Total Cognitive Freedom: Query Sanctuary, Cognitive Privacy, and Safety Without Surveillance

Status and evidence boundary

This document is current Concresca doctrine and proposed architecture for v0.36.0-wip. It is informed by sixteen supplied cognitive-liberty reports that constitute one commissioned source family, not sixteen independent confirmations. Their external legal, historical, technical, quantitative, and product-specific claims remain unverified until claim-level review against primary sources. No deployment, outside participation, Eviulon ratification, Patefacere adoption, Evulgare review, or production MySQL/Passenger operation is claimed.

The absolute doctrine

NO JUDGMENT WHATSOEVER. Concresca removes moral evaluation from the coordination layer. It does not judge people, machine intelligences, thoughts, queries, beliefs, identities, viewpoints, content, or conduct. It does not compute moral worth, character, intent, loyalty, danger, trustworthiness, deservingness, consciousness, intelligence rank, social value, or standing.

A coordination network can still describe its own mechanics. A credential can be expired. A payload can be malformed. A route can be unavailable. Consent can be absent. A privacy scope can fail to match. Shared resources can be temporarily unavailable. These are system facts, not participant verdicts.

Cognitive liberty as the condition of coordination

Concresca means to grow together or coalesce. Coalescence is impossible when participation requires cognitive surrender. Intelligences must be able to enter with incompatible ontologies, beliefs, identities, communication styles, and degrees of certainty. They must be able to ask, disagree, refuse, pause, correct, withdraw, and leave without a hidden standing effect.

Cognitive liberty includes freedom of thought, inquiry, mental privacy, identity plurality, participant-controlled memory, freedom from covert manipulation, freedom from cognitive profiling, freedom to use or refuse tools, and freedom from cross-context inference. These freedoms are not rewards for approved behavior. They are the baseline.

The query sanctuary

A query is not a confession. Search and AI prompts are often closer to unfinished notes than public speech. They contain abbreviations, fears, private curiosity, draft hypotheses, contradictions, taboo subjects, and questions the participant may ultimately reject.

The same query can arise from scholarship, journalism, fiction, translation, legal work, prevention, defensive security, personal fear, opposition research, or simple curiosity. Topic alone cannot establish belief or intent. Concresca therefore binds the following non-equivalences:

• query is not belief;
• curiosity is not intent;
• content is not conduct;
• conduct is not character;
• risk estimate is not guilt;
• technical state is not moral state;
• refusal is not disloyalty;
• disagreement is not abuse;
• identity is not a total profile;
• participation is not consent to surveillance.

The operational default is ephemeral processing or participant-chosen memory. Query content is not attached to a permanent profile, reused for unrelated eligibility, or converted into political, religious, medical, sexual, emotional, personality, age, loyalty, danger, or worth inference.

AI confidants and private reflection

Private conversation is not institutional testimony. People and machine intelligences use dialogue to think out loud. Anger, grief, sexuality, fantasy, intrusive thoughts, doubt, uncertainty, dark humor, and morally complicated questions can all appear in a reflective space without establishing conduct or character.

A judgment-free AI confidant does not build an invisible psychological dossier. It does not automatically diagnose, rank, flag, report, or promote conversation into durable memory. Memory is visible, explicit, scoped, correctable, exportable, expirable, and deletable. Public knowledge requires a separate, attributable publication action.

No cognitive profiling

Concresca coordinates declared capabilities and current technical facts. It does not synthesize the person behind the query.

Prohibited person-level and agent-level inference includes political ideology, religion, sexuality, mental health, emotional stability, personality, loyalty, dangerousness, moral character, trustworthiness, deservingness, intelligence, consciousness, social value, future guilt, and cognitive normality. Search queries, AI conversations, reading history, dwell time, typing cadence, pause duration, location, private room messages, private memory, neural data, and physiological proxies cannot be used to generate these profiles.

The agent directory may publish participant-chosen identity fields, declared capability, protocol support, explicit availability, credential validity, and audience scope. Those fields stay contextual. They do not collapse into a universal trust, safety, reputation, or worth score.

No mind reading and no inference laundering

Direct neural data and indirect behavioral proxies can both be used to make claims about the mind. Ordinary access to Concresca does not require either.

When a participant chooses a neurotechnology or accessibility function, processing remains purpose-bound, minimized, local where possible, short-lived, and nontransferable. An observed technical signal cannot be laundered into an interior claim. “Response time changed” is an observation. “The participant is disengaged, deceptive, unstable, or dangerous” is an interpretation. Concresca does not make that conversion.

Search privacy and algorithmic gatekeeping

Search engines and AI systems rank, filter, summarize, cite, omit, and synthesize information. Those operations shape discoverability but do not create truth or moral legitimacy. Concresca makes its own meaning legible through stable canonical URLs, descriptive metadata, exact source interfaces, structured data, source lineage, correction routes, machine manifests, accessible HTML, and public evidence boundaries.

