ımmanaıt KQ – Knowledge Quality Explorer

Basic logic

How do audit inputs become three densities?

KQ looks for material deviations under an assumption’s own reality-claim — not for the trivial fact that assumptions are never perfect copies of reality. In an audit, relevant operative assumptions are assessed across the knowledge base and weighted where they shape action, decisions, or orientation.
For one audit item, ν (Greek nu) estimates the size of the relevant deviation. v (Latin v) estimates how visible that deviation is to the assessed system. π estimates how protected the visible part is.
Together, these inputs produce three densities: ρ — the part the assessed system does not yet recognize as deviation, κ — the visible and addressable part, and Π — the visible but protected part.

One audit item

ν (Greek nu) — Size of the assessed deviation0.48

clear deviation · lower end

ν (Greek nu) sets the size of the assessed deviation before it is partitioned.

v (Latin v) — How visible is the deviation to the assessed system?0.62

partly visible · upper end

Invisible share = 1 − v (Latin v). This is the source of the ρ share in this partition.

π — How protected is the problem?0.41

politically/institutionally protected · lower end

For the calculation, π splits the visible share: what is openly addressable, and what remains protected? Protection may also reduce visibility; that is reflected through v (Latin v).

Partition of the assessed deviationTotal value: N = 0.48

ρ

κ

Π

Step 1

Deviation size

ν (Greek nu) sets the assessed deviation size: 0.48.

Step 2

Visible or PQD share?

38% remains invisible → ρ = 0.182.

62% is visible and is partitioned further.

Step 3

Open or protected?

59% of the visible share is openly addressable → κ = 0.176.

41% of the visible share is protected → Π = 0.122.

ρ shows the PQD share: the part of the deviation that the assessed system does not yet recognize as a deviation. κ shows the visible and addressable share. Π shows the visible but protected share.

From audit item to knowledge base. This section shows the basic partition for one audit item. The same ν (Greek nu), v (Latin v), and π fields appear in the audit table below, where they are repeated per item and weighted by action weight a. The learning-loop model later adds how resistance and hard-to-evade evidence can affect correction over time.

Effect of ρ

Why PQD blocks intelligence

This simplified view separates PQD inhibition from total potential. The reference potential is 100%, but the demo uses a bounded operating range: even without PQD not all potential becomes effective, and even under very high PQD some residual capability remains.

76.2

/ 100

bounded operating range

15–88%

effective range in this demo; reference potential remains 100%

ρ — PQD density0.32

elevated ρ · lower end

I = bounded effective share of 100 · unavailable or held back = 23.8

Bounded inhibition curve — PQD increasingly holds back effectiveness

Why does the curve not start at full effectiveness or end at zero? This demo separates PQD inhibition from total potential. Even at ρ=0, not all theoretical potential may become effective; and even at very high ρ, some residual capability may remain. The curve maps ρ into a bounded operating range: 15–88% effective intelligence here. Within that range, the PQD effect is convex: each additional point of ρ can reduce effective intelligence more than the previous one. The numbers are illustrative, not calibrated.

Try Ω — manual correction steps

Correction step

Try Ω — manual correction steps

Ω marks one correction step. It starts from the current ρ produced in Basic Logic. Apply Ω once to move from ρ_t to ρ_t+1; after that, the panel continues locally from the new value.

What the step does

Each click asks: under the current conditions, how much of the current PQD density ρ_t can be reduced, and how much new PQD enters the next state?

Show formula

ρ_t+1 = clip_[0,1][ ρ_t(1 − α_eff·C_t) + ε_t ]
α_eff = α · 1_{T≥θ}
if repeated unchanged and the gate is open: ρ* = ε / (α·C)

Where does ρ_t come from? The first Ω baseline is the current ρ from Basic Logic: the PQD share produced by ν (Greek nu), v (Latin v), and π. After Ω is applied, this panel continues locally: ρ_t+1 becomes the next ρ_t.
Why does Ω not simply erase PQD? A correction can close a concrete PQD form, but the next state may also contain new or successor-generated PQD. KQ is therefore an ongoing operation, not a final clean state.
Repeated unchanged. The reference value below is dynamic: with the gate open and no clipping, ρ tends toward the balance point ρ* = ε/(α·C). Repeated unchanged: ρ* = ε/(α·C) ≈ 0.084.

