The thesis of Kant's third antinomy is based on the fact that, if the antithesis was true (i.e. there is no causality through freedom and thus only causality by natural laws) then, for any given effect, we must assume an infinite series of causes and effects (P1). Yet, natural law is precisely the fact that "nothing happens without a sufficiently determined cause a piori" (P2).

From here, Kant concludes that (P1) and (P2) are contradictory. This is precisely the part I'm struggling with: why isn't an infinite series of causes (and effects) a sufficiently determined cause a priori?

From my understanding, if I consider the general example of an object falling, I would consider such a series of causes to the point of a given natural law (which is gravity) for which I cannot give any cause (meaning that I can only observe its necessity, as a natural law). Would it be what Kant means then? That for any given effect, we would always have to regress back to a sufficiently determined law of nature a priori which, in itself, could only be considered as a spontaneous cause (and thus, free)?

  • The introduction of a 'spontaneous cause' does not solve Kant's objection to an infinite chain of causes. Beause it prompts at once the question for the cause of the spontaneous cause. The latter does not exist by definition. Hence the proposed explanation ends with an indetermined first event. - The concept of an 'spontaneous cause' seems to me a contradiction in adiecto.
    – Jo Wehler
    Commented Apr 28, 2022 at 7:34
  • @JoWehler Spontaneous means by definition self-sufficient (literally: out of itself), thus a spontaneous cause is the only way to have a non-infinite causal chain (strictly speaking within Kantian framework here, not defending the argument as such).
    – Philip Klöcking
    Commented Apr 28, 2022 at 12:48
  • @PhilipKlöcking I agree that ‚spontaneous cause‘ means self-sufficient cause. I would equate it with ‚causa sui‘. I uphold my objection that it is a contradictio in adiecto. The term attemps to save causality by introducing non-causal causality, named causality from freedom (Kausalität aus Freiheit, transzendentale Freiheit).
    – Jo Wehler
    Commented Apr 28, 2022 at 14:32
  • Indeed, it is not causality that has to be saved since it is an empirical fact. For Kant, it is a place for freedom that needs saving
    – Philip Klöcking
    Commented Apr 28, 2022 at 15:22
  • @Philip Klöcking OK. Then let’s turn the tables. The problem of explaining our basic feeling of free will cannot be solved by postulating ‚non-causal causality‘. Begging the question is no substitute for explanation. Hence I cannot recognize Kant's contribution to solve the problem.
    – Jo Wehler
    Commented Apr 28, 2022 at 17:12

1 Answer 1


To answer this, we need to be aware of several points:

These arguments are not Kant's own

For all antinomies, Kant paraphrases historical arguments that have been made on the respective problems. These are not Kant's own arguments. Rather, Kant does not "solve" but rather "dissolve" the antinomies by offering a third way which makes the theses and antitheses only superficially contradictory.

This particular argument is Aristotelian

The argument against an infinite regress of causes is from his Metaphysics and is only paraphrased (and slightly modified, as all arguments in the antinomies...) here. Thus, the question should not be directed at Kant, but at Aristotle.

How can the argument state it that way?

The sufficiency of determination is an explicatory one: Only if a cause can completely explain the current state of affairs, it is sufficient. But if every cause in itself has a cause, then it is literally not self-sufficient, ie. it needs another cause to explain how it came to be etc.

Thus, assuming an infinite regress of causes, even the whole chain does never sufficiently explain the outcome as we will never reach an ultimate cause that itself is sufficiently explicatory. The explanatory force gets handed down infinitely, as it were.

What does Kant himself offer?

Kant, like discussed shortly in this answer of mine, with a pointer to where Kant writes about that does distinguish between infinite and indefinite. For any causal chain of an empirical event, it would be wrong to say there is an infinite chain of causes since the event and the causal chain are mere representations of reality and thus our picture of it. According to Kant, we always have to keep in mind that we cannot easily apply what and how we think about the world to metaphysical truth. Indeed, the whole book is about how we must end this in order to finally provide a solid fundament for any proper philosophy. Accordingly, he says that all we can say about any empirical causal chain is that no matter how far we try to go back, we so far have ever found another cause behind the last one. Thus, we can say there is an indefinitely long causal chain here.

The argument about the explanatory insufficiency does apply all the same though.

  • Thank you for this historical perspective and detailled answer, it was what I was looking for.
    – kronenbouh
    Commented Apr 28, 2022 at 13:35

You must log in to answer this question.

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