Each of the resources in a given composite expression in the static case are physically proximal, because in order to be included, they must exist in a physical form, right on the final expression. One can not display a picture on a newspaper page without printing a copy of that picture in ink on all of the printed newspapers. This may sound stupid, but it will make more sense when contrasted with the software domain.
In particular, the most important example of physical locality is between the consumer and the expression itself. In the old, static case (think 1970's again), there is no way for many people to consume an expression at a distance. A fully independent copy of the expression must be delivered physically to the user in order for the user to experience the message; this is why the Reader is shown at the top of the derivation tree in the figure, because without the reader this is all an exercise in futility. For any media the user can capture and use (video via VCR, physical possession of a book, etc.), the user will always possess a copy, no matter what the original copyright holder may desire, short of theft or confiscation by the government.