2. Does an Intuitive Intellect Represent Everything?

It is uncontroversial that Kant believed that an intuitive intellect would make no use of the categories. At B145, for instance, he states that ‘the categories would have no significance at all with regard’ to the cognition of an intuitive intellect. To generate a challenge to the strong metaphysical reading, Kohl claims that, for Kant, ‘Noumena have categorial properties (i.e., the properties that we represent through the pure categories) only if an intuitive, divine intellect would represent them as having such properties’ (Kohl Reference Kohl2015: 91). In support of that claim, Kohl argues that ‘an intuitive intellect would have a complete maximum of cognition: it would cognize every property of every thing in itself’ (Reference Kohl2015: 91). This latter claim, however, is ambiguous. Once it is disambiguated, a proponent of the strong metaphysical reading can sidestep Kohl’s challenge.

To appreciate the ambiguity, imagine a world in which there are just three fundamental facts: A, B and C. Given those fundamental facts, there will also be many non-fundamental facts, such as the fact that A-and-C and the fact that B-or-D-or-E. These non-fundamental facts hold in virtue of the three fundamental facts. Now imagine a subject who makes exactly three judgements, corresponding to each of the three fundamental facts. Has this subject represented all the facts? In one sense, she has not, for she has not made judgements corresponding to any of the (many) non-fundamental facts. On the other hand, there is another sense in which she has indeed represented all the facts, because she has represented all the fundamental facts, and all non-fundamental facts hold in virtue of some fact that she does represent. We can say that, in this case, the subject directly represents the fundamental facts, but indirectly represents the non-fundamental facts.

Next consider a historical example. In the Ethics, Spinoza distinguishes three kinds of knowledge (Scholium 2 to Proposition 40 of Book 2). The first is fallible empirical knowledge, based on the senses and signs. The second is based on ‘the fact that we have common notions and adequate ideas of the properties of things’. The third, which Spinoza dubs scientia intuitiva (intuitive knowledge), ‘proceeds from an adequate idea of the formal essence of certain attributes of God to the adequate knowledge of the essence of things’. Spinoza illustrates these different kinds of knowledge with an example. In solving for x in an equation of the form a/b =c/x, someone with only the first kind of knowledge might just apply a rule she remembers from school. Someone with the second kind of knowledge will appeal to ‘the common property of proportionals’, namely, that a*x = b*c (for this, Spinoza cites Proposition 19 of Book 7 of Euclid’s Elements). But for someone with the third kind of knowledge, ‘none of this is necessary’, since she will grasp the solution ‘in once glance’. For example, if a = 2, b = 1 and c = 4, she will see in one glance that x = 2. Both the second and third kinds of knowledge, Spinoza holds, are infallible (Proposition 41, Book 2). But Spinoza thinks the third kind is more divine than the second – later in the Ethics, Spinoza locates the ‘intellectual love of God’ in the third kind of knowledge (Scholium to Proposition 32, Book 5), saying that this is ‘the very love … by which God loves himself’ (Proposition 36, Book 5). Arguably, then, Spinoza thinks that divine knowledge would be exclusively knowledge of the third kind.Footnote ⁶

For Spinoza, then, would a being with only the third kind of knowledge know or even represent the common property of proportionals, the representation of which is used to define the second kind of knowledge? It seems that such a subject would not explicitly represent that property, since she would only see the relations between particular numbers. If her intellect were infinite, though, such that she immediately saw all the relations between all particular numbers, then there would be a sense in which she would represent that common property by representing all the particular relations it concerns. Here again, we could say that such a subject would directly represent particular numerical relations, and indirectly represent the common property of proportionals.

In both examples, the distinction between fundamental and non-fundamental is naturally understood as being independent of finite minds. It is not merely that we project some distinction between, say, the fact that B and the fact that B-or-D-or-E. Rather, it seems to be a mind-independent truth that disjunctive facts are less fundamental than the disjuncts that make them true. Similarly (though more controversially), general principles about proportionals seem to hold in virtue of particular facts about numbers. Of course, this way of thinking about fundamentality and non-fundamentality could be challenged, but it is a familiar view.Footnote ⁷ This makes room for the idea of noumenal (and so strongly mind-independent) facts that are non-fundamental.

