The Oxford Handbook of German Philosophy in the Nineteenth Century

by Michael N Forster

  Empiricism can now be read, in a first attempt to articulate its very idea, as a critique of any metaphysical defence of beliefs in a transcendent after-world and in things or events behind the only real, empirical, world. The problem, however, now turns to the very meaning of the word ‘transcendent’, just because, as Kant realizes, our general and theoretical knowledge transcends by far merely empirical knowledge, especially when we identify the domain of empirical assertions with the traditional domain of historia: empirical assertions, even when they concern the future, are true only ex post or a posteriori. We have to wait and see, so to speak, whether they turn out to be true. But the generic truths of the theoretical sciences have a different logical status, even though their very meaning and function is narrowly connected with experience taken in a very wide sense of this word, which surpasses by far what we ‘experience empirically by our senses’. As a result, we have to be aware of the multitude of meanings of the words ‘empirical’ and ‘experience’, and, what is even more important, of Kant’s new distinction between the transcendent beliefs of ‘bad’ metaphysics and an explication of transcendental conditions of reasonable judgements concerning our joint and general experience. More precisely, Kant sees that a priori truths express such conditions of possibility of empirical meaning and immanent reference to real objects and states of affairs or events in the world.

  In fact, it does not help to propose not to read, or even to burn, all books that do not have only ‘empirical’ content, as David Hume proposes: we will thus not get rid of the traditional problems of metaphysics and of the fact that objective knowledge and reasonable belief ‘transcends’ or ‘goes beyond’ merely subjective, that is, arbitrary, belief. Kant’s critical philosophy tries to overcome this attitude of a merely empiricist and pragmatist ‘quietism’, as it is supported by a too hasty ‘sceptical’ rejection of all traditional ‘metaphysical’ sentences.

  Kant’s positive use of the word ‘metaphysics’ even as a title for systematic philosophy altogether stands in the tradition of Aristotle’s books: they investigate the meaning of ‘being’ and ‘truth’. Kant investigates conceptual and material preconditions of different forms of knowledge and the very constitution of a domain of discourse, for example, in his metaphysics of the natural sciences or in his metaphysics of morals. We therefore must distinguish a negative or critical from a positive or reconstructive use of the words ‘metaphysical’ and ‘metaphysics’. According to the first, the negative line, Kant and Hegel both use the words, like Hume, when criticizing dogmatic claims with transcendent content. Any such claim turns out to be a totally arbitrary opinion without sufficient reasons or without an adequate understanding of the still possible function of the corresponding words and texts in the only world to which we can apply them, namely, the real world in which we live. In this sense, having ‘transcendent opinions’ is senseless or at least lacks the usual assessment of reliability, when having the opposite opinion is as good and true, or rather, as bad and false.

  In other words, metaphysics in the negative sense is something that has to be criticized because it is merely ‘belief-philosophy’ or arbitrary opinion: all arguments in favour of really transcendent contentions are, by their very logical status, totally subjective and, as such, void of any reliable orientation. A certain consequence of this is that a merely subjective attitude to religious belief, together with the idea that there is no ‘true’ religion or theology, already renders such a belief as senseless. The same holds for any determinate belief about a ‘true’ world of things an sich, in Kant’s terminological sense: such a thing or world ‘in itself’ is, by Kant’s definition, totally detached from any access by our senses. The expression ‘an sich’ or ‘in itself’ thus marks a form of ‘abstraction’ by which it becomes tautologically true that we cannot know anything empirically about such ‘things’. Nevertheless, we sometimes can talk about such things, which thus turn into entities of a mere mundus intelligibilis, that is, of a domain of abstract things of mere thought, so-called noumena.

  The question now is in which sense it may be reasonable to believe that some of the objects of mere thought somehow really exist—even though they never can be directly perceived. Kant claims, for example, that free will is such a noumenon. The third chapter of the Phenomenology of Spirit shows that all forces, efficient causes and dispositions are, as such or an sich in Hegel’s sense, supersensible, as he nicely says: at least if we look at the way we have ‘immediate’ access to them, they are products of our intellect (Verstand) and, hence, of the abstract form of Kant’s noumena. Nevertheless we say that gravitational force, for example, really exists—as-such-and-for itself (an-und-für-sich), as Hegel would say—and even explain ex ante empirical results that we can perceive only ex post.

  In modern terms, forces are theoretical entities. Hegel makes the interesting point that, as such, their logical status is just like the soul, free will, or even God. According to this reading, we do not have to wait for the twentieth century in order to see that there are many different domains of merely theoretical entities that are, as such, mere objects of pure thought, produced by our ways of talking, for example, of using name-like expressions in our thinking, such as in all mathematical theories and models. When we say, for example, that there is a greatest pair of prime numbers n, m with m = n + 2, we already refer to (and quantify over) the abstract domain of the (pure natural) numbers. But when we say that numbers exist, we rather mean that such a domain is formally or internally well constituted and externally useful, as Carnap already pointed out: we use the sentences of pure arithmetic that are declared true as admissible rules for calculation. In their application to sets of empirical objects, however, we have to check how well the ‘ideal’ conditions of pure set theory are fulfilled.

