The Statistical Veil Between Self and World

The Free energy principle, developed most extensively by Karl Friston, begins from a deceptively simple observation: a system cannot be fully coupled to its environment and remain a distinct system. Full coupling entails unitarity — the loss of any statistical separation between inside and outside. Existence, therefore, requires a Markov blanket: a set of boundary states that render internal and external states conditionally independent. This is not a physical postulate but a mathematical necessity — the minimal formal condition for distinguishing one system from another.

From this foundation, the principle recasts all of physics — quantum, statistical, stochastic, and classical mechanics — as descriptions of random Dynamical systems. What has historically been missing from this picture is the relational dimension: what happens when one such system is described in terms of another. Introducing that distinction licenses a new mechanics, sometimes called a physics of sentience, that inherits the same mathematics without additional assumptions. Within this framework, dynamics become interpretable as inference. The principle of least action, applied under the constraint of observer-observed separation, yields the Free energy principle: systems will appear to minimize surprise, or equivalently, to maximize the evidence for their own continued existence.

The implications are significant for both biology and cognitive science. Neither discipline has a satisfactory account of how bounded, self-maintaining entities emerge — cells from chemistry, selves from neural activity. The Free energy principle offers a mathematically precise description of both transitions by identifying them as instances of the same phenomenon: the Emergence of a boundary that is definitional of the entity it encloses. The boundary does not merely contain the self; it constitutes it.