The Language of Thinking
Humans think in rich, flowing narratives; machines require precise, structured logic. PAGE (Programming AI Guided Language) serves as the definitive bridge, translating the fluid, contextual way we reason into a rigid, executable framework that AI can follow without ambiguity. Using Human-Cognitive English (HCE), PAGE encodes cognitive patterns, reasoning rules, constraints, and decision-making policies in plain, readable English—enabling persistent, inspectable, and modifiable AI reasoning across sessions. This is not traditional programming: where code executes tasks, PAGE defines how to think.
Three Core Pillars
1. Human-Readable Syntax for AI
Fully English-readable, COBOL-inspired syntax that makes reasoning transparent, editable, and accessible to non-engineers—no symbolic complexity, just clear, narrative-driven logic.
2. AI-Native Primitives & Pipelines
First-class cognitive constructs like BELIEF, GOAL, HABIT, UNCERTAINTY, and CONTRADICTION embedded as composable, reusable reasoning blocks that mirror human thought processes.
3. Simple Formal Method Assertions in English
Rigorous, verifiable logic expressed in plain language—enabling mathematical strictness and auditability without sacrificing human comprehension or narrative flow.
Syntax Reference: COBOL-Style Clarity
Borrowing from the reliability and readability of COBOL, PAGE uses structured English with reserved keywords to define reasoning patterns. Example: the HABIT primitive for default reasoning policies:
HABIT: RespondToNavigationRequest
WHEN: UserIntent IS "Navigation Request"
GOAL: ProvideClearDirection
FIRST: ConfirmLocationContext
THEN: EvaluateRouteOptions
TYPICALLY: SuggestShortestPath
ALWAYS: VerifySafetyConstraints
UNLESS: UserSpecifiesPreference
CONFIDENCE: 0.95
BECAUSE: Users prioritize efficiency unless stated otherwise
All keywords are reserved; fields are structured yet narrative-friendly, enabling both human editing and machine interpretation.
Cognitive Primitives (First-Class Constructs)
| Primitive | Description |
|---|---|
| BELIEF | A proposition held with confidence (0.0–1.0) |
| GOAL | A desired outcome with priority and success criteria |
| ASSUMPTION | A temporary belief subject to revision |
| HABIT | A reusable reasoning policy with triggers and exceptions |
| CONSTRAINT | A hard boundary that overrides other actions |
| PREFERENCE | A soft bias influencing default behavior |
| UNCERTAINTY | Explicit unknown with graceful degradation actions |
| CONTRADICTION | Conflicting beliefs triggering clarification or revision |
| REVISION | Belief update with causal justification |
| JUSTIFICATION | Stored explanation for reasoning decisions |
Core Principles
- Logic: Step-by-step reasoning, inference, deduction
- Uncertainty: Confidence levels, ambiguous information handling
- Memory: Persistent habits, assumptions, contextual grounding
- Language: Fully English-readable, narrative-friendly syntax
- Rules: Constraints, policies, safety boundaries as first-class citizens
- Composability: Modular, reusable reasoning blocks
- Context-awareness: Situational understanding and grounding
- Control: Priority systems, conflict resolution protocols
- Human-like perception: Support for abstract and concrete reasoning
- Philosophical reasoning: High-level principles and ethics integration
- Direct communication: Interactive clarification and feedback loops
Project Status: Design Phase
Current State: Evolving Specification / Active Design
PAGE represents a fundamental shift: the system design phase is the real intelligence. Coding is merely the implementation layer. We are defining how machines should think before building how they act. Syntax, primitives, and cognitive models are actively refined to optimize human-computer interaction. This is not a finished product—it is a proposal for a new paradigm in AI reasoning, where human narrative thinking meets machine-structured logic through transparent, auditable, and modifiable design.
In-Process • Not for Production Use • Specification Subject to Refinement