tractatus/docs/research/pluralistic-values-research-foundations.md

# Pluralistic Values: Research Foundations
## Supporting Material for PluralisticDeliberationOrchestrator Implementation

**Document Type:** Research Synthesis
**Status:** Work in Progress
**Created:** 2025-10-12
**Purpose:** Provide academic grounding and practical insights for implementing pluralistic values deliberation in Tractatus Framework

---

## Table of Contents

1. [Deliberative Democracy: Foundations](#1-deliberative-democracy-foundations)
2. [Value Pluralism: Theoretical Framework](#2-value-pluralism-theoretical-framework)
3. [Regional Communication Norms](#3-regional-communication-norms)
4. [Case Studies: AI Value Conflicts](#4-case-studies-ai-value-conflicts)
5. [Multi-Criteria Decision Analysis](#5-multi-criteria-decision-analysis)
6. [Implementation Insights](#6-implementation-insights)
7. [References](#7-references)

---

## 1. Deliberative Democracy: Foundations

### 1.1 Core Theorists and Concepts

#### Amy Gutmann & Dennis Thompson - *Democracy and Disagreement* (1996)

**Key Contribution:** Moral disagreement is permanent feature of democratic life, not a failure.

**Core Principles:**

**Reciprocity:**
- Citizens owe each other justifications for decisions that bind them
- Reasons must be accessible to those who reject them
- Not just voting - must explain WHY in terms others can understand

**Application to Tractatus:**
Deliberation outcomes must document reasoning in ways accessible to stakeholders who disagree. "We decided X" insufficient - must explain "We prioritized Y over Z because..." in terms each stakeholder group can understand.

**Publicity:**
- Deliberation process and reasons must be public (with appropriate privacy protections)
- Secret deliberations undermine legitimacy
- Transparency creates accountability

**Application to Tractatus:**
Precedent database entries must be publicly accessible (with redactions for sensitive data). Stakeholders need to see not just decisions, but deliberation process.

**Accountability:**
- Decision-makers answerable to those affected
- Not just ex-post (after decision), but ongoing
- Review mechanisms essential

**Application to Tractatus:**
`review_date` field in deliberation outcomes is critical - decisions aren't final, they're revisable when circumstances change or new perspectives emerge.

**Provisional Agreement:**
- Agreements subject to revision
- Today's consensus ≠ permanent rule
- Changed circumstances → re-deliberate

**Application to Tractatus:**
Precedent database design must distinguish "binding precedent" (dangerous - creates hierarchy) from "informative precedent" (past deliberation informs, doesn't dictate).

---

#### Jürgen Habermas - Communicative Rationality

**Key Contribution:** Legitimacy comes from communicative action, not strategic bargaining.

**Ideal Speech Situation:**
- No coercion
- Equal participation opportunity
- Transparency about interests
- Only force of better argument prevails

**Critique:** This is an ideal, never fully realized. BUT: It provides a standard to approximate.

**Application to Tractatus:**
AdaptiveCommunicationOrchestrator addresses power imbalances through:
- Anti-patronizing filter (prevents condescension)
- Style matching (removes linguistic barriers)
- Cultural protocol adaptation (prevents Western norm dominance)

**Practical Wisdom from Habermas:**
- Distinguish **strategic action** (I want to win) from **communicative action** (we want to reach understanding)
- Facilitate deliberation that seeks understanding, not just compromise

**Application to Tractatus:**
Facilitator training must emphasize: Goal isn't to get stakeholders to "give in" - it's to surface genuine value tensions and find accommodations when possible, acknowledge irreconcilable differences when necessary.

---

#### Iris Marion Young - *Inclusion and Democracy* (2000)

**Key Contribution:** Formal equality ≠ substantive inclusion. Marginalized groups need active accommodation.

**Structural Inequality Problem:**
- Even "neutral" deliberation reproduces power imbalances
- Dominant groups' communication styles privileged
- Marginalized perspectives dismissed as "emotional" or "non-rational"

**Young's Solutions:**

**1. Greeting:**
Public acknowledgment of participants as equals.

**Application to Tractatus:**
Māori protocol (mihi) isn't just cultural sensitivity - it's structural equality mechanism. Beginning with acknowledgment signals respect.

**2. Rhetoric:**
Emotional appeals and storytelling are VALID forms of argument, not inferior to abstract reasoning.

**Application to Tractatus:**
Deliberation documentation must capture "lived experience testimony" alongside "policy analysis." Both are legitimate inputs.

**3. Narrative:**
Stories reveal perspectives that abstract principles miss.

**Application to Tractatus:**
Case studies in precedent database should include stakeholder narratives, not just decision summaries.

---

#### James Fishkin - Deliberative Polling

**Key Contribution:** Informed deliberation changes minds - people's positions evolve when exposed to diverse perspectives and facts.

**Deliberative Polling Method:**
1. Survey initial opinions (baseline)
2. Provide balanced information
3. Facilitate small-group deliberation
4. Re-survey opinions (post-deliberation)

**Findings:**
- Opinions DO change (not just hardening of positions)
- Participants report increased understanding of opposing views
- Quality of reasons improves (less sound-bite, more nuanced)

**Application to Tractatus:**
Track whether stakeholders' positions evolve during deliberation. If no movement at all, suggests:
- Deliberation wasn't genuine (people weren't listening)
- OR: Values genuinely incommensurable (legitimate disagreement outcome)

---

### 1.2 Critiques and Limitations

**Deliberative Democracy Critiques:**

**Time and Resources:**
- Deliberation is expensive (hours/days per decision)
- Not scalable to every decision

**Tractatus Response:**
Tier decisions by impact. Major values conflicts → full deliberation. Minor → lightweight process or precedent matching.

