shall have something to say in a moment; but it is not because facts are 'selected' and 'accepted' according to our purposes, even though selection is 'immensely arbitrary,' and many facts are 'allowed to drop out and so to become unreal.' All who wish to explain or to theorise choose those facts which are relevant to the purpose in hand; but an honest enquirer will 'accept' indifferently facts which confirm, and facts which militate against, the hypothesis which is to be proved; while only those which are irrelevant will be neglected. Why purpose should be supposed to 'make' or to 'distort' fact I cannot understand. It makes us generally alert and observant; it determines the relevance of fact it causes us to 'accept' certain facts and neglect others, and to consider those we do accept in relation to our hypothesis and to neglect other relations and significances which the same facts may have, but which are irrelevant. In short, it determines what Professor Stout has called the 'intent' of thought. But the accepted facts are facts for all that; and their content is not affected by our choice of them in preference to others.^{4}

Almost all the difficulties which the pragmatist finds in the 'rigidity' and 'coerciveness' of fact arise from a confusion between content and intent. Take, for instance, the following dilemma in which Dr. Schiller sums up his arguments against the realist's position on this point

If our choice, selection, and congé d'élire does not affect the rigidity of fact, it is an illusion which ought not even to seem to exist, and we certainly have no right to talk about it: if, on the other hand, there really is 'selection,' will it not stultify the assumption of a rigid fact, introduce a possible arbitrary manipulation, and lead to alternative constructions of reality?^{5}

In reply to the first horn of this dilemma, we may grant that the congé d'élire exists and is no illusion, but really determines the intent of thought, i.e., it settles what facts are to be admitted as relevant but in reply to the second horn of the dilemma, we must deny that the rigidity of fact is stultified by our selection, since its content and nature is not affected. Nor can we admit that selection introduces arbitrary manipulations, since it is the facts themselves which must determine our manipulation of them if we are seeking truth; or that it necessarily leads to alternative constructions, since alternative constructions imply ignorance or neglect of relevant facts, and this is precisely what selection is intended to avoid.

§ 415. There seems to be no reason, then, why purposive selection or any other kind of purposive activity should lead to error. On the contrary, since, like habit, it facilitates both perception and conception by defining the point at issue and excluding irrelevant matter, it seems to tend rather the other way and to be, in fact, a function of the greatest service in the acquisition of truth. No intellectual function of itself can lead us astray. Abstraction is trustworthy, for in it we consider entities, which, though they do not exist as such a parte rei, nevertheless exist as aspects of reality, and by reality indirectly through the phantasm the intellect is determined to act. Abstraction is not falsification, unless we assume with the Hegelian that all aspects of reality intrinsically modify one another, or with the pragmatist that selection implies transformation: both of which assumptions are wholly gratuitous and unwarranted, and lead inevitably to Scepticism. Judgment, too, is trustworthy. For, if a posteriori and contingent, the conjunction of subject and predicate in a certain relation is directly determined by facts; and, if necessary, the predicated relation is implied in the abstract entities the nature of which the intellect has apprehended. In like manner, inference is trustworthy, for we apprehend explicitly what is already contained implicitly in the premisses, as when by means of the syllogism we apply a general law to a particular concrete case. We have no reason, then, for distrusting the results of our intellectual operations, which per se are reliable provided and in so far as they claim our assent.

But intellectual operations do not always claim an unconditional assent. Not infrequently judgment and inference is tentative, hesitating, doubtful. We are not sure that we have apprehended things correctly or that the conclusions we are inclined to draw are really contained in the premisses. This is the case not only in regard to inferences based on a complexus of hypotheses or principles, all interrelated and inter-dependent; but it is true also of those immediate judgments in which we apprehend a relation as holding universally and necessarily between subject and predicate. On this point Aquinas remarks that the recognition of the truth of principles varies according to the intellectual capacity of each individual.^{6} Sometimes our immediate judgments and our inferences are certain, as when the entities or laws with which we are dealing are comparatively simple but sometimes they are doubtful and present only a claim to truth which requires verification. Now a doubtful judgment, a doubtful inference, or a mere postulate is not an error, since, in so far as it is doubtful, it is not accompanied by assent. The activity of thought, therefore, even when inferential and constructive, is not per se a cause of error, though it may be so per accidens, as I shall now proceed to explain.

