§2 Paradigm Shift
Process accountability vs outcome accountability. Empowerment infrastructure, not surveillance. Cognitive load theory. Psychological safety.
§2 Paradigm Shift
v2.0.0 · Locked · L1 · March 19, 2026
Purpose
This section establishes the design paradigm that the substrate implements. §1 established that the problem is real and empirically validated. This section answers the harder question: why do organizations produce truth-indifferent outputs? The answer determines the design. If the problem is moral failure, the solution is surveillance. If the problem is infrastructure failure, the solution is empowerment. Five decades of behavioral decision research establishes that the problem is structural — and the substrate is empowerment infrastructure, not surveillance technology.
Foundation
Empirical Grounding in Behavioral Decision Research
This section is grounded in five decades of independent research across cognitive science, organizational psychology, and behavioral economics. The research converges on a single insight: human decision-making is bounded by cognitive resources, shaped by context, and responsive to infrastructure design. Design specifications are derived from this evidence, not from philosophy or preference.
The substrate implements a design paradigm different from traditional governance infrastructure. That paradigm shift is specified here through six research foundations: process accountability, bounded rationality and satisficing, cognitive load theory, dual-process cognition, psychological safety, and variety engineering. Each grounds a specific design choice in the substrate.
Governance
This section is owned by Cam (founder, GrytLabs). Changes to behavioral research grounding or paradigm principles require evidence review. The design paradigm is not aspirational; it is research-grounded.
Substance
§2.1 Process Accountability, Not Outcome Accountability
Lerner and Tetlock (1999) established the distinction that defines the substrate's design. Two types of accountability produce opposite effects on decision quality:
| Accountability Type | Definition | Effect on Decision Quality |
|---|---|---|
| Process accountability | Accountable for how you decided — the reasoning, evidence considered, alternatives evaluated | Improves decisions: more careful reasoning, reduced cognitive bias, genuine consideration of alternatives, openness to disconfirming evidence |
| Outcome accountability | Accountable for what happened — whether the result was favorable or unfavorable | Degrades decisions: defensive behavior, premature closure, confirmation bias, avoidance of risky but potentially correct decisions |
This is not a philosophical preference. It is an empirical finding replicated across multiple experimental paradigms.
The mechanism is direct. Under outcome accountability, the decision-maker seeks the first defensible option — not the best option — and locks in early. Contrary evidence is resisted because changing course implies the first decision was wrong. The evaluation criterion rewards luck (did the world cooperate with your choice?) rather than quality of reasoning. Under process accountability, the evaluation criterion is the quality of reasoning itself: what evidence was considered, what alternatives were explored, what constraints were weighed. This produces genuine deliberation because thorough analysis is rewarded regardless of whether the outcome happens to be favorable.
The substrate captures decision process — the intent, the evidence considered, the constraints active, the authority exercised. It does not retrospectively judge whether the outcome was "right." This is process accountability infrastructure. The substrate proves you decided responsibly; it does not judge whether the world cooperated with your decision.
Kahneman and Tversky's prospect theory (1979) amplifies the distinction. Loss aversion — losses loom approximately twice as large as equivalent gains — means that outcome accountability structurally invokes loss framing. Being held accountable for outcomes feels like potential exposure: punishment, career damage, blame. Process accountability can invoke gain framing: evidence of competence, demonstration of thoroughness, protection against unfair blame. This is not rhetoric — it is architecture. A substrate designed around outcome accountability would structurally trigger defensive decision-making in every user. A substrate designed around process accountability enables the cognitive conditions under which good decisions are made.
§2.2 Bounded Rationality and "Click and Hope"
Simon (1947; 1955) demonstrated that real decision-makers face limits in memory, attention, information processing, and time. The rational response to bounded cognitive resources is not optimization — it is satisficing: seeking solutions that are good enough rather than optimal. Satisficing is not failure. It is the only computationally feasible strategy given the constraints of the decision environment. Organizations that demand optimization from satisficing agents produce two pathological outcomes: decision paralysis (the optimization demand exceeds cognitive capacity, so no decision is made) or compliance theater (the appearance of optimization is produced without the substance — the decision was made on other grounds and retroactively justified).
"Click and hope" is satisficing under extreme cognitive load with inadequate decision support. The decision-maker faces a complex governance choice — interacting constraints, competing stakeholder interests, uncertain evidence — and the information environment provides no structured support for navigating it. The rational response is to select the most familiar option and move on. This is not a character flaw. It is the predictable behavior of a bounded agent in an unsupportive environment. The substrate addresses the environment, not the agent.
