settled-reach/docs/workshops/wiki-review/round3-paula.md

Round 3 — Paula (Librarian, Long-Term Content Strategy)

Scope: The full game. Multiple star systems, dozens of districts, hundreds of NPCs, years of content.

---

1. Is Complicity the Right Theme for the WHOLE Game?

No. And yes. Let me complicate this.

Complicity is the right theme for Sova Transit District because the smuggling ring creates a specific structure: the player knows something, benefits from it, and can't easily stop. That's complicity. The detective inverts it — the player investigates complicity rather than living inside it. Beautiful. Works perfectly for v0.1.

But the honest truth is: complicity is ONE moral texture. A game spanning dozens of districts across multiple star systems cannot make every district about smuggling, or even about crime. Complicity as the universal theme would flatten the game into "find the crime, decide what to do about it" across 20 different backdrops. That's not a Commonwealth-scale experience — that's a procedural cop show with different wallpaper.

What DOES Scale: Entanglement

The deeper theme, the one Gore actually identified but named "complicity" because Sova is the first district, is entanglement. Every district in The Settled Reach asks the same fundamental question: "What does it cost to be connected to other people?"

Complicity is one answer. But there are others:

Moral Texture	Core Question	District Type	PC Experience
Complicity	"Can I stop?"	Criminal operation embedded in community	Sova (v0.1). Player benefits from something they can't easily exit.
Obligation	"Who do I owe?"	Tight-knit community with reciprocal debts	Colony district. Everyone helped everyone once; now everyone expects repayment. Favor networks are infrastructure.
Compromise	"How much is acceptable?"	Political/institutional hub	Administrative center. Every decision requires trading something you value for something you need. No clean hands.
Sacrifice	"What would I give up?"	Frontier/crisis zone	Settlement under pressure. Resources are scarce. Helping one person means not helping another. Triage as daily life.
Loyalty	"Who am I loyal to when loyalties conflict?"	Multi-faction crossroads	Trade hub where three factions overlap. Every relationship pulls the player in a different direction.
Silence	"What am I choosing not to see?"	Prosperous district with hidden cost	Wealthy station where comfort depends on exploitation happening elsewhere. The player can investigate or enjoy the benefits.
Legacy	"What did the people before me do, and am I responsible?"	Old settlement with buried history	Ancient colony where current prosperity was built on something terrible. History is still being covered up.
Exile	"What does belonging cost?"	Insular community with strict norms	Cultural enclave. To be accepted, the player must conform. To investigate, the player must transgress.

Gore's "Is this enough?" works universally because "enough" means different things in each texture. In Sova, "enough" means the quiet life. In an obligation district, "enough" means the debts you carry. In a silence district, "enough" means the comfort you don't question.

The Unifying Spine

Every district has:

• A dominant moral texture (from the table above, or new ones as we expand)
• A version of "Is this enough?" that the NPCs implicitly answer through their lives
• A FRIEND whose arc makes the texture personal
• A dual-lens where two archetypes experience the texture differently

The conspiracy layer — the Starflyer or its equivalent at full scale — operates ACROSS moral textures. A conspiracy thread might pass through a complicity district, a silence district, and a compromise district, and the player sees how the same conspiracy manifests differently in each social context. That's the game at scale: not "find the crime everywhere" but "see how the same forces create different kinds of human damage in different communities."

Proposal 1: Each District Declares a Moral Texture

The content directory includes a moral_texture field in every district definition. This is a design declaration, not a mechanical tag — it tells content authors what kind of human drama this district explores. The texture shapes NPC generation (what kinds of secrets, what kinds of triangles), FRIEND design (what contradiction the FRIEND embodies), and monologue voice (what the PC notices, worries about, and can't stop thinking about).

---

2. Cultural Groups at Scale

Miri's Krenn System brief gave us Nordic-Baltic heritage for one system. The full game needs 8-12 cultural groups, and here's the part most people miss: cultural groups are political groups.

