Wednesday, 27 February 2008

algebraic number theory - Is there a notion of Galois extension for Z / p^2?

The above title is in fact a special case of what I want to ask.



Certainly we have a well defined notion of Galois extension for mathbbQp. The intersections of these extensions to the ring of integer of the absolute algebraic closure of mathbbQp give us a notion of Galois extensions for mathbbZp. ( I know that there is a notion of Galois extension for commutative rings, and I believe that it should give us this. Am I correct?)



Let's go further. Let AK be the ring of integer in a finite Galois extension K of mathbbQp. Let e be the ramification degree of K over mathbbQp. The injection of mathbbZp into AK will induce an injection of mathbbZ/pn into AK/mathfrakpen. In this picture, there seems to be some desire to say that AK/mathfrakpen is the correct notion Galois of extension of mathbbZ/pn. But there are problems; taking this notion of Galois extension, if K is has ramification degree e>1, the corresponding extension AK/pe is not a field (it is not even an integral domain).



Question 1: Is there any notion of Galois extensions corresponding to what I desire?



Question 2: Can a class field theory (i.e a nice description of absolute abelian Galois extension) of mathbbZ/pn be developed in this context? Is there any relationship between this and the local class field theory of mathbbQp ( which is the same as that of mathbbZp)?

No comments:

Post a Comment