A logic for arguing about probabilities in measure teams (Q2402954)

In this short paper, the authors use sets of assignments, which they call \textit{teams}, and measures on these teams to study the logic of probabilities of first-order formulae. The first section is a short introduction to teams which are a natural way to explicate ways in which values of variables may change (over time). The focus of this paper is on continuous changes of time and thus (at least) continuously large teams. The authors acknowledge that experiments do normally not produce such rich data but point to the fact that variables in the physical sciences (time, pressure, velocity, etc.) are taken to be continuous. Section 2 then formally introduces the pivotal notions: team, measure team and the probability of a sentences \(\phi\) as follows: \([\phi]_X\) is the probability that an assignment \(X\) from a team satisfies \(\phi\). The next section is devoted to setting up the logic as well as semantics and the consequence relation. The deductive system contains first-order logic, the axioms of \(\mathrm{RCF}^*\), three simple axioms and one rule. The fourth and penultimate section demonstrates the fruitfulness of the approach by showing how it can deal with three examples: measurement in quantum theory, homogeneous Markov chains and the Hardy-Weinberg principle in biology. The final section provides the main technical result that the deductive system of this paper is complete with respect to the given semantics for positively bounded theories. The boundedness condition is demonstrated to be necessary. For finite theories matters are smoother: the system is complete and sound.