"The longest chain" is a pretty common false argument by forkers as it tries to mix up two concepts - consensus within a protocol and community consensus on what chain is Bitcoin. The first means that the longest chain is the valid chain - this is just the basis of how Bitcoin works and how all nodes are in sync but when there's a fork that changes protocol rules, it doesn't matter if how many blocks the new chain has, if people don't start switching to that software it's as if that chain hasn't existed. Miners can't dictate to users what chain is Bitcoin, in fact it's the users who can create incentives for mining by trading coins on open markets.
I agree that you have a plausibly reasoned theoretical argument on how bitcoin should work but, in the unstructured commercial world, enforcement is an important issue as well. And the key enforcement in Bitcoin is majority hash power. Hence we seem to be increasingly likely to see Core Development running a low hash power alt coin or Core Development "walking away" It seems that Miners (in league with declared hashfollowing exchanges) can indeed dictate what chain is Bitcoin and identified as BTC on the exchanges. If the miners hang together when i say "Hey Google - what is the price of Bitcoin" I'm going to hear what the price of Bitcoin-from-segwit2x is at that moment, regardless of the intelligence, blood, sweat and tears that Core has put into Bitcoin.
"If people don't start switching to that software" - Tons of people buy bitcoin from Coinbase and Gemini leaving them on the exchange or using SPV wallets - the first is an NYA signatory the second is a stated majority hash power follower. Tons of people pay bitcoin to commercial vendors using Bitpay another NYA signatory. I'm not seeing a problem for Bitcoin-from-segwit2x here.