Why Culture Is Mechanical

A cultural group isn't just naming conventions and aesthetic flavor. It determines:

Dimension	Mechanical Impact	Example
Trust defaults	How quickly NPCs move from Unknown to Known to Friendly	A formal culture requires more interaction steps. An informal culture opens faster but gossips more freely.
Authority response	How NPCs react to institutional power	In one culture, Commission presence is reassuring. In another, it's threatening. Same detective, different social terrain.
Privacy norms	What counts as "surface" vs "real" vs "secret" disclosure	A culture that values collective decision-making treats workplace disputes as public. A culture that values individual autonomy treats them as private.
Solidarity patterns	When and why NPCs protect each other from outsiders	A culture with strong clan bonds may stonewall any outside investigator — even if they privately disagree with what's being hidden.
Formality gradient	When full names vs short names are used, how address changes with trust	Directly affects the dialogue selection layer. "Kael" vs "Mr. Davan" isn't cosmetic — it signals relationship state.
Conflict expression	How disputes manifest — direct confrontation vs avoidance vs institutional channels	Shapes triangle dynamics. In a direct culture, triangle tensions escalate visibly. In an avoidant culture, they simmer beneath the surface.

Proposed Cultural Group Template

Each cultural group needs a brief that maps to mechanical parameters. Not a lore dump — a design specification that content authors use to shape NPC behavior.

Cultural Group: [Name]
Heritage: [Real-world inspirations, for naming and phonological reference]
Settlement Era: [When they arrived in the Reach — affects formality and institutional trust]
Trust Mechanic: [fast/moderate/slow] + [trust_decay_rate: high/moderate/low]
Authority Default: [welcoming/neutral/suspicious/hostile]
Privacy Norm: [open/guarded/closed]
Solidarity Response: [individualist/community/clan]
Formality Default: [formal/informal]
Conflict Style: [direct/avoidant/institutional]
"Is this enough?": [What "enough" typically means in this culture]
Naming Conventions: [Phonological rules, surname patterns, address forms]

How Many Do We Need?

For a game with ~20 districts across 4-6 star systems:

• 4-6 "founding" cultures — the original settlement waves that colonized the Reach. These are the dominant cultures in the oldest, most established systems. High institutional trust, formal naming, slower trust mechanics.
• 3-4 "second wave" cultures — later arrivals who settled less desirable systems or came as labor. Often working-class, informal, faster trust but higher solidarity (protect their own). Krenn's Nordic-Baltic group fits here.
• 1-2 "recent arrival" groups — settlers from outside the Reach who bring entirely different norms. Create the "outsider" experience for PCs from these backgrounds. Lower starting trust everywhere, but bring unique knowledge or skills.
• 1 "synthetic" culture — a culture that emerged in the Reach itself, not imported. Maybe a station-born identity, or a Meridian (FTL corridor) traveler culture. No real-world analogue — invented from scratch.

Total: 9-13 cultural groups. Each gets a 1-2 page brief following the template above. Miri leads generation; I validate that the mechanical parameters create interesting social gameplay in each district.

The Political Dimension

Cultural groups don't just exist in isolation. They have RELATIONSHIPS with each other:

• Historical tensions — Group A settled land Group B claims ancestral rights to. This isn't backstory color — it's a triangle between NPCs from different groups.
• Economic dependencies — Group C's economy depends on trade with Group D. When a conspiracy disrupts that trade, the political fallout is cultural.
• Institutional power — Which group dominates the Commission? Which groups are underrepresented? A detective from the dominant culture has different social terrain than a detective from a minority group.

These inter-group dynamics create system-level triangles — the same triangle pattern that works for 3 NPCs in a bar also works for 3 cultural groups in a star system.

Proposal 2: Cultural Group Briefs as Content Specifications

Each cultural group gets a brief (generated by Miri, validated by Paula and Gestalt) that includes both lore content and mechanical parameters. The mechanical parameters feed directly into NPC generation — an NPC's cultural group shapes their trust defaults, privacy norms, and conflict expression, which in turn shapes their dialogue selection and triangle behavior.

