Sentience and the Origins of Consciousness: From Cartesian Duality to Markovian Monism

In terms of the philosophy of science, perhaps the most tenable way of treating a dual aspect information geometry is under structural realism. We mean this in the sense that the mathematical and geometric form (i.e., structure)—afforded by the mathematical analysis above—allows one to say something about the relationship between (probabilistic) beliefs and the (statistical) physics of internal states that ‘hold’ or ‘represent’ those beliefs. Structural realism takes the pressure off any strong ontological commitments to the mapping between information structures and their content. However, this information structure implies a lawful dependency of probabilistic beliefs (about external states) and parameterised probability distributions (over internal states), in the following sense. Any movement on the internal statistical manifold will necessarily be accompanied by a movement in belief space, as measured by the information length or distance between the beliefs that are parameterised by expected internal states. Furthermore, because these internal states lie upon a statistical manifold of conditional expectations, they must play the role of thermodynamic variables. It follows that belief updating and statistical thermodynamics both supervene on the same internal manifold. Note, the claim here is that physics (i.e., statistical thermodynamics) supervenes on the same statistical manifold as belief updating. This supervenience—on the same statistical manifold—from which both information geometries inherit their structure could be read as the philosophical formulation of the mathematical conjugacy implied by intrinsic and extrinsic information geometries.

In terms of the ontological commitments beyond this structural (realism) argument, any claims would have to be argued much more carefully. It is tenable to associate physics (in the sense of quantum, statistical and classical) mechanics with the intrinsic information geometry. Indeed, this is common parlance in statistical physics [62,64,143]. The more delicate issues arise in terms of commitments to—or interpretation of—the second (extrinsic) sort of information geometry that underwrites Bayesian mechanics. One can avoid any strong ontological commitments here and simply note that should there be any philosophical sentience (i.e., ‘qualia’) in play, they are more likely to be an attribute of belief updating—and therefore part of Bayesian mechanics. We have approached this issue by suggesting Markovian monism entails a gradual difference between non-conscious and conscious entities, and—in this sense—consciousness is a vague concept.

Conceptualization, K.J.F, W.W. and J.A.H.; formal analysis, K.J.F and W.W.; writing—original draft preparation, K.J.F. and W.W.; writing—review and editing, K.J.F, W.W. and J.A.H.; supervision, J.A.H.. All authors have read and agreed to the published version of the manuscript.

K.J.F was funded by a Wellcome Trust Principal Research Fellowship (Ref: 088130/Z/09/Z)