GUIDO KREIS | Negative Dialektik des Unendlichen: Kant, Hegel, Cantor | Berlin: Suhrkamp, 2015


By Franz Knappik

I have greatly enjoyed reading Guido Kreis’s new book Negative Dialektik des Unendlichen. Kant, Hegel, Cantor has much to offer to its reader: a book-length argument in the logic of infinity, careful reconstructions and novel interpretations of key doctrines and arguments in Kant and Hegel, as well as an insightful analysis of important types of dialectic. The book draws on a broad array of sources that ranges from Kant and Hegel through Cantor’s work on the mathematics of infinity to the most recent debates on unrestricted quantification. Despite the complexity and abstractness of the topics that it deals with, Negative Dialektik des Unendlichen is written throughout in a very readable and clear manner. 

The main argument of the book goes as follows. Kant realises that our attempts to think infinity lead to paradoxes: it seems possible to derive contradictory truths from the concept of the total series of real conditions for given spatiotemporal objects.[1] In dealing with the antinomies, Kant uses a strategy that Kreis calls “limitative dialectic”: Kant resolves the apparent contradictions by adopting Transcendental Idealism, and hence by restricting the range of what we can know. On Kreis’s reading, Kant’s Transcendental Idealism is a version of semantic antirealism. It denies that our assertions about objects that are necessarily experience-transcendent have a truth-value. Hegel, by contrast, develops an alternative, ‘positive’ form of dialectic by generalising and radicalising Kant’s antinomies. For Kreis’s Hegel, contradictions that arise from particular categories—such as those of the qualitatively and the quantitatively infinite (corresponding to Kant’s first two antinomies)—have to be resolved by the construction of ‘concretely universal’ concepts, which in their turn have to be examined for contradictions. This ultimately allows Hegel to develop a complete system of categories—the ‘Concept’—which is itself free from contradictions, and which is valid for (and allows us in principle to cognise) everything. Yet this last claim of Hegel’s, Kreis argues, is highly problematic: for the notion of ‘everything’, too, yields paradoxes, as has been shown by Cantor for the set of all things, and, more recently, by Patrick Grim and Timothy Williams for ‘all facts’ and for unrestricted universal quantification. Kreis argues that these paradoxes can only be resolved if some fundamental assumption of our ordinary worldview is abandoned. He therefore recommends a ‘negative’ dialectic which diagnoses the paradoxes of infinity and analyses possible remedies, but which does not actually prefer a particular solution, and remains aporetic instead.

While I find many of the systematic and exegetical claims that Kreis advances in developing this argument very plausible, there are also a number of points about which I have doubts. In the following comments, I first raise some objections to Kreis’s interpretation of Kant’s Transcendental Idealism (Section 1). Afterwards, I shall discuss his reading of Hegel’s idealism (Section 2).

1. Kant’s Transcendental Idealism

It is customary in the recent literature on Kant’s Transcendental Idealism to distinguish three main exegetical approaches to this doctrine: first, ‘two-world’ readings, on which appearances and things in themselves are numerically distinct entities that constitute two ontologically distinct realms—a realm of things in themselves and a realm of appearances (understood as mental states, or constructions out of mental states, of subjects) (e.g. Van Cleve 1999); second, ‘ontological two-aspect readings’, according to which the distinction between appearances and things in themselves refers to an ontological distinction between two different aspects of numerically identical entities (e.g. Langton 1998, Rosefeldt 2007, Allais 2015); and ‘epistemological’ or ‘methodological two-aspect readings’, on which the distinction between appearances and things in themselves refers to a distinction between two ways of thinking about or referring to things, namely, one that takes into account necessary conditions for (human) knowledge of those things, and one that abstracts from such conditions (e.g. Prauss 1974, Allison 2004).

Kreis proposes a semantic version of the third approach. This semantic version understands the distinction between appearances and things in themselves as a distinction between two classes of assertions: namely, “objectively valid” assertions, which are assertions about entities that can be given in our experience; and assertions that are “not objectively valid”, which are assertions about entities that cannot be given in our experience (pp. 114–15). But the distinctive feature of Kreis’s reading of Transcendental Idealism is not so much this semantic twist (which itself is compatible with existing versions of the methodological two-aspects reading) but the further claim that for Kant, assertions that count as “not objectively valid” are assertions that lack a truth-value. This has the following important consequence: not only is it not possible for us to know whether it is true or not that there are things in themselves that ground the contents of our experience, or that the world in itself has a temporal beginning, or that God exists. Rather, assertions with these contents are not even candidates for truth or falsity. (As Kreis emphasises, this is not a form of a verificationist reading [such as Strawson’s 1966], since assertions can lack a truth-value and still be meaningful [pp. 94–5].)

While Kreis’s discussion certainly adds an interesting new facet to the debate on Transcendental Idealism, I have doubts about how attractive it is as a reading of Kant. To explain why, I shall first present two problems for this reading (a–b). Afterwards, I examine the reasons that Kreis gives in favour of his interpretation, and argue that they are not compelling (c–e).

