For binary linear codes of length , the bar product of binary linear codes is . The parametrization is injective and linear, so its dimension is . If , a nonzero word has Hamming weight at least ; if , . Taking and of minimum weight shows
For , the inner product of and is , since . The spaces have complementary dimensions, proving .
Define the Reed-Muller code by evaluating all multilinear polynomials of total degree at most at the points of . The squarefree monomials are independent functions, giving dimension . Splitting according to the last variable gives the Reed-Muller bar-product recursion , with the endpoint codes interpreted in the usual way. The distance recursion gives .
To identify the Dual of a Reed-Muller code, multiply a monomial of degree at most by one of degree at most . The product omits some variable, so its sum over is even and its binary inner product is zero. The binomial dimension identity then gives
A square-free monomial in variables of degree evaluates to one at exactly points of , an even number. Since parity of Hamming weight is a linear functional over , every sum of such evaluation words also has even weight. Hence every word in has even weight.
If and have degrees at most and , then has degree at most . The binary inner product of their evaluation words is the parity of the weight of the evaluation of , and is therefore zero. Thus
Finally, binomial symmetry gives
so the inclusion has equal dimensions and
This is the Dual of a Reed-Muller code.