By Scott Edgar
The first thing for me to say is how much I enjoyed Corey Dyck’s book, and how much I learned from it. Dyck spends a lot of time emphasizing that there are arguments Kant makes in the First Critique that we cannot understand properly unless we understand the positions of his rationalist predecessors. I, at least, need no convincing. And the wealth of historical detail Dyck gives about Wolff’s and his followers’ philosophical methods, and about their conceptions of the relation between reason and experience, is incredibly interesting in its own right, but also makes sense of passages of the First Critique that I have never understood.
A lot of the historical work Dyck does are surveys of parts of Wolff’s, Knutzen’s, Mendelssohn’s, Baumgarten’s and others’, including the younger Kant’s, doctrines of rational psychology. He mobilizes that historical work to a few different ends. But one of his aims is to set Kant’s arguments in the Paralogisms against their Wolffian background in order to highlight certain continuities between Kant’s mature position and the views of his rationalist predecessors.
So, to take one important example, Dyck emphasizes that the mature Kant agrees with Wolff and various of Wolff’s followers, including Kant himself in his younger days, about why rational psychology is rational. (And it turns out that lots of contemporary readers of the Paralogisms have understood this point entirely wrong.) For Wolff, the Wolffians, and the mature Kant, rational psychology as a discipline is not defined by having a method that is purely a priori. Rather, Wolff, both the immature and mature Kant
, and others, take rational psychology to be a discipline that begins with a concept of the soul that is empirical. On this conception, it is rational psychology because it uses reason to infer truths about the soul from that initial concept (p. 80).
I do not want to take issue with any of the specific points of continuity that Dyck identifies between Kant and his rationalist predecessors. But I do want to suggest that his account also reveals the hard limits of those continuities. I am going to suggest that, although Dyck does not draw attention to this fact, his interpretation of the Paralogisms makes Kant’s central premise in those arguments an essentially Humean commitment. This in itself is not an objection to Dyck’s interpretations of the Paralogisms—since I am inclined to think it is a virtue of an interpretation of Kant that it shows clearly how deeply his Humean commitments run.
However, seeing clearly the Humean commitment at work in Kant’s Paralogisms—on Dyck’s interpretation of them—does raise one potential worry. The worry is this. On Dyck’s account, Kant aims for his arguments in the Paralogisms to respond to the rational psychological doctrines of his rationalist predecessors. But—and here is the problem—Kant’s rationalist predecessors mostly reject the Humean commitment in question. As a consequence, Dyck’s interpretation of Kant’s arguments in the Paralogisms make those arguments look potentially inert against their rationalist targets.
Now before I say anything else, I need to head off a possible misunderstanding of what I am saying. I am not going to be arguing that Kant is committed to any significant Humeanism about the metaphysics of the self. So the Hume in 1.4.6 of the Treatise is not the Hume I am going to saddle (Dyck’s interpretation of) Kant with. Instead, I am going to argue that Kant’s arguments in the Paralogisms, on Dyck’s interpretation, turn out to depend on Humean views about the possibility of metaphysical knowledge in general.
So here is what I am going to say. First, I need draw out what, on Dyck’s interpretation, turns out to be the essential premise of Kant’s arguments in the Paralogisms, and I am going to raise the question of what justifies that premise for Kant. Then I am going to talk about the doctrine of transcendental illusion, in order to make the point that it cannot justify the essential premise of the Paralogisms. Then I am going to sketch an argument that I think does justify the essential premise of the Paralogisms. And that is when we are going to see Kant’s Humean commitments.
1. How Kant, on Dyck’s interpretation, seems simply to assert without argument precisely what his rationalist predecessors deny
So first, in order to see the key premise of the Paralogisms on Dyck’s interpretation of them, we need to start with his account of the rational–psychological views Kant is targeting. Then I am also going to have to say something quick about Dyck’s account of Kant’s arguments in the Paralogisms.
Here is the point about rational psychology I need to emphasize. On Dyck’s interpretation of Kant’s rationalist predecessors, their rational psychology takes as its starting point a concept of the soul that has its origin in experience. That is, they take the object of rational psychology to be the soul conceived as the object we have intuitions of in inner sense. Then, on their conception of the task of rational psychology, we can use reason to infer various metaphysical truths about that soul’s substantiality, its simplicity, and its numerical identity over time, etc. But the whole project starts with a representation of the soul that derives from inner sense and so experience.
So, for example, as Dyck makes clear, Wolff thinks our consciousness of ourselves consists most fundamentally in a kind of inner experience. That consciousness is “indubitable”, and for familiar Cartesian reasons—namely, that any attempt to doubt that I am having an experience of myself (or even that I am having experiences of external things) already presupposes that I am having the experiences. But still, even if my consciousness of myself is “indubitable” for Wolff, that consciousness is nevertheless experience. In fact, it is Wolff’s experience of himself that provides him with the basis for his initial definition of the soul. Here it is: for Wolff, the soul just is “that thing which is conscious of itself and things outside of itself” (p. 29).