(a) The theses and antitheses of the antinomies are claims about the world, understood as thing in itself (A504/B532). On Kreis’s reading, they therefore have to count as not objectively valid, and hence as lacking a truth-value. But Kant explicitly states that according to his solution of the antinomies, both the theses and the antitheses of the mathematical antinomies are false. (Kreis suggests that Kant asserts this only in the Prolegomena, while he merely “considers” it as a possibility, or put it forward as a “conjecture”, in the First Critique [pp. 96, 99, citing A504/532]. But this is not correct: at A531/B559, looking back on the mathematical antinomies, Kant writes that in them, “both dialectically opposed assertions had to be declared false”.)

Moreover, regarding the dynamic antinomies, Kant holds that in a “corrected” form, the theses and antitheses “may both be true” (A532/B560; cf. A560/B588, A562/B590). At least with regard to the theses, it is clear that this “corrected” form[2] amounts to claims about entities that necessarily transcend experience: the relevant claims, of which Kant argues at great length that they are not rendered necessarily false by the antinomies or by Transcendental Idealism (A532ff./B560ff.), hold that (i) there are unconditioned causes (namely, we as free agents), and that (ii) there is a being that exists with unconditional necessity. If Kreis’s reading were correct, we should expect Kant to write that (i) and (ii) lack a truth-value, rather than holding that they “may […] be true”.

It should also be noted that the passages from which I have cited are not incidental remarks which might not express Kant’s considered position. Rather, Kant assigns the asymmetry between the mathematical and dynamic antinomies great importance[3]: for it is exactly this asymmetry which enables him to argue for the non-impossibility of transcendental freedom and God’s existence (topics that I shall turn to immediately). We therefore should take very seriously what Kant writes about the truth and falsity of the theses and antitheses in this context.

(b) Together with the immortality of the soul, transcendental freedom and God’s existence are the postulates of ‘rational faith’. In the section ‘On Having an Opinion, Knowing, and Believing’ of the First Critique, Kant analyses Glauben (translated variously as ‘belief’ or ‘faith’) as a form of “taking something to be true” (Fürwahrhalten) that obtains when a subject takes itself to have “subjectively sufficient”, but “objectively insufficient” grounds for endorsing the content in question (A822/B850). In the specific case of rational faith, the “subjectively sufficient” grounds consist in the facts that we are obliged by moral law, and that we can only make sense of this obligation if we believe in free will, God, and immortality (A828/B856).[4] Kant thinks that such rational faith is not an optional, but a “necessary” attitude that is rationally obligatory for us (A824/B852 together with A828ff./B856ff.).

Since rational faith is a form of “taking something to be true”, it requires us to take the postulates to be true. It is not simply a matter of acting as if they were true, or of adopting a practical viewpoint with a corresponding “alternative normative description” of our actions, as Kreis suggests in a related context (p. 208n). Rather, we have rational faith only if we believe that God exists and that we are free and immortal, and if we are aware that we are rationally obliged to believe this for practical reasons, even though we have no sufficient theoretical reasons (“objectively sufficient grounds”) to do so. Now if Kreis’s reading were correct, Transcendental Idealism would teach us that the postulates lack a truth-value. But it is certainly irrational to take a proposition to be true of which one knows at the same time (thanks to Transcendental Idealism) that it does not have a truth-value at all, and hence cannot be true. Kreis’s reading therefore has the consequence that morality puts irrational demands on us. Yet this cannot be acceptable for Kant, since the entire purpose of the theory of postulates (and of the discussion of the dynamic antinomies that prepares it) is to render the moral viewpoint rationally coherent.

Hence, there are substantive problems with Kreis’s semantic reading of Transcendental Idealism. Let us see now what reasons Kreis offers to motivate this reading, and whether they are strong enough to recommend the reading even in the face of such obstacles.

(c) To begin with, the only more or less direct textual support that Kreis offers for his reading is a passage in which Kant argues that questions about experience-transcendent objects are void and empty (“gänzlich nichtig und leer”), because no object of the question is given (A478–9/B506–7n). Kreis suggests that it is natural to infer from this that judgements about experience-transcending objects lack objective validity, which, in its turn, he interprets as meaning that they lack a truth-value (p. 97); and at a later point (p. 126), Kreis cites the passage at A478–9/B506–7n as direct support for his semantic reading.

Before considering whether this passage actually supports Kreis’s reading, notice that the passage in question occurs in a section whose overall aim it is to argue that transcendental philosophy is able to answer all questions that can be asked about the objects of the ideas of reason. In transcendental philosophy, one therefore has no “right to avoid answering” such questions “decisively” (A477/B505). In particular, this means that transcendental philosophy cannot simply avoid resolving the antinomies by leaving open the cosmological questions. The passage cited by Kreis explains the way in which transcendental philosophy answers most questions about the objects of reason: namely, by identifying them as void, since no object is given for them. In such cases, Kant writes, “no answer is an answer” (A478–9/B506–7n).

