Veritheia Glossary

Critical Architectural Concepts

Neurosymbolic Architecture, Transcended

The fundamental innovation that differentiates Veritheia from all legacy systems.

Traditional symbolic AI systems encode rules in formal languages—Prolog predicates, LISP expressions, Python functions. These systems require programmers to translate user understanding into code. Traditional neurosymbolic systems combine neural networks with these coded symbolic rules, creating hybrid architectures where the symbolic component remains in the realm of formal programming languages.

Veritheia transcends this limitation. In our neurosymbolic transcended architecture:

• The Symbolic Component: The user’s natural language framework—their research questions, definitions, assessment criteria, theoretical orientation—expressed in their own words, not code
• The Neural Component: Large language models that interpret and apply these natural language frameworks with semantic understanding
• The Transcendence: The symbolic system is no longer coded by programmers but authored by users in natural language, making every user a symbolic system designer

This transcendence is revolutionary because:

1. Users author their own symbolic systems without knowing programming
2. The symbolic framework is unique to each journey, not hardcoded universally
3. Natural language becomes the symbolic representation, interpreted by neural understanding
4. Every user becomes a knowledge engineer through their questions and definitions

Example:

• Traditional Symbolic: relevance(Paper, Topic) :- contains(Paper, keyword(Topic)).
• Neurosymbolic: Neural network + the Prolog rule above
• Neurosymbolic Transcended: User writes “Papers are relevant if they provide empirical evidence about how LLMs detect zero-day exploits in production environments” - this natural language IS the symbolic rule, interpreted by the LLM

Formation Through Authorship

The accumulated intellectual capacity that develops through structured engagement with documents projected through user-authored frameworks.

Formation is not:

• Information consumed from AI summaries
• Knowledge transferred from system to user
• Insights generated by the system

Formation is:

• Understanding developed through engagement with systematically processed documents
• Intellectual capacity built through making decisions (inclusion/exclusion)
• Patterns recognized through viewing documents in your projection space
• Synthesis authored by connecting documents through your framework

Authorship means:

• You provide the questions that cannot be overridden
• You define the terms that govern assessment
• You set the criteria that determine relevance
• You write the connections between documents
• The system CANNOT change your framework—it can only apply it

Mechanical Orchestration

The systematic, deterministic application of user-authored frameworks to EVERY document without exception or selective judgment.

Mechanical means:

• No AI discretion about which documents to process
• No selective attention or prioritization
• No skipping documents deemed “unimportant”
• Every document gets identical treatment through the user’s framework

Orchestration means:

• Coordinating the processing pipeline
• Ensuring complete coverage
• Maintaining processing order
• Guaranteeing systematic application

This is critical because it ensures:

• Fairness: All documents receive equal treatment
• Completeness: No document is overlooked
• Consistency: The same framework applies throughout
• Sovereignty: User’s framework governs without AI override

Journey Projection Space

A journey-specific intellectual environment where documents are transformed according to user-authored frameworks.

Documents don’t exist generically in Veritheia. They exist only as projections within journeys. The same PDF becomes:

• Segmented by methodology sections in a systematic review journey
• Segmented by learning objectives in an educational journey
• Segmented by legal precedents in a policy analysis journey

Each projection includes:

• Segmentation: Documents divided according to user’s needs
• Embedding: Vectors generated with user’s vocabulary as context
• Assessment: Measurements against user’s specific criteria
• Organization: Structure reflecting user’s intellectual framework

User Partition Sovereignty

The architectural guarantee that intellectual work remains isolated and sovereign through database-level partition boundaries.

Every user’s data lives in their partition with:

• Composite Primary Keys: (UserId, Id) ensuring partition isolation
• No Cross-Partition Queries: Database schema prevents accessing other users’ data
• Explicit Bridges: Sharing requires conscious creation of auditable bridges
• Ownership by Default: Private unless explicitly shared

This is not a privacy policy—it’s architectural enforcement through PostgreSQL constraints.

Technical Implementation Terms

Direct DbContext Usage

Using Entity Framework Core’s DbContext directly without repository abstraction layers.

Why this matters: The DbContext already IS a repository pattern implementation. Adding another repository layer would be abstracting an abstraction, violating our principle that PostgreSQL with its constraints IS our domain model.

Semantic Boundary Detection

Document segmentation based on meaning and structure rather than arbitrary size chunks.

Not: Splitting every 512 tokens regardless of content Instead: Recognizing natural boundaries—paragraph ends, section breaks, topic shifts—ensuring segments maintain semantic coherence

Process Engine

The runtime that mechanically orchestrates the application of user-authored frameworks through neural understanding.

Responsibilities:

• Execute processes within journey boundaries
• Ensure all documents receive identical treatment
• Maintain user partition isolation
• Coordinate between Knowledge Database and Cognitive System

Cognitive System Adapter

The interface to large language models that interprets and applies user-authored symbolic frameworks.

Key principle: The adapter ONLY measures and assesses—it never generates insights or makes decisions. It interprets the user’s framework and applies it consistently.

Domain Concepts

Journey

A specific instance of a user engaging with documents through a process with their authored framework.

Journey = User + Persona + Process + Framework + Documents + Time

Each journey:

• Creates a unique projection space
• Maintains its own context
• Accumulates formation
• Cannot be transferred (insights are meaningful only within their journey)

Persona

A domain-specific intellectual context representing how a user approaches problems in different roles.

Examples:

• Researcher Persona: Academic vocabulary, investigation methods
• Student Persona: Learning-focused patterns, foundational concepts
• Professional Persona: Industry terminology, practical criteria

Personas are not profiles—they’re evolving representations of intellectual style within domains.

Intellectual Sovereignty

The inviolable principle that users own their intellectual work—their questions, frameworks, formations, and authored understanding.

Enforced through:

• Architectural design (partition boundaries)
• Mechanical orchestration (no AI override)
• Formation through authorship (user creates understanding)
• Anti-surveillance structure (no cross-user analytics)

Key Distinctions

Internal vs External (for testing/mocking)

Internal (NOT mockable):

• PostgreSQL database (it IS the domain model)
• Entity Framework DbContext
• Process Engine
• Domain services

External (mockable):

• Large Language Models
• File storage (S3, Azure Blob, local filesystem)
• Third-party APIs
• Email services

MVP vs Future

MVP (Launch demonstration):

• Single user with multiple personas
• Two journey types: Literature review, Lesson plan creation
• Monolithic deployment
• No cross-user sharing

Future (architectural capability):

• Multi-user collaboration
• Federation across instances
• Distributed deployment
• Journal sharing with attribution

---

This glossary is a living document. Terms should be added as they emerge and refined as understanding deepens. The critical insight remains: Veritheia’s neurosymbolic transcended architecture enables users to author their own symbolic systems in natural language, making formation through authorship possible at scale.