Science

Via the Zeitgeister, a fun panel discussion at the Perimeter Institute between Seth Lloyd, Leonard Susskind, Christopher Fuchs and Sir Tony Leggett, moderated by Bob McDonald of CBC Radio’s Quirks & Quarks program. The topic is “The Physics of Information,” and as anyone familiar with the participants might guess, it’s a lively and provocative discussion.

A few of the panel members tried to pin down Seth Lloyd on one of his favorite catchphrases, “The universe is a computer.” I tackled this one myself at one point, at least half-seriously. If the universe is a computer, what is it computing? Its own evolution, apparently, according to the laws of physics. Tony Leggett got right to the heart of the matter, however, by asking “What kind of process would not count as a computer?” To which Lloyd merely answered, “Yeah, good question.” (But he did have a good line — “If the universe is a computer, why isn’t it running Windows?” Insert your own “blue screen of death” joke here.)

So I tried to look up the definition of a “computer.” You can open a standard text on quantum computation, but “computer” doesn’t appear in the index. The dictionary is either unhelpful — “a device that computes” — or too specific — “an electronic device designed to accept data, perform prescribed mathematical and logical operations at high speed, and display the results of these operations.” Wikipedia tells me that a computer is a machine that manipulates data according to a list of instructions. Again, too specific to include this universe, unless you interpret “machine” to mean “object.”

I think the most general definition of “computer” that would be useful is “a system that takes a set of input and deterministically produces a set of output.” The big assumption being that the same input always produces the same output, but I don’t think that’s overly restrictive for our present purposes. In that sense, the laws of physics act as a computer: given some data in the form of an initial configuration, the laws of physics will evolve the configuration into some output in the form of a final configuration. Setting aside the tricky business of wavefunction collapse, you have something like a computer. I suppose you could argue about whether the laws of physics are “the software” or the computer itself, but I think you are revealing the limitations of the metaphor rather than learning something interesting.

But if we take the metaphor at face value, it makes more sense to me to think of the universe as a calculation rather than as a computer. We have input data in the form of the conditions at early times, and the universe has calculated our current state. It could have been very different, with different input data.

And what precise good does it do to think in this way? Yeah, good question. (Which is not to imply that there isn’t an answer.)

We’ve previously celebrated Father Georges-Henri Lemaitre on this very blog, for taking seriously the idea of the Big Bang. His name has come up again in the post expressing thanks for Hubble’s Law — several commenters, including John Farrell, who wrote the book and should know — mentioned that it was actually Lemaitre, not Hubble, who first derived the law. That offered me a chance to haughtily dismiss these folks as being unable to distinguish between a theoretical prediction (Lemaitre was one of the first to understand the equations governing relativistic cosmology) and an observational discovery. But it turns out that Lemaitre did actually look at the data! Shows you how much you should listen to me.

I received an email from Albert Bosma that cleared up the issue a bit. Indeed, it was not just a theoretical prediction (as I had wrongly presumed) — Lemaitre definitely used data to estimate Hubble’s constant in a 1927 paper. He obtained a value of about 625 km/sec/Mpc, not too different from Hubble’s ultimate value. Of course, Lemaitre’s paper was in French, so it might as well be in Martian. Arthur Eddington translated the paper for the Monthly Notices of the Royal Astronomical Society, in 1931, but left out the (one-sentence) discussion of the data!

Here is a slide from Albert containing the original paragraphs. Click for a larger version, and you can compare the French to the English. Thanks to Albert for sending it along.

However — Lemaitre didn’t have very good data (and what he did was partly from Hubble, I gather). And for whatever reason, he did not plot velocity vs. distance. Instead, he seems to have taken the average velocity (which was known since the work of Vesto Slipher to be nonzero) and divided by some estimated average distance! If Hubble’s Law — the linear relation between velocity and distance — is true, that will correctly get you Hubble’s constant, but it’s definitely not enough to establish Hubble’s Law. If you have derived the law theoretically from the principles of general relativity applied to an expanding universe, and are convinced you are correct, maybe all you care about is fixing the value of the one free parameter in your model. But I think it’s still correct to say that credit for Hubble’s Law goes to Hubble — although it’s equally correct to remind people of the crucial role that Lemaitre played in the development of modern cosmology.

Further update: Albert has now sent me more snippets from Lemaitre’s original paper (one, two, three), and the English translation. (All jpg images.)

Last year we gave thanks for the Lagrangian of the Standard Model of Particle Physics. This year, we give thanks for Hubble’s Law, the linear relationship between velocity and distance of faraway galaxies:

v = H₀ d.

