Division II of Transcendental Logic is entitled “Transcendental Dialectic.” The definition of transcendental logic, and the distinction of transcendental logic into analytic and dialectic, have already been briefly discussed. The goal of the Dialectic is to uncover a natural “dialectical illusion,” which results from the attempt to apply the principles of the understanding beyond the limits of the understanding, to “objects that are not given, or indeed . . . objects that perhaps cannot be given in any way at all” (A63/B88).
Specifically, the objects to which the understanding (mistakenly) makes “hyperphysical” application of its principles are three: the rational soul, the totality of the world, and God. Corresponding to these problematic objects there are in the philosophy of Wolff three branches of “special metaphysics”: rational psychology, rational cosmology, and rational theology, respectively. The critical treatment of the claims of special metaphysics is found in the “Paralogisms,” the “Antinomy,” and the “Ideal” of pure reason. This document will cover both the introductory material in the Dialectic and the Paralogisms (as the latter are exposited in the second edition of the Critique).
The general goal of the Dialectic is to expose the sources of a kind of cognitive illusion, in which something appears to be what it is not. The senses and understanding operating by themselves (“without the influence of some other cause”) are not capable of such error. The senses make no judgments at all, and so make no judgments about how things are, and any judgment made by the understanding according to its own laws must be correct, since accordance with those laws is the criterion of truth. Since neither sensibility nor undertanding can err on its own (and since sensibility makes not judgments), where there is illusion, the cause is “sensibility’s unnoticed influence on the understanding” (A294/B350).
It might be thought that transcendental illusion is the result of the transcendental use of the categories, which takes place when the understanding unknowingly oversteps its bounds and tries to apply the categories beyond the sphere of objects of experience. But Kant’s target is instead principles which require that there be no distinction between objects of experience and other objects. Kant calls these principles “transcendent,” removing the limits of experience altogether. The principles defended in the Analytic are, by contrast, “immanent” (A296/B352). In the Prolegomena, Kant claims that the transcendent use of the categories produces transcendental illusion, “by which metaphysics has been deceived hitherto and misled into childish efforts of catching at bubbles” (First Part of the Main Transcendental Problem, Remark III, Ak 4:292).
The illusion engendered by transcendent principles is, according to Kant, unavoidable. It arises because the understanding mistakes principles based on cognitions which arise solely subjectively for principles that are based on objective cognitions. The fact that the source of the error lies within us means that the “transcendental illusion” to which it gives rise “attaches to human reason unpreventably” (A298/B354) and will always deceive us when we let our guard down. This unavoidable tendency to fall into error is contrasted with mere errors in inference, which occur through bad reasoning or attempts to deceive.
Kant regarded reason as the highest faculty of the human mind, ranking above sensibility and understanding. While the understanding unites what is given in sensibility, reason’s function is to unite the cognitions of the understanding. In its logical use, reason makes inferences which bring together material thought separately in different judgments. If I judge that all composite things are changeable, and that all bodies are composite things, I can conclude that all bodies are changeable (A330/B387). The concept of a body has been brought into unity with the concept of the changeable through the act of inference. The result may be called a principle. “Reason in making inferences seeks to reduce the great manifoldness of understanding’s cognition to the smallest number of principles (universal conditions) and thereby to bring about the highest unity of this cognition” (A305/B361).
Besides its logical use, reason has a real or pure use. In the case of inference, the unification of the concepts in the conclusion of the inference occurs on the basis of premises which themselves are unities. So the premise that all composite things are changeable would be the conclusion of another inference, which has its own premises, which serve as its conditions. The task of reason in its logical use is “to find, for understanding’s conditioned cognition, the unconditioned whereby the cognition’s unity is completed” (A307/B364). This is merely an injunction that we not be satisfied in the process of inference until we have found some condition that is not itself conditioned.
This guiding “maxim” for inference attains its status as a principle
of pure reason only on the assumption “that, if the conditioned is given, then the entire series of conditions subordinated to one another—a series that is hence itself unconditioned—is given” (A307-8/B364). This assumption is more than mere guidance in the conduct of inference. It is a synthetic principle that tells us that there is an unconditioned given along with the conditioned. Without this principle, for all we know our search for absolute unity of the cognitions of the understanding might be in vain. (And indeed, Kant thought that it is in vain.)
This “supreme principle” (A308/B365) of reason is transcendent with respect to appearances. The rules of the understanding only require a condition for any conditioned (e.g., a cause for any change of the state of an object), but do not dictate that the entire series (e.g., of causes) be given along with any change. So the question arises as to whether this principle is correct, either with respect to the synthesis of the series of appearances or with respect to the synthesis “of the thinking of things as such” (A308/B365). Or is the more modest rule that we should search for completeness the only justifiable principle of reason? The outcome of the Dialectic will be that the so-called “supreme principle” of reason is the result of transcendental illusion.
