visualization IV.

cohort gradient

the framework's most direct empirical prediction, made visible as a falsifiable shape

The framework asserts a structural prediction: users with earlier age-of-first-exposure to closed-loop platforms retain less $I (τ)$ at long lags, after controlling for total exposure. Older cohorts had a baseline self-evaluative distribution $ν_{u}$ constructed under an architectural regime that included extra-platform inputs; the artisanal mode has something to maintain. Younger cohorts have $ν_{u}$ approximating $π_{\hat{R}}$ at the outset.

Prediction. The framework predicts conditional autocovariance at reflective lags is moderated by developmental cohort: earlier age-of-first-exposure produces lower

I (τ)

even after controlling for total exposure. The solid black curve is the prediction; the shaded band is the framework's hedge on magnitude; the oxblood dashed line is the falsification threshold. The three dotted curves show what confirming, null, and disconfirming data would look like. Hover a cohort age to see the exact predicted

I (60 s)

value.

what to look for

The solid black curve is the framework's prediction, monotonically decreasing from cohort age 25 (~0.7) to cohort age 3 (~0.1). The oxblood dots mark the six cohort bins; hover any one to see the exact predicted value at that age. The shaded paper-85 band is the framework's hedge on magnitude — not a confidence interval in the classical sense, but an acknowledgement that the formal mechanism is asserted, not derived from equations (1)–(5) alone (see §8.2 open questions).

The oxblood dashed horizontal line is the falsification threshold. The three dotted curves are the alternative empirical outcomes: confirming (close to the prediction), null (flat — no cohort effect), and disconfirming (inverted — younger cohorts retain more autocovariance, the opposite of the prediction). Data that produces the null or disconfirming shape would falsify this part of the framework. This is the framework's most direct empirical commitment.

The prediction sits beneath the framework's structural claims, not above them. If the data look flat or inverted, the framework is wrong on the cohort gradient — the rest of the formal apparatus would survive that disconfirmation, but the most testable empirical commitment would not.