After elucidating the principle, Kant notes (in the second edition) that it is threatened by idealism. He undertook to remove the threat in interpolated passage. Kant had tried in the first edition of the Critique to refute idealism in another location. The original attempt is found in the Fourth Paralogism. This section was removed from the second edition in favor of the Refutation of Idealism. Kant notes in the Preface to the second edition that the change is only in “the kind of proof I offer” (Bxxxviii). In a footnote he famously asserted that it is a “scandal for philosophy and human reason in general” that the existence of things outside us had heretofore been based on faith rather than proof (Bxl).
There is a certain urgency in Kant’s attack on idealism in the second edition. The first review of the first edition, by Garve and Feder in 1782, had dismissed Kant’s position as nothing but a “freshened up” Berkeleyan idealism. In the Prolegomena, Kant denounced this interpretation and tried to distinguish his position clearly from what he took Berkeley’s to be. He began by noting that his system was described as idealistic.
In order to take a position from which my reviewer could most easily set the whole work in a most unfavorable light, without venturing to trouble himself with any special investigation, he begins and ends by saying: “This work is a system of transcendental (or as he translates it, of higher) idealism” (Prolegomena, Appendix, Ak 4:373).Kant takes pains in the Prolegomena to distinguish other kinds of idealism from his kind of idealism (which he even re-names from “transcendental” to “critical” at Ak 4:294).
In the second edition Preface, Kant called attention to several “improvements” he had made in the exposition of the doctrines of the Critique. One such “improvement” was “to remedy the misinterpretation of the the paralogisms advanced against rational psychology” (Bxxxviii). Kant described the change as the only one that could properly be described as an “addition” to the Critique, and he minimized the extent to which it was an addition by stating that it involved only the kind of proof offered against a kind of idealism. The new Refutation of Idealism is indeed markedly different from the argument in the first edition of the Critique, and it is not to be found in the Prolegomena, which was written between the publication of the two editions of the Critique.
One of the difficulties in understanding Kant’s attitude toward idealism is that he distinguished several kinds of idealism, and he used a broad array of adjectives (some of them co-extensive) to describe them. In this section, the various types of idealism identified in each of the three texts will be defined. Before turning to the texts themselves, we will lay out a framework for classifying types of idealism.
The kind of idealism Kant was trying to dispatch is called “material” in the Refutation of Idealism (B274). Material idealism and material realism are concerned with matter, or alternatively, with extended, impenetrable bodies. There are two basic forms that material idealism can take. One kind is metaphysical, concerning the existence of matter, and the other is epistemological, concerning the possibility of certainty that matter exists. Thus, the metaphysical material idealist denies the existence of matter, and the epistemological material idealist claims that the existence of matter cannot be known with certainty.
In the Fourth Paralogism, idealism is taken to be an epistemological thesis, which is contrasted with “dualism.” More specifically, idealism is said to be a “doctrine” of the “ideality of outer appearances,” which in turn is an “uncertainty” according to which “the existence of all the objects of outer senses is doubtful” (A367). Kant goes on to describe what an “idealist” is:
Hence by idealist we must mean, not someone who denies the existence of external objects of the senses, but someone who merely does not grant that this existence is cognized through direct perception, and who infers from this that we can never through any possible experience be completely certain of their actuality. (A368-9)Opposed to this idealism is the “dualist,” who “can grant the existence of matter without going outside mere self-consciousness and without assuming anything more than the certainty of presentations in me and hence the cogito, ergo sum” (A370). Kant describes the dualist as an “empirical realist,” which suggests that the “idealist” described here is an “empirical idealist.” Kant also notes that it is (at least in some cases) “psychologists who adhere to empirical idealism” (A372). The reference to the cogito, ergo sum suggests that Descartes is taken by Kant to be such a psychologist. So we can take it that ‘empirical,’ and ‘psychological’ are both words used to describe epistemological material idealism. The evidence for the existence of matter is empirical, and the fact that the existence of the self is supposed to be the starting point for our knowledge of matter makes the investigation psychological.
Later in the Fourth Paralogism, Kant makes a new distinction. “The dogmatic idealist would be the one who denies the existence of matter; the skeptical idealist the one who doubts matter’s existence because he considers such existence to be unprovable” (A377). It seems that the “empirical idealist” here is being identified with the “skeptical idealist.” Dogmatic idealism is a version of metaphysical material idealism, while skeptical idealism is a version of epistemological material idealism. The term ‘skeptical’ refers to the outcome of the investigation of matter which has an empirical, psychological, starting-point.
The dogmatic idealist is one who “believes that he finds contradictions in the possibility of a matter as such,” and he will be dealt with later, in the Antinomy of Pure Reason. The apparent target of the later refutation is Leibniz, as noted below.
