Kant on the Status of Ideas and Principles of Reason

1 Introduction

The Transcendental Dialectic of the Critique of Pure Reason has for a long time been regarded as implementing a mainly destructive project. After all, it establishes that reason, understood as the faculty of inference, is not able to attain theoretical cognitions of its own. This is fatal, or at least seriously damaging, for the prospects of rational psychology, rational cosmology and rational theology. Yet, the Transcendental Dialectic also identifies a positive role of reason for cognition, which Kant explores especially in the Appendix to the Transcendental Dialectic. There, he identifies a “regulative” use of reason in which some of reason’s ideas and principles are put to the service of our empirical research into nature. Over the past few decades, the interest in this positive role of reason has dramatically increased, where this has contributed to highlighting the constructive side of the Transcendental Dialectic.

Over the past few years, although our understanding of the regulative use of reason has significantly improved, agreement regarding basic issues is still lacking. One example of this difficulty is the discussion regarding the attitude one takes when one assumes something “regulatively.” For example, does this involve “belief” (in a broad sense) that a certain proposition is true, or does it imply a weaker commitment? Another difficulty, the one I will focus on in this article, concerns the status of regulative ideas and principles. More precisely, it is not clear whether regulative ideas and principles are merely “fictions” – where we assume propositions that we know are false because of their heuristic utility – or whether they imply the assumption of true, or at least possibly true, propositions.^[1] Here, one might wonder what it means that ideas imply propositions, since, according to Kant, they are concepts. In a first sense, they imply propositions because through them we represent an object as having certain properties. Accordingly, our idea of the soul implies the proposition “the soul is substance.” In a second sense, ideas give rise to the proposition “the object x exists” when they are taken as possibly referring to objects.

Interpreters are still holding very different views on this issue.^[2] To a large extent, this divergence of views is due to Kant himself, who describes the status of regulative principles and ideas in apparently inconsistent ways. For example, he writes that they have merely a “heuristic” role, but he also submits that they have “objective” but “indeterminate” validity. In this article, I suggest that progress in determining the status of regulative ideas and principles can be made if we give up the presupposition that only one answer regarding their status is correct. Rather, I will suggest that it is not the case that all regulative principles and ideas have the same status. In my view, while assuming some regulative ideas and principles do involve assuming propositions that we know are false, this is not true for all regulative ideas and principles. At least in some cases, assuming regulative ideas and principles involves assuming possibly true propositions.

In Section 2, I sketch three competing readings of the status of ideas and principles and present the passages on which they rely. In Section 3, I discuss a point recently made by Marcus Willaschek, who argues that it is wrong to conflate the regulative/constitutive and the logical/transcendental distinctions. This approach makes one reading presented in Section 3 less plausible, namely, the reading that claims that assuming regulative ideas and principles implies assuming propositions that we know are true. Section 4 also starts from Willaschek’s analysis of regulative ideas and principles, since Willaschek’s account seems to oscillate between two readings of their status: one according to which regulative ideas and principles consist of possibly true propositions and one according to which regulative ideas and principles consist of propositions that we know are false. Finally, in Section 6, I argue that this oscillation is due to Kant himself. This does not mean that Kant is inconsistent. Rather, it means that it is wrong to presuppose that all regulative ideas and principles have the same status.

2 Three possible readings and what supports them

Within the Appendix to the Transcendental Dialectic, Kant claims that reason can be used regulatively in connection to a multiplicity of ideas and principles. In its first part, “On the regulative use of the ideas of pure reason,” he focuses on the idea of the systematicity of nature and the related principles of homogeneity, continuity and specification. In its second part, “On the final aim of the natural dialectic of human reason,” he instead discusses the ideas of the soul, the world and God. He argues that ideas of reason can only be used regulatively in other sections of the Dialectic, too. This happens in the Antinomy of Pure Reason (KrV, A508-15/B536-43) and the Ideal of Pure Reason (KrV, A616-20/B644-8).^[3] It is natural to expect that in all these cases ideas and principles of reason will have similar characteristics, such that they cannot be assumed “constitutively” and can instead be assumed “regulatively.” However, obtaining this unity of interpretation is not easy, since Kant often makes conflicting claims. Unsurprisingly, interpreters who have attempted to provide a unitary reading have found plenty of passages to support their views, but also passages that speak against them. I will now briefly survey what I take to be the three main readings of the status of ideas and principles that can be assumed regulatively.

