II.1. The Discipline of Pure Reason in regard to Hypotheses and Proofs

Since criticism of our reason has at last taught us that we cannot by means of its pure and speculative employment arrive at any knowledge whatsoever, may it not seem that a proportionately wider field is opened for hypotheses For are we not at liberty, where we cannot make assertions, at least to invent theories and to have opinions?

If the imagination is not simply to be visionary, but is to be inventive under the strict surveillance of reason, there must always previously be something that is completely certain, and not invented or merely a matter of opinion, namely, the possibility of the object itself. Once that is established, it is then permissible to have recourse to opinion in regard to its actuality; but this opinion, if it is not to be groundless, must be brought into connection with what is actually given and so far certain, as serving to account for what is thus given. Then, and only then, can the supposition be entitled an hypothesis.

As we cannot form the least conception a priori of the possibility of dynamical connection, and as the categories of the pure understanding do not suffice for devising any such conception, but only for apprehending it when met with in experience, we cannot, in accordance with these categories, creatively imagine any object in terms of any new quality that does not allow of being given in experience; and we cannot, therefore, make use of such an object in any legitimate hypothesis; otherwise we should be resting reason on empty figments of the brain,’ and not on concepts of things. Thus it is not permissible to invent any new original powers, as, for instance, an understanding capable of intuiting its objects without the aid of senses; or a force of attraction without any contact; or a new kind of substance existing in space and yet not impenetrable. Nor is it legitimate to postulate a form of communion of substances which is different from any revealed in experience, a presence that is not spatial, a duration that is not temporal. In a word, our reason can employ as conditions of the possibility of things only the conditions of possible experience; it can never proceed to form concepts of things quite independently of these conditions. Such concepts, though not self-contradictory, would be without an object.

The concepts of reason are, as we have said, mere ideas, and have no object that can be met with in any experience. None the less they do not on this account signify objects that having been invented are thereupon assumed to be possible. They are thought only problematically, in order that upon them (as heuristic fictions), we may base regulative principles of the systematic employment of the understanding in the field of experience. Save in this connection they are merely thought-entities, the possibility of which is not demonstrable, and which therefore do not allow of being employed, in the character of hypotheses, in explanation of the actual appearances. It is quite permissible to think the soul as simple, in order, in conformity with this idea, to employ as the principle of our interpretation of its inner appearances a complete and necessary unity of all its faculties; and this in spite of the fact that this unity is such as can never be apprehended in concreto. But to assume the soul as a simple substance (a transcendent concept), would be [to propound] a proposition which is not only indemonstrable—as is the case with many physical hypotheses—but is hazarded in a quite blind and arbitrary fashion. For the simple can never be met with in any experience whatsoever; and if by substance be here meant the permanent object of sensible intuition, the possibility of a simple appearance is quite incomprehensible. Reason does not afford any sufficient ground for assuming, [even] as a matter of opinion, merely intelligible beings, or merely intelligible properties of things belonging to the sensible world, although (as we have no concepts of their possibility or impossibility) we also cannot lay claim to any insight that justifies us in dogmatically denying them.

In the explanation of given appearances, no things or grounds of explanation can be adduced other than those which have been found to stand in connection with given appearances in accordance with the already known laws of the appearances. A transcendental hypothesis, in which a mere idea of reason is used in explanation of natural existences, would really be no explanation; so to proceed would be to explain something, which in terms of known empirical principles we do not understand sufficiently, by something which we do not understand at all. Moreover, the principle of such an hypothesis would at most serve only for the satisfaction of reason, not for the furtherance of the employment of the understanding in respect of objects. Order and purposiveness in nature must themselves be explained from natural grounds and according to natural laws; and the wildest hypotheses, if only they are physical, are here more tolerable than a hyperphysical hypothesis, such as the appeal to a divine Author, assumed simply in order that we may have an explanation. That would be a principle of ignava ratio; for we should be passing over all causes the objective reality of which, at least as regards their possibility, can be ascertained in the course of experience, in order to rest in a mere idea—an idea that is very comforting to reason. As regards the absolute totality of the ground of explanation of the series of these causes, such totality need suggest no difficulty in respect of natural existences; since these existences are nothing but appearances, we need never look to them for any kind of completeness in the synthesis of the series of conditions.