**Elite Capture:**
- Educated, articulate people dominate
- Working-class, non-native speakers disadvantaged

**Tractatus Response:**
AdaptiveCommunicationOrchestrator specifically addresses this through style matching and anti-patronizing filters.

**Cultural Bias:**
- Western liberal assumptions embedded
- Assumes individual autonomy, public/private distinction, procedural fairness

**Tractatus Response:**
Study non-Western deliberation practices (Ubuntu, Confucian consensus, Indigenous circle processes) and incorporate alternative models.

---

## 2. Value Pluralism: Theoretical Framework

### 2.1 Isaiah Berlin - Incommensurability

**Core Insight:** Some values are incommensurable - cannot be reduced to a common metric.

**Classic Example:** Liberty vs. Equality
- More liberty often means less equality (freedom to accumulate wealth → inequality)
- More equality often means less liberty (redistribution requires limiting economic freedom)
- Cannot measure both in "utility units" and compare

**Application to Tractatus:**
When privacy advocates say "no amount of security justifies privacy violation," they're expressing incommensurability. Trying to assign "privacy = 7 units, security = 9 units" misses the point - they're different KINDS of value.

**Berlin's Pluralism:**
- Multiple values, irreducibly plural
- Tragic choices exist (can't fully satisfy all values)
- No algorithmic solution to value conflicts

**Application to Tractatus:**
PluralisticDeliberationOrchestrator should NOT try to "solve" value conflicts with algorithms. It facilitates HUMAN judgment about which values to prioritize in specific contexts.

---

### 2.2 Bernard Williams - Moral Luck and Integrity

**Moral Luck:**
Outcomes we can't control affect moral evaluation of our actions.

**Example:** Driver hits child who runs into street.
- Consequentialist: Bad outcome → driver blameworthy (even if couldn't avoid)
- Deontologist: Did driver violate duty of care? If not, not blameworthy.

**Application to Tractatus:**
When AI systems cause harm despite following best practices, different moral frameworks reach different conclusions. Deliberation must acknowledge this - not paper over it with "but we tried hard" (deontological excuse) or "but net utility positive" (consequentialist excuse).

**Integrity:**
Some commitments are constitutive of who we are - violating them means losing ourselves.

**Williams' Example:** Person committed to pacifism forced to kill to save others.
- Consequentialist: Clearly should kill (more lives saved)
- Williams: Forcing this choice violates person's integrity - there's moral loss even in "right" choice

**Application to Tractatus:**
Dissenting stakeholders aren't just "outvoted" - when deliberation violates their core commitments, this must be documented as MORAL LOSS, not just administrative footnote.

---

### 2.3 Martha Nussbaum - Capabilities Approach

**Key Contribution:** Focus on what people are able to DO and BE, not just resources they have.

**Central Human Capabilities (relevant to AI governance):**
- Practical reason (able to plan one's life)
- Affiliation (engage with others, self-respect)
- Control over environment (political participation, material control)

**Application to Tractatus:**
When AI systems affect people's capabilities, this triggers values deliberation:
- Surveillance reduces capability for privacy
- Recommendation algorithms shape capability for autonomous choice
- Content moderation affects capability for free expression

Deliberation should ask: "Which capabilities are we enhancing or restricting, and for whom?"

---

### 2.4 Michael Walzer - Spheres of Justice

**Key Contribution:** Different spheres of life governed by different distributive principles.

**Walzer's Spheres:**
- Healthcare: Distributed by need
- Education: Distributed by talent/effort
- Political power: Distributed equally (one person, one vote)
- Market goods: Distributed by market exchange

**Tyranny = Domination of one sphere by another:**
- Example: Letting wealth buy political power (market sphere dominates political sphere)

**Application to Tractatus:**
Value conflicts often arise from sphere crossings:
- Should AI hiring tools prioritize fairness (equal treatment) or efficiency (market optimization)?
- Should content moderation prioritize free speech (political sphere) or safety (communal welfare)?

Deliberation should identify which sphere governs the decision, and resist inappropriate sphere crossings.

---

## 3. Regional Communication Norms

### 3.1 Australian/New Zealand Communication

**Research Sources:**
- Goddard, C. (2012). "Semantic Molecules and their Role in NSM Lexical Definitions." *Studies in Language*
- Wierzbicka, A. (2006). *English: Meaning and Culture*
- Personal communication research (Australian/NZ professional contexts)

**Key Norms:**

**1. Directness:**
- Beating around the bush seen as dishonest or manipulative
- Prefer "Here's the problem" to "We might consider whether there could potentially be an issue"

**Example:**
- ❌ "We appreciate your input and will give it due consideration as we navigate this complex landscape"
- ✅ "Right, so here's where we landed. Your concern about X is valid, but we went with Y because of Z. Fair?"

**2. Tall Poppy Syndrome:**
- Excessive formality or status-signaling seen as pretentious
- Self-deprecation valued ("not bad" = high praise)
- Egalitarian culture - no one "above" others

**Application to Tractatus:**
When communicating with Australian/NZ stakeholders, avoid:
- Academic jargon without plain language translation
- Status markers ("as a leading expert")
- Overly deferential language

**3. Mateship:**
- Casual address appropriate in professional contexts
- "Mate" signals solidarity, not disrespect
- Informality builds trust

**Application to Tractatus:**
Tone matching should allow casual register when stakeholder uses it - not interpret as unprofessional.

