Candidates for the Transcendental Function of ThIs

If we are to start afresh by focusing upon identifying the transcendental function ThIs play, what clues do we find in Kant’s text? A transcendental function is expressed in terms of a necessary condition, and we find two types of such claims in relation to ThIs. The first is summarized in Kant’s claim that ‘the word “appearance” must already indicate a relation to something … which in itself must be something, i.e. an object independent of sensibility’ (A252). We shall examine this passage in detail in section 2, but what is important for now is to note that the text makes the claim that it is through the relation of appearing that we are led to consider ThIs. On its own, however, this is uninformative since it amounts to saying that ThIs are what appears as appearances.

The second transcendental function is what is referred to as transcendental affection in the literature. This is the requirement that something that is independent of the mind should affect our sensibility, something that Jacobi (Reference Jacobi and di Giovanni1994) famously claimed was needed to enter the Critical system. In the first section of the Transcendental Aesthetic, Kant produces a short transcendental argument by claiming that the object’s being given to us is ‘possible only if it affects the mind in a certain way’ (A19/B33). While a straightforward identification of the object with the ThI, which is the most widespread interpretation of this passage (Gardner Reference Gardner1999: 287, Longuenesse Reference Longuenesse1998: 22, Westphal Reference Westphal2004: 52) is perhaps questionable,Footnote ³ the claim that the mind is affected by something implies that there is some mind-independent thing producing sensations, and this can only be the ThI. There are several other occurrences of this claim in Kant’s theoretical philosophy (e.g. B207, A253/B309; Prolegomena (hereafter, P), 4: 314–15, 318).Footnote ⁴

A third type of textual resource should also be considered for, although it does not explicitly define a transcendental condition, it defines another possible transcendental function of ThIs. This is the claim that appearances are grounded in ThIs, thereby suggesting that their function lies precisely in this grounding role: ‘[t]he transcendental object … is … an unknown ground of … appearances’Footnote ⁵ (A379–80, see also A537/B565; P, 4: 315).

A Unifying Account of the Transcendental Function of ThIs

Calling the above three candidate claims for the transcendental function of ThIs (T1), (T2) and (T3) respectively, one must ask how they are related. Looking at (T2) and (T3) first, since they sometimes appear together e.g. in the Prolegomena (4: 314–15), it is often assumed that (T2) entails (T3), as indeed that passage might suggest. But that would mean that the fact that ThIs cause sensations accounts for their being the ground of appearances.Footnote ⁶ This entailment is not obvious because it derives a causal link between ThIs and appearances, i.e. empirical objects, from a causal link between ThIs and sensations. The missing link from sensations to empirical objects can hardly be described as causal, since it involves acts of spontaneity bringing syntheses under the unity of apperception. It is therefore not clear how to infer (T3) from (T2). Looking at the text from the Prolegomena (4: 314–15), Kant says that concluding to (T3) is ‘unavoidable’ (unvermeidlich); this is perhaps best read as claiming that (T3) is the only possible explanation for (T2). Adopting that interpretation, we need to find a way of interpreting (T3), i.e. the claim that ThIs ground appearances in such a way that it explains how it is that ThIs affect our sensibility.

To do that, it is useful to note that Kant has an empirical account of the origin of sensations (A28, B1, A213/B260; see also Aquila Reference Aquila1981: 13). This is commonly referred to as empirical affection and is the fairly uncontroversial claim that our sensations occur in time as a result of the causal impact of spatiotemporal objects. Drawing upon this notion at this stage might appear to complicate matters insofar as it seemingly brings with it the famous problem of reconciling the two types of affection, transcendental and empirical, i.e. the problem of double affection (cf. Adickes 1924). But in fact, we can avoid this problem precisely by showing in what way empirical affection helps us understand how the transcendental grounding of appearances (T3) explains transcendental affection (T2).

We therefore have, in effect, claims about three causal links. These links are: the grounding of appearances in ThIs, i.e. (T3), ThIs causing sensations through transcendental affection, i.e. (T2), and empirical objects/appearances causing sensations through empirical affection. I propose to interpret the grounding claim (T3) as the grounding of empirical causality:Footnote ⁷ this implies that all empirical causal links, especially those of empirical affection, are thereby grounded in the ThI.Footnote ⁸ This in turn entails that the effect of such empirical affection, the sensation, is thereby causally grounded in the ThI. That is (T3), interpreted in this way, provides an explanation for the claim of transcendental affection (T2).Footnote ⁹