It can never be permissible, in the speculative employment of reason, to resort to transcendental hypotheses, and to presume that we can make good the lack of physical grounds of explanation by appealing to the hyperphysical. The objection to such procedure is twofold: partly, that reason, so far from being in the least advanced thereby, is cut off from all progress in its own employment; partly, that this license would in the end deprive reason of all the fruits that spring from the cultivation of its own proper domain, namely, that of experience. For whenever the explanation of natural existences is found to be difficult, there is always at hand a transcendental ground of explanation which relieves us from further investigation, and our enquiry is brought to an end not through insight, but by the aid of a principle which while utterly incomprehensible has from the start been so constructed as necessarily to contain the concept of what is absolutely primordial.

The second requirement for the admissibility of an hypothesis is its adequacy in accounting a priori for those consequences which are [de facto] given. If for this purpose we have to call in auxiliary hypotheses, they give rise to the suspicion that they are mere fictions; for each of them requires the same justification as is necessary in the case of the fundamental hypothesis, and they are not, therefore, in a position to bear reliable testimony. If we assume an absolutely perfect cause, we need not be at a loss in explaining the purposiveness, order, and vastness which are displayed in the world; but in view of what, judged at least by our concepts, are the obvious deviations and evils, other new hypotheses are required in order to uphold the original hypothesis in face of the objections which these suggest. If the simple self-sufficiency of the human soul has been employed to account for its appearances, it is controverted by certain difficulties, due to those phenomena which are similar to the changes that take place in matter (growth and decay), and we have therefore to seek the aid of new hypotheses, which are not indeed without plausibility, but which yet have no credentials save what is conferred upon them by that opinion—the fundamental hypothesis—which they have themselves been called in to support.

If the instances here cited as examples of the assertions made by reason—the incorporeal unity of the soul and the existence of a supreme being—are propounded not as hypotheses, but as dogmas proved a priori, I am not at present concerned with them, save to remark that in that case care must be taken that the proof has the apodeictic certainty of a demonstration. For to set out to show no more than that the reality of such ideas is probable is as absurd as to think of proving a proposition of geometry merely as a probability. Reason, when employed apart from all experience, can know propositions entirely a priori, and as necessary, or it can know nothing at all. Its judgments, therefore, are never opinions; either it must abstain from all judgment, or must affirm with apodeictic certainty. Opinions and probable judgments as to what belongs to things can be propounded only in explanation of what is actually given, or as consequences that follow in accordance with empirical laws from what underlies the actually given. They are therefore concerned only with the series of the objects of experience. Outside this field, to form opinions is merely to play with thoughts. For we should then have to presuppose yet another opinion—the opinion that we may perhaps arrive at the truth by a road that is uncertain.

