A physical system is synergistic if it cannot be reduced to its constituents. Intuitively this is paraphrased into the common statement that 'the whole is greater than the sum of its parts'. In this manner, many basic elements in combination may give rise to some unexpected collective behavior. A paradigmatic example of such phenomenon is information. Several sources, which are already known individually, may provide some new knowledge when joined together. Here we take the trivial case of discrete random variables and explore whether and how it is possible get more information out of lesser parts. Our approach is inspired by set theory as the fundamental description of part-whole relations. If taken unaltered, synergistic behavior is forbidden by the set theoretical axioms. Indeed, the union of sets cannot contain extra elements not found in any particular one of them. However, random variables are not a perfect analogy of sets. We formalise the distinction, finding a single broken axiom - union/intersection distributivity. Nevertheless, it remains possible to describe information using Venn-type diagrams. We directly connect the existence of synergy to the failure of distributivity for random variables. When compared to the partial information decomposition framework (PID), our technique fully reproduces previous results while resolving the self-contradictions that plagued the field and providing additional constraints on the solutions. This opens the way towards quantifying emergence in large systems.
翻译:暂无翻译