Then on Wolff’s conception, that concept of the soul—a concept he thinks has empirical content—constitutes the starting point for rational psychology. Rational psychology seeks to infer truths about the soul from that concept. In so doing, it seeks, whenever possible, to confirm truths about the soul that have been discovered in experience, but also to derive truths about the soul from that initial concept that are not available in experience. So, Wolff thinks, he can use that initial empirical concept of the soul as the starting point to derive claims about, for example, the soul’s substantiality and simplicity (p. 32).
As Dyck shows us, various German philosophers in Wolff’s wake—and not just those who self-consciously allied themselves with Wolff—had roughly this picture. In different ways, Baumgarten, Meier, Mendelssohn, Lambert, and Tetens all held that rational psychology aims to infer truths about the soul from a concept of the soul that gets its content from or by reference to experience. Even Kant in the 1770s has the Wolffian view that rational psychology begins by taking from experience the claim that we have a soul, and then seeks to infer various metaphysical truths about the soul from that initial, empirical representation (p. 66).
So now I want to pivot to Kant’s Paralogisms. The first thing to say is that, in Dyck’s discussion of each of the Paralogisms, he is perfectly clear that Kant sometimes has multiple targets and deploys multiple, distinct arguments against those different targets.
But at the same time, Dyck nevertheless sees one overarching argumentative pattern in the Paralogisms; and on his account, the arguments that fit that pattern are all directed at the Wolffian picture of rational psychology I have just been talking about. Kant’s overarching objection in the Paralogisms is this: his rationalist predecessors are wrong that rational psychology begins with a concept of the soul that has empirical content; they have mistaken the concept of the soul for a representation that has empirical content, when actually it is derived from the merely formal I think—that is, the mere form of thinking in general. But then, since the concept of the soul is derived from a representation that has no empirical content (and, in fact, no synthetic content), that concept cannot serve to ground any (theoretical) cognition of the claims the rationalists want to make in their rational psychologies.
So for example, on Dyck’s account, the First Paralogism takes aim at the rational psychologists’ claim to have found an empirical basis for cognition of the soul’s substantiality. As Dyck shows us, Kant had endorsed that claim in the 1770s. But in the first Paralogism—on Dyck’s account of it—Kant argues that the concept of the soul cannot ground the judgement that the soul is a substance, because it cannot ground the judgement that the soul persists through time. And it cannot ground that judgement, because Kant now maintains that the concept of the soul has its source in the I think—which, since it is purely formal, contains no representation of time and cannot be the basis of any cognition of persistence.
On Dyck’s account, the Second Paralogism has an analogous form. Here, Dyck distinguishes between two different arguments for the simplicity of the soul, both of which Kant targets. There is the so-called ‘Achilles argument’, in which the rational psychologist infers the simplicity of the soul from the singularity or unity of the thinking subject. But second, on Dyck’s account, Kant also targets an argument that we could discover the simplicity of the soul by attending to the singularity of our own activity of thinking—activity which, on this argument, is disclosed to us in inner sense. Kant now objects in the Second Paralogism that the singularity of the ‘I’ we are conscious of is the singularity of the ‘I’ in the I think. But Kant now maintains that the I think is merely the form of thinking in general, and as such has no intuitive content. So it follows for him that, when we are conscious of the singularity of the ‘I’ in the I think, we are not conscious of anything disclosed to us in intuition.
Finally, we see a similar form of argument in Dyck’s account of the Third Paralogism. Again, Dyck distinguishes between a couple of different targets of Kant’s arguments. But he argues that one of them is the Wolffian view of how the soul is conscious of (or at least, has the capacity to be conscious of) its numerical identity over time. So on this account, Kant agrees with the Wolffian rational psychologist that we are conscious of our soul’s identity over time; but Kant thinks the Wolffian has the wrong explanation of how we come to have that consciousness. On the view Kant targets, the soul’s consciousness of its identity at different times has an empirical basis, and so it grounds a cognition of that identity over time. But, as Dyck has the argument, Kant now insists that the soul’s consciousness of its identity over time is actually grounded in its consciousness of the logical identity of the ‘I’ in the I think. Once again, the I think is just the form of thinking in general, and has no empirical (and indeed, no synthetic) content. So, Kant argues against the rationalist, it cannot be the basis for any (theoretical) cognition of the soul’s identity.
So just to sum all of that up: on Dyck’s interpretation, the key premise of Kant’s arguments in the Paralogisms is his mature view that the concept of the soul is derived not from experience in inner sense, but from the I think—that is, from the mere form of thinking in general.