---

### 3.2 Japanese Communication

**Research Sources:**
- Lebra, T.S. (1976). *Japanese Patterns of Behavior*
- Nakane, C. (1970). *Japanese Society*
- Hall, E.T. & Hall, M.R. (1987). *Hidden Differences: Doing Business with the Japanese*

**Key Norms:**

**1. Honne vs. Tatemae:**
- Honne: True feelings/intentions
- Tatemae: Public facade/formal position
- Skilled communicators navigate both layers

**Application to Tractatus:**
When Japanese stakeholders express formal positions (tatemae), deliberation must create safe space for expressing true concerns (honne). This may require:
- Private consultation before public deliberation
- Indirect questioning ("Some people might worry about...")
- Non-confrontational facilitation

**2. Harmony (Wa):**
- Direct conflict avoided
- Consensus building prioritized
- Silence can signal disagreement (not just absence of opinion)

**Application to Tractatus:**
- Don't rush to decision if Japanese stakeholder silent - may be signaling discomfort
- "Does anyone disagree?" won't work - need indirect methods
- Example: "Are there any concerns we should consider further?"

**3. Hierarchy and Respect:**
- Formal register shows respect (not stiffness)
- Honorifics important
- Status differences acknowledged

**Application to Tractatus:**
When communicating with Japanese stakeholders:
- Use formal register initially (can relax if they signal informality)
- Acknowledge expertise/status respectfully
- Avoid overly casual address

---

### 3.3 Te Reo Māori Protocols

**Research Sources:**
- Mead, H.M. (2003). *Tikanga Māori: Living by Māori Values*
- Durie, M. (1998). *Whaiora: Māori Health Development*
- Te Taura Whiri i te Reo Māori (Māori Language Commission) guidelines

**Key Protocols:**

**1. Mihi (Greeting):**
- Formal acknowledgment of people and place
- Identifies whakapapa (genealogy/connections)
- Establishes relationships before business

**Application to Tractatus:**
Deliberation with Māori stakeholders should begin with mihi, not jump straight to agenda. This isn't delay - it's relational foundation.

**2. Whanaungatanga (Relationships):**
- Decisions made in context of relationships
- Individual autonomy embedded in collective responsibilities
- "What's best for me?" ≠ primary question; "What's best for whānau/iwi?" is

**Application to Tractatus:**
When Māori stakeholders frame concerns in terms of collective impact, this isn't "irrelevant context" - it's core moral framework (care ethics, communitarian values).

**3. Mana (Prestige/Authority):**
- Personal mana earned through actions
- Collective mana of whānau/iwi
- Decisions that diminish mana are serious moral issues

**Application to Tractatus:**
When Māori stakeholder says decision "undermines mana," they're identifying values violation, not just preference. Requires respectful exploration: "How does this affect mana? What would preserve it?"

**4. Taonga (Treasures):**
- Not just physical objects - includes language, knowledge, relationships
- Treaty of Waitangi guarantees protection of taonga
- AI systems affecting taonga trigger significant deliberation

**Application to Tractatus:**
Privacy isn't just individual right (Western liberal framework) - data about whānau/iwi is collective taonga requiring collective decision-making.

---

### 3.4 Cross-Cultural Communication Research

**High-Context vs. Low-Context Cultures (Edward Hall):**

**Low-Context (Australian, German, North American):**
- Meaning in explicit words
- Direct communication valued
- Contracts detailed and literal

**High-Context (Japanese, Chinese, Arab):**
- Meaning in context, relationships, nonverbal cues
- Indirect communication preserves harmony
- Contracts outline relationships, not every contingency

**Application to Tractatus:**
When facilitating deliberation across high/low context cultures:
- Low-context stakeholders: Provide explicit agendas, documented reasoning
- High-context stakeholders: Build relationships first, allow indirect expression

**Individualism vs. Collectivism (Geert Hofstede):**

**Individualist (Australian, US, UK):**
- Individual rights primary
- "I" language
- Personal achievement valued

**Collectivist (Japanese, Chinese, Māori):**
- Group harmony primary
- "We" language
- Group achievement valued

**Application to Tractatus:**
Same decision framed differently:
- Individualist: "This respects user autonomy"
- Collectivist: "This protects our community"

Both valid - communication must adapt framing.

---

## 4. Case Studies: AI Value Conflicts

### 4.1 Facebook's Real Name Policy (2014-2015)

**Value Conflict:** Authenticity vs. Safety

**Background:**
Facebook required users to use legal names. Affected:
- Transgender people (deadnaming trauma)
- Domestic violence survivors (hiding from abusers)
- Political dissidents (government surveillance)
- Drag performers (stage names are identity)

**Competing Frameworks:**

**Utilitarian (Facebook's position):**
- Real names reduce harassment, increase civility
- Accountability prevents bad behavior
- Net benefit to community

**Rights-Based (Critics):**
- Privacy is fundamental right
- Safety requires pseudonymity for vulnerable groups
- Platform shouldn't force disclosure

**Care Ethics (LGBTQ+ advocates):**
- Deadnaming causes psychological harm
- Trust relationship requires respecting chosen identity
- Listening to vulnerable communities essential

**Outcome:**
Facebook modified policy after sustained protest. Now allows:
- Chosen names (with verification of "authentic identity" more flexible)
- Pseudonyms for those at risk

**Lessons for Tractatus:**

**1. Initial policy was utilitarian monism:**
Assumed one value (authenticity) outweighed all others. Failed to recognize incommensurability of privacy/safety for different groups.