a. Ideas and principles that can be assumed “regulatively” imply propositions that we know are false. One first way to read the status of ideas and principles that can be assumed regulatively is to say that they imply propositions that we know are false, but are nonetheless instrumental for improving our empirical knowledge. This reading of ideas is defended, among others, by Michelle Grier. She distinguishes two principles of reason: P₁ is a logical principle that prescribes unity to “conditioned” cognitions by pushing us to seek for the “unconditioned” form from which they can be derived.^[4] P₂ is an illusory principle, in which we take P₁ as valid and applying to objects themselves.^[5] According to Grier, when we assume ideas regulatively, we do not only have to do with P₁. In a certain sense, we also assume P₂. This does not mean that we take to be true that there is an “unconditioned” that we can know. Rather, we assume the idea of an “unconditioned” objectively given as a guide for our inquiry, even though we recognize that assumption is false and illusory. Grier writes: “The regulative status of these principles is thus intimately linked up with the fact that they carry with them a certain ‘subjectively necessary’ illusion. Because such ideas are illusory, they can only serve as standards in terms of which we conduct our inquiries.”^[6]

A reading such as Grier’s finds support in Kant’s discussion of the regulative use of reason in the Antinomy of Pure Reason. Kant submits that cosmological ideas arise because we think that there is a given “totality of appearances.” The latter is a totality of “conditions” for a given conditioned appearance where every member of the series of conditions is itself an appearance. A conditioned appearance is an object or state of experience that, as such, must stand under the subjective conditions of experience (the forms of intuitions and the categories). This totality is considered either finite (in the theses of the four antinomies) or infinite (in their antitheses). A fundamental step in the solution of all antinomies consists in the claim that it is wrong to assume that there is a “totality of appearances” in the first place.^[7] It is this claim that lies at the basis of Kant’s contention that both theses and antitheses of the first two antinomies are false: “But if it is said that the world is either infinite or finite (not-infinite), then both propositions could be false. For then I regard the world as determined in itself regarding its magnitude.” The assumption that the world is “determined in its magnitude” would turn out to be “false, if, namely, the world were not given at all as a thing in itself, and hence, as regards its magnitude, neither as infinite nor as finite” (KrV, A504/B532). Therefore, the world could have a determinate magnitude if it were a thing in itself. But since the world is, for us, a “totality of appearances,” it means that it does not have a determinate magnitude, which also means that it is false to assume that there is a “given” totality of appearances in the first place.

Once Kant has established that assuming that there is such a totality implies assuming a false proposition, he turns to identifying a legitimate “regulative” use for cosmological ideas. Kant argues that the regress from conditioned to the series of its conditions should not be though as either finite or infinite, but rather as “indefinite” (KrV, A510-15/B538-43). However, since this regress contributes to an improvement of our cognition, and since we cannot really go through this regress without referring to the idea of a complete series as a guidance, we are legitimated to use it regulatively.

Since through the cosmological principle of totality no maximum in the series of conditions in a world of sense, as a thing in itself, is given, but rather this maximum can merely be given as a problem in the regress of this series, the principle of pure reason we are thinking of retains its genuine validity only in a corrected significance: not indeed as an axiom for thinking the totality in the object as real, but as a problem for the understanding, thus for the subject in initiating and continuing, in accordance with the completeness of the idea, the regress in the series of conditions for a given conditioned (KrV, A508/B536).