Now there are at least two reasons why the passage at A478–9/B506–7n does not support a reading in terms of semantic anti-realism. First, Kant explicitly states that cosmological questions are excepted from the charge of voidness: as he points out, there is a sense in which the object of the cosmological idea is given to us, since the world is the series of conditions for given appearances, and (most of)[5] these conditions are themselves possible objects of experience (whereas God and the soul do not show up in experience at all) (A478–9/B506–7). Kant hastens to add that the series of conditions is not given in experience as a whole. Nevertheless, he thinks that the fact that each condition on its own can become an object of experience suffices for cosmological questions to count as non-void. Importantly, this seems to include the questions about the infinity vs. finitude of the total series of conditions, for these are the only cosmological questions that Kant is concerned with in the Antinomies. So even if one could validly reason from the voidness of a question to the lack of a truth-value for candidate answers to that question, the passage cited by Kreis would not show that claims about experience-transcending objects in general lack a truth-value: cosmological claims concern experience-transcending objects, but the charge of voidness does not apply to questions about them.

Secondly, Kant restricts the charge of voidness also in another respect: he levels it only against questions that ask “[w]hat kind of constitution transcendent objects have (A478–9/B506–7n; emphasis added). Kant does not claim that questions about the mere existence of transcendental objects are void. So the passage in question cannot support Kreis’s view, on which claims about the existence, too, of experience-transcending objects lack a truth-value (p. 100).

But why, it may be replied on Kreis’s behalf, does Kant use the term ‘void’ at all in order to characterise questions about transcendent objects in the passage in question? Why does he not simply describe those questions as going beyond the limits of what is knowable for us? Does the charge of voidness not suggest something stronger than that—such as precisely a lack of truth-value for the possible answers to those questions? I think that this reply loses its force if the context of the passage is taken into account once more. In the section in which the passage occurs, Kant contrasts transcendental philosophy with empirical sciences. It is legitimate for an empirical science to leave a question open because we do not (and perhaps will never) have sufficient empirical knowledge to decide the question (A476–7/B504–5). By contrast, transcendental philosophy, as I have already mentioned, cannot similarly leave open the questions about transcendent objects with which it is concerned in the Dialectic. Rather, it has to, and is able to, definitely settle them.

Given this context, the reason why Kant talks here about ‘void’ questions, rather than questions about objects that go beyond the boundaries of possible knowledge, may simply be a rhetorical need to make the contrast with empirical sciences sufficiently clear. The contrast case of empirical sciences is itself characterised in terms of an appeal to ignorance, so if Kant were here to describe transcendental philosophy too, in terms of boundaries of knowledge or the like, it might be more difficult for the reader to grasp the contrast that Kant is after. But this is entirely compatible with the view that the reason for which transcendental philosophy, according to Kant, declares questions about transcendent objects as void (and thereby definitely settles them) is precisely that we cannot possibly have knowledge about such objects.

(d) In addition to the passage from A478–9/B506–7n that I have been discussing so far, Kreis argues that a reading of Transcendental Idealism in terms of semantic anti-realism is suggested by Kant’s theory of judgement. His idea is that for Kant, a judgement is a normative act of claiming objective, and hence also intersubjective, validity for a particular content. That a judgement possesses a truth-value is a “consequence” of this normative act (p. 98). Yet since, for Kant, we “cannot have any obligation to do what we cannot know” (A476/B504), we cannot perform this normative act with regard to objects that are necessarily experience-transcendent. Hence, judgements about such objects have no truth-value (p. 99).

I think that there are several problems with this argument. One problem is that Kant’s principle that we “cannot have any obligation to do what we cannot know” is formulated, as it stands, only for legal and/or moral obligations,[6] and it would take much more argument to show that Kant is committed to an analogous principle for acts of judging. Furthermore, even if the latter could be shown, it would not be clear how semantic anti-realism results from this. Presumably, Kreis thinks that in judgements we establish normative demands for ourselves and others to do particular things—e.g. we oblige ourselves to offer reasons for the judgement if challenged, and we expect others to agree with our judgement (p. 98)—and that necessary ignorance about the truth or falsity of the judgement makes it impossible for us to establish such demands (p. 99). But it seems clear from the context in which Kant formulates the above normative principle that it means roughly the following: we can only be obliged to do something if we can know about this obligation.[7] So the demands that are established in acts of judgement could be undermined by ignorance only if this ignorance concerns specifically the fact that the judging subject raises such-and-such demands, e.g. that she obliges herself to give reasons when challenged, or that she expects others to agree. In other words, we can judge, on this view, only if we and others can be aware of the normative demands that the judging subject raises for herself and for others in virtue of this act of judgement. But it seems to be typically unproblematic both for a subject who judges that p and for others (e.g. those who witness public expressions of the judgment) to have knowledge about this aspect of judgment—i.e. knowledge about the fact that the judging subject raises such-and-such normative demands, regardless of whether it is possible for any subject to know whether p is true or not. Hence, it is unclear why our necessary ignorance about transcendent entities should make it impossible for us to claim objective validity for particular thoughts about these entities. (Of course, such claims will always be unwarranted for Kant, but Kreis needs the much stronger point that they are impossible.)