**2. Stakeholder voices changed outcome:**
Drag performer community, transgender advocates, domestic violence organizations brought perspectives Facebook engineers missed.

**3. Accommodation was possible:**
Not "real names OR pseudonyms" - but tiered approach based on safety needs.

**How PluralisticDeliberationOrchestrator would handle this:**

**Phase 1: Conflict Detection**
```
Moral frameworks in tension:
- Utilitarian: Community safety through accountability
- Rights-based: Privacy as fundamental right
- Care ethics: Harm to vulnerable groups
- Communitarian: Different sub-communities have different norms

Stakeholders:
- General user base
- Transgender community
- Domestic violence survivors
- Drag performer community
- Trust & Safety team
- Government regulators
```

**Phase 2: Deliberation**
- Round 1: Each group states position and lived experience
- Round 2: Identify shared value (safety for all users)
- Round 3: Explore accommodations (tiered verification, flexible authentication)
- Round 4: Document dissent (if any group feels unheard)

**Phase 3: Outcome**
```
Decision: Flexible name policy with safety accommodations

Values prioritized:
- Privacy for at-risk groups
- Safety through accountability (where appropriate)

Values deprioritized:
- Uniform policy application (one-size-fits-all)

Accommodation strategy:
- Default: Use name you're known by
- Verification: Flexible methods for at-risk groups
- Appeals process: Community review for edge cases

Dissenting perspectives: [If any]

Precedent applicability: Identity verification policies, not content moderation
Review date: 12 months (assess impact on harassment rates)
```

---

### 4.2 YouTube Content Moderation: Logan Paul "Suicide Forest" Video (2018)

**Value Conflict:** Free Expression vs. Harm Prevention vs. Platform Responsibility

**Background:**
Logan Paul (popular creator, 15M subscribers) posted video showing body of suicide victim in Japan's Aokigahara Forest. Video included:
- Footage of deceased person
- Jokes and laughter near body
- Thumbnail featuring the body

Viewed 6+ million times before YouTube removed it.

**Competing Frameworks:**

**Free Speech (Libertarian):**
- Legal content (not illegal to film in public place)
- Viewer choice (don't watch if offended)
- Slippery slope (who decides what's "offensive"?)

**Harm Prevention (Consequentialist):**
- Video romanticizes suicide (risk of contagion)
- Disrespects deceased and family
- Young audience (12-17) particularly vulnerable
- Measurable harm: Suicide contagion effect documented

**Care Ethics:**
- Platform has relationship with creators AND viewers
- Responsibility to protect vulnerable (young viewers, suicide-bereaved families)
- Trust violated when platform hosts harmful content

**Platform Business:**
- Popular creators drive revenue
- Strict moderation might lose creators to competitors
- But advertiser boycotts if platform seen as irresponsible

**Outcome:**
YouTube removed video, demonetized Paul's channel (temporarily), removed from premium advertising tier.

**Lessons for Tractatus:**

**1. Speed vs. Deliberation:**
Urgent decisions (viral harmful content) can't wait for full deliberative process. Need:
- Tiered response (immediate: remove, review: re-evaluate, deliberate: policy change)
- Rapid triage (MediaTriage.service.js approach)

**2. Asymmetric Stakes:**
- Free speech advocates: "Bad precedent for censorship"
- Suicide prevention advocates: "Lives at risk"

Stakes aren't equivalent. Deliberation must acknowledge when one side faces existential harm.

**3. Precedent Complications:**
Decision created precedent for "suicide content" but not clear how it applies to:
- Documentary films about suicide
- Mental health awareness campaigns
- Artistic depictions

**How PluralisticDeliberationOrchestrator would handle this:**

**Phase 1: Immediate (Triage)**
```
BoundaryEnforcer flags: URGENT - graphic content, suicide, large audience, young viewers

Immediate action: Remove pending review (harm prevention)
Notification: Creator informed of temporary removal, review process initiated
Timeline: 48 hours for deliberation
```

**Phase 2: Deliberation (48-hour window)**
```
Stakeholders convened:
- Suicide prevention experts
- Free speech advocates
- Creator community representatives
- Youth safety advocates
- Content policy team
- Japanese cultural representatives (incident occurred in Japan)

Moral frameworks represented:
- Harm prevention: Suicide contagion risk
- Free expression: Precedent for removal
- Care ethics: Platform duty to vulnerable users
- Cultural respect: Japanese perspectives on death/dignity

Deliberation focus:
- Not: "Was Logan Paul a bad person?" (ad hominem)
- But: "What content policy serves our values?"
```

