This is crappy armchair sociology. The number and age of unique references in modern journal papers is a measurable number, and I'm sure these folks have done a

This is crappy armchair sociology. The number and age of unique references in modern journal papers is a measurable number, and I'm sure these folks have done a great job of measuring it. But where's the evidence for the conclusion that changes in this number represent changes in the nature of scientific thought itself? Have these people even worked as scientists? Here are some reasons why papers get cited: - Many journal articles open with one or two Ancient Authorities -- articles from 1899, or a famous article written by the founder of your field in the 1950s. These are there as a kind of mission statement: If you cite Folkman et al from the 1970s, you're saying "This paper is about antiangiogenesis". In a world where computers can tell you "here's a list of all the papers published this year that cited Folkman et al, 1971", putting an Ancient Authority citation on your paper is an awesome form of SEO, much better than a keyword. (It's not polite to point this out... but there's a good chance that most of a paper's readers, and even some of the authors, haven't even read the Ancient Authority paper. That's often not a problem, because the first paper in a field is usually an even mix of (a) things everybody knows by now and (b) things that turned out to be wrong, later.) - You cite work that directly anticipates yours. For this, you look for the most recently published relevant references that you can find, partly because it shows that you're paying attention and partly because if the person reviewing your paper has published on the same topic, and you don't cite their most recent work, they will find a way to make you wish you had. - You cite works as a shorthand for the ideas contained in them. This is like me citing PG essays here on news.yc, or like the great Star Trek:TNG episode "Darmok" where the alien language is composed entirely of brief references to (alien) classical history. When you're doing this, it is once again important to cite the most recent works that you can -- it's like the difference between commenting on a recent blog post and commenting on a blog post from 2001, or the difference between referencing LOLcats (a stale cliche), referencing ALL YOUR BASE (a painfully stale cliche) and referencing, um, ST:TNG (which makes me sound like I've been trapped in a box since 1990). And you want to name-drop popular articles, because what's the sense in trying to reference things that nobody has read? This factor alone probably accounts for the observed changes in citation frequency. The Web is good at spreading new memes quickly! And that works for science too! It used to take years for new fads to sweep the nation; now a guy can claim he's Familiar With All Internet Traditions on Tuesday and there'll be T-shirts on sale by Friday night. Oh, and let me not forget: - People cite their own recent work to try and encourage others to read it (and to boost their own personal impact factor). I doubt that the rise of technology has changed that.