By Andrew Brook
I do not disagree with Corey Dyck about most of the things he says, indeed I am not even competent to comment on a lot of the history of philosophy that he does. So I will focus on passages in Kant about which he could, perhaps, have said more than he did. Given the time constraint, I will have to pick my spots. Dyck’s book is about rational psychology. Kant’s critiques of the Second and Third Paralogisms being the most important parts of his treatment of rational psychology, in my view, I will spend most of my time on Dyck’s two chapters on these critiques. As I said, I am going to focus on things Dyck could have said but didn’t, though I do disagree with him about one aspect of the Third Paralogism, as we will see in due course.
There are at least three ways to study a philosopher from an earlier time. One can mine the person’s work for interesting ideas, interpret the work as a whole, and relate the work to the intellectual history of the person’s time. These are not exclusive categories: from one side, all miners will have to interpret, indeed do a lot of interpretation, and, from the other, so will all historians. And of course a few among us have the energy and luck to do all three.
That said, the distinctions are still useful. A miner will search for “Oh, that is interesting!” passages and will tend to be on the lookout for connections to later philosophers whereas an historian will search for passages to which she reacts, “Oh, that must have been influenced by …!” and so will tend to look for connections to earlier work.
Dyck’s work and mine exhibit exactly this contrast. It often results in us emphasizing quite different things in key passages. Dyck is an historian and a very good one. I am mostly a miner, though I have made a few forays into interpretation for its own sake, pure exegesis. And we do indeed emphasize different things. Whereas I think the analysis of reference to self on A355 in the A-edition critique of the Second Paralogism is at the very centre of Kant’s critique, for example, Dyck hardly discusses it.
Each approach comes with attendant risks. If miners and interpreters do not know the history, they can miss or misunderstand why and how a topic was important to a philosopher. Historians, by contrast, can fail to take into account internal requirements and ways in which a work went beyond what had been done previously. In the case of the Paralogisms chapter, there were three important internal requirements:
- To protect the doctrine that we can know nothing about anything as it is in itself from Rationalist claims that that is not true of the mind.
- To beat back temptations to over-interpret the functional claims about the mind of the Deduction as having implications for the structure of the mind. (The functional claims by themselves already push requirement 1. fairly hard.)
- To “deny knowledge so as to make room for faith” (Bxxx). Kant seems to have thought that if argument and evidence can touch the questions of “God, freedom, and immortality” at all, argument and evidence will tend to show that we have no reason to believe in any of them. So the only way to keep the three available to us is to insulate them from all argument and evidence so that we are free to accept them on faith.
As to ways in which Kant’s chapter on the Paralogisms goes beyond previous work, let me just say for now that I find breakthroughs of astonishing originality in it. This I will try to demonstrate by example as we go. A study focused on the history of a philosopher does not have to consider whether or how the work goes beyond its history, of course—you can’t do everything in one book, but the question is an important one.
Before I end these introductory remarks, I want to say a word about Dyck’s historical work. To the extent that I can assess it, it seems to me to be splendid. In particular, his work on Wolff and other figures immediately prior to Kant was eye-opening. I now see Wolff and his contemporaries in a completely different way. I am not competent to examine this aspect of Dyck’s work and in any case my co-symposiasts will do so, but I think that most people reading Dyck’s account would have a reaction similar to mine. In short, Dyck’s book will make a difference.
The Second Paralogism is about why certain arguments that the soul is simple, i.e., is not a composite of parts, do not work. (As Dyck sees, there is
more than one such argument, not just the one that Kant called the Achilles of rational psychology.) The subject-matter of the critique of the Third Paralogism is not as clear but it is at least about why certain arguments for the soul’s persistence as itself over time—and ultimately beyond the death of the body—do not work. Dyck makes a case that it is also about other aspects of what is distinctive to personhood, a claim that I will not examine. The Second Paralogism is about synchronic singularity, specifically the claim that the soul is simple, and the Third Paralogism is about diachronic singularity, specifically the claim that the soul persists in a special, what Butler called a “strict and philosophical”, way. We will turn to them in a moment. But I want to start with two matters preliminary to both of them, topics that Dyck discusses in Chapter 3 of his book.
