In this lecture series a definition of a tensor is given roughly as follows:
(Definition 1):
[we] define tensors as variants that transform from one coordinate system to another by a very special rule. A variant $T_i$ is called a covariant tensor if its values $T_i$ and $T_{i'}$ in the coordinate systems $Z^i$ and $Z^{i'}$ are related by $$T_{i'}= T_i J^i_{i'}$$ [then a similar definition for a contravariant tensor]
How is this reconciled with the definition that seems to me to be much more general:
(Definition 2):
Given vector space $V$, a tensor is a multilinear map $V^* \times ... \times V^*\times V \times ... \times V \to \mathbf{R}$
It seems to me that the first definition is extremely restrictive compared to the second one. Aren't there other types of linear transformations that tensors can do (acc. to the second definition) that are not specifically transformations between coordinate systems? (even if you ignore the fact that the first definition is restrictive in terms of the dimensions of the tensor).
What explains this difference?