1. Introduction
2.2 Realistic Presentation of the Thesis and Two-World Reading of Its Modification of
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 Modification of Transcendental Idealism
Schopenhauer is correct in asserting that when state A is the entirely sufficient cause of state B in the sense that state A is the synchronically complete set of conditions amounting to state B, the completeness of the causal relation between the two states has no direct connection to the question of how the state A itself came to be.
However, his criticism that Kant's “universalization” of the demand for a completeness of the successive series in time is an over-simplification in the following two senses: (1) as also mentioned by Allison, Jonathan Bennett (1974:
185), following Heimsoeth (1967),6 pointed out that the claim requires the cause to be sufficiently “determined,” not merely “sufficient.” The demand for completeness entails the determination of the causal relation in the effective domain of such a law, viz., reality, not merely the explanation. From this point of view, the
self-contradiction indeed falls on the shoulder of the position of transcendental realism;
despite what Schopenhauer attempted to disparage, the contradiction only appears to be feasible in the position of transcendental idealism that spatial synchronical
completeness must be conflated with temporal successive completeness – in contrast to the realistic position that the appearance A and the appearance B can form a closed causal/temporal relation not further in relation to the pre-condition of appearance A.
More precisely, whereas Schopenhauer believed that the transcendental realists can
6 This group of critics maintains that the “a priori” in Step 6 is the pre-Kantian use (transcendental realistic use), i.e., in advance or prior to experience, rather than the Kantian use, i.e., as the necessary formal condition of experience.
This position maintains that in the presentation of the thesis, Kant intended to show how this question is regarded entirely in the view of transcendental realism so that the self-contradiction challenges this position; meanwhile, in his resolution, Kant presented his modification of the problem of the thesis and showed that there is no contradiction in transcendental idealism.
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
survive the challenge of the contradiction by distinguishing between the
completeness in coexistence at state A and the completeness in temporal succession that leads to state A, the self-contradiction is drawn entirely in the position of
transcendental realism. “Unfortunately,” as Allison also wrote, “Bennett does not explain what the requirement that a cause must be sufficiently determined prior to its occurrence amounts to or how the argument based on it is supposed to work” (1990:
16). Here, we follow up and spell out the requirement: upon the position of transcendental realism that the world in itself is assumed to be the cause of (its appearance in) our cognition, whether or not the appearance is realized, the world itself requires such a complete determination of a complete causal exhaustion in itself that amounts to any state that takes place in time, so that the principle of the natural causal laws that any event that takes place must have its cause in the previous time point is accountable. Without the first cause in the world in itself, such a
determination of even just the state A in the world in itself is impossible. Hence, the inference from the assumption of the self-contradictory conclusion in the thesis is entirely within transcendental realism and not in need of any perspective of
transcendental idealism. (2) The distinction between the completeness in coexistence and the completeness in temporal succession is itself not a transcendental realistic distinction but a transcendental idealistic one: either we take space and time as in the world in itself as the (transcendental) realists maintain, so that the inference in the thesis falls into self-contradiction as aforementioned in (1), i.e., the completeness of the traditionally ontological causal exhaustion in the world itself is required, whether or not in relation to our knowledge; or we take space and time as the mere forms of sensibility as Kant's transcendental idealism proposes (i.e., as to understand “a priori”
in Step 6 in the Kantian sense), so that the natural causality itself is taken to be effective in the phenomenal world and that the synchronical completeness and successive completeness would seem to be considered separately. This leads to an ambiguity between the world interpretation (e.g., Schopenhauer) and the two-aspect interpretation (Allison). Schopenhauer, unlike Kant who insisted on the a
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
priori unity of intuition and concept, openly believed that we directly intuit via desire/will without the need for any conceptualization; this position was further developed in (Nietzsche and) Heidegger and specifically crystalized in the notion of