To what extent can we understand TA?
I agree with Dyck, in CPR Kant sharply distinguishes awareness of self in transcendental apperception (TA) from awareness of self in inner sense (p. 72). Indeed, probably no contrast is more important to him. I also agree with Dyck that Kant makes a good case for saying that TA is non-empirical. In particular, to use an idea of Jonathan Bennett’s, a judgment by TA does not divide experience into parts, the part to which it applies and the rest. Consciousness of self via TA is available in all experience, and in exactly the same way. If so, TA is empty of content. That said, TA is a form of consciousness of self. We agree on this, too (“I think […] constitute[s] the basis of reason’s representation of the […] I or soul [p. 73]). In my view, that TA is a form of consciousness of self is a clue to what the arguments that Kant goes after will look like.
To what extent can we understand what TA is like? In his well-known circularity argument (346/B404), Kant argues that, contrary to what rational psychologists thought, not much can be said about the consciousness of self that apperception yields:
Through this I or he or it (the thing) which thinks, nothing further is represented than a transcendental subject of the thoughts = X. It is known only through the thoughts which are its predicates, and of it, apart from them, we cannot have any concept whatsoever, but can only revolve in a perpetual circle, since any judgment upon it has always already made use of its representation. (A346/B404)
Dyck endorses Kant’s claim here (p. 79) and so do I. That said, this seems to me to be a place where Kant is on to something new and original. If so, more could have been said. In my view, the passage anticipates the idea of the essential indexical articulated by John Perry some 200 years later.
The last clause is the key: “any judgment upon it has always already made use of its representation.” Kant seems to be saying that in order to study my representations of myself or my references to myself, I must know that the representation or act of reference is mine. That is to say, I must self-ascribe in order to study self-ascription. This is very much like the claim that use of an indexical is essential. Well, half the claim, anyway—that reference to self of some kind is essential. If so, there is more to Kant’s argument than Dyck suggests. Dyck suggests that Kant relies on the fact that the soul is not available to itself in TA (p. 80) but Kant’s point is more radical than that. Because in using TA to understand what the self is, one must refer to oneself, such a reference is part of anything that could be used to unpack what a self is—and we can only ‘revolve in a perpetual circle’, just as Kant said. (I think that Kant also saw that reference to self is indexical, but we will examine that question later.)
Kant says that the ideas that the soul is simple and persists in a special way arise from “an illusion that cannot be avoided, though it may […] be rendered harmless” (A341/B400). Kant clearly considered this claim to be important. Even though he rewrote the bulk of the chapter on the Paralogisms for the second edition, he kept this passage and some other important introductory remarks, including the one that we just considered.
That said, in and around the passage just quoted, Kant gives us little to work with when we try to figure out what the illusion is and how it arises—and he hardly takes up the topic again in A until A402! Nonetheless, what the illusion yields is pretty clear. In the case of the two Paralogisms that we are considering, it is that the soul or self is simple (not a composite) and that it has ‘personality’, in particular persisting in some strong way throughout life and indeed beyond the end of bodily life. From this we should be able to infer some pretty good candidates for what, in Kant’s view, the illusion/s consist/s in.
So, what gives rise to the illusion/s? What Kant says about this is rather general. In A (in the passage just mentioned on A402), he says: “Nothing is more natural and seductive than the illusion from taking the unity in the synthesis of thoughts for a perceived unity in the subject of these thoughts” (A402). He then tells us that the mistake here is a subreption, which is the error of mistaking a contribution of the understanding for a deliverance of the senses (or vice versa but the opposite mistake is not apposite here). In B he says: “The unity of consciousness, which underlies the categories, is here mistaken for an intuition of the subject as object […]” (B421).
Dyck’s view of what gives rise to the illusion follows Kant’s lead. He writes: “Kant is […] quite clear that the [illusion about] the soul consists in its apparent givenness to inner experience” (p. 88). The idea is this: One notices that one’s experiences, including experiences of oneself, are unified, are one “general experience” (A110) of many things all together. One then mistakes this “unity in the synthesis of thoughts” for a “perceived unity in the subject” (A402; introspected unity might have been a better wording). One misinterprets a general feature of experience and then takes it to be a property of the subject of experience.
In my view, this reading is fine as far as it goes. However, it does not go far enough. Why does perceiving or introspecting ourselves “unavoidably” lead us to believe that we are simple at a time and identical in some special way over time? Well, the arguments for simplicity and persistence that Kant attacks were for him, presumably, products of the illusion/s, so by examining these arguments, we should be able to get some clues as to what the illusion/s are.
As Dyck brings out nicely (pp. 105, 137), Kant examines two kinds of argument, inferences from the unity of consciousness and inferences from how we appear to ourselves in acts of TA. The two are quite different: the first makes no appeal to experience, the second does (though by appealing to a feature of experience that is necessarily part of all experience, its unity, the appeal is not empirical—see the Section immediately above). Kant goes after an argument—or more than one argument—of each type in connection with each Paralogism.
In connection with the Second Paralogism, the argument of the first type is the famous Achilles argument: thinking is not something that could be done by a system made up of parts because if one distributed the words, say, of a verse, one word to each component, nothing would be aware of the whole verse. Here is Kant:
[…] suppose it be the composite that thinks: then every part of it would be a part of the thought, and only all of them together would constitute the whole thought. But this cannot be consistently maintained. For because representations (say, the single words of a verse), distributed among different things, never make up a whole thought (a verse), the thought can never inhere in something composite as composite. (A352)
There are at least two arguments of the second type. What exactly they are is hard to pin down exactly, but the first is that one cannot think of oneself except as one (A107; cf. A116, A122, and B132), so one cannot think of oneself as a collection of components, and another is that as one appears to oneself, no parts appear. The conclusion in both cases is that one is simple.
What the arguments are in connection with the Third Paralogism has to be a matter of speculation because Kant’s account is extremely compressed in the A-edition and even more so in B. That said, the argument of the first type appears to be that because I must unify earlier with later experiences to think and can unify with my current experiences only earlier experiences that were mine, I must have existed over the period of time for which I have (remembered) experiences that I can unite with current ones. The argument of the second type appears to be that when I recall earlier experiences as had, earlier feelings as felt, earlier actions as done, I recall myself as their subject and agent, so I had to have existed for that period of time. (I will do more than I have done here to justify this reading of the arguments when we turn to Kant’s critique of the Third Paralogism.)
Now we can see what illusions Kant must have had in mind:
- In connection with the Second Paralogism: that being composite rules out being a thinking being; that if one cannot think of oneself as a plurality, one could not be a plurality; and, that if no parts are apparent in one’s awareness of oneself, one has no parts.
- In connection with the Third Paralogism and even more obviously: if I can unify with my current experiences only earlier experiences that were mine, I had to have existed throughout that period of time.
This account of transcendental illusion has a satisfying level of detail. It also connects the Paralogisms chapter very closely to the Deduction—a good thing, since the Paralogisms chapter was designed at least in part to block irresistible but ultimately fatal over-interpretations of the conclusions of the Deduction. With this account in place, we can turn to the critique of the Second Paralogism. We have done a lot of the necessary spadework, so we can proceed more quickly.
Critique of the Second Paralogism
The history that Dyck provides of the idea that the soul is simple (not a composite of parts) and immaterial is particularly penetrating and helpful. Kant explores the version where the implication runs from simplicity to immateriality but both Wolff and Reid (whom Dyck does not discuss) used the version in which the implication runs from immateriality to simplicity.
As we saw and as Dyck makes clear (pp. 105, 137), Kant presents rational psychology as using two kinds of argument, one consisting of inferences from the concept of unified experience, the other based on how we appear and must appear to ourselves when we refer to ourselves using ‘I’. (So saying, as Kant does, that ‘I think’ is the sole text of rational psychology (A354) is a bit over-simplified—Dyck discusses the claim on p. 137.)
