The Dot Product, the Cross Product and Duality
Until now we have been discussing “how vectors are added and how they are scaled by a scalar”. This chapter introduces multiplication between vectors for the first time. The dot product gives a scalar and measures the angle and the length of the projection between two vectors; the cross product gives a new vector perpendicular to both. These two operations look like unrelated tricks, but in fact both are special cases of matrix operations in the simplest dimensions.
The chapter begins with the dot product, focusing on its geometric meaning — projection — rather than on “adding up a few products in the formula”. It then uses the right-hand rule to build up the intuition for the cross product, so that you understand why its result is perpendicular to the two original vectors and why its norm equals the area of the parallelogram. Finally it brings out the real point of the chapter: duality — there is a strict correspondence between the dot product and matrix-vector multiplication, and this correspondence exposes a common algebraic structure behind a large family of operations that look entirely different.
Duality looks abstract at first, but it explains why so many apparently unrelated problems — least squares, orthogonal projection, Markov chains — end up with answers of a similar form.
