Non-perturbative aspects in quantum field theory: Proceedings of Les Houches Winter Advanced Study Institute,March 1978

Until recent years field theories were only studied from standard perturbation series. These are equivalent to small fluctuation expansions around the classical ground state configurations in which the fields are space-time independent. It has been gradually realized that this procedure may miss some crucial physical features of the theory. In particular the absence of free quarks cannot be explained in quantum chromodynamics from this perturbative viewpoint. One is led to deal with the non-linearities in an essential way. This issue aims at covering the recent advances in this direction typically since 1975 and the issue 23C of Physics Reports.One way of going beyond perturbation theory is to make use of non-perturbative classical solutions which are in general very difficult to obtain. In Yang-Mills theories remarkable advances have been made in the search for self-dual solutions. They are described in the lectures (I, II) by E. Corrigan and R. Stora et al. The quantum meaning of these solutions and the technical problems raised by quantizing fluctuations around them are discussed by J.L. Gervais (III) and D. Gross (IV). The most recent topics of this subject deal with imaginary time solutions, the so-called instantons, which describe quantum tunneling in the semi-classical approximation. Other field configurations such as merons, are also discussed in relation with quark confinement. An attempt to build a general picture of strong interactions on these grounds is displayed in (IV).Recent advances concerning the solitons which are real time, classical Minkowski solutions, are discussed by D. Olive (V). These solutions describe generalized Dirac magnetic monopoles in non-Abelian gauge theories. The resulting model contains confined magnetic charges and unconfined electric charges. If the magnetic and elastic characters can be interchanged by a dual transformation, as discussed by F. Englert et al. in (VI), this would provide a model for quark confinement. This duality transformation has been spelled out in (IX) by C. Itzykson who proves the existence of a confinement phase transition in a lattice model with Z₂ gauge symmetry.Another non-linear feature of Yang-Mills theory is that no single gauge condition can be implemented over large field fluctuations. This is discussed in (VII) by Sciuto and also in (III).Another mechanism of confinement is proposed by McCoy and Wu (VIII) in which the propagator has a cut instead of a pole, as it does happen in the two-dimensional Ising model.The lectures of J. Zinn-Justin and G. Parisi, (X) and (XI) deal with the asymptotic estimation of the large order behaviour of perturbation theory, which one can obtain by semi-classical methods. The use of these methods is either to provide a way of improving practical perturbative calculations, or to characterize possible ambiguities due to vacuum tunneling. Additional problems raised in renormalizable theories are discussed in (XI).Another non-perturbative approach is to examine the large N limit of an SU(N) gauge theory. At leading order, only planar Feynman diagrams contribute. They are discussed in (XII) by E. Brézin. There are very subtle questions about the corresponding exact summation in two dimensions which are studied by T.T. Wu in (XV).In two-dimensional field theories exact solutions for the S-matrix have been recently discovered. They are discussed by Karowski in (XIII). In connection with this problem, Lüscher (XIV) has shown the existence of non-linear conserved charges which ensure the absence of particle production. These models are particularly interesting since the mass spectrum cannot be obtained from coupling constant expansion.Finally several related topics in statistical physics are discussed. D. Nelson (XVI) reviews the beautiful properties of the two-dimensional X-Y model which is an ideal example of mechanisms invoked in hadronic field theories. A. Luther (XVII) discussed an attempt to extend to a higher number of dimensions the bosonization of fermion theories which is so powerful in two dimensions. G. Toulouse (XVIII) describes modern ideas in spin glass phase transitions, a possible testing ground for gauge field theories besides its obvious intrinsic interest.