“the will to power” (der Wille zur Macht). It follows readily from this fundamental deviation that the distinction between the phenomenon and the noumenon in Kant is interpreted as two worlds (two different ontological states: phenomenal world via sensibility and noumenal world via intelligence). Based on this, the completeness in a set of synchronical conditions and the completeness in the successive series of causes seem to be able to separate; since the appearance in state A has already had all the conditions amounting to the appearance in state B in the phenomenal world alone, the principle of the natural causal laws is already accountable, and this causal relation is itself entirely sufficient without needing to take into consideration the previous appearances as conditions that lead to the appearance in state A. In this case, the Third Antinomy can be resolved in the fashion of the mathematical antinomies, whereby we refuse the thesis and modify the anti-thesis with transcendental idealism, i.e., modestly replace the rational demand for the universalizability of the natural causal law with the principle (idea): Always seek further conditions. Although sticking to Kant's own epistemological distinction between the phenomenon and the noumenon (since the a priori unity between intuition and concept is by principle maintained) necessarily means that the form of space and the form of time can never be separated empirically, i.e., can never be separated into different ontological states – the distinction of the two types of completeness in “appearance alone” makes no sense. From this perspective, it becomes quite clear that the two world interpretation of the distinction between the phenomenon and the noumenon cannot but push the interpretation itself away from transcendental idealism back to transcendental
realism, in which the world in itself becomes the necessary ontological postulate that we either can never really know (a point that numerous Kantian scholars collectively worry about in transcendental idealism) or will require another different kind of cognitive capacity. One example is the Schopenhauerian will, in contrast to, as the
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
genuine transcendental idealism in Kant maintains, the necessary idea that we do consciously know the world itself (not dogmatically but critically) via understanding and reason within the empirical reality (resulting cognition), i.e., via understanding and reason that are incorporated with/in the appearance.
In Bennett's critical examination of the commentators following the Schopenhauer-Kemp Smith line, he essentially concluded that this line of interpretation has little textual support in Kant and identified Ewing's commentary as the most promising (Allison, 1990: 16; Ewing, 1938/1967: 218). Ewing argued for the thesis whereby
“if the cause is viewed as explaining or giving the reason of the effect, then this suggests that there must be something which is its own cause, otherwise causation gives no ultimate explanation or reason at all, just as it would be futile to give a chain of reasons for accepting a proposition if none of them could, any more than the original proposition, be seemed to be true in their own right.” Yet, he himself was not satisfied with the explanation argument, because the explanation itself is “by preceding events,” viz., the explanation itself requires the anti-thesis, which seems to us to anticipate Bennett's stress on “determination”. Besides, Ewing also thought that the thesis is based on the impossibility of a completed infinite; since such a concept is impossible, the uncaused cause must somehow be admitted in order, as Allison (ibid.:
16) aptly suggested, to leave “something further to be explained” or to provide “a resting place for thought.” This understanding continues to follow the trend of understanding the thesis in the position of the (transcendental) realists, and Ewing still believed that the thesis and the anti-thesis contradicts each other, for the same basis (the principle of natural causality that demands a necessary preceding cause for any occurrence of an event) leads to both the requirement and the rejection of the uncaused cause in the thesis and anti-thesis respectively. The resolution remains the
“third alternative,” i.e., that “the world is not in space and time at all,” even though he confessed that he was not convinced by the transcendental idealism of Kant in the footnote (ibid.: 222-3). Bennett was of course dissatisfied with the explanation
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
argument and with Schopenhauer, because this suggestion of Ewing for a mere
“resting place for thought” actually maintains both that there is only natural causality (the assumption in Step 1) and that every event must have an ultimate explanation (the interpretation of Step 6), with this interpretation pushing the opponent into a place of “such an obvious straw man that Kant cannot have taken it seriously or supposed that the thesis-arguer would do so” (ibid.: 186). The “determination” of the causal chain is that which accurately matters for such a demand for completeness. In other words, Kant was not hasty in treating the thesis by conflating the synchronical completeness and the successive completeness of the natural causality or simply sophistic in assuming that the opponent of the thesis only demands theoretical completeness. By staying loyal to the role of a realist, Kant continued to confront a realistic predicament, viz., without the uncaused cause, the determination of any effective cause of the occurrence of an event in natural causality alone is problematic, whether in theory (explanation) or in reality (determination).