The Transcendental Dialectic is divided into two Books, the first being very short, while the second is quite lengthy. The topic of the first Book is the cognitions specific to reason, which Kant calls “ideas.” The second examines the principles governing the ideas. In this way, the structure of the Dialectic parallels that of the Analytic, which begins with a treatment of the categories of the understanding and then considers principles which apply them to objects.
Kant took over the word ‘idea’ from Plato, who intended it to signify something which cannot be found in experience, and to which objects of experience are never “congruent” (A313/B370). (By contrast, Aristotle’s categories are congruent with objects of experience.) Platonic ideas, as Kant conceived them, are “archetypes of things themselves, and not merely keys to possible experiences, as are the categories” (A313/B370). They flow from divine reason, which imparts them to human reason, which in turn must laboriously recall them. Plato invoked the idea because he recognized a need of our cognitive power, a need which is not satisfied merely with unifying appearances into experience. Rather, it is compelled to look beyond experience to objects with which the objects of experience could not be congruent (A314/B271).
Kant rejected as mysticism what he took to be Plato’s theory of our relation to the ideas (the doctrine of their divine origin and of recollection) as well as his hypostatization of them (making them into separately existing objects). They are of greatest use as models for behavior, as with the idea of virtue, which, according to Kant’s version of Plato, is required for “any judgment about moral value or lack of value” (A315/B372). Plato extended his use of the forms to nature, as for example as an explanation of the organization of living things. Kant’s moral philosophy and his philosophy of nature will reflect these Platonic endeavors, though in quite a different way from Plato’s.
After eulogizing Plato, Kant provides a hierarchical chart (Stufenleiter) of presentations (A320/B376-7). He defines a notion as a pure concept (category) which is not thought through a schema. Then he defines an idea as “a concept framed from notions and surpassing the possibility of experience” (A320/B377). So an idea is a pure concept of reason, whose origin lies solely in the understanding, untouched by sensibility. Kant decries the use of ‘idea’ to describe “the presentation of some red color” as a debasement of what was initially conceived as the pinnacle of human cognition (A320/B377).
Just as the table of forms of judgments is supposed to guide the formation of the table of categories in the metaphysical deduction, the a classification of the forms of inference is expected to generate the “transcendental” ideas (A321/B378). The generation of the ideas has two phases. In the first phase, we arrive at a single transcendental concept of reason, that of a totality of conditions. The second phase is to apply this notion to the three relational forms of judgment: the categorical, the hypothetical, and the disjunctive.
The argument in the first phase depends on the semantical analysis of a special kind of inference, in which an individual is brought under a concept.
Consider the categorical judgment “All human beings are mortal.” We find that the subject of that judgment is the predicate of the judgment “Caius is a human being.” We could say that the concept human being is conditioned in this respect. But the intuition of Caius is not conditioned, since it is not the predicate of any subject. This is the basis for the first transcendental idea, the psychological idea of the soul.
In a hypothetical judgment, if p then q, the consequent can be said to be conditioned by the antecedent. So the unconditioned would be an antecedent that is not the consequent of any antecedent, or “the presupposition that presupposes nothing further” (A323/B380). This gives rise to the cosmological idea of the world as a totality that is not part of anything else.
In a disjunctive judgment, as Kant understood it, the subject A of the judgment is joined to an exhaustive list of exclusive predicates. For example, “The world exists either through blind chance, or through internal necessity, or through an external cause” (A74/B99). If we think of each disjunct as a condition, the unconditioned corresponding to the disjunctive judgment would be something in which there is an exhaustive list of predicates which are not separated by disjunction. The idea of such an object is the idea of a “most real being” or God.
Kant speaks of the unconditioned as an absolute totality. Here by ‘absolute’ he means what is not comparative or restricted in any way. He notes that the word had recently been used to indicate what is intrinsic to a thing, which leads to confusion, particularly in classifying types of modality.
It is true, Kant says, that the transcendental ideas are “only ideas.” There is no object of experience that corresponds to them. But they have great use in morality and even some use for the regulation of thought in experience, as we will discuss in the final lecture.
One feature of an absolute totality is that it applies only to the regressive or ascending series, and not to the progressive or descending series. Thus if the series is one of causes, the idea is that of the totality of prior causes. Reason claims that there must be such a totality, in order that the present state could have occurred at all. On the other hand, even reason does not presume to have an idea of the totality of the later states, which exist only potentially (A337/B394).
