Transcendental illusion

According to Kant, empirical illusion (empirischer Schein) occurs when the empirical use of the understanding is misled by the influence of the imagination (1998: 406–7; CPR A295/B351–2). Kant gives the example of the moon appearing to be larger when rising (1998: 408; CPR A297/B354). In such a case, the imagination associates appearances in a way that is only subjectively necessary, though it can easily be taken to be objectively necessary. Although judgements based on illusion are mistaken, the illusion itself is not necessarily deceptive. Error only enters when one makes a certain judgement about appearances, but appearances themselves cannot be mistaken (Kant Reference Kant1998: 405–6; CPR A293–4/B349–50). That is, if one judges that what seems to be true really is true, then one has been deceived by the illusion. However, even if one does not judge falsely, the illusion persists. Although the astronomer knows that the moon's true size is not affected by whether it is rising or setting, it still appears to him to be larger when rising (Kant Reference Kant1998: 408; CPR A297/B354).Footnote 4

Kant claims that logical illusion (logischer Schein) consists in the mere imitation of the form of reason and results from a lack of attention to logical rules. As soon as one's attention is sharpened, then the logical illusion disappears (Kant Reference Kant1998: 408; CPR A296–7/B353). For example, it might seem to one who has been skipping one's logic class that, given ‘A→B’ and ‘B’, one may conclude ‘A’. This illusion will dissipate once one pays closer attention. One might define logical illusion as a subjectively necessary connection of concepts in a judgement. Unlike empirical illusion, logical illusion cannot persist once it is known to be an illusion. That is because logic is nothing but the form of thinking for Kant, so there can be no logical seeming apart from thinking of some kind (Kant Reference Kant1998: 26; CPR Bxxiii).

Like empirical illusions but in contrast to logical illusions, transcendental illusions are not dissipated once one realizes that the illusion is an illusion through a critique (Kant Reference Kant1998: 408; CPR A296–7/B353). Kant defines transcendental illusion (transzendentaler Schein) as follows:

Thus, any maxim of the use of reason which is taken to be objectively necessary when it is only subjectively necessary is a transcendental illusion. The subjective necessity arises from the human epistemic faculties (Erkenntnisvermögen), just as in the case of empirical illusion (Kant Reference Kant1998: 408; CPR A297/B353–4). According to Kant, reason is merely the subjective law of the management of our understanding to bring about the most general use of the understanding by subsuming all concepts under the smallest possible number of general concepts. Reason thus has only indirect relation to objects via the understanding (Kant Reference Kant1998: 415; CPR A306/B362–3). Its conclusions are those of deduction from what the understanding gives it, not of immediate knowledge such as the understanding has (Kant Reference Kant1998: 412–14; CPR A303–5/B359–61).

The Idea of the Systematic Unity of Nature as a Transcendental Illusion

Although Kant does not use the term transzendentaler Schein in the Appendix, he does refer to the Illusion that an idea refers to an object (1998: 710; CPR A644/B672).Footnote 5 Kant says that ideas (Ideen) cannot be constitutive and must always be regulative (1998: 709–10; CPR A644/B672). Their regulative use, however, is indispensably necessary for reason to orientate the understanding towards a goal outside the boundaries of possible experience. This goal is the greatest unity of the concepts of the understanding (Kant Reference Kant1998: 709–10; CPR A643–4/B671–2).

Kant makes an analogy between this goal and a mirror image understood as an imaginary focal point (focus imaginarius). When one looks into a mirror, one is seeing a virtual image which results from the reflection of light from the mirror. The virtual image appears to be a certain distance behind the mirror (Kant Reference Kant1998: 710; CPR A644–5/B672–3). The virtual object appears to be where the light rays reflected from the mirror would converge if they were not reflected from the mirror. This optical illusion itself is not a mistake, since only a judgement about the illusion can be mistaken. For example, one does not usually think when one sees oneself in a mirror that one is seeing another person.Footnote 6 One merely uses the virtual image for the purpose of, say, combing one's hair.

Similarly, the systematic unity of nature is only an idea, a projection from the known to a point beyond the bounds of possible knowledge. As Kant says, this serves as a goal for the understanding:

It appears to be the convergence of all concepts in a single idea beyond all possible experience.