To summarize, in the context of the Antinomy, Kant seems to argue that the idea of a complete series of conditions that we can legitimately assume regulatively consists of a proposition that we know is false. The assumption is legitimate because, on the one hand, we recognize that it is false and, on the other, it plays a fundamental role in the indefinite regress from conditioned to its conditions which is a source of new cognitions.

b. Ideas and principles that can be assumed “regulatively” imply propositions that we know are true. A second way to read the status of ideas and principles that can be assumed regulatively is to say that they imply propositions that we know are true but in an indeterminate way. For example, we might know that nature is systematic, but this knowledge could be indeterminate in the sense that we do not know how and to what extent we can describe nature in that way. This reading is an attempt to spell out the meaning of Kant’s claim in the Appendix that reason’s ideas and principles have “objective” but “indeterminate” validity (see KrV, A663/B691; A669/B698; A680/B708).

This approach finds support in Kant’s account of the principles of homogeneity, continuity and specification of forms in the Appendix. Kant argues that these are first of all logical “maxims” that guide reason’s research into nature such that we seek: (a) to bring different species concepts under higher genus concepts; (b) to find intermediate species concepts among given species concepts;^[8] and (c) to find species concepts under the known genus concepts. Kant adds that when we investigate nature by following these maxims, we must also presuppose the validity of some corresponding “transcendental” principles according to which homogeneity, continuity and specification are principles that are not simply demands we put on our concepts, but also characteristics of nature itself (see KrV, A654/B682; A657/B685; A660/B688). That these principles must have “transcendental validity” because we must assume that nature itself has those characteristics suggests that they are at least possibly true propositions. Thomas Wartenberg and Jim O’Shea have argued that they are propositions that we know are true, even though to an indeterminate extent.^[9] Let us focus on O’Shea’s account. The idea is the following: the valid universal laws of nature prescribed by the understanding through the categories already presuppose that there are particular empirical laws of nature.^[10] However, if nature did not conform to the principles of homogeneity, continuity and specification, it might turn out that we are not able to identify particular empirical laws. But since this is already presupposed by the a priori legislation of the understanding, it means that nature must necessarily agree to the aforementioned principles, at least to a certain degree.^[11]

c. Ideas and principles that can be assumed “regulatively” imply propositions that can be true or false. Finally, a third way to read the status of ideas and principles that can be assumed regulatively is to say that they imply propositions that, as far as we know, can be either true or false. This reading builds on our ignorance of things in themselves. Let us set aside principles and focus on ideas of reason for the moment. These represent objects that would only be cognizable as things in themselves. Because we cannot cognize objects as they are in themselves, it seems that at least for many propositions concerning these things, we cannot tell whether those propositions are true or false.^[12] But since ideas represent objects that would only be cognizable as things in themselves, it is expectable that they will involve propositions for which we cannot tell whether they are true or false.

Support for this view can be found in Kant’s characterization of the regulative use of reason as “hypothetical”:

The hypothetical use of reason, on the basis of ideas as problematic concepts, is not properly constitutive, that is, not such that if one judges in all strictness the truth of the universal rule assumed as a hypothesis thereby follows; for how is one to know all possible consequences, which would prove the universality of the assumed principle if they followed from it? Rather, this use of reason is only regulative, bringing unity into particular cognitions as far as possible and thereby approximating the rule to universality. (KrV, A647/B675).

If, in using ideas and principles regulatively, we use them as “hypotheses” that guide our inquiry, it seems correct to maintain that, if these hypotheses need to be rational, they must consist of possibly true propositions. The intuition is that it would be odd to claim that our inquiry into nature should be guided by hypotheses that we know are false. More plausibly, it must at least be indeterminate whether they are true or false. A reading along these lines has recently been defended by Willaschek, who argues that assuming a principle regulatively means treating it as a hypothesis. Take the principle of homogeneity as one of such hypotheses. We can use that principle to generate more specific hypotheses, as for example one that says: “there exists a genus x of which y and z are species.” If, by means of our empirical research, we found out that genus x does indeed exist, this would first of all prove that our specific hypothesis was right. Indirectly, this would also provide some support to the general hypothesis that was guiding us, that is the one assuming the principle of homogeneity. In other words, for Willaschek, by treating ideas and principles of reason “regulatively,” we assume as guiding hypotheses propositions that we do not know whether they are true or false but that can be true. Indeed, in the process of drawing conclusions from these hypotheses, we might even gather some evidential support in their favour, but we will never be in a position to prove that they are actually true.^[13]