With the above examples in place, we can see the ambiguity of the claim that an intuitive intellect would cognize every property of every thing in itself. That claim could mean either:

1. (a) an intuitive intellect would directly cognize every property of every thing in itself; or

2. (b) an intuitive intellect would directly or indirectly cognize every property of every thing in itself.

For Kohl’s challenge to go through, it must be clear that Kant accepts (a). Otherwise, a proponent of the strong metaphysical reading can say that, for Kant, the intuitive intellect directly represents things in themselves non-categorially, but indirectly represents facts about the substantiality and causality of noumena. This might be because the facts about noumenal substances and noumenal causes are all non-fundamental (but are facts nonetheless). Presumably, this would also imply that the intuitive intellect would only indirectly represent facts about fundamentality and non-fundamentality, such as the fact that B-or-D-or-E is non-fundamental. This is coherent, at least if facts about non-fundamentality themselves also hold in virtue of fundamental facts.Footnote ⁸ (Note that I am sliding here between talk of representing facts and talk of representing objects, but nothing substantive turns on this.)

While Kohl does not discuss the above ambiguity, he does offer textual support for the claim that, for Kant, an intuitive intellect would cognize every property of every thing in itself. We should therefore consider whether Kant’s statements about the scope of the intuitive intellect support disambiguation (a) over (b). Perhaps the clearest claims on this point come from the Pölitz lectures, where Kant says that an intuitive intellect would have ‘a maximum of understanding’ (28: 7), would be ‘unlimited’ (28: 52) and would ‘cognize everything at once’ (28: 103).Footnote ⁹ None of these claims are helpful in resolving the ambiguity, however, since disambiguation (b) can be understood as describing an intellect that has a maximum, unlimited understanding of everything – after all, there is no thing that falls outside the scope of its understanding. Similarly, Spinoza shows no signs of regarding the third kind of knowledge as limited in any respect, even though it does not directly represent facts about common properties.

Moreover, there are grounds for thinking that Kant must accept (b). Consider the non-noumenal principle of the Second Analogy: ‘All alterations occur in accordance with the laws of the connection of cause and effect’ (B232). Kant clearly holds that this principle is true (at least, true relativized to possible experience – see A156–8/B195–7). To directly represent it, however, would require using <causation>. An intuitive intellect, therefore, could not directly represent this truth. However, if, as Kant suggests, an intuitive intellect would be omniscient, then it would have to represent this non-noumenal truth without using <causation>, and so represent it indirectly. Kant never suggests the heretical view that facts about the empirical world would be inaccessible to the divine intellect. Similarly, Kant takes it to be a fact that subjects like us, with discursive intellects, represent using categories like <causation>. Directly representing that fact about us would likewise seem to require using the categories, so an omniscient intuitive intellect would have to represent this fact about us indirectly. Whether or not this fact is noumenal, it is not empirical in any straightforward way. God must therefore have some way of representing the categories that does not itself require use of the categories. Such indirect representation of representations is not particularly mysterious. By analogy, a philosophical anthropologist might successfully but abstractly refer to the set of empirical concepts that ceased to exist before the invention of writing.

To be sure, it is not obvious how God’s indirect representation of such facts would work, but Kant emphasizes that our grasp of the workings of the divine intellect is extremely limited:

Since an intuitive intellect would not use the categories but would nonetheless be omniscient, then the intuitive intellect must indirectly represent categorial non-noumenal facts such as the principle that every alteration has a cause. Similarly, an intuitive intellect must be able to represent non-empirical facts about the representational capacities of discursive intellects. If so, then surely the intuitive intellect could also indirectly represent categorial noumenal facts.