The Prolegomena introduces a different scheme of classification. Here idealism as such is defined as a metaphysical, not an epistemological doctrine.
Idealism consists in the assertion that there are none but thinking thing beings; all other things which we believe are perceived in intuitions are nothing but presentations in the thinking things, to which no object external to them in fact corresponds. (Ak 4:288-9)This kind of idealism is also called “genuine idealism,” in which “the existence of the thing that appears is . . . destroyed” (Ak 4:289). In the Appendix, this “genuine idealism” is associated with the Eleatics, Plato and Berkeley. Kant also calls this idealism “mystical and visionary,” and a “phantasm” (Ak 4:293). On the other hand, “empirical idealism” (also, “dreaming idealism”), which appears to be a version of epistemological idealism in the A edition of the Critique, cannot fall under this classification, because it is a doctrine that doubts the existence of things, though as a species of “idealism” it seems that it ought to be taken as a metaphysical doctrine, given the general definition of idealism given in this part of the Prolegomena.
The target in the Preface to the B edition is described as “psychological idealism” (Bxl). (In the Refutation itself it is called “material idealism” (B274).) The apparent reason it is called “psychological” is that it begins with psychological facts (e.g., that I myself exist) and claims that such facts constitute an insufficient basis for drawing conclusions about non-psychological, material, objects.
As noted above, material idealism is a doctrine about material things, things in space. Kant was trying to refute one of two views. The first is the epistemological view, problematic idealism,
that we cannot know that there are material things: “if it occurs to anyone to doubt [the existence of things outside us], we have no satisfactory proof to oppose him” (Bxl). Descartes is singled out as its representative (B274). The second is the metaphysical view that the existence of objects in space is “false and impossible” (B274). Berkeley is held up as the paradigmatic dogmatic idealist.
Berkeley’s thesis is said to be that, “space, with all the things to which space attaches as inseparable condition, is something to be regarded as in itself impossible” (B274). As he had already demonstrated the possibility of space in the Aesthetic, Kant considered dogmatic idealism as thereby refuted. The Cartesian thesis is that “we are unable to prove by direct experience an existence apart from our own” (B275). This is “reasonable and is in accordance with a thorough philosophical way of thinking—viz., in permitting no decisive judgment before a sufficient proof has been found” (B275). Thus, it is the target of the Refutation.
Before we turn to the argument, it is worth developing Kant’s characterization of idealism in more detail. Kant openly espoused the thesis of the “transcendental ideality” of space and time, and hence the ideality of objects in space and time. As a self-avowed idealist, Kant was immediately put on the defensive and forced to distinguish his version of idealism from material idealism. He did this in various ways in the Prolegomena.
Kant dealt extensively with Berkeley in the Prolegomena. He had not even mentioned Berkeley in the first edition of the Critique, though he had identified Descartes as a skeptical idealist there. The dogmatic idealist was taken to be someone whose grounds could only be that “he believes that he finds contradictions in the concept of matter as such” (A377). It is pretty clear that Kant had Leibniz in mind in this context. The ground for dogmatic idealism was supposed to be removed in the next chapter, and in the Second Antinomy, Kant had presented the argument for monads, which are simple substances. Since everything in space is divisible, monads are not spatial and hence not material.
In the Appendix to the Prolegomena, Kant claimed that “genuine” idealism or idealism “proper” is really mystical. By this, Kant meant that the idealist counts as real only what can be cognized a priori. On this view, as Kant described it, our access to the a priori is through an intellectual intuition, which Kant denied that we have, calling it nothing more than an alleged mystical vision.
The dictum of all genuine idealists, from the Eleatic school to Bishop Berkeley, is contained in this formula: “All cognition through the senses and experience is nothing but sheer illusion, and only in the ideas of pure understanding and reason is there truth.” (Ak 4:374)Berkeley is thus classified with Parmenides and Plato as an extreme rationalist. The argument for this “mystical and visionary” idealism would be that space is cognized only through experience, in which case it, as well as all things in it, are illusory. Thus, the mystical idealist “convert[s] actual things (not appearances) into mere presentations” (§13, Ak 4:293). Kant had refuted the premise of this argument in the Aesthetic by treating space as an a priori form of intuition. It appears that Kant understood Berkeley as an extreme rationalist on the basis of one of Berkeley’s later works, Siris (1744).
In the second edition of the Critique, Kant presented Berkeley’s argument quite differently. Here, the problem with space is not that it is merely empirical, but that it is metaphysically suspect. If space is regarded as a thing in itself, then it is neither substance nor inherent in a substance, and so it is nothing at all. In that case, “we can hardly blame the good Berkeley for degrading bodies to mere illusion” (B71).