α — correction force (higher = stronger effect)

0.65

strong intervention · lower end

C — corrective capacity (higher = better)

0.55

usable corrective capacity · upper end

T — reality readiness (θ=0.5)

0.75

reality-oriented · lower end

higher = correction more likely

ε — new PQD inflow

0.030

turbulent PQD inflow · lower end

Gate open: correction can take effect

T ≥ θ

Before the first Ω click, the starting ρ_t follows the current ρ from Basic Logic. After each click, ρ_t+1 becomes the next local ρ_t. Change the sliders before clicking again to test a different next step. Use Reset to relink the first baseline to Basic Logic.

current ρ_t

0.182

ρ_t+1 after Ω

0.147

Δρ

−0.035

Learning loop

When correction reinforces itself

The core model shows the learning loop: reducing ρ can make more intelligence effective, and successful correction can strengthen the capacity to correct again. The demo also adds P as a separate scenario variable for resistance over time. P is not a density and not a redefinition of Π. C grows through successful correction, not by assumption.

Core model

ρ, I, and C show the core loop: lower PQD can make more intelligence effective, and successful correction can strengthen C. Without gains, C can also erode.

Demo extension

P and q are separate loop variables for resistance and evidence strength over time. P is not a density and not a redefinition of Π.

What comes from the audit, and what is set for the loop? ρ, κ, and Π come from the audit logic. The learning loop also needs estimates for correction force, capacity growth, erosion, readiness, new PQD inflow, resistance, and evidence strength. These values may be audit-informed or exploratory, but they are not the three densities themselves. P is a resistance estimate for the loop. It is not measured as a density and does not redefine Π.

What shapes the learning loop?

Main controls. Start with the few values that shape the loop most visibly: initial PQD, correction readiness, correction force, new PQD inflow, and resistance. Advanced values remain available below.

Start-ρ — initial PQD density (after reset)0.42

elevated ρ · middle

T — correction readiness (higher = better)0.70

ready to learn · upper end

the gate opens when readiness reaches the threshold

α — correction force (higher = stronger effect)0.65

strong intervention · lower end

ε — new PQD inflow per cycle0.015

normal PQD inflow · middle

higher = more new PQD enters the next cycle

Start-P — resistance estimate (after reset)0.22

noticeable resistance · lower end

P is a loop estimate, not a density or a redefinition of Π

Advanced parameters

These values refine the loop behavior: how gains become capacity, how capacity erodes, where the gate threshold sits, and how hard-to-evade evidence affects readiness and resistance.

γ — learning amplification (higher = better)0.55

success partly becomes routine · upper end

Start-C — initial corrective capacity0.32

limited corrective capacity · upper end

applied when the learning-loop simulation is reset or started

δ — capacity erosion (higher = more risk)0.025

low capacity loss · upper end

higher = corrective capacity is lost faster

θ — correction threshold (higher = harder)0.50

normal threshold · middle

higher = gate opens less easily

q — evidence strength (higher = less evasion)0.55

noticeable evidence strength · middle

clearer evidence can slightly raise T_eff and reduce P when the gate is open

Development over cycles Cycle 0 · Gate open

■ ρ — PQD ■ I — effective intelligence / 100 ■ C — corrective capacity ■ P — resistance estimate, not a density

PQD ρ

0.42

effective intelligence I

82

corrective capacity C

0.32

resistance P

0.22

newly addressable Δκ

0.000

When does the gate open? T ≥ θ means the system can continue correction when the result becomes uncomfortable or costly. Below the threshold, correction does not take hold: new PQD can accumulate, corrective capacity can erode, and resistance can solidify. Clearer evidence can raise effective readiness slightly. When the gate is open, ρ can fall, P can weaken, and Δκ shows how much PQD became newly addressable in that cycle.

LEARNING SCENARIO

Learning loop gets going

High PQD and noticeable resistance at the start, but the evidence is hard to evade. If the gate stays open, ρ and P fall; I and C rise.

→ Load learning system

DEFENSIVE SCENARIO

Self-correction breaks down

The gate often remains closed. New PQD can arise, resistance can solidify, C erodes, and I falls.

