[conal]

Lazy and eager evaluation are both sequential evaluation schemes, and are properties of language implementations, not of languages. Haskell is thus not lazy (nor is any other language). Rather, Haskell is a non-strict language, meaning that lambda is just abstraction, not abstraction-plus-strictification as in "strict" languages. Non-strictness does not imply sequential evaluation or indeed any particular evaluation order. (Ditto for strictness.) For instance, speculative evaluation is a parallel strategy that can implement non-strictness.