The argument from unified consciousness is the one that Kant calls the Achilles of rational psychology, by which he means, presumably, that if it does not work, the whole enterprise is doomed. Let us call it A. We saw it earlier; it is a very simple argument:
… because representations (say, the single words of a verse), distributed among different things, never make up a whole thought (a verse), the thought can never inhere in something composite as composite. (A352)
Kant accepts that the consciousness of a verse unifies all the words of it into a single experience. He merely holds that this phenomenon does not entail that the soul is simple. Indeed, Kant dispatches A quite swiftly. Since Dyck did not quote Kant’s rebuttal, I will:
The unity of a thought, which consists of many representations, is collective and, as far as mere concepts can show, may relate just as well to the collective unity of different substances acting together […] as to the absolute unity [Kant means simplicity] of the subject. (A353)
I have no quarrel with anything that Dyck says about A.
The argument from how we appear and must appear to ourselves (let us call it B) raises a whole host of complicated issues about consciousness of and reference to self, a host too complicated to go into in any detail here. Dyck devotes most of the last section of Chapter 4 to them. Once again, I have no quarrel with what he says but think that more could have been said.
To show that B, how we appear and must appear to ourselves, does not entail that we are simple, Kant focuses mainly on the act of referring to oneself using ‘I’, the apperceptive act that provides unified consciousness of self. Here is the crucial passage:
In attaching ‘I’ to our thoughts, we designate the subject of inherence only transcendentally […] without noting in it any quality whatsoever—in fact, without knowing anything of it either directly or by inference. (A355; trans. Kemp Smith, which is here the best one)
He said similar things in other places. In such consciousness of self, “nothing manifold is given” (B135). In the kind of reference in which we gain this consciousness of self, we “denote” but do not describe ourselves (A382).
Kant then dispatches B as swiftly as he had dispatched A: In such consciousness of self, “the simplicity of the representation of the subject is not knowledge of the simplicity of the subject itself” (A355).
This group of statements, it seems to me, lays out ideas about reference to self that next saw the light of day only some two hundred years later in the work of philosophers such as Sydney Shoemaker and John Perry. As Shoemaker (1968) argues, one can be conscious of something as oneself without identifying it as oneself via properties that one has ascribed to the thing (self-reference without identification).
My use of the word ‘I’ as the subject of statements such as “‘I feel pain’ or ‘I see a canary’ […] is not due to my having identified as myself something [otherwise recognized] of which I know, or believe, or wish to say, that the predicate of my statement applies to it” (Shoemaker 1968:558).
And as Perry (1979) argues, first-person indexicals (I, me, my, mine) cannot be analyzed out in favour of anything else, anything description-like (the essential indexical). To know that I wrote a certain book a few years ago, for example, it is not enough to know that someone over six feet tall wrote that book, or that someone who teaches philosophy at a particular university wrote that book, or … or … or … , for I could know all these things without knowing that it was me who has these properties (and I could know that it was me who wrote that book and not know that any of these things are properties of me). Shoemaker summarizes the point this way: “[…] no matter how detailed a token-reflexive-free description of a person is, […] it cannot possibly entail that I am that person” (1968:560). Kant’s ideas about the consciousness of self that uses of ‘I’ yield
were, in my judgment, truly prescient (for more on them, see Brook 1994 and 2001).
If apperceptive consciousness of self is indexical and non-descriptive, that would give us a complete explanation of why, as Kant put it, “through the ‘I’, as simple representation, nothing manifold is given” (B135). And also of why this special way of appearing to oneself contains no implications for (nor against, of course) the soul being simple.
Transcendental designation was important to Kant for another reason, one that he managed to articulate clearly only in the B-edition. It allowed him to accept that reference to self is reference to self, not just to a highly doctored appearance of self, and yet make the key claim of the Paralogisms chapter that the consciousness of oneself resulting from reference to oneself using ‘I’ is not knowledge of oneself. As he put it in the second edition Deduction, “consciousness of self is […] very far from being a knowledge of the self” (B158).
