1. Introduction
2.1 The Problem of the Third Antinomy
國
立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
2.1 The Problem of the Third Antinomy
There are four antinomies in Transzendentale Dialektik, KrV: the antinomy on the limitation of the universe in space and time, the antinomy of atomism (mereology:
whole-parts problem), the antinomy of the uncaused cause and the universal causal exhaustion (freedom) and the antinomy of the existence of the necessary being (God).
All of them attempt to demonstrate the inadequacy of the assumption of
transcendental realism, i.e., the assumption that space and time are the ontological form in the world itself while the cause of cognition dwells within that space and time; the conflict in each antinomy is the consequence of the assumption. If the thesis and the anti-thesis of each antinomy are modified in transcendental idealism, i.e., space and time are assumed to be mere forms of our intuition (sensibility) while the cause of cognition is not in time and space, the antinomy will be resolved in the sense that (1) in the former two instances (those concerning the object of sensibility and the theoretical employment of reason: mathematical antinomies), both their thesis and anti-thesis are rejected so that the antinomy no longer remains and (2) in the latter two instances (in relation to the practical employment of reason that transcends the boundary of sensibility: dynamical antinomies), both their thesis and anti-thesis may survive the challenge of the self-critics of reason with the condition of all
possible cognitions. Each antinomy is provided in form of reductio ad absurdum, in an attempt to illustrate that the non-critical theses and anti-theses are logically
contrary to our experience, while the modified critical theses and anti-theses are not.
By touting this virtue, Kant attempted to persuade us to accept transcendental idealism together with his transcendental method, which goes hand in hand with empirical realism (that as the resulting cognition and hence within the limit of
experience, empirical reality with its phenomenal causal exhaustion is systematically accountable), and to anticipate the admission of morality in humanity.
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
The Thesis of the Third Antinomy is that, aside from the causal law of nature whereby any event taking place must have its cause in the previous time point, it is necessary also to admit the causal law of freedom, namely that there must be a cause without any preceding cause. One can make the argument that if there exists no free (uncaused) cause, there can then be no sufficient determination of the previous causal conditions and thus nothing can happen upon a solid ground. The Anti-thesis is that there exists no freedom; anything that occurs must have its cause(s) in the previous time stage – (the law in the Second Analogy of his first Kritik). Likewise, one can make the argument that if there exists any free cause, the law of natural causality (Second Analogy) loses its universal validity because of contradiction. Here we will only focus on the thesis, since the anti-thesis is relatively uncontroversial. The two sides of transcendental idealism and problematic transcendental idealism, which is ultimately just transcendental realism, both demand the phenomenal causal
exhaustion within the scope of experience, and the disagreement rests with the fact that transcendental idealism proposes the free cause in another kind of causality, while transcendental realism rejects that possibility (freedom is in no way factual;
there is only one causality) and maintains that phenomenal causal exhaustion alone is sufficient to account for our experience. In other words, we can simply give up the thesis and preserve the anti-thesis, so that it goes back to the problem of the thesis.
Both sides call for the incorporation of the phenomenal causal exhaustion in the anti-thesis, so the controversy rests mainly and squarely in the thesis.
As Allison (1990:11) so elegantly construed it, Kant's central claim in the Third Antinomy is that “it is only because the resolution of this antinomy leaves a
conceptual space for an incompatibilist4 conception of freedom that it is possible to give the claims of practical reason a hearing.” However, the reception of this attempt among his commentators, as Allison (ibid.) yet again so succinctly outlined, was
4 Incompatibilism: freedom is not compatible at all with the completely deterministic universe in natural causality.
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
either “ignoring [this cosmological dimension] completely or dismissing it as one more example of Kant being deflected from his proper philosophical course by architectonic considerations,” for such a dimension, viz., the treatise about the antinomy between the uncaused cause and the causal exhaustion in the cosmos (the world of the appearance), seems to have nothing to do with the issue of free well. Of course, this has largely to do with the poor reception of Kant's transcendental
idealism among the German idealists (and later continental philosophers as well) and the school of Kantian philosophy as launched in the North America by Strawson (1966), because, in Allison's opinion, these mainstream commentators rather
unfortunately took the position to understand Kant's division of the noumenal and the phenomenal as two worlds (two distinct ontological systems) when it is actually better to take the position, in Allison's and our opinion, to understand the division as two aspects, or more precisely, two epistemological aspects; for more details, see Allison (1983/2004). In general, owing to the assumption of the ontological
distinction of the two-world view among the German idealists and their preference for the ideas (the appearance) over the world in itself, such idealism cannot but fall into empirical idealism and hence transcendental realism in Kant's vocabulary, thereby pushing the concept of the world in itself either into the opposite position of mere appearance in dualism (e.g., the absolute idealism of Fichte and Schelling as well as Schopenhauer) or into nihilism (e.g., F. H. Jacobi, Nietsche and Heidegger).