After having given the logical form of the transcendental ideas, Kant must now show how to generate the list of objects to which these forms apply. It is clear that he had in mind the three objects of Wolffian special metaphysics, the soul, the world, and God. He generates his result somewhat artificially, by distinguishing “what is universal about all reference that our presentations can have” (A333/B390). We get the following list:
Corresponding to each of the three forms of inference that generate the ideas is a “subtly reasoning inference” whose premise is something familiar and whose conclusion is the existence of an object corresponding to the idea. We go from a concept which is legitimate to an object of which we have no concept. The absolute unity of the thinking subject is generated by a transcendental paralogism, the absolute unity of the series of appearances by a transcendental antinomy, and a “being of all beings” (God) by a transcendental ideal. The remainder of this document will be concerned with the paralogisms of pure reason. Subsequent lectures will discuss the antinomy and ideal.
The first of the three types of subtly reasoning inference is paralogism. Logical paralogism is what we would now call formal fallacy, such as affirming the consequent. Committing the fallacy is an error that is not prompted by the content of the inference. But transcendental paralogism is a fallacy which is prompted by its content. As such, it will “have its basis in the nature of human reason, and will carry with it an illusion that is unavoidable but not unresolvable” (A341/B399).
The dialectical inferences of pure reason are supposed to start with a legitimate concept and take us to an object of which we have no concept. In the set of inferences which concern relation of concepts as predicates of a subject, the legitimate concept of a subject with which we begin is “I think.” Since “I think” is strictly speaking a judgment, Kant really has in mind “I, as a thinking being” (A342/B400). This is not one of the categories, but it belongs with them as “the vehicle of all concepts as such” (A341/B399). Since the categories are transcendental (a priori) in character, “I, as thinking being” is a transcendental concept as well. Its function is “only to bring forward all thought as belonging to consciousness” (A341/B400).
If this concept of the thinking I is treated without relation to experience, then it is the subject of rational (not empirical) psychology. The fact that the I is given in inner sense does not detract from the purity of rational psychology, because the thinking I is a condition even for inner sense, and so is prior to any determination of the I in inner sense. So if one tries to apply the concept of the thinking I to an object, one would have to apply “transcendental predicates” to the object. These transcendental predicates would be taken from the general classification of the first two types of category (the “mathematical” categories) and one category each from the second two types (the “dynamical” categories).
Following this procedure, Kant produces a table of transcendental predicates attached to the thinking I (now called the soul). He begins with the category of substance and proceeds backwards through the table of categories. Each of these predicates brings with it a further concept that reveals the real interests of rational psychology.
| Category/Categories | Transcendental Predicate | Concept of Interest | |
| 1 | Substance | Substance | Immaterial |
| 2 | Quality | Simple | Incorruptible |
| 3 | Quantity | Unity (over time) | Personal |
| 4 | Possibility | Related to possible objects in space | Interacting with bodies |
While the transcendental predicates in items 2 and 3 make some sense as bases for the stated doctrines in psychology, the categories in the other two items are a stretch. For example, many philosophers granted that there can be material substances, so merely classifying the thinking I as a substance seems to imply nothing about its immateriality, while item 4 insinuates that the soul is distinct from body.
The basis of all the paralogisms that move from the concept of the thinking I to the soul is “nothing but the simple, and by itself quite empty, presentation I, of which we cannot even say that it is a concept, but only that it is a mere consciousness accompanying all concepts” (A345/B403). Kant calls it “the transcendental subject of thoughts = x,” which can only be thought as an object through the thoughts it has. There is no presentation of it apart from its thoughts because it is the mere “form of presentation as such” (A346/B404). This form is valid for all thinking things (even though it is presented indexically as “I”) through a “transfer of this consciousness of mine to other things, which thereby alone are presented as thinking beings” (A347/B405).
For the exposition of the paralogisms, we follow the second edition. We are first told that if any inferences are made from the proposition “I think,” they would involve a transcendental use of the categories, which had already been condemned. In the second edition, Kant cites brevity as his reason for treating all the paralogisms together, which departs from his having assigned them separate sections in the first edition.
A preliminary remark reveals the inappropriateness of the inferences. The thinking I is what makes determination of intuition possible, and determination of intuition is what makes for cognition of an object. So the “I think” is “consciousness of the determining self” (B407). Rational psychology demands a cognition of the soul, which means that there must be an intuition which is determined in a certain way. This would be the intuition of myself in inner sense. Unlike the thinking I, the I of inner sense is determinable, not determining. (That is, it is capable of taking on properties, but does not determine which properties it takes on.)