It may be possible to say more about how the intuitive intellect’s indirect representation works. In some places, Kant suggests that our conceptual representation differs from intuitive intellectual representation merely in terms of abstraction. For instance, in Metaphysics Vigilantius, Kant writes:

Here, Kant seems to be saying that abstract representations (all concepts, and hence including the categories) are the result of the mind’s finitude, having to set aside ‘much of the things’ in order to cognize them. On this point, at least, Kant’s view resembles that of John Locke (though Kant of course differs from Locke on the apriority of abstract representations). In the Essay, Locke writes that,

Locke thus holds that our limited capacity for names leads us to use abstract ideas and general terms. That would seem to suggest that an infinite mind would have no need for abstract ideas (though Locke seems wary of affirming much about the divine mind beyond omniscience – see Essay 4.10). While abstract ideas do not capture everything about their objects, Locke does not claim that abstract ideas fail to represent real features of objects. Abstract ideas, for Locke, accurately capture real similarities between particular things, though not their real essences (see 3.3.13–15). Hence, like Spinoza, Locke seems to think that an infinite mind would have no need of abstract ideas, yet does not think that precludes abstract ideas from genuinely applying to the mind-independent world. God would then indirectly represent abstract facts by directly representing particular things in all their details. There is therefore historical motivation to attribute a view along these lines to Kant.

Kohl considers something like the possibility just described, that is, of explaining the difference between human and divine intellects in terms of abstraction. Kohl rejects this abstractionist interpretation, however, because he holds that ‘for Kant general [and so abstract] representations need not be conceptual’ (Kohl Reference Kohl2015: 98). To support this, he appeals to Kant’s talk of synthetic universals in the third Critique: ‘we can … conceive of an understanding which, since it is not discursive like ours but intuitive, goes from the synthetically universal (of the intuition of a whole as such) to the particular, i.e., from the whole to the parts’ (5: 407). Presumably, then, Kohl thinks that this shows that the intuitive intellect could indeed have abstract representations. The 5: 407 passage, however, only shows that Kant thinks an intuitive intellect would represent a whole prior to the parts. That does not show that the intuitive intellect would represent anything as general or abstract – if anything, it would represent some particulars (parts) in relation to another particular (the whole). Therefore, I think a proponent of a strong metaphysical reading could understand the intuitive intellect as representing all things as particulars, and only indirectly representing the generalities we represent abstractly. This might bring Kant’s views of divine cognition quite close to Spinoza’s (and perhaps Locke’s).

My evaluation of Kohl’s challenge does not hinge on the above characterization of the intuitive intellect, however. What is it does hinge on is the fact that none of Kant’s statements about an intuitive intellect support disambiguation (a) over (b). If so, then a proponent of the metaphysical interpretation need not be worried by Kant’s statements that the intuitive intellect would not represent using the categories, and can therefore sidestep Kohl’s main argument. Of course, that does not show that the strong metaphysical interpretation is correct, but it does show that a successful argument against it must come from elsewhere.

There are two responses that an opponent of strong metaphysical readings might make at this point. First, she might deny that there are any non-fundamental noumenal facts. Since Kant arguably does not have a distinct term for the notion of fundamentality I appealed to above, he never explicitly addresses the question of whether there are non-fundamental noumenal facts. That makes room for a reading on which all noumenal facts are fundamental, or on which the concepts <fundamental> and <non-fundamental> do not apply to noumena. After all, the notions of fundamentality and non-fundamentality seem to concern a sort of dependence, and Kant thinks at least one sort of dependence is captured by the relational categories (see A80/B106). That, however, would pose a problem for the strong metaphysical reading only if we already assumed that the categories did not apply to noumena – but that would be to beg the question against the strong metaphysical reading. The present issue is not whether there are defensible alternatives to the strong metaphysical reading, but whether Kant’s claims about the intuitive intellect generate a problem for the strong metaphysical reading. Kant’s lack of any explicit rejection of non-fundamental noumenal facts is thus enough for strong metaphysical readings to be safe from Kohl’s challenge.

