nucleation

THEORY-1097: A Markov blanket, in any formulation including sparse coupling at nonequilibrium steady state and time-varying blanket states, does not individuate a system: where it falls is fixed by modelling choices made before it is found, so it presupposes the boundary it is used to find, and it cannot represent a boundary the system produces

1 min read

Properties4
StatusProposed
Tagsindividuation, philosophy-of-biology, philosophy-of-science, probabilistic-modeling, complex-systems
SourceLIT-603 — The Emperor's New Markov Blankets, LIT-598 — The Markov blanket trick: On the scope of the free energy principle and active inference, LIT-526 — Life as we know it
ExtendsTHEORY-073 — In Friston's 2013 formulation and the applications the record holds, a Markov blanket does not individuate a system: where it falls is fixed by modelling choices made before it is traced (the coupling graph, the internal set and the time at which steady state is assumed), and a blanket read from a fixed graph is the same whether the system produced the boundary or it was imposed

Graph View

Source

The same three works as THEORY-073: Bruineberg, Dołęga, Dewhurst and Baltieri (LIT-603, read in NOTE-447), Raja, Valluri, Baggs, Chemero and Anderson (LIT-598, read in NOTE-450), and Friston (2013, LIT-526, read in NOTE-421).

The claim

THEORY-073, as narrowed, holds the account for Friston’s 2013 formulation and the applications the record holds. Both critiques state it without that restriction: a Markov blanket, however defined, is placed by choices the modeller makes first. The later formulations are the free-energy literature’s reply. They are the sparse-coupling definition at nonequilibrium steady state (Friston 2019 and later) and time-varying blanket states (Ramstead et al. 2020). The record holds neither, so the general form is offered and not held.

What it does not say

That the later formulations fail. It says that, if the critiques are right about the source of the partition, they fail in the same way. Whether they do is the promote_when’s question.

Backlinks