Descartes is mentioned in both editions of the Critique as well as in the Prolegomena. His alleged idealism is called “skeptical” (Appendix) and “empirical” (§13). It is called empirical rather than dogmatic because it takes the empirical evidence for the existence of things in space to be inadequate, though unlike the mystical idealist, the empirical idealist would accept an empirical proof if he could find one. But according to Kant, Descartes did not find one, and so “he thought every one at liberty to deny the existence of the corporeal world because it could never be proved satisfactorily” (§13, Ak 4:293). This is a puzzling claim, because Descartes in fact asserted he had proved that the corporeal world exists, in the Sixth Meditation.
In an apparent allusion to Descartes’s claim in Meditation One that he could not infallibly distinguish between waking and dreaming, Kant also makes a reference to “dreaming idealism.” This is an idealism that “changes mere presentations into things” (Ak 4:293-4). What Kant seems to have meant is that because of the possibility that our experience might all be a dream, the presentations of dreams are elevated to the same status as that of material objects.
In the first edition Fourth Paralogism, Kant gave the argument for empirical idealism.
That whose existence can only be inferred as the existence of a cause for given perceptions has a merely doubtful existence.This argument has its seeds in the Third Meditation, where Descartes finds that, “As to my ideas of corporeal things, I can see nothing in them which is so great as to make it seem impossible that it originated in myself.” Descartes could accept the first premise, though with the proviso that “perceptions” be limited only to “sense perceptions.” He would not accept it on the grounds Kant adduced in its favor, that “the inference from a given effect to a determinate cause is always uncertain, because the effect may have arisen from more than one cause” (A268). Descartes believed that one could infer from the idea of God to God as a cause, but that the way in which this is done does not apply to the inference from ideas of bodies to bodies as causes. Only the idea of God is “great” enough for us to be able to infer a cause greater than ourselves as its origin (Third Meditation).
Now all outer appearances are of such a kind that their existence cannot be preceived directly, but that these appearances can only be inferred as the cause of given perceptions.
Therefore, the existence of all objects of outer senses is doubtful. (A366-7)
Even granting that Descartes could accept the first premise, he would have to reject the second premise. He thought that the inference to the existence to bodies is not based on the ideas they allegedly cause, but rather on his conviction that they exist. He could only explain the strength of this belief through their actually existing, since God created him and is no deceiver (Sixth Meditation). Since Kant did not allow that God’s existence can be proved metaphysically, he might have thought that Descartes had no grounds for rejecting the second premise.
The problem with the argument, Kant maintained, is that the argument is sound only if by “outer” we mean “something that as a thing in itself exists as distinct from us” (A373). The objects of the outer senses, therefore, would be things in themselves and would be unknowable.
For, indeed, if we regard outer appearances as presentations produced in us by their objects where these objects are taken as things that are in themselves outside us, then it is impossible to see how we can cognize the existence of these objects otherwise than by the inference from effect to cause; and in this inference it must always remain doubtful whether the cause is in us or outside us. (A372)On the other hand, if we take “outer” (or “outside us”) as meaning simply “in space,” then the argument falls. Specifically, the second premise is false, because outer objects can be perceived directly. Since space is nothing but presentation, “All outer perception, therefore, directly proves something actual in space—or is, rather, the actual itself” (A375). This last statement is strongly reminiscent of the Berkeley of Principles of Human Knowledge and Three Dialogues, who asserted that the being of bodies is to be perceived. Kant granted that, “It is a proposition that must indeed sound strange: that a thing can exist only in the presentation of it” (A375).
Although the proposition that a thing can exist only in its presentation sounds strange, Kant thought that he could make it plausible by noting that the “thing” in this case is only an appearance, and an appearance is a presentation. This is the doctrine of transcendental idealism as stated in the first edition. It is consistent with the doctrine of the existence of bodies, which is empirical realism. On the other hand, the empirical idealist is a transcendental realist. That is, the empirical idealist regards bodies as things in themselves and as a result either dogmatically denies that they are possible or skeptically denies that their existence can be proved.
In the Prolegomena, Kant reiterated the position of the first edition of the Critique.
I grant by all means that there are bodies without us, that is, things which, though quite unknown to us as to what they are in themselves, we yet know by the representations which their influence on our sensibility procures us. These presentations we call “bodies,” a term signifying merely the appearance of the thing which is unknown to us, but not therefore less actual. (§13, Ak 4:289)Here, Kant seems to be tying the reality of bodies to the reality of things in themselves, inasmuch as bodies are the appearances of actual things.
Kant seems to rely on an argument with the same structure as the argument for empirical idealism, which posits an “influence” as the basis of our knowledge of the reality of things which are otherwise unknown. In the Fourth Paralogism he advances a similar argument: “Now we may, indeed, grant that something that may in the transcendental meaning of the term be outside us is the cause of our outer intuitions” (A372). This he calls the “transcendental object.” But the first premise of the argument for empirical idealism was that what can only be inferred as a cause “has a merely doubtful existence.”