This interpretation of the ThI as the ground of empirical causality avoids any problem of double affection since transcendental affection is a particular feature of the ThI’s grounding of all empirical causality, namely that it thereby grounds empirical affection. It is thus by being grounded in the ThI that the empirical affection of our sensibility takes on that transcendental dimension which is transcendental affection. In this respect, this proposal has a distinct advantage in comparison with the traditional interpretations. This is particularly true with respect to TW, which is often accompanied by an attempt to resolve the problem of double affection,Footnote ¹⁰ but it is also true for dual-aspect theories: to claim that something is produced simultaneously by two different aspects of things also suggests two distinct causal stories that need reconciling.Footnote ¹¹

Where does that leave claim (T1)? This claim, that ThIs are what appears as appearance, has not been given much attention in the literature, although it clearly needs elucidating if it is not to be tautologous. The interpretation I propose suggests a straightforward answer. Insofar as we have an account of how ThIs ground a key feature of empirical substances, namely their causality, and insofar as this explains how we are affected transcendentally so that we can thereby cognize objects as appearing to us, it follows that this grounding is best interpreted as the relation of appearing.

Filling in the Interpretation

I have proposed to view ThIs as the ground of empirical causality. Because this interpretation does not involve referring to a ThI as a distinct substance, it is natural to view it as a type of metaphysical dual-aspect interpretation. It is distinguished from existing versions of MDA in that it makes no claims at all under the category of substance.Footnote ¹² The ThI is just posited as causal ground; as a result, it is also not possible to differentiate ThIs on the basis of empirical differences: that is why I refer indifferently to the ThI or ThIs, since we know nothing about it under the categories of quantity. All that is claimed is that some causal ground of empirical causality exists in-itself. Worries about the possibility of such knowledge in a Kantian context are examined in section 4.

One might ask whether this grounding of empirical causality is only one part of a more general grounding of phenomena (T3). Here it is important to emphasize the key role of causality in Kant’s critical system. All objective determinations of substance, insofar as they are actualized at some point in time, have an empirical cause according to the Second Analogy. This means that they are thereby also grounded in the ThI. So, aside from the very existence of substance, this function of grounding empirical causality accounts for how all objective determinations of phenomena are grounded in ThIs.Footnote ¹³

The Grounding of Causality

In this sub-section, I examine further textual support for my claim that the ThI grounds empirical causality, independently of the dual-aspect claim. At A537/B565, in the Third Antinomy, Kant explicitly draws upon this grounding to provide a general solution to this antinomial conflict before, in the next two sections of that discussion, focusing upon the causal account of free actions.

The general solution must be one that finds a first (uncaused) cause for empirical causal regresses in line with the Thesis’s requirement for completeness of the causal determination of events (A446/B474), while conforming to the Antithesis’s claim that all events in the world are governed by the laws of nature and must therefore have a cause (A445/B473). Kant’s solution consists in showing that a reconciliation is possible if the first cause is outside the causally governed domain of appearances. Since the First Antinomy excludes this first cause being only effective at a first moment in time, it must be the case that the ‘causality of the cause’ (A444/B472) of any event considered in the causal regress of the Thesis can be referred to such an in-itself ground. That is why Kant spells out his general solution as follows: ‘appearances … [i.e.] representations connected in accordance with empirical laws … must have grounds that are not appearances’ (A537/B565).

This amounts to a grounding of the causality of (empirical) causes in the ThI, i.e. a clear statement of the fact that the ThI’s grounding of appearances involves a grounding of empirical causality. There is no space to give a more detailed analysis of the Antinomy in this article,Footnote ¹⁴ but as further support, note that in the Groundwork, Kant states: ‘But because the world of understanding [the noumenal world] contains the ground of the world of sense and so too its laws…’ (4: 453). We here find a similar emphasis on the causal laws being grounded in the ThI: it is because the grounding of the world of sense is one of empirical causality that causal laws are thereby grounded.

In the next two sub-sections, I shall focus upon the Phenomena and Noumena (PN) chapter of the first Critique, where Kant looks back upon what he has achieved in delimiting the ‘land of truth’ (A235/B294–5) with the intention of consolidating the results he has obtained, and thereby of clearly identifying the limits of this domain. This will provide support for the metaphysical dual-aspect claims I made in section 1, independently of the grounding claim, and enable a further development of this interpretation.