  29.3 WIDENING TRANSCENDENTAL ANALYSIS

  As we have seen, Kant shares the empiricist’s contention that the world of sensible appearances (mundus sensibilis) decides what there actually is and what is ‘only’ an abstract, pure, or ideal thing of thought. To the world of appearances we have access by so-called intuition (Anschauung), which I take to be already something like joint observation of objective things and events. Thus a main difference between Kant and the classical empiricists rests on the distinction between unstructured (animal) sensations and jointly formed (human) intuition. When twentieth-century logical positivism in the philosophy of Russell, Carnap, and Ayer talks of ideal sense-data, the problem is that in the actual world there are no things corresponding to these merely theoretical entities as such—with the consequence that the movement of neo-empiricism is just a cluster of sweeping metaphysical doctrines.

  Things beyond appearances are always, at least at first, things of mere thought. We cannot observe these abstract entities as such. We cannot even evaluate the formal truths of pure number theory or of pure geometry merely by empirical observation. Nevertheless, the truths of geometry play an important role in empirical judgements about the non-pure gestalts of bodies or other spatial relations.

  Kant’s transcendental philosophy can now be understood as a speculative (i.e. in Kant’s and Hegel’s language: high-level) reflection on the very conditions of making true statements or of arguing for a belief. One of these conditions is that we know already, in this rather civil sense of a priori, how to use language, for example how to calculate with formal analytical truths that express merely conventional rules of verbal inference, as, for instance, in the case of deriving the sentence ‘Peter is unmarried’ from the sentence ‘Peter is a bachelor’. Empiricists agree of course that there are analytical truths expressing such rules. But they claim that the only a priori truths we should accept are of this merely formal sort. This is precisely what Kant denies by saying that the truths of pure arithmetic and geometry have a peculiar a priori status insofar as they express general norms of inference and not just empirical or historical facts a posteriori. They are not analytic but synthetic insofar as they are not true just by verbal st
ipulations: we cannot ‘define’ the truth of arithmetical or geometrical sentences by arbitrarily chosen axioms. We do not prove them by empirical observation either. Nevertheless, synthetic a priori truths are situation-invariant and time general; they express presupposed norms of inference which we use when we apply empirical concepts. In geometry, for example, we make reproducible spatial form(s) of things and a spatial order of relatively moving things in some ideal way explicit. A priori truths thus are part and parcel of the very content of empirical statements in which we say something about the objective world. They even express infinite or categorical truths about general types in contrast to the ‘finite’ truths of empirical statements about singular tokens in particular cases.2

  Hegel follows Kant in separating finite knowledge and beliefs in the sense of reference to already localized empirical things or events in the actual world from an infinite reflection or speculation that is concerned with the whole of the world and the generic forms of our local access to it. The latter includes our performative attitudes towards the world. And he sees that the so-called totality notions like God or the (immortal, that is time-general) soul always already had and still have a certain function in such speculative reflections—even though this function was and still is almost always partially misunderstood. The task is to make the immanent meaning, role, and ‘truth’ of ‘speculative’ sentences containing these words in our actual life explicit.

  Indeed, Hegel’s predecessors, Johann Gottlieb Fichte and Friedrich Wilhelm Joseph Schelling already saw a need to widen Kant’s set of synthetic a priori truths to a system of materially conceptual truths that function as ‘transcendental’ pre-conditions not only for the empirical content of informational statements about possible facts in the empirical world of particular events, but also for leading a reasonable life and for placing scientific claims into our system of cultural practices. Theological texts, religious poems, literature, and metaphysical reflections on their inner-worldly sense are part of this practice. If we read them in this way, we can distinguish religious and metaphysical teachings that transport ‘true’ from those that entail ‘misleading’ or ‘wrong’ orientations.

  In other words, infinite or speculative sentences reflect on the general forms of (human) life in the world and on practical attitudes towards them in our life. We have to distinguish them from ‘assertoric’ assumptions about nature as the domain of empirical ‘things’ that behave in a way independent of our actions and attitudes and from our constructions of scientific theories in which we articulate forces and default dispositions for various things in corresponding situations. Unfortunately, the word ‘speculation’ later gave way to abusive polemics—because of its use in the contexts of all too dangerous bets, for example, in casinos or on a religious after-life according to Pascal’s wager.3 Therefore, we have to return to understanding speculative metaphysics in the positive sense as a logical topography of different practices and different language games, as we can say today in a dense and thick way by making use of some Wittgensteinian labels.4

  29.4 A LONG NINETEENTH CENTURY OF PHILOSOPHY

  The continuation of Kant’s critical thinking by Hegel stands in the centre of the debate of the whole nineteenth century, as we shall soon see in more detail. The ‘long’ nineteenth century—which we may for our purpose allow to last almost 150 years from 1781 to 1927, that is, from the publication of Immanuel Kant’s Critique of Pure Reason to Martin Heidegger’s Being and Time—is, in philosophical terms, an ongoing debate about Kant’s criticisms of metaphysical rationalism and empiricism and his ethical interpretation of transcendent theology and other religious mythology: in Kant’s transcendental turn, our traditional talk about God and the soul gets a new world-immanent reading.