But although, in dealing with the merely speculative questions of pure reason, hypotheses are not available for the purposes of basing propositions upon them, they are yet entirely permissible for the purposes of defending propositions; that is to say, they may not be employed in any dogmatic, but only in polemical fashion. By the defence of propositions I do not mean the addition of fresh grounds for their assertion, but merely the nullifying of the sophistical arguments by which our opponent professes to invalidate this assertion. Now all synthetic propositions of pure reason have this peculiarity, that while in asserting the reality of this or that idea we can never have knowledge sufficient to give certainty to our proposition, our opponent is just as little able to assert the opposite. This equality of fortune [in the ventures] of human reason does not, in speculative modes of knowledge, favour either of the two parties, and it is consequently the fitting battle-ground for their never-ending feuds. But as will be shown, reason has, in respect of its practical employment, the right to postulate what in the field of mere speculation it can have no kind of right to assume without sufficient proof. For while all such assumptions do violence to [the principle of] completeness of speculation, that is a principle with which the practical interest is not at all concerned. In the practical sphere reason has rights of possession, of which it does not require to offer proof, and of which, in fact, it could not supply proof. The burden of proof accordingly rests upon the opponent. But since the latter knows just as little of the object under question, in trying to prove its non-existence, as does the former in maintaining its reality, it is evident that the former, who is asserting something as a practically necessary supposition, is at an advantage (melior est conditio possidentis). For he is at liberty to employ, as it were in self-defence, on behalf of his own good cause, the very same weapons that his opponent employs against that cause, that is, hypotheses. These are not intended to strengthen the proof of his position, but only to show that the opposing party has much too little understanding of the matter in dispute to allow of his flattering himself that he has the advantage in respect of speculative insight.

Hypotheses are therefore, in the domain of pure reason, permissible only as weapons of war, and only for the purpose of defending a right, not in order to establish it. But the opposing party we must always look for in ourselves. For speculative reason in its transcendental employment is in itself dialectical; the objections which we have to fear lie in ourselves. We must seek them out, just as we would do in the case of claims that, while old, have never become superannuated, in order that by annulling them we may establish a permanent peace. External quiescence is merely specious. The root of these disturbances, which lies deep in the nature of human reason, must be removed. But how can we do so unless we give it freedom, nay, nourishment, to send out shoots so that it may discover itself to our eyes, and that it may then be entirely destroyed? We must, therefore, bethink ourselves of objections which have never yet occurred to any opponent, and indeed lend him our weapons, and grant him the most favourable position which he could possibly desire. We have nothing to fear in all this, but much to hope for; namely, that we may gain for ourselves a possession which can never again be contested.

Thus for our complete equipment we require among other things the hypotheses of pure reason. For although they are but leaden weapons, since they are not steeled by any law of experience, they are yet as effective as those which our opponents can employ against us. If, therefore, having assumed (in some non-speculative connection) the nature of the soul to be immaterial and not subject to any corporeal change, we are met by the difficulty that nevertheless experience seems to prove that the exaltation and the derangement of our mental powers are alike in being merely diverse modifications of our organs, we can weaken the force of this proof by postulating that our body may be nothing more than a fundamental appearance which in this our present state (in this life) serves as a condition of our whole faculty of sensibility, and therewith of all our thought, and that separation from the body may therefore be regarded as the end of this sensible employment of our faculty of knowledge and the beginning of its intellectual employment. Thus regarded, the body would not be the cause of thought, but merely a restrictive condition of it, and therefore, while indeed furthering the sensible and animal life, it would because of this very fact have to be considered a hindrance to the pure and spiritual life. The dependence of the animal and sensible upon the bodily constitution would then in nowise prove the dependence of our entire life upon the state of our organs. We might go yet further, and discover quite new objections, which either have never been suggested or have never been sufficiently developed.

Generation, in man as in non-rational creatures, is dependent upon opportunity, often indeed upon sufficiency of food, upon the moods and caprices of rulers, nay, even upon vice. And this makes it very difficult to suppose that a creature whose life has its first beginning in circumstances so trivial and so entirely dependent upon our own choice, should have an existence that extends to all eternity. As regards the continuance (here on earth) of the species as a whole, this difficulty is negligible, since accident in the individual case is still subject to a general law, but as regards each individual it certainly seems highly questionable to expect so potent an effect from causes so insignificant. But to meet these objections we can propound a transcendental hypothesis, namely, that all life is, strictly speaking, intelligible only, is not subject to changes of time, and neither begins in birth nor ends in death; that this life is an appearance only, that is, a sensible representation of the purely spiritual life, and that the whole sensible world is a mere picture which in our present mode of knowledge hovers before us, and like a dream has in itself no objective reality; that if we could intuit ourselves and things as they are, we should see ourselves in a world of spiritual beings, our sole and true community with which has not begun through birth and will not cease through bodily death—both birth and death being mere appearances.

