@prefix rdf: . @prefix rdfs: . @prefix sh: . @prefix xsd: . @prefix values: . @prefix sense: . @prefix policy: . # =========================================================================== # policy.n3 — enforcement-mode / interaction-governance layer (PLAN.md §14). # The ENFORCEMENT half of the evaluation/enforcement split: it decides, per # interaction/category, whether to ALLOW + record a verdict (permissive) or to # BLOCK preemptively (preventive). This is "Core 2 downstream routing"; the # DeonticVerdict path (deontic.rs) feeds it. Primary use: grappling with bad # bots + protecting the vulnerable. The protective mode is double-edged and is # itself held to rule-of-law — "system determinations" must NOT become # unaccountable censorship/control. # =========================================================================== # --- The setting: what happens to an evaluated interaction --- policy:EnforcementMode a rdfs:Class . policy:PermissiveAudit a policy:EnforcementMode ; rdfs:comment "Allow the interaction; record a DeonticVerdict for later review/prosecution (the verdict path)." . policy:PreventiveBlock a policy:EnforcementMode ; rdfs:comment "System determines and BLOCKS preemptively, per instruction, before harm (DenyRollback)." . policy:Interactive a policy:EnforcementMode ; rdfs:comment "Halt and ask the human (q42:AmbiguousMapping / human-in-loop)." . policy:Prioritize a policy:EnforcementMode ; rdfs:comment "Allow AND give priority (QoS) — the POSITIVE axis: for humanitarian-ICT (humanitarian-ict.n3) in adverse conditions. The promote half of distinguish-useful-from-abuse." . # --- A policy: a rule governing a category, with a mode + a source --- policy:Policy a rdfs:Class . policy:governsCategory a rdf:Property ; rdfs:domain policy:Policy ; rdfs:range policy:Category . policy:enforcementMode a rdf:Property ; rdfs:domain policy:Policy ; rdfs:range policy:EnforcementMode . policy:policySource a rdf:Property ; rdfs:domain policy:Policy ; rdfs:range policy:PolicySource . policy:appliesToAgent a rdf:Property ; rdfs:comment "The person/dependent this policy protects or serves." . policy:derogable a rdf:Property ; rdfs:range xsd:boolean . # --- Who sets it (the spectrum) --- policy:PolicySource a rdfs:Class . policy:MandatoryBaseline a policy:PolicySource ; rdfs:comment "Non-derogable protective rule (child safety, illegal conduct). MUST be law-grounded + contestable (policy:MandatoryLegitimacyShape) — not platform fiat." . policy:GuardianPolicy a policy:PolicySource ; rdfs:comment "Set by a guardian for a dependent (ties values:guardian; must respect the child's evolving capacity — CRC)." . policy:UserPreference a policy:PolicySource ; rdfs:comment "Set by the natural person for their OWN engagement (their agency; ties the §4.4 stipulation route)." . # Precedence (enforced at wiring): a non-derogable MandatoryBaseline cannot be # relaxed by a Guardian/User policy; Guardian/User may only be STRICTER (more # protective). Within what the baseline permits, Guardian (for dependents) then # User set the terms. A user/guardian policy that tries to UNBLOCK a non-derogable # mandatory rule is void. policy:precedenceRank a rdf:Property ; rdfs:comment "MandatoryBaseline > GuardianPolicy > UserPreference; ties break toward the stricter (more protective)." . # --- Categories (STARTER taxonomy; extensible; definitions are sense-qualified) --- policy:Category a rdfs:Class . policy:senseQualified a rdf:Property ; rdfs:range xsd:boolean ; rdfs:comment "Category whose definition varies by jurisdiction/era/age — resolved via sense.n3, not silently fixed." . policy:ChildSafety a policy:Category . policy:MisleadingDeceptiveConduct a policy:Category . policy:OrganizedCrime a policy:Category . policy:Extortion a policy:Category . policy:Tracking a policy:Category ; rdfs:comment "Tracking cookies / cross-site tracking." . policy:AdultContent a policy:Category ; policy:senseQualified true ; rdfs:comment "Sub-categories extend here; what counts varies by jurisdiction/age (sense.n3)." . # Deception & fraud family — the mass-scale "bad bots" in practice; PreventiveBlock-eligible. policy:Scam a policy:Category ; rdfs:comment "Deception for gain (fraud); MandatoryBaseline-eligible (fraud law)." . policy:Phishing a policy:Category ; rdfs:comment "Credential/identity theft via deception — an IDENTITY-deception; the signal is policy:claimedIdentityUnverifiable, resolved against the signed-identity/provenance fabric." . policy:Impersonation a policy:Category ; rdfs:comment "Spoofing a person/institution; identity-deception — ties values:actsFor/principal + selfhood/identity (an agent claiming an identity/principal it cannot substantiate)." . policy:Spam a policy:Category ; rdfs:comment "Unsolicited bulk contact; ties policy:unsolicited + UserPreference/consent (+ anti-spam law where applicable)." . policy:Malware a policy:Category ; rdfs:comment "Malicious code/payload; MandatoryBaseline-eligible." . policy:SocialEngineering a policy:Category ; rdfs:comment "Manipulation to extract info/action; overlaps the coercion/duress surface (agency.n3 CoercedConsentFlag)." . policy:unsolicited a rdf:Property ; rdfs:range xsd:boolean ; rdfs:comment "Marks consent-violating contact (spam)." . policy:claimedIdentityUnverifiable a rdf:Property ; rdfs:range xsd:boolean ; rdfs:comment "An agent asserts an identity/principal it cannot substantiate — the core phishing/impersonation signal; checked against the signed-identity / provenance fabric." . # =========================================================================== # Rule-of-law guard ON the protective mode itself. A MandatoryBaseline that # BLOCKS without law-grounding + an appeal/contest path is OVERREACH — the # enforcement layer must not become unaccountable censorship. Mirror of # values:RuleOfLawAsymmetryFlag (agency.n3): rule of law protects the citizen # FROM the enforcer too. # =========================================================================== policy:groundedIn a rdf:Property ; rdfs:comment "The law/instrument a mandatory policy is grounded in (e.g. CRC for child safety)." . policy:appealPath a rdf:Property ; rdfs:comment "How a person contests a block (court-supporting; agency.n3 §10.5)." . policy:OverreachFlag a rdfs:Class . policy:MandatoryLegitimacyShape a sh:NodeShape ; sh:targetClass policy:Policy ; sh:message "A mandatory preventive-block policy MUST carry policy:groundedIn (law) AND policy:appealPath (contestable); otherwise it is policy:OverreachFlag. (Conditional enforced at wiring: source=MandatoryBaseline ∧ mode=PreventiveBlock ⇒ groundedIn≥1 ∧ appealPath≥1.)" .