(e) A last point that Kreis puts forward in favour of his reading of Transcendental Idealism concerns the concept of existence. Kreis argues as follows: for Kant, existence is a category (that of actuality), whose application is governed by a principle (the Second Postulate of Empirical Thinking) according to which “[t]hat which is connected with the material conditions of experience (of sensation) is actual” (A218/B266). For Kreis, it follows from this that an application of the predicate of existence to objects which necessarily transcend experience cannot yield true or false judgements (p. 123). I do not think this conclusion follows, since there are at least two alternatives to it. First, one could hold that the Second Postulate of Empirical Thinking has the consequence that experience-transcendent objects do not exist. Secondly, one can distinguish the schematised category of actuality or existence from its unschematised counterpart, and allow that the latter be applied to experience-transcendent objects (without thereby making any claims to substantive knowledge about them). There is strong reason to prefer this second option over the previous alternatives. For Kant often claims that appearances need to be grounded in, or caused by, things in themselves,[8] and it seems that such a relation of grounding or causation can obtain only if both of its relata exist.

Hence, not only does Kreis’s reading of Kant’s Transcendental Idealism face severe problems, it also lacks compelling arguments in its favour. I therefore do not presently see strong reasons to accept this reading.

2. Hegel’s Idealism

The chapters on Hegel in Negative Dialektik des Unendlichen contain (1) a reading of Hegel’s idealism, (2) an interpretation of his dialectical method, (3) a reconstruction of the chapters ‘Quality’ and ‘Quantity’ in the Science of Logic, and (4) a critique of Hegel’s claim that the Concept is everything. To my mind, what Kreis writes about (2) and (3) may well be the best treatment of these topics in the literature. His impressive interpretation of the dialectic of qualitative and quantitative infinity is able to make sense of the details of Hegel’s arguments and terminology in these sections, and even to improve the arguments where they fail to be compelling. Similarly, Kreis’s account of Hegel’s dialectic offers an accessible and unprejudiced reconstruction of the basic ideas and structures of Hegel’s dialectic that both makes sense of Hegel’s views and does justice to his texts. Kreis’s discussion of (4) elegantly shows how the paradoxes formulated by Cantor, Grim and Williamson are relevant to Hegel’s project in the Science of Logic (even though one may wonder whether the paradoxes are any more pressing for Hegel’s theory than for just any ontological theory that makes claims about everything that there is). I am less convinced by Kreis’s discussion of (1), Hegel’s idealism. The present section will therefore be focussed on this part of his reading.

Kreis sees Hegel’s idealism as a variant of Kant’s transcendental theory of categories (even though for Kreis, Hegel worked with a reading of Kant that made him fail to see this similarity). For the sake of the following discussion, it will be helpful to distinguish some central claims of Kreis’s reading of Hegel’s idealism:

(1) Like Kant’s transcendental philosophy, Hegel’s system examines the basic features of our system of categories—it is a form of descriptive metaphysics (p. 219).

(2) For Hegel, our categories are at the same time the basic structure of reality (p. 347).

Hegel’s reason for (2) is that

(3) For Hegel (like for Kant), there are no thoughts with a truth-value about objects that go beyond possible experience (p. 210–11), and hence beyond the scope of our categories.

(As Kreis nowhere indicates otherwise, it is reasonable to assume that he understands the notion of ‘possible experience’ in the context of his interpretation of Hegel in the same way as in his interpretation of Kant. My following discussion will rest on the assumption that on Kreis’s reading, Hegel shares the relevant notion of ‘possible experience’ with Kant. I shall come back to this point at the end of this section.)

(4) Hegel acknowledges an ‘Absolute’ only in the following sense: philosophy—the examination of the basic features of our system of categories—is ‘absolute’ (namely, ‘absolute spirit’) because it constitutes an activity of objective spirit in which the latter grasps its own nature (p. 216).

(5) Absolute spirit is not itself an experience-transcendent entity (p. 218).

(6) There are no experience-transcendent entities among the contents of what absolute spirit thinks (p. 218).

Hegel’s reason for (6) (as well as for [3]) is that

(7) Absolute spirit is the highest realisation of our freedom, and any thought about an inaccessible “in-itself” would be incompatible with this freedom (p. 210–11).

This reading shares with some ‘anti-metaphysical’ readings of Hegel’s idealism like Robert Pippin’s the view that Hegel denies the possibility of valid thoughts about objects that necessarily go beyond experience (e.g. Pippin 1989:91), while it agrees with recent ‘moderate metaphysical’ readings (e.g. Stern 1990, 2009; Kreines 2015) that Hegel’s idealism offers a genuinely metaphysical theory (in the sense of a theory about being as such) (p. 209).[9] With both strands of scholarship, Kreis has in common, in virtue of his claim (4), the rejection of ‘strong metaphysical’ readings (e.g. Horstmann 1984; Beiser 2005; Bowman 2013), according to which Hegel sees reality as grounded by one ultimate, necessarily existing explanatory principle (usually identified with the Idea).

For my own part, I think that only strong metaphysical readings can fully do justice to Hegel’s texts (Knappik 2013, chs 5, 9; Knappik 2016b). I therefore disagree with Kreis’s claims (3), (4), and (6), while I do agree with the rest of his above claims.[10] For the sake of the following discussion, however, I shall bracket my disagreement with claim (4), and focus only on claims (3) and (6). These claims are the most controversial parts of Kreis’s reading because they contradict not only strong metaphysical readings, but also the most prominent versions of moderate metaphysical readings. They therefore stand in particular need of discussion.

The main reason why advocates of moderate metaphysical readings like Kreines and Stern would deny (3) and (6) is that for them (as for some versions of strong metaphysical readings: e.g. deVries 1991; Knappik 2013, 2016a), Hegel sees the basic structure of reality as governed by realistically understood universals. These universals constitute the essences of the individuals that instantiate them, and form the basis of explanations for the determinate features of those individuals (Stern 1990, 2009; Kreines 2008, 2015).[11] Examples for such universals (or, in Hegel’s terminology, “objective determinate concepts”: Kreines 2008, 2015; Knappik 2013:294ff., 2016a) are chemical and biological kinds, but also MATTER as basic kind of mechanical nature, and the kinds that Hegel investigates in his philosophy of spirit, e.g. REASON, STATE or ART. To cite just one passage that strongly suggests a metaphysical picture like this:

[I]f […] the nature, the specific essence, that which is truly permanent and substantial in the manifold and accidentality of appearance and transitory manifestation, is the concept of the thing, that in the thing itself which is universal, just as each human being, although infinitely unique, has the principle of all his particular features in himself, in his being human, and just as each individual animal has its principle in its being animal: if this is true, then there is no saying what such an individual could still be if this foundation were removed from it, no matter how many the predicates with which it would still be otherwise adorned –even though such a foundation can, like the other features, be called a predicate. (Science of Logic5:26 [Hegel 2010a:16–17; trans. amended])

Adherents of both moderate and strong metaphysical readings of Hegel’s idealism have argued that it is one of the central aims of Hegel’s Science of Logic to establish such a metaphysics of realistically and essentialistically understood universals.[12] In particular, it has been suggested that in the Logic of Essence and the Logic of the Concept Hegel discusses candidates for metaphysical entities that might support explanations for observed phenomena (e.g. laws of nature, forces, mechanical systems etc.), and that his own form of essentialism is meant to offer a superior alternative to them (Kreines 2004, 2015; Knappik 2016a). On the resulting view, Hegel holds that we can grasp the structure of reality by identifying the relevant universals, and explaining the determinate features of reality on their basis.

Now Kant clearly holds that (real) essences go beyond the sphere of possible experience (Hanna 2006:160ff.). For example, he writes in a letter to Reinhold of 12 May 1789:

[T]he real essence (the nature) of any object, that is, the primary inner ground of all that necessarily belongs to a given thing, this is impossible for man to discover in regard to any object […]. [T]o know the real essence of matter, the primary, inner, sufficient ground of all that necessarily belongs to matter, this far exceeds all human capacities. We cannot discover the essence of water, of earth, or the essence of any other empirical object […]. (Br, AA 11:36–7)[13]

The First Critique presents a straightforward rationale for this position. While Kant allows for knowledge of entities that are not directly observable but postulated by empirical science (e.g. “magnetic matter”: A226/B273), he states two conditions for the possibility of such knowledge: the fact that we cannot directly observe such entities must be a matter of the contingent constitution of our sense-organs; and the justification for postulating such theoretical entities must be an application of the Analogies of Experience (ibid.). (Real) essences satisfy neither of these two conditions.

Quite independently of whether Hegel is read as accepting or not a form of ontological monism, an Absolute that necessarily exists etc., moderate and strong metaphysical readings (in the versions that I have cited) therefore agree that Hegel is committed to the existence and knowability of experience-transcendent entities (in Kant’s sense), namely, realistically and essentialistically understood universals. Hence, the readings in question all contradict Kreis’s claims (3) and (6), regardless of whether they agree or not with (4). At the same time, I think it is fair to say that advocates of the essentialist view of Hegel have put forward robust textual evidence for their interpretation (see again Stern 1990, 2007; Kreines 2008, 2015; Knappik 2016a). So Kreis needs strong arguments in order to rebut the essentialist reading, and to make plausible his own claims (3) and (6).[14]

Kreis nowhere directly addresses essentialist interpretations of Hegel. Instead, the only argument that he offers for claims (3) and (6) is summed up in claim (7)—the claim that absolute spirit is the highest realisation of our freedom, and that any thought about an inaccessible “in-itself” would be incompatible with this freedom (pp. 210–11). I fully agree that the assumption of an “in-itself” that is cognitively inaccessible to us would constitute for Hegel a limitation of our freedom, and that on his account, absolute spirit—especially in the form of philosophy—sets us free from such limitations. But it seems that this thought (as well as Kreis’s argument for it in Section 7.4) supports a reading of Hegel’s idealism such as Kreis’s only if the possibility of experience of an entity is seen as a necessary condition for the cognitive accessibility of this entity—and hence, if claims (3) and (6) are already presupposed. If, by contrast, (3) and (6) are denied, there is no reason why reference in philosophy to an experience-transcendent but cognitively accessible entity should compromise our freedom.[15] (Of course, such reference requires an explanation of how we can gain cognitive access to experience-transcendent entities like essences. But both in the literature on Hegel and in current epistemology and philosophy of science, there are well-established candidates for such an explanation, including accounts in terms of inferences to the best explanation [e.g. Kreines 2015; Ellis 2001], and accounts in terms of an intuitive source of modal knowledge [cf. Förster 2012; Bealer 2002].)