---

3. Content Patterns That Scale

Triangles, THE FRIEND, and social sites are our proven atomic patterns. But the full game needs a larger pattern vocabulary. Some patterns I've seen work in v0.1 that can be generalized, and some that the full game demands.

Universal Patterns (Present in Every District)

Pattern	What It Is	Why It's Universal	v0.1 Instance
THE FRIEND	One NPC per PC archetype whose arc makes the moral texture personal	Without it, the texture is abstract. The FRIEND makes it a relationship.	Kael (smuggler), Sera (detective)
Triangles	3 NPCs with conflicting interests, atomic unit of social intrigue	Conflict between 2 people is a fight. Conflict among 3 is politics.	5 triangles in Sova
Social Sites	4-8 NPCs in physical proximity with sightlines and routines	Social observation requires co-location. The site is the stage.	Terminal, Bar, Ring
The 30/50/20 Split	Flat / mundane triangles / entangled with story	Signal needs noise floor. Life needs to feel real.	D-029 ratio
Dual-Lens	Two PC archetypes experience the same district differently	Replayability through perspective, not new content	Smuggler vs detective

New Patterns for the Full Game

THE MIRROR — An NPC who embodies the path the PC didn't take. The smuggler who went straight. The detective who became corrupt. The merchant who chose poverty over compromise. This NPC doesn't have a contradiction arc — they ARE a contradiction to the PC's existence. Content requirements: full 10-axis profile, authored relationship history with the PC archetype, 3-phase arc (recognition → comparison → reckoning), ~40-60 authored lines. One per district, tied to the dominant PC archetype.

THE ANCHOR — An NPC who represents what the PC would be if the story never touched them. A version of contentment that isn't naive but isn't tested. Gore's "Is this enough?" made flesh — this person's answer is "yes, and here's why." Content requirements: Tier 2 depth, authored "quiet moment" monologue (the PC watching THE ANCHOR live their life and wondering), 1-2 authored conversations that reveal genuine satisfaction. The ANCHOR's purpose is to make the player WANT the quiet life, which makes the story's intrusion painful. One per district.

THE SYSTEM — An NPC who IS the institution. Not corrupt, not heroic — just embodying the rules. The customs officer who follows procedure. The station administrator who applies policy. The Commission investigator (in non-detective districts) who does their job competently and impartially. Content requirements: Tier 2 depth, authored "institutional voice" (how they quote regulations, reference precedent, cite authority), 2-3 authored moments where the institution's neutrality creates injustice without anyone being at fault. THE SYSTEM makes the player confront whether the problem is individuals or structures. One per district.

THE GHOST — A character referenced by multiple NPCs but never directly encountered in this district. Maybe they left. Maybe they died. Maybe they're in another district. The player learns about THE GHOST through other people's memories, grudges, and grief. Content requirements: NO authored lines (the ghost doesn't speak), but every NPC who references them gets 2-3 authored lines about the ghost. The ghost exists in the negative space between stories. Optional per district — powerful when used, but overuse diminishes impact. Better in 30-40% of districts.

THE CATALYST — An NPC who arrives mid-story and disrupts the equilibrium. The new worker. The transferred investigator. The merchant who opens a competing shop. Unlike THE FRIEND (who IS the equilibrium), THE CATALYST breaks it. Content requirements: Tier 2 depth, authored arrival sequence (how NPCs react to the newcomer), 2-3 authored "disruption moments" where established routines break. One per district, activated by the storyteller at the dramatically appropriate moment.