  Whereas the younger Ludwig Wittgenstein still follows Hume in his Tractatus (1921) in declaring that only statements about empirical states of affairs with truth evaluations ex post (according to some formal truth value semantics) are meaningful and claiming that there is no fact of the matter of sufficient cause on the one side, of free will on the other, Hegel had already seen that the theoretical sentences of the sciences do not belong to a domain of ‘historical’ sentences with a posteriori truth conditions but express situation-invariant default inferences. The background insight is this: generically admissible prima facie ‘deductions’ and normatively canonized default expectations are expressed by conceptual rules or sentences which we use in conceptual inferences conditioned by empirical classifications. They count as valid or admissible if they are reliable in standard situations. But this means that they often only hold if nothing goes wrong. Only few generic inferences hold universally. Even in the case of gravitation, for example, we must consider particular side-conditions, for example friction or diffusion.

  Causality thus belongs categorically to scientia as a special form of generic knowledge, not merely to narrative historia. In theories we do not just quantify over all past and future situations. It is not even clear what it could mean to quantify over all situations or possible worlds. Any reasonable estimation of possibilities already depends on theoretical causal fore-knowledge.5

  Since real scientific theories are our free linguistic constructions and do not directly represent the empirical course the actual world has taken and will take, but only help us in orienting our rational expectations, there is no—indeed there cannot be any—contradiction between free action and the principle of causal explanation, that is, the merely hopeful idea that every (actual or possible) event could be explained by ‘its’ efficient cause, that is, by actual preceding events and some causal law. The difficult word ‘could’ already refers to some ideal or full explanation. To believe that this ideal can or could be ‘absolutely’ fulfilled and that the causal principle ‘really’ holds in the real world (just because we succeed in some fields of physics in reproducing forms of movements causally) is a category mistake. A corresponding ‘ontic’ materialism (physicalism, naturalism) thus ironically turns out to be an idealistic hope and a metaphysical opinion. It must be counterbalanced by a realistic use of the causal principle. If we wish, we can label any ‘real logic of causal explanation’ with the phrase ‘objective idealism’, which we then should not confuse with subjective idealism as it leads from Descartes’s substance dualism via Berkeley’s empiricism to Kant’s somewhat unclear duality of receptivity and spontaneity in perception and thinking.

  The most widespread problem for understanding speculative reflection properly, however, is the difficult use of nominal expressions or title words in our logical topographies, for example, when we talk about the will or any capacity, about reason or spirit, but also about nature and the world. To be more precise, neither the expression ‘free will’ nor the expression ‘the cause of the event X’ name an ‘entity’. They rather are vague labels, as for example when we indicate, in the first case, that some event is the consequence of some free action which, as such, cannot be fully explained without an account of some governing intention to perform a certain generic action and not another (which does not apply to merely ‘natural’ events or merely ‘animal behaviour’), or when we say, in the second case, that some individual event of some particular type (together with the circumstances) is enough to explain causally what happened or will happen.

  The ‘antinomy’ of freedom and causality is the leading problem of Hume, Kant, and Hegel. Hume did not account sufficiently for the fact that our scientific research is guided by the quest for determining causes, as Kant clearly sees. But Kant’s own aspect dualism between Actor and Spectator (Lewis White Beck) does not solve the problem either, because, if every event really could be causally explained by the sciences, as Kant still claims, there would be no room left for free decisions and intentional acts in the world, as Hegel points out. As a result, Kant’s merely formalist use of the dualism between an empirical world and a merely intelligible world of pure thought is not good enough for understanding ethical responsibility at all, which, in contras
t to merely automatically reactive behaviour and other passive occurrences, rests on the notion of free intentional action. If our philosophical taste is good enough, we therefore should not be content either with Hume’s or with Kant’s ‘solutions’.

  Hegel’s next step of reflection shows the possible error of any sceptical fear of making possible errors. Scepticism can mislead us into a much too narrow idea of the ‘bounds of sense’ (Peter F. Strawson)6 in our thinking and language use. No meaningful and true knowledge is such that we are ‘forced’ to accept it. Reasonable judgement is always already free judgement. As such, it is defeasible. From this it does not follow that there is no knowledge at all. Rather, we must distinguish between claims we reasonably put into question and a merely arbitrary decision not to take part in a common practice of accepting and working with the standing sentences by which we express the best available generic knowledge at a given time. In other words, we have to consider the fact that a sceptic just defects from a reasonable debate by stopping using the cooperative device of judging which questions are still relevant and which must already be accepted as settled, or which competence and knowledge we presuppose and which claim is still under assessment. The sceptical sophist is not willing (or not able) to understand the methodological order of conceptual presuppositions defining the material meaning of a claim and the assessment of a sufficient satisfaction of its validity or ‘truth’. As such, he is ‘too smart’ to be reasonable: he wants to be ‘forced’ to accept a truth, which is impossible because such recognition is a free cooperative act. Another problem lies in lack of good judgement, as in mixing up ideal, pure, or merely formal conditions of truth with actually relevant conditions.