The unity posited by the idea is an ordered combination or system, as opposed to an aggregate of concepts (Kant Reference Kant1998: 711; CPR A645/B673). One does not get this idea from nature, but one demands that nature conform to it. This means that if our experience does not conform to it, we consider it to be defective and search for unity (Kant Reference Kant1998: 711; CPR A645–6/B673–4). Thus, while the unity is actually subjectively necessary, the thought of it may be said to appear to be objectively necessary (Kant Reference Kant1998: 408; CPR A298/B354). That is what makes it an illusion.

Textual Support

Though it would be tedious to discuss how the doctrine of transcendental illusion can account for every apparent contradiction in the Appendix, I will consider several illustrative passages. In many of these cases Kant himself does not state that the idea of the systematic unity of nature or its derivative principles have objective reality, but that they pretend to or we assume that they do. For example, Kant says of the idea of a fundamental force underlying all various appearances that it ‘pretends to have objective reality’ (1998: 714; CPR A650/B678; emphasis added). This is because even if all attempts to find unity in diversity fail, ‘we assume’ we will find it. ‘Reason assumes systematic unity of manifold forces because particular natural laws stand under more general ones, and the economy of principles becomes not just an economical principle of reason, but an inner law of nature’ (Kant Reference Kant1998: 715; CPR A650/B678; emphasis added). This passage is cited by commentators as arguing that the idea is objectively valid.Footnote 7 But as the italicized words show, Kant makes no such claim. Rather, we assume that this principle is not one merely of economy, but that it is something more – a law of nature.

Kant says that, though one might believe that maxims such as Ockham's Razor are merely for saving effort, ‘such a selfish purpose is very easy to distinguish from the idea that everyone presupposes: this unity of reason fits nature itself, that reason does not here beg but command, even though it cannot determine the boundaries of this unity’ (1998: 717; CPR A653/B681). Commentators are divided on whether this passage is saying that the unity of reason is subjective or objective.Footnote 8 But that everyone presupposes that reason does not here beg but command does not mean that nature really does not here beg but command. Kant only says here that the assumption that the idea is objectively valid is easy to distinguish from the idea being merely subjectively valid.

Another case concerns the principles of genera, species, and continuity, which are entailed by the idea of the systematic unity of nature. According to the principle of genera, homogeneity must be found in the manifold of appearances in order for empirical concepts to be possible (Kant Reference Kant1998: 717; CPR A653–4/B681–2). According to the principle of species, heterogeneity must be found in the manifold of appearances in order for empirical concepts to be possible (Kant Reference Kant1998: 719; CPR A656/B684). The conjunction of these two principles entails that once all appearances can be subsumed under general concepts, all appearances fit into a system of hierarchically organized concepts (Kant Reference Kant1998: 720–1; CPR A657–8/B685–6).

Kant says of these principles:

That is, Kant says that the principles of genera, species, and continuity judge nature to be unified under these ideas. This is perfectly compatible with this unity being a transcendental illusion. The remark that these principles carry their recommendation with themselves and not as Handgriffe of method is only a result of the judgement that they are appropriate to nature.

One passage says that the same principles of genera, species, and continuity ‘may appear (scheinen) to be objective’ (Kant Reference Kant1998: 727; CPR A666/B694).Footnote 9 As another passage says, these principles:

First of all, Kant repeats that these principles ‘appear’ transcendental, ‘even though’ they are only used asymptotically. He then says that ‘as synthetic a priori propositions’ they have ‘objective, though indeterminate validity’. The principles’ a priori status is not that of a priori knowledge, but of mere independence of experience. That is, they are assumed a priori (Kant Reference Kant1998: 715; CPR A650–1/B678–9). What is assumed independently of experience is not necessarily known independently of experience. The use of ‘actually’ shows that Kant goes from talking of the illusion of objective validity to talking of the actual subjective use of the principle as heuristic.Footnote 10