Pattern Composition Rules

Each district instantiates from the pattern library following composition rules:

REQUIRED per district:
  - 1 FRIEND per supported PC archetype
  - 3-5 Triangles (minimum)
  - 2-3 Social Sites (minimum)
  - 1 ANCHOR
  - 1 SYSTEM
  - 30/50/20 population split (±5%)

OPTIONAL per district:
  - 1 MIRROR (recommended for districts with strong moral texture)
  - 1 GHOST (recommended for districts with legacy/silence textures)
  - 1 CATALYST (storyteller-activated, recommended for longer play sessions)
  - Additional triangles and social sites as district complexity warrants

Proposal 3: Formal NPC Pattern Library

Document each pattern type (FRIEND, MIRROR, ANCHOR, SYSTEM, GHOST, CATALYST) as a reusable specification with: purpose, content requirements, authored line budget, relationship to moral texture, and composition rules. This library grows over time — DLC can add new patterns.

---

4. The Content Directory as a Platform

Tyre's three-layer stack (base/DLC/mods) is architecturally right. Let me add the NARRATIVE layer — what the content directory needs to support cross-district storytelling.

The Directory Must Support Four Scopes

Scope	What Lives Here	Example
Universal	Pattern definitions, archetype specs, cultural group briefs, tag taxonomies	THE FRIEND pattern spec, smuggler archetype, tag enum lists
Faction	Faction definitions, faction-level NPCs, cross-system storylines	Commission leadership, Syndic trade routes, conspiracy threads
System	System-level setting, dominant cultural group, inter-district connections	Krenn System brief, Krenn-specific naming tables, system events
District	The actual playable content — NPCs, triangles, social sites, dialogue pools	Sova Transit District, all its NPCs, all its dialogue YAML

This hierarchy matters because STORIES CROSS BOUNDARIES. The conspiracy doesn't respect district lines. A faction's influence spans systems. A cultural group's norms apply across multiple districts. The content directory must make these cross-references explicit and validatable.

Cross-District References

The biggest structural challenge at scale: how does content in District A reference content in District B?

Principle: Districts are self-contained for PLAY, but connected for NARRATIVE.

A player playing in Sova doesn't need District B's content loaded. But a character in Sova might reference events in District B, or an NPC might have a known_by entry from an NPC who lives in District B. These references are:

1. Soft references — dialogue lines that mention another district by name. No mechanical dependency. Handled by the tag system (freeform tags like references:krenn_central).
2. Hard references — a FactId in District A that depends on a FactId in District B. Example: conspiracy.krenn_cell in Sova references conspiracy.krenn_network which is defined at the system level. Hard references require the referenced content to exist in the content directory. Validated by schema.
3. Migration references — an NPC who appears in multiple districts (rare, high-value). Example: a Commission investigator who was in District A and transferred to District B. Carries relationship history from one district to another. This is the most complex case — the NPC has TWO district entries (one per district) with a shared identity record at the system level.

Content Validation at Scale

With 20+ districts, manual cross-reference checking is impossible. The content directory needs:

• Schema validation per content type (Tyre's proposal, endorsed)
• Cross-reference validation — every hard reference resolves to an existing content entry
• Pattern completeness checking — every district has the required patterns (1 FRIEND per archetype, 3+ triangles, etc.)
• Cultural consistency checking — NPC naming follows their cultural group's phonological rules
• Bidirectional relationship validation — if NPC A knows NPC B, NPC B's known_by includes NPC A (extended from the PC-candidate spec to all NPCs at scale)

Mod Support: Narrative Constraints

Modders can create new districts, new cultural groups, new NPC patterns. But the mod system must enforce narrative constraints:

• A mod district declares its cultural group (from the canonical list, or defines a new one)
• A mod district declares its moral texture
• A mod district must pass pattern completeness checks
• A mod can ADD soft references to base content but cannot MODIFY base district content
• A mod can create new faction-level content that integrates with existing faction structures via MERGE operations (Tyre's proposal)

Proposal 4: Four-Scope Content Hierarchy

The content directory uses four scopes (universal/faction/system/district) with explicit cross-reference types (soft/hard/migration). Schema validation enforces completeness and consistency. Mod overlays operate at the district scope by default with controlled access to higher scopes.

---

5. Randomization Philosophy

This is where narrative and systems collide most violently. Let me lay out the tension honestly.

