[FOM] Who was the first to accept undefinable individuals in mathematics?

Vaughan Pratt pratt at cs.stanford.edu
Wed Mar 18 04:22:21 EDT 2009

Kreinovich, Vladik wrote:
> This description makes the enumeration of algebraic numbers perfectly
> definable. 

When I proposed a rational multiple of the Liouville constant as a 
transcendental number in [a,b], it was with Bill's mention of Sturm's 
theorem in mind.  Like Vladik I wondered why Sturm's theorem was needed, 
and concluded that Bill wasn't content to *define* a transcendental in 
[a,b] but must have wanted to *construct* one.  For those puzzled by my 
width-2 intervals, my goal was to be impeccably constructive.  A 
function f(a,b) producing a transcendental in [a,b] cannot be 
*constructed* without excluding a set of measure zero (interpret that as 
you will).  However it can be *defined* everywhere (on the upper left 
triangle of R^2 defined by a < b), as Vladik indicated (though again 
Liouville's constant gives a simpler definition).

(I'm not actually a constructivist, I'm just trying to play one on FOM.)

Vaughan Pratt

More information about the FOM mailing list