Regardless of the direction, the key issue centers around the dissatisfaction with Kant's sharp distinction between intuition and concept yet somewhat contradictory demand for a priori unity between them. On the contrary, in the development among the Kantians in North America after Prichard and Strawson, the transcendental
idealism itself was “removed” in order to preserve the empirical realism, a move that Kant himself also desired. This was because, as pointed out by Prichard, Kant's distinction between the world in itself and the world that appears to us engenders a dilemma, which leads right back to “a Cartesian skeptic malgré lui” (Allison, 1983/2004: 6). Kant claimed that space and time are just the mere forms the world
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
can ever appear to us and also wished to preserve the idea of the world in itself, and in Prichard's view, these constructs together cornered Kant into the predicament whereby his “the world seems to us” means that for Kant, we cannot really know anything at all about the world (in itself) (Prichard, 1909: 78- 79; Allison, 1983/2004:
6). Following up on this reading by Prichard, Strawson (1966/1989: 240) initiated the trend of “separability thesis” (coined by Allison), whereby giving up Kant's claim that space and time are mere forms of cognition means that Kant's “transcendental argument” is no longer being “contaminated” by any idealistic premises (i.e.,
empirical realism alone, without being in relation to transcendental idealism) and can then be used for defending Kant's anti-sceptic position, and that for Kant, we do know things about the world in itself, thereby resolving the transcendental dilemma in Prichard's view. According to Allison, this line of thought is taken up and adhered to by Guyer, Langton and many other Kantians in North America (1983/2004: 6 – 11).
In response to these two scenarios, Allison defended transcendental idealism by interpreting the transcendental ideality introduced in KrV as “a bold, even
revolutionary, theory of epistemic conditions” (1983/2004: 19) and claiming that the transcendental condition for all possible experience should be understood as the epistemic condition of all possible cognitions, not the ontological condition. In other words, the distinction between the noumenon and the phenomenon should be simply epistemological (ideal), not ontological (real). This quickly gained popularity in this field as the two-aspect view vs. two-world view. Here, we shall skip the details, but this attempt by Allison is remarkable in (a) that the epistemology in KrV demands greater intellectual appreciation in contrast to the inappropriate mainstream
interpretations that are inclined to understand the distinction as one between two different ontological layers or two worlds, and (b) that those expressions in his transcendental philosophy should be understood not as the expression of the
unspeakable but as the “metalanguage” (1983/2004: 73), whose “correlates are to be
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
understood as technical terms within this metalanguage rather than as terms referring to transcendentally real entities.” In his treatise of the Third Antinomy, Allison (1990) put forth a crystalized demonstration of how transcendental idealism as the theory of the epistemological, rather than the ontological, conditions of cognition can resolve the antinomy, via which his defense of two-aspect view can be derived.
Moreover, the fundamental form of critical epistemology and the epistemological holism as its content will also serve as the backup to support Allison's defense.
The argument for the thesis of Kant's Third Antinomy was analyzed and summarized by Allison (1990: 15) into the following seven steps:
1. The assumption of the opposing view: “There is no other causality than that in accordance with the laws of nature” (KrV: A444/B472).
2. This means that “everything which takes place presupposes a preceding state upon which it inevitably follows according to a rule” (KrV: A444/B472).
3. But this entails (by universalization) the assumption that the preceding state must itself have come into existence in time. [Should this be denied, i.e., by assuming that the preceding state has always existed, then its consequence (the succeeding state) would likewise have always existed. But this contradicts the assumption that the latter has come into existence in time, i.e., taken place.]
4. Since the “causality of the cause through which something takes place is itself
… something that has taken place,” it “presupposes, in accordance with the law of nature” (nach dem Gesetz der Natur) (KrV: A444/B472) its own antecedent cause, and so on.
5. Consequently, given the assumption that “everything takes place in accordance with mere laws of nature” [nach blossen Gesetzen der Natur), there will
always be only a relative [subalternen] and never a first beginning, and consequently no completeness of the series on the side of the causes” (KrV:
A444-6/B472-4).
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
6. “But the law of nature [Gesetz der Natur] is just this, that nothing takes place without a cause sufficiently determined a priori” (KrV: A446/B479).