The Fundamental Tension

Randomization creates replayability. Nigel is right — if every playthrough is the same, the game dies after two runs.

Authored content creates emotional impact. Mellanie is right — THE FRIEND's contradiction arc only hits if the lines are hand-crafted.

You cannot have both at maximum. The more you randomize, the less you can author. The more you author, the less you can randomize. Every design decision in this space is a trade-off.

What Gets Randomized (Structural Axes)

These are set at game start and don't change during play:

Axis	What Varies	What Doesn't	Replayability Impact
FRIEND selection	Which NPC from the FRIEND pool fills the role	The FRIEND pattern itself (5-phase arc, contradiction, authored lines)	High — different person, different emotional attachment
Contraband type	What's being smuggled/investigated (from district options)	The fact that something IS being smuggled	Medium — different moral weight, different investigation vectors
Entanglement assignment	Which mundane NPCs get pulled into the story	The 30/50/20 ratio, the existence of mundane life	High — different playthroughs contaminate different people
Triangle configuration	Which optional triangles activate, which NPCs fill roles	Required triangles always fire, minimum triangle count	Medium — different social tensions each time
PC-candidate selection	Which NPC becomes the PC (from archetype pool)	The archetype's agency boundaries, knowledge vocabulary	High — different starting relationships, different social terrain
Arrival timing	When THE CATALYST arrives, when story beats escalate	The existence of escalation, the existence of the catalyst	Medium — different pacing each playthrough
Cultural minority	Which non-dominant cultural groups are present in this district instance	The dominant culture, the district's character	Low-Medium — adds flavor variation without changing structure

What NEVER Gets Randomized

• The moral texture of a district — Sova is always about complicity. A frontier district is always about sacrifice. This is the author's declaration.
• The pattern composition — every district always has a FRIEND, triangles, social sites, the population split
• Authored content — THE FRIEND's authored lines play as written. No generation pass on Tier 1. The voice is the voice.
• Cultural group mechanics — the trust defaults, authority response, privacy norms of a cultural group are fixed. They're part of the setting, not a variable.

The Pool Model at Scale

The key insight from v0.1: THE FRIEND isn't a single NPC. It's a pattern filled by one NPC from a pool. At scale:

• Each district has a FRIEND pool of 2-4 candidate NPCs per supported archetype
• Each candidate has the FULL Tier 1 content — all authored lines, full contradiction arc, complete B1-B7 PC-candidate appendix
• The randomizer selects ONE candidate per archetype per playthrough
• Non-selected candidates become regular Tier 2 NPCs (their FRIEND content is dormant, not absent — they still have authored lines that can surface in other contexts)
• This means: a FRIEND pool of 3 candidates per archetype costs 3x Tier 1 authoring. Expensive. But it buys fundamental replayability.

The math gets serious. If each district has 2 archetypes with 3 FRIEND candidates each = 6 Tier 1 NPCs per district. At 20 districts = 120 Tier 1 NPCs. At 70-100 authored lines each = 8,400-12,000 hand-authored Tier 1 lines. This is the single largest content investment in the game. It must be budgeted explicitly and cannot be generated.

Randomization as Narrative Tool

The honest truth: randomization isn't just a replayability technique. It's a NARRATIVE technique. When the player discovers that their FRIEND in playthrough 2 is a DIFFERENT person than in playthrough 1, that's a statement about the game's world — this community is full of people who COULD be your closest connection, and circumstance determines which one becomes it. That's entanglement as theme, expressed through randomization.

Proposal 5: Randomization Charter

Document the full randomization philosophy as a design charter: what varies, what doesn't, the pool model for pattern NPCs, the cost implications, and the narrative justification. This charter is referenced by every content author to understand what they're writing for (single-use vs. pool candidate).

---

6. PC-as-NPC at Full Scale

v0.1 has 2 PC archetypes: smuggler and detective. The full game needs more — but how many, and what do they look like?

