To provide context and pointers to previous literature. Very often it happens that only some of the insights from papers fall into fashion and when you go back and read the original you find other things of value that didn't get picked up.

As a reader, I'm often left trying to figure out whether the author is presenting something new, or is rehashing old ideas (which, btw, is a good thing sometimes, but I want to know that's someone tried this once before). I want to know exactly what the new contribution of the paper is, and how it fits in the universe of contributions.