Why functional languages really need parallelism

The essence of the desirability of functional languages is their great extensibility, compared with other language classes, on account of their formally greater effective (as opposed to ultimate Church-Turing-theoretical) expressive power. A straight-forward extension exercise shows that mere "functional" languages ultimately capitalise rather than promise, that certain functionality is beyond their effective expressiveness. In order to close the functional expressiveness gap, parallelism, hitherto regarded as an interesting implementation device, emerges as a necessary semantic enhancement to functional languages.<
>