Once Kant's doctrine of transcendental illusion is understood, the Appendix no longer appears to contradict itself. Close examination reveals that every passage in which Kant appears to claim objective validity or ascribe transcendental significance to the idea of the systematic unity of nature refers to an assumption or presupposition. The assumption or presupposition in question is that the idea of the systematic unity of nature is transcendental or objectively valid. As Kant declares at the beginning of the Appendix, this is a transcendental illusion (1998: 709–10; CPR A643–5/B671–3). However, I have not yet considered the two most problematic passages (Kant Reference Kant1998: 715, 717; CPR A651/B679, A653–4/B681–2).Footnote 11 Since they are pivotal to the claims that the idea of the systematic unity of nature is necessary for the formation of empirical concepts, I will discuss them in the next section in connection with those claims.

A Criterion of Empirical Truth

Henry Allison says with regard to the idea of the systematic unity of nature and its derivative principles that ‘at times Kant … insists on their indispensability for the operation of the understanding’ (2004: 424). Allison refers to Kant's argument at A651/B679 (Kant Reference Kant1998: 715). There Kant says that the only way that the principle of the logical unity of rules can exist is if the systematic unity of the objects themselves is a priori assumed to be necessary. Kant gives the example of positing a single fundamental force that underlies all apparent forces. If all forces were completely heterogeneous, then the assumption that nature was unified would set a goal at variance with reality. Kant continues:

Paul Guyer says this passage suggests ‘without explanation that reason and its idea of systematicity are somehow directly involved in the generation of empirical concepts and cognition’.Footnote 12 Guyer also says concerning this passage that ‘Kant's statement is resounding. Unfortunately, he does not explain how or why systematicity is required in order to have an empirical criterion of truth [sic]’ (Reference Guyer1990: 28). Aside from Guyer's obvious slip in calling the criterion of empirical truth an empirical criterion of truth, he raises an important point.

To say that reason demands that one seek unity in nature is understandable enough, given Kant's definition of reason. And to say that there is no coherent use of the understanding without reason is also merely a matter of stipulation. Up to this point, the argument is clear: without reason, concepts that the understanding discovers cannot be united into higher order concepts. But what does it mean to say that there is no adequate criterion of empirical truth without reason? Kant does not elaborate.

It is difficult to imagine how the law of reason to seek the unity of concepts could serve as a criterion of empirical truth in one sense. Concepts that admit of unification are not for that reason applicable to a sensory manifold. As Reference AllisonAllison rightly points out:

The ‘truth’ Kant is concerned with here seems to be not at the level of the understanding forming concepts, but at the level of reason ordering concepts. That is, given that reason requires the ordering of concepts, there must be a criterion for doing so. The idea of the systematic unity of nature could be the criterion for deciding among rival systematizations.Footnote 13 The passage means only that the idea of the systematic unity of nature is the goal or end point to which reason strives. This does not mean that it is indispensable for the operation of the understanding, but only indispensable for the unity of the concepts of the understanding.

In a similar passage, Kant says that ‘the systematic unity of empirical cognition of the understanding … is the touchstone (Probierstein) of the truth of rules’ (1998: 712; CPR A647/B675). Again, the sense of the ‘truth’ of ‘rules’ must be specified. Immediately after saying that the idea of the systematic unity of nature is the touchstone of the truth of rules, Kant says that the unity is ‘only a projected unity’ that ‘must be seen as only as a problem’ (1998: 712; CPR A647/B675). That is, if the touchstone of the truth of rules is only projected and assumed to exist as a problem, then there is no actual touchstone of the truth of rules. ‘This systematic unity or unity of reason is a logical principle that helps the understanding where it cannot attain rules to bring its rules to unity under a single principle and thereby achieve coherence insofar as it is possible’ (Kant Reference Kant1998: 713; CPR A648/B676; emphasis added). That is, in the system that reason outlines for the understanding, there will always be gaps. Reason helps the understanding see that there are such gaps and where they are. Reason provides a goal to guide the understanding to find affinities among empirical concepts to form more general concepts. For example, Kant says that Ockham's Razor has led chemists to class all salts into two main genera (acidic and alkaline) and to seek to unite those two genera under one genus (Kant Reference Kant1998: 716; CPR A652–3/B680–1).

