I believe the human population will experience huge deflation in the coming decades, so a tail emission is counterproductive (unless you really believe the human population will keep increasing).
I agree with most of your post, but tail emission could be countered by the "lost coins" if it's sufficiently low, and thus I don't really think it would incide really in tokenomics, even with a reducing population or a stagnating or even gradually shrinking economy.
The beauty of a "sidechain-driven" tail emission (by merge-mining and miners getting rewarded by sidechain utility tokens) is however that the emission schedule would be much more flexible than a hard-coded tail emission. For example, if the sidechain token inflation resulted to be too high to maintain the sidechain utility tokens' value, sidechains have much lower hurdles to hardfork into a lower emission schedule than the main chain, and of course vice versa too.
This isn't that relevant for the Monero <-> Bitcoin discussion (which is pointless and borderline OT) as Monero has already had a history of hardforks and thus could hardfork again and again to lower their supply, but for Bitcoin even a single hardfork could be very difficult. Thus, the "sidechain-driven" tail emission model would be the concept least intrusive to BTC's "core value proposition", which includes its "deflationary" nature. The sidechain utility tokens' supply would not affect Bitcoin's emission schedule and its deflationary nature at all.