§ 416. When the pragmatist asserts that the function of the will and of the emotions is of the greatest importance in cognition, he is perfectly correct. Not only is almost all cognition governed by purposes which we strive to realise; but the will may also directly affect assent, provided we have that to which assent may be given.

A rational act [says Aquinas] can be considered in two ways first in reference to the exercise of the act and regarded from this aspect a rational act is always under the imperium of the will; and, secondly, in reference to the object, in respect of which there are two rational acts to be considered. The first is the apprehension of the truth about anything; and in this the intellect is not under our control. For it takes place by virtue of some natural or supernatural light. Hence in this respect the intellect is not under our control, nor is it under the imperium of the will. There is another rational act, however, in which the intellect assents to what it apprehends. If, therefore, what is apprehended is such that the intellect naturally assents to it (as in the case of first principles) such assent or dissent is not under our control, but pertains to the order of nature, and so, properly speaking, is subject to the imperium of nature. There are, however, certain things which we apprehend, which do not compel the intellect to assent or to dissent, or, at any rate, do not prevent it from withholding assent or dissent for some reason; and in such cases assent or dissent is under our control, and is subject to the imperium of the will.^{7}

It lies in our power, therefore, not only to think or not to think, but also to give or to withhold our assent in a doubtful case, i.e., in a case where the evidence ex parte objecti is insufficient to establish more than a claim to truth ; and if assent is given in such a case error is the probable, though not the necessary, result.

§ 417. There are various ways in which the will may influence assent. We may be unwilling to take the trouble to examine facts carefully or to think a difficult problem out. We may be in too great a hurry to pay due attention to the observation of facts and their circumstances, or too impatient in our desire to arrive at a solution to consider carefully the pros and cons. Or, again, we may be anxious to prove that we are right and somebody else is in the wrong. Many, indeed, and varied are the motives which may impel us to give a premature assent or dissent to a matter that has only been half thought out. Errors due to inadvertence and carelessness, to a kind of mental inertia, are the most common of all. Facts, inferences and postulates, though recognised as doubtful at first, tend to become 'true' by force of habit, if we do not examine and reject their claims. For what is constantly before the mind and yet is never examined comes to be regarded as part of the body of truths which have already been established and validated. Errors due to the emotions and to the 'will to believe' are also not uncommon, especially in cases where personal judgments are concerned, and where in consequence there is more scope for emotional likes and dislikes. We are inclined to believe what is derogatory to the character of an enemy and to disbelieve what is derogatory to the character of a friend. Again, if the object of our affection or dislike is a class or an institution, it is easy to generalise what we have found to the credit or to the discredit of one or two individuals, and to extend it to the whole. Moreover, errors due to the influence of these and other emotions tend to accumulate and so to form a system of erroneous beliefs which hang together and mutually support one another, and which, acting as bias or prejudice, prevent the apperception of what has in reality a valid claim to our assent. There can be no question but that there are systems of error as well as systems of truth; and systems of error are peculiarly difficult to get rid of; for though they must inevitably lead to contradiction somewhere, either internally or with facts, the emotions backed up by the system itself may be sufficiently powerful to cause us to slur over the contradiction, or, if it is a question of fact, to deny the fact that threatens to destroy our cherished beliefs.

It is needless, however, to develop further the unfortunate effect of the influxus voluntatis in the matter of truth. It is the formal cause of most, I think we may say of all, error; for strictly the disconformity between thought and reality in which error consists does not exist until the content of thought has been definitely referred to some object. The material causes of error account for the disconformity, but they do not account for the assent that accompanies it. We know that under abnormal conditions perception is unreliable, and we are also aware that doubtful inferences and hypothetical constructions may or may not correspond with their objects. In such cases, therefore, assent implies either inadvertence and carelessness -- a negative influxus of the will, as it were -- or else it is conditioned directly by emotional influences and the 'will to believe.'