But now, notice what this interpretation does to the Paralogisms. On Dyck’s interpretation, the Paralogism’s key, essential premise is Kant’s claim that the concept of the soul has its source in the merely formal I think. But that is a premise that Kant’s rationalist predecessors would reject. And if they can reject that premise, it looks as though they would not have to accept Kant’s criticism of them in the Paralogisms.
So for Kant’s arguments to have any force against the rationalist positions he is targeting, he needs an independent argument for why—contrary to the claims of his rationalist targets—the concept of the soul has its source in the merely formal I think. In the absence of some argument for why the concept of the soul has to be derived from the purely formal I think, Kant’s arguments in the Paralogisms seems not to engage with the positions of his rationalist predecessors.
2. The doctrine of transcendental illusion
One might think that Kant’s doctrine of transcendental illusion can help here. In fact, Dyck does not suggest this, and we need to take a few minutes to consider why he is right not to.
Dyck’s identification of the work done by the doctrine of transcendental illusion in the Paralogisms is very illuminating. That doctrine is the one more familiar from the Antinomies. It is the doctrine that the human mind, in its attempts to reason about and discover metaphysical truths, is subject to an unavoidable temptation to make a certain kind of mistake. The mistake in question is the mistake of trying to use the pure concepts of the understanding beyond the sphere of their valid application, namely, beyond the limits of possible experience. On Kant’s picture, we face an inevitable temptation to take a representation that’s given to us merely as a problem or task—for example, the idea of an unconditioned condition, like, say, a first cause—and to mistake that representation for an object given in experience. We thus end up with the mistaken metaphysical view that we could have cognition of a first cause, when actually, there is no way for that concept to be applied within the boundaries of possible experience.
On Dyck’s illuminating account, this doctrine is not just at work in the Antinomies; it is also at work in the Paralogisms. Why does Dyck think the concept of the soul is subject to this inevitable transcendental illusion? It is because the I think is originally something connected in an important way to inner sense. After all, it is the I think that I can connect to any representation that is mine. We come to represent the I think as the mere form of thinking in general only when we can abstract it away from every thought it can be connected to in inner sense. But because the I think has that original connection to inner sense, Dyck argues that it always has the illusory appearance of having been given in inner sense (p. 89), rather than appearing to be what it actually is, namely, the form of the logical unity of any thinking in general.
But now notice that that mistake is exactly what is at work in Kant’s criticisms of his rationalist predecessors in the Paralogisms—on Dyck’s interpretation of them. In each of the Paralogisms, Dyck sees Kant objecting that the rational psychologist mistakes the representation of the soul for the concept of an object that is given in inner sense. But actually, the representation of the soul has its source in the merely formal I think. That mistake—the mistake that Kant is pointing out on Dyck’s account of the Paralogisms—is thus an instance of transcendental illusion.
Dyck’s account of the doctrine of transcendental illusion in the Paralogisms is very illuminating. If Dyck is right about it, then his account makes the Transcendental Dialectic more deeply coherent and deeply unified than it otherwise appears to be. Transcendental illusion turns out to be a thread that unifies the Paralogisms and the Antinomies in a previously underappreciated way. I’m not going to take issue with any of that.
My point is this. The doctrine of transcendental illusion cannot supply the independent argument Kant needs, on Dyck’s interpretation of him, for the claim that the concept of the soul has its source in the merely formal I think, and not in experience.
Why not? Kant’s doctrine of transcendental illusion is a diagnostic account of how its target views go wrong. So on Dyck’s account, the doctrine of transcendental illusion provides the explanation for why so many rationalist philosophers made the mistake of thinking that the concept of the soul has its source in inner sense, when really it has its source in the merely formal I think.
But the doctrine does not provide any argument or warrant for the claim that its target views are wrong. Rather, it proceeds from the presumption that those target views are wrong, and it seeks to provide the explanation for how the error got made.
So on Dyck’s account of the doctrine of transcendental illusion in the Paralogisms, it provides an account of why Kant’s rationalist predecessors made the mistake of thinking the concept of the soul has its source in experience. But it does not provide any warrant for the claim that it is a mistake to think the concept of the soul has its source in experience. And that is the argument we are looking for.
3. So what is the argument?
What, on Dyck’s account, is Kant’s argument for the claim that the concept of the soul has its origin in the merely formal I think, and not—as his rationalist predecessors thought—in inner sense? If I have understood Dyck correctly, the argument is this:
The first premise is about rational psychology’s aims. Namely, rational psychology does not aim to discover just any truths about the soul. It aims to discover truths about the soul that are universal. That was, for example, Wolff’s aim for rational psychology (p. 185).
