Difficulty understanding Linear maps between vector spaces as $(1,1)$ tensors on the target vector space

48 Views Asked by At

In Nakahara's book "Geometry, Topology, and Physics" there is an exercise that asks to show that if $f$ is a Linear Map between vector spaces: $f: V \rightarrow W$ then it is a tensor of type $(1, 1)$.

The exercise is fairly simple because there isn't a lot of ways to "force" $f$ to fit this definition, clearly for $v \in V$ we have $f(v) \in W$ and hence $f(v)$ is a type $(1,0)$ tensor on $W$. So, by virtue of this we can go on to say that since $f(v)$ acts on a co-vector $\omega \in W^{\ast}$ we have a map that takes pairs of the form $(\omega, f(v))$ to some underlying field $K$ and hence establish that $f$ is a $(1, 1)$ tensor.

My difficulty with this first and foremost, is that $f$ is not guaranteed to be surjective on $W$, since it is only said to be a Linear Map and nothing further. So can it be a proper Tensor on $W$ at all in such a case?

A secondary problem is how contrived it appears to force $f$ to fit this definition, and I am not sure why and where is it useful to do this. I'll be glad if someone can help me to clarify this.

1

There are 1 best solutions below

12
On BEST ANSWER

In case of finite dimensions and if you have basis $$V={\rm span}\{b_1,...,b_n\},$$ $$W={\rm span}\{c_1,...,c_m\},$$ then one defines $L_{f\otimes w}:V\to W$ via $$L_{f\otimes w}x=f(x)w,$$
where $V^*={\rm span}\{\beta^1,...,\beta^n\}$ is the dual of $V$ and so $\beta^k(b_j)=\delta^k_j$.

That transformation is linear and the element $f\otimes w$ is in $V^*\otimes W$.

The task of checking that $L:V^*\otimes W \to \hom(V,W)$, mapping $$f\otimes w\mapsto L_{f\otimes w},$$ is a vector spaces isomorphism, would be a routine.

Observation:

The elements $\beta^k\otimes c_j$ serve to generate $V^*\otimes W$.