Let me note that it is possible to accept Willaschek’s account of the status of ideas and principles of reason, that is one that claims that they imply propositions that can be either true or false, without accepting his view on the attitude that is appropriate towards them. For example, Andrew Chignell has argued that the attitude we take towards regulative ideas and principles is one of “belief,” in the Kantian sense of the term.^[14] This is a form of “taking-to-be-true” in which we are legitimately convinced of the truth of a certain proposition on the basis of grounds that are not evidential. A condition for rational belief is that the proposition we belief can, as far as we know, be true. In case we knew or had strong evidence that the proposition is false, our belief would be irrational. Therefore, when we maintain that assuming “regulatively” implies “believing,” we take a position on the status of ideas and principles of reason according to which they imply propositions that, as far as we know, can be either true or false.

3 Willaschek on the regulative/constitutive and the logical/transcendental distinctions

We have now three alternative takes on the status of ideas and principles of reason that can be assumed regulatively. Which one should we adopt? In this section, my aim will be to show that the reading of the status of ideas and principles according to which they imply true but indeterminate propositions is less plausible than it initially seems. In order to do that, I will rely on an argument put forward by Willaschek, according to which it is wrong to conflate the regulative/constitutive and the logical/transcendental distinctions.

In the previous section, we have seen that readings of the status of ideas and principles that affirm that they involve propositions that we know are true rely on Kant’s analysis of the principles of homogeneity, continuity and specification of nature. They take Kant’s claim that these principles are “transcendental” and have “objective but indeterminate validity” to imply that the principles involve some valid cognition of nature. This approach does seem to imply a conflation of the logical/transcendental and the regulative/constitute distinctions. The reasoning appears to be the following: if the transition from a logical maxim to a transcendental principle implies a valid yet indeterminate cognition of nature, it is correct to describe the principle as constitutive. True, both Wartenberg and O’Shea claim that their readings of regulative ideas and principles do not make them constitutive.^[15] In their view, what keeps ideas and principles “regulative” is that they are “indeterminate.” But this seems insufficient to support the claim of “regulativity.”^[16] In fact, it appears that much of the intuitive appeal of the second reading of the status of ideas and principles I presented above rests on the implicit identification of the “transcendental” with the “constitutive.”

Willaschek has recently argued against this often-implicit conflation of the logical/transcendental and the regulative/constitutive distinctions.^[17] He has suggested that the concept “transcendental” should not be treated as a synonym of “constitutive.” Rather, “constitutive” and “regulative” are both uses that a transcendental principle can take. While a logical principle is a prescriptive rule that tells us how to proceed in our inquiry, when Kant speaks of a “transcendental” principle in the Appendix, “transcendental” means something like “which refers to objects and not only to concepts or cognitions.”^[18] In fact, this well suits what Kant says on the principles of homogeneity, continuity and specification. It is because we must assume that homogeneity, continuity and specification are characteristics that pertain to nature as an object that these principles gain “transcendental” rank. But once we have modelled these principles as “referring to objects,” we can either treat them as providing actual cognition of those objects, thus forming propositions that we know are true, or as propositions that we can assume as a guide in our inquiry without taking them as cognition.^[19] The former interpretation of the transcendental principles would lend constitutive principles, the second only regulative ones.