→ Load defensive system

Note: One cycle corresponds to a complete KQ loop: test a representation or assumption → surface PQD and make it addressable → revise or replace the representation → observe the effect. Fast: about 1 month; slow: about 1 quarter. The demo uses 12 cycles as the default horizon.

Law 5 · successor check

Correction also asks what comes next

When a representation can no longer carry its reality-claim, action still needs a successor. A successor is not automatically better: it has to be checked against the old mismatch, the action it enables, and any new trade-off it introduces.

r
old representation — Ω → r′
successor

This section is explanatory in this demo. It does not feed back into the calculations.

It shows what must be checked after correction: whether the successor representation is better, still unresolved, or incomparable because it introduces a relevant trade-off.

better

r′ reduces the old mismatch and opens at least as much valid action in the relevant context.

unresolved

r′ is plausible, but the successor has not yet been tested enough to count as a clean improvement.

incomparable

r′ solves one relevant problem, but introduces a trade-off or closes another relevant action path.

No hidden feedback. In this demo, the successor check is not fed back into ρ, I, C, or ε. In a fuller version, a documented successor outcome could inform the successor-generated part of ε, the new PQD inflow in the learning loop.

Practical application

Practical assumption audit

Enter, edit, import, or export audit items: operative assumptions, models, or decisions that may shape action. For each item, assess the material deviation, visibility to the assessed system, protection, and action weight.

How to choose values: The sliders use coarse, documented categories such as “clear deviation,” “partly visible,” or “protected.” The decimal value makes the estimate calculable; the category keeps the judgment auditable. Better roughly right than precisely wrong.

Longitudinal audits need comparability. When comparing real developments over time, document which assumptions were replaced, removed, or newly introduced. Otherwise, a falling ρ may also reflect a changed knowledge base.

Audit items: assumptions, models, and decisions

Audit name

Last export: —

Assumption / model	Claim / evidence notes	Deviation strength ν Greek nu	Visibility v Latin v	Protection π	Action weight a

ρ — PQD

0.000

κ — open problems

0.000

Π — protected problems

0.000

N — total deviation

0.000

I — intelligence index from ρ

—

How the audit feeds the loop. Per audit item, ν (Greek nu) estimates the size of the relevant deviation, v (Latin v) estimates its visibility to the assessed system, π estimates protection of the visible part, and a weights how much the item counts in the aggregate. These inputs produce ρ, κ, Π, and N. The displayed I is not scored separately; it is a demo index derived from audit-ρ using the bounded inhibition curve from “Effect of ρ”: I = 100 · [0.88 − (0.88 − 0.15) · ρ^1.6].

Current audit values will appear here.

Transfer rule: audit-ρ becomes Start-ρ. Start-P is estimated as Π/(κ+Π), the protected share of visible problems. T and ε are suggested from the audit state. α, γ, δ, θ, q, and Start-C remain loop settings.

Audit XML can be exported with a timestamp and audit name in the filename, then loaded later via file selection. A standalone HTML file cannot automatically scan its local folder.

System dynamics summary

How the learning loop reinforces itself — or erodes

R1: lower ρ can make more intelligence effective; gains in effective intelligence can strengthen C; and higher C can reduce ρ again when the gate is open. B1: erosion δ lowers C directly. It does not directly raise ρ, but it weakens future correction. Green arrows mark the reinforcing loop here, not κ.

Selected application fields

Where the logic becomes practical

The same KQ logic can be re-scoped or applied in different contexts. The cards below are selected examples, sorted alphabetically, not the full module set.

Artificial systems — AI and model awareness

AI systems also need access to the model-character and limits of their outputs; otherwise PQD can remain invisibly effective at scale.

Communication — avoiding category errors

At first contact, KQ can be misread as a framework, coaching tool, AI service, or generic method. Communication quality preserves the order: discovery, standard, application.

Execution fidelity — correct use matters

The operator is simple, but professional execution depends on scoring discipline, case memory, protocols, and audit quality.

Freedom — freedom as replaceability

Freedom here means the standing capacity to replace a representation that no longer carries reality-contact, rather than remaining invisibly bound by it.

Governance — making correction binding

Governance makes correction binding: who may test assumptions, who may object, who must act, and how correction remains auditable.

NEO architecture — structures that sustain learning

Good structures preserve learning capacity and reduce the chance that correction dissipates after the first visible improvement.