§ 418. Nevertheless, the cognitive faculties per se and when functioning under normal conditions, are trustworthy; and in so far as this is the case their deliverances do not present merely a claim to truth, but they are true, since they are under these circumstances determined by the object itself. The criterion of truth here, therefore, is nothing more nor less than objective evidence. We assent because we are forced so to do by the object itself; because it is the object itself and not some other object or cause which seems to have determined the content of our thoughts, and so to have manifested itself to our mind. We assent because that to which we assent is 'obvious' and we cannot help assenting. Ridiculous and empty as this criterion of truth may seem to some, it is nevertheless the only criterion which the ordinary man uses, and the only ultimate criterion that there is. There are many negative criteria: we must not assent if we have reason for doubting the normal condition of the senses, if the formal laws of logic have been violated, or if we have contravened the norms of scientific method. But these criteria are criteria of error rather than of truth. Moreover, their value as means by which error may be detected presupposes the validity of the methods to which they pertain; and this we can only know because the truth of the principles on which they are based is objectively evident in the sense and under the conditions above explained.

§ 419. What has been said above of the trustworthiness of human faculties in general applies to my neighbours' as well as to my own. Consequently, although information gained from others, whether by word of mouth or by reading, at first presents only a claim to truth, we are justified in accepting the testimony of others in regard to facts, provided we have reason to believe that their observations were made with due care, and provided we have no reason to suspect any motive which would have led them to state what they knew to be false.

The theoretical constructions and explanations of others are in a different case; for in the construction of theory the possibility of error is so great that we cannot accept as authoritative the testimony of one of the workings of whose mind we are ignorant. We must examine the evidence for ourselves, and if we can detect no error, and both data and inferences seem to us 'objectively evident,' then we have rational grounds for giving our assent. We may apply as many negative criteria as we like, but in the end we shall come back to objective evidence, and shall ask ourselves is it or is it not clear to my mind that the reasonings of this author or this lecturer are valid, that his data are accurate, his inferences logical, his conclusions consistent with facts; in a word, is it obvious to me that the theory there before me is indirectly reality itself manifesting its nature to my mind, or does there seem to be mingled with it something subjective, something personal, something due merely to the constructive activity of the thinker who propounds the theory which claims my assent?

§ 420. The case is not very different when it is a question of verifying our own deductions or theoretical constructions. When we have taken due care to avoid mistakes, when the ordinary logical criteria and the special criteria and methods of the particular science with which we are concerned, have been applied, we are forced back finally upon objective evidence. Is it or is it not clear to my mind that the content of my thoughts has been determined by the object? Is the evidence sufficient to justify assent? Do my theoretical conclusions and the complexus of experimental facts harmonise sufficiently to justify me in giving my assent to the truth of the former, or have I still reasonable ground for doubt? These are the questions which I ask myself when judging of the truth of physical and other scientific theories, and according to the answer which the evidence compels me to make, I give or withhold my assent. The evidence in such a case is not, nor can it be, complete. The human mind is finite. It cannot analyse a theory into a complete system of propositions logically connected together and arranged in order of generality. It cannot deduce from each proposition all possible conclusions; nor can it discover all the facts with which these conclusions must coincide if the theory is to be a complete and adequate explanation. We cannot hope that any construction of ours shall in this sense be verified and validated, for that would imply omniscience. Nevertheless, granted that abstraction is not falsification and that systems of fact may be considered in comparative isolation without destroying their nature, we are justified in holding that theoretical constructions are true, provided the system of theoretical inferences coincides approximately with the system of known facts and nowhere leads to contradiction and truth here is real truth in spite of its approximate character and its incompleteness. For an approximate truth is not one that is subject to indefinite modification, but one that is permanently true within known limits of error. Such a truth expresses laws which really govern the operation of real things, and which really belong to the structure and plan of the universe, though there may be other laws which co-operate with these of the nature of which we are ignorant. Evidence, then, without being complete may be sufficient to justify assent, provided it is objective. And it is objective in the case of theoretical constructions, even though the latter may have begun as postulates due in part to the activity of mind. For our facts are objectively valid and our principle that like effects presuppose formaliter or eminenter a similar cause is also objectively valid. Hence in so far as our ideal laws are expressed in particular conclusions, the real laws which express themselves in particular facts that correspond to these conclusions must be in nature one and the same.