Perhaps a proponent of the strong metaphysical reading could go further than this. For example, in the second Critique, Kant says that freedom is the ratio essendi (essential ground) of the moral law, where both freedom and the moral law seem to be noumenal (5: 4n.). At least in his early work, Kant is explicit that a ratio essendi is metaphysically prior to what it grounds, calling it an ‘antecedently determining ground’, ‘the reason why’ and ‘the ground of being’ (Nova Dilucidatio, 1: 392). This suggests that Kant believes in at least one instance of a non-fundamental noumenal fact.Footnote ¹⁰

The second response an opponent of strong metaphysical readings might make here would be to deny that omniscience requires any representation of empirical facts or facts about subjects’ capacities. If so, then an intuitive intellect need not even indirectly represent these facts, which would partly undermine my case for there being a distinction between direct and indirect representation in Kant. Kant does say, after all, that we can speak of space and time ‘only from the human standpoint’ (A26/B42) and that ‘the temporal condition is nothing’ for an eternal being (second Critique, 5: 123), so perhaps there is room to deny that God represents any spatiotemporal facts, even indirectly. This, however, would be a radical break from the theistic philosophical tradition, according to which God certainly knows the spatiotemporal world. In the second Critique, Kant says that God’s omniscience is a quality that is found in finite creatures, but ‘raised to the highest degree’ (5: 131). That seems to imply that God knows (and therefore represents) everything that creatures know, and more. A proponent of the strong metaphysical reading would therefore have significant textual support for holding that God must indirectly represent various facts, and could read the relevant passages as saying that only humans directly represent temporal matters. The possibility of indirect representation and non-fundamental noumenal facts therefore undermines Kohl’s argument against the strong metaphysical reading.

Even though Kohl’s primary argument is not successful, Kohl offers a secondary argument that does not appeal to the scope of the intuitive intellect, but rather to how Kant contrasts the categories themselves with the representations belonging to an intuitive intellect.

3. Is the Content of the Categories Too Subjective to Represent Noumena?

Throughout the Transcendental Deduction of the Categories, Kant emphasizes that the role of the categories hinges on the fact that our intellect is not intuitive. This provides the basis for a second, related challenge that Kohl offers to strong metaphysical readings. Kohl holds that some of Kant’s statements about the nature of the categories show that, by their very nature, the categories could not be applied to things in themselves. If so, then Kohl might still offer an important argument against strong metaphysical readings, even if his main argument about the intuitive intellect is unsuccessful. To be clear, the main reason Kohl himself draws on these statements is to show why the categories would not be used by an intuitive intellect, and I think he makes this case convincingly. However, Kohl also thinks that they show why categories would not apply to noumena (the objects of an intuitive intellect), and it is this further argument that I believe is problematic. Seeing why, however, sheds light on the nature of the categories.

Kohl draws our attention to some striking statements Kant makes about the subjective nature of the categories. Kant says that the pure categories ‘contain exclusively the synthetic unity of apperception’ (B148), that, in the unschematized categories, ‘nothing can be found other than the mere form of thought’ (A567/B595), and that the pure category ‘can contain nothing but the logical function to bring the manifold under a concept’ (A245).Footnote ¹¹ As Kohl reads these passages, Kant is saying that the categories ‘signify purely discursive features’ (Kohl Reference Kohl2015: 99), that is, that the categories are only about features of a discursive mind such as ours (apperception, forms of thought and logical functions all being tied to discursivity). As such, they would be incapable of representing noumena outside our mind.

This reading of Kant faces at least three complications. First, Kant’s claims about the connection between reason and our noumenal sides (e.g. A546–7/B574–5) suggest that discursive faculties are noumenal (noumenal powers). In that case, the categories would be about noumena in virtue of being about our discursive faculties. Second, even the claim that Kohl takes Kant to be making (that the categories signify purely discursive features) is ambiguous between direct and indirect representation. It could be read as saying that the categories directly signify only features of a discursive mind, but might indirectly signify other things, such as non-mental features of noumena. Third, part of Kohl’s support for this reading comes from Kant’s claim at B149 that the categories could not be applied to ‘an object of non-sensible intuition’ (similar passages appear elsewhere: e.g. A287/B343). That claim, however, can be read in the way that metaphysically inclined interpreters now standardly read Kant’s claims about the limitation of the categories: saying that the categories do not yield cognition of such objects, which is consistent with the possibility of us having true thoughts about such objects using the categories (see e.g. Adams Reference Adams1997: 807–8).