Now of all this we have not the least knowledge. We plead it only in hypothetical fashion, to meet the attack; we are not actually asserting it. For it is not even an idea of reason, but is a concept devised merely for the purposes of self-defence. None the less we are here proceeding in entire conformity with reason. Our opponent falsely represents the absence of empirical conditions as itself amounting to proof of the total impossibility of our belief, and is therefore proceeding on the assumption that he has exhausted all the possibilities. What we are doing is merely to show that it is just as little possible for him to comprehend the whole field of possible things through mere laws of experience as it is for us to reach, outside experience, any conclusions justifiable for our reason. Anyone who employs such hypothetical means of defence against the rash and presumptuous negations of his opponent must not be considered to intend the adoption of these opinions as his own; he abandons them, as soon as he has disposed of the dogmatic pretensions of his opponent. For though a merely negative attitude to the assertions of others may seem very modest and moderate, to proceed to represent the objections to an assertion as proofs of the counter-assertion is to make claims no less presumptuous and visionary than if the positive position and its affirmations had been adopted.

It is evident, therefore, that in the speculative employment of reason hypotheses, regarded as opinions, have no validity in themselves, but only relatively to the transcendent pretensions of the opposite party. For to make principles of possible experience conditions of the possibility of things in general is just as transcendent a procedure as to assert the objective reality of [transcendent] concepts, the objects of which cannot be found anywhere save outside the limits of all possible experience. What pure reason judges assertorically must (like everything that reason knows) be necessary; otherwise nothing at all is asserted. Accordingly, pure reason does not, in point of fact, contain any opinions whatsoever. The hypotheses, above referred to, are merely problematic judgments, which at least cannot be refuted, although they do not indeed allow of any proof. They are therefore nothing but private opinions. Nevertheless, we cannot properly dispense with them as weapons against the misgivings which are apt to occur; they are necessary even to secure our inner tranquillity. We must preserve to them this character, carefully guarding against the assumption of their independent authority or absolute validity, since otherwise they would drown reason in fictions and delusions.

What distinguishes the proofs of transcendental synthetic propositions from all other proofs which yield an a priori synthetic knowledge is that, in the case of the former, reason may not apply itself, by means of its concepts, directly to the object, but must first establish the objective validity of the concepts and the possibility of their a priori synthesis. This rule is not made necessary merely by considerations of prudence, but is essential to the very possibility of the proofs themselves. If I am to pass a priori beyond the concept of an object, I can do so only with the help of some special guidance, supplied from outside this concept. In mathematics it is a priori intuition which guides my synthesis; and thereby all our conclusions can be drawn immediately from pure intuition. In transcendental knowledge, so long as we are concerned only with concepts of the understanding, our guide is the possibility of experience. Such proof does not show that the given concept (for instance, of that which happens) leads directly to another concept (that of a cause); for such a transition would be a saltus which could not be justified. The proof proceeds by showing that experience itself, and therefore the object of experience, would be impossible without a connection of this kind. Accordingly, the proof must also at the same time show the possibility of arriving synthetically and a priori at some knowledge of things which was not contained in the concepts of them. Unless this requirement be met, the proofs, like streams which break their banks, run wildly at random, whithersoever the current of hidden association may chance to lead them. The semblance of conviction which rests upon subjective causes of association, and which is regarded as insight into a natural affinity, cannot balance the misgivings to which so hazardous a course must rightly give rise. On this account, all attempts to prove the principle of sufficient reason have, by the universal admission of those concerned, been fruitless; and prior to our own transcendental criticism it was considered better, since that principle could not be surrendered, boldly to appeal to the common sense of mankind—an expedient which always is a sign that the cause of reason is in desperate straits—rather than to attempt new dogmatic proofs.

