This is a bit more abstract and advanced but one way of looking at it is this: Given an n-dimensional vector space, V, over field F, the set of all functions from V to F, the "dual space" to V, is itself a vector space, V*, with addition defined by (f+ g)(v)= f(v)+ g(v) and (af)v= a(f(v)), also of dimenion n. We can then represent functions in V* as "row matrices" and the vectors in V as "column matrices" so that the operation f(v) is a matrix multipication.
However, because it is still true that V and V*, both being n-dimensional vector spaces, are isomorphic we can identify one with the other, the row and column matrices as both representing vectors and think of the matrix multiplication as an "inner product" on a vector space.