(We could be sticklers and call it the “effective velocity as inferred from the cosmological redshift,” but it’s a holiday and we’re in an expansive mood.) Here is the original plot, from Hubble 1929:

And here is a modern version, from Riess, Press and Kirshner 1996 (figure from Ned Wright’s cosmology tutorial):

Note that Hubble’s distance scale goes out to about two million parsecs, whereas the modern one goes out to 500 million parsecs. Note also that Hubble mis-labeled the vertical axis, expressing velocity in units of kilometers, but he discovered the expansion of the universe so we can forgive him. And yes, the link above is to Hubble’s original paper in the Proceedings of the National Academy of Sciences. Only 146 citations! He’d never get tenure these days. (Over 1000 citations for Freedman et al., the final paper from the Hubble Key Project to measure H₀.)

Hubble was helped along in his investigations by Milton Humason; together they wrote a longer follow-up paper. (Some habits don’t change.) Here is a sobering sentence from an article about Humason: “During the period from 1930 until his retirement in 1957, he measured the velocities of 620 galaxies.” These days projects measure millions of velocities. So let’s give thanks for better telescopes, CCD cameras, and software, while we’re at it.

Hubble’s Law is an empirical fact about photons we receive in our telescopes, but it’s implications are profound: the universe is expanding. This discovery marks a seismic shift in how we think about the cosmos, as profound as the Copernican displacement away from the center. It was so important, Einstein felt the need to visit Hubble on Mt. Wilson and check that he wasn’t making any mistakes.

Garrett Lisi has a new paper, “An Exceptionally Simple Theory of Everything.” Many people seem to think that I should have an opinion about it, but I don’t. It’s received a good deal of publicity, in part because of Lisi’s personal story — if you can write an story with lines like “A. Garrett Lisi, a physicist who divides his time between surfing in Maui and teaching snowboarding in Lake Tahoe, has come up with what may be the Grand Unified Theory,” you do it.

The paper seems to involve a novel mix-up between internal symmetries and spacetime symmetries, including adding particles of different spin. This runs against the spirit, if not precisely the letter, of the Coleman-Mandula theorem. Okay, maybe there is a miraculous new way of using loopholes in that theorem to do fun things. But I would be much more likely to invest time trying to understand a paper that was devoted to how we can use such loopholes to mix up bosons and fermions in an unexpected way, and explained clearly why this was possible even though you might initially be skeptical, than in a paper that purports to be a theory of everything and mixes up bosons and fermions so casually.

So I’m sufficiently pessimistic about the prospects for this idea that I’m going to spend my time reading other papers. I could certainly be guessing wrong. But you can’t read every paper, and my own judgment is all I have to go on. Someone who understands this stuff much better than I do will dig into it and report back, and it will all shake out in the end. Science! It works, bitches.

For a discussion that manages to include some physics content, see Bee’s post and the comments at Backreaction.

Monica Dunford, at the US/LHC Blog, has a great metaphor: thinking of the “trigger” in a particle detector as a spam filter. The trigger, you will remember, is the combination of hardware and software that works to separate potentially interesting events from boring old background. If my rough numbers are anywhere near right (experts should chime in if not), the LHC will create about a billion collisions per second, and only about 100 of them will actually get stored on hard disk. Doesn’t sound like much, but we’re talking about a megabyte of data per event, so you’re writing a gigabyte to disk every ten seconds. Just not practical to keep every piece of data, so the trigger makes some snap judgments about what events are fun (like the simulated supersymmetric ATLAS event below) and which are just the usual workings of the Standard Model.

The spam-filter analogy is pretty apt: you’re being deluged with data, and most of it is irrelevant, and you can’t afford to look at all of it individually. So you have to come up with some automated system that decides what to keep and what to toss out. And of course you have exactly the same concern that you would have with any spam filter: the worry that you’re tossing out interesting stuff! You don’t want any job offers to get lost amidst the ads for C!al1$.

A great deal of work, therefore, goes into deciding what the trigger should keep and what it should toss out. Perhaps that helps explain the graph in a previous post of Monica’s:

That would be “meetings as a function of time,” in case it wasn’t obvious. Over 4500 scheduled meetings in 2007, and that’s just for the ATLAS collaboration. The other general-purpose LHC experiment, CMS, has a similar graph, but they only had about 1000 meetings. Whether it is a tribute to their greater efficiency or a mismatch in accounting procedures remains an open question.