Temporal factorisation: A unifying principle for dynamics of the world and of mental states

Abstract

Dynamics play a major role in many phenomena, addressed in a variety of disciplines.

Adaptive Network Modeling for Criterial Causation

“…Criterial causation as introduced by Tse [25] describes how as a form of plasticity, in the brain, configurations emerge that provide informational criteria for future processing and behaviour. In the current paper, first it was shown how this notion relates to the more general notion of temporal factorisation based on mediating state properties to describe the world's dynamics as introduced in [17]. The core of both of these two notions is that by some adaptive process, over time a (past) brain or world pattern leads to the formation of a present brain or world configuration that in turn drives the (future) brain or world pattern.…”

“…The temporal factorisation principle [17,18] states that any systematic temporal 'past pattern implies future pattern' relationship a ) b between a past pattern a and a future pattern b can be factorised in the form of two temporal relationships a ) p and p ) b for some state property p (called mediating state property) of the present world state; see Fig. 1, left hand side.…”

“…The temporal factorisation principle claims that the present world state configuration p encodes sufficient information so that the world can forget about the temporal pattern a in the past if it makes temporal pattern b occur in the future; therefore it essentially is a claim that world state configurations are sufficiently rich to encode all (future-)relevant information on the past (which in theory could concern an almost infinite number of world states, with their temporal relations) in some condensed form in one state configuration. In [17] it is discussed in more detail how this principle relates to views on the world's dynamics from Descartes [8], Laplace [13], Ashby [2] and van Gelder and Port [27]. Descartes [8] puts forward that systematic relationships (laws of nature) exist for world states over time, in the sense that past world states imply future world states (called the clockwork universe).…”

“…It only applies to aspects of the world that happen to be deterministic (as expressed by the conditional a ) b). In [17,18], the temporal factorisation principle is modeled in a formalised manner, and by many examples it is shown how the temporal factorisation principle plays its role in the world's dynamics, taken from Physics and from Cognitive Science.…”

“…This paper presents a computational adaptive network model that makes the above more precise, and illustrates it by a simulation for an example scenario. Moreover, the paper shows how the perspective as sketched can be considered a special case of the more general perspective on the dynamics of the world based on temporal factorisation by mediating state properties [17,18]. Mediating state properties on the one hand encode in the present world configuration, information on the past pattern in the world, and on the other hand they determine the possible future patterns for the world from there.…”