Problem is, most papers aren't that good or are not relevant to what you're doing, but it's usually hard to tell based on the abstract. When I do research in a subject, I find ~10 papers out of which I'll read (parts of) maybe one. If I were to buy each of them for $15, I'd end up paying $150, and only get a few pages of read useful material out of it.
Exactly, without a strong reference from someone you trust who also knows the subject, you have no idea if the article is worth $15. And then you could probably just make a copy of their copy.