Allison also found Ewing's rejection of the target of the thesis as a “straw man” to be
“premature” (Allison, 1990: 17). Alternatively, Allison interpreted Ewing's
suggestion of the “resting place of thought,” or his “principle of sufficient reason” to be what Bennett termed a “principle of causality,” so that we can arrive at a
Leibnizian understanding of the thesis as it was articulated in Leibniz's polemic with Clarke. Leibniz maintained that “every occurrence has a sufficient reason both in the sense that it has an antecedent cause and in the sense that it has an ultimate
explanation (accessible only to God) based on its role within the total context of the possible world actualized by the divine will” (Allison, 1990: 17; Allison's emphasis).
Furthermore, “much to the dismay of Clarke, Leibniz also extends this principle to the divine will itself, maintaining that it is determined (although not necessitated) by what the divine intellect recognizes as the best.”7 Such a Leibnizian response to Clarke showcased exactly the challenge that the thesis sought to address: the strict
7 Cf., e.g., Leibniz's Fifth Letter to Clarke, esp. secs. 1 – 20, The Leibniz-Clarke Correspondences, H. G. Alexander, ed. and trans., pp. 55-60 (Allison, 1990: 251).
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
insistence on the universal entailment of the principle of natural causality helps with the rejection of the spontaneity that Clarke wanted to be necessary for the
understanding of the agency in actions, be it divine or human. From this point of view, the criticisms of Kant's presentation of the thesis from Schopenhauer and Ewing are no longer trivial. Furthermore, Allison highlighted a quote from Leibniz that spells out directly the double implications in the completeness: “to anything's existing … to any event's happening … to any truth's taking place” (Leibniz, 1956:
95). This principle of sufficient reason with its universal entailment in such double implications “equally applies” (Allison, 1990: 17), meaning in Kantian terms that
“[Leibniz] construed it (the principle of sufficient reason) both as a logical principle requiring adequate grounds for any conclusion and as a real or causal principle requiring sufficient preconditions for every occurrence” (ibid.). In other words, in Allison's opinion, a “cause sufficiently determined a priori” in Step 6 can be equated with Leibniz's “sufficient reason,” so that Schopenhauer, Ewing and Kemp Smith's questioning of the weight of this step, especially the conflating of the synchronical completeness and the successive completeness, as well as the simple demand for spontaneity as “a resting place for thought,” is an over-simplification of Kant's
serious confrontation in presenting the thesis from the angle of transcendental realism such as Leibniz's, especially in front of the full Leibnizian argument.
Allison (1990: 18) succinctly pointed out that the law of natural causality appears twice in the argument for the thesis: the first in Step 4 expresses the requirement that every causal relation that takes place must have its own antecedent cause, since any exception to this requirement would evoke the violation of the principle of sufficient reason. As Allison understood it, “the requirement can be seen as an expression of that principle.” This leads to the incompleteness of the causal series in Step 5. In its second appearance in Step 6, the law expresses something more than before: the double demands in the Leibnizian notion of sufficient reason, i.e., the demand for an accurate antecedent cause taking place in (previous) time (i.e., “universalizability
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
requirement”) and the demand for an ultimate and unconditioned explanation (i.e.,
“completeness requirement”). In the end, such two-fold “conflicting demands of a single principle” amount to the self-contradiction that Kant stated in Step 7.
The Leibnizian notion of sufficient reason invited objection, however, in its
“circularity” and “dogmatic appeal” (Röttges, 1974: 36; Ortwein, 1983: 24-6;
Allison, 1990: 18). It quickly becomes clear that the argument in the thesis, construed in such a line of Leibnizian sufficient reason, takes aim at a real target, namely Leibniz and not a straw man. But, as Allison pointed out, “for that very reason its scope is limited to philosophies with similar metaphysical commitments, in particular, a commitment to a strong form of the principle of sufficient reason”
(ibid.). Once we reject that rationalistic commitment, we can reject the thesis, as what Schopenhauer and Ewing have done, by pointing out that the thesis conflates these two demands; in other words, as a realist, one still retains the opportunity to maintain that the natural causality alone (universalizability requirement) can suffice for our experience as such without the need for the uncaused cause (completeness requirement). Essentially, as Allison also highlighted, Kant's fundamental principle underlying all these four antinomical conflicts as a whole is being challenged, i.e., “if the conditioned is given, the entire sum of conditions, and consequently the absolute unconditioned (through which alone the conditioned has been possible) is also given.”