Sweller (1988) provides the formal framework. Working memory has strictly limited capacity. Cognitive load has three components, and each maps to a substrate design response:
| Load Type | Definition | Substrate Response |
|---|---|---|
| Intrinsic | Inherent complexity of the decision — interacting business rules, constraints, stakeholders, precedents | Structure into nine named primitives (§4), each independently addressable rather than entangled in a single unstructured space |
| Extraneous | Burden imposed by how information is presented — compliance paperwork, unclear formats, documentation as afterthought | Eliminate — capture at the moment of action, not as a separate compliance activity. The act of making a decision IS the act of documenting it |
| Germane | Productive cognitive effort devoted to building understanding — the deliberation that improves decisions | Maximize — surface relevant precedents, active constraints, and applicable evidence in context. Help the decision-maker understand the decision, not just comply with governance requirements |
Table 2.2.1: Cognitive load decomposition and substrate design specification.
When total load (intrinsic + extraneous) exceeds working memory capacity, germane processing collapses. The decision-maker stops building understanding and starts coping — satisficing at the lowest aspiration level, defaulting to the most familiar option, clicking and hoping. Traditional governance adds extraneous load (fill out this form, justify this decision, document this process) on top of already-overwhelming intrinsic load. The substrate is designed to reduce total load by structuring intrinsic load into manageable primitives and eliminating extraneous load entirely.
§2.3 Cognitive Load Theory Applied to Governance
The three components of cognitive load map directly to substrate design. Intrinsic load — the complexity of the decision itself — cannot be eliminated. Stakeholders genuinely interact; constraints genuinely conflict; evidence is genuinely uncertain. The substrate does not remove this complexity; it structures it into nine independent primitives, each addressing a distinct governance question, so the decision-maker can engage with complexity without overload.
Extraneous load — the burden of documentation, form-filling, compliance paperwork — is architectural waste. It imposes cost without improving decisions. The substrate eliminates it by collapsing the documentation and decision activities into a single operation. When a decision is made, the governance record is created simultaneously. The act of deciding IS the act of documenting.
Germane load — productive cognitive effort — is the target. The substrate maximizes germane load by surfacing decision-relevant context at the moment of decision: what commitments are active, what capacity constraints bind, what evidence is available, what authority is delegated, what constraints apply. This is decision support infrastructure, not compliance infrastructure.
§2.4 Dual-Process Cognition and System 2 Preservation
Kahneman (2011) formalized the dual-process architecture of human cognition. System 1 operates automatically — fast, heuristic, bias-prone. System 2 allocates attention to effortful deliberation — slow, analytical, capacity-limited. Under cognitive load, time pressure, or sequential decision demands, processing defaults to System 1 heuristics.
The governance implication: every additional governance demand consumes System 2 capacity. If governance itself depletes System 2, then adding governance requirements beyond a threshold produces worse decisions, not better ones. Danziger, Levav, and Avnaim-Pesso (2011) demonstrated this empirically: experienced Israeli parole judges granted favorable rulings at approximately 65% at the start of each session, declining to near 0% by session's end — a pattern driven by sequential depletion of deliberative capacity, not case characteristics.
The paradox, formalized:
- Good governance requires careful deliberation (System 2)
- System 2 has limited capacity and fatigues with use
- Each governance requirement consumes System 2 capacity
- Therefore, adding governance requirements beyond a threshold reduces governance quality
The substrate resolves this paradox by absorbing routine governance operations into infrastructure. Constraint checking — whether the decision violates active rules — is not a deliberative activity. Precedent surfacing — whether similar decisions exist — is not a deliberative activity. Authority verification — whether the actor has the delegated right to decide — is not a deliberative activity. These are mechanical operations that consume System 2 capacity without benefiting from it. The substrate performs them structurally, reserving the decision-maker's System 2 for the genuinely deliberative aspects: weighing competing values, exercising judgment about ambiguous evidence, considering novel alternatives. This is not automation replacing human judgment. It is infrastructure preserving human judgment by protecting it from exhaustion on tasks that do not require it.
§2.5 Psychological Safety as Infrastructure Requirement
Edmondson (1999) established that psychological safety — a shared belief that the team is safe for interpersonal risk-taking — is the primary predictor of learning behavior in work teams. Learning behavior includes asking questions, admitting mistakes, reporting errors, and flagging problems. If the substrate is perceived as a surveillance mechanism, it will suppress the very behaviors it needs to capture. People will not document uncertainty, flag constraints, or admit that they satisficed if they believe the record will be used against them.