Positing the Transcendental Object

In the A version of the PN chapter, Kant makes the following claim already cited in section 1: ‘If there is not to be a constant circle, the word “appearance” must already indicate a relation to something … which in itself must be something, i.e. an object independent of sensibility’ (A252). The interpretation outlined in section 1 led to understanding the relation of appearing as the grounding of empirical causality; here, I want to consider this claim in the context of the PN chapter and independently of that interpretation. The principle of charity invites us to consider that this statement must be more than a merely semantic, and thereby analytic, claim about what is entailed by calling something an appearance. Because the obvious reply is, then, why call the objects of our discursive knowledge appearances in the first place? This claim is presumably nothing new: in this chapter, Kant is, as already indicated, clarifying the cognitive limits he has already identified. The statements preceding this claim, about the role of the transcendental object (TO), confirm that what is here spelt out is a consequence of a core result of the Transcendental Deduction in its A version (A-TD).

Kant defines the TO as the ‘entirely undetermined thought of something in general’ (A253). It is also ‘the object to which I relate appearance in general’ (A253). This is where Kant refers to the A-TD, where he characterizes it as the object to which appearances refer insofar as they are ‘not things in themselves, but themselves only representations’ (A108–9).Footnote ¹⁵

But how did Kant introduce the key role of the TO in the TD? After presenting the threefold synthesis, Kant asks the key question which motivated his whole Critical turn, namely ‘[w]hat does one mean, then, if one speaks of an object corresponding to and therefore also distinct from the cognition?’ (A104). His answer is that ‘this object must be thought as something in general = X, since outside our cognition we have nothing we could set over against this cognition as corresponding to it’ (A104). That is, since X is outside our representations, it is nothing for us, and the concept of this X is defined by ‘the formal unity of consciousness in the synthesis of the manifold of the representations’ (A105).Footnote ¹⁶ But importantly, X is posited as something that is not a representation, i.e. lies outside our representations. This locates it as a thing beyond the mental domain, and must therefore imply some existence claim. Although the TO has the function it has precisely because we have no cognitive access to anything beyond our representations, it fulfils this function by being posited, i.e. through some existence claim. To be sure, this claim is internal to the context of the judgement, which may be a problematic judgement. But if the notion of X being posited outside our representations is to have any meaning at all, it must involve some such (problematic) existence claim. That is, insofar as I am determining a possible object as having property ϕ, I am considering the possibility of some ThI (that is thereby posited)Footnote ¹⁷ appearing as an object with property ϕ. If I make an assertoric judgement, this possibility becomes actuality.

So, while it is the case that ‘in all of our cognitions [it] is really one and the same = X’ (A109), a judgement involves positing this TO as some thing to which appearances, qua mere representations, refer, i.e. some thing that is thereby considered as existing outside the mind.

In the PN chapter, Kant draws upon these results to conclude to the need for ‘appearance’ to refer to something ‘which in itself … must be something’ (A252). He is thus confirming that in being cognitively related to an appearance, one also thereby posits something in-itself. The interpretation I have given of the role of the TO in the A-TD to explain this passage will perhaps not convince a reader who sees the TO as a merely methodological tool. While this reader would then have to produce another account of A252, greater clarity about the key role of existence requires looking at the B version of PN.

Negative Noumena

In the B-edition, Kant replaces the role of the concept of transcendental object with the claims that:

In this sentence, Kant makes two distinctions. First, starting with an empirical object, he considers objects that one can think (‘conceived in accordance with [the understanding’s] constitution’), and adds that we cannot intuit them.

In the following paragraph, Kant sheds light on the function of this distinction that is at the heart of the notion of TO in the A-TD. He explains that ‘the understanding, when it calls an object in a relation mere phenomenon [i.e. the empirical object], simultaneously makes for itself, beyond this relation, another representation of an object in itself’ (B306). Of the latter, he says, as for the TO, that we can only form an ‘entirely undetermined concept’ of it (B307). The point of doing this is to warn against the error that characterized his residual Leibnizian viewpoint in the Dissertation, of taking this ‘for a determinate concept of a being that we could cognize through the understanding’ (B307).

The second distinction is between the thing that is cognized as an empirical object and ‘other things’ (B306). For our immediate purposes, this sheds light upon the first distinction as concerning the same thing but conceived as different objects, namely what the A-TD refers to as empirical object and TO. But in the second-edition PN chapter, Kant does not have the A-TD to back this claim. So we might ask: what allows him to assert that one can conceive of the thing given as empirical object, as an object of the understanding?

