@edgarblythe,
Well, you can look at Feynman diagrams, the advanced and retarded components of electromagnetic waves, and more broadly, Wheeler-Feynman time-symmetric theory, to get non-linear temporal interactions (i.e. future and past states rolled up in current states).
More fundamentally, it occurs to me that in quantum theory, the state of a system is actually a superposition of possibilities, each having its own probability, called eigenstates. Only when a system is observed (i.e. only when a particular quantum value such as momentum, position, etc. is measured) does the system revert to a single eigenstate (i.e. the particular measurement value observed).
Now, in order to know what time it is you have to make a measurement. It might be the distance that a clock hand has moved, or the number of vibrations of a quartz crystal in a digital watch, or something else: but you have to make an observation before you can talk about what time it is.
But because that observation involves quantum uncertainties, the system has no eigenstate until you conduct the measurement. But the measurement determines the time. And the particular eigenstate that the system reverts to is a matter of randomness and probability, according to quantum theory. So there is no one time until you make an observation that creates a particular time value measurement.