Saturday, October 31, 2009

The dark side of the landscape

In the previous post I had presented the multiverse in a way that made it look almost innocuous. As I have said a few times in this blog I had heard about how the landscape (existence of a large number of vacua) in string theory made it unavailable to make predictions.

Despite it the actual articles I had read didn't give to me that impression, so I suspected that I was missing something, the problem is that I didn't know what. Reading a recent entry in Lubos blog titled A small Hodge three-generation Calabi-Yau I faced again that problem of missing information. So I reared once again the KKTLT paper and I searched for bibliography that would give me some cloud.

At last I was lead to the correct paper, The statistics of string/M theory vacua by Michael R. Douglas (he is not the actor, of course). The abstract of the paper says all it:

We discuss systematic approaches to the classification of string/M theory vacua, and physical
questions this might help us resolve. To this end, we initiate the study of ensembles of
effective Lagrangians, which can be used to precisely study the predictive power of string
theory, and in simple examples can lead to universality results. Using these ideas, we outline
an approach to estimating the number of vacua of string/M theory which can realize
the Standard Model.

I still haven't finished to read the paper, but the image is clear. Yes, one can have a chaotic/eternal inflation scenario that creates an infinite of universes, or one can go from one to another thought some kind of CDL (Coleman de Luccia) or Hawkings instantons among deSitter vacua or whatever mechanism to create an universe for whatever vacua of the one available in string theory. An yes, every new universe would have an smaller cosmological constant that the previous one. In that way one has an universe with the small cosmological constant (cc) observed in ours. The anthropic principle (or ideology as prefer to name it Lubos) says that in universes with large cc there are no observers so it is not that bizarre that we observe such an small cc, despite the fact that naturally theories with broken supersymmetry would have a big one to begin with.

The real problem is that in that paper is argued that even with the restrictions of an small cc and the observed gauge content (the standard model one) one still has a large number of solutions with the values of the coupling constants, masses of the particles and etc. in the observed margin of the standard model. I have made quick search in the paper to see if it was here where it appeared the famous $10^{500}$ but I couldn't find it (the search feature of acrobat seems to not work with math expressions) but in the text appear ofthem $10^{100 }~ 10^{400}$ so it is in the right order of magnitude. I have intention of reading this paper soon, as well as another by Kallosh and Linde, Landscape, the scale of SUSY breaking, and inlation

It is not that I like the idea of the landscape, I don't, and that's why I hadn't found sooner this papers and I had searched other lines of investigations, such as the ones mentioned in this blog. But like it seems that cosmology is a such a hot topic nowadays, mainly because the large amount of data available, I think it is a good idea to know this kind of things in some detail.

As I had said previously, in other blog entries, I was aware that there were some concrete approaches that tried to disprove the landscape, understood in the sense presented here-that is, too many vacuums compatible with the standard model, not just too many vacuums compatible with an small cosmological constant-. Some of that papers are The String Landscape and the Swampland discussed by Lubos here and also discussed by Distler in his entry YOU CAN’T ALWAYS GET WHAT YOU WANT. More entries in Lubos blog discussing papers against the (SM)landscape are Ooguri and Vafa's swampland conjectures. He also has a paper with C. Vafa and Nima Arkani-Hamed titled The String Landscape, Black Holes and Gravity as the Weakest Force. I think that I have seen a blog entry of him about that paper, but it doesn't appear in the trackback for some reason.

Well, I leave this entry as a loosely discussed bibliography of the real problems of the landscape ideology. As I said there are possibly good reasons to expect a good "vacuum selection method" as M. Douglas calls it and so one wouldn't care too much about it. Possible the LHC could give a cloud of it. It is good to know that beams are beginning to circulate in it again, at least partially, and thatvry soon-if everything goes ok-it will be giving data.


Anonymous said...

Este es el texto más extenso que he encontrado sobre la teoría F.

¿Podrías tú (ya que sabes inglés) y conoces el tema y la terminología traducirlo?


Anonymous said...

Este es el texto más extenso que he encontrado sobre la teoría F.

¿Podrías tú (ya que sabes inglés) y conoces el tema y la terminología traducirlo?


Javier said...