Allison (Reference Allison1983: 237–42; Reference Allison2004: 51–6) and Prauss (Reference Prauss1974) understand this claim as having the methodological function of contrasting our cognition with one that would have access to the same things as they are in themselves, and thereby underscoring the limits of our discursive cognition. But as many commentators have pointed out (e.g. Gardner Reference Gardner1999: 293; Schulting Reference Schulting2017a: 391), no justification for being able to conceive the same things independently of the conditions of sensibility is thereby provided. As Gardner (Reference Gardner1999: 293) indicates, nothing remains if we consider the number 2 minus its property of being even.

The text at B306 does however contain the inference ticket for the move from the thing as object of my sensible cognition to the same thing considered in-itself through the understanding. For there is a conditional clause preceding the above citation: ‘if we call certain objects, as appearances, beings of sense (phaenomena)’ (B306). This is a more explicit and clearer appeal to existence than in the A version, because of the use of ‘being’.Footnote ¹⁸ However, as in the A version, it does not involve the claim that something exists, but the positing of a thing, as Kant spells out at A598/B626: ‘Being … is the mere positing of the thing or of certain determinations in themselves’. This positing occurs in judgement, i.e. it is a conditional existence claim, as the conditional form ‘if we call certain objects … beings of sense’ indicates. And indeed, it suggests a reference to the context of a judgement: when do we fulfil this condition if not in some judgement?

Kant uses this positing of a thing to move from the consideration of a being of sense to that of an intelligible being (‘being … of understanding’). The text at B306 thus presents the two ways in which such an intelligible being can be conceived, and the first is as the same thing qua object of the understanding. This is the noumenon in the negative sense which takes over the role of the TO in the first-edition PN chapter. Kant thereby explains that the consideration of this noumenon is already implicit in the consideration of the being of sense, insofar as one distinguishes that in the thing which is in-itself from the way it is intuited by us. So the posited existence of the thing (as being of sense) is the inference ticket enabling this transition to be made.

The pivotal role of the notion of positing/existence thus enables us to move from the empirical object to the same thing considered as that of which it is the appearance (B306). This means that the claim that a ThI is posited is implicit in any judgement that posits objects qua appearances. If the judgement is assertoric, then we find that a single claim of existence has two aspects to it: the existence of the empirical object and that of the ThI that grounds it. This explains why, for Kant, claiming the existence of an object is claiming that something exists tout court,Footnote ¹⁹ and therefore that some thing exists in itself, as he states in the third Critique: ‘our entire distinction between the merely possible and the actual rests on this: in saying that a thing is possible we are positing only the presentation of it … but in saying that a thing is actual we are positing the thing itself (an sich selbst)’ (5: 402). Existence is thus never qualified with an epithet of ‘empirical’, unlike the determinations of the object. Claiming that an object exists is therefore claiming that there is a thing which is the object of my empirical knowledge and that exists (in itself).Footnote ²⁰ This is confirmed by Kant’s very clear statements in the B-TD: ‘appearances are only representations of things that exist without cognition of what they might be in themselves’ (B164) and in the Prolegomena: ‘I leave to things as we obtain them by the senses their actuality and only limit our sensuous intuition of these things’ (4: 292–3).

The second-edition PN text therefore gives us metaphysical grounds, rather than methodological grounds, for a dual-aspect theory of the relation between ThIs and appearances. In the next sub-section, I fill in this dual-aspect account, first by showing how it sheds light upon the grounding function of the TO; and second, by comparing it with existing dual-aspect theories.

Filling in the Dual-Aspect Picture

Gardner (Reference Gardner1999: 290) claims that the understanding of the ThI as TO or noumenon in the negative sense, i.e. as that of which appearances are appearances (see also B164) is bound up with their being grounded in ThIs. This claim can be substantiated insofar as a metaphysical dual-aspect theory is endorsed. This is the case because the appearance aspect of a thing is merely the way a ThI is manifested to our discursive cognition at a particular place and time. As such a manifestation, it follows trivially that it must be grounded in the way the thing is in-itself. This provides support for the grounding claims about the TO independently of the Transcendental Aesthetic and the Antinomy.Footnote ²¹

But we can show that it also provides support for a further specification of Gardner’s claim, namely that the grounding in question must be one of empirical causality. Indeed, this grounding role of the TO discloses the existence of ThIs: as we have seen in the previous sub-section, the positing of the empirical object is simultaneously a positing of a thing as it is in-itself, so that the claim that it exists is made when the empirical object is grasped as actual. But what justifies such a claim of actuality? The Postulates of Empirical Thought connect actuality with perception (and therefore the occurrence of a sensation):Footnote ²² there must be a causal relation (direct or not) of this object to sensation. Since this causal relation must be grounded in the same thing as it is in-itself, we thus reach the same conclusion as in section 1. Namely, the ThI is disclosed as what grounds an empirical aspect, causality, and thereby affects us (transcendentally) in that, through this empirical causality, we are affected (empirically), which means that the grounding in Gardner’s claim must be one of empirical causality. The way empirical affection thus takes on transcendental significance is confirmed by Kant’s claim in the Prolegomena: ‘if we view the objects of the senses as mere appearances, as is fitting, then we thereby admit at the very same time that a thing in itself underlies them, although we are … acquainted … only with … the way in which our senses are affected by this unknown something’ (4: 314–15).