The Logical Principle of Genera

Earlier I mentioned that, according to Kant's definition of the principle of genera, if there were no homogeneity in appearances, empirical concepts would not be possible. Kant says more about this. I will quote this passage at length, since it is relied upon heavily by the commentators I will discuss.

This passage seems to say two things that are problematic in themselves and that are at variance with the transcendental analytic. First, it seems to say that, without the presupposition of necessary homogeneity, no empirical concepts would be possible. Second, it seems to say that without empirical concepts, experience would be impossible. If both these statements are true, then without the presupposition of necessary homogeneity, experience would not be possible. This would make the presupposition of necessary homogeneity necessary for experience.

However, the Transcendental Analytic and the Jäsche Logik provide an account of the formation of empirical concepts which makes no mention of a presupposition of necessary homogeneity. Further, the Transcendental Analytic and the Jäsche Logik provide an account of how experience is possible without empirical concepts. Besides conflicting with Kant's positions enunciated elsewhere, this argument is problematic in its own right. There is no reason why empirical concepts require a presupposition of a necessary homogeneity or why experience is impossible without empirical concepts. I will provide support for these claims in my discussion of Allison's and Ido Geiger's readings of this passage.

According to Allison, the presuppositions of necessary homogeneity and heterogeneity of the manifold of possible experience are necessary for ‘the operation of the understanding’. As he says,

What Allison says here is right, as far as it goes. But this does not entitle him to say that the principles of genera, species, and continuity are ‘indispensable for the operation of the understanding, rather than [being] mere desiderata of reason’ (Allison Reference Allison2004: 434).

The passage concerns the possibility that the manifold of possible experience is completely heterogeneous. Of course, if this were true, then neither empirical concepts nor experience would be possible.Footnote 14 But all Kant has supposed in this counterfactual statement is there being no similarity at all among appearances. Thus from the fact of experience only ‘a necessary homogeneity in the manifold of possible experience’ follows (Kant Reference Kant1998: 717; CPR A654/B682; emphasis added). The law of genera as Kant states it requires not some minimum of homogeneity but the maximum. The logical law of genera demands that all concepts must have something in common (Kant Reference Kant1998: 716; CPR A651–2/B679–80). Since experience is possible even if the logical law of genera does not obtain, the fact of experience does not permit a transcendental deduction of the logical law of genera.Footnote 15 Thus understood, the passage in question is no support for holding that the principle of genera ‘is an application condition of the understanding in the sense that it is indispensable to the latter's endeavor to form empirical concepts’, as Allison does (2004: 436).

Geiger's argument is much more involved. He takes the idea of the systematic unity of nature to be necessary for experience because it ‘is a necessary condition of the very meaningfulness of empirical concepts’, which he takes to be necessary for experience (Geiger Reference Geiger2003: 285). First I will explain why Geiger thinks the idea of the systematic unity of nature is necessary for the meaning of empirical concepts. Then I will explain why Geiger thinks that, without empirical concepts, experience would be impossible.Footnote 16

Geiger contrasts what he takes to be Kant's theory of the meaning of empirical concepts with that of Hume. He describes Hume's theory as follows: ‘Empirical concepts owe their content or meaning, ultimately, to individual impingements of reality on our senses’ (Geiger Reference Geiger2003: 287). According to Geiger, attributing such a view of empirical concept formation to Kant is wrong because it entails that perception of nonconceptual content is possible. As he says, ‘this commitment surely flies in the face of one of Kant's most important and best known dicta: “intuitions without concepts are blind” ’ (Geiger Reference Geiger2003: 288). According to Geiger, the meaning of empirical concepts is not provided by intuition, but by the relation of empirical concepts to one another. That is,

