BIP-110 doesn't attack taproot, only big tapscripts. they are neutered substantially but compact formats can still be used to do shitcoinery. there was only one reason why they introduced taproot. they removed the hash on the pubkey to make it "cheaper" and also exposed it to quantum secret key cracking, and it piggybacks on the witness discount of segwit. and BIP-110 rule three seems to be basically claiming the right to gatekeep new segwit witness programs. I suppose the claim is that is to stop another thing even worse than taproot being added but it's unquestionably true that a schnorr signature benefits lightning and ark by making it possible to construct transactions that look the same as normal spends, pay in, pay out for closure. making a witness program that just gives us actual schnorr signatures that satisfies the needs for these uses is also now obstructed by BIP-110. in theory. redundantly, because actually nothing is stopping anyone from doing anything because BIP-110 is only a mempool patch not a consensus addition. why it's there one can speculate about but you can't deny that there is a psychological manipulation factor in the way it words it, redundantly, because already it's the case that nodes can't spend from unknown witness program outputs without the actual verifier that confirms the spend of that utxo is valid.
if you ask me, there is a trojan horse behind this specific part of bip-110 and nobody is talking about it, *or* the question of whether taproot should be deprecated (we deprecated legacy p2pkh, there's no reason we can't deprecate anything that has abuse potential and weakens the network). ultimately that is a filter on the mempool too, for taproot that would mean "can spend out of taproot utxo, can't spend into it".