I was thinking about this during the microeconometrics seminar yesterday.
A lot of people have noted how rough economics presentations can be relative to other disciplines. It occurred to me that this is a lot like the mentality of a courtroom. It's not that economists are mean. It's just that the adversarial approach is judged to be efficient for getting at "the truth" (whatever the hell that is). In some ways it's effective to bounce ideas off each other constructively, but another approach is to adopt very specific roles. The author of a paper wants the paper to be convincing - that's why she wrote it. And she thought it was convincing otherwise she wouldn't have had an interest in writing. She is in the best position to defend that argument. Now the audience might have a high view of the paper as well but they are not in as good a position to defend it. It's not clear there's any value added in defending it when the author is in the room.
There is value in trying to break the paper.
I'm often critical of papers I think are awesome in seminars. Why? Because the authors know why the paper is awesome. They don't need me to tell them it's awesome. And more importantly, it does nothing for their paper, which presumably we all want to be the strongest piece of scientific writing that it can be. So I take on the role of critic and try to break the paper in whatever way it seems most sensible to break it. It's not completely adversarial - if there's an obvious way to suggest how to deal with the problem, of course you mention that too. And you don't want to come across mean, so you might even preface by saying "I think this is an awesome paper, but...". But aside from that you don't dwell on it, because like I said - the author is very well aware of the various ways in which the paper is awesome.
The important thing is that these are roles that people take on, not positions that they hold. We provide our positions on a paper in literature reviews, for example (i.e. "So-and-so provides an awesome investigation of such-and-such that speaks to what I'm trying to do here"). But in the seminar room when you run across the exact same paper, it's sort of understood that your most productive role is to be an adversary.
Obviously this story is adjusted for context. When it comes to comment periods for my colleagues' papers I try to comment on both what I like and don't like. We're colleagues and we're all still feeling our way around this profession so there's value in building that sense of community. Presumably other adjustments for context make sense as well. But overall, I think the adversarial system works.
I kind of take an adversarial approach to blogging too, although I try to do my best to share things I like as well as what I don't like.