Hurkyl
Nov16-04, 05:52 PM
Conjecture:
Let F be a field ordered by <
Let F(p) be ordered by <'
Let F(q) be ordered by <''
(These may be either algebraic or transcendental extensions)
Suppose that F(p) and F(q) are isomorphic as fields (with p mapping to q)
Suppose that the restrictions of <' and <'' to F coincide with <.
Suppose that for any b in F:
b <' p iff b <'' q
Then, we may conclude that F(p) and F(q) are isomorphic as ordered fields.
------------------
In the case that F(p) has a positive element smaller than any positive element of F, I think the proof is straightforward, so I've been working entirely on the remaining case... but I'm having difficulty: I keep tripping over myself because everything seems so obvious. :frown: Anyone have any ideas?
In case you're wondering, this came up because I was interested in generalizations of metrics where the metric function is into any ordered ring, instead of the reals. I think that nonstandard models of the reals might make a nice class of alternatives, and I think that any ordered ring can be embedded into one such model (and this is what I'm really trying to prove)
Let F be a field ordered by <
Let F(p) be ordered by <'
Let F(q) be ordered by <''
(These may be either algebraic or transcendental extensions)
Suppose that F(p) and F(q) are isomorphic as fields (with p mapping to q)
Suppose that the restrictions of <' and <'' to F coincide with <.
Suppose that for any b in F:
b <' p iff b <'' q
Then, we may conclude that F(p) and F(q) are isomorphic as ordered fields.
------------------
In the case that F(p) has a positive element smaller than any positive element of F, I think the proof is straightforward, so I've been working entirely on the remaining case... but I'm having difficulty: I keep tripping over myself because everything seems so obvious. :frown: Anyone have any ideas?
In case you're wondering, this came up because I was interested in generalizations of metrics where the metric function is into any ordered ring, instead of the reals. I think that nonstandard models of the reals might make a nice class of alternatives, and I think that any ordered ring can be embedded into one such model (and this is what I'm really trying to prove)