There are two problems with this view. First, if empirical concepts do not get their meaning from intuition but from their relation to one another, we need more than the idea of a systematic unity of nature in order to have meaningful empirical concepts. Rather, we would need to actually have a system of empirical concepts if any empirical concept were to have any meaning whatsoever. The idea of a system is not the same thing as a system. Second, on Kant's view empirical concepts do get their content or meaning from intuition. Kant says when defining empirical concepts that they contain sensation (Kant Reference Kant1998: 129; CPR, B74). He also says that the comparison of representations precedes the concept of the thing (Kant Reference Kant1998: 385; CPR A269/B325).Footnote 17 Additionally, Kant says in his Jäsche Logik that ‘the empirical concept springs from the senses through comparison of the objects of experience and receives only the form of generality from the understanding. The reality of these concepts rests on actual experience, from which they are created with regard to their content’ (1923: 92; JL 9: 92).Footnote 18 These statements make clear that empirical concepts get their content or meaning from intuition.Footnote 19

According to Geiger, experience is impossible without empirical concepts (2003: 284, 288–90).Footnote 20 His argument for this is as follows. The rule for applying a given empirical concept cannot be sensible, nor can it be derived from the categories. It can only be a matter of relation to other empirical concepts (Geiger Reference Geiger2003: 288–9). As Geiger puts it, yellowness cannot be experienced without the empirical concept ‘yellow’ (2003: 288). This is what Geiger takes Kant to mean when Kant says ‘intuitions without concepts are blind’ (Kant Reference Kant1998: 130; CPR B75). As Geiger puts it, ‘intuitions without empirical concepts are still blind’ (2003: 290). When considering the claim that only the categories make it possible for empirical comparison and generalization to take place, he declares that this is ‘contrary to Kant's explicit claim’ in A653–4/B681–2 (Kant Reference Kant1998: 717), which he takes to be that, without empirical concepts, experience would be impossible.

This claim rides entirely on the ‘therefore’ (mithin) in the clause ‘without [a necessary homogeneity in the manifold of possible experience] no empirical concepts would be possible, therefore no experience would be possible’ (Kant Reference Kant1998: 717; CPR A654/B682). While this may be easily read as entailing that experience is impossible without any empirical concepts, Kant nowhere says why experience is impossible without empirical concepts. Furthermore, this ‘therefore’ can be understood differently. One is not compelled to read it as saying that experience is impossible without empirical concepts. Rather, one can read it as saying that, if empirical concepts are impossible, then so is experience. That is, if there were no homogeneity in the manifold of possible experience, neither empirical concepts nor experience would be possible. This passage is therefore weak evidence for the view that Kant believed experience to be impossible without empirical concepts.

That empirical concepts get their meaning or content from intuition and that empirical concepts are not required for experience both entail that perception of nonconceptual content is possible. While I cannot give a thorough account of the place of nonconceptual content in Kant's Critical philosophy, I will refer to a passage from the Jäsche Logik:

One might argue that the ‘savage’ has the empirical concepts such as ‘wood’ or ‘stone’ that allow him to make some sense of what he sees, just not as well as someone who has the empirical concept ‘house’. However, Kant states that the act of knowing for the savage consists of intuition only and no concepts at all.Footnote 21

Since intuition without concepts is not meaningless, Geiger's claim that the rules for empirical concept application must be conceptual must be rejected. Upon close consideration, A653–4/B681–2 (Kant Reference Kant1998: 717) does not say that the logical law of genera is necessary for the formation of empirical concepts or that experience is impossible without empirical concepts.

Necessary

What does it mean to say that the idea of the systematic unity of nature is ‘necessary’?Footnote 22 I have already argued that it is not necessary for experience. This would make the idea objectively necessary or necessary for the knowledge of any object (Kant Reference Kant1998: 172; CPR A93/B125–6). However, I have not yet considered whether the idea of the systematic unity of nature is necessary for knowledge of some particular object. One might think that without the idea of a systematic unity of nature some particular empirical concept would never be discovered by anyone. That is, the idea of the systematic unity of nature can make some empirical concept possible for someone by pointing the understanding in a direction in which it would not have gone otherwise. Kant rules out this possibility when he says that the understanding might well stumble upon the empirical concept in question without reason telling it what to look for, however unlikely that might be.Footnote 23

The idea of the systematic unity of nature can only be subjectively necessary for the knowledge of an object. According to Kant, a rule that connects representations is merely subjectively necessary when the rule is due not to the object, but only to the subject (Kant Reference Kant1998: 205; CPR B168).Footnote 24 That is, it could be the case for a particular person in a particular circumstance that only with the aid of the idea of the systematic unity of nature could that person discover a given empirical concept. There is no reason why anyone in any circumstance could not have discovered a particular empirical concept without using the idea of the systematic unity of nature as a guide.