§ 421. In metaphysical theory our ultimate criterion is the same. Our last appeal is always to objective evidence. Many of the principles upon which metaphysics is based are, or should be, necessarily true. They should, as Mr. Bradley says, be such that in denying them we implicitly affirm their truth. Yet the test of necessary truth is not the psychological impossibility of thinking the contradictory, for we may still ask why this should be a test of truth. The final criterion is self-evidence, the self-manifestation of the nature of the object to a rational mind. From self-evident principles and from others that are less clearly self-evident we make deductions, which we may test by the laws of formal logic. But the validity of the laws of formal logic itself demands a criterion, and that can only be self-evidence. Finally, comparing principle with principle and deduction with deduction we ask whether they are consistent; or applying our theory in the concrete we enquire how far it explains or fails to explain or contradicts the data of experience. Should it manifest self-discrepancy or prove incompatible with facts we pronounce it erroneous; should it merely fail to explain we pronounce it inadequate and incomplete. But our criteria are not ultimate. Contradiction certainly shows that somewhere there is error. Indeed, contradiction is our only certain test of error, for inadvertence and the influence of subjective conditions do not necessarily invalidate a theory: they merely render it doubtful. But, on the other hand, the absence of contradiction does not establish truth. We cannot say that a thing is true simply because it is coherent; unless we take coherence in a very wide sense as including harmony with facts. And even then our criterion is not ultimate, for we may still ask (1) how we recognise facts; (2) how we recognise that a theory is coherent and does harmonise with facts; and (3) how we know that reality must be consistent and coherent and must not be contradictory and self-discrepant. And to all these questions our final answer must be the same, it is obvious or evident.

We have said that if a metaphysical theory is incoherent or incompatible with facts it must contain error. Can we say conversely that it is true if it is coherent and does harmonise with facts? This will depend upon what we understand by a metaphysical theory. If we understand by it a theory which shall explain all facts and co-ordinate all sciences, clearly there is no such theory. But if we understand by it a certain complexus of doctrines which explain facts in general, then these may be established as true. From the general fact of finite existence we may argue to the infinite, from the contingent to the necessary, from the caused to the uncaused or self-existent, and from the existence of a rational plan in the universe to an intellect which constructed and devised it. These arguments, I shall be told, are old-fashioned and at best only probable. That they are old-fashioned I admit; but that they are only probable I cannot grant, for to me their validity is evident. To discuss this question, however, would take us outside the theory of knowledge. What I wish here to point out, is that even if the arguments by which the existence of God is commonly proved are, when taken individually, only probable, they may yet be valid when taken together, for then they form not merely a sum of probabilities, but, as the Hegelian would say, a coherent whole. The universe is a coherent system of inter-related parts, and the only condition which can satisfy this fact, the only 'hypothesis' that can explain it, is the doctrine of the existence of a necessary Being from whom it proceeds by an unchanging yet creative act.

§ 422. What I have said above of objective evidence as the ultimate and in many cases the only criterion of truth is in substance what Newman has said in his Grammar of Assent of what he calls the sanction of the Illative Sense. "The course of inference," he says, "is ever more or less obscure, while assent is ever distinct and definite, and yet what is in its nature thus absolute does, in fact, follow upon what in outward manifestation is thus complex, indirect, and recondite." Hence he infers that we must take things as they are, and "instead of devising, what cannot be, some sufficient science of reasoning which may compel certitude in concrete conclusions, must confess that there is no ultimate test of truth besides the testimony born to truth by the mind itself."^{8} Newman treats the subject of truth psychologically, whereas I have treated it epistemologically. He does not ask 'how comes it about that we can be certain;'^{9} he accepts certitude as a fact and then enquires what is its nature, what its conditions, what the processes that lead up to it as a psychical act of the mind. Consequently he does not base certitude on objective evidence as I have done, but rather on the trustworthiness of the faculty of reason. Nevertheless, Newman maintains the certainty of knowledge and the real objective validity of knowledge as earnestly as does any realist. Certitude is a natural and normal state of the mind. It is not something to hide away and be ashamed of. Our very possession of certitude is a proof that it is not a weakness nor an absurdity to be certain.^{10} And the object of certitude and assent is something real, something the nature of which we apprehend; it is not a mere notion or symbol or word, the product of abstract and formal reasoning which is out of touch with reality.

