Often it makes no sense to multiply together two elements of a vector space, but for some pairs of linear maps a useful product exists.
Product of linear maps
If and , then the product is defined by
for all .
Thus is just the usual composition of two functions, but when both functions are linear, we usually write instead of . The product notation makes the distributive properties (see below) seem natural.
Note that is defined only when maps into the domain of . We can verify that is indeed a linear map from to whenever and .
Algebraic properties of products of linear maps
- Associativity: whenever , , and are linear maps such that the products make sense (meaning maps into the domain of , and maps into the domain of ).
- Identity: whenever ; here, the first is the identity operator on , and the second is the identity operator on .
- Distributive properties: and whenever and .
Multiplication of linear maps is not commutative. In other words, it is not necessarily true that , even if both sides of the equation makes sense.
Example: Two noncommuting linear maps from to .
Suppose is the differentiation map (as we saw in here), and is the multiplication by map we saw. Let’s consider applying these maps to some polynomial .
Then:
but
Thus, – differentiating and then multiplying by is not the same as multiplying by and then differentiating.