The Third Antinomy represents the attempt of reason to apply the understanding’s category of causality to the totality of the synthesis of appearance. Kant had already argued in the Second Analogy that every event occurring within the series of appearances is subject to causal law. The issue for reason is whether the totality of the series is achieved in virtue of the occurrence of an event which is not subject to the causal laws that govern nature or is instead itself an infinite series of events, all of which are subject to those laws.
The thesis is not quite what one would expect, i.e., that there is an uncaused cause. Instead, the thesis is that there is a kind of cause different from “natural” causes, a “causality through freedom,” which is necessary to “explain” the appearances (A444/B472). The following is a reconstruction of the argument. It will be helpful if we first establish a lemma. (A lemma is the conclusion of an argument that in turn plays a key role in another argument. In the present case, the lemma is the conclusion of the first paragraph of the “proof” of the thesis.)
For a given cause of an effect, we may speak of the “causality of the cause” (Kausalität der Ursache). This “causality” is a state of the cause, previous to the effect “upon which it unfailingly follows according to a rule” (A444/B472).
With the lemma in hand the main argument may be stated.
Kant goes on in the Comment to note that this argument requires only one spontaneous causality, i.e., one that is needed for the series of all natural causes to begin. But he allows that once we grant that there is spontaneous causality in one case, it is permissible to allow spontaneous causality in other cases as well. He has in mind the idea that human will acts as a spontaneous cause of a limited series of events in the world. The spontaneous, non-natural causality of the human will is the basis for his doctrine of human freedom, to be developed later.
An objection immediately arises at this point. Suppose that state A is said to be a spontaneous cause of state B in an object O. Because state A is in time, it is subject to natural causality, and hence its existence is determined by an earlier state of O, in which case it is not really spontaneous.
This objection undermines the attempt to prove the freedom of anything but the original spontaneous cause of the world (which Kant compares to Aristotle’s prime mover). Kant had a vital interest in fighting the objection off, because transcendental freedom is the lynch-pin of his moral theory. Here he gives a way of preserving human freedom in the context of what he acknowledges to be an unsound argument for a first cause of all appearances in the world. But he is setting himself up for a defense of freedom along similar lines.
His strategy was to say that when spontaneous causes occur, “the determinative natural causes entirely cease above them,” so that although the causality of state A follows those states, it “does not result from them” and so can be called “an absolutely first beginning of a series of appearances” (A450/B578). For example, if I freely decide to rise from my chair, I do so “without the influence of natural causes” (since this is a spontaneous decision), and the states of myself prior to the decision have no causal influence on my decision at all. The old series of changes of states of appearances has been broken, and a new series has been initiated.
The doctrine of the antithesis is that there is no transcendental freedom: everything that occurs in the world does so through natural causality. The argument is much simpler than the thesis argument and can be reconstructed as follows.
In the Comment on the Third Antinomy, Kant makes some connections between the antithesis doctrine and other results he had already established or was to establish. In the solution to the First Antinomy, he had held that the world has no first beginning in time. If so, then why should it have a first cause? And in the First Analogy, he had claimed that substance has always existed. Why not, then, agree that the series of changes has always existed as well?
Kant adds that even if it is conceded that there is a first cause of the world, it would have to stand outside the time-series, not to mention outside of the reach of any perception. This claim is bold. But it is not as bad as placing a first cause in the series of changes, because that would destroy the coherence of experience to the extent that reality would not be distinguishable from dreaming. The laws of nature would be “altered incessantly by the influences of freedom, and the play of appearances—which according to mere nature would be regular and uniform—is thereby rendered confused and incoherent” (A451/B479).
The defense of the antithesis is made entirely on the basis of doctrines that Kant had incorporated into his principles of the understanding, so it is no surprise that Kant accepted the antithesis as being true. For example, Kant states that:
As for the principle concerning the thoroughgoing connection of all events in the world of sense according to immutable natural laws, its correctness is already established as a principle of the Transcendental Analytic and tolerates no impairment. (A536/B564, cf. A542/B570)Since the thesis directly contradicts the antithesis, the thesis is false. There is no transcendental freedom in the world of appearances. There is, to be sure, a “pure transcendental idea” of freedom, a creation of reason. But this idea has no reference in the world of appearances.