This is the one principle that Kant at first raised to challenge the dogmatic
metaphysics of the transcendental realists and later undercut with his own philosophy of transcendental idealism. Since both camps in the dispute are transcendental
realists, neither side can actually avoid the accusation of appealing to such a principle dogmatically. As was pointed out previously, either the world in itself alone or the divine will has to follow such a principle so that the empirical reality that we are now experiencing (the resulting cognition) can be identified with the transcendent reality (the cause of cognition in itself), unless we abandon both the assumption of the world in itself and the divine will. By leveling accusations at the dogmatic employment of
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
such a principle in transcendental realism, Kant diagnosed the former two
mathematical antinomies to be cases whereby both the thesis and the anti-thesis are not adequate demands (since the world in itself is in no way a legitimate existence, the dogmatic requirement of the completeness of the conditions that lead to the given conditioned makes no sense). For the latter two dynamical antinomies, the diagnosis is that both the thesis and the anti-thesis may be true or somehow compatible in the modification of transcendental idealism (that since the required complete set of conditions need not be real but ideal, i.e., in reason, the demand for it may be compatible with the demand for empirical reality). Such a scheme of Kant was judged unconvincing by the Schopenhauer-Kemp Smith camp, particularly by Strawson.
Strawson challenged the methodological structure of these four antinomies and their resolution of transcendental idealism. In particular, with regard to the systematic distinction between the former two mathematical antinomies and the latter two dynamical antinomies, Strawson (1966: 209) wrote:
It seems obvious what the correct “critical” solution of this conflict should be.
Since things in space and time are appearances, the series of ever more remote causes should no more be regarded as existing as a whole than the series of ever more remote temporal states of the world or the series of ever more remote spatial regions of the world. Since the series does not exist as a whole, there is no question of its existing either as an infinite whole or, as is asserted in the thesis, as a finite whole with a first, uncaused member. Every member of the series which is actually “met with” in experience, however, may and must, be taken to have an antecedent cause. The thesis, then, is false, the antithesis true.
At this point, Strawson, following the Schopenhauer-Kemp Smith line of
interpretation, no longer complained about the (transcendental) realistic presentation
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
of the thesis and the anti-thesis but instead focused on the modification of
transcendental idealism, which Kant believed to be the resolution to the antinomy, and furthermore, the systematic methodology for such critical resolutions. Since transcendental idealism maintains that space and time are the mere forms of the appearance and that natural causality (the demand for the Second Analogy) is phenomenal, the difficulty in the determination of the complete causal series in the world itself, together with the demand for either the spatial completeness of the causal condition (Schopenhauer's complete set of conditions at state A) or the temporal completeness of the causal condition (the completeness of the successive series on the cause side), does not bother us any more. As Schopenhauer noted, the appearance in state A, which already contains a complete set of conditions amounting to the appearance in state B and is not in need of any reference to the preconditions leading to state A, suffices to account for the causal determination between the
appearance in A and the appearance in B. Thus, it seems, in light of such a version of transcendental idealism, that the thesis should be rejected and the anti-thesis alone should stay, as it was clearly stated in the last line of the quoted text of Strawson above; the Leibnizian rationalistic demand for the double completeness does not seem to suit the modification of transcendental idealism, since such a demand
contains assumptions that the modification should reject – assumptions including (a) the cause's being independent from the appearance and our cognition (i.e., God, or divine will), in contrast to being coined in the appearance via the realization of our
contains assumptions that the modification should reject – assumptions including (a) the cause's being independent from the appearance and our cognition (i.e., God, or divine will), in contrast to being coined in the appearance via the realization of our