>>11191946arXiv started as a way to accumulate different physics departments' preprint systems. in physics it was very important to circulate a preprint that would go on the record before actual journal publication because in physics it is very easy for others to find the same result (we are just dealing with math [in theoretical physics] or data [in experiment] which is hardly "proprietary")
so in physics it became necessary to disseminate results before they are peer-reviewed or disseminated in journals because otherwise douchebags could scoop you by publishing the less controversial/challenging subsets of your results before the more comprehensive results managed to get published.
why does this not affect medicine or biology or psychology? probably because they have less to do with theory or reproducible data. they design their experiments very specifically and can develop very specific methodologies. in theory they care very little about experiment so their "intellectual property" is completely non-proprietary and anybody else can figure it out, and in experiment, the experiments are so huge that the data they get is necessarily shared by thousands of researchers, so individual claims are impossible.
therefore my basic point is that arxiv is the marketplace for people to publish original ideas that authors are proud of THINKING of before anyone else, whereas in bio and psych and med, the "thinking" is a side-note to whoever does whatever specialized study. so there is no pressure to get things published when the real requirement is doing specialized studies