1. (First Paralogism
in A.) If I call the soul substance, I am determining it as an object or as a being that subsists by itself. It might seem that this can be done because the I can exist only as subject and never as predicate of a subject, and hence falls under the category of substance. But this fact implies nothing about the way in which the subject exists as an object. “This latter claim goes very far, and hence it also requires data that are in no way found in thought” (B407).
2. (Second Paralogism
in A.) If the soul, now taken as substance, is determined as simple substance, it is determined with respect to whether it has component parts. To be sure, there is a logical “simplicity” to the thinking I, in that it cannot be resolved into a number of distinct subjects. Again, this implies nothing with regard to the manner of the soul’s existence. The simplicity of the thinking I is the result of its impoverishment, its lack of content. Kant noted that it would be miraculous if from this empty concept one were able to infer the answer to a knotty metaphysical issue, “as if through a revelation” (B408).
3. (Third Paralogism
in A.) If the soul, now taken as simple substance, is determined as identical through time, it is determined with respect to a form of intuition, from which we abstract when we think of the thinking I. This does not follow from the analytic truth that I am the same through all of my thoughts. A number of synthetic principles (not specified by Kant) would have to be invoked in order to prove my identity through time.
4. (Replacement for Fourth Paralogism
in A.) Finally, if the soul is determined as a merely thinking thing which can exist without any relation to bodies (as Descartes had claimed), I am determining the relation of the soul to other things. Now it is an analytic truth that I as a thinking being am distinct from other things which are not thinking, such as my body. But this says nothing about the relation of dependence between myself and these distinct things.
It should be noted that here Kant has shifted the fourth topic from the one announced in the table given above. There, the soul was portrayed as interacting with bodies. Here, it is portrayed as existing independently of bodies. Both these accounts of the soul were held by Descartes. In the first edition Fourth Paralogism, Kant sought to demonstrate the existence of things existing outside me, as we have seen.
If these inferences were valid, we would have generated by them a priori synthetic judgments about the way in which the soul exists. This would subvert the entire critical project by allowing the understanding to extend its reach to (one kind of) noumena, to “things as such and in themselves” (B410). But that danger is slight, as can be seen by examination of the following inference to the substantiality of the soul.
What cannot be thought otherwise than as subject also does not exist otherwise than as subject, and therefore is substance.There is an ambiguity in the use of the expression “thought as subject” in the inference. In the first premise (the major), it is used in its most general way, applying to any object, including an object of intuition. In the second premise (the minor), which mentions being “thought as subject,” there is no application to an object, but only “the reference to oneself as subject (as the form of thought)” (B411). So there is no reference to any intuition whereby it can be given as an object. The only thing that really follows from the argument is “that in thinking my existence I can use myself only as the judgment’s subject” (B412). I am the thing that thinks my own existence.
Now a thinking being, considered merely as such, cannot be thought otherwise than as subject.
Therefore it exists only as a subject, i.e., as substance. (B410-11)
This criticism of the argument is confirmed by claims Kant had made earlier. He had said that the category of substance makes no reference to objects except through intuition, which in the case of substance would be intuition of a permanent. There is nothing permanent in inner sense, and only inner sense could provide the intuition of an object that could be called the soul. “Hence if one remains with mere thinking, then one also lacks the necessary condition for applying the concept of substance—i.e., the concept of a self-subsistent subject—to oneself as a thinking being” (B413). Without the substantiality of the thinking I, its simplicity could never be proved, and we are left only with the the logical simplicity of the subject.
The practical interest of rational psychology lies in the demonstration of the immortality of the soul. If Kant is right, the fallaciousness of the argument for the simplicity of the soul undercuts the demonstration of immortality. He went further though, and argued that even the assumption that the soul is simple could not establish that it is immortal.
The standard argument for immortality is that a simple being, which has no parts, is incapable of division, and division is the only way in which an object can be naturally destroyed. (Philosophers such as Leibniz held that it could still be supernaturally destroyed by an act of God’s will.) Against this, Kant’s contemporary Moses Mendelssohn noted that the soul might be destroyed by vanishing even if it were simple.
The argument begins with the premise that the soul has no parts. If it has no parts, then it cannot diminish, since it would have no parts to lose. So if it were to vanish from existence, there would be a time at which it fully exists, followed immediately by a time in which it does not exist at all. But this is impossible, as any transition from existence to non-existence must be gradual. So, if the soul is simple, it cannot vanish.
Kant’s response is that one of Mendelsshn’s premises is false. A thing which has no parts has no extensive magnitude, but it might still have an intensive magnitude. And in fact, the empirical self that is the object of inner intuition does have degrees of consciousness. So it could gradually make the transition from existence to non-existence through an erosion of its powers of perception.