Critique of the Third Paralogism
Although I find Dyck’s suggestion that Kant deals with personhood in general (what Dyck calls personality) in the critique of the Third Paralogism, not just personal persistence (what philosophers tend to call personal identity), plausible, for reasons of space I will restrict myself to the latter.
I have more problems with Dyck’s treatment of the critique of the Third Paralogism than I had with his treatment of the Second. Indeed, he may be guilty of an over-interpretation of just the kind that Kant was attempting to guard against. Dyck suggests that in the critique of the Third Paralogism, Kant claims that we do know something about the referent of I think. It is unlikely, I think, that Kant made any such claim. The results of the chapter on the Paralogisms are meant to be purely negative, purely critical. Kant’s aim is to deny knowledge of the soul, in order, as we saw, to make room for faith (Bxxx). To pinpoint the disagreement, let me quote a fairly lengthy passage from Chapter 5 of Dyck’s book:
[T]he soul, as the I of the I think, is necessarily represented as numerically identical. This is because the possibility of the cognition of objects requires the capacity to combine my representations, which itself presupposes that all my representations belong to a numerically identical consciousness. That the consciousness of this numerical identity of apperception also amounts to a consciousness of the numerical identity of myself in different times is evident upon a consideration of the relation between time, considered as the form of inner sense, and the unity of apperception:
But now I am an object of inner sense and all time is merely the form of inner sense. Consequently, I relate each and every one of my successive determinations to the numerically identical self in all time, i.e., in the form of the inner intuition of myself. (A362)
Time is a form of intuition […] and as such its unity must be taken to originate in the unity of apperception. This implies that all determinate times are necessarily related to the numerically identical self, and so that the self of whose identity I am conscious is necessarily identical throughout all the various times of which I am possibly conscious (pp. 162–3; Dyck’s emphases).
Dyck is correct in saying that Kant regarded time—all time, earlier and later as well as present, as properties of myself. He is also correct that, as I would put it, any current experience of which I am aware directly and non-inferentially or from the point of view of having it, will be one of my experiences. It will be part of current unified consciousness. And Dyck exposes a tension between all time, including all earlier time, being ‘in me’ and some memories possibly being of experiences that someone else had (A363), indeed exposes this tension in a way that I have not seen before.
All that said, even if it is true that “all determinate times are necessarily related to the numerically identical self”, would it follow that “the self of whose identity I am conscious is necessarily identical throughout all the various times of which I am possibly conscious”? I think not. My only sensuous access to earlier times is in memories. Even if the earlier times that I remember are in some way properties of me, so that the time when the moon was new last month is a property of me (in some nontemporal way), it does not follow that the content of that memory, in this case the moon appearing as a semi-circular sliver, is a property of me. And if Kant would have said that the earlier moon is a property of me, somehow (see the critique of the Fourth Paralogism for at least suggestions along these lines), the distinction between the form of time and the content of experience is even clearer with respect to access to myself in memories.