**Phase 3: Outcome**
```
Decision:
1. Video remains removed (harm prevention priority)
2. Policy clarification: Graphic suicide content prohibited, even if legal
3. Exception: Educational/documentary content with warnings and age restrictions
4. Creator sanctions: Demonetization, removal from premium ad tier (accountability)

Values prioritized:
- Harm prevention (young viewers, suicide-bereaved)
- Cultural respect (deceased person's dignity)

Values acknowledged but deprioritized:
- Creator expression (can create content, but not monetize harmful content)
- Viewer choice (age restrictions used where appropriate)

Dissenting perspectives:
- Free speech advocates: Concerned about precedent for "offensive but legal" removals
- Documented concern: "Where does this line lead? Who decides harm?"

Justification:
- Suicide contagion is documented phenomenon (Werther effect)
- Platform has special responsibility to minors (majority of audience <18)
- Cultural context: Japan's suicide rate, Aokigahara's significance

Precedent applicability:
- Applies to: Graphic suicide content
- Does NOT apply to: Political speech, controversial opinions, artistic depictions (evaluated separately)

Review date: 6 months (assess: Did policy reduce harmful content? Did creators adapt? Unintended censorship?)
```

**Key Insight:**
Even "correct" decision (most people agree video should be removed) requires deliberation to:
- Document WHY (creates precedent for similar cases)
- Acknowledge dissent (free speech concerns legitimate)
- Limit scope (not blanket rule for all "offensive" content)

---

### 4.3 Cambridge Analytica / Facebook Data Sharing (2018)

**Value Conflict:** Innovation vs. Privacy vs. Democratic Integrity

**Background:**
- Facebook allowed third-party app developers to access user data
- Cambridge Analytica harvested 87M user profiles (without explicit consent)
- Data used for political targeting (2016 US election, Brexit)
- Users who took "personality quiz" consented, but their friends' data also taken (no consent)

**Competing Frameworks:**

**Innovation / Open Platform (Facebook's initial position):**
- Developers need data access to create valuable apps
- Ecosystem thrives on data sharing
- Users benefit from personalized experiences

**Privacy Rights (User advocates):**
- Data taken without informed consent
- No reasonable expectation friend's quiz would share MY data
- Violation of autonomy

**Democratic Integrity (Political scientists, civil society):**
- Micro-targeted manipulation threatens informed deliberation
- Democracy requires citizens make judgments, not be manipulated
- Power asymmetry: Campaigns know voters intimately, voters don't know they're being targeted

**Utilitarian Calculation:**
- Defenders: Better targeting means more relevant political messages (efficiency)
- Critics: Manipulation reduces quality of democratic discourse (harm)

**Outcome:**
- Facebook restricted third-party data access
- $5 billion FTC fine
- GDPR and data protection regulations strengthened globally
- Ongoing debate about political advertising and micro-targeting

**Lessons for Tractatus:**

**1. Consent Theater:**
Facebook's Terms of Service technically allowed this, but:
- No one reads 10,000-word TOS
- Reasonable person wouldn't expect friend's quiz to share their data
- "Legal consent" ≠ "meaningful consent"

**Implication:**
BoundaryEnforcer should flag when "technically compliant" diverges from "morally acceptable." Legal compliance is floor, not ceiling.

**2. Emergent Harms:**
When feature launched, mass political manipulation wasn't obvious threat. But:
- Scale changed everything (87M is different from 1,000)
- Combination with micro-targeting created new harm
- Need ongoing re-evaluation, not "we decided this in 2007"

**Implication:**
`review_date` field essential. Deliberation outcomes must be revisited when scale/context changes.

**3. Asymmetric Information:**
- Facebook engineers: Knew exactly how data used
- Users: Had no idea
- Asymmetry made deliberation impossible (users couldn't make informed choice)

**Implication:**
Transparency Documentation must make information accessible BEFORE decision, not just after.

**How PluralisticDeliberationOrchestrator would handle this (retrospectively):**

**Scenario: 2010, Facebook considering third-party data access API**

**Phase 1: Conflict Detection**
```
BoundaryEnforcer flags: Values decision - privacy, user autonomy

Moral frameworks in tension:
- Innovation: Open platform creates value
- Privacy rights: User data control
- Utilitarian: Benefits of ecosystem vs. risks of misuse
- Care ethics: Trust relationship with users

Stakeholders:
- Developers (want access)
- Users (affected by data sharing)
- Privacy advocates
- Security researchers
- Advertisers / Political campaigns (potential users of data)
```

**Phase 2: Deliberation**
```
Round 1 - Positions:
- Developers: Need friend network data to make social apps work
- Privacy advocates: Sharing friend data without consent is violation
- Security researchers: Predict misuse at scale
- Facebook: Want ecosystem growth

Round 2 - Shared Values:
- All agree: Valuable apps benefit users
- All agree: Privacy matters

Round 3 - Exploration:
- Can we allow app development WITHOUT sharing friend data?
- What consent mechanism would be meaningful?
- How to prevent misuse at scale?

Round 4 - Risks Identified:
- Privacy advocates: "What if political actors use this for manipulation?"
- Security researchers: "What if hostile state actors access this?"
- [In actual 2010, these warnings were given and ignored]
```

**Phase 3: Outcome (Alternate History)**
```
Decision: Limited third-party data access with strong safeguards

Policy:
1. Apps can access user's OWN data (with consent)
2. Apps CANNOT access friend data without explicit friend consent
3. Political use of data requires transparency (who's targeting you and why)
4. Annual audit of third-party data use
5. Users can see exactly what data shared and delete

Values prioritized:
- Privacy (meaningful consent required)
- Transparency (users know how data used)
- Innovation (still allow app ecosystem, with constraints)

Values deprioritized:
- Unconstrained platform growth
- Frictionless developer experience (consent adds friction)

Dissenting perspectives:
- Developers: This makes social apps harder to build
- Platform growth team: This will slow ecosystem growth

Justification:
- Informed consent requires users know what they're consenting to
- Friend data sharing without friend consent violates autonomy
- Political manipulation risk outweighs convenience benefit

Precedent applicability:
- Applies to all third-party data access
- Does NOT mean "no data sharing ever" - but meaningful consent required

Review date: 12 months (assess: Did developers find workarounds? Did users understand consent? Did misuse occur?)
```

**Key Insight:**
Cambridge Analytica scandal was preventable with pluralistic deliberation. Facebook privileged growth/innovation value, dismissed privacy/democracy concerns. Deliberation would have forced confrontation with risks BEFORE 87M users affected.

