Two quantities describe the aperture's operation from moment to moment. The Markov blanket is where the boundary is. ΔF is how much friction is present across it.
Both quantities come from Karl Friston's free energy principle and the theory of active inference built on it, which supplies the Framework's biological dynamics; the geometry in which they are measured comes from information geometry, and the higher-dimensional field they filter is described by string theory and the holographic principle. The three sources and how they overlap are set out in §2.1. This section introduces the two quantities as mechanism.
The blanket
Every self-organising system, from a cell to a brain, maintains a statistical boundary separating its internal states from the external world. Inputs cross the boundary as sensation; outputs cross it as action. The boundary is what makes the organism a distinct node rather than a smear of the field. In this manual, "the aperture" and "the Markov blanket" refer to the same structure viewed from two angles: the aperture is the blanket's rendering function; the blanket is the aperture's boundary.
Free energy
To stay alive, every such system must resist thermodynamic decay, and it does so by minimising prediction error: the mismatch between what its internal model expects and what actually arrives across the blanket. That mismatch is variational free energy, written ΔF. High ΔF means the model and the world disagree; the organism must spend metabolic energy either updating the model (perception, learning) or changing the world (action) until they agree again.
The interface's top-down priors are the model's expectations projected onto incoming data. Strong priors compress well and act fast; they also filter heavily and, when defended rather than updated, generate friction. Bottom-up telemetry is the raw sensory stream arriving from the field. The balance between the two, how much weight each is given, is called precision weighting, and it is the single dial that most of this manual's procedures adjust.
- Markov blanket
- The statistical boundary separating an organism's internal states from external states; crossed inward by sensation and outward by action. (Friston.)
- Variational free energy (ΔF)
- Prediction error: the divergence between the internal model's expectation and incoming telemetry. Used in this manual as the working measure of internal friction, noise and stress. (Friston's free energy principle.)
- Active inference
- The continuous process by which a node minimises ΔF, either by updating its model to match the world or by acting on the world to match its model.
- Ground truth
- What is actually arriving across the blanket, before defensive priors distort it. Alignment is measured as agreement with ground truth.
Sit or stand anywhere. Do not close your eyes.
- Notice one physical fact about the body: weight on the feet or seat, temperature of the air on the skin. That is telemetry crossing the blanket. Name it silently as body.
- Notice the scene: the room, objects, the sense of being located somewhere in it, and whatever commentary is running. That is the rendered interface. Name it render.
- Ask: what is aware of both? Do not answer in words. Whatever notices the question is not the body and is not the render. Name nothing; rest there for three breaths.
- Return to the room and continue whatever you were doing. Total duration under one minute. Repeat at any point in ordinary activity.
Purpose: to make the stack (§1.4) a perceptual fact rather than a diagram. The exercise is complete when step 3 produces no effort.