Thursday, March 01, 2007

Quantum Gravity in the Lab

On Tuesday, we had the first meeting of the discussion group on 'Quantum Gravity in the Lab' that I've set up here at PI. As I've mentioned earlier, I exported the topic to a blog on its own. This is a copy of the entry from the first meeting. Comments are off here, because I want to encourage you to leave comments over there.

I had the intention to briefly summarize all the good reasons for the discussion group, but obviously I forgot half of what I wanted to say. So, if you're reading this you definitely have an advantage.

The reason for me setting up the discussion group was that even for those who are smart enough not to work on it, quantum gravity is the holy grail of theoretical physics in the 21st century. And here we are, standing on the shoulders of giants, trying to make a step without falling down. Luckily, over the last decade, we were able to move the giants a bit, and it's become quite fashionable to work on not derived, but well motivated models that incorporate the one or the other feature of the pursued full theory: like a minimal length scale, extra dimensions, modified dispersion relations, decoherence in a Planck-scale foamy background etc.

However, when I read papers about phenomenological implications of quantum gravity, they most often start with sentences like 'the recently proposed model...' or, 'Recently, it has been shown...', and I'm wondering if there's a decay time for the word 'recently'. E.g. the idea of having the 'true' Planck scale around a TeV dates back about a decade; DSR in it's present version came up around 00/01; early ideas about curved momentum spaces, modified dispersion relations etc. date back far longer.

Yes, it's a new field. But there are more and more people working on it every year, and it's about time that we get an overview about who is doing what, where progress has been made, and what problems have been faced. Part of this discussion group's intention is to summarize these efforts and see how promising they are (though this is surely a subjective judgement that I don't remotely mean to make objective).

The other part of the reason for the meetings is that I, being a particle physicist, find myself a bit lost in the quantum gravity seminars. It is clear that most of these approaches are very strong on the theoretical side, but lack some application. And I repeatedly ask myself if one couldn't investigate this or that feature. So, I would really like to bridge the gap between theory and experiment, to hear your ideas - however weird they might be - how one could test quantum gravity. (On my desk, if possible.)

Roughly speaking, the discussion group is therefore the forum for you to ask all the really stupid questions, and to spit out all the half-cooked topics that you have in mind. I deliberately have not scheduled speakers for the whole term. I want to see what issues come up and be flexible to focus on those that attract the most interest.

In the first meeting, I want to outline what we're going to talk about. Not so much because I don't like vivid discussions, but mostly because I'm not a good philosopher. It's not only that I constantly mix up words that end on -ism, but I generally don't like to argue about words whose meaning hasn't been clarified. So, to begin with, we should try to figure out what we mean with 'quantum gravity' and it's 'phenomenology'.

To me quantum gravity is the try to resolve the apparent disagreement between general relativity and quantum field theory. I say 'apparent' because, well, I'm pretty sure nature knows how to deal with that, so there ought to be a solution to the problem. Since three out of the four interactions in the standard model are quantized, it's natural to expect that also the fourth one - gravity - has to be quantized. In the meaning that the metric is no longer a classical field.

However, I think one should keep in mind that we actually don't know whether gravity is quantized at all. That doesn't solve the problem how it goes along with quantum field theory, but it's nevertheless a possibility. In this case one has a semi-classical theory with an unquantized gravitational field. The question is then how quantum field theory couples to the classical field. This comes with all the issues about renormalization, the vacuum, and as Rafael points out one has to think about what happens to the quantum nature (determinism) of the theory. (This is a topic I've also scratched in the comments at CV here.) Yes, I'm not claiming I know how to deal with that, I just don't a priory see why this is a totally inconsistent try. I don't think anybody knows what happens to the gravitational field of a 'collapsing' wave-function, so there's definitely room to play around there. So, apart from the quantization of the metric, one has the question what happens to the quantization itself.

Further, if one thinks about quantum gravity as stuff that happens when the curvature gets strong, one also has to consider modifications of gravity itself. These might, or might not be motivated by a quantization approach. Issues like singularity avoidance, the cosmological constant, dark matter etc could fall in this range.

One way or the other, as the status is right know, if one starts attacking quantum gravity and wants to investigate phenomenological implications, one makes some more or less motivated approximations, and cooks up an effective model that one can use to make a prediction. I want to point out (if you've read my recent paper you know why) that besides making predictions for new experiments, one shouldn't forget to reproduce all the old and boring stuff in the standard model. You know. Like, apples falling to the ground, photon self-energy, nuclear decay. These things.

The important question is then, what information got lost in the intermediate steps from the 'full theory' to the prediction. That is, we have kind of an inverse problem. Given that we were lucky enough to actually measure something, what could we learn about the full theory? Therefore, questions we should ask about all of the proposed phenomenological approaches are

  • How stringent is the top-down approach (derivation/motivation)?
  • How good works the bottom-up connection (reproduces standard model)?
  • Exactly what is predicted under which assumptions?
  • What could be learned from the outcome of the experiment (positive as well as negative)?

Lee mentions that I'm being too strict and since everything is going so badly we should try something unusual and not set too many requirements to our models. Well, for one I don't think everything is going badly, but more importantly, I haven't set any requirements. These are just the points I think one should investigate about the presently discussed approaches. See, this is just great about being at PI. Here, it's me who's considered being too conservative.

This is obviously a completely biased summary of the meeting, and I've dropped every comment that I didn't like or didn't understand.