Tier 1
The Qurʾān
Primary revelation — the first and weightiest reference for every chain of reasoning. 6,236 ayat, byte-verified against the Tanzil.net reference standard.
Tahqiq.
تحقيق
Rigorous, transparent, source-first reasoning.
Cognitive Stack
12 Cognitive Operations
Every question passes through a structured sequence of reasoning steps — visible, auditable, reproducible.
1
Decomposition
Breaking complex questions into analyzable sub-questions
2
Premise Identification
Extracting the unstated assumptions behind a question
3
Implication Tracing
Following logical consequences through textual evidence
4
Contradiction Detection
Identifying tensions between sources or conclusions
5
Evidence Weighing
Ranking sources by authenticity, relevance, and specificity
6
Scope Determination
Defining what the evidence does and does not address
7
Temporal Sequencing
Ordering rulings and contexts chronologically
8
Absence Reasoning
Analyzing what the sources deliberately omit
9
Analogical Mapping
Connecting novel cases to established precedents
10
Confidence Calibration
Assigning proportionate certainty to each conclusion
11
Error Recognition
Detecting when reasoning has overreached the evidence
12
Strategy Selection
Choosing the right cognitive approach for each sub-question
From principles to training data
These 12 operations aren't abstract ideals. Each one is encoded into 250,000 synthetic training examples in Eve-Genesis™ (Uṣūl Edition), the dataset created by HYVE Labs specifically for Islamic reasoning. When Theo applies Decomposition or Evidence Weighing, it's executing reasoning patterns learned from this dataset — patterns that formalize what Islamic scholars have done for centuries.
The source hierarchy described above isn't a suggestion. It's enforced architecturally. Eve-Theology™ F5/reasoner always retrieves from the IPSC corpus before generating a response, and Chain of Verification (CoVe) checks every citation against the corpus after generation.
The distinction between an EBO and a fatwa isn't just a disclaimer — it's enforced by methodology guarding, a verification layer that detects when the AI drifts from evidence-based reasoning into unsupported assertion.
Foundation
The Source Hierarchy
Every reasoning chain Theo traces begins here — weighed, graded, transparent.
Tier 2
Authenticated Ḥadīth
449,415 narrations across 61 collections. Each graded, each narrator assessed, each chain structurally parsed. Theo cites proof-text with grading context visible.
Tier 3
Reasoned Analysis
Transparent inference — how premises connect, what is analogized, and where humility is owed. Every step of reasoning is visible and inspectable.
EBO vs Fatwā
An Evidence-Based Opinion is a computational analysis. A fatwā is a binding legal verdict. Theo produces the former, never the latter.
What an EBO is
- An evidence-anchored view with visible reasoning and source strata.
- A structured map of where schools align, partially agree, or diverge.
- A proportionate confidence signal tied to textual strength, not rhetoric.
- An invitation to verify: citations, grades, and inference on display.
What an EBO is NOT
- A binding legal verdict tailored to your private circumstances.
- A substitute for a qualified muftī or pastoral counsel.
- An opaque recommendation with no inspectable chain of evidence.
- A universal claim that erases legitimate ikhtilāf without naming it.
The Corpus Foundation
Every claim Theo makes traces back to the Islamic Primary Source Corpus™ (IPSC™) — 668,436 structured, attributed documents across four indexes. When Theo cites a hadith, it cites from a corpus where the isnad has been separated from the matn, the narrator has been assessed across 20 rijal works, and the grade traces to a specific scholar's opinion in a specific book.
Explore the IPSC™Experience Tahqiq reasoning
Ask a question. See the evidence. Follow the reasoning. Decide for yourself.