But if the proposition to be proved is an assertion of pure reason, and if I am therefore proposing to pass beyond my empirical concepts by means of mere ideas, justification of such a step in synthesis (supposing it to be possible) is all the more necessary as a precondition of any attempt to prove the proposition itself. However plausible the alleged proof of the simple nature of our thinking substance, derived from the unity of apperception, may be, it is faced by the unavoidable difficulty, that since the [notion of] absolute simplicity is not a concept which can be immediately related to a perception, but, as an idea, would have to be inferred, there can be no understanding how the bare consciousness (which is, or at least can be, contained in all thought), though it is indeed so far a simple representation, should conduct us to the consciousness and the knowledge of a thing in which thought alone can be contained. If I represent to myself the power of a body in motion, it is so far for me absolute unity, and my representation of it is simple; and I can therefore express this representation by the motion of a point—for the volume of the body is not here a relevant consideration, and can be thought, without diminution of the moving power, as small as we please, and therefore even as existing in a point. But I may not therefore conclude that if nothing be given to me but the moving power of a body, the body can be thought as simple substance—merely because its representation abstracts from the magnitude of its volume and is consequently simple. The simple arrived at by abstraction is entirely different from the simple as an object; though the ‘I’, taken in abstraction, can contain in itself no manifold, in its other meaning, as signifying the soul itself, it can be a highly complex concept, as containing under itself, and as denoting, what is very composite. I thus detect in these arguments a paralogism. But in order to be armed against this paralogism (for without some forewarning we should not entertain any suspicion in regard to the proof), it is indispensably necessary to have constantly at hand a criterion of the possibility of those synthetic propositions which are intended to prove more than experience yields. This criterion consists in the requirement that proof should not proceed directly to the desired predicate but only by means of a principle that will demonstrate the possibility of extending our given concept in an a priori manner to ideas, and of realising the latter. If this precaution be always observed, if before attempting any proof, we discreetly take thought as to how, and with what ground for hope, we may expect such an extension through pure reason, and whence, in such a case, this insight, which is not developed from concepts, and also cannot be anticipated in reference to any possible experience, is yet to be derived, we can by so doing spare ourselves much difficult and yet fruitless labour, not expecting from reason what obviously exceeds its power—or rather, since reason, when obsessed by passionate desire for the speculative enlargement of its domain, is not easily to be restrained, by subjecting it to the discipline of self-control.

The first rule is, therefore, not to attempt any transcendental proofs until we have considered, with a view to obtaining justification for them, from what source we propose to derive the principles on which the proofs are to be based, and with what right we may expect success in our inferences. If they are principles of the understanding (for instance, that of causality), it is useless to attempt, by means of them, to attain to ideas of pure reason; such principles are valid only for objects of possible experience. If they are principles of pure reason, it is again labour lost. Reason has indeed principles of its own; but regarded as objective principles, they are one and all dialectical, and can have no validity save as regulative principles for its employment in experience, with a view to making experience systematically coherent. But if such professed proofs are propounded, we must meet their deceptive power of persuasion with the non liquet of our matured judgment; and although we may not be able to detect the illusion involved, we are yet entirely within our rights in demanding a deduction of the principles employed in them; and if these principles have their source in reason alone, the demand is one which can never be met. And there is thus no need for us to concern ourselves with the particular nature and with the refutation of each and every groundless illusion; at the tribunal of a critical reason, which insists upon laws, this entire dialectic, so inexhaustible in its artifices, can be disposed of in bulk.

The second peculiarity of transcendental proofs is that only one proof can be found for each transcendental proposition. If I am inferring not from concepts but from the intuition which corresponds to a concept, be it a pure intuition as in mathematics, or an empirical intuition as in natural science, the intuition which serves as the basis of the inference supplies me with manifold material for synthetic propositions, material which I can connect in more than one way, so that, as it is permissible for me to start from more than one point, I can arrive at the same proposition by different paths.