So far, our analysis has led to upholding a metaphysical dual-aspect theory of the relation between ThIs and appearances in which the second aspect grounds the first’s causality through which it affects us. In what I have proposed, there is no requirement for us to think of ThIs as substances, except perhaps in the minimal sense of a placeholder for that which has the property of grounding empirical causality. I take this to distinguish this theory from a covert assumption of contemporary MDA theories. To take the most recent version of MDA, it is claimed that Kant starts with ‘things, things which have intrinsic natures and relational properties, and he argues that we cannot know these things’ intrinsic natures and can know only their relational properties’ (Allais Reference Allais2015: 72–3). This spells out the two aspects, and in so doing these things are implicitly thought as bearers of properties, i.e. substances (see also Allais Reference Allais2015: 34). Below, we shall find grounds for not considering the intrinsic properties of a ThI to be what constitutes the second (in-itself) aspect of things, which thus eliminates any motivation for thinking of it primarily as substance.Footnote ²³

At this point, against any dual-aspect interpretation, it will be objected that Kant also speaks of ‘other things’ under the class of ‘beings of the understanding’ in the text (B306) cited above. How could the consideration of such ThIs ‘which are not objects of our senses at all’ be justified? The answer is straightforward: once we allow for ThIs that exist, and of which we have access to their appearances in space and time, there is no further difficulty in thinking of other ThIs which might not so appear. As he clarifies further, ‘there may even be beings of the understanding to which our sensible faculty of intuition has no relation at all’ (B308–9). To invoke a two-world theory for these logically possible ThIs would be an option, but it would distinguish between ThIs that appear and those which do not, not between ThIs and appearances.

The Issue of our Cognitive Perspective

In what we have so far examined in the PN chapter, the ThI has been considered in terms of our faculty of understanding. The distinction between two aspects is that between what we know through sensible intuition and that which we can think about in abstraction from this intuition. The latter amounts to considering the ThI qua object of thought for our understanding, whose existence has a grounding function.Footnote ²⁴

Although this is an issue that is mostly overlooked in the literature, it is important to point out that it says nothing about how ThIs are qua ThIs, i.e. how these things are in themselves. Indeed, the dual-aspect position I have outlined is only meant to account for the relation of ThIs to appearances, a relation that is defined by (and indeed only makes sense from) the perspective of human cognition. This follows from the fact that the ThI qua object of thought is thereby considered in terms of the categories of our discursive cognition. It is therefore thought in relation to a set of universal concepts, the categories, and not as the particular that it is.

To describe these as relational properties does not mean that only relational categories are involved. Rather it points to the fact that these are properties for a certain type of cognition, namely a discursive one that brings unity to a manifold. The constituents of the manifold are thereby related in a unity, i.e. related to one another and related to the thinking I. In thinking the ThI, I abstract from the fact that no manifold is actually given, to consider how a manifold would be synthesized under the transcendental unity of apperception, so the ThI is thereby thought in relation to a possible apperceiving subject (and its categories). This invites us therefore to consider whether there is another cognitive perspective from which the intrinsic nature of ThIs could be thought.Footnote ²⁵

Noumena in the Negative and Positive Senses

This is precisely what Kant does next in the PN chapter. He accomplishes this in two steps which describe how the concept of a noumenon arises. As we saw above, the noumenon is first defined negatively by reference to ‘the thinking of something in general, in which I abstract from all form of sensible intuition’ (A252). In that sense, as noted earlier, it corresponds to the function of the concept of TO, i.e. that of the ThI to which appearances refer, but it sheds light upon the TO’s role in the constitution of objectivity.Footnote ²⁶ Like the concept of TO, this concept does not address ThIs qua ThIs.Footnote ²⁷

The concept of noumenon however takes on a positive form which is clearly distinct from the TO (A253) and is defined as the object of another kind of understanding which cognizes ‘its object not discursively through the categories, but intuitively in a non-sensible intuition’ (A256/B311–12 – see also B178). This is now the shift of perspective needed to make sense of reference to the ThI qua ThI: to make sense of talk about its intrinsic properties,Footnote ²⁸ we must consider the possibility of an intuitive understanding. As Kant explains in the TD, ‘the categories would have no significance at all with regard to such a cognition’ (B145). This is therefore clearly different from the conception of the ThI as the second (in-itself) aspect of that which we cognize through its first aspect, as empirical object, since, as indicated earlier, we can think this second aspect by making use of the categories.

