It is the objective of this group to i) discuss the nature of information ii) characterize the variant concepts in different fields and iii) evaluate to conequences for research data management.
In his classical paper More is different, P.W. Anderson (1972) discusses the hierarchical levels of nature. Each level is made of and obeys the laws of its constituents from one level below. All levels are connected by quantitative relations to their lower and higher hierarchical neighbor which links the physical entities of all hierarchies. In simple cases, when the constitutive relations are linear or conceptually readily apparent, we may deduce exact ab initio equations providing material properties as a function of its constituents. In most cases, however, we encounter the formation of complex systems with emergent properties. Typically, there is a branch of science for each level of complexity, since, for each complexity level, there seem to be objects for which it is appropriate and fitting to build up vocabulary for the respective levels of description leading to formation of disciplinary language. It is the central idea of causal emergence (Hoel 2017) that on each level defining a major discipline there is an optimal degree of coarse graining to define those objects in such a way that causality becomes maximal between them.
It is the nature of those objects which defnes the kind of information needed to describe them.
It has been argued that nature exists at a self-organized critical point to harness fluctuations in evolution (Gleiser et al. 2000). Intra and inter level coarse graining is thus reminiscent of renormalization group flow in interaction parameter space leading to universality classes characterized by the same fixpoint. Thus, we may perceive the world around us in progressively more structured entities.
For research data management, there are two major conclusions. First, since experimental data are created by dynamic processes classified by the symmetry of their respective universality class, one expects to find a finite set of spatio-temporal network structures considering all possible data. This is the theoretical foundation to strive for a minimally heterogeneous research data management by considering universality classes of data structures derived from system universality. Second, the discussion of structured information with appropriate measures of complexity opens up new avenues in semantic analysis, In fact, it seems possible to eliminate the qualitative by quantifying it with semantic metrics.
Key to measuring complexity is the amount and especially the kind of information [Floridi 2011, Ay et al. 2020, Klamut et al 2020, Jost 2020, Ellis 2020] different disciplines gather about their objects of study. In Statistical Physics, we find that the classical Shannon entropy measures system order which in equilibrium just equals the heat exchanged with the environment. In cell biology, we realize that each protein carries certain functional information. Cognitive Science asserts how we perceive our environment and control behavior accordingly. Psychology is concerned with the assessment of our perceptions and ensuing actions. Finally, Philosophy builds logical constructs and formulates principles, in effect transforming facts into knowledge.
As a first step, we acquire chairs for the major disciplines physics - chemistry - biology - neurocience - cognitive science/psychology - social science - philosophy
A RDA-WG with a related agenda is certainly https://www.rd-alliance.org/groups/data-granularity-wg