Of course, one can disagree with the details of Willaschek’s interpretation. However, since it provides a viable way for separating the “transcendental” and the “constitutive,” it makes the work of those who argue that regulative ideas and principles imply propositions that we know are true more difficult. As we saw, the intuitive appeal of the latter approach often rests on a conflation of the “transcendental” and the “constitutive.”

4 A tension in Willaschek’s account of the status of ideas and principles

In the previous section, we have set aside one reading of the status of ideas and principles as less convincing. What should we do with the remaining two readings? Both seem compatible with Willaschek’s take on the regulative/constitutive distinction. According to him, a transcendental regulative principle is a description of objects that we do not take as a cognition of those objects, but only as a guide in our inquiry. Unless we specify further what these “descriptions” are, it seems possible that they are either propositions that we know are false or propositions that we do not know whether they are true or false. Therefore, the first and the third reading of the status of ideas and principles are both still on the table.

We know which side Willaschek takes. Since he regards ideas and principles as consisting of propositions that we assume “hypothetically,” he characterizes these propositions as propositions that, as far as we know, can be either true or false. In this section, I would like to highlight a tension in Willaschek’s account of the status of ideas and principles. While he argues that ideas and principles of reason assumed regulatively consist of propositions that are possibly true, his general account of the Supreme Principle of reason – a principle that governs the development of reason’s ideas and more specific principles – suggests that ideas and principles consist of propositions that we know are false.

One key thesis in Willaschek’s interpretation of the Dialectic is that Kant provides an interesting and plausible story regarding how our interest in metaphysical questions is rooted in the very nature of our reason as a faculty of inference. This story unfolds through the transition from reason’s “Logical Maxim” to its “Supreme Principle.” The former is a prescriptive rule which states, roughly, that for every cognition that is inferentially or epistemically conditioned, we must seek its inferential or epistemic conditions up to its unconditioned conditions.^[20] The latter is a descriptive proposition, which says that if there is something that is “real”-conditioned, then there is the totality of its “real”-conditions.^[21] Note that in the Logical Maxim, conditions and conditioned are cognitions, whereas in the Supreme Principle, they are objects or states of affairs. Moreover, the conditioning relation is in the former case inferential or epistemic, while in the latter case, it is “real,” in the sense that it is a form of metaphysical dependence.

To see how Willaschek’s account of the Supreme Principle leans towards an understanding of the status of ideas and principles according to which they imply false propositions, we need to focus on two claims. First, Willaschek argues that the transition from the Logical Maxim to the Supreme Principle unfolds in two steps. The former is a transition from the Logical Maxim to the regulative employment of the Supreme Principle. The latter is a transition from this legitimate regulative employment of the principle to an illegitimate constitutive employment of it. The second claim on which we need to focus is the one according to which the Supreme Principle is false for appearances. Accordingly, Willaschek claims that the Supreme Principle is valid for things in themselves, but is invalid for appearances. “The only conditioned objects of which we can have cognition, according to Kant, are appearances, not things in themselves. But the constitutive Supreme Principle that we need for the inference to the unconditioned holds only for things in themselves (noumena in the positive sense). Hence, the inference is invalid.”^[22]

Since the Supreme Principle is invalid for the objects of our cognition, it seems to follow that when we employ that principle regulatively, we are “assuming” a principle that we know is false. Moreover, since ideas and specific principles of reason result from the application of the Supreme Principle, it seems that they must have a similar status, too. How can Willaschek then claim that they consist of propositions that are possibly true?^[23] I believe that Willaschek can say this because he, in fact, does not understand the regulative and constitutive Supreme Principle as two different employments of the same principle, but rather as two different principles with different contents. Accordingly, he sometimes speaks of two versions of the Supreme Principle,^[24] where the regulative version treats the Principle as a “problem,” in the sense that it does not assume that there is an actually given totality of conditions, but that we ‘“approximate it to universality’ by finding as many conditions for conditioned objects as possible (this being the ‘problem’ or task).”^[25]