From the perspective of an intuitive understanding, there is also no role for the object to be considered as having a sensible aspect: it is an object of an intellectual and thus non-sensible intuition which has a single aspect, namely the way it is in-itself for this intellectual intuition.

This perspective is of course entirely speculative, but nevertheless ‘unavoidable’ (A311/B256) in the Critical system: it is by contrasting this perspective upon a ThI with that of our cognition that we express the sense in which the reality of appearances is a limitation of knowledge as compared with another (logically) possible type of knowledge of things.Footnote ²⁹

This last claim may seem to clash with the (correct) view that Kant already delimits our knowledge within the theory of judgement.Footnote ³⁰ But these are two distinct ways of delimiting our cognition. In the theory of judgement (to which Kant refers in PN at A248/B305), it is the distinction between two aspects that is involved in delimiting what is possible knowledge for our understanding – namely the result of applying it to sensible intuition – from what is not, but can only be an object of thought because no sensible intuition is involved. This is a distinction in which our understanding is fixed and we consider the difference between what can be achieved cognitively with and without sensible intuition.

The distinction between two perspectives, on the other hand, involves considering another type of understanding (that is simultaneously intuitive), which is ‘problematic’ (A256/B311), but has the function of defining the kind of knowledge that could be had by such an understanding, namely knowledge of ThIs qua ThIs.

The former distinction shows the limitations of our knowledge of universal determinations of objects insofar as it requires sensible input. Without that input, we can only think the object under the categories. Both these aspects describe our discursive cognitive perspective: how things appear in space-time (first aspect) and how their nature as ThIs grounds these appearances (second aspect). The latter distinction points to the kind of knowledge that could be had by a different kind of understanding which would know the ThI in its particularity.

Intrinsic and Extrinsic Properties in a Kantian Context

These two distinctions suggest differentiating between intrinsic^K and extrinsic^K properties in a Kantian context. Unlike the transcendental realist use of the distinction between ‘intrinsic’ and ‘extrinsic’, which assumes an ‘ontological’ view from nowhere (e.g. Langton Reference Langton1998), the Kantian version of these terms would involve distinguishing two perspectives upon things. These are, respectively, the perspective of an intuitive intellect that would grasp them as they are in themselves, i.e. their intrinsic^K properties, and that of a discursive intellect that thinks and cognizes their relation to a discursive cognition, i.e. their extrinsic^K properties.Footnote ³¹ Intrinsic^K properties are therefore intrinsic properties, but calling them intrinsic^K draws attention to the fact that they exist only if such an intuitive intellectual perspective is (really) possible.Footnote ³² Extrinsic^K properties are unlike extrinsic properties in that they are perspectival, from the perspective of some non-intuitive understanding such as ours,Footnote ³³ while extrinsic properties are defined from a perspective from nowhere in that they are ontological properties.

On this proposal, the empirical properties of appearances are a type of extrinsic^K properties of ThIs. Langton (Reference Langton1998) had proposed to view them as extrinsic properties of ThIs. The inadequacy of the intrinsic/extrinsic distinction as an account of the distinction ThIs/appearances has been pointed out by Allais (Reference Allais2006) in her critique of Langton’s proposal: such a distinction leaves out the idealism.Footnote ³⁴ However, her alternative proposal, namely to consider that appearances are relational properties of ThIs, while it introduces the all-important relation to a cognizing subject that characterizes idealism, still does not capture what is essentially perspectival about our knowledge of appearances (see Schulting Reference Schulting2017b). For a relational property, even when asymmetric (e.g. cause–effect), does not feature the kind of dependence characterizing the status of the object with respect to the subject, i.e. that it is real only for a possible subject. In a relation involved in a relational property, each relatum could be considered on its own terms without the other, which is not the case for the subject–object relation.