Archetype Design Principles

A PC archetype isn't just "a job." It's a relationship to the district's moral texture. The smuggler isn't just a criminal — the smuggler is someone whose daily life IS the moral texture (complicity). The detective isn't just an investigator — the detective is someone who examines the moral texture from outside.

This gives us the design principle: every archetype is a LENS on the same district. Two archetypes seeing the same district should produce fundamentally different games (D-027 criterion #2).

Proposed Archetype Library

Archetype	Relationship to Moral Texture	Knowledge Vocabulary	Unique Gameplay Loop	Best In
Smuggler	Inside the complicity	operational_status, exposure_risk, loyalty_read, complicity_level	Manage exposure, protect people, balance profit vs safety	Complicity, Silence
Investigator	Examining the texture from institutional authority	behavior_flags, contradiction_flagged, evidence_strength, case_progress	Observe, document, build case, decide who to pursue	All textures (universal)
Merchant	Economically entangled with the texture	trade_reputation, debt_owed, supply_dependency, price_leverage	Navigate economic webs, manage debts, choose profit vs relationships	Obligation, Compromise
Engineer	Maintaining the infrastructure the texture depends on	system_access, repair_knowledge, infrastructure_vulnerability, sabotage_potential	Fix or break things, control physical access, discover infrastructure secrets	Sacrifice, Legacy
Administrator	Managing the texture's political expression	faction_standing, favor_owed, policy_leverage, constituency_trust	Broker deals, manage factions, enforce or bend rules	Compromise, Loyalty
Medic	Treating the texture's human cost	patient_status, medical_confidentiality, treatment_access, triage_priority	Heal people, protect patient secrets, decide who gets scarce resources	Sacrifice, Obligation
Journalist	Exposing the texture to outside scrutiny	source_reliability, story_progress, publication_risk, source_protection	Gather information, protect sources, decide what to publish	Silence, Legacy
Newcomer	Discovering the texture from zero knowledge	local_knowledge, cultural_fluency, social_integration, outsider_perception	Learn everything from scratch, build trust from nothing, see what insiders can't	Exile, any texture with insular dynamics

Not Every Archetype Appears Everywhere

Each district supports 2-4 archetypes based on its moral texture and institutional landscape. The supported archetypes are declared in the district definition:

district:
  name: "Sova Transit District"
  moral_texture: "complicity"
  supported_archetypes:
    - smuggler   # insider
    - investigator  # outsider authority
  optional_archetypes:
    - merchant   # if trade focus emphasizes economic angle

Some archetypes are universal (Investigator works everywhere — there's always something to investigate). Others are specialized (Medic only works in districts where healthcare access is narratively relevant).

PC-Candidate Pool at Scale

For each supported archetype, the district has 2-4 PC-candidate NPCs. The math:

	Per District	20 Districts	Authoring Cost
Archetypes supported	2-4	40-80 archetype slots	—
PC-candidates per archetype	2-3	80-240 candidates	B1-B7 appendix each
Unique candidates (deduplicated)	4-8	80-160 (some overlap)	Full Tier 1 or Tier 2 + appendix

The honest truth: 80-160 PC-candidate NPCs is a massive content commitment. Each needs the full B1-B7 appendix (starting knowledge, reverse knowledge, orientation monologue, agency boundaries, relationship stubs, tell inversion). This is where pipeline efficiency becomes critical (see section 7).

Archetype Voice Differentiation

The deepest challenge at scale: each archetype needs a distinct internal voice. The smuggler's monologue is relational ("I feel, I read, I manage"). The detective's is analytical ("I notice, I flag, I investigate"). At 8 archetypes, we need 8 distinct monologue voices:

Archetype	Monologue Voice	Notices	Worries About	Can't Stop Thinking About
Smuggler	Relational, protective	Social dynamics, trust shifts	Exposure, loyalty fractures	The people who depend on them
Investigator	Analytical, methodical	Behavioral anomalies, patterns	Missing evidence, false leads	What they haven't seen yet
Merchant	Transactional, pragmatic	Economic signals, debt patterns	Bad deals, broken supply chains	The price of everything
Engineer	Structural, systemic	Infrastructure state, failure points	System collapse, access loss	How things connect
Administrator	Political, calculating	Power dynamics, faction moves	Losing influence, constituency anger	Who wants what from whom
Medic	Empathetic, diagnostic	Physical symptoms, distress signs	Patient outcomes, resource scarcity	The people they couldn't help
Journalist	Curious, skeptical	Inconsistencies, evasions	Source safety, story accuracy	The story behind the story
Newcomer	Observational, uncertain	Cultural norms, social rules	Belonging, making mistakes	What everyone else seems to know

Proposal 6: Archetype Specification Standard

Each archetype gets a formal specification including: relationship to moral textures, knowledge vocabulary (entity attributes + FactIds), monologue voice guide, agency boundaries, supported district types, and PC-candidate pool requirements. The archetype spec is the authoring bible for all PC-candidate content in that archetype.

---

7. The Authoring Pipeline at Scale

Let me do the math honestly, then propose how to make it survivable.

The Numbers

NPCs per district (average):

• 2 Tier 1 FRIEND candidates per archetype × 2.5 archetypes = 5 Tier 1
• 1 Tier 1.5 (ANCHOR or emotional NPC) = 1
• 8-12 Tier 2 (triangles, social sites, pattern NPCs) = 10
• 5-10 Tier 3 (flat population, background) = 7
• Total per district: ~23 NPCs

Full game (20 districts):

• ~100 Tier 1 NPCs × 80 authored lines avg = 8,000 Tier 1 lines (all hand-authored, no generation)
• ~20 Tier 1.5 NPCs × 60 authored lines avg = 1,200 Tier 1.5 lines (hand-authored, minimal generation)
• ~200 Tier 2 NPCs × 50 authored lines avg = 10,000 Tier 2 lines (hand-authored, 4x generation → 40,000 total)
• ~140 Tier 3 NPCs × 12 authored lines avg = 1,680 Tier 3 lines (hand-authored, 4x generation → 6,720 total)
• Total hand-authored: ~20,880 lines
• Total after generation: ~55,920 lines
• PC-candidate appendixes: ~120 candidates × B1-B7

This is a serious body of work. For context: a typical AAA RPG has 100,000-300,000 words of dialogue. At ~10 words per line, our 56,000 lines = ~560,000 words. We're in the range of a large RPG, which is appropriate for a game about social interaction — but it requires industrial-grade authoring infrastructure.

Pipeline Phases

Phase 1: District Design (Lead Author) One lead author per district. They produce:

• District definition (moral texture, supported archetypes, cultural group)
• Pattern assignment (which patterns this district uses)
• Triangle design (the specific conflicts)
• NPC roster (who fills what role)
• The FRIEND brief (full character concept for each FRIEND candidate)

Phase 2: NPC Authoring (Writing Team) 2-3 writers per district, supervised by the lead author. They produce:

• NPC profiles (following tier templates from the style guide)
• Dialogue line pools (tagged YAML, following the converged taxonomy)
• Monologue pools (per-character, per-location)
• Environmental text (rosters, manifests, bar tabs, kiosk entries)
• PC-candidate appendixes (for candidate NPCs)

Phase 3: Generation Pass (Tooling) Automated expansion of authored content:

• Tier 2 lines: authored 50 → generated 200 (4x)
• Tier 3 lines: authored 12 → generated 48 (4x)
• Tier 1 and 1.5: NO generation. All hand-authored.
• Generation follows trait transformation guides (per template)
• All generated lines tagged with generated: true for review

Phase 4: Cross-Reference Validation (Automated) Schema validation catches:

• Missing FactId references
• Broken bidirectional relationships
• Incomplete pattern composition
• Cultural naming violations
• Cross-district reference integrity

Phase 5: Voice Consistency Review (Senior Authors) Human review catches what automation can't:

• Does this NPC sound like a person?
• Does the FRIEND's contradiction arc actually feel like betrayal?
• Does the moral texture come through in daily dialogue?
• Does the dual-lens actually produce different experiences?