Hola anonymous. No estoy seguro si es la primera vez que comentas algo aqui o eres elque pregunto hace un tiempo si podia hacer una entrada en español, sore teoria F. Si es así decir que no he olvidado esa propuesta, pero, como puedes ver, ahora he andado mirando otras cosas y tenia la teoria F momentaneamente aparcada.

Conozco esa entrada del blog de Distler, hace referencia al primero de los papers que salió de las F-theory GUT's.

La ve3rdad es que lo quecomenta ahi Distler es interesante, perosolo si ya sabes mucho sobre el asunto tratado. Además ese artícuoes astante formal y las aplicaciones del formalismo surgieron lueg, en los restantes papers.

Sabiendo eso te pregunto si sigues interesado en que la traduzca literalmente. Tambien seriaa interesante saber tu nivel general en fisica, y sobre cuerdas en particular, para intentar adaptar la entrda a ese nivel.

Grcias por tu interés.

Anonymous said...

Sí, soy la misma persona. A falta de otro material ya iría bien una traducción de ese enlace. Menos es nada. Eso no quita que cuando dispongas de más tiempo e información nos puedas deleitar con otro texto más explicativo para los que no estamos duchos en física. En todo caso: gracias.

Lumo said...

Hola Javier,

just a vaguely related comment, concerning an issue you nicely raised on other blogs.

You suggested that the F-theory and maybe even type IIA-based braneworlds are much more strongly falsifiable, and one could reduce the options to the heterotic world after a few measurements of SUSY in our world.

I think yours is a kind of idealized truth. However, the real world of string theory is more complex.

The stringy 4D vacua reduce to effective low-energy field theories that cover a big portion of the low-energy EFT parameter space. So unless we're able to calculate the accurate values which are more sharply predicted by the (discrete) stringy vacua, the frameworks of EFT and string theory are "operationally" equivalent.

However, they still provide us with wildly different optics how to look at the probability distributions of different measures. EFT wants you to think about a quasi-uniform distribution on the low-energy couplings and masses.

The anthropic principle wants you to look at "quasi-democratic" probabilities for different discrete stringy vacua, which prefers the regions of the low-energy parameter space with a high density of stringy vacua.

That's what some people think is more naturally implied by string theory.

However, non-landscape constructions such as Vafa-Heckman also deviate from the low-energy effective field theory measure. But in some sense, they deviate in the opposite direction than the anthropic principle. They favor very special vacua with big symmetries, intense decoupling of various portions of physics, and so forth.

Clearly, none of these two philosophies - Vafa and/or anthropic - can actually be derived from string theory. They're philosophies independent of string theory. Unlike the case of string theory itself, we don't really know which of them is correct.

There are clear advantages of each: the advantage of the anthropic principle is that its vacua are flexible while the disadvantage are that they're flexible (and not predictive).

Similarly, the more unique vacua of F-theory bottom-up phenomenology are predictive, which is nice, but the disadvantage is that they're predictive (and can therefore be falsified pretty soon haha).

Clearly, the truth may combine these two philosophies and others in ways that are not fully known today. When people run into trouble and falsify the simplest Vafa-Heckman decoupled theories, they will relax some assumptions, generalize the class, and try additional ones.

On the opposite side, the anthropic people may try to look for some features of the scenarios/classes of vacua that will focus their attention.

This progress from both sides - special and anthropic - may continue for a while, and both sides may eventually agree on the right model. And they may find some new principle - which is neither anthropic nor special - that picks the right vacuum from a rather special class.

At any rate, you correctly identified which classes of the vacua are special. F-theory is perhaps the most predictive ones, and the IIA braneworlds are also pretty predictive.

Heterotic string or Horava-Witten theories are slightly less predictive, but maybe more so than the generic KKLT IIB flux vacua. Heterotic strings are closest to the effective field theory approach, in a sense, but it's partly because we aren't able to calculate their detailed features terribly well (except for Yukawa couplings which already show a more detailed knowledge than pure effective QFT with random couplings).

I wouldn't say that F-theory bottom-up models may be "completely" falsified because what the class really contains is a somewhat ill-defined question and people's creativity may always surprise us.

Best wishes

Javier said...

Hi Lubos, wellcome here again and thanks a lot for your clarifications :).

For the readers of this blog just to inform that Lubos is doing a clarification that I solicited in the last part of an answer that I gave in this entry (see comment 22) of the antistring blog "not even wrong".