Rather than reject Allais’s proposal of viewing them as relational properties, I would add a qualification, namely that these relations be understood as valid from a certain perspective. That is, empirical objective determinations are properties of the appearances of ThIs in space-time as seen from our cognitive perspective. As such, they can be described as secondary extrinsic^K properties of the ThI in contrast to the basic extrinsic^K properties of the ThI, which are its grounding of empirical causality and any features thereof which we may think but not know.Footnote ³⁵

So, to summarize, there are three types of properties of ThIs:

• intrinsic^K properties, i.e. properties of ThIs qua ThIs as grasped from the perspective of an intuitive intellect;

• basic extrinsic^K properties, i.e. the property(ies) ThIs have from our perspective as ground of empirical causality;

• secondary extrinsic^K properties characterizing how ThIs appear to a subject in the form of objective determinations in space-time.

Summary of the Position and the Secondary-Quality Analogy

To summarize, a complete theory of ThIs and appearances must take on board two perspectives. From the one perspective, the actual perspective of our cognition, we have an object which is a thing that has two aspects: one empirical (secondary extrinsic^K properties of the ThI from the perspective of our cognition), i.e. that of being empirically determined as an object; and the other, at least partly (as characterized above) intelligible, i.e. that of being a completely indeterminate causal ground that exists in-itself (a basic extrinsic^K property of ThIs from our cognitive perspective). What it grounds is the object’s empirical causality through which we are affected. The other perspective is the problematic perspective of an intuitive intellect. It defines a grasp of an object with a single aspect, the ThI as it is in-itself (i.e. its intrinsic^K properties). This then amounts to a two perspective dual-aspect theory.

In a seminal paper, Allais (Reference Allais2007; see also Rosefeldt Reference Rosefeldt and Stolzenberg2007) shows how the sometimes discredited analogy Kant draws between the relationship of secondary to primary qualities and that of appearances to ThIs is in fact a powerful illustration of the latter relation. In line with the issues I raised about Allais’s position, I therefore want to show how the secondary quality analogy should be understood on the interpretation I propose. I agree with Allais that the intrinsic properties of ThIs, or more accurately, the intrinsic^K properties, map onto the primary qualities, and appearances onto the secondary qualities. That is what Kant claims (Prolegomena, 4: 289) and it forms the core of this analogy.

But unlike Allais, I do not view the primary/secondary quality distinction as that between two ways of describing things, i.e. two aspects of things. Rather, the analogy maps the intrinsic^K properties of ThIs onto the perspective from which primary qualities are graspable, i.e. the perspective of the physicist. This is because these are the properties of ThIs that an intuitive understanding cognizes. Calling the physicist’s understanding a perspective rather than an aspect is not a mere linguistic issue. Indeed, a perspective does not necessarily refer to any other perspective as it is a perspective on something. On the contrary, an aspect does so refer as it is (merely) one aspect of something (see n. 24). Here, in the physicist’s perspective, there is no reference to another perspective on reality. Insofar as physics reveals the primary qualities of things, it has no place for secondary qualities: physics defines a single aspect. In the analogy, the corresponding perspective is merely problematic for Kant.

From the perspective of our cognition, there is a dual aspect: insofar as the ‘appearance’ aspect (secondary extrinsic^K properties) maps onto secondary qualities, the ‘in-itself’ aspect (basic extrinsic^K properties), on my interpretation, maps onto the ground of these secondary qualities. Unlike Allais’s account, this grounding role does not define primary qualities, but something that lies between primary and secondary qualities, namely the primary qualities’ ability to bring about secondary qualities. An account of how physical properties give rise to colours does not belong to physics, because these are only properties for our kind of perception.Footnote ³⁶ It is therefore this capacity of objects with primary qualities to give rise to secondary properties in us that the ‘in-itself’ aspect of our perspective maps onto.

Jacobi’s Worries

Jacobi famously complained that he had been unable to enter into the Critical system without assuming the existence of ThIs, but that with that presupposition he had found it difficult to remain therein (Jacobi Reference Jacobi and di Giovanni1994). Jacobi’s complaint is that, to enter into the Critical system, one needs to assume that something real (in-itself) is given as precondition for having experience, i.e. as ground of our sensations, and once the system is in place, this reality turns out to lie beyond the bounds of possible knowledge.

