[hist-analytic] Defiending Kant Against Kripke

Baynesr at comcast.net Baynesr at comcast.net
Mon Nov 9 14:27:17 EST 2009

In a couple of posts I argued that Kripke has not 
refuted Kant on the matter of the, alleged, impossibility 
of the contingent a priori. Essentially, my argument at 
one point was this: from the truth of an identity statement, 
even if necessity follows, you can't know that the 
statement is necessary in virtue of experience, alone, and, 
so the truth or judgment cannot be necessary a posteriori; 
and this is because you need more than knowledge that the 
proposition is true in order to know it is necessary. 
Since you don't know it is necessary a posteriori, the truth 
is not a posteriori. Kripke in his closing remarks in NN 
p. 159 provides further reasons for thinking I'm right on 
this. Take a look. 

Kripke addresses my very concern, which I think (of course!) 
is much to his credit. He says, 

"Nor can Kant argue that experience can tell us that a 
mathematical proposition is *true*, but not that it is 

At this point the crucial question is: "Why can't Kant 
argue this way. Here is Kripke's answer: 

"...for the peculiar character of mathematical propositions 
...is that one knows (a priori) that they cannot be 
contingently true; a mathematical statement, if true, is 

But this is hardly reason to deny to Kant the option of 
arguing that whereas it may be the "peculiar" character of 
mathematical propositions" that they be known a priori, 
that we do know them a priori is precisely why (!) we cannot 
be said to know them a posteriori! In the very next 
line Kripke says that this is not a feature peculiar to 
mathematics; it holds for "all the cases of necessary a 
posteriori advocated in the text." If this true, then there 
is a problem, it seems. 

Suppose, then, that I know that the theory of types is 
consistent because my teacher told me so - this is Kripke's 

argument not mine, although I supply the example. Now I 
have some difficulty accepting this, but let's let it pass. 

Now the quesion is this: if I have empirical knowledge of 
the truth of some identity statement, 'p', how do I know 
that 'Nec(p)' is true? Kant will say I cannot know this 
by way of experience. Kripke quotes Kant, and he selects, 
thanks to Stroud, the perfect quotation. It goes: "Experience 
teaches us that a thing is so, but not that it cannot be 
otherwise." So DOES experience teach me that 'Nec(p)', which 
is just another way of saying p could not be otherwise? 
If Kripke is right, then since it is necessary a posteriori, 
it should be knowable a posteriori. However, by his own 
admission there is reason to doubt this. 

According to Kripke, and he actually says this (p. 159), 
such necessary identities are known to be necessary 
because "Philosophical analysis tells us that they cannot 
be contingently true..." But how can we equate this with 
knowing a posteriori?! Something has gone wrong, or so 
it seems. But wait! There's more. Kripke says that IF we 
accept this reliance on "philosophical analysis" then 
empirical knowledge of the truth of these propositions 
is "automatically empirical knowledge that they are necessary." 
But this is simply incorrect. The use of "automatic" is the 
tip off. 

What we have here is an inference to knowledge of the necessity 
of these truth from "philosophical analysis," NOT experience. 
Again, the identity may be necessary, and THAT it is necessary 
may follow from analysis or logic etc. But it does NOT follow 
from *experience* as an a posteriori truth. Ergo, Kripke is 
wrong. That is my argument. It should be noted that Kripke, 
unlike many of his epigones, is most definitely not doctrinaire. 
He has advanced these issue to new heights, particularly in the 
semantics of counterfactuals and the model theoretic treatment 
of the modalities; but he has not refuted Kant! One other point 
on Kant. 

We can't expect most folks, philosophers or otherwise, to be 
Kant scholars. The table of contents of the First Critique 
is an education of sorts in its own rights, but if anyone 
maintains that Kant has been refuted then a few well memorized 
misunderstood facts about Kant will not suffice. Here's the 
point. Kant view of the a priori differs from Leibniz's, but 
he isn't always clear where he moves from earlier to latter 
views (his) on the matter. Kant realized that his synthetic 
a priori judgments concerned knowledge of this, our human 
species. He knew or believed that other minded entities may not 
share the same set of Categories of the Pure Understanding etc. 
So in a very real sense, Kant knew all along that there is a 
contingent a priori, where "a priori" is laden with the notion 
of the "transcendental," something that supplies a good laugh 
particularly among those who haven't read Kant very closely, as 
a rule. So claiming a contingent a priori is no refutation of 


STeve Bayne 
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://rbjones.com/pipermail/hist-analytic_rbjones.com/attachments/20091109/e52c1687/attachment-0002.html>

More information about the hist-analytic mailing list