Transcendental freedom is now distinguished from practical freedom. A person acts through practical freedom when the act is not determined by any “sensible impulses” (A534/B562). Since sensible impulses are all natural causes—and are the only natural causes that influence human action—to act freely of sensible impulses is to act independently of natural causes. But this is what the antithesis appears to rule out. Kant claims that if transcendental freedom is not real, then practical freedom is not real either. Since practical freedom was for Kant central to moral action, transcendental freedom must be upheld in the face of the antithesis.
Kant begins his project of reconciling transcendental freedom and natural necessity by recalling the distinction between mathematical and dynamical categories. In the first two antinomies, mathematical categories are applied to appearances taken as a totality, and this application is always illegitimate. In the third antinomy, the dynamical category of causality is applied. Its principle is only regulative, not constitutive. The rule that every change has a cause applies without limit, since it does not require that an infinite synthesis take place, as with the first two antinomies (especially the first).
What is more important for present purposes is that because the causal principle is only regulative, “we can abstract also from the magnitude of the series of conditions, and what matters in their case is merely the dynamical relation between condition and conditioned” (A535-6/B563-4). This means that we can think of causality as being a character not only of appearances, but also of things in themselves. Here Kant claims that appearances “must themselves still have bases that are not appearances,” each of which basis (or ground) he calls an “intelligible cause” (A537/B565). These bases can be thought as free from all natural causality, which applies only to appearances.
On the other hand, if “we were to adhere obstinately to the reality of appearances,” there would be no relief from the inexorable march of natural causes. So the distinction between appearances and things in themselves, which is based in the ideality of space and time, provides an opening for a solution to the conflict between freedom and determinism, though only in broad outline at this point. “This distinction, when set forth in a universal way and quite abstractly, must appear extremely subtle and obscure, but it will become clear in its application” (A537/B565). (One might say that at least it will become more clear in its application: it is hard to make a case that the distinction is ever made very clear.) At any rate, Kant spends some twenty-one pages, in two titled but not numbered sections, elucidating his compatibilist thesis that natural necessity does not preclude transcendental freedom.
The human mind appears to itself in inner sense as a series of states. At the basis of this sensible or empirical self is something in itself, something which Kant calls “intelligible” (A538/B566). We can think this intelligible character of the mind as a “power which is not an object of sensible intuition but through which it can still be the cause of appearances” (A538/B566). The causality of the being that has this power could then be viewed as intelligible, while its effect is sensible.
There is no barrier whatsoever to attributing this power to what Kant calls the “transcendental object” that lies at the basis of the self.
For since these appearances are not things in themselves, they must be based on a transcendental object determining them as mere presentations; and hence nothing prevents us from attributing to this transcendental object, besides the property through which it appears, also a causality that is not appearance although its effect is nonetheless encountered in appearance. (A538/B566)The seat of this postulated causality in the transcendental object is its character, which is intelligible and “could also be called . . . the character of the thing in itself” (A539/B567). Such a character would not stand under the condition of time, which applies to appearances only, and so it is not subject to natural causality, which is bound to time.
Kant adds that not only may we think ourselves as having an intelligible character, we must think ourselves as having one.
We could not indeed, ever become acquainted with this intelligible character directly, because we cannot perceive anything except insofar as it appears; but we would still have to think it in accordance with the empirical character, just as in general we must—in thought—lay a transcendental object at the basis of appearances although we know nothing about this object as to what it is in itself. (A540/B568)The idea here is that the intelligible character would be the correlate of the empirical character, and that its intelligible causality would be the correlate of the empirical causality of appearances. We can only have the general concept of such a character as a noumenon or being of the understanding, but we can say that “insofar as this subject is noumenon, nothing occurs in it and there is found in it no change requiring dynamical time determination and hence no connection with appearances as causes” (A541/B569). Because the intelligible character of the subject is free from natural causality, “we would say quite correctly that it begins its effects in the world of sense on its own, without the action’s beginning in the subject itself” (A541/B569). (Insofar as the intelligible character is not determined in time, the action does not begin in the intelligible character.)