Jacobi thereby raises two concerns: the purported problem of using the category of causality outside the realm of possible experience and that of the status of the claim that ThIs exist. Starting with the first issue, there is overwhelming evidence that Kant allows a use of the categories outside the spatiotemporal domain of sensible intuition, while not allowing that it provides any determinate objective knowledge (e.g. Adams Reference Adams1997: 810). Indeed, although Kant claims that, when they are not applied to sensible intuition, the categories are without meaning (A239/B298), this is in the sense of determinate objective meaning. So the claim, when spelt out, is that without such intuition the categories are ‘representations that we cannot relate to any determinate object’ (A258/B314). But, as Kant says in the same chapter, ‘[t]he pure categories, without formal conditions of sensibility have merely transcendental significance, but are not of any transcendental use’ (A248/B305). This implies that ‘the categories extend further than sensible intuition since they think objects in general’ (B309/A254). This use of the categories is, moreover, essential to forming the Ideas of Pure Reason, so if rejected, central claims of the Transcendental Dialectic would have to go. This is all that is needed to confirm the possibility of a use of the category of causality outside the domain of possible experience.

Turning to Jacobi’s second worry, some commentators view the claim of existence of ThIs as lying beyond the bounds of possible knowledge (Senderowicz Reference Senderowicz2005, Bird Reference Bird2006: 553–80). However, the knowledge Kant analyses in the first Critique involves determination of empirical objects, of laws governing their behaviour (physics), or laws governing pure spatial (geometry) or also possibly non-spatial (arithmetic) magnitudes. The bounds that Kant identifies are thus bounds of possible determinate knowledge. This contrasts with the purported knowledge of ThIs that is indeterminate.Footnote ³⁷ As such, it is not excluded (see Ameriks Reference Ameriks2000); indeed Kant clearly states: ‘without the condition of sensible intuition, the categories have no relation to any determinate object’ (A246; see also A258/B314). This does not exclude indeterminate existence claims.Footnote ³⁸

One might object that the claim that the ThI must be thought as causal ground defines a determination of ThIs. But this claim only concerns the extrinsic^K properties of ThIs, and says nothing about what ThIs are qua ThIs.Footnote ³⁹ It is made from the perspective of our cognition and tells us how ThIs feature as ‘material’ conditions of knowledge.

Other Objections

Competing theories about the relation of ThIs to appearances have already been criticized earlier in the article and in notes. To summarize the main points:

1. (i) TW assumes a view from nowhere from which ThIs and appearances would define two objects which could then either be distinct or not; but such an ‘ontological’ point of view is not available.

2. (ii) EDA does not provide a warrant for considering the in-itself aspect of that which is determined as empirical object in our cognition.

3. (iii) Existing MDAs implicitly think of ThIs primarily as substances and do not provide a satisfactory account of how the ideality of the appearance aspect of things is to be understood.

While this article cannot extensively discuss the replies that could be made to these objections, it will briefly consider two further objections, one from a TW theorist and one from an EDA theorist.

Stang (Reference Stang2014) argues against what he calls identity theories, by which he refers to dual-aspect theories, pointing out that if objects are identical, they must share all properties. But first, the very characterization of them as ‘identity’ theories already involves an appeal to a view from nowhere that does not exist. And second, in my dual-aspect theory, there is an asymmetry between the intelligible and the empirical aspects of things that would not arise were identity at stake. Rather, one set of properties (those of how things appear) is grounded in another (in-itself causality), which avoids the purported worry.

The EDA proponent is likely to dispute my assigning a pivotal role to the positing of a thing, in warranting the consideration of things as they are in themselves: he will deny that any kind of existence claim (even conditional) is implied by positing the TO (A-edition PN) or an intelligible being (B-edition PN). Rather than dispute this on both systematic and textual grounds, I shall point to a difficulty that he encounters by denying that this appeal to existence provides the warrant for considering ThIs.

The problem that arises for the EDA theorist is that of how to understand the judgement that an object exists. As the second Postulate of Empirical Thought characterizes it (A218/B266), actuality obtains when there is a (causal) connection with perception. So something is actual whenever it is (directly or not) causally responsible for the subject’s reception of sensations. For this EDA theorist, establishing the existence of an empirical object therefore amounts to determining the object as the cause (direct or not) of certain sensations I experience.

But this contradicts Kant’s claim that judging that some object exists ‘does not augment the concept’ of this object (B266/A219; see also Schulting Reference Schulting2017a) or that ‘[b]eing is obviously not … a concept of something that could add to the concept of a thing’ (A598/B626). Indeed, the EDA theorist’s claim that something exists does involve adding to its concept, namely the determination of its being causally linked to certain sensations.Footnote ⁴⁰ This problem does not, however, arise for the two-perspective dual-aspect account: judging that an object exists amounts to determining (through the aforementioned causal link) a time and place for an existence that is already posited in the mere problematic consideration of this empirical object. This temporal determination does augment the concept of the object, but this is not through adding ‘existence’ to the object: ‘existence’ was already there and it is now assigned a spatiotemporal location.