While this makes Willaschek’s position on the status of ideas and principles certainly coherent, it is not clear whether it squares well with Kant’s discussion of the regulative use of reason. Specifically, when Kant argues that ideas and principles of reason cannot be used constitutively, but should instead be assumed regulatively, what it seems he is doing is distinguishing different attitudes we can take towards the same concept or proposition. Accordingly, we can take a concept or proposition as expressing a piece of knowledge on an existing object. Alternatively, we can take a concept of, or a proposition on, an object as a guiding principle in our inquiry while being aware that concept or proposition cannot be considered knowledge. For example, this picture is suggested by formulations such as:

To take the regulative principle of the systematic unity of nature for a constitutive one, and to presuppose hypostatically, as a cause, what is only in the idea as a ground for the harmonious use of reason, is only to confuse reason. (KrV, A693/B721).

Here, what generates the difference between the regulative and the constitutive principle of the systematic unity of nature is that we take a different attitude towards the same principle. We “make” the principle “constitutive” by “hypostatizing” the regulative principle, where I take this to mean that we treat the regulative principle as a piece of knowledge.

I am not interested in determining whether Willaschek’s distinction between two versions of ideas and principles of reason is correct. What is interesting to me is the “tension” that this distinction is devised to solve. Since the Supreme Principle, at least in its constitutive form, is false for appearances, it seems that by assuming it regulatively, we would be assuming a proposition that we know is false. It is only by specifying that, when we assume the Supreme Principle regulatively, we are in fact assuming a different principle that Willaschek can argue, in accordance with his account of the “hypothetical” assumption of ideas and principles, that we can describe the proposition we assume as one that is possibly true. I do not think that this tension is due to a problem in Willaschek’s account. Rather, in the next section, I argue that the tension between these two characterizations of reason’s ideas and principles, one according to which these imply propositions that we know are false and one according to which they imply propositions that, as far as we know, can be true or false, is rooted in Kant’s own account.

5 Two types of ideas of reason

In the Critique of Pure Reason, Kant often describes the state of metaphysics at his time as a “battlefield” in which different parties make contrary but equally unfounded assertions. The question regarding how we should deal with these metaphysical disputes is central in the section of the Discipline of Pure Reason dedicated to its “polemical” use. Kant defines the polemical use of pure reason as “the defense of its propositions against dogmatic denials of them” (KrV, A739/B767). The polemic to which Kant refers concerns objects of pure reason, namely, objects that cannot be given in experience but could only be given through reason, as for example God or the soul. Since, according to the Transcendental Dialectic, we cannot cognize these objects, it makes no sense to enter in a metaphysical dispute regarding their existence. Therefore, it would also be illegitimate to react to an unfounded denial that these objects exist with an equally unfounded affirmation that they do.

Why is Kant’s treatment of the “polemical” use of reason interesting for our purposes? It is interesting because Kant distinguishes between two different ways to react to reason’s polemic, which correspond to two different forms the polemic takes. As we will see, this distinction between two forms of the polemic points towards a fundamental difference between two types of ideas of reason: one type consists of propositions that we know are false and another type implies propositions that we do not know whether they are true or false. But let us begin by presenting the two forms of the polemic. The first form corresponds to the conflict of reason that animates the antinomies.

It is worrisome and depressing that there should be an antithetic of pure reason at all, and that pure reason, though it represents the supreme court of justice for all disputes, should still come into conflict with itself. We had such an apparent antithetic of reason before us above, to be sure, but it turned out that it rested on a misunderstanding, namely that of taking, in accord with common prejudice, appearances for things in themselves, and then demanding an absolute completeness in their synthesis, in one or another way (which were both equally impossible), which could hardly be expected in the case of appearances. There was thus in that case no real contradiction of reason with itself in the propositions “The series of appearances given in themselves has an absolutely first beginning” and “This series is absolutely and in itself without any beginning”; for both propositions are quite compatible, since appearances, as regards their existence (as appearances) in themselves are nothing at all, i.e. something contradictory, and thus their presupposition must naturally be followed by contradictory consequences. (KrV, A740/B768).

