Cognitive Mechanics as a Systems Architecture: Load, Timing, and Persuadability

Question

@kungfufk

STATUS: SAMPLE / ILLUSTRATIVE MODEL — NOT EMPIRICALLY VALIDATED
This extends the wave/energy model and the bounded-rationality piece by re-expressing the same math in computing/cybersecurity vocabulary. It still predicts behavior, not truth or belief content. Every free parameter is still unfit. Renaming variables does not add evidence.

Scope note: This document describes why the model predicts that load and timing correlate with persuadability, in the same way the source documents describe why contradiction produces disturbance. It does not include a targeting methodology — no vulnerability-selection step, no procedure for timing an attack against a specific person, no escalation-when-resistant mechanic, no delusion-exploitation case. Part 2 exists specifically to explain why that layer is excluded and how to recognize it if someone else is using it on you.

---

Part 1 — Recasting the Model as a Computing System

1.0 Why recast at all

The wave/energy model already describes a bounded-capacity system doing time-limited processing on multiple concurrent inputs. That's structurally similar to a computer handling concurrent I/O under a fixed CPU/RAM budget. Recasting it doesn't add predictive power — it's a translation layer, useful if bandwidth/latency/thrashing language is more intuitive to reason with than sine waves. Nothing here is new physics or new psychology; it's the same equations from the source documents, relabeled.

1.1 Component mapping

Wave-model term	Systems-architecture analog	Role
Active traits $K, Em, Tr, At, Cu$	Input channels / concurrent data streams	What the system is processing right now
Confidence $Cf$	Output / status register	A readable value, checkable against a computed expected value
Fatigue $Fa$	Thermal throttling	Reduces effective clock speed the longer/harder the system runs
Memory Load $ML$	Working buffer occupancy	Consumes the same bounded RAM/cache the active traits need
Capacity $C(t)$	Total compute budget (cycles + RAM) available this moment	The shared ceiling
$E_{cognitive}(t)$	Current load	Sum of what's actively being processed
$R(t) = E/C$	CPU utilization	$R>1$ = thrashing (system is oversubscribed)
$N_{cycles}$	Processing cycles to resolve a decision	Same quantity as the source model, renamed
Incoming information $\Phi_{info}$	Incoming packet / request	New data arriving from outside
Interference $D(t)$	Write conflict between new packet and cached state	Cost of reconciling two disagreeing values
Trust wave $\Psi_{Tr}$	Cached session state	What's already "committed" that new input gets checked against

1.2 Equations, restated in this vocabulary

Utilization (same as the source model's load ratio):

$$R(t) = \frac{E_{cognitive}(t)}{C(t)}, \qquad C(t) = C_0 - \gamma_{Fa}A_{Fa}(t) - \gamma_{ML}A_{ML}(t)$$

$R \ll 1$: idle capacity, full deliberate processing available. $R \approx 1$: system is running near its throttled ceiling — this is the "bounded rationality" regime. $R > 1$: forced degradation — the system sheds work by falling back to cheaper (heuristic) processing, the cognitive equivalent of a server shedding load under an oversubscribed request queue.

Processing latency:

$$N_{cycles}(t) \approx \frac{N_0\,(1+R(t))^{\beta}}{A_K(t)}$$

More cached knowledge ($A_K$) lowers cycles needed to resolve a decision; higher utilization raises it — the same trade-off as a cache hit vs. a cache miss under load.

Write conflict (interference) between new input and cached state:

$$D(t) = A_{info}\cdot A_{Tr}(t)\cdot\bigl[1-\cos(\phi_{info}-\phi_{Tr})\bigr]$$

Low disagreement between new packet and cached state → near-zero conflict cost, resolves cheaply. High disagreement → maximal conflict cost, expensive to reconcile (more cycles, more risk of an inconsistent state until resolved).