Transcendental

Kant says that the idea of the systematic unity of nature is assumed to be or is presupposed as ‘transcendental’,Footnote 25 and I have argued above that he means only that it is a transcendental illusion. What sense of ‘transcendental’ is at work here? One must distinguish the logical from the transcendental use of both the idea of the systematic unity of nature and the principles that it entails. The logical use brings about the advancement of empirical knowledge by showing to what extent our existing empirical concepts do and do not admit of systematization. Kant describes the logical use as follows: ordering empirical concepts insofar as they permit of ordering (Kant Reference Kant1998: 709, 724–5; CPR A643/B671; A663/B691), making logical prescriptions in order to approach the completeness of conditions and attain the greatest possible unity of our knowledge (Kant Reference Kant1998: 417; CPR A308–9/B365), and bringing about the smallest possible number of the most general empirical concepts by comparing them with one another (Kant Reference Kant1998: 415; CPR A306/B362–3).

Kant says the logical principle (that all concepts can be unified into higher concepts and divided into lower concepts to form a system with no gaps) presupposes the transcendental principle (that nature itself admits of systematization) (Kant Reference Kant1998: 715; CPR A650–1/B678–9). That is, the logical principle would not always advance empirical knowledge if the transcendental principle were false. In such a case, the use of the logical principle would sooner or later lead the understanding astray in attempting to systematize something which did not admit of systematization (Kant Reference Kant1998: 722–3; CPR A660/B688). If that were true, one could only use the logical principle as a heuristic guide to scientific investigation. That nature is systematizable can never be known, so it can never be more than an assumption that nature is systematizable (Kant Reference Kant1998: 719; CPR A656/B684). When Kant talks about assuming that nature is systematizable, he means nothing more than assuming that this principle is transcendental. Of course, assuming that something is the case does not mean that it is the case.

Does this use of ‘transcendental’ differ from that used in the Transcendental Analytic? Kant defines ‘transcendental’ as what grounds the a priori possibility of knowledge of objects (Kant Reference Kant1998: 83, 191; CPR B25, B150). Since Kant also claims that transcendental ideas are regulative and not constitutive (Kant Reference Kant1998: 709; CPR A644/B672), the obvious question is how what does not constitute objects can nevertheless make them a priori possible. In the absence of overdetermination, what makes something possible is a necessary condition of that thing. So, if the idea of the systematic unity of nature makes knowledge of objects possible, in the absence of anything else that could do the same job, that would be a necessary condition of knowledge of objects. But the assumption that nature is systematizable is not necessary for the knowledge of objects (or, equivalently, for experience). While some have argued that the idea of the systematic unity of nature is required for experience because it is required for empirical concept formation and empirical concepts are required for experience, I have rejected that argument.

Kant seems to be using the term ‘transcendental’ in a different way when describing in what sense a transcendental deduction can be given of the psychological, cosmological, and theological ideas. He says that while they are not constitutive but only regulative, if they lead up to (auf) systematic unity and always extend empirical knowledge, it is a ‘necessary maxim of reason to proceed according to the same ideas’ (Kant Reference Kant1998: 731; CPR A671/B699). ‘Transcendental’ in this watered-down sense applies to any principle whose use always extends empirical knowledge. One might question whether the use of the idea of the systematic unity of nature will always extend empirical knowledge. In fact, Kant admits that it does not always extend empirical knowledge when he says if the unity one seeks is not found in a particular case, one nevertheless assumes that it eventually will be found (Kant Reference Kant1998: 714–15; CPR A650/B678). So the idea of the systematic unity of nature is not even transcendental in this watered-down sense. The idea is only assumed to be or presupposed as transcendental in this watered-down sense. One presupposes that the idea is transcendental in this sense in order to justify use the logical principles of genera, species, and continuity. That these principles seem to always extend empirical knowledge (to be transcendental) may thus be termed a transcendental illusion, but one need not be deceived by this illusion in order to see the usefulness of these principles.