In the case of transcendental propositions, however, we start always from one concept only, and assert the synthetic condition of the possibility of the object in accordance with this concept. Since outside this concept there is nothing further through which the object could be determined, there can therefore be only one ground of proof. The proof can contain nothing more than the determination of an object in general in accordance with this one single concept. In the Transcendental Analytic, for instance, we derived the principle that everything which happens has a cause, from the condition under which alone a concept of happening in general is objectively possible—namely, by showing that the determination of an event in time, and therefore the event as belonging to experience, would be impossible save as standing under such a dynamical rule. This is the sole possible ground of proof; for the event, in being represented, has objective validity, that is, truth, only in so far as an object is determined for the concept by means of the law of causality. Other proofs of this principle have, indeed, been attempted, for instance, from the contingency [of that which happens]. But on examining this argument, we can discover no mark of contingency save only the happenings that is, the existence of the object preceded by its non-existence, and thus are brought back to the same ground of proof as before. Similarly, if the proposition, that everything which thinks is simple, is to be proved, we leave out of account the manifold of thought, and hold only to the concept of the ‘I’, which is simple and to which all thought is related. The same is true of the transcendental proof of the existence of God; it is based solely on the coincidence, of the concepts of the most real being and of necessary being, and is not to be looked for anywhere else.

This caution reduces the criticism of the assertions of reason to very small compass. When reason is conducting its business through concepts only, there is but one possible proof, if, that is to say, there be any possible proof at all. If, therefore, we observe the dogmatist coming forward with ten proofs, we can be quite sure that he really has none. For had he one that yielded—as must always be required in matters of pure reason—apodeictic proof, what need would he have of the others? His purpose can only be that of the parliamentary advocate, who intends his various arguments for different groups, in order to take advantage of the weakness of those before whom he is pleading—shearers who, without entering deeply into the matter, desire to be soon quit of it, and therefore seize upon whatever may first happen to attract their attention, and decide accordingly.

The third rule peculiar to pure reason, in so far as it is to be subjected to a discipline in respect of transcendental proofs, is that its proofs must never be apagogical, but always ostensive. The direct or ostensive proof, in every kind of knowledge, is that which combines with the conviction of its truth insight into the sources of its truth; the apagogical proof, on the other hand, while it can indeed yield certainty, cannot enable us to comprehend truth in its connection with the grounds of its possibility. The latter is therefore to be regarded rather as a last resort than as a mode of procedure which satisfies all the requirements of reason. In respect of convincing power, it has, however, this advantage over the direct proofs, that contradiction always carries with it more clearness of representation than the best connection, and so approximates to the intuitional certainty of a demonstration.

The real reason why apagogical proofs are employed in various sciences would seem to be this. When the grounds from which this or that knowledge has to be derived are too numerous or too deeply concealed, we try whether we may not arrive at the knowledge in question through its consequences. Now this modus ponens, that is, the inference to the truth of an assertion from the truth of its consequences, is only permissible when all its possible consequences are [known to be] true; for in that case there is only one possible ground for this being so, and that ground must also be true. But this procedure is impracticable; to discover all possible consequences of any given proposition exceeds our powers. None the less this mode of reasoning is resorted to, although indeed with a certain special modification, when we endeavour to prove something merely as an hypothesis. The modification made is that we admit the conclusion as holding according to analogy, namely, on the ground that if all the many consequences examined by us agree with an assumed ground, all other possible consequences will also agree with it. But from the nature of the argument, it is obvious that an hypothesis can never, on such evidence, be transformed into demonstrated truth. The modus tollens of reasoning, which proceeds from consequences to their grounds, is not only a quite rigorous but also an extremely easy mode of proof. For if even a single false consequence can be drawn from a proposition, the proposition is itself false. Instead, then, as in an ostensive proof, of reviewing the whole series of grounds that can lead us to the truth of a proposition, by means of a complete insight into its possibility, we require only to show that a single one of the consequences resulting from its opposite is false, in order to prove that this opposite is itself false, and that the proposition which we had to prove is therefore true.

