Overcoming nonrenormalizability (Q1402286): Difference between revisions

As is well known, lattice space regularizations of Euclidean \(\phi^4\) theories in dimensions \(n> 4\) become trivial in the continuum limit. There is strong evidence from computer studies that the same might be true for \(n= 4\). All told that the conventional path of constructive field theory fails in these cases. In his article, Klauder suggests an unconventional approach to nontriviality. A suitable counterterm in the Euclidean lattice action is combined with several additional procedures to arrange for the required uniform rescaling of all correlation functions. The additive extra term in the action is formed with help of two functions \(A(a)\) and \(B(a)\) of the lattice spacing \(a\), where \(B(a)\) tends to zero when \(a\) does, and \(A(a)\) is determined by the choice of \(B(a)\). Another feature is the introduction of multiple copies leading to a reducible algebra generated by the scalar field at a fixed time. In Klauder's words, ``clearly the entire construction rests heavily on choosing a suitable function \(B(a)\), and then finding the right function \(A(a)\) to go with that choice''. In the last section, Klauder addresses this problem in a much simpler model, the Schrödinger equation with an inverse square potential.