[Strilanc]

Whether or not you drift there or instantaneously teleport there, there are still causality issues. You can make it into your own past by jumping once, changing your speed to tilt the frame of reference you're moving "instantaneously" in, then jumping again.

http://casa.colorado.edu/~ajsh/sr/simultaneous.html

[logfromblammo]

I suggest that encountering yourself in your own past would be on the same order of difficulty as plotting the course of a photon across an entire galaxy, to hit a "mirror", precisely positioned to return that photon back across that same galaxy, this time apparently as an antiphoton, to another "mirror", that would bounce the photon back towards itself, prior to the time at which you redirected it across the galaxy. The precision required makes it effectively impossible.

But, given the nature of probabilities, if enough trials occurred, eventually one would succeed in causing something to interact with its past self. How would you even detect if that occurs? The particles that have the highest likelihood of doing this by chance all look pretty much the same anyway.

[Someone]

That argument assumes that our understanding of physics is correct and complete. If FTL is possible, that is an incorrect assumption. But yes, any theory including FTL travel must explain how nature behaves if this happens.

[jerf]

No, it's much weaker than that... that assumes that our understanding of relativity is even remotely correct. You can obtain the FTL/causality problem from even a brutally simplified version of general relativity, such that one can literally draw the problem on a piece of paper. For FTL not to break causality essentially requires that A: you can only FTL jump once and only once in the entire history of the universe as long as you don't leap into your own past light cone, B: FTL is impossible or C: Causality is not absolute.

I've often thought that the announcement that we've developed FTL would actually be a very bad thing, as fun as it sounds at first. Sort of like this story, where it turns out a proof of P = NP is just about the worst possible thing: http://www.antipope.org/charlie/blog-static/fiction/toast/to...