To change Bitcoin everyone (basically a super majority), users, miners, and merchants, have to accept the change, otherwise you get a fork.
It's not "otherwise you get a fork." You do get a fork in that case, even if you get a super majority to switch over. You really just end up with two coins that are both claiming the name "bitcoin," and you're hoping that enough will switch over that the new one is enough more popular to make the old one more or less irrelevant. But miners that spent money on special hardware that isn't going to be usable with the new coins are going to be highly motivated to keep the old fork alive as its own alt coin, and users will have coins in both the old and new forks so nothing stops them from using both, right? So practically speaking getting a super majority to stop using the old fork is going to be very difficult. If the miners with the sole-purpose hardware represent a majority of the hashrate then getting the hard fork to happen at all is going to be a tough sell.
Furthermore, if there's a compelling new technology implemented in a new altcoin built from the ground up, and it is sufficiently superior to merit a hard fork to try to duplicate that in bitcoin, then before you could coordinate the miners into agreeing to the hard fork and put out new client software and merchant software and exchange software to support the new fork, and publicize the change so that everyone will know to get the new software and start using the new fork, and stop using the old fork, you'll have people moving their money to the new altcoin that's completely independent.
And once that starts it's going to accelerate. Digital currencies aren't reliable stores of value because the exchange rate depends solely on market share, and market share is fickle. If there's any sign that a currency is going to lose market share you'd better get out ahead of the crowd or you'll lose big. Announcing that a hard fork is going to occur on such and such a date in order to adopt the desirable technology currently found in zerocoin (or whatever) seems like a good way to guarantee an accelerating stampede out of bitcoin prior to the switchover date.
Furthermore, it's a big leap to assume that future digital currencies will use the blockchain architecture at all. The global ledger is a clever idea, but there's no reason to think that it's the only possible architecture. And if some future idea is fundamentally different on that point and offers significant advantages as a result, then it's the sort of change that bitcoin simply wouldn't be able to adopt.
Basically a hard fork for the kind of enhancement we're talking about (as opposed to an emergency fix to blockchain corruption or something needed just to keep bitcoin working at all) doesn't seem likely to be feasible in practice.