A lot of things, but this paper points to one issue:
* In order to get an academic position, you need to have letters from a research community.
* In order to get citations, you need your papers to be used by a research community.
This gives a strong advantage to work done within existing lines of research. Virtually anything outside of one of the existing "academic cottage industries" or "mutual adoration societies" is at a huge disadvantage.
You end up with research communities going down long rabbit holes with tunnel vision, and big broad areas of potential research are never explored.
* In order to get an academic position, you need to have letters from a research community.
* In order to get citations, you need your papers to be used by a research community.
This gives a strong advantage to work done within existing lines of research. Virtually anything outside of one of the existing "academic cottage industries" or "mutual adoration societies" is at a huge disadvantage.
You end up with research communities going down long rabbit holes with tunnel vision, and big broad areas of potential research are never explored.