7. Thus, when “taken in unlimited universality,” the claim that “no causality is possible save in accordance with laws of nature” (nach Naturgesetzen) (KrV:
A446/B474) is self-contradictory. Consequently, it cannot be regarded as the sole type of causality.
Among these seven steps, the first five steps straightforwardly amount to the less controversial implication that given this assumption (1) the necessary consequence is that there is no complete determination of the series on the side of the causes (5) upon the basis of the universalization (3) of the law of natural causality (2). The problem is also very ostensive: while the proponent of the thesis maintains that there exists a self-contradiction between the assumption and the consequence that there is no
complete determination of the series on the side of causes (6), the opponent maintains that the complete determination is not necessary so that the anti-thesis (which only demands the phenomenal causal exhaustion) alone is sufficient to account for the empirical reality. As Allison wrote, “this step bears almost the entire weight of the argument. Unfortunately, it is far from clear that it is capable of supporting this burden” (1990: 16).
The crux of the problem in this step is the implication of the complete determination of the series on the side of the causes according to the necessary demands of the principle of natural causal laws (that everything that takes place must have its cause in the previous time point) and its universal entailment. The difficulty in following the line of thought on the proponent side is what this notion of “sufficient cause”
approximates. Allison (ibid.) believed, and we concur, that the mainstream
interpretation understands the thesis as: (1) both sides of the thesis and anti-thesis (as well as both the proponent and opponent sides of the thesis) are transcendental
realists who maintain that the cause (condition) of our experience is the self-existing
‧
world in itself and (2) the sufficient cause comprises of (a) a complete set of the simultaneous conditions that amount to the effect-event and (b) a completion in the successive series of antecedent causes leading up to the effect-event. While Kant believes that the transcendental realists cannot survive the challenge proposed in the thesis and thus brings forth his resolution of transcendental idealism, mainstream critics (e.g., Schopenhauer in his The World as Will and Representation) were quick to raise their objections. As Allison noted in vol. 1 (Payne, 1958: 498),
Schopenhauer understood the thesis in this composition of the two senses of sufficiency and complained that “only by conflating these two claims can the argument, on this interpretation, get off the ground” (Allison, 1990: 16). In the appendix of Schopenhauer's book, we can see how he presented a more direct and pointed criticism of Kant's treatment of the thesis. Specifically, Schopenhauer's objection was that as a transcendental realist, one can perfectly explain the
occurrence of an event with its complete set of simultaneous conditions, of which we do not need our complete cognition in order to account for such a completion in the world in itself. Moreover, the further requirement of the complete successive series remains in the scope of the unknown world in itself that transcends our capacity for experience and thus does not require any explanation.5 Such a line of thought was followed up by Kemp Smith: “each natural cause is sufficient for its effect. That is to say, the causation is sufficient at each stage. That the series of antecedent causes cannot be completed is due to its actual infinitude, not to any insufficiency in the
5 Cf. Schopenhauer, World as Will and Idea (Werke, ed. Frauenstädt, ii. p. 590; Eng. trans. by Haldane, ii. pp. 111-12). “The argument for the third thesis is a very fine sophism, and is really Kant's pretended principle of pure reason itself entirely unadulterated and unchanged. It tries to prove the finiteness of the series of causes by saying that, in order to be sufficient, a cause must contain the complete sum of the conditions from which the succeeding state, the effect, proceeds. For the completeness of the determinations present together in the state which is the cause, the argument now substitutes the completeness of the series of causes by which that state itself was brought to actuality;
and because completeness presupposes the condition of being rounded off or closed in, and this again presupposes finiteness, the argument infers from this a first cause, closing the series and therefore unconditioned. But the juggling is obvious. In order to conceive the state A as the sufficient cause of the state B, I assume that it contains the sum of the necessary determinations from the coexistence of which the state B inevitably follows. Now by this my demand upon it as a sufficient cause is entirely satisfied, and has no direct connection with the question how the state A itself came to be; this rather belongs to an entirely different consideration, in which I regard the said state A no more as a cause, but as itself an effect; in which case, another state again must be related to it, just as it was related to B. The assumption of the finiteness of the series of causes and effects, and accordingly of a first
beginning, appears nowhere in this as necessary, any more than the presentness of the present moment requires us to assume a beginning of time itself.”
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
causality which it embodies” (1962: 493). Allison's reconstruction is clearer: “the question of whether A itself is sufficiently explained is distinct from the question whether A (being given) is sufficient to explain B.”
2.2 Realistic Presentation of the Thesis and Two-World Reading of Its