One way out of this apparent difficulty is to claim that the premise does not imply that it is doubtful that there is a cause, but only doubtful that the cause is of a certain kind. This is indicated by the fact that Kant took the problem as lying in the fact that we cannot tell from the presentation what kind of cause exists. Thus he could say that there is a transcendental object, but he could not say what kind of object it is.
There is still a large issue looming over whether Kant is even entitled to the claim that there is a transcendental object that is the cause of outer intuition. This is because the category of causality is supposed to be meaningful for humans only through the rules which connect objects in time. As the transcendental object would lie outside time, there seems to be no basis for any claim that it is in any way a cause of outer intuition.
Kant seems to have thought that part of the problem in understanding his new kind of idealism lay in his use of the term “transcendental” to describe it. So in the Prolegomena, he suggests the substitute term “critical.” This is just a distraction. As far as he was concerned, the real issue was whether his idealism was “genuine,” and he simply defined “genuine” idealism (which is dogmatic) as that which treats bodies as illusion because they are not objects of the pure understanding.
So he tried to squeeze between skeptical and dogmatic idealism. He can deny skeptical idealism by allowing that bodies exist, though only as appearances, which are presentations. Then he can deny dogmatic idealism by claiming that presentations are not illusion, but instead are appearances of something real.
In the second edition of the Critique, Kant abandoned his earlier attack on empirical idealism in favor of a new one, which is modeled on the form of argument used in the Analogies. This is the Refutation of Idealism. The key difference between the old and the new arguments is that the new one does not depend on transcendental idealism, though the doctrine of transcendental idealism itself is not abandoned.
Some critics have held that Kant was in fact running away from transcendental idealism by 1787. Arthur Schopenhauer, born a year after the publication of the second edition, attributed its difference from the first to “a visible fear of decided idealism” (The World as Will and Representation, Appendix, “Criticism of the Kantian Philosophy”). Schopenhauer held that no one who has read only the second or subsequent editions can appreciate the unity of Kant’s thought, because he would have “read only a mutilated, spoilt, and, to a certain extent ungenuine text.” Schopenhauer believed that transcendental idealism is fundamentally correct. Most philosophers believe the opposite, and so they welcome any attempt to prove the existence of the external world that does not concede the need for some kind of idealism, even if it is not material idealism.
The argument of the refutation is required to show that our presentations give us experience of objects rather than merely being the play of imagination. Kant grants with Descartes that inner experience is indubitable. What is to be shown is that inner experience is possible only if outer experience is presupposed.
As this is supposed to be a theorem, it is accompanied by a proof, as was provided for the relational categorial principles.TheoremThe mere, but empirically determined consciousness of my own existence proves the existence of objects in space outside me. (B275).
The argument will be reconstructed using the modification Kant proposed in the footnote to the Preface, Bxxxix.
The I which is determined in time is not to be identified with original apperception, the consciousness of the “I” as such. This self-consciousness is wholly undetermined and is not a cognition of the self. We cognize ourselves, that is, present ourselves, as a determined object, only through inner intuition. Inner intuition is subject to the form of time, but time itself does not present the order of our intuited inner states.
Kant claimed that this argument establishes that we perceive outer objects directly and our own inner states only indirectly. In this way, it turns the tables on the argument for empirical idealism, according to which knowledge of outer objects is indirect and dependent on an inference from direct inner intuition. Kant here recites the same argument as in the Fourth Paralogism, that “we infer [external things] only unreliably, as happens whenever we infer determinate causes from given effects, because the cause of the presentation that we ascribe—perhaps falsely—to external things may also reside in ourselves” (B276).
So has Kant redeemed philosophy and reason in general from the scandal in which it found itself? Unfortunately, it seems that he has not. The main problem with this argument is the same as with the arguments for the Analogies. Even if we grant that determining the order of past perceptions requires reference to permanent objects, it does not seem that determining the order of present perceptions does so. Something must be given, something must be the matter of inner intuitions. This will be a factor in the order of our presentations of ourselves. Kant has given no reason to deny that the perceived order of inner intuition is not sufficient for us to determine their order.
Kant might respond by claiming that without determination by reference to the perception of a permanent, we would have no experience of ourselves. He would then be trading on his special sense of “experience” as connected perception. But as with the Analogies, one may just be skeptical here. Perhaps there is no “experience” in the sense Kant has laid down. Perhaps I am not conscious of myself as determined in time, but rather am only conscious of an apparent sequence of my mental states. Hume might weigh in here, by asking how it is that present perception could have any connection at all with past perception, since it can only present the past through the present perception (A Treatise of Human Nature, Book I, Part IV, Section 6, “Of Personal Identity”).