Kant's Third Antinomy

Either Kant cut the Gordian knot vis-a-vis free will, or he violated non-contradiction.

No, that was not nonsense, but a preview to the upcoming discussion of the Third Antinomy in the Critique of Pure Reason. (If you wonder how many posts one can make about this book, the Hackett edition is 774 pages with a 38-page bibliography; blog posts are nothing!) It will take some time to explain exactly what the antinomies are, what place they serve in the Critical philosophy, and why the third in particular is so interesting.

The First Critique has two rather large sections dealing with very different sorts of judgments: the Transcendental Analytic, on the one hand, and the Transcendental Dialectic, on the other. The Analytic is distinguished by being merely a rigorous analysis of the proper function of the mind, without regard to the correspondence of its judgments to an independent reality:

Transcendental analytic consists in the dissection of our entire a priori cognition into the elements of understanding's pure cognition.

Because Kant, in the Transcendental Aesthetic, has already shown that the reality perceived by the senses is not entirely independent, but is the result of the filtration by the mind of pure data into spatially and temporally extended experience, it makes sense that Transcendental Analytic will continue the project of showing what preexisting conceptual arrangements in the mind allow that mind to shape its experience to conform to the instrument. Thus, although the Analytic does not deal with some mystical "real" reality, it deals precisely with reality as it must be for us, and thus with things that are (for all we can ever tell) as they are.

By contrast, the Transcendental Dialectic seeks to break free of cognition's earthbound limitation:

...this general logic, which is merely a canon for judging, has been used - like an organon, as it were - for the actual production of at least deceptive objective assertions, and thus has in fact been misused. Now general logic, when used as supposed organon, is called dialectic...Now we may note (as a sure and useful warning) that general logic, when regarded as an organon, is always a logic of illusion, i.e., it is always dialectical. For general logic teaches us nothing whatever about the content of cognition; it teaches us merely the formal conditions for the agreement [of cognition] with the understanding, and there conditions are wholly inconsequential otherwise, i.e., as regards the [cognition's] objects.

That may be rather confusing, so I will attempt to shed some more light on it, if at all possible. "General logic" is an exhaustive inquiry into the way cognition is able to take unorganized experience and make it into a coherent whole with parts interacting in regular ways. Whenever we think, we use logical relations to make sense of our thoughts, and without logic, nothing would be seen to be in any relation to any other thing, much less a law-governed relation. I think a reasonably enlightening analogy can be found in a computer program - say QBasic.* The program knows that, upon seeing the GOTO command, it needs to call up another line whenever the GOTO line is executed. In a variety of other ways, the program relates two lines to each other, by using a number created by one subroutine as the value for "x" in a formula in another line in another subroutine, and so on. The program has a warehouse of relations that it can use when called upon by the use of certain terms in lines of the program.

But the program will not do anything if no lines are written into it. Open QBasic, write nothing, and then run. Nothing happens. Similarly, cognition cannot occur without some external objects' being presented to it for it to manipulate and think about. Transcendental Dialectic is analogous to taking the commands of QBasic itself and trying to achieve a result, without writing any lines of code. Though space, time, and the categories (the conditions for cognition in the First Critique) contribute to experience, and indeed make it possible by importing organization to raw experience, they produce nothing without some data coming from somewhere to stimulate them to operate in the first place. A mind without external objects is entirely contentless.**

It seems like the proper goal of the Transcendental Dialectic would be, by exposing the illusion of the improper use of reason, to end such unfounded speculation permanently. Kant claims that this is not possible:

But that the illusion should even vanish as well (as does logical illusion) and cease to be an illusion - this the transcendental dialectic can never accomplish. For here we are dealing with a natural and unavoidable illusion that itself rests on subjective principles and foists them on us as objective ones, whereas a logical dialectic in resolving fallacious inferences deals only with a mistake in the compliance with principles, or with an artificial illusion created in imitating such inferences. Hence there is a natural and unavoidable dialectic of pure reason.

The mind is compelled to consider metaphysical questions, and because the mind has only the organizing principles it has to deal with reality, it must consider metaphysical questions in those terms. But things cannot actually be in themselves what cognition thinks them as, because cognition alters its objects to conform to itself. Metaphysics attempts to perceive things as they are in themselves. Kant views metaphysics as a natural, unavoidable urge of all rational beings, and one with some goal, at least; metaphysics can give ideal models of how things work, without actually bringing concrete, logically sound knowledge of those things to us. The model is a regulative, not a constitutive concept.

