The Inverse Matrix, the Column Space and the Null Space

The rank of the previous chapter answered only “how many independent directions are there”. This chapter has to answer another question: where does the transformation send a vector? These are two different questions, and they need two sets of tools — the column space and the null space.

Geometrically you can imagine a linear transformation as a machine: the column space is the set of all the results the machine can “spit out”, that is, the image of the input space; the null space is the input directions that the machine squashes to zero, which vanish completely in the output. When the sum of the dimensions of the column space and the null space equals the dimension of the domain (the rank-nullity theorem), the machine has wasted no dimension at all — this is where the theorem’s name comes from.

The chapter first builds the intuition for the inverse matrix: it is the operation that turns the machine back, so the inverse exists exactly when the machine squashes no direction, which is the same fact as “the determinant is non-zero”. It then works through the column space and the null space and explains why they are a pair of concepts — there is a correspondence between the row space of a matrix and its column space and null space. Finally the rank-nullity theorem strings the rank, the dimension of the column space and the dimension of the null space together into a single equation.

Chapters

Found an issue with this content?

Tell us what could be improved.