The antinomies are formed by seemingly contradictory propositions that describe “totality of appearances” in opposing ways, either as forming a “finite” set or as forming an “infinite” set. Kant argues that in the case of this polemic, there is no “real contradiction of reason.” As we saw in Section 2, Kant’s point is that it is wrong to assume that there is a given “totality of appearances” in the first place, which means that both the proposition that describes this totality as finite and the proposition that describes it as infinite are false. Because both propositions can be shown to be false, the polemic dissolves. It is in this sense that the polemic is only apparent.^[26]

We have a polemic of a totally different form when the propositions forming the conflict of reason concern either God or the soul:

However, such a misunderstanding cannot be alleged and the conflict of reason thereby set aside if, say, it is asserted theistically There is a highest being and asserted atheistically, on the contrary, There is no highest being, or when it is asserted, in psychology, “Everything that thinks is of absolutely persistent unity and therefore distinct from all transitory material unity,” against which someone else asserts, “The soul is not an immaterial unity and cannot be exempted from all transitoriness.” For the object of the question is here free of anything foreign that contradicts its nature, and the understanding is concerned only with things in themselves and not with appearances. There would thus certainly be a genuine conflict here, if only pure reason had any thing to say on the negative side that would approximate the ground for an assertion[…]. (KrV, A741/B769).

In the case of propositions affirming and denying the existence of God, or propositions affirming and denying the substantiality of the soul, there can be a real polemic. This is the case because both propositions can be true. How can we stop the polemic in this case? Simply, while both the “affirming” and the “denying” propositions are possibly true, we do not have any ground for taking a side between these alternatives. Any claim of cognition regarding the existence or the substantiality of these objects would be equally unjustified. Moreover, while there is no reason whatsoever to take the side that denies either the existence of God or the substantiality of the soul, affirming that these are given has “at least the interest of reason in their behalf” (KrV, A741/B769). Here, Kant alludes to the fact that while no side of the polemic can present theoretical grounds in their favour, that is grounds that would support a claim to cognition or knowledge,^[27] the side affirming the existence of God or the substantiality of the soul has at least practical grounds in their support. These are grounds that depend on the importance of those objects for our morality.

At this point, it should be clear why this distinction between two forms of the polemic of reason is relevant for our discussion. The distinction points towards a fundamental difference between two types of ideas of reason. While cosmological ideas consist of propositions that we know are false, since they illegitimately assume that a “totality of appearances” can and must be given, psychological and theological ideas imply propositions that can be true. Given that there is this fundamental difference between two types of ideas, it means that any attempt to provide a univocal account of the status of ideas and principles that can be assumed regulatively is flawed from the start.

We still need a better grasp of the ideas that imply or consist of propositions that we know are false. For if it is evident that Kant has cosmological ideas in mind, in the context of the Antinomy of Pure Reason, he also discusses ideas that can be used in true propositions. This happens in the third and fourth antinomies, where Kant asserts that the theses (and the antitheses) can be true. Accordingly, there can be freedom, as an uncaused cause that closes a series of causes, and there can be a necessary being that closes the series of contingent conditioned conditions of a conditioned appearance. In other words, in the case of freedom and the necessary being, it appears to be possible that there is a “totality of conditions for a conditioned appearance.”

Here, I want to emphasize two things. First, obviously, in this case the “condition” that would close the series would not be an appearance, but a thing in itself. Second, even though, if there were freedom and a necessary being, we would have “totality of conditions for a given conditioned appearance,” it would be illegitimate to affirm that freedom and the necessary being exist because “there must be a given totality of conditions for a given conditioned appearance.” This follows from the lack of validity of the Supreme Principle when the conditioned from which we start is an appearance. Still, when, within a series of conditions, there can be conditions of conditioned appearances that are not themselves appearances, the idea that there can be a totality of those conditions is not contradictory. Simply, we cannot claim that there must be such totality.