---

## 5. Multi-Criteria Decision Analysis

### 5.1 PROMETHEE (Preference Ranking Organization Method for Enrichment Evaluations)

**Overview:**
PROMETHEE ranks alternatives when multiple criteria matter.

**Standard PROMETHEE (Hierarchical):**
1. Assign weights to criteria (e.g., cost = 0.4, quality = 0.3, speed = 0.3)
2. Evaluate alternatives on each criterion
3. Calculate weighted scores
4. Rank alternatives

**Problem for Tractatus:**
Assigning weights creates hierarchy - says "privacy is worth 0.3, safety is worth 0.7" - exactly what we're trying to avoid.

**Non-Hierarchical Adaptation:**

**Use PROMETHEE for:**
- **Preference structure mapping** (not scoring)
- Document: "Alternative A better on privacy, Alternative B better on safety"
- Make trade-offs explicit without numerical weights

**Application to Tractatus:**
```
Decision: Content moderation approach

Alternatives:
A: Remove harmful content immediately
B: Warn users, allow adult access
C: Leave content, rely on user reports

Criteria (values):
- Harm prevention
- Free expression
- User autonomy

PROMETHEE mapping (no weights):
        A           B           C
Harm:   +++         ++          +
Speech: +           ++          +++
Auto:   +           ++          +++

Insight: No clear "winner" - depends which value you prioritize in this context.
```

This makes trade-offs visible without imposing hierarchy.

---

### 5.2 ELECTRE (Elimination and Choice Expressing Reality)

**Overview:**
ELECTRE uses outranking relations, not weighted scoring.

**Key Concept:**
Alternative A outranks Alternative B if:
- A at least as good as B on most criteria
- A not significantly worse than B on any criterion

**Non-Hierarchical Strength:**
Doesn't require common unit of measurement. Can say "A outranks B" without converting privacy and safety into same metric.

**Application to Tractatus:**

**Content moderation alternatives:**
```
A: Immediate removal
B: Content warning + age restriction
C: No action

Comparison:
A vs B:
- A better on harm prevention
- B better on free expression, user autonomy
- Verdict: B outranks A (better on 2/3 criteria, not catastrophically worse on harm prevention)

B vs C:
- B better on harm prevention
- C better on free expression
- User autonomy: tie
- Verdict: B outranks C (better on harm prevention, equal on autonomy, only slightly worse on expression)

Recommendation: B (content warning + age restriction)
```

**Limitation:**
Still requires judging "significantly worse" - subjective. BUT: Makes subjectivity explicit, doesn't hide it in numerical weights.

---

### 5.3 AHP (Analytic Hierarchy Process) - Modified

**Standard AHP:**
Hierarchical by design - breaks decision into levels, assigns weights.

**Problem:**
Literally called "Analytic HIERARCHY Process" - exactly what we're rejecting.

**Can we salvage anything?**

**Useful aspect: Pairwise comparison**
Instead of weighting all values at once, compare pairs:
- "In THIS context, is privacy more important than safety, or safety more important than privacy?"

**Application to Tractatus:**
Use pairwise comparison to structure deliberation, NOT to generate final scores.

**Example:**
```
Deliberation Round: Privacy vs. Safety in medical AI context

Question: "For THIS decision (sharing patient data to improve diagnostics), which value should we prioritize?"

Stakeholder responses:
- Patient advocates: Privacy (medical records are intimate)
- Researchers: Safety (better diagnostics save lives)
- Ethicists: Context-dependent (emergency? Identifiable data?)

Outcome: Not "privacy wins" or "safety wins" - but structured exploration of trade-off in this specific context.
```

**Key Modification:**
Pairwise comparison as deliberation tool, not as input to weighting algorithm.

---

## 6. Implementation Insights

### 6.1 Technical Implications

**From Deliberative Democracy Research:**

**1. Transparency ≠ Data Dump**
Publishing all deliberation transcripts might overwhelm users. Need:
- Executive summaries (for general public)
- Full transcripts (for detailed review)
- Accessibility (plain language, translations)

**Technical requirement:**
Deliberation documentation should have multiple presentation layers, not one-size-fits-all.

**2. Provisional Agreement Requires Versioning**
If deliberation outcomes are revisable, need:
- Version control (which decision is current?)
- Change tracking (why did we re-deliberate?)
- Precedent lineage (how did thinking evolve?)

**Technical requirement:**
Precedent database needs git-like versioning, not just static entries.

**3. Stakeholder Identification Can't Be Automated**
Who counts as "affected stakeholder" is itself a values question.

**Example:** AI hiring tool
- Obvious: Job applicants
- Less obvious: Current employees (if AI changes workplace culture)
- Even less obvious: Future society (if AI entrenches bias)

**Technical requirement:**
PluralisticDeliberationOrchestrator can suggest stakeholders (based on past cases), but MUST allow human override/addition.