So much for the purpose of the Dialectic. Now a bit about the structure of it - the Dialectic is divided into three parts: the Paralogisms of Pure Reason, the Antinomy of Pure Reason, and the Ideal of Pure Reason. The Paralogisms are false inferences about the self that rationalist metaphysics (with Descartes as the patriarch) made. The Antimony is a list of four pairs of contradictory conclusions about the world as a whole that reason can provide equal proof for. The Ideal is the idea of a necessary being (God), and proof that that being is necessary and real by use of reason itself. As you might imagine, Kant is devastating in this section.

The Antinomy consists, as I said, of four antinomies, that is, four pairs of contradictory assertions. Kant takes it upon himself to state these pairs, and then to defend each assertion, showing that both sides of each issue have substantial rational basis, so that one cannot prove, by pure reason, one or the other. The conclusion is that thinking about the world as a whole is impossible, because the world as a whole could never be an object of experience. So the principles of thought that apply to experience cannot apply to the whole of experience, considered as an object. The Third Antinomy is especially fun:

Thesis

The causality according to laws of nature is not the only causality, from which the appearances of the world can thus one and all be derived. In order to explain these appearances, it is necessary to assume also a causality through freedom.

And the antithesis:

Antithesis

There is no freedom, but everything in the world occurs solely according to laws of nature.

Following both thesis and antithesis are proofs, which the curious can find elsewhere, if they want. The reader is supposed to conclude, along with Kant, that neither free will nor determinism is a plausible theory about how things really occur. But in denying that either theory is true of everything, Kant affirms both theories for some (mutually exclusive) parts of reality:

Hence reason is the permanent condition of all the voluntary actions under which the human being appears. Each of these actions, even before it occurs, is predetermined in the human being's empirical character. But in regard to the intelligible character, of which the empirical character is only the sensible schema, no before or after holds, and every action - regardless of its time relation to other appearances - is the direct effect of the intelligible character of pure reason.

An example is helpful. Take an action - for instance, my closing the Critique of Pure Reason after typing that last blockquote. In one sense, that action is entirely conditioned on the operation of natural events in accordance with physical laws. I closed that book at that time because a chain of events, one causing the next, which next serves as the cause for another effect, had the eventual result that I applied certain force to a certain object in a certain way that can be described as "I closed that book." In this empirical aspect, my action was no different than the action of the wind I hear blowing outside, something that occurs according to laws of nature with no irregularity, no consciousness, and no spontaneity.

On the other hand, I am a rational thing, unlike the wind, and further, I have the freedom to choose what to do - or so we think. But how to reconcile this freedom with nature? Recall that empirical experience is subject to cognition and its conditions. Even the succession of events, of effect following cause, is only possible because our experience is organized temporally. So to say that one thing occurs later than another is to make a claim about experience, not about things-in-themselves. Considered as a free actor in myself, my choice to close the book was entirely free, and indeed could not be caused by any prior event, because in the world of things-in-themselves, time does not exist. Events in that world are not related temporally, and thus cannot be subject to the limits of causation, because causation and time are simply attached to raw experience by the mind, and do not attach to things considered as independent realities.

That has a curious consequence - each action a person takes is both entirely determined by nature and entirely free. When I closed the book, I could not have done anything else; but then, considered as a thing-in-itself, I could have done literally anything else. When we feel compelled by nature to do things against our will, we are in fact no less free than when we feel totally unencumbered in our choices. The person with a gun to his head is free to do anything he wants, and even if he does exactly what his assailant demands, his choice was entirely free. What this view does not address is our perceived interaction with the world of experience, of our capacity to make free choices that nevertheless have some operation in the physical world. If Kant's philosophy is right, and if a person can really adhere to it faithfully, then interaction is impossible and in fact irrational to consider. Something that is not physical cannot cause something physical to happen, because the two types of beings operate on different levels. Perhaps "levels" is not an accurate description, because there is not a division between classes of reality, but a division between aspects of a single reality. The human mind, considered as a thing-in-itself, is free, but the same thing, considered as object of experience, is entirely unfree. The ideal mind cannot cause empirical events because causation is applicable only to relations among empirical events, not to relations between things-in-themselves and empirical events, nor even to relations among things-in-themselves.

The Critique of Practical Reason attempts to give an account of human freedom and morality, but I wonder if Kant did not simply abandon the solution to the Third Antinomy in writing that second book. It is certainly hard to reconcile the solution with any coherent theory of ethics, because it cuts off ethics entirely from experience. Commentary would be appreciated, if you've made it this far.

*No, I'm not very good at computer programming. But I can write a mean QBasic program. Oh yes.

**Even the mind itself has to be presented as an object, and thus subject to cognition's conditions, to be an experience; therefore, even the mind cannot be experienced except as a sort of object of nature, and thus as an external object to the mind-in-itself.