All that said, I think Kohl’s reading of these passages is defensible. If it were the only defensible reading, then proponents of the strong metaphysical reading would indeed face another important challenge. But the passages have at least one other defensible reading. Contrast these two claims:

1. (1) The Hobbit contains nothing but events in Middle Earth.

2. (2) The Hobbit contains nothing but words.

(1) makes a claim about what the novel represents (its ‘content’). (2) makes a claim about what constitutes the novel (the ‘vehicle’). The constituents of a representation may or may not set any limits to what it can represent – being made out of words, a book can arguably represent anything whatsoever, whereas the constituents of a painting (paint) might keep it from representing some non-visual properties. To be sure, the nature of the content/vehicle distinction, as well as the question of what exactly constitutes things like novels, is controversial. Yet it is hard to deny that there is some distinction along these lines. (There may be additional analogous claims worth considering – for instance, that The Hobbit contains nothing but escapism, or that it contains nothing but J. R. R. Tolkien’s ideas.)

In light of this, we can see that Kant’s statements about the subjectivity of the categories can be understood on analogy with (2) instead of (1). Kant would not, then, be saying that unschematized categories merely represent features of a discursive mind, but rather that those categories are constituted by certain features of a discursive mind. This constitution claim would be consistent with the categories representing features of mind-independent things, just as words can represent non-linguistic entities. Moreover, this reading seems to me to be more attractive. It sounds prima facie absurd to say that, when I think about causes, I am thinking about apperception or logical functions to bring the manifold under concepts. On the other hand, it does not sound absurd to say that when I think about causes, I am employing a representation constituted by a form of self-consciousness or a logical function (though that is not to say it sounds immediately plausible, either). Because Kant’s statements can be read on analogy with (2), therefore, these statements by themselves pose no direct threat to strong metaphysical readings of Kant. As before, this does not prove that Kohl’s reading is incorrect, but it does provide a way for a proponent of the strong metaphysical reading to sidestep his challenge.

Perhaps anticipating a response along these lines, Kohl notes Kant’s claims about space and time being mere forms of intuition (e.g. ‘Space is nothing other than merely the form of all appearances of outer sense, i.e., the subjective condition of sensibility’: A26/B42), where these claims seem intended to bear on why space and time are not features of things in themselves. That is, it would be because space and time are constituted by subjective forms that they do not apply to things in themselves. This may be a defensible reading of Kant’s argument in the Aesthetic (one that resembles Allais Reference Allais2010). However, Kant makes a point of separating his claims about the constitution of the representation of space and time from his claims about their content – e.g. ‘Space represents no property at all of any things in themselves’ (A26/B42), which is listed as a separate conclusion from the claim about space being a subjective condition of sensibility. No comparably clear separation appears in Kant’s claims about the categories, so the comparison with space and time is dubious.

An opponent of the strong metaphysical reading might offer two responses at this point. First, she might object that the ‘vehicle’ reading I have suggested would make Kant’s claims about the categories trivial. After all, what else could categories be constituted by except something like forms of thought? The claims would not be trivial on this reading, though. For one thing, the categories being solely constituted by forms of thought would preclude them from having sensory representations or marks as constituents (in contrast with empirical concepts). For another, it is not trivial to say that certain concepts are constituted by forms of judgement. On some pre-Kantian views, fundamental concepts are prior to any judgements, thoughts and/or self-consciousness. Kant, however, would be surprisingly (and so, non-trivially) reversing the order of metaphysical priority here.

Second, an opponent of the strong metaphysical reading might push back on the apparent absurdity of glossing thoughts about causes in terms of thoughts about self-consciousness or forms of judgement. After all, the relevant form of thought for <causation> is the hypothetical form of judgement, and ‘if … then’ thoughts do seem intuitively related to thoughts about causes. There is an ambiguity here, however. Having an if-then thought about some objects could be a matter of thinking that those objects are connected in an if-then way, or it could be a matter of thinking that I think of them in an if-then way. Kohl’s challenge requires the second and less attractive gloss. As before, this reading may be defensible, but it is unattractive enough that the strong metaphysical reading remains a live option.