To sum up, Kant clearly thinks that there are two fundamental types of ideas. One type implies propositions that we know are false. Another type implies propositions that, as far as we know, can be either true or false. Ideas of the first type assume that there is a given “totality of appearances.” This is a totality of “conditions” for a given conditioned appearance where every member of the series of conditions is itself an appearance. Obviously, this type of totality is the one assumed in the first and second antinomy. The second type of ideas might take more than one form. As we have seen, it involves a “totality of conditions for a given conditioned appearance,” where we do think the “totality” starting from a “conditioned” that is an appearance, but where there can be members of the totality that are not appearances. This makes it possible that there is such totality, even though it is illegitimate to assume that the latter must be given. It is not clear whether the second type of ideas also involve ideas that do represent an “unconditioned” condition of a given conditioned, but where we do not arrive at the postulation of that “unconditioned” condition through the idea of a “totality of conditions.” Think of the different ways in which Kant describes reason’s sophisma figurae dictionis in the Antinomy and the Paralogism. The reasoning that brings us to cosmological ideas essentially involves the assumption of a “totality of conditions for a given conditioned.” This is evident because in this case the major premise of the sophisma simply is reason’s Supreme Principle (KrV, A498-501/B526-9). By contrast, the assumption of a “totality of conditions” does not seem to be required for the psychological sophisma. There is no reference to a “totality of conditions” in either its major or minor premises. The major premise describes substance as “what cannot be thought otherwise than as subject” (KrV, B410). This brings us to conceiving the soul as an “unconditioned” object, but not through the idea of a “totality of conditions.”^[28]

6 Conclusion

In this article, I have argued that it is wrong to assume that all ideas and principles of reason have the same status. Rather, we have seen that, according to Kant, there are at least two types of ideas and principles: one first type implies or consists of propositions that we know are false. Ideas and principles have this status when they involve a “totality of conditions” for a given conditioned appearance where every member of the totality is also an appearance. A second type of ideas and principles implies or consists of propositions that, as far as we know, can be either true or false.

Kant’s distinction between these two types of ideas and principles has important consequences for how we should conduct our inquiry into their status. First of all, I believe that it would be more productive if efforts were directed towards distinguishing more clearly between these types of ideas and principles, rather than in devising new arguments in favour of a univocal account of their status. For, while it is clear that there are these two types of ideas, it is not clear which items fall under each type, and for what reasons. For example, the ideas animating the first and second antinomies are not the only representations of reason that involve propositions that we know are false. In the Appendix to the Dialectic, Kant attributes a similar status to the idea of the continuity of the forms of nature:

But it is easy to see that this continuity of forms is a mere idea, for which a corresponding object can by no means be displayed in experience because the species in nature are really partitioned and therefore in themselves have to constitute a quantum discretum, and if the graduated progress in their affinity were continuous, they would also have to contain a true infinity of intermediate members between any two given species, which is impossible. (KrV, A661/B689).

Therefore, when we address the status of the principle of continuity, the challenges are, first, how the principle of continuity is related to the first two cosmological ideas and, second, whether the principle might also involve the thought of a “totality of appearances” similar to the one identified in the Antinomy of Pure Reason.

A second consequence of Kant’s distinction between two types of ideas concerns Kant’s derivation of the transcendental ideas in the Dialectic. Especially in the Introduction and the first chapter of the Dialectic, Kant gives the impression that all ideas have a similar origin and that the Supreme Principle plays a central role in it. However, if it is true that, according to what Kant says in the Discipline, there are two fundamentally different types of ideas, it might turn out that what Kant suggests at the beginning of the Dialectic is misleading and that the derivation of the ideas is not traceable back to a single pattern. To be clear, I am not arguing that there is not such a single pattern of derivation. Rather, I am suggesting that Kant’s contentions on reason’s polemic and the different types of ideas animating it show that it cannot be presupposed that there is one.