The apagogic method of proof is, however, permissible only in those sciences where it is impossible mistakenly to substitute what is subjective in our representations for what is objective, that is, for the knowledge of that which is in the object. Where such substitution tends to occur, it must often happen that the opposite of a given proposition contradicts only the subjective conditions of thought, and not the object, or that the two propositions contradict each other only under a subjective condition which is falsely treated as being objective; the condition being false, both can be false, without it being possible to infer from the falsity of the one to the truth of the other.

In mathematics this subreption is impossible; and it is there, therefore, that apagogical proofs have their true place. In natural science, where all our knowledge is based upon empirical intuitions, the subreption can generally be guarded against through repeated comparison of observations; but in this field this mode of proof is for the most part of little importance. The transcendental enterprises of pure reason, however, are one and all carried on within the domain proper to dialectical illusion, that is, within the domain of the subjective, which in its premisses presents itself to reason, nay, forces itself upon reason, as being objective. In this field, therefore, it can never be permissible, so far as synthetic propositions are concerned, to justify assertions by disproving their opposite. For either this refutation is nothing but the mere representation of the conflict of the opposite opinion with the subjective conditions under which alone anything can be conceived by our reason, which does not in the least contribute to the disproof of the thing itself—just as, for instance, we must recognise that while the unconditioned necessity of the existence of a being is altogether inconceivable to us, and that every speculative proof of a necessary supreme being is therefore rightly to be opposed on subjective grounds, we have yet no right to deny the possibility of such a primordial being in itself—or else both parties, those who adopt the affirmative no less than those who adopt the negative position, have been deceived by transcendental illusion, and base their assertions upon an impossible concept of the object. In that case we can apply the rule: non entis nulla sunt predicata; that is, all that is asserted of the object, whether affirmatively or negatively, is erroneous, and consequently we cannot arrive apagogically at knowledge of the truth through refutation of the opposite. If, for instance, it be assumed that the sensible world is given in itself in its totality, it is false that it must be either infinite in space or finite and limited. Both contentions are false. For appearances (as mere representations) which yet are to be given in themselves (as objects) are something impossible; and though the infinitude of this imaginary whole would indeed be unconditioned, it would contradict (since everything in appearances is conditioned) the unconditioned determination of magnitude, [that is, of totality], which is presupposed in the concept.

The apagogic method of proof is the real deluding influence by which those who reason dogmatically have always held their admirers. It may be compared to a champion who seeks to uphold the honour and incontestable rights of his adopted party by offering battle to all who would question them. Such boasting proves nothing, however, in regard to the merits of the issue but only in regard to the respective strength of the combatants, and this indeed only in respect of those who take the offensive. The spectators, observing that each party is alternately conqueror and conquered, are often led to have sceptical doubts in regard to the very object of the dispute. They are not, however, justified in adopting such an attitude; it is sufficient to declare to the combatants: non defensoribus istis tempus eget. Everyone must defend his position directly, by a legitimate proof that carries with it a transcendental deduction of the grounds upon which it is itself made to rest. Only when this has been done, are we in a position to decide how far its claims allow of rational justification. If an opponent relies on subjective grounds, it is an easy matter to refute him. The dogmatist cannot, however, profit by this advantage. His own judgments are, as a rule, no less dependent upon subjective influences; and he can himself in turn be similarly cornered by his opponent. But if both parties proceed by the direct method, either they will soon discover the difficulty, nay, the impossibility, of showing ground for their assertions, and will be left with no resort save to appeal to some form of prescriptive authority; or our criticism will easily discover the illusion to which their dogmatic procedure is due, compelling pure reason to relinquish its exaggerated pretensions in the realm of speculation, and to withdraw within the limits of its proper territory—that of practical principles.