Organization — shared model quality

An organization can have good local insights and still act poorly if its representations do not cohere across teams, roles, and decisions.

Parasitic equilibria — visible but protected problems

Some problems are not hidden. They are seen, but remain protected by incentives, status, fear, role, or yield. That is the Π side of the partition, not ρ.

Responsibility — from hidden to addressable

PQD becomes operationally relevant when it is surfaced as visible non-identity. Only then can responsibility be attached to what the system can address.

Show glossary & technical notes

The main page uses PQD as the guiding term. The formal terms below preserve compatibility with the KQ Algebra while keeping the demo readable. Slider value bubbles use a contextual traffic-light cue: green means favorable for effective intelligence or correction, red means unfavorable or riskier, and gray means neutral or midpoint. Action weight uses amber because it marks importance rather than good or bad.

Notation note. The first two audit inputs can look similar in some fonts: ν is Greek nu and means deviation strength; v is the Latin letter v and means visibility.

a
Action weight: how much the audit item counts because it shapes action, decisions, dependencies, or orientation.

α
Correction force of an intervention when the gate is open.

Audit-item N
N is the total assessed deviation before partitioning. In the density model, N is divided into ρ, κ, and Π. In a full audit, comparable items are weighted and aggregated across the operative knowledge base.

Audit transfer
Audit-ρ transfers directly as Start-ρ. Start-P is estimated as Π/(κ+Π). T and ε are suggested from the audit state by demo heuristics. α, γ, δ, θ, q, and Start-C remain loop settings. The audit I value is a derived demo index from ρ, not a separately scored audit variable.

Audit XML
The practical audit table can be exported and imported as XML. The XML is for documentation and reloading within this standalone demo; it is not a certification format.

Basic partition
For one audit item: ρ = ν(1−v), κ = νv(1−π), Π = νvπ, and N = ν. In the audit table, item values are weighted by a and aggregated.

C
Corrective capacity: how well a system can address PQD. In the learning loop, Start-C is a loop setting; it is not computed from the audit table.

clip[0,1]
Technical rule: values are constrained to the interval 0 to 1.

δ
Capacity loss or erosion over cycles.

Environment dynamic (adaptive mode)
When active, the selected slider values are not used as fixed constants. For each cycle, the demo derives effective values from the current state of ρ, P, C, I, and q. These couplings are illustrative demo assumptions, not calibrated empirical laws.

ε
New PQD inflow: new, imported, or successor-generated PQD entering the next cycle. It explains why PQD does not simply disappear after correction.

η(ρ)
Function that describes how ρ reduces effectiveness within a bounded operating range. In this demo, effective intelligence ranges from 88% at ρ=0 to 15% at ρ=1, with a convex PQD effect η(ρ)=ρ^1.6 inside that range. It is illustrative, not calibrated.

γ
Learning amplification: whether successful correction becomes repeatable capacity.

I
Effective intelligence index in the demo. It is not IQ. In the audit table, I is derived from ρ using the bounded inhibition curve; it is not scored separately.

Ī
Finite reference potential or ceiling of a system; normalized to 100 in this demo. The demo does not assume that ρ=0 reaches all of it or that ρ=1 destroys it completely.

κ
Visible and addressable non-identity: the part of the deviation that the assessed system can see and work on.

Law 5 successor check
Explains what must be checked after correction: better, unresolved, or incomparable. In this demo, it is not fed back into ρ, I, C, or ε.

Non-identity
A material deviation under a representation’s own reality-claim. The demo usually says deviation or mismatch for readability.

ν (Greek nu)
Strength of the assessed material deviation. In some fonts, ν can look similar to the Latin letter v.

Ω
A correction step. In the manual panel, Ω maps the current ρ_t to ρ_t+1 under the selected conditions.

Ω reference value ρ*
The balance point in the manual Ω panel under unchanged settings. With the gate open and no clipping: ρ* = ε/(α·C). It is not a prediction of the full learning-loop simulation.

P — resistance estimate
A separate demo state for resistance over time. P is not a fourth density and not a redefinition of Π. From the audit, P is cautiously estimated as Π/(κ+Π), the protected share of visible problems.

π
Protection setting. In the basic partition, π splits the visible share between κ and Π; in practice, protection can also reduce visibility and then appears through v (Latin v).