I therefore think that Kreis’s argument will not, as it stands, convince adherents of the essentialist reading. However, there is an alternative point that might be made in favour of claims (3) and (6): there are passages in Hegel that strongly suggest at least something close to (3) and (6). One such passage, which I briefly discuss in the remainder of this section, is the following: “There lies in empiricism this great principle that what is true must exist in actuality and be there for perception” (Encyclopædia §38, Remark, 8:108 [Hegel 2010b:79]).

At first sight, this passage certainly seems to commit Hegel to a view on which only entities that are objects of possible experience really exist. This is still weaker than Kreis’s claim (3), as it leaves untouched the possibility of truth-values for statements about experience-transcendent objects. Nevertheless, the resulting view would be at least similar to the position Kreis ascribes to Hegel, and it would contradict the essentialist view of Hegel with its claim that Hegel accepts experience-transcendent entities as constituting the basic structure of reality.

However, closer inspection shows that there is reason to be sceptical about this reading of the passage. To begin with, the addition to the paragraph reports the following remark on empiricism:

[P]erception is the form in which matters are supposed to be comprehended, and this is the deficiency of empiricism. Perception as such is always something individual and transitory; knowing, however, does not end with this but in the perceived individual seeks the universal, that which abides, and this is the progression from mere perception to experience. (Encyclopædia §38, Addition, 8:109 [Hegel 2010b:80; trans. emended])

As this remark shows, that “what is true must […] be there for perception” cannot mean that all truths about really existing entities are directly revealed in perception. Rather, genuine knowledge requires that what is perceptually given be transformed into universal contents of experience.

So far, this is entirely compatible, of course, with Kant’s views on cognition (both on Kreis’s and on other readings). But when it comes to the question of what concepts enable us to bring about the transformation of perception into experience, Hegel not only cites concepts like ‘causation’, ‘force’ or ‘matter’—i.e. concepts referring to theoretical entities that Kant admits among the objects of possible experience (see above), but also the concept of ‘genus’ (Gattung) (Encyclopædia §12, Remark, 8:57; §21, Addition, 8:77–8), which for Hegel is arguably only another term for objective determinate concepts, i.e. realistically and essentialistically understood universals (Knappik 2016a, n. 14).[16] Moreover, it can be argued (as I have already mentioned) that according to the Science of Logic such universals are superior in their explanatory power to all other types of theoretical entities that Hegel cites in this context.

We therefore can conclude that the principle “what is true must […] be there for perception” must be sufficiently flexible to allow that experience-transcendent entities (in Kant’s sense) like realistically and essentialistically understood universals can “be there for perception”. This is the case if the principle is spelt out in the following sense: something can really exist only if it is at least indirectly accessible to perception—i.e. only if it either is itself accessible to perception, or if it (as ground) has consequences that can be given in perception, and that can serve as basis for an explanatory inference which establishes the existence of the ground. If the passage from Encyclopædia §38 Remark is understood in this sense, it is entirely compatible with essentialist readings of Hegel.

Indeed, our discussion of that passage also warrants a stronger conclusion: far from speaking in favour of a reading similar to Kreis’s and against the essentialist reading, Hegel’s discussion of empiricism actually counts in favour of the essentialist reading. As I have argued, the principle “what is true must […] be there for perception” has to be seen in the context of Hegel’s view that the perceptually given is a mere starting point for cognition, and that genuine knowledge requires a transformation of perceptual contents into something ‘universal’. Given that Hegel counts concepts of realistically and essentialistically understood universals among the concepts that make possible such a transformation (and arguably considers them superior to all alternative candidates for that task), his views on empiricism and his essentialism are parts of one coherent position. By contrast, Kreis’s interpretation cannot do justice to the role that Hegel ascribes to concepts of realistically and essentialistically understood universals for our transformation of perceptual into universal contents: since they are experience-transcendent concepts (in Kant’s sense), they cannot make a constitutive contribution to genuine knowledge on his reading.

