We all know "the problem of affection" raised by Schulze:

"if causality is apriori structure of the mind and exist inside the mind, how can we claim thing-in-themselves cause phenomena when thing-in-themselves are outside the mind?"

How did Neo Kantians and Post Kantians like Hermann Cohen and Ernst Cassirer and Schopenhauer and Heidegger! responded to this issue?

I only know that Schopenhauer claim that (unlike kant) thing-in-themselves can be known to some extend, for example Schopenhauer says Will is the thing-in itself. And it seems like Cohen claims that our Sensibility isn't passive but active as well in creating the Phenomena, I haven't read much about him so i don't know whether he would reject the existence of Thing-in-themselves or like Schopenhauer would claim that it exist but it can be known to some extend.

But I don't understand how these are relevant to "the problem of affection".

Does anybody know their arguments? Is there any book or website that talks about these things?

  • Marburg Neo-Kantian responses are not interesting because they simply discarded the unknowable thing-in-itself, like Fichte and Hegel did earlier. IEP reviews more interesting responses that try to stay closer to Kant. They either suggest that the affectors are not things-in-themselves but "mind-independent appearances" (Bird, Prauss) or relax the unknowability thesis. Allison makes them "thinkable" with affection in some "logical" sense, Chignel and Watkins distinguish knowledge and cognition, etc.
    – Conifold
    Commented May 16 at 3:41
  • I wonder if this is related to recent questions of Identity and Uniqueness?
    – Scott Rowe
    Commented May 16 at 11:09

2 Answers 2


I realize that you are asking about the accounts of this matter from people besides Kant, but I would like to respond to the "problem of affection" from the point of view of Kant's own writing. The first thing is that according to the first Critique, the concept of causality is metaphysically interpolated from the if-then relation in pure general logic. Secondly, Kant's thing-talk is not fully split between "things as they appear to us" and "things in themselves" but also includes the phrase "things in general," e.g.:

  • 1. "In the mere conception of a thing in general this is really the case, but not in things as phenomena."

  • 2. "And thus these limitations prove that the representation of an object as a thing in general is not only insufficient, but, without sensuous determination and independently of empirical conditions, self-contradictory; that we must therefore make abstraction of all objects, as in logic, or, admitting them, must think them under conditions of sensuous intuition; that, consequently, the intelligible requires an altogether peculiar intuition, which we do not possess, and in the absence of which it is for us nothing; while, on the other hand phenomena cannot be objects in themselves. For, when I merely think things in general, the difference in their external relations cannot constitute a difference in the things themselves; on the contrary, the former presupposes the latter, and if the conception of one of two things is not internally different from that of the other, I am merely thinking the same thing in different relations."

(Those quotations are from the sort-of-infamous Meiklejohn translation, as available on Wikisource, but so I don't know how much difference other, "better" translations make to the interpretation and application of those passages...) Moreover, Kant sees fit to compartmentalize the first six [edited because I said "eight" originally] categories under the meta-category "mathematical" and the second six under "dynamical," which goes to show that he has a notion with causal overtones that is not quite reducible to the same concept of causation as the category proper is supposed to mark out.

Finally, despite being a noumenal domain par excellence, the moral sphere in Kant has for itself the "categories of freedom" (see also here) yet another interpolation of logical and metaphysical forms.

The "upshot" is that, "A, a thing-in-itself, causes phenomenon B," is not the same kind of statement, despite the wording, as, "A, a phenomenon, causes another phenomenon B." The meaning of "causes" in the first statement is little more than a (trivial!) conditional like, "If a being that is conscious in general is consciously related to a thing in general, then the thing in general appears to the conscious being," which is little more than an overly complicated way of saying that things that can appear can appear. So cause-talk can be mingled with thing-in-general talk a little bit, without there being a full-fledged application of the category of causality proper to things-in-themselves. Or then the little bit of mingling can relate even thing-in-itself talk to cause-talk, without transgressing the strictures of transcendental idealism.

Incidentally, I did find "Kant and double affection: solving the problem", which seems relevant to your actual question. "Kant and the problem of affection" looks like another recent response to the issue.

  • thanks for the reply, are you trying to say kant didn't make the claim that thing-in-themselves causes Phenomena? Commented May 15 at 18:18
  • @ParsaFakhar I would say that there are weaker and stronger uses of the word/concept "cause" (or whatever the German word for that is) in Kant's account, down to the non-committal use of the word "because" even to represent deductive inference. Moreover, even as a category, causality has a special noumenal counterpart (among the categories of freedom). On the other hand, Kant was part of the counter to scholasticism, incl. about substances, so he did not think a meager reference to conditional representation would have substantive implications for physics as such. Commented May 15 at 18:34
  • 1
    @ParsaFakhar "it can only mean one thing" how often do words generally ever mean only one thing, much less words as used by philosophers like Kant? Why are you asking for a "layman's terms" explanation of Kant? Commented May 16 at 0:54
  • 2
    I think what you are outlining is Allison's interpretation, but I do not think it can be unequivocally read into Kant's own writing. Kant is all over the place on this with parts of the text variably suggesting that "causing" is ascribed to "transcendental objects" rather than things in themselves, that sensibility is exceptional with respect to 'dynamical' categories like causing, that "unknowable" does not mean what it is conventionally taken to mean, etc. Indications are that his position is incomplete and he never confronted the costs any coherent completion would incur on his system.
    – Conifold
    Commented May 16 at 3:54
  • 1
    @ParsaFakhar The evening star and the morning star are the same thing, yet "appears in the evening" is only a determination of the former. This is called one world - two aspects interpretation of Kant and it is pretty popular. Appearances are things in themselves in extension, but not in intension (aspect). Their determination as appearances is manufactured by our cognitive apparatus and their inner nature is inaccessible to us. Just as the fact that the "star" is actually a planet was inaccessible to early humans.
    – Conifold
    Commented May 17 at 3:47

Re. "how can we claim thing-in-themselves cause phenomena"?

The thing-in-itself that causes a phenomenon is simply unknown beyond appearance, that's the problem. It's not that it doesn't exist (notionally), it's just that one doesn't know at all what is it in order to make a judgement on it, which is what would make it objectively exist in the Kantian sense. I.e. note 36 B287 :

"[existence] is the conjunction of the [notional] thing with perception."

Re. "Schopenhauer says Will is the thing-in itself"

The one thing-in-itself that one can know is oneself. However, previously the topic was objective existence. Now this is subjective existence, aka Being, so the goal posts are shifting unreasonably.

Heidegger discusses affection—that which causes an effect—in Kant and the Problem of Metaphysics. Since the "I" of cogito ergo sum is self-contained any source of effect must come from the self. That is to say, in Descartes' pared-back cogito, the assumption is that it is complete in itself. So all source of effects come from itself. In its own movement it creates time.

Time is, by nature, pure affection of itself. (Kantbuch, p. 194)

There is nothing 'outside' Descartes' ego cogito that it is affected by, as by definition.

Therefore, it is absolutely untrue that the mind exists in such a way that, among other beings, it relates certain things to itself and in so doing posits itself [Selbstsetzungen ausübt]. Rather, this line of orientation from the self toward . . . and back to [the self] first constitutes the mental character of the mind as a finite self.

It is at once obvious, therefore, that time as pure self-affection is not found "in the mind" "beside" pure apperception. On the contrary, as the basis of the possibility of selfhood, time is already included in pure apperception and first enables the mind to be what it is. (Kantbuch, p. 196-7)

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .