The whole idea of etymology is that you can figure out what a word means from its roots.

That was the idea in ancient Greece when the name of the endavour/field was created (etymon = “true”). In the 19th century when linguistics became a serious science it was effectively becoming abandoned, and quite clearly criticised by 20th century linguists. Words’ meanings and forms shift inevitably, they’ve always been shifting, and trying to pick one single stage of this process as the right one is basically like saying that the earth is flat because from any individual vantage point it looks flat to you.

If you throw all that out, you give up the scaffolding that makes words make any sense.

No, you don’t. 99% of people don’t know the etymology of 99% of the words they use. Not even linguists have definitive answers for the etymology of words such as ‘boy’ and ‘dog’. Words’ meanings are actually established by usage, by tradition as it’s handed down to us, with some leeway in how we accept and modify the tradition. (These mechanisms are many and affect various levels of language.) Note that cultures that don’t have scientific etymology still have perfectly functional languages.

It seems like the argument for descriptivism is “let’s not be elitist when people become less competent with the rules of a language”

That’s one of the arguments, but as you can see I don’t think it’s crucial.

I suspect there is also a body of professional linguists who oppose your point for the same reasons.

There are some professional linguists who are active as prescriptivists. Their number varies depending on the country, in Anglophone countries their number is miniscule. In countries with a more pronounced prescriptivist tradition (as in mine, I’m from Croatia) their number declines through time as academia accepts and integrates modern linguistic theories, and the remaining prescriptivists’ positions soften. And I can’t help but notice that many of the current prescriptivists are shoddy linguists and politically motivated.

The prescriptivists are actually quite thin on the justifications for their approach. They won’t theoretically or empirically defend prescriptivism, arguments for it amount to vague and unscientific claims of a need for order and clarity in language (which exist regardless of prescriptivist intervention), and such stuff. But even they usually don’t dare to go so far as to claim etymology is the source of correct meanings, because they know that holding such a position would immediately lead into absurdity and extremism. Leaf through an etymological dictionary and try to stick to the oldest meaning described there. You’ll quickly realise that the source of correct meanings can’t be the words and forms from 500, 1000, or 4000 years ago.

A book recommendation, if you’re interested: L. Bauer and P. Trudgill, Language Myths.