From: vadimpaster@virtue.com (Vadim Pasternak) Subject: Is this function polynomial ? Date: 1 Jun 2001 07:39:05 -0400 Newsgroups: sci.math Summary: Functions on products which restrict to polynomials on lines Suppose f:Q^2 -> Q (Q is the rationals) is a function of two variables x and y such that f(x,y) is a polynomial in y for each fixed x, and a polynomial in x for each fixed y . Is f(x,y) a polynomial in both x and y ? i.e f(x,y) = p(x,y) where p is polynomial in 2 variables ? Thanks ============================================================================== From: Douglas Zare Subject: Re: Is this function polynomial ? Date: Fri, 01 Jun 2001 23:00:04 -0400 Newsgroups: sci.math Azmi Tamid wrote: > [...] > Douglas can you please share the construction with the rest of us that > do not know what is the experts site ? Well, since someone else asked, for any countable subset of the complex numbers S, there are functions on SxS which are not polynomials of two variables but whose restriction to any {a}xS or Sx{b} is a polynomial. Here is a construction: Enumerate S: S={s_0,s_1,s_2,...}. Choose values for the diagonal {(s_i,s_i)}. Let the function be constant on {s_0}xS and Sx{s_0}. Let the function be linear on {s_1}xS and Sx{s_1} (interpolating through the values at s_0 and s_1), at most quadratic on {s_2}xS and Sx{s_2} (using the values at s_0, s_1, and s_2) etc. For some choices of the values on the diagonal, the result will be a polynomial, but this will not be the case if the choice is f(s_i,s_i)=i or (-1)^i for any enumeration of the rationals. One could also choose (s_i,s_i) at the ith step to force the polynomial function on {s_i}xS to have degree exactly i. Douglas Zare [reformatted --djr]