As I said, memory is my only sensuous access to the past. This is what “consciousness of the numerical identity of myself in different times” consists in. When I remember having had an earlier experience, having done an earlier act, it will indeed appear to me to be me who had the experience or did the action. That the earlier subject or agent was me is something that I will “automatically assume”, as Parfit once put it (1971:15). As a result, it appears that “I can trace my identity, quite independent of the identity of my body” (Wittgenstein 1968:264–5). Instead of ‘automatically assume’, Kant said: “In my own consciousness, identity of person is unfailingly met with” (A362). This creates the central illusion of the Third Paralogism. The reason it is an illusion is that I will think that the earlier person who had the experience that I remember having had or did the action that I remember having done was me whether that person actually was me or not. ‘Unfailingly’ may appear to be an over-statement—the illusions of rational psychology are not impossible to resist—but, as Kant makes clear in the next paragraph, he means that the appearance of identity is unfailingly met with:
The identity of the consciousness of myself at different times is […] only a formal condition of my thoughts and their coherence, and in no way proves the numerical identity of my subject. Despite the logical identity of the ‘I’, such a change may have occurred as does not allow the continuation of identity, and yet we may ascribe to it the same-sounding ‘I’. (A363; my emphasis and, in part, translation)
Same-sounding ‘I’. Kant and Parfit appear to be saying pretty much the same thing. They also draw the same inference: Even if I will automatically assume that it was me who had an earlier experience that I now remember having had, it does not follow that it was me who had it. As Kant put it in the famous footnote to A363, even if the earlier experience had been had or the earlier action had been done by someone else, the person who is now remembering them would “be conscious of the states of the previous substances [subjects of experience] as being its own states [so long as] they […] have been transferred to it together with the consciousness of them (A364, my emphases). In the same way as we appear to ourselves to be simple (A354), we appear to ourselves to go back at least as far as the earliest experience that we remember having, or action that we remember doing. For Kant, such appearances are not to be trusted. In short, I see no positive claims at all in Kant’s critique of the Third Paralogism.
Moreover, and this is crucial for reconciling this sceptical argument about identity across time with the unity demands of the Deduction, someone else appearing in a memory of mine as though he or she were me would be just as good for synthesizing the earlier, now-remembered experience or action with current experience as the earlier person actually having been me. For me to synthesize with current experience, all that is needed is that I remember earlier experiences and actions as though they had been mine (A672/B700). It does not matter whether or not they actually had been mine.
A final point. What did Kant mean when he said that the appearance of identity of consciousness across time is a ‘formal condition’ of thinking? Perhaps this. In the same way that time is part of the form of my experiences, my appearing to be the single common subject of my experiences is part of the form of them. (Space, too, for representations of ‘outer’ objects.) For me to synthesize an experience with my current experiences, it must appear in the same way as my experiences appear to me—whether it actually was one of my experiences or not. Having a place in time and appearing to be mine are common to experiences not as common properties that experiences share (concepts) but as a single common form within which experiences are located. This is what Kant meant, I think, when he referred to consciousness as the “subjective form of all our concepts” (A361). We could add, ‘and experiences’.
In his treatment of the Third Paralogism, Dyck takes up interesting issues that I have not touched on so far. Kant says, for example, it is the “external observer who first considers me as in time; in apperception time is properly represented only in me” (A362). I would love to discuss this strange saying with Dyck. Unfortunately, I have run out of … time.
An earlier draft of this essay was read at an Author Meets Critics symposium at the Canadian Philosophical Association, on June 3rd, 2015.
 The first argument makes no essential reference to one’s experience of oneself, the second does. The second is the basis of the para-syllogism; the first dominates in the early paragraphs of the critique.
Brook, A. (1994) Kant and the Mind (Cambridge UP).
—— (2001) ‘Kant, Self-reference, and Self-awareness’, in A. Brook & R. DeVidi (eds), Self-reference and Self-awareness (Amsterdam: John Benjamins Publishing), pp. 9–30.
Parfit, D. (1971) ‘Personal Identity’, Philosophical Review 80: 3–27.
Perry, J. (1979) ‘The Problem of the Essential Indexical’, reprinted in Brook & DeVidi, pp. 143–62.
Shoemaker, S. (1968) ‘Self-Reference and Self-Awareness’, Journal of Philosophy 65 (19): 555–67.
Wittgenstein, L. (1968) ‘Notes for Lectures on “Private Experience” and “Sense-Data”‘, ed. by R. Rhees, Philosophical Review 77 (3): 275–320.
© Andrew Brook, 2015.
Andrew Brook is Chancellor’s Professor of Philosopher and Cognitive Science at Carleton University, Canada. Among many publications, he published Kant and the Mind (Cambridge UP, 1994). A book entitled A Unified Theory of Consciousness (co-authored with P. Raymont) is forthcoming from MIT Press.