By way of conclusion, it might be asked if Kreis could not respond to the argument in this section by interpreting the notion of ‘experience-transcendent’ in claims (3) and (6) in the sense of Hegel’s notion of experience, which has turned out to be much broader than Kant’s (for as we have seen, ‘experience’ in Hegel’s sense is the result of a transformation of perceptual contents that can use concepts like that of ‘objective determinate concepts’, which are experience-transcendent in Kant’s sense.) This would render Kreis’s reading compatible with the essentialist view, and he could simply sidestep the objections from this section. The problem with this move is that it would render the restrictions of thought and philosophy to the objects of possible experience in claims (3) and (6) extremely weak. For once some realistically and essentialistically understood universals are admitted among the possible objects of experience, there is no clear reason why more fundamental universals should not also count as possible objects of experience. And given that some strong metaphysical readings hold that Hegel’s Absolute consists precisely in a most fundamental universal that is the basis of all objective determinate concepts (and hence also of all individual entities) (e.g. deVries 1991), Kreis’s modified reading would have to count the Absolute among the possible objects of experience, which is certainly not in the spirit of his proposal.

Invited: 28 April 2016; received: 27 December 2016; revised: 24 September 2017.


[1] Kreis replaces Kant’s notion of “conditions” in this context by that of “causes” (“Ursachen“, pp. 42ff.), but this is misleading: on Kant’s understanding of “cause”, only the conditions that are relevant for the third antinomy are causes, while the other antinomies refer to other types of conditions (earlier points of time, limiting areas in space, material parts, and necessitating conditions; see A411ff./B438ff.).

[2] That is, a form in which the antithesis, but not the thesis, is restricted to objects of possible experience (cf. A560/B588).

[3] When it comes to resolving the dynamic antinomies, Kant writes, the distinction between the mathematical and the dynamic ideas “comes to be important, and opens up for us an entirely new prospect in regard to the suit in which reason has become implicated” (A529/B557).

[4] I am presupposing here a reading of the distinction between “objectively sufficient” and merely “subjectively sufficient” grounds of judgement, on which the latter are reasons that the subject takes to be truth-conducive, while merely subjectively sufficient grounds are reasons that are not seen as truth-conducive by the subject. Merely subjective sufficiency of the grounds for a judgement in this sense does not rule out that such grounds can make a judgement rationally obligatory for all subjects (also by the subject’s own lights), as is the case when these grounds are rooted in the moral law. For a defence of this reading, see Knappik & Mayr,  forthcoming, Section VI.2.

[5] The dynamic ideas leave open the possibility that there is a last condition in the series that is not itself an appearance (A530/B558).

[6] Kant formulates the principle while giving an example of questions (apart from those of transcendental philosophy) that must not be left open—namely, questions about “what is lawful [Recht] or unlawful [Unrecht]” (trans. amended), by which he may mean either moral or legal questions, or both.

[7] For Kant, this is the reason why moral/legal questions must not be left open through an appeal to ignorance (cf. the preceding footnote): it must always be possible to decide questions about whether an action is lawful or not, for we cannot be necessarily ignorant about a law, and still be bound by it.

[8] E.g. Bxxvi–xxvii, A288/B344, A538/B566.

[9] Cf. also Kreis’s critique of Pippin in footnote 42 on p. 206.

[10] I shall presently say more on how this applies to (7).—It may seem that (1) is not compatible with a strong metaphysical reading: for one might think that experience-transcendent claims play simply no role in our ordinary conceptual schemes that are the target of Hegel’s descriptive metaphysics. But a historically accurate reading should take into account that when Hegel aims to reconstruct the basic categories of human thought, he can use as a basis for this reconstruction only human thought as it was performed and expressed before and in his own time. And within this frame of reference, it was entirely normal for people to use pairs of categories like ‘God’ and ‘finite (created) world’ to articulate their world-views—and to enact these aspects of their world-view in the practices of art, religion and philosophy that Hegel analyses as forms of absolute spirit. That a majority of the population in the Western countries has abandoned the relevant categories and practices since Hegel’s time may suggest that Hegel was wrong to consider them basic (and hence presumably: necessary) aspects of our world-views; but it does not show that Hegel’s metaphysics is not at the same time ‘strong’ and descriptive.

[11] Strictly speaking, such Hegelian universals are linked in two ways to essences: they are themselves non-artificial (‘natural’) kinds with essences; and it is part of the essence of any individual instance of such a universal that it is an instance of this universal (cf. Knappik 2016a). As nothing hinges on this in the present context, I shall in the following bracket this point.

[12] On readings like these, such universals are seen along Aristotelian rather than Platonist lines (e.g. Stern 1990:59), and the relevant notion of essence is distinguished from the targets of Hegel’s criticisms in the Logic of Essence (Knappik 2016a).

[13] Cf. also Log, AA 9:143–4 and A277–8/B333–4.

[14] An additional challenge to Kreis’s reading, which I shall not pursue further here, comes from Hegel’s discussions of Kant. Hegel explicitly rejects the claim that the possibility of knowledge (and, a fortiori, of thoughts with a truth-value) should be seen as conditioned by experience: “Turned against reason, […] understanding behaves in the manner of ordinary common sense, and asserts its view that truth rests on sensuous reality, that thoughts are only thoughts, that is, that only sense perception gives content and reality to them; that reason, in so far as it abides in and for itself, generates only mental figments. In this self-renunciation of reason, the concept of truth is lost […]” (Science of Logic, 5:38 [Hegel 2010a:25; trans. amended]).

