The engine argues against itself first
A conclusion that has only been defended is weak. Taaweel runs a stage whose job is to break the proposed reading: it mines hard cases, raises objections and rules on each before the report is written.
What it does
The audit has two halves. A deterministic miner looks for the verses most likely to resist the reading, and a language-model red team writes objections. Together they fill an objections ledger.
A reading is then put on trial against linguistic, contextual, cross-verse, physical and eschatological tests, one judging call per candidate. The verdict is computed in code: any failed test, any survival only by interpretive rescue, or any contradiction with a recorded property rejects the reading, however widely held it is.
Audit is not approval
Passing the audit means a reading survived these tests. It does not make it correct, and it is not endorsement by any scholarly body. The objections and rulings are published in the report so qualified readers can judge them.
- A failed principle revises the element it touches; it never hides the answer.
- Surviving readings are kept when valid at a different scope.
- Rulings bind the claims, the architecture and the final audit.
Step by step
Mine hard cases
Find the passages most likely to strain the reading.
Raise objections
A red team writes objections into a ledger.
Try the readings
Test the inherited and alternative readings against five kinds of test.
Check principles
Audit against 14 uṣūl principles and revise what fails.
Rule
Each objection is answered, sustained or carried forward.
What the audit covers
14 uṣūl principles
Wording, Quran with Quran, Sunnah, Companions, language, context, conditions of ta'wīl, creed, restraint on the unseen, multiplicity of scopes, labelling and more.
Verdict in code
The accept or reject decision is computed from test results, not left to model discretion.
Objections ledger
Every objection is recorded with its ruling so a reader can disagree on the record.
Restraint on the unseen
The audit enforces caution on matters the text does not specify.
Common questions
Who decides what the principles are?
The 14 principles are implemented as explicit checks in the engine and shown in a compliance table in the report, so they can be reviewed and challenged.
Can the audit overturn a popular reading?
It can reject any reading that fails the tests, and it says why. Readings that fail only at one scope can be kept as valid at another.
Is a human in the loop?
Yes. The report leads with the answer but keeps objections, rulings and evidence so a scholar can review them.
See Taaweel run on your question
Tell us the question you are researching. We will walk through a run and what the evidence ledger and audit look like.