Kant’s Question: How is synthetic a priori knowledge possible?
(Note: synthetic statements are statements whose negations are not logically
contradictory. Synthetic statements are not “true by definition,” but purport to say
something substantive (non-trivial) about how the world is.
We have a priori knowledge of a statement when we do not need to consult sense
experience to justify our belief that the statement is true. We have a posteriori knowledge
of a statement if we do need to consult sense experience to justify our belief that the
statement is true.
According to Kant, one thing that Hume has correctly shown is that we do not have a
posteriori knowledge of any claim that involves either necessity or universality, e.g., that
when billiard ball B hits billiard ball A, A necessarily moves; or that all events have
causes (universality). Kant agrees with Hume that sense experience can only show us
that something did happen (not that it happened necessarily) and that experience can
never justify any claim purported to be universally true, since we have not experienced
everything. So, Kant thinks that any claim that purports to be necessarily true or
universally true must be known a priori, if known at all.
Analytic statements (if true) are necessarily true, so they must be known a priori. The
disagreement between Kant and Hume concerns whether or not we have a priori
knowledge of any synthetic statements. Hume did not see how this could be possible.
Kant’s goal is to explain how it could be possible.
Finally, metaphysical knowledge, -if we have any-, would be synthetic a priori
knowledge—non-trivial knowledge about reality that can be justified without appeal to
sense experience. So, if Kant can show how synthetic a priori knowledge is possible, he
will have shown how metaphysical knowledge is possible. But not all synthetic a priori
knowledge is metaphysical. The claims of arithmetic and geometry are synthetic a priori,
but not metaphysical.)
Kant’s answer: Synthetic a priori knowledge is possible because all knowledge is only of
appearances (which must conform to our modes of experience) and not of independently
real things in themselves (which are independent of our modes of experience). This
claim, that we know only appearances and not things in themselves, is known as Kant’s
transcendental idealism. So Kant’s claim is that if in experience we knew things as they
were in themselves, then Hume would be correct, and there could be no synthetic a priori
knowledge. But since we know instead only appearances (i.e., how things appear to us),
and since how things appear to us must conform to our subjective modes of experience,
we can know in advance of experience that these experiences will (necessarily and
universally) conform to the only ways in which we could possibly know or experience
objects. So, synthetic a priori knowledge is possible, but only because certain aspects of
our experience of objects reflects something that we (i.e., our mind’s) contribute to that