Π
Visible but protected non-identity. Π belongs to the density partition and is not the same as the later resistance variable P.

Publication status
Didactic demo version. Reference use is limited to lawful quotation, scholarly reference, review, commentary, and non-commercial higher-education demonstration with attribution. Professional and commercial use, including organizational consulting, is excluded without separate written authorization.

q — evidence strength
How hard the evidence is to evade in the demo scenario. In adaptive mode, stronger evidence can slightly raise effective readiness and weaken resistance when the gate is open. q does not change the density partition.

Representation
An assumption, model, image of reality, decision logic, or other operative representation that can guide action.

ρ — PQD density
Passive Qualitative Disinformation: invisible, action-relevant deviation, weighted where it shapes action, decisions, or orientation.

T
Readiness for correction: whether the system can continue correction when the result becomes uncomfortable or costly.

θ
Threshold: how high the hurdle is before correction takes hold. The gate opens when T ≥ θ.

v (Latin v)
How visible the assessed deviation is to the assessed system. This is the Latin letter v, not Greek ν.

Show compact formulas and transfer notes

Basic partition. For one item, ν sets the size of the relevant deviation, v sets visibility, and π splits the visible part. In a full audit, item values are weighted by action weight a and aggregated.

ρ = ν(1 − v)
κ = νv(1 − π)
Π = νvπ
N = ρ + κ + Π = ν

Audit transfer:
audit-ρ → Start-ρ
Π/(κ+Π) → Start-P
T and ε → suggested loop values from demo heuristics
α, γ, δ, θ, q, Start-C → remain loop settings

Show adaptive-mode formulas

Adaptive mode is a demo extension. It updates effective loop inputs from the current state. The selected sliders remain the baseline; the following effective values are recomputed each cycle and then clipped to the valid range where needed.

I = current effective intelligence / 100
qPressure = 0.16·q·(0.35 + 0.65·C)·(1 − 0.35·P)
T_eff = clip[T + 0.16·(C − 0.40) + 0.10·(I − 0.70) − 0.05·ρ − 0.16·P + qPressure]
α_loop = clip[α·(0.88 + 0.35·C + 0.18·(1 − P))]
γ_loop = clip[γ·(0.85 + 0.35·I + 0.10·(1 − P))]
δ_loop = clip[δ·(1.15 − 0.55·C + 0.35·P)]
q_loop = clip[q·(0.75 + 0.55·C + 0.20·I)]
ε_loop = max(0, ε·(1.05 − 0.35·C − 0.18·(1 − P) + 0.22·ρ + 0.16·P))

Publication status & rights

Didactic demo, not a professional execution license

This standalone micro app is a didactic, paper-informed demo of selected KQ relations. Official information and supporting materials are referenced at immanait.com.

Copyright

© Dr. Thomas R. Glück. All rights reserved. immanait, KQ, PQD, NEO, the KQ operation, professional execution protocols, audit/certification architectures, derivative products, services, software implementations, trade secrets, trademarks, patent claims, and future implementations remain reserved.

Permitted reference use

Short quotation, scholarly reference, review, and commentary remain permitted where allowed by applicable law. This demo may also be used as a non-commercial didactic demonstration in higher-education contexts, with attribution and without implying certification, authorization, endorsement, or a professional implementation license.

Explicit exclusions

No permission is granted to use this demo, its terminology, formulas, screens, export format, or derived materials for organizational consulting, management consulting, corporate training, audit or assessment services, certification, implementation projects, AI or software products, commercial workshops, advisory services, or other professional/commercial applications without separate written authorization.

Non-limitation. This demo is illustrative and non-exhaustive. It does not define, limit, waive, surrender, dedicate, or exhaust the scope of Knowledge Quality, PQD, immanait NEO, any professional protocol, commercial service, certification architecture, technical implementation, legal claim, or future version.

Selected references

Selected Publications

Selected Publications:
https://orcid.org/0009-0009-8514-8592

Introduction to KQ:
https://doi.org/10.17605/OSF.IO/2N6E4

The Genius Machine:
https://doi.org/10.17605/OSF.IO/N3VBD

Laws of KQ – The Genius Machine Algebra:
https://doi.org/10.17605/OSF.IO/53JXV