Paradoxes [Digest urth.v024.n023]
It's not true that all time-travel (properly considered)
has to lead to paradoxes.
First, forward-in-time travel is non-paradoxical. The
proof is that we are doing it right now! Vinge did a lot
with fowards-only time travel, as have others (the oldest
such story I know of is Sleeping_Beauty, but Rip_van_Winkel is more aware of the difference between the "now" of the sleeper and the present).
Second, in the many-worlds model, timelines split at "change points", either for all change points [to support quantuum mechanics] or just for "imposed" points due to time travellers; in that case, backwards time travel is non-paradoxical:
The mad scientist Dr. Parah travels from 1999 to
2002 and grabs copies of the best-selling novels
from a bookstore. He than travels back to 1999
and (having copied them) sends the manuscripts
to publishers as his own work.
The original author still wrote the book during 2000 in
timeline 1, and though the book appears full-fledged (with
no apparent author) in 1999 in timeline 2, that's not a
paradox. Note that the theft from the bookstore only
happens in timeline 1.
"Many worlds" is prodigal of universes, but logically consistent. The movie "Sliding Doors" is an example.
Third is the one I think more Wolfean: rather than a huge number
of timelines, we have a universe in which history has near-repeats. Similiar causes, similiar effects: so there are ages of barbarism and of civilization, ages of discovery and of forgetting.
Then you have a Severian, in one cycle, write a book but not read it. Another Severian (very much like the first) reads the book and may or may not write another, etc.
See the discussion in Steven Pinker's How_the_Mind_Works of how repetition in time (via generations) can mimic backwards or circular causation.
-John Bishop