Jonathan, you are confusing functionalism with behaviorism.

I think what you are really saying is as follows: Suppose that Bob hears input Q at 2:00, thinks about it for a minute, from 2:00 to 2:01, and then gives output A.  You seem to think that functionalism holds that Bob's consciousness during that time depends only on Q and A.  So if someone shoots Bob dead one second before 2:01, so he never gives A, that would mean his consciousness from 2:00 until that time would be retrocausally removed or changed.

Of course that is nonsense.  Functionalism has nothing to do with input and output.  It holds that the internal functioning of Bob's brain, the states it passes through and the transition rules, is what causes his consciousness during those times.  Those things are obviously unaffected by future events.