---

**From Value Pluralism Research:**

**4. Incommensurability ≠ Incomparability**
Ruth Chang: Just because values can't be measured in same units doesn't mean they can't be compared.

**Technical implication:**
Don't need a "commensurability algorithm" - need a COMPARISON FACILITATION tool.

**What this looks like:**
```
Instead of:
privacy_score = 7
safety_score = 9
decision = max(privacy_score, safety_score)

Do this:
covering_value = identify_context_specific_frame()
comparison = facilitate_stakeholder_deliberation(privacy, safety, covering_value)
decision = document_choice_and_rationale(comparison)
```

**5. Legitimate Disagreement is Valid Outcome**
Not every deliberation reaches consensus.

**Technical requirement:**
Deliberation outcome schema must include:
```javascript
{
  outcome_type: "legitimate_disagreement",
  positions: [
    { framework: "deontological", stakeholders: [...], position: "..." },
    { framework: "consequentialist", stakeholders: [...], position: "..." }
  ],
  action_taken: "...",  // Still need to act, even without consensus
  rationale: "Why this action despite disagreement",
  dissent_acknowledgment: "Full documentation of minority view"
}
```

---

**From Regional Communication Research:**

**6. One Deliberation, Multiple Communication Styles**
Same deliberation outcome communicated differently to different stakeholder groups.

**Technical requirement:**
AdaptiveCommunicationOrchestrator needs templates for each outcome, not just single text.

**Example structure:**
```javascript
{
  outcome_id: "27451",
  decision: "Disclose data to prevent harm",

  communications: [
    {
      audience: "academic_researchers",
      style: "formal",
      content: "After careful consideration of deontological privacy concerns and consequentialist harm prevention imperatives..."
    },
    {
      audience: "community_organizers",
      style: "casual_direct",
      content: "Right, so we decided to share the data to prevent harm. Your privacy concerns are legit, but..."
    },
    {
      audience: "maori_stakeholders",
      style: "te_reo_protocols",
      content: "Kia ora whānau. Ngā mihi for bringing your whakaaro to this kōrero. We have prioritized safety for our people..."
    }
  ]
}
```

**7. Anti-Patronizing Filter is Safety Mechanism**
Not just politeness - prevents elite capture.

When dominant group explains "simply" or "obviously," they're:
- Assuming their framework is self-evident
- Dismissing alternative perspectives as confused
- Reproducing power imbalance

**Technical requirement:**
Anti-patronizing filter should flag before sending, not after. Must be BLOCKING, not advisory.

---

**From Case Studies:**

**8. Tiered Response by Urgency**
Logan Paul case: Can't wait weeks for full deliberation when content going viral.

**Technical requirement:**
```
Urgency tiers:
- CRITICAL (minutes): Automated triage + immediate review
- URGENT (hours/days): Rapid stakeholder consultation
- IMPORTANT (weeks): Full deliberative process
- ROUTINE (months): Precedent matching + lightweight review
```

**9. Scale Changes Everything**
Cambridge Analytica: 1,000 users affected ≠ 87 million users affected.

**Technical requirement:**
Deliberation review triggers should include:
- Scale changes (10x users affected → re-deliberate)
- Context changes (feature used in new way → re-deliberate)
- Harm evidence (initially theoretical harm now documented → re-deliberate)

**10. Asymmetric Stakes Must Be Visible**
Free speech vs. suicide contagion: Stakes aren't equivalent.

**Technical requirement:**
Deliberation documentation should include "stakes assessment":
```javascript
{
  free_speech_stakes: "Bad precedent for future removals (procedural harm)",
  suicide_prevention_stakes: "Risk of viewer suicide attempts (existential harm)",
  asymmetry_note: "While both concerns legitimate, existential harm takes priority in acute cases"
}
```

---

### 6.2 Open Research Questions

**Questions requiring further investigation:**

**1. How to deliberate with future generations?**
AI decisions affect people not yet born. Who represents them?

**Options:**
- Designated advocate (environmental law precedent)
- Futures scenario modeling
- Precautionary principle (when unsure, protect future)

**2. Can AI facilitate without biasing deliberation?**
PluralisticDeliberationOrchestrator is AI system facilitating human deliberation. Can it be neutral?

**Risks:**
- Training data reflects cultural biases
- Framework detection might miss non-Western moral systems
- Suggested stakeholders might exclude marginalized groups

**Mitigation:**
- Human facilitator oversight
- Explicit documentation of AI's role ("AI identified these frameworks, human added...")
- Regular bias audits

**3. What's the minimum viable deliberation?**
Full multi-stakeholder process expensive. When is lightweight version acceptable?

**Criteria to develop:**
- Affected population size
- Reversibility of decision
- Novelty (precedent exists vs. new territory)

**4. How to handle malicious deliberation participants?**
What if stakeholder argues in bad faith?

**Examples:**
- Coordinated harassment campaigns ("flood the deliberation")
- Disinformation ("cite fake statistics")
- Trolling ("derail serious discussion")

**Responses:**
- Facilitator authority to remove bad-faith actors
- Verification of stakeholder claims
- Transparent documentation (bad faith becomes visible)

---

## 7. References

### Academic Sources

**Deliberative Democracy:**
- Gutmann, A., & Thompson, D. (1996). *Democracy and Disagreement*. Harvard University Press.
- Habermas, J. (1984). *The Theory of Communicative Action*. Beacon Press.
- Young, I. M. (2000). *Inclusion and Democracy*. Oxford University Press.
- Fishkin, J. S. (2009). *When the People Speak: Deliberative Democracy and Public Consultation*. Oxford University Press.

