There isn't going to be a fork. Come early 2016 either most will support > 1MB blocks or most won't. No one is going to be want to be on the wrong side of the argument - it's just a losing proposition.
Then there might be a fork. Any change to the max block size will require a fork.
No. Introducing a software change that has the *potential* for a hard fork does not mean that there will be one. Ask yourself this: If 90% of nodes are supporting > 1MB blocks, including miners - would you hold tight and not upgrade? Same will be true for vice versa. Once > 50% of the nodes go one way, everyone else will too.