[15] Indeed, I have argued elsewhere (Knappik 2013, ch. 9) that according to Hegel’s analysis of the specific form of freedom that we gain through philosophy, such freedom even requires reference to experience-transcendent entities, including a form of ontological monism.

[16] Notice that the context in the relevant passages makes it clear that Hegel is thinking of a constitutive, not a merely regulative use of this concept.


Allais, L. (2015), Manifest Reality: Kant’s Idealism and His Realism (Oxford: Oxford University Press).

Allison, H. (2004), Kant’s Transcendental Idealism. An Interpretation and Defense, revised and enlarged edition (New Haven: Yale University Press).

Bealer, G. (2002), ‘Modal Epistemology and the Rationalist Renaissance’, in T. Gendler and J. Hawthorne (eds), Conceivability and Possibility (Oxford: Oxford University Press), pp. 71–125.

Beiser, F. (2005), Hegel (New York/Abingdon: Routledge).

Bowman, B. (2013), Hegel and the Metaphysics of Absolute Negativity (Cambridge: Cambridge University Press).

DeVries, W. (1991), ‘The Dialectic of Teleology’, Philosophical Topics 19(2): 51–70.

Ellis, B. (2001), Scientific Essentialism (Cambridge: Cambridge University Press).

Förster, E. (2012), The 25 Years of Philosophy: A Systematic Reconstruction, trans. B. Bowman (Cambridge, MA: Harvard University Press).

Hanna, R. (2006), Kant, Science, and Human Nature (Oxford: Oxford University Press).

Hegel, G. W. F. (1986), Werke, 20 vols, ed. E. Moldenhauer and K.M. Michel, (Frankfurt: Suhrkamp).

——— (2010a), The Science of Logic, trans. and ed. G. di Giovanni (Cambridge: Cambridge University Press).

——— (2010b), Encyclopedia of the Philosophical Sciences in Basic Outline. Part I: Science of Logic, trans. and ed. K. Brinkmann & D. Dahlstrom (Cambridge: Cambridge University Press).

Horstmann, R.-P. (1984), Ontologie und Relationen. Hegel, Bradley, Russell und die Kontroverse über interne und externe Beziehungen (Königstein: Athenäum/Hain).

Knappik, F. (2013) Im Reich der Freiheit. Hegels Theorie autonomer Vernunft (Berlin/Boston: de Gruyter).

——— (2016a), ‘Hegel’s Essentialism. Natural Kinds and the Metaphysics of Explanation in Hegel’s Theory of “the Concept”’, European Journal of Philosophy 24(4): 760–87.

——— (2016b), ‘And Yet He is a Monist. Comments on James Kreines, Reason in the Word’, Hegel Bulletin, online first: link, accessed 27 December 2016.

Knappik, F. & E. Mayr. (forthcoming), ‘“An Erring Conscience is an Absurdity”. The Later Kant on Moral Judgment, Certainty and the Infallibility of Conscience’, Archiv für Geschichte der Philosophie.

Kreines, J. (2004), ‘Hegel’s Critique of Pure Mechanism and the Philosophical Appeal of the Logic Project’, European Journal of Philosophy 12(1): 38–74.

——— (2008), ‘The Logic of Life: Hegel’s Philosophical Defense of Teleological Explanation of Living Beings’, in F. Beiser (ed.), The Cambridge Companion to Hegel and Nineteenth-Century Philosophy (Cambridge: Cambridge University Press), pp. 344–77.

——— (2015), Reason in the World. Hegel’s Metaphysics and its Philosophical Appeal (New York: Oxford University Press).  

Langton, R. (1998), Kantian Humility. Our Ignorance of Things in Themselves (Oxford: Clarendon Press).

Pippin, R. (1989), Hegel’s Idealism. The Satisfactions of Self-Consciousness (Cambridge: Cambridge University Press).

Prauss, G. (1974), Kant und das Problem der Dinge an sich (Bonn: Bouvier).

Rosefeldt, T. (2007), ‘Dinge an sich und sekundäre Qualitäten’, in J. Stolzenberg (ed.), Kant in der Gegenwart (Berlin/New York: de Gruyter), pp. 167–209.

Stern, R. (1990), Hegel, Kant and the Structure of the Object (London/New York: Routledge).

——— (2007), Hegelian Metaphysics (Oxford: Oxford University Press).

Strawson, P.F. (1966), The Bounds of Sense. An Essay on Kant’s Critique of Pure Reason (London: Methuen).

Van Cleve, J. (1999), Problems from Kant (New York: Oxford University Press).

© Franz Knappik, 2017.

Franz Knappik is Associate Professor of Philosophy at the University of Bergen, Norway. He specialises in Kant and German Idealism, and in philosophy of mind. His publications include: ‘Hegel’s Essentialism. Natural Kinds and the Metaphysics of Explanation in Hegel’s Theory of “the Concept”, European Journal of Philosophy (2016) and ‘”An Erring Conscience is an Absurdity”: The Later Kant on Certainty, Moral Judgment and the Infallibility of Conscience’ (co-authored with Erasmus Mayr), forthcoming in Archiv für Geschichte der Philosophie.