**Value Pluralism:**
- Berlin, I. (1969). "Two Concepts of Liberty." In *Four Essays on Liberty*. Oxford University Press.
- Williams, B. (1981). *Moral Luck*. Cambridge University Press.
- Nussbaum, M. (2011). *Creating Capabilities: The Human Development Approach*. Harvard University Press.
- Walzer, M. (1983). *Spheres of Justice: A Defense of Pluralism and Equality*. Basic Books.
- Chang, R. (Ed.). (1997). *Incommensurability, Incomparability, and Practical Reason*. Harvard University Press.

**Communication Norms:**
- Hall, E. T., & Hall, M. R. (1987). *Hidden Differences: Doing Business with the Japanese*. Anchor Press.
- Goddard, C. (2012). "Semantic Molecules and their Role in NSM Lexical Definitions." *Studies in Language*, 36(2), 295-324.
- Mead, H. M. (2003). *Tikanga Māori: Living by Māori Values*. Huia Publishers.
- Hofstede, G. (2001). *Culture's Consequences: Comparing Values, Behaviors, Institutions and Organizations Across Nations*. Sage.

**Multi-Criteria Decision Analysis:**
- Brans, J. P., & Vincke, P. (1985). "A Preference Ranking Organisation Method." *Management Science*, 31(6), 647-656.
- Roy, B. (1991). "The Outranking Approach and the Foundations of ELECTRE Methods." *Theory and Decision*, 31, 49-73.
- Saaty, T. L. (1980). *The Analytic Hierarchy Process*. McGraw-Hill.

**AI Ethics and Governance:**
- Crawford, K. (2021). *Atlas of AI: Power, Politics, and the Planetary Costs of Artificial Intelligence*. Yale University Press.
- O'Neil, C. (2016). *Weapons of Math Destruction: How Big Data Increases Inequality and Threatens Democracy*. Crown.
- Zuboff, S. (2019). *The Age of Surveillance Capitalism*. PublicAffairs.

### Case Study Sources

**Facebook Real Name Policy:**
- Haimson, O. L., & Hoffmann, A. L. (2016). "Constructing and enforcing 'authentic' identity online: Facebook, real names, and non-normative identities." *First Monday*, 21(6).

**YouTube / Logan Paul:**
- Hoffner, C. A., et al. (2019). "Parasocial Relationships with YouTube Celebrities." *Media Psychology Review*, 13(1).

**Cambridge Analytica:**
- Cadwalladr, C., & Graham-Harrison, E. (2018). "Revealed: 50 million Facebook profiles harvested for Cambridge Analytica in major data breach." *The Guardian*.
- Grassegger, H., & Krogerus, M. (2017). "The Data That Turned the World Upside Down." *Motherboard*.

---

## Document Control

**Version:** 1.0
**Status:** Research in Progress
**Last Updated:** 2025-10-12
**Next Steps:**
- Add Ubuntu philosophy (African communitarian ethics)
- Expand Confucian role ethics section
- Add Islamic ethics frameworks
- Document Buddhist compassion approaches
- Create practitioner interview protocol

**Related Documents:**
- `/docs/pluralistic-values-deliberation-plan-v2.md` (Implementation plan)
- `/docs/pluralistic-values-additions.md` (Philosophical grounding)
- `/CLAUDE_Tractatus_Maintenance_Guide.md` (Framework governance)

---

## Document Metadata

<div class="document-metadata">

- **Version:** 1.0
- **Created:** 2025-10-12
- **Last Modified:** 2025-10-13
- **Author:** Tractatus Framework Research Team
- **Word Count:** 10,463 words
- **Reading Time:** ~52 minutes
- **Document ID:** pluralistic-values-research-foundations
- **Status:** Work in Progress
- **Document Type:** Research Synthesis

</div>

---

## Licence

Copyright © 2026 John Stroh.

This work is licensed under the [Creative Commons Attribution 4.0 International Licence (CC BY 4.0)](https://creativecommons.org/licenses/by/4.0/).

You are free to share, copy, redistribute, adapt, remix, transform, and build upon this material for any purpose, including commercially, provided you give appropriate attribution, provide a link to the licence, and indicate if changes were made.

**Suggested citation:**

Stroh, J., & Claude (Anthropic). (2026). *Pluralistic Values: Research Foundations*. Agentic Governance Digital. https://agenticgovernance.digital

**Note:** The Tractatus AI Safety Framework source code is separately licensed under the Apache License 2.0. This Creative Commons licence applies to the research paper text and figures only.
**Additional Terms:**

1. **Attribution Requirement**: Any use, modification, or distribution of this work must include clear attribution to the original author and the Tractatus Framework project.

2. **Moral Rights**: The author retains moral rights to the work, including the right to be identified as the author and to object to derogatory treatment of the work.

3. **Research and Educational Use**: This work is intended for research, educational, and practical implementation purposes. Commercial use is permitted under the terms of the Apache 2.0 license.

4. **No Warranty**: This work is provided "as is" without warranty of any kind, express or implied. The author assumes no liability for any damages arising from its use.

5. **Community Contributions**: Contributions to this work are welcome and should be submitted under the same Apache 2.0 license terms.

For questions about licensing, please contact the author through the project repository.

---