原文読んでみたけども、最初からパラメトリック多相の文脈で話してるし
問題の部分のちょっと前で暗黙どころか明確に否定している気が…

> The intuitive expIanation of this result is that r must work on lists of X
> for any type X. Since r is provided with no operations on values of type X,
> all it can do is rearrange such lists, independent of the values containedin them.
> Thus applying a to each element of a list and then rearranging yields
> the same result as rearranging and then applying a to each element.