A map of evidence, not a pile of documents.
The Engine stores what each piece of evidence can support, where it stops, how it relates to other evidence and what would have to be true for a conclusion to change.
Claims
Small enough to test honestly. “Jesus was buried” is separate from “Joseph buried him in his own tomb.”
Sources
Not just citations. Date, authorship, preservation, motive, genre, independence and limits.
Relationships
Does this source support the claim, oppose it, merely make it possible, or depend on another source?
Make the complex simple.
This is a corpus rule, not just a writing preference. Complexity belongs in the data model so the conversation can stay human.
Inside the engine
Source provenance, dependence, competing scenarios, expected traces, motive, relevance, uncertainty, version history.
What the user hears
“This Roman rule matters, but it does not tell us what Pilate actually did. Here are the other things that bear on that decision.”
The pilot is teaching us where apparently simple claims hide different questions.
A rule, a belief, a practice and an event are different things
Jewish law could forbid leaving a corpse hanging overnight. That does not automatically show Jews could enforce the rule, Rome routinely accommodated it, or Pilate did so in one case.
Archaeology has visibility bias
Crucifixion disproportionately affected lower-status people, while recoverable rock-cut tombs and ossuaries disproportionately represent wealthier families.
Separate jurists can share one surviving pipeline
Two ancient legal opinions may be intellectually independent while both reach us through a later compilation and editorial history.
Early belief is not the same as event proof
A very early burial formula tells us burial belief appeared early. That is important. It still does not automatically prove the exact burial event or later tomb details.
The same act can mean something different depending on the reason behind it.
Cleopas does not reduce people to motives. It records possible motives with their own uncertainty and tests alternatives.
Historical actor
What pressures, incentives, fears or beliefs could explain the action?
Source
What would the writer or community gain from preserving, shaping or inventing this detail?
Reader
Why does this modern person trust a Roman document automatically but scrutinise a religious one, or vice versa?
Ask what we should expect to see if the opposite were true.
Missing evidence matters only when there was a good reason to expect the evidence to exist, be recorded and survive.
Was the burial place known? Was the body identifiable? Could opponents access it? Would producing it have been persuasive? Would anyone record the response? Would that record survive? Only after those questions can the absence carry weight.
No person, source or internal score becomes the authority.
No expert vote
Specialists can add facts, challenges and assumptions. Their status does not become the answer