Understanding how societies and civilizations reacted to different pressures and how the structure and nature of their institutions evolved and handled those pressures is much more important than remembering what happened where.
Those who fail to learn from the mistakes of history are doomed to repeat them, as they say.
I frequently forget the trig identities, but I know the basic definitions of trigonometry and that there ought to be certain identities, so I know I can re-derive them or look them up. But, I only need to do this when I teach calculus.
> You can also re-discover the dates by doing your own historical research
My point was that you can put a mathematician in a closed room with only pencil and paper, and they could redevelop an entire theory if needed. A historian in the same position would be unable to do any work -- they need primary sources. I went with your "learning lists of dates" as a stand-in, since, like primary sources, you can't derive dates from scratch.
Sure, history has a "how" of research methods, but the research products must rest on the "what." Mathematics is almost entirely "how." The "what" for a mathematician is mainly knowing what's been done and what needs doing, for the purposes of directing research and giving attribution.
Also, I am pretty sure G.C. Rota was being somewhat facetious in quite a lot of the article. These are "lessons of an MIT education." Lessons aren't necessarily truth.
Often mathematicians will make informal arguments but the idea is that you should be able to add rigor when needed.