I had a couple things to add to Rob’s excellent article. First, see Drezner and Joyner. I endorse pretty much all of Rob’s argument, but the particularly structure of academic publishing makes his point particularly important.
A point I brought up in my recent podcast with Paul is that — while there’s no way of knowing to an absolute certainty — given the download stats I would bet that my unpublished scholarly articles have been far more widely red than my published ones. It’s a perversity of the current system that official approval immediately makes articles inaccessible to much of the interested public. If I want to reference my work on countermobilization, because it’s relevant to questions about the Supreme Court and same-sex marriage currently being widely discussed, I can put up a link to an article anyone can read if they’re so inclined. But if I want to reference the article I wrote with George Lovell about the relationship between the Supreme Court and political branches — equally relevant to these debates — I can only offer a link that might get you access to the article if you have access to academic databases and won’t if you don’t. The fact that articles become much less accessible to the public as soon as they’ve passed a peer review process isn’t easy to justify.
I generally prefer the social science blind peer review system to the law review non-blind student-edited system. But one advantage of the latter is that authors seem to retain the right to make published articles available to the public. It seems to me that this should be the norm.
And, as I’ve said before, there’s another factor that makes the social science publishing system particularly irrational. One could object that copyright by definition makes information less accessible to the public. But in many cases, there’s good reason for that: copyright makes it more likely that creators of work get paid for it. I think this a serious issue, and I’m disdainful of “information wants to be freeeeeeeeee” arguments that handwave that point away. But in academic article publishing, it’s obviously mostly inapplicable. Writers of articles don’t get compensated, and reviewers don’t get compensated (and generally don’t even get the indirect compensation of having their efforts make a non-trivial contribution to tenure and promotion.) So in academic publishing copyright protections have all of the downside with none of the potential upside.