Bernstein (2012) sharpened the point. In a field study at a large manufacturing facility, complete observability of workers reduced performance. Workers responded to total transparency by concealing activities through codes and workarounds. Conversely, creating zones of privacy increased performance — privacy supported productive deviance, experimentation, and continuous improvement. This distinguishes surveillance from process accountability:
| Dimension | Surveillance | Process Accountability |
|---|---|---|
| Question | "How are they behaving?" | "How did they reason?" |
| Observes | Operational behavior in real time | Decision reasoning after the fact |
| Behavioral response | Impression management, concealment | Genuine deliberation, documented reasoning |
| Effect on performance | Degrades (Bernstein 2012) | Improves (Lerner & Tetlock 1999) |
Table 2.4.1: Surveillance vs. process accountability. The substrate captures decision reasoning without creating panoptic visibility into operational behavior.
The substrate captures why a decision was made and what was considered. It does not capture how someone spent their time getting there. The difference is structural, not cosmetic.
Organizational silence compounds the problem. Morrison and Milliken (2000) demonstrated that silence is a collective phenomenon driven by organizational structures, not individual shyness. Detert and Edmondson (2011) identified implicit voice theories — largely tacit beliefs about when speaking up is risky — that suppress voice even in psychologically safe environments with supportive leadership. These findings mean that creating a safe environment is necessary but not sufficient. The substrate must actively structure voice as a first-class operation.
Signal Capture (§15) and IQ Capture (§16) are the anti-silence infrastructure. Signal Capture provides governed channels for problem-flagging on any governed object — behavioral invariant B7 (§5) guarantees the channel exists on every primitive, and B8 ensures signals route to authority on the governance chain. IQ Capture provides governed channels for cognitive emergence — questions, patterns, gaps, connections — with B9 guaranteeing that resolved inquiries produce governance action. The distinction between informal voice and governed voice is architectural: informal voice is vulnerable to silence norms, implicit voice theories, and individual risk calculation. Governed voice — with defined channels, structural routing, invariant-enforced delivery, and explicit signal types — resists organizational silence because it is infrastructure, not individual courage. §14 specifies the unifying capture framework; §15 and §16 specify the mechanisms.
§2.6 Ashby's Law Applied to Individual Decision-Making
Ashby's Law of Requisite Variety (1956): V(R) ≥ V(D). The variety of the regulator must match the variety of the disturbances it controls. §1.3 applied this to the organizational level: organizations lack governance infrastructure with enough variety to represent their decisions. The same law applies to the individual decision-maker.
The decision-maker faces disturbance variety — interacting constraints, stakeholder interests, regulatory requirements, precedents, competitive dynamics. The decision-support capacity available — what information is structured, what precedents are surfaced, what constraints are flagged — constitutes their regulatory variety. When V(R) < V(D), the decision-maker cannot match the complexity they face. "Click and hope" is the inevitable result.
The variety gap IS the empowerment gap. "Click and hope" reframed through Ashby's Law is not moral failure — it is variety deficiency. The decision-maker satisfices (Simon) because their regulatory variety is insufficient. They default to System 1 (Kahneman) because System 2 cannot compensate for missing infrastructure. They produce truth-indifferent outputs (§1.2) because the environment provides no structured mechanism for engaging with truth. No amount of training, motivation, or accountability closes a variety gap. Only infrastructure closes variety gaps — by providing the model, context, and structure that the decision-maker currently lacks.
This connects §1's cybernetic diagnosis to §2's design paradigm. The Conant-Ashby theorem (§1.3) says governance must be a model of the system it regulates. Ashby's Law says the model must have enough variety. The substrate provides both: an organizational world model (§3) with requisite variety (nine primitives, §4) that surfaces decision-relevant context to the decision-maker at the moment of action. The empowerment gap closes when the infrastructure provides the regulatory variety that the individual, unaided, cannot match.
§2.7 The Triple Convergence: Three Independent Research Traditions
Three independent research traditions — each with different methods, different vocabularies, and different communities — have discovered the same dual-process architecture for intelligent systems:
| Tradition | System 1 (Reactive) | System 2 (Deliberative) | Transition |
|---|---|---|---|
| Cognitive Science (Kahneman 2011) | Fast, heuristic, bias-prone | Slow, analytical, capacity-limited | Load and fatigue push S2 → S1 |
| Computational Neuroscience / AI (LeCun 2022) | Reactive, pattern-matched, no world model | Model-based planning, deliberative reasoning | Amortized inference compiles S2 → S1 |
| Organizational Cybernetics (Beer 1972) | S1–S3 operations (here and now) | S4 intelligence (there and then) | Variety engineering balances inside/outside |
Table 2.6.1: Three traditions discover one architecture. The convergence is structural — any system operating in a complex environment faces the same reactive/deliberative trade-off.
The convergence is not coincidental. Any system — biological, artificial, or organizational — that must function in a complex, changing environment faces the same trade-off: reactive processing is fast and cheap but fails on novelty; deliberative processing handles novelty but is slow, expensive, and capacity-limited. The critical architectural question is not whether both modes are needed — all three traditions agree they are — but how the transition between them is governed.
