Traces, dispersions of states and hidden variables (Q1768573)

The concept of trace has been well studied because of its importance both to matrix theory and operator algebras, and to applications of quantum theory and statistical physics. Here the focus is on the trace and its relation to ``the physically relevant concept of dispersion'' (p. 582), specifically to the minimal dispersion of states on the projections of von Neumann and C*-algebras. The motivation is made clear at the outset. By showing that dispersions of states on von Neumann algebras with zero abelian part are uniformly bounded from below, the author ``excludes the existence of so-called e-hidden variables (measures with arbitrarily small dispersions) proposed by Mackey and considered by Jauch'' (p. 583). The short introduction clearly outlines the argument and motivation for the results of this paper, and presents some preliminary definitions, (albeit with patchy proof-reading). Section~2, ``Dispersions of traces on von Neumann algebras'', contains the main results. First the smallest possible dispersions of states are linked to the normalised trace, and next the main results are established. As promised in the introduction a theorem shows ``that a state \(j\) on a von Neumann \(M\) having neither nonzero abelian part nor Type 12 part is tracial if and only if \(j\) has the smallest possible dispersion on all subalgebras of \(M\) isomorphic to the algebra of all \(3\times3\) complex matrices. An analogous result for weights provides a new characterisation of normal semifinite traces on von Neumann algebras'' (p. 582). The significance is noted after the theorem has been proved: ``Theorem 2.7 demonstrates the difference between quantum and classical measure theory. In the standard probability model the states with the minimal dispersion property are Dirac measures, i.e. pure states. In contrast to this in quantum mechanics, states having the minimal dispersion property are traces, i.e. uniform mixtures of pure states'' (p. 592). Section 3, ``Dispersions and quasi-traces on \(C^*\)-algebras'' takes up the relation between the minimal dispersion property and invariance with respect to unitary groups, seeking to ``improve'' the arguments of von Neumann ``on the non-existence of normal dispersion free states on infinite Type I factors. Motivated by these results we prove that on the \(C^*\)-version of infinite factors there is no quasi-state with the minimal dispersion property'' (p. 594). The paper ends with no further discussion of the significance of these results, but as the author noted in the introduction ``These results deepen and globally explain hitherto known no-go theorems for hidden variables in both \(W^*\)- and \(C^*\)-approach to quantum theory'' (p. 583).