Before turning to Kant’s “Elucidation” of his reconciliation, a pause for reflection is in order. Kant had firmly set down in the chapter on Phenomena and Noumena that the concepts of the understanding have no signification apart from their application to experience. If we want to consider things apart from their spatio-temporal character, our thought of them as noumena is negative only.
Here, Kant seems to treat noumena in a sense that is positive. Yet this is not what he called “positive noumena” in the earlier chapter: that title was restricted to what would be objects of a non-sensible intuition, which we humans do not have. Still, Kant wants to attribute some character to the noumena, specifically the character of causality. Apart from time, all this category could mean is something, the existence of which brings about the existence of something else. This seems to be a meaningful concept, but Kant had claimed in the earlier chapter that “the concept would have no determination whatever as to how it fits any object” (A243/B301).
Perhaps there is no real conflict here. But there is another problem with Kant’s set-up that seems incapable of solution. Suppose it is granted that intelligible causes correspond to empirical causes, and that the action of the intelligible cause does not begin in the subject. This is not a sufficient reason to conclude that the action is begun by the subject “on its own.” There is no reason to think that the intelligible causality must be spontaneous, since it could be the effect of some other intelligible causality. We must interpret Kant as claiming only that an intelligible cause can be spontaneous, at least for all we know.
In the case of the human being, Kant believed that we must postulate an intelligible causality. Although the various determinable states of the self in time are appearances, the human mind acts in ways that cannot be presented in inner sense. Specifically, the understanding (power of judgment) and reason act in such a way that “cannot be classed at all with the receptivity of sensibility” (A546/B574). From the very start of the Critique, Kant had distinguished the spontaneity of the understanding from the receptivity of sensibility. Now he adds that reason determines the understanding and so is even more fundamentally spontaneous in its actions.
But what makes it “evident” that reason has causal powers is that it lays down rules, imperatives dictating what ought to be. This cannot be done by the understanding, which is concerned only with what is, what occurs in nature. “We cannot ask what ought to happen in nature, any more than what properties a circle ought to have, but can ask only what happens in nature, or what properties the circle has” (A457/B575). The basis of action is a mere concept of how things ought to be, and this is something that can never be the product of nature. Reason “makes for itself an order of its own according to ideas” (A548/B576).
Still, the “power of choice” (A549/B577) which acts on reason’s imperatives has an empirical character corresponding to its intelligible character. This character is in the grips of natural causality, and so determines every action of a human being. “If we could explore all appearances of his power of choice down to the bottom, there would not be a single human action that we could not with certainty predict and cognize as necessary from its preceding conditions” (A550/B578). The actions of the subject of empirical psychology, then, are deterministic. (This claim had been made by Hume, though with a weaker notion of causality in mind. See An Enquiry Concerning Human Understanding, Section 8.)
So, the same action might have both an intelligible cause (one acted because reason dictated that he ought to) and an empirical cause (one acted because one was in a certain prior state). Kant held that if reason does have this causal power, it can be said to act freely, despite the empirical determinism. The action would be the outcome of two different causalities, one intelligible and one empirical. We may never know through experience what the intelligible cause is, and thus whether the action is blameworthy, praiseworthy, or morally neutral.
At this point, we can see that Kant’s reconciliation of freedom and determinism is different from that proposed on the side of the thesis. The dogmatic doctrine required that a free cause initiate a new series of appearances by breaking off the old series and beginning anew. Kant’s critical doctrine keeps the old series intact, although he describes the situation similarly to the dogmatist’s way, as “the power of reason to begin on its own a series of events” (A554/B581). This possible beginning of action could only be a redundant beginning, though, because the events prior to that series must be said to begin it.