1.3 What a "conflict" predicts — and what it doesn't

The model treats a write conflict symmetrically: it's a property of how far apart two data streams are, not of who sent which one or why. It doesn't distinguish "new evidence from a reliable source" from "an attempt to manipulate" — both look identical to the equation as a disagreement between $\Phi_{info}$ and $\Psi_{Tr}$. That's consistent with the source model's own Prediction 3 and Prediction 7: it describes what a conflict costs to resolve, not how to construct one on purpose against a target. The math has no "attacker" role built into it — adding one is a design decision, not something the equations imply.

---

Part 2 — Why Load and Timing Correlate With Persuadability, and Where This Stops

2.1 The real psychology behind the correlation

The reason $R(t)$ (utilization) plausibly correlates with persuadability isn't new to this model — it restates three well-established findings:

• Dual-process theory (Kahneman): fast, heuristic processing dominates when deliberate processing capacity is scarce — under fatigue, time pressure, or high concurrent load. High $R(t)$ is the model's stand-in for "capacity scarce."
• Elaboration Likelihood Model (Petty & Cacioppo): when motivation or capacity to process an argument carefully is low, peripheral cues (authority, emotion, repetition) carry more persuasive weight than the argument's actual quality. This is the same mechanism as $Cf_{predicted}$ dropping as $R(t)$ rises in the source model.
• Cognitive dissonance theory (Festinger): holding two contradictory cognitions at once is uncomfortable and motivates some resolution — but the theory does not say resolution goes in a chosen direction. People under contradiction often entrench their original belief further rather than update it (the backfire effect / attitude polarization is itself well documented).

All three are real research areas. The wave/energy notation doesn't add predictive power beyond what's already established in this literature — it's a different notation for the same correlation.

2.2 Naming the pattern: Cognitive Dissonance (as one theory-application, not a general method)

The wave model's basic claim, stripped to Festinger's actual theory: holding two contradictory cognitions at once creates uncomfortable tension, and that tension motivates some resolution — but not necessarily toward truth. It can resolve by:

• updating the belief (rational correction), or
• rejecting the new input and entrenching harder (backfire effect), or
• distorting one of the two inputs until they seem to agree (rationalization)

This is the one place the "wave" framing is legitimate to use as illustration, because dissonance really is a tension between two signals that must be reconciled somehow — that's genuinely wave-like as a metaphor. It's intentionally not generalized here to the other trait-waves (trust, emotion, knowledge) as manipulable targets, because that's the part that turns description into a targeting method. Dissonance-as-signal-conflict is fine as a sample application precisely because the thing being acted on is your own resolution process, not someone else's.

2.3 The solution: Kalman filtering as a personal mental model

Cognitively, dissonance is what constructive and destructive interference look like when they arrive within the same short interval — agreement reinforcing your existing state, contradiction canceling against it — and a Kalman filter is precisely the mechanism for weighting two such signals instead of letting either one win outright.

This is a genuinely good metaphor, and it works better than "just be skeptical."

A Kalman filter maintains two things at once: a prediction (your current belief, plus an estimate of how uncertain it is) and an incoming measurement (new, possibly noisy information). It doesn't fully trust either one. It computes a gain — how much to weight the new measurement vs. your existing estimate — based on how noisy the new input is and how uncertain your prior belief already was. Then it updates smoothly, rather than jumping fully to the new input or ignoring it.

Applied to yourself under dissonance pressure:

• Don't treat every contradictory input as equally weighted evidence. Ask: how noisy/unreliable is this specific new information, and how urgent is the person making it feel? Manufactured urgency is itself noise, not signal — a real update rarely needs to happen in the next ten minutes.
• Keep an explicit uncertainty estimate on your own belief, rather than either full confidence or panic. "I'm fairly confident, but I know my error bars are wide when I'm tired" is the mental equivalent of the filter's covariance term.
• Update proportionally, not maximally. A single contradictory data point should nudge your belief a little, not flip it entirely — a real Kalman gain is rarely 1.0. If someone is pushing you toward a full, immediate flip, that pressure itself is diagnostic of a manipulation attempt rather than of you having encountered decisive new evidence.
• High "process noise" (fatigue, exhaustion, hypervigilance) should make you trust the filter less, not more. When you know you're running in a degraded state, the correct move is to widen your uncertainty and defer the update — literally "wait until I've slept" — not to let a single urgent contradictory input dominate the estimate.
• Get an outside "sensor." A second observer (a friend, therapist, someone uninvolved) is a second measurement source. Real Kalman filters get more accurate with more independent sensors; a single noisy channel talking to you under time pressure is the worst-case input, not the best.

That's the actual defensive skill: not detecting the wave-math structure of an attack, but noticing when you're being asked to update at gain = 1.0, under artificial time pressure, from a single uncorroborated source, while your own error bars are unusually wide. All five of those are things you can check about your own state without needing to model anyone else's internal variables at all — which is also why this version doesn't require, or benefit from, any targeting layer to work.

2.4 Why "engineer it against a specific person" doesn't survive contact with reality, and why it's harmful regardless

• The model assumes clean, legible variables — a known $\phi_{Tr}$, a known $A_{Tr}$. In reality you cannot read someone's internal trust-phase; belief systems aren't observable at that resolution from outside.
• Because dissonance resolution doesn't have a guaranteed direction, deliberately timed contradiction under stress more often produces defensive entrenchment or a stress response than a clean "flip" — the technique the equations might seem to license doesn't reliably produce the outcome it's aimed at, even on its own terms.
• Deliberately targeting people already identified as fatigued, overloaded, or emotionally heightened, and pushing timed contradictions at them, is the actual mechanism behind several well-documented, often illegal harms: coercive control within abusive relationships, cult recruitment cycles (love-bombing followed by induced doubt and isolation), elder financial exploitation, and high-pressure sales/scam tactics that rely on manufactured urgency. These aren't hypothetical failure modes of the model — they're the real-world pattern the model would be describing if it were built as a targeting tool.
• For someone experiencing a break from consensus reality, deliberately introducing more contradictory stimulus isn't a persuasion technique — it's a way to worsen an active psychiatric crisis. That's a situation calling for professional mental-health support, not more engineered dissonance.

This is why Part 1 stops at description: the same variables that make the model interesting (load, timing, contradiction) are exactly the variables that make a targeting version of it dangerous.

2.5 Recognizing the pattern

Signs this dynamic may be happening to you or someone you know:

• Big decisions being pushed during exhaustion, grief, illness, or overwhelm
• Contradictory, urgent information delivered together, paired with pressure to decide immediately
• Pressure that increases specifically when you ask for time or push back
• Being discouraged from checking in with other people ("they won't understand," "don't tell them yet")
• Alternating warmth and sudden doubt/criticism (idealize-then-destabilize cycles)

Protective responses:

• Treat "decide now, don't check with anyone" as a red flag on its own, separate from the content of the decision
• Default to delaying significant decisions made while fatigued or overwhelmed, when delay is possible
• Get an outside read from someone not involved in the situation
• If someone's grip on shared reality seems to be slipping — not just disagreeing with you, but losing contact with consensus reality — professional support is the right next step, not more argument

---

Limitations

• Part 1 inherits every limitation already flagged in the two source documents: free parameters are unfit, the sinusoidal representation of traits is unproven, the shared-frequency assumption is the weakest structural piece. Relabeling variables as CPU/I-O terms doesn't fix any of that — it's the same unvalidated math.
• Part 2's underlying psychology (dual-process theory, the Elaboration Likelihood Model, cognitive dissonance theory) is genuinely established research, but the wave model doesn't add anything to it beyond restating a known correlation in different notation.
• No targeting or execution layer is included, by design, for the reasons in 2.4 — this is a description of a correlation and a guide to recognizing it, not an operating procedure.