The second premise is about the possible origins of necessary or universal truths—namely, that experience alone can never provide the warrant for necessary or universal truths. Because Kant thinks that, he thinks the rational psychologist can never derive necessary or universal truths from empirical concepts. So to the extent that a concept of the soul is derived from inner sense, the rational psychologist will not be able to derive necessary or universal truths from it (p. 185).
But then, it follows for Kant that rational psychology must properly be understood to take as its starting point a concept of the soul that is not derived from experience. For, if it were concerned with an empirical concept of the soul, it could not satisfy its aim to discover truths that are universal. So in effect, Kant’s argument is that his rationalist predecessors have, at the centre of their rational psychology, a concept of the soul that decisively undercuts their rational psychology’s ability to satisfy the aims that they themselves have for it.
On this view, if rational psychology is ever going to satisfy the aims that Wolff and other rationalists had for it—if it is ever going to discover any truths about the soul that are universal—then it cannot take as its starting point a concept of the soul that is derived from inner sense. Instead, its starting point will have to be a concept of the soul, the source of which is entirely a priori. And for Kant, a concept of the soul that is derived from the merely formal I think fits that bill.
So if I have got Dyck right, that gives us the argument we are looking for—Kant’s argument against his rationalist predecessors for why the concept of the soul cannot have its source in experience.
4. Hume and the limits of Kant’s rationalism
I started out with the way Dyck’s account emphasizes the continuities between Kant’s views in the Paralogisms and the rational psychology of his rationalist predecessors. But now I think we have bumped up against the limits of Kant’s rationalism in the Paralogisms. On Dyck’s interpretation, at least one strain of argument running through the Paralogisms depends on Kant’s claim that the concept of the soul has its origins in the merely formal I think, and not—as his rationalist predecessors thought—in inner sense. But if I have got Dyck right, that claim turns out to depend on an essentially Humean premise—namely, the premise that experience can never provide the warrant for universal truths. So whatever continuities there are between Kant and his rationalist predecessors on Dyck’s account, that is where the continuities end.
In case it needs to be said explicitly, it is clear that Wolffian rationalists did reject the Humean view that experience cannot provide the warrant for necessary or universal truths. Dyck’s account of Wolff includes a very interesting account of his epistemology, on which certain kinds of experiences can be “indubitable” and can provide the warrant for highly general philosophical claims (p. 22ff.). Dyck’s account of Meier likewise makes clear that, on Meier’s view, generalizations from experience can provide the warrant for truths that are fully universal. And indeed, there were other Wolffian rationalists in Kant’s own time—such as Mendelssohn and J.G. Sulzer—who explicitly attacked and rejected Hume’s argument that experience can never provide the warrant for necessary and universal truths. Sulzer did so in the commentary he wrote for his 1755 German edition of Hume’s first Enquiry—which was presumably the edition Kant himself read. And Mendelssohn sketched his anti-Humean arguments a year later, in his essay ‘Thoughts on Probability’.
So now we can see how Kant’s rationalist predecessors might have responded to his arguments in the Paralogisms—on Dyck’s interpretation of them. If Wolffian rationalists rejected the Humean view that experience never provides the warrant for universal truths, then they would not have had to accept the argument I have just sketched above for Kant’s claim that the concept of the soul has its source in the merely formal I think, rather than in inner sense. But then they would not have had to accept Kant’s arguments in the Paralogisms. So there is at least the possibility here that Kant’s arguments in the Paralogisms, on Dyck’s interpretation of them, have little or no force against their targets.
I want to conclude by suggesting why this point about a rationalist response to Kant’s Paralogisms does not necessarily amount to an objection to Dyck’s interpretation of Kant. First, Kant obviously is committed in his mature philosophy to the Humean view that experience cannot provide the warrant for necessary or universal truths. So it is no objection to an interpretation of Kant that it attributes this view to him.
But second—and here I am speaking for myself—I am inclined to agree with Kant that Hume is just right that experience can never provide the warrant for necessary or universal truths; and Wolff, Meier, Sulzer et al are just wrong to disagree. But then, anyone who agrees with Kant about that Humean commitment can find Kant’s arguments in the Paralogisms convincing against his rationalist predecessors, even if those rationalist predecessors would have disagreed.
An earlier draft of this essay was read at an Author Meets Critics symposium at the Canadian Philosophical Association, on June 3rd, 2015.
© Scott Edgar, 2015.
Scott Edgar obtained his Ph.D. in philosophy from the University of Pennsylvania in 2009 and is Assistant Professor of Philosophy at Saint Mary’s University, Halifax, Canada. His primary interests are modern philosophy, Kant, and neo-Kantianism. He published in, among others, the British Journal of the History of Philosophy, Canadian Journal of Philosophy, and Studies in History and Philosophy of Science.