To illustrate his compatibility thesis, Kant uses the case of someone who has told a harmful malicious lie. There may be an explanation of this action by reference to the person’s past circumstances. “But although we believe the action to be determined by these causes, we nevertheless blame the perpetrator” (A555/B583). We do this because we regard his action as subject to an intelligible causality in his intelligible character—a causality which we take to be entirely free. This is why we say that the liar could have refrained from lying, even though the whole course of his circumstances move him toward lying. We disregard his past and regard the act of lying “as if the perpetrator starts with it a series of consequences completely on his own” (AA555/B583).
In imputing freedom to human action, we do not thereby make any claim about how the intelligible and empirical characters are related to each other. “In judging free actions with regard to their causality we can get only as far as the intelligible cause, but not beyond it” (A557/B585). This is of no concern to Kant here, since all he was trying to do was to establish the possibility of freedom in the face of natural determinism. And this possibility is only a logical, not a real possibility. “For this also would not have succeeded, because in general we cannot from mere a priori concepts cognize the possibility of any real basis and any causality” (A558/B586).
Kant’s doctrine of the compatibility of human freedom and natural determinism has sparked continuing debate since he proposed it. It is faced with various problems. Suppose that the liar was naturally determined to tell his lie. He could not avoid doing so, at least in one sense. Then of what value is his transcendental freedom? Even if his use of reason was itself not determined by anything in time, it still could not have produced anything but the lie without destroying the coherence of experience. Yet Kant was, as we saw, absolutely adamant about the need to preserve the coherence of experience. So the possibility of doing otherwise embodied in transcendental freedom turns out to be a logical, not a real possibility.
The dispute in the Fourth Antinomy concerns the existence in the world of an absolutely necessary being. The argument of the thesis is a version of the so-called “cosmological argument,” which was most famously used in modern philosophy by Leibniz. Reason demands that the series of contingent beings terminate in a being that is necessary. However, since the antinomies involve only totality of the synthesis of appearances, this being must be within the world itself. The pure
version of the cosmological argument leaves it open as to whether the necessary being is in the world or outside of it. The subordinate version which purports to prove the existence of God as a transcendent being, and so not merely the totality of the world, is treated in the next chapter.
One point raised in the Comment on the thesis is worth mentioning. What gets the argument started is the fact that changes in appearances are empirically contingent, which means that they are dependent on prior states of objects. “Change proves only empirical contingency, i.e., it proves only that the new state by itself, without a cause that belongs to the previous time, could not—in consequence of the law of causality—have taken place at all” (A460/B488). Because no state of an object in the series is not empirically contingent, philosophers looked elsewhere to complete the series.
Thus, they appealed to the pure concept of contingency, “that whose contradictory opposite is possible” (A460/B488). That is, they regarded the series of appearances as intelligible and thus subject to the pure concept of contingency. In this sense, each state of the objects, as intelligible, might not have been. To explain why they exist, it is required to postulate a necessary being. Kant notes that from the empirical concept of contingency, the intelligible cannot be inferred. Change does not prove the possibility of the opposite of the present state at the present time. The state at a later time is not an opposite state of the thing at the present time. In fact, all empirically contingent states of things are hypothetically necessary.
Opposing the thesis is the claim that nature is uniform in the sense that all appearances are contingent. The argument here is essentially the same as for the antithesis of the Third Antinomy. A necessary being in the world would be an uncaused cause, which conflicts with the rules for the unity of experience. A necessary being outside the world would have to begin to act, in which case it is not outside the world after all.
The solution to the Fourth Antinomy is similar to that of the Third. The antithesis is true of all appearances: they are one and all empirically contingent. If we regard things in themselves, however, there is the possibility of a necessary being. But this necessary being could not be a contingent being that is considered in itself. Rather, it would have to lie entirely outside the series, because it would have to be a condition of the whole series.
[ Next Lecture | Philosophy 175 Home Page | Lexicon ]