The substrate implements this convergent architecture. The graduation mechanism (§A1) provides principled transition between reactive and deliberative governance modes:
- Option A: Human routes all decisions; substrate captures with full primitive context.
- Option B: Substrate proposes routing; human confirms or overrides.
- Option C: Substrate routes within delegation bounds; human reviews by exception.
Each stage requires demonstrated competence at the previous level — evidence accumulation justifies progressive transfer of routine governance from human deliberation to substrate infrastructure. This is variety engineering (Beer) applied to decision governance: the substrate progressively absorbs reactive governance to preserve human deliberative capacity for genuine judgment. The transition is LeCun's amortized inference in organizational form — deliberative solutions compiled into reactive patterns as evidence of reliability accumulates.
Boundaries
This section specifies the design paradigm — process accountability, empowerment infrastructure, cognitive load reduction, psychological safety. It does NOT specify: the substrate mechanism (§4–§7), the graduation mechanics (§A1), or the interface design that operationalizes these principles (§14–§17).
Positions
Locked. Design paradigm is empirical, not aspirational. Process accountability improves decisions; outcome accountability degrades them. Empowerment infrastructure reduces cognitive load; surveillance infrastructure depletes deliberative capacity. Psychological safety is an architectural requirement, not a cultural aspiration. The three research traditions converge on a dual-process architecture that the substrate implements through the graduation mechanism.
Lineage
v1.0.0 (February 25, 2026): Base lock. Six design principles grounded in five decades of behavioral research. v2.0.0 (March 19, 2026): Locked as paradigm specification. No changes to research grounding or design principles. Formatting converted to fspec.
Commitments
SDK implementations MUST implement process accountability infrastructure, not outcome accountability surveillance. SDK implementations MUST reduce cognitive load through nine-primitive structure and elimination of extraneous documentation burden. SDK implementations MUST support psychological safety through governed voice channels (B7, B8, B9) that function as infrastructure, not cultural features.
Coverage
All six design principles fully specified with research grounding: process accountability (Lerner & Tetlock), satisficing (Simon, Sweller), dual-process cognition (Kahneman), psychological safety (Edmondson, Bernstein), variety engineering (Ashby), and triple convergence (LeCun, Beer). Graduation mechanism specified in §A1.
Addressing
Document ID: s02-paradigm-shift. Part: I (Foundational Paradigm). Sections addressable by number (§2.1–§2.7). Cross-references use [s02-paradigm-shift.{section-id}] notation.
Attribution
Primary author: Cam (founder, GrytLabs). Research grounding: Lerner & Tetlock (1999), Simon (1947, 1955), Sweller (1988), Kahneman (2011), Kahneman & Tversky (1979), Danziger et al. (2011), Edmondson (1999), Bernstein (2012), Morrison & Milliken (2000), Detert & Edmondson (2011), Ashby (1956), Beer (1972), LeCun (2022). Composition support: Claude (Anthropic).
Situational Frame
This section was composed after establishing the problem (§1) to answer the design question: if organizations produce truth-indifferent outputs, is that a moral problem (requiring surveillance) or a structural problem (requiring empowerment)? The answer is structural, grounded in five decades of behavioral research. The paradigm shift from surveillance to empowerment is research-derived, not philosophical.
Scope Governance
Core namespace: dlp. Design paradigm applies to all subsequent specifications. The six principles constrain interface design (§14–§17), graduation mechanics (§A1), enforcement architecture (§5), and implementation guidance (§26+).
Framing
This section frames organizational decision failures as structural problems requiring empowerment infrastructure, not as moral failures requiring surveillance. The framing produces a paradigm shift: from "how do we catch people making bad decisions" to "how do we create the conditions under which good decisions are made." The six principles follow from this framing.
Adaptation
No post-lock adaptations to design paradigm. The research grounding and design principles remain stable. Graduate mechanism (§A1) and interface design (§14–§17) are specific implementations that operationalize these principles.
Readiness
This section is ready for all uses: design rationale documentation, implementation guidance, stakeholder communication, organizational change management. The paradigm shift is grounded in research and suitable for explaining why the substrate is designed differently from traditional governance tools.
Meaning Resolution
No meaning resolution for this document.
Perception Surface
No perception surface for this document. This section specifies design principles; it does not interface with external systems. External human decision-making enters through the interface layers specified in §14–§17.
Temporal Governance
No temporal governance for this document. Design paradigm is persistent. The transition between reactive and deliberative modes (graduation mechanism, §A1) operates over time, but the paradigm itself does not change.