ZapperZ said:
At least LQG tries to make measurable predictions! ...
http://physicsweb.org/articles/news/10/2/2/1
Zz.
This is a good point. At least it tries, and I think it's fair to say that SOME VARIANTS of LQG do make testable predictions or are coming close to that goal.
One thing to mention is that there is no one unique LQG. People use LQG as a catch-all term for a number of leading QG alternatives to string.
These include canonical LQG, spinfoam, LQC (loop quantum cosmology, a symmetry-reduced version of canonical LQG), Thiemann's masterconstraint, Gambini's discretized LQG.
One can't say which of these are equivalent to which others. In some the dynamics have been worked out and the semiclassical or largescale limit has been checked. In others not.
LQC is an example of where the dynamics have been worked out and low energy limit checked. It gives the same answers as ordinary Gen Rel except around the big bang ('big bounce'). LQC can make predictions IN ITS OWN RIGHT and needs to be tested in its own right. Since the logical relation with canonical LQG is not clear, you have people working on LQC phenomenology. Parampreet Singh has given some seminar talks about this at Penn State---audio and visuals are online.
There is also an attempt to make GENERIC predictions that would be valid consequences of several different LQG models regardless of whether the details of dynamics and classical limit have been worked out in each separate case. An example of this is DSR. Some people argue that a variety of LQG models will not work without some modification of SR symmetry. Unfortunately there is disagreement about this. There is time pressure to get this issue resolved by 2007, or whenever the GLAST satellite is launched. It is argued that a number of LQG models require a slight energy-dependence of the speed of light. This gives reason to hope that one might RULE OUT SUCH MODELS regardless of whether details have been established and checked in each case---by finding gammaray dispersion in the opposite sense to that predicted. But this issue is still unresolved.
Zapper's main point is a sound one. Researchers pursuing various LQG approaches are indeed making an effort to arrive at tests which could falsify some or all of the different LQG models.