The fact that [a] is a linked list is exploited in Data.List.tails as well as in https://hackage.haskell.org/package/suffixtree where suffixes and internal segments of the tree all use Cons cells of the original list, so the overhead is only in the tree structure itself. In fact, exploiting the location of the internal tree nodes relative to the original list, one could compress the list similarly to the Lempel-Ziv 78 algorithm. This way, bioinformaticians manage to create full-text index structures [*] that are not much larger, or even smaller than, the original text. Olaf [*] Lempel-Ziv 78 creates a suffix tree of the text holding only a subset of all suffixes, namely those starting at a block boundary. This already suffices to find all occurrences of substrings spanning a block boundary.