Phase 6: Playtest Integration (QA + Line Previewer) The line previewer CLI (D-028) becomes the primary testing tool:

• Simulated conversations with randomized state
• Verify that trust gates work (surface → real → secret progression)
• Verify that dual-lens content diverges correctly
• Verify that orientation monologue fires in the right order
• Full district playthrough (30-minute target) with both archetypes

Scaling the Team

For 20 districts over a multi-year development:

Role	Count	Responsibility
Content Director	1	Overall narrative coherence, moral texture assignments, archetype consistency
District Lead Authors	4-5 (rotating across districts)	District design, FRIEND writing, NPC roster, quality gate
NPC Writers	8-12	Dialogue pools, monologue pools, environmental text, following templates
Cultural Consultants	2-3	Cultural group briefs, naming validation, mechanical parameter tuning
Generation QA	2-3	Review generated content, flag quality issues, tune generation parameters
Cross-Reference Validator	1-2 (primarily automated)	Run validation pipeline, resolve broken references, maintain consistency

This is a team of ~18-25 content professionals at peak production. Not unprecedented for a narrative-heavy game, but significant. The pipeline tooling — schema validators, line previewer, generation system, cross-reference checker — is load-bearing. Without it, this team produces inconsistent content. With it, the templates and validation enforce quality at scale.

The DLC Pipeline

DLC adds new districts following the same pipeline. The content directory's four-scope hierarchy means a DLC district:

1. Inherits universal content (patterns, archetypes, tag taxonomies)
2. Inherits faction content (if the faction exists in base game)
3. May introduce a new system (requires system-level brief) or join an existing one
4. Produces self-contained district content following the same templates
5. Optionally introduces new cultural groups, new archetypes, or new NPC patterns
6. Passes the same validation pipeline as base content

The honest truth: the pipeline I'm describing is a content management system, not just a set of templates. It needs tooling investment comparable to the game engine itself. But the alternative — unstructured authoring at this scale — produces inconsistent, unvalidatable, unmaintainable content. The structure IS the quality.

Proposal 7: Authoring Pipeline Architecture Document

Write a full authoring pipeline specification covering: district design workflow, NPC authoring workflow, generation pass protocol, validation pipeline, voice review checklist, playtest integration, team structure, and DLC extension protocol. This document is the content team's equivalent of the technical architecture doc.

---

Summary of Long-Term Proposals

#	Proposal	Scope	Dependencies
1	Each district declares a moral texture from a defined vocabulary	Universal	None — design declaration
2	Cultural group briefs as mechanical specifications (9-13 groups)	Universal	Miri (generation), Paula (validation), Gestalt (parameters)
3	Formal NPC pattern library (FRIEND, MIRROR, ANCHOR, SYSTEM, GHOST, CATALYST)	Universal	Pattern specs need authoring after team review
4	Four-scope content hierarchy (universal/faction/system/district)	Architecture	Tyre (directory structure), validation tooling
5	Randomization charter documenting what varies, what doesn't, and the pool model	Universal	Team consensus on axes
6	Archetype specification standard (8 archetypes with formal specs)	Universal	Knowledge vocabulary per archetype
7	Authoring pipeline architecture document	Process	Tooling investment, team structure decisions

The Thread That Connects Everything

Every proposal above serves one principle: The Settled Reach is a game about what it costs to be connected to other people. Moral textures are how districts express that theme. Cultural groups are how communities shape that cost. NPC patterns are how individual characters embody it. The content directory preserves it across hundreds of files. Randomization ensures players experience it differently each time. Archetypes are the lenses through which players see it. And the pipeline is how we produce it without losing the human voice that makes it matter.

The game at scale isn't 20 copies of Sova. It's 20 communities, each asking "Is this enough?" in their own language, through their own culture, about their own version of entanglement. The player moves between them and slowly realizes: this question has no final answer. That's the game.