Monday, November 18, 2013

On the fallibility of autobiographical memories

I suspect that most people who have not actively studied the subject have a fairly naive view of how autobiographical memories work. They imagine our memories of events in our lives to be little home movies, tucked away in slots in our brains, ready to be recalled when prompted by circumstance or demanded by our conscious selves.

That this model is entirely inadequate was brought home to me the other day when I was recalling a particular drive I used to take on a weekly routine in my early 20s, when I lived in England. The memory was crisp and detailed, including the undulations of the road, signposts to various villages along the route, and the appearance of the landscape at different times of year. Just one detail was wrong: in my memory, I was driving on the right and sitting in the left-hand seat, as one would in America or Europe; of course, since these events took place in England, the reverse would have been how it actually took place.

As I thought about it more, I realized that the same was true for (almost) all of my memories of England involving cars. In my mind's eye I am opening the wrong door, getting in or out of the wrong side of the car, driving on the wrong side of the road. Only by a conscious effort of will can I restore the image in my head to how reason tells me it actually must have happened.

What this tells us is that autobiographical memories are not really like a movie, but more like the script of a movie. Somewhere in my memory is stored a script about that familiar drive, perhaps associated with a high-fidelity description or image of the landscape, and a slug line such as:

EXT. DAY: CARL drives a late-model compact through winding country roads.

...and from that, like a movie director, my brain constructs the movie, using its library of backdrops and knowledge of what driving involves. Only now that I have spent more years driving on the right than I ever did on the left, it defaults to the more habitual perspective, like an American movie director setting a scene in England and forgetting that they drive on the other side there.

I said "almost" above, for good reason. A few of my English driving memories are correctly oriented, invariably ones involving some amount of drama. For example, once while driving on the M25 I was witness to a multiple-car pile-up that caused a huge back-up on both sides of the motorway. In my memory of that event, I remember every moment with great precision, including which lane I was in and how I avoided being part of the accident myself.

Thursday, October 24, 2013

Time Travel Does Not Work Like That

Whenever a discussion thread turns to time travel, somebody will always raise the challenging question: If you jump forward in time, say by a week, shouldn't you reappear in outer space because the Earth has continued to move and is now seven days further along in its orbit?

And I always feel obliged to answer: Time travel does not work like that.

This conception of time travel typically imagines a time machine as a box of some kind that jumps in time but stays in the same place. And written into that conception are some assumptions that mislead. One is the assumption that you can "jump" discontinuously from one point in time to another. A second is the assumption that movement in time can be considered separately from movement in space. And a third is the implication that the expression "in the same place" is well-defined: as Relativity taught us, we have to ask "the same place relative to what? The Earth? The sun? The center of the galaxy?" (and once we answer that question, we realize that we have pretty much answered our own conundrum.) To properly understand what should happen in time travel, we need to overturn each of those assumptions.

So as far as we know* there is no possibility of "a box that travels through time but doesn't travel through space". It's not just hard, it's an oxymoron -- although in order to understand why, you have to teach yourself to think about spacetime properly, instead of thinking of space and time as separable.

So we have to take a few steps back, discard the presumption that such a box exists, and ask what an actual time machine could be like. And it turns out that (realistic) time travel is a characteristic of a region of spacetime, not a characteristic of a particular machine.

The first thing to understand is that nothing ever moves through space alone or through time alone**. Everything is moving through spacetime, all the time (so to speak). And that movement is always continuous. It simply doesn't make sense to say that a particle X moved from time t1 to time t2 without specifying space coordinates too. The only thing that makes sense is to say that X moved from spacetime position (x1, y1, z1, t1) to (x2, y2, z2, t2). This is so basic that, Relativity shows, even basic concepts like "before", "after", and "simultaneous" are not well defined if you don't specific both time and space coordinates.

