An informal approach to formal inner products (Q2501671)

This article is about an alternative classroom presentation of the concept of the inner product on a real or complex vector space. Instead of defining the inner product by the traditional method as a symmetric, bilinear, positive-definite function over a vector space, the author suggests an alternative approach (which turns out to be equivalent). More precisely, a finite-dimensional vector space \(V\) over a field \(\mathfrak F\) (\(\mathfrak F\) is either the set of real or complex numbers) is endowed with a length function, a symmetric relation (corresponding to the orthogonality of vectors), and a concept of similarity of triangles. Let \(w\in V\). For \(v\in V\) let \(E_w(v)\) be the projection of \(v\) in the one-dimensional subspace \(\text{span} (w)\). Then the inner product on \(V\) is defined as a function \(\langle\;,\;\rangle : V\times V\to\mathfrak F\), so that \(E_w(v)= \langle v, w\rangle\frac{w}{| w| ^2}\). If \(w=0\) set \(\langle v, w\rangle =0\). According to the author, such a presentation of the inner product is more pedagogical. This needs to be tested more efficiently though. The reason being, is that the usual inner product for vectors in \(\mathfrak R^3\) (even in \(\mathfrak R^n\)), as very naturally defined by the use of the angle between any two vectors, requires less cognitive effort than the concept of the projection. The easily verified properties of being symmetric, bilinear, and positive-definite can be used to generalize this definition to any (real or complex) vector space.