Objectively But Indeterminately Valid

Given that the principles of genera, species, and continuity are not transcendental, it is no surprise that Kant says that no transcendental deduction of them can be given (Kant Reference Kant1998: 725; CPR A663–4/B691–2). Kant refers to the same reason he gives for why there can be no ‘objective deduction’ of the psychological, cosmological, and theological ideas: ideas have no relation to an object that can correspond to them (Kant Reference Kant1998: 439; CPR A336/B393). So it is surprising that he says that the principles of genera, species, and continuity nevertheless have objective but indeterminate validity (Kant Reference Kant1998: 725; CPR A663/B691).

Kant does not immediately say what he means by this. But in the second section of the Appendix, Kant says that a transcendental deduction of the psychological, cosmological, and theological ideas is possible. Just as with the principles of genera, species, and continuity, he says it cannot be of the kind he gave of the objective validity of the categories (Kant Reference Kant1998: 730; CPR A669–670/B697–698). His reasons are as follows. He says that ideas of reason act as schemata not to determine objects, but to indirectly allow their representation to us through the systematic unity of the idea in question (Kant Reference Kant1998: 730; CPR A669–70/B697–8). This indirectness is the cause of the indeterminateness of the objective validity of the transcendental ideas. Kant says the acts of the understanding can only be made determinate by schemata. Since there can be no schema for the systematic unity of all empirical concepts, that unity is indeterminate. Nevertheless, ‘the idea of the maximum of division and unity of knowledge of the understanding (Verstandeserkenntnis) in one principle’ is an analogue of a schema for bringing about the unity of concepts. The idea of a maximum is itself determinate, though it cannot be determinately applied to any object, since it only applies to an object indirectly (Kant Reference Kant1998: 726–7; CPR A665/B693). The indeterminateness of the use of this idea means, among other things, that it cannot always extend knowledge. Since the goal of the systematic unity of nature is indeterminate, one cannot be certain in a given case that one has found what one is looking for.

So now we know why Kant says the principles entailed by the idea of the systematic unity of nature are objectively but indeterminately valid. ‘Objectively valid’, like ‘transcendental’, is used here in the watered-down sense of being a condition for the knowledge of some but not all objects. But even in that case, there is, just like the watered-down sense of ‘transcendental’, no reason to believe that the idea of the systematic unity of nature is ‘objectively valid’ in the sense of being a condition of the knowledge of any particular object. Because the use of the idea of the systematic unity of nature to guide scientific research more often than not extends knowledge, the idea may be assumed to always extend knowledge. That is why Kant speaks of the idea being assumed to be or presupposed as objectively valid.Footnote 26

Summary and Conclusion

Allow me to summarize the points of this section. The idea of the systematic unity of nature is:

1. 1. Not objectively necessary but only subjectively necessary;

2. 2. Not transcendental, but is only assumed to be or presupposed as transcendental; and

3. 3. Not objectively valid, but is assumed to be objectively valid (or, if you will, objectively but indeterminately valid).

Strictly speaking, the idea is not transcendental, objectively necessary, or objectively valid in any sense. However, according to Kant, the benefits of assuming that the idea of the systematic unity of nature is all three of these things in a certain sense make the assumption unavoidable as a heuristic or maxim in scientific practice. This is what makes it a transcendental illusion. As long as one keeps the difference firmly in mind between assuming that the idea is transcendental and the idea really being transcendental, one is not deceived by the illusion. In fact, the illusion can be used to great effect, like the illusion that what I see in the mirror is myself.

Thus I hope to have demonstrated the following points. The Appendix to the Transcendental Dialectic of the first Critique does not contradict itself with regard to whether the idea of the systematic unity of nature is necessary, transcendental, or objectively valid. According to Kant, the idea of the systematic unity of nature is a transcendental illusion which may nevertheless be useful, provided one keeps certain distinctions in mind. Furthermore, the idea of the systematic unity of nature is not required for the formation of empirical concepts or for experience.Footnote 27