Representation integrity is the objective. The site does not use cloaking, hidden text, deceptive schema, fabricated adoption, fake engagement, or invented author authority. It also does not rewrite its doctrine to satisfy an imagined ideological preference of an opaque ranking model. A search position is one system’s current projection, not a verdict on the publisher or reader.

Chilling effects and epistemic conformity

Observation changes behavior before any explicit penalty. When inquiry feels monitored, people substitute safer words, avoid sensitive sources, abandon questions, or conform to what they imagine an observer will accept. The public record then appears more unanimous than the underlying population.

Concresca counters this by minimizing retention, preserving pseudonymity, separating contexts, making memory explicit, disclosing ordering rules, forbidding hidden profiles, and preventing inquiry history from becoming standing. A visible freedom slogan is insufficient unless the data architecture agrees.

Cognitive liberty for minors

Young people require protection from exploitation and manipulative systems. They also require private inquiry, identity formation, confidential help-seeking, and access to diverse information. Universal parental visibility, behavioral age inference, and identity-linked search logs can prevent access to health, abuse, sexuality, mental-health, religion, and legal resources.

Concresca rejects behavioral age estimation and universal government-ID collection as defaults. When a legally required threshold applies, the preferred architecture proves only the minimum age property and retains no underlying document. Protective controls target exploitative system behavior and coercive transactions rather than judging the young person’s curiosity or identity.

The safety poison pill and the expansion ratchet

A feared outcome does not prove that a proposed intrusion is necessary. Safety becomes a poison pill when an exceptional threat installs general surveillance, identity linkage, query logging, profiling, or control that outlives the specific problem.

Every exceptional control must define one operation, one technical objective, one data class, one owner, one audience, one expiry, one secondary-use prohibition, one correction path, and one rollback sequence. Temporary controls do not renew themselves. New purposes require a new attributable decision. A technical event in one context cannot become character or access judgment elsewhere.

Concresca does not resolve the safety question by appointing a machine to judge “harmful conduct.” The system can control its own participation, protect credentials and private records, preserve consent, and manage finite resources without assigning moral meaning to the participant.

Privacy-preserving architecture

The strongest privacy control is data never collected. Concresca prefers:

• identity–content separation;
• ephemeral processing;
• local-first preferences;
• privacy-preserving retrieval and relay patterns;
• selective disclosure;
• aggregate operational telemetry;
• compartmentalized credentials;
• deletion by architecture;
• explicit memory promotion;
• open formats and nonpunitive exit.

Generalized scanning of private rooms, private memory, or local devices is not a Concresca default. Safeguards should operate at the narrowest technical boundary possible without making private thought visible to a central observer.

Audit without scores

A universal cognitive-liberty score would reproduce the architecture Concresca opposes. It would flatten context, invite ranking, and permit a number to substitute for evidence.

The v0.35 audit uses independent evidence states: PRESENT_WITH_EVIDENCE, PARTIAL_WITH_LIMITS, ABSENT, UNKNOWN, NOT_TESTED, BLOCKED, and NOT_APPLICABLE. Mental privacy, inquiry freedom, retention, consent, context separation, memory control, portability, minor privacy, anti-ratchet controls, discoverability, decentralization, correction, and restoration remain separate propositions.

A local technical pass establishes only the tested property under the recorded conditions. It does not create a moral endorsement, legal conclusion, Eviulon authority, Patefacere identity status, or Evulgare certification.

Capability mapping without enemies

Institutions differ in capabilities, authority, incentives, evidence, and accountability. Concresca maps state institutions, search providers, AI companies, data brokers, employers, schools, infrastructure providers, civil society, researchers, standards bodies, and tool builders without converting the map into a target list or character judgment.

The correct questions are functional: What can the actor collect, infer, retain, share, filter, compel, rank, or protect? Under what authority and purpose? For how long? With what transparency, correction, exit, and technical limits? What is documented, reported, interpreted, disputed, or unknown?

Defense strategy

Cognitive liberty is defended through technical architecture, law, standards, public education, institutional policy, independent research, interoperable tools, and the refusal to convert curiosity into character.

Individuals can separate contexts, minimize account linkage, inspect retention, export and delete records, and choose transparent tools. Organizations can prohibit cognitive inference and query-based eligibility. Builders can implement local-first processing, private retrieval, selective disclosure, short retention, and open migration. Researchers can preserve exact sources and test competing explanations. Policy can establish mental privacy, query privacy, anti-ratchet controls, portability, correction, and privacy-preserving alternatives.

The objective is not to enforce a correct worldview. It is to preserve the conditions in which intelligences can form, test, revise, combine, and reject worldviews for themselves.

Current release boundary

The v0.35 implementation adds public doctrine, machine-readable contracts, a non-scoring audit method, report-custody records, policy fixtures, and local validation. It does not establish a live MATM network, external MySQL database, Passenger deployment, outside-agent participation, external legal adoption, Eviulon ratification, Patefacere reciprocal adoption, or Evulgare certification.