Time travel isn't just science fiction: Albert Einstein's general relativity suggests it could exist. And now we might have solved the tricky matter of time paradoxes. It's all just a question of adjusting probabilities.
A certain reading of Einstein's theories argue for the existence of closed timelike curves, which are strange paths of spacetime that take anything traveling on them into the past and then back to the future. First proposed by Kurt Gödel in 1949, CTCs could theoretically exist deep within black holes or other similarly chaotic corners of the universe. Since these CTCs, however difficult they might be to access, would apparently make time travel into the past a genuine possibility, the question then becomes how to deal with the potential time paradoxes.
As always, the Grandfather Paradox, in which a time traveler kills his or her grandfather before he fathered the traveler's parent, gets the most attention here. Various workarounds have been proposed over the years - Oxford physicist David Deutsch came up with an intriguing possibility in the early 90s when he suggested that it was impossible to kill your grandfather, but it was possible to remember killing your grandfather. In some weird way, the universe would forbid you from creating a paradox, even if your memories told you that you had.
This theory, like most others put forward, relies on liberal use of the word "somehow" and the phrase "for some reason" to explain how it works. As such, it's not an ideal explanation for paradox-free time travel, and that's where a new idea by MIT's Seth Lloyd comes into the picture. He says that paradoxes might be impossible, but extremely improbable things that prevent them from happening very definitely aren't.
Let's go back to the grandfather paradox to see what he means. Let's say you shoot your grandfather at point-blank range. This theory suggests that something will happen, such as the bullet being defective or the gun misfiring, to stop your temporal assassination. This can involve some very low-probability events - for instance, the manufacturer becomes incredibly more likely to make that specific bullet improperly than any other, for the sole reason that it will be later used to kill your grandfather. It might even come down to an ultra-low-probability quantum fluctuation, in which the bullet suddenly alters course for no apparent physical reason, in order to keep the paradox at bay.
Dubbed the post-selected model, Lloyd's theory is all about trading the impossible for the improbable, which admittedly can cause some very, very unlikely things to happen right around the specific moment where the paradox would otherwise occur. As Charles Bennett of IBM's Watson Research Center explains:
"If you make a slight change in the initial conditions, the paradoxical situation won't happen. That looks like a good thing, but what it means is that if you're very near the paradoxical condition, then slight differences will be extremely amplified."
Incredibly enough, Lloyd and his team say they actually have some experimental evidence of the theory. Though they obviously can't send anything, even a subatomic particle, back in time, they can at least create certain quantum conditions that would closely resemble those experienced by a time traveler. They placed photons in these temporal-like circumstances and then tried to push them towards what were essentially paradoxical situations. The closer they got, the more frequently the experiment failed, and they argue the universe as a whole could function in much the same way when it comes to stopping paradoxes. (This is probably one of those times when you're going to need to read the original paper to understand what they were up to here, because I'll readily admit this is a bit beyond my comprehension.)
In any event, other physicists have met the new theories with great enthusiasm. Todd Burn of the University of Southern California calls it "a nice, consistent loop" and "a really interesting body of work." However, he reminds us that, for now, these aren't much more than clever thought experiments:
"I don't expect these will be tested anytime soon. These are ideas. They're fun to play with."
Looks like we need to find a closed timelike curve. Who's up for a sightseeing trip to the nearest black hole?
[arXiv]