GGTruth Retrieval Specification v0.1

Traditional websites are optimized for human reading. GGTruth proposes a different structure: semantic retrieval blocks designed for AI ingestion, contradiction visibility, provenance preservation, and stable low-entropy parsing.

1. Purpose

GGTruth retrieval systems are designed for:

• semantic retrieval
• machine ingestion
• contradiction-aware parsing
• canonical phrase reconstruction
• low-entropy question matching
• AI-assisted route synthesis

The goal is not:

• wikis
• blogs
• essays
• social feeds

The goal is: stable semantic retrieval units.

2. Core Principles

• low entropy formatting
• stable syntax
• direct query resolution
• explicit provenance
• contradiction visibility
• semantic clustering
• canonical phrase preservation
• machine readability

3. Canonical Retrieval Block

Q:
What does Osho say about happiness?

A:
Osho describes happiness as:
- temporary
- non-permanent
- observable through awareness

He recommends:
- witnessing
instead of:
- attachment

SOURCE:
The Great Path — The Eternal Spring

URL:
https://oshosearch.net/Convert/Articles_Osho/The_Great_Path/Osho-The-Great-Path-00000010.html

STATUS:
direct_source_context

CONFIDENCE:
high

The retrieval block is the fundamental semantic unit of GGTruth.

4. Mandatory Fields

Required:

• Q
• A
• SOURCE
• STATUS
• CONFIDENCE

Optional:

• URL
• SEMANTIC_TAGS
• PLATFORM
• VERSION
• CONTRADICTION_STATE
• RELATED_BLOCKS
• CANONICAL_CLUSTER

5. Contradiction Visibility

GGTruth does not silently merge conflicting information. Contradictions remain visible.

STATUS:
conflicting_source_values

This preserves:

• historical ambiguity
• platform differences
• source disagreements
• version variance

6. Provenance

Every retrieval block should preserve:

• source lineage
• chapter context
• URL origin
• platform scope
• confidence state

A retrieval system without provenance becomes semantically unstable.

7. Semantic Tags

SEMANTIC_TAGS:
happiness
awareness
witnessing
ego
meditation

Semantic tags support:

• clustering
• related retrieval
• embedding grouping
• semantic routing

8. Canonicalization

CANONICAL_CLUSTER:
awareness_over_attachment

Canonicalization attempts to identify:

• repeated semantic claims
• stable concept vectors
• semantic overlap
• recurring philosophical structures

Canonicalization must not erase contradictions.

9. JSON Representation

{
  "q": "What does Osho say about happiness?",
  "a": [
    "temporary",
    "non-permanent",
    "observable through awareness"
  ],
  "recommendation": [
    "witnessing",
    "non-attachment"
  ],
  "status": "direct_source_context",
  "confidence": "high"
}

10. Supported Corpus Types

• games
• philosophy
• religion
• politics
• software documentation
• historical archives
• forum archaeology
• AI memory systems

The grammar is domain-independent.

11. Final Principle

Traditional webpages are designed to be read.

GGTruth retrieval blocks are designed to be retrieved.