When the continuous growth will flatten what will be the attractivity if not cheaper borderless uncensored usage?
People are already working on that. Maybe not as much as it should, but they do. Until then, you're required to pay an extra buck. The inherent nature of things should make it clear that achieving universal satisfaction is impossible. It is unrealistic to expect a payment network to be simultaneously free-of-charge, uncensored, decentralized, and immutable, without any drawbacks that I cannot even think of.
There are tradeoffs we deal with. If you prioritize transaction cost, use banks. If you prioritize anonymity, use physical cash. If you prioritize the ability to send money via the Internet, in an censorship resilient and borderless fashion use Bitcoin.
Reducing the size would trigger incompatibility, all the previously mined blocks would still be valid under an 8 or 16 MB rule.
Prior-version nodes will reject the new blocks, while post-version nodes will continue normally. That is the definition of a hard fork.
Hundreds of academic research mean nothing when you have to put food on the table, just as nobody gives a crap about the world ending in a climate catastrophe when the only way to go to work is by driving a wheeled furnace, or whatever the name of vw models is right now!
So you're going to dismiss all that research, and assert that there isn't a scaling solution other than tinkering with the block size limit?
Few care about academic research when they have to pay 100sat/b or wait for one month till the fees go down!
And apparently[1][2][3] even fewer are so naive, to use what dismisses all that research of how Bitcoin is proposed to work.
[1]
https://coinmarketcap.com/currencies/bitcoin-cash/[2]
https://coinmarketcap.com/currencies/bitcoin-sv/[3] <insert other hard fork crap>