In a footnote to B415, Kant takes the issue further and notes that the doctrine of the simplicity of the soul does not rule out its division into several substances or the melding of several souls into one. If the powers of the soul have intensive magnitudes, their diminution might be a kind of leakage, with the quantity lost making up the beginnings of a new soul. The division and subsequent melding might explain how the souls of children are formed from the souls of their parents.
Of course, Kant regarded such hypotheses as inadmissibly transgressing the limits of experience. His point in bringing them up is that a materialist might seize on such an explanation to account for the generation of new minds from the “simple substances that give us the appearance of matter” (B417). This explanation would not be drawn from experience, nor could any behavior of matter account for how what is lost from the parents is replaced. The point in discussing it is that the idea of the soul does not automatically favor the “rationalist” (such as Wolff) over the materialist. In fact, someone favoring the immortality of the soul would be better off simply to “admit that he does not know how to explain the possibility of a thinking nature” (B417).
The doctrine of rational psychology is open to the possibility that the soul exists by itself, without there being any material things. First, one can determine that one exists simply from the concept of the thinking I. Also from this concept, one can deduce that one is a permanent being. Thus, the existence of the I as a permanent being is settled without any appeal to objects in space. This makes problematic or skeptical idealism “unavoidable.” That is, “if the existence of external things is not thus required for determining one’s own, then, by the same token, one is only assuming their existence quite gratuitously, without ever being able to give a proof of it” (B418). Kant thought that problematic idealism can be defeated if rational psychology is rejected.
Kant also claimed that materialism can be defeated if we make legitimate use of the concept of the thinking I. We can correctly say that the I is simple, in the logical sense. In the guise of the unity of apperception, it is that to which all combination and separation refer. This simplicity cannot be derived from anything material, because there is nothing simple in material nature. Even a mathematical point is not simple, but only a limit.
The defeat of materialism does not, however, imply the truth of spiritualism, the doctrine that there are no material things. The existence of the thinking I that is simple is determined empirically. It is not based on the general principle that every thinking thing exists, since that would make every thinking thing a necessary being: existing solely by virtue of its concept. Instead, I know that I exist thinking because I am aware of thoughts that I have. But to determine myself as thinking, I must determine myself in time, and this requires reference to something permanent. This permanent is material, and so spiritualism is refuted.
In a concluding section, Kant finally pursues the issue raised in the original fourth topic, which is the interaction between soul and matter. In Kant’s terms, the difference between soul and matter is that between the object of inner sense and the object of outer sense, respectively. The former is in time only while the latter is in space as well. Kant notes that this difference is found only at the level of appearance, and that “what it is, as thing in itself, that underlies the appearance of matter might perhaps not be so heterogeneous” (B428). So it is at least logically possible for the interaction of soul and matter to have a basis in something common. As for how appearances can interact, Kant refers us to the Third Analogy.
The true use of rational psychology is not as a doctrine, but as a discipline. Properly conceived, the thinking I must itself be thought in a way that keeps us from “throwing ourselves into the lap of soulless materialism” and “from getting lost while roving about in spiritualism” (B421). If we adhere to the discipline and refrain from speculation, we can turn our attention to moral reasons for assuming immortality. Speculative proofs have never convinced common human reason and are precarious even in the schools.
Reason has its own realm: not the illusory noumena, but “the order of purposes, which is yet simultaneously an order of nature” (B425). The human being, who is “the ultimate final purpose of the order” (B425), is not bound by it. He can formulate purposes which go beyond the rewards and advantages to be found in natural life. What is ultimately important to the human being is “the moral law within him” (B425), which motivates him “to be a citizen of a better world, which he has in his idea” (B426). This is the basis for a belief in a future life: an ideal of unlimited improvement.
Even though rational psychology fails in its attempt to determine the thinking I through pure concepts of understanding, it may be that the I can be determined by principles of practical reason. If these principles, which spring from our own reason, tell us what we ought to do, then in a sense they “determine” our existence. This would give us a way to determine our existence “in reference to an intelligible (although only thought) world” (B431). This determination would not, though, involve the predicates of rational psychology, which are predicates of theoretical reason.
In the last paragraph, Kant gives us a hint that the categories of the understanding might have application beyond experience in the service of practical reason. “I would be entitled to apply these concepts, in accordance with the signification analogous to the one found in reason’s theoretical use, to freedom and its subject” (B431). Thus I could think of cause as basis of free action, even if I understand this action apart from the conditions of experience. This theme will be developed at length in our discussion of the Third Antinomy.