This thought occurs to me and it’s kind of disturbing: what if one engages in scientific study of a system so complex (e.g. Earth) that it takes decades to understand. Understanding accumulates so slowly that it does not really advance. The reason is that people forget what they knew 5 years ago as they learn new things today. So a synthesis of old and new knowledge does not really occur, because one does not keep all the relevant facts in one’s head at one time.
This is particularly troubling if past knowledge was tentative, and not conclusive. In that case, the past knowledge is never really confirmed or refuted. This is another hindrance to advancement.