New Scientist must hire someone to trawl through the arXiv in the hopes of getting the science news one step ahead of everyone else. Unfortunately, their record for distinguishing good science from bad science is not all that good, so I was pretty skeptical when I was pointed to an article on a new theory of everything™.
This paper is actually a very impressive piece of work, though it is thoroughly overhyped and completely un-understandable to anyone who doesn't regularly do particle physics and group theory. The author, A. Garrett Lisi, has proposed a back-to-the-future approach to uniting quantum mechanics and gravity. Throughout the 20th century (and late 19th century) it was discovered that the electric force, the magnetic force, the weak force (responsible for radioactivity), and the strong force (responsible for holding the nucleus together) could all be described by a single theory. The different forces, their properties and associated particles could all be obtained from different symmetry operations (think rotations and reflections) of an algebraic system. This very successful approach has withstood the test of time, with absolutely every experimental test falling within error bars of the calculated results. However, gravity stands apart as the force which does not get included in this set and its inclusion (or a totally new theory) would constitute a theory of everything.
Much of the early work focused on exploring higher symmetry algebraic systems that might include gravity. Several were found, but none actually survived contact with reality. This approach has largely fallen out of favor because any object with sufficient symmetry operations can be made to unite gravity with everything else while still not agree with reality as we measure it. Lisi has revived this approach by looking at the shadows cast by an extremely complicated symmetry group (called E8). Unsurprisingly, if you choose (by hand) the right starting methodology and ignore a few possible problems with units, a selection of symmetry operations will result in symmetry subgroups that correspond to those from particle physics as we know it, something that might be the symmetry operations of gravity, and some other stuff.
There are a couple of reasons to be skeptical of Lisi's work. First, he has a problem with units at the very start of the paper, which could cascade through the whole work. It is not yet clear that this problem is fatal, but at this point we just don't know. He has found that the chosen symmetry operations correspond to the symmetry groups of particles—not that surprising, considering the number of symmetry operations he has at his disposal—but at this point, he can't calculate the properties of predicted particles (really, these are just degrees of freedom in the symmetry class), because that requires what look to be extraordinarily difficult calculations. If the unit problem is serious, this will get worse, meaning no predictions at all.
In the Observatory thread on this topic, posters have pointed to a blog article that puts Lisi firmly in the crank category, which I initially agreed with but I have since been persuaded that I was being too hasty. I still think he is probably wrong (because of his early problem with units) and that his predicted particles are not predictions at all since he can't actually tell us where to look for them. In fact nothing beyond their symmetry group is predicted by Lisi at this point. If he can answer the question over units early in the paper, find a way to put masses on his predicted particles, and show that the predicted half-life of the proton is sufficiently long, then his theory will start to look like a winner.
Personally, I dislike New Scientist's arXiv trawling ways; by putting Lisi into the role of the anti-establishment hero (he's a surfer not a scientist...) and publicizing his work before peer review, they (along other media reporting, including our own) may well have compromised Lisi's chance at getting a fair shot a peer review.
Many thanks to Geon and his post in the Observatory.
reader comments