Hm, that's not really what I was trying to say. It does not seem incoherent to me. I feel like I don't understand what you are saying though.
My concern is something like this: observing and modelling patterns in the universe can only take place within time. If your models derived from the observed patterns then allow you to model a universe that does not have time, then there is something fishy with the model -- you've bootstrapped a timeless universe that is only intelligible from within time.