Talks

In quantum field theory it is possible to create negative local energy densities. This would violate the Generalized Second Law (GSL) unless there is some sort of energy condition requiring the negative energy to be counterbalanced by positive energy. TO explore what this energy condition is, I will assume that the GSL holds in semiclassical gravity for all future causal horizons. From CPT symmetry it follows that the time-reverse of the GSL, properly understood, holds for all past causal horizons. These two conditions together then imply that the Averaged Null Energy Condition (ANEC) holds on any null line, i.e. a complete achronal lightlike null geodesic. In curved spacetimes, the ANEC can be violated on general geodesics. But even if the ANEC only holds on null lines, theorems by Sorkin, Penrose and Woolgar, and by Graham and Olum imply that semiclassical gravity should satisfy positivity of energy, topological censorship, and should not admit closed timelike curves. These results can thus be seen as consequences of the GSL. However, these theorems don't apply when gravitational fluctuations are taken into account. In that case, the GSL argument suggests a modification to the ANEC which may make these theorems applicable to perturbative quantum gravity.

The scaling analysis in the large spin limit of Feynman amplitudes for the Bosonic colored group field theory are considered in any dimension starting with dimension 4. By an explicit integration of two colors, we show that the model is positive. This formulation could be useful for the constructive analysis of this type of models.

Spin foam models aim at defining non-perturbative and background independent amplitudes for quantum gravity. In this work, I argue that the dynamics and the geometric properties of spin foam models can be nicely studied using recursion relations. In 3d gravity and in the 4d Ooguri model, the topological invariance leads to recursion relations for the amplitudes. I also derive recursions from the action of holonomy operators on spin network functionals. Their geometric content is discussed in terms of elementary moves on simplices, and is related to the classical constraints of the underlying theories. Interestingly, our recurrence relations apply to any SU(2) invariant symbol. Another method is considered for non-topological objects, and applied to the 10j-symbol which defines the Barrett-Crane spin foam model.