§ 423. There is scarcely a single doctrine now upheld by the pragmatists which is not to be found verbally stated in the Grammar of Assent; yet Newman was not a pragmatist. His standpoint is psychological and human; nay, more, he acknowledges the personal element in truth; yet his standpoint is perfectly compatible with Realism because the results of his psychological analyses are not exaggerated. The real nature of truth is not confused with its pragmatic value. Product is not confused with process, content with intent, the various processes and methods by means of which truth is attained with the real objective validity of truth itself. He had as great a horror of verbal arguments as any pragmatist, yet he insists that certitude must follow on investigation and proof.^{11} He admits that certitude is accompanied by intellectual satisfaction and repose; yet any additional strength given to it by the emotions he regards as adventitious and accidental. He shows how knowledge is a growth, how in knowing a truth we are active, and how in order to understand and appreciate a truth we must live it; but the active recognition of propositions as true must be exercised at the bidding of reason, and reason never bids us be certain except on absolute proof.^{12} No merely formal or verbal argument is valid for Newman. And while, with Simmel, he acknowledges that in a series of reflex judgments each in turn may exercise a critical function towards those of the series that precede it; yet the certitude which results from this series of proofs when taken in the concrete -- not as a sum, but as a systematic whole -- is certitude, indefectible and irreversible.^{13} Lastly, everyone who reasons is his own centre, and no expedient for attaining a common measure of minds can reverse this truth, yet the personal action of the ratiocinative faculty, the perfection of which is called the Illative Sense, is capable of pronouncing a final judgment on the validity of inference in concrete matters such as may warrant that certitude is rightly elicited in favour of a proposition we infer.^{14}

§ 424. For Newman, as for all realists, there is only one ultimate and universal criterion of truth, the evidence which results from a careful examination and study of that which we wish to know; and as to the sufficiency of that evidence we are forced to trust and may rightly trust the deliverance of reason itself. Other criteria of error and of truth are useful and have methodological validity, but taken singly and in isolation from the living process of reasoning about objective real things, they are inadequate criteria. "Our enquiries," he says, "spontaneously fall into scientific sequence, and we think in logic as we talk in prose, without aiming at doing so." But such a tangible defence of what we hold, though it so fortifies and illustrates our holdings that it acts as a vivid apprehension acts, giving them luminousness and force, yet considered as an analysis of our ratiocination in its length and breadth, is necessarily inadequate.^{15} "Thought is too keen and manifold, its path too personal, delicate, and circuitous, its subject-matter too various and intricate, to admit of the trammels of any language, of whatever subtlety and of whatever compass."^{16} And precisely because thought is so subtle and intricate and personal, the human mind is capable of attaining truth and yet is incapable of proving the truth of what it holds by formal arguments to another who fails to appreciate the concrete evidence upon which that truth is based, and the concrete and intricate process of reasoning by which it has been established.

{1} Nature of Truth, p. 73.

{2} cf. Wallace, Hegel's Philosophy of Mind; and Humanism, pp. 45, 46. This objection is simply absurd, since it totally misrepresents the realist's doctrine.

{3} Cf. supra, § 47.

{4} cf. Professor Stout's excellent essay on 'Error' in Personal Idealism, with which I find myself almost wholly in agreement.

{5} Studies in Humanism, p. 185.

{6} Summa Theologica, 2. 2, q. 5, a. 4, ad 3.

{7} Summa Theologica, 1. 2, q. 17, a. 6.

{8} Grammar of Assent, p. 350. (Italics mine.)

{9} Ibid., p. 344.

{10} loc. cit., pp. 344, 209.

{11} Ibid., p. 258.

{12} Ibid., p. 345.

{13} Ibid., p.255.

{14} Ibid., p. 345.

{15} Ibid., pp. 286, 287.

{16} Ibid., p. 284.