What are the arguments against defining learning as a change in either cognitive structures or in the nervous system?

What are the arguments against defining learning as a change in either cognitive structures or in the nervous system?