[FOM] Could PA be inconsistent?

Richard Heck heck at fas.harvard.edu
Fri May 14 13:03:04 EDT 2004

> There are now two threads going on FOM contemplating this possibility. 
> No one has mentioned that we possess proofs that PA is consistent. 
> These proofs can be carried out in quantifier-free primitive recursive 
> arithmetic with induction permitted on a primitive recursive 
> well-ordering of the natural numbers of order type epsilon-0. These 
> proofs (Gentzen, Ackermann, G\"odel) have gone unchallenged for decades.
> Martin

Crispin Wright once compared the situation here to that with 
philosophical skepticism. We have what we take to be good reasons to 
believe e.g. that the speed of light is finite, that the sun will rise 
tomorrow, and, indeed, that there are other people in the world. But, as 
Descartes famously pointed out, our reasons could conceivably be 
revealed as fraudulent. It has now become the common wisdom in 
epistemology, however, that the mere fact that such an eventuality is in 
some sense conceivable does not undermine our claim to know such things 
as those mentioned. For similar reasons, the mere fact that the proofs 
of PA's consistency Prof Davis mentions depend upon assumptions that 
are, in a reasonably well-defined sense, stronger than those of PA 
itself does not undermine the reason-giving force of the proof. /If/ PA 
were shown to be  inconsistent, then the reasons we take ourselves to 
have for believing it is consistent would, of course, thereby have been 
shown to be fraudulent. But that is no more interesting a fact than 
that, /if/ it were shown that all my experience was the result of an 
evil demon's manipulations, my reasons for believing Prof Davis exists 
will have been shown fraudulent.

Richard Heck

More information about the FOM mailing list