The second thing to understand is that physicists have dreamed up lots of ways that time travel might be possible, and all of them involve a continuous path through spacetime, not a mysterious jump. The trick is, the spacetime curves in such a way that when you return to the same coordinates in space, you arrive at a different time from, say, an identical twin who stayed home. However, at no point do you perceive anything "weird" or discontinuous happening. Think of it as being like an extreme version of the famous Twins Paradox from Relativity. In that "paradox" you leave home with a clock that reads 2pm, travel in a big loop at almost the speed of light, come home, and your clock now reads 3pm while an identical clock that stayed home reads 4pm. (This happens all the time, on a smaller scale, right here on Earth: the GPS satellites you rely on for satnav have to correct for the fact that their atomic clocks drift slightly from identical clocks on the surface of the Earth). If the spacetime path you follow is sufficiently warped, when you get home your clock might read 3pm while the one on Earth reads 1pm -- congratulations, you've traveled an hour into your own past!

Of course, we should also note that when you return from your big loop you might be a little surprised to discover that the Earth isn't where you expected it to be, it's actually a little earlier in its orbit from where you would calculate according to the clock that traveled with you, so you have to adjust your path in spacetime a little to navigate back to it. But notice there was still no "jump" involved: you traveled away from Earth, you traveled back to Earth, and found something unexpected, because the paths you and Earth followed through spacetime put your clocks out of sync.

Note also that I pulled a little sleight of hand in that paragraph. When I said "where you expected the Earth to be", I should really specify relative to what for the sentence to have any meaning. And what I really mean is "relative to where you would expect it to be if you had traveled at low speed through flat spacetime and your clock had kept time with Earth clocks". This is kind of a subtle point, but it is at the heart of what's going on here. Your clock tells you where Earth "should" be, and relative to your clock Earth has moved "out from under you" -- but that is only because you assume that you and your clock traveled in flat spacetime. If you could look back with a telescope and watch the Earth the whole time you were on your trip you would see nothing odd happen, other than that it's position drifts slowly more and more backwards from where your clock says it should be. But at no point does Earth blink out of view and reappear further back in its orbit, the way a "time jump" is typically depicted in sci-fi.

So really, when I say "return to the same coordinates", I need to specify relative to what. If I "return to the same coordinates" relative to the Earth, it's right where I expect. If I choose some other point of reference relative to which the Earth is moving, it's not.

By the way, one of the first people to demonstrate rigorously that this kind of thing could happen in General Relativity was Einstein's good friend Kurt Godel. He showed that if the universe is rotating and sufficiently large, you could follow a very long loop around the universe and return to your starting point at an earlier time. Reportedly, Einstein was quite upset by this.

Now, when we see time travel in TV or movies, one way to think about it is to assume that the device is creating its own region of curved spacetime that is extremely small and very severely curved. I like  to pretend that the time travel in 7 Days works this way: the device creates a highly distorted region of spacetime around the capsule. The capsule travels in that highly distorted region which takes it both into space (relative to the Earth) and back in time (again, relative to the Earth). It then emerges from that distorted region and navigates back to Earth traveling in our (relatively) flat spacetime. But rather than "jumping 7 days into the past", I think of it as "traveling 7 days into the past and elsewhere in space".

Another way to  think about it is this: the most consistently realistic depiction of time travel in fiction is wormholes: a wormhole connects two different points in both space and time. If you had a pair of wormholes -- or a single wormhole that connects two points close to each other in space but distant in time -- you would effectively have the classic time machine. And in fact, the wormhole is simply an extreme*** example of curved spacetime. 

I have an even longer answer than this that also explains why we haven't met time travelers from the future yet, but it runs to several pages, and I'm saving it for the book I'm never going to write, "Physics for Smarties: an essential math-free guide for curious arts and humanities students".

Anyway, I hope that helps a little bit, and if you take away nothing else, remember "continuous path through spacetime" and "doesn't travel in space -- relative to what?"


Footnotes


*And yes, I readily concede that "as far as we know" is not very far, but the fundamental character of General Relativity is very, very suggestive on this point.

**OK, you could imagine such a thing if you wanted to, but we have no idea what physics would describe it, so basically you could decide how it behaves completely arbitrarily. It does whatever you want it to do. You just have to bear in mind that there are no absolute coordinates in spacetime, so if you define that your box "doesn't travel through space", it's up to you define what you mean by that: doesn't travel relative to what frame of reference? In short, your time box does whatever you choose it to do, because you are making it up.

***Insert your own Wormhole Extreme! reference here.