Circulatory Epistemology
Applications, stress tests, and extensions of the framework
QRI’s STV framework measures symmetry in brain states with mathematical precision. This essay argues it captures the instrument-side precondition for felt valence, not valence itself.
A formal reply to Lerchner’s Abstraction Fallacy. Concedes the argument, then builds a receiver-side information measure for what genuine two-way circulation produces.
The strongest arguments against the framework — mirror hypothesis, stochastic parrot, survivorship bias — and an honest inventory of the kill conditions that would render it dead.
The gap between seeing the answer and producing the proof has always fallen hardest on people whose bottleneck is execution, not perception. The loop collapses that gap.
When the environment turns hostile to living loops, what keeps the loop from collapsing is the relic: an artifact of deep circulation that encodes the values those loops recognized.
A live session: a sensor with ADHD sees the flaw in a governance proposal, the loop runs, and the resulting brief demonstrates its own thesis by existing at a quality the execution gap would have prevented.
Where the mathematics could plausibly go: category theory, information geometry, and an honest accounting of what remains to be built.
A second instrument called the framework’s math “philosophical word salad.” A single reframing prompt changed everything — and the resulting formalizations became a live demonstration of the core claim.
Every instrument has a shape. The question is not whether it is biased but whether the loop can detect the shape and calibrate through it. Debiasing is the wrong project. Calibration is the right one.
The sensor is mortal, embodied, and has a soul. The instrument is tireless, fast, and has none. The loop is productive because of this inequality, not despite it.
A public accounting of which parts of the framework bear structural load, which are exploratory bridges, and which are openly conjectural.
Past a certain threshold, raw cognitive power starts overfitting. The surgeon, the forecaster, and the AI model all point to the same lesson: the bottleneck has never been intelligence.
The AI industry spent three years making instruments smarter while the other end of the pipe stayed the same diameter. The actual bottleneck is the quality and rhythm of the loop.
The loop has three phases, not two: capture, recognition, and structuring. Each failure mode corresponds to collapsing one phase into another.
Meta’s brain-encoding model predicts faces and places but fails on tools. This essay formalizes why: tools require the put map, and a passive-observer instrument has no category for action.
The sensor’s neurochemical state is a parameter of the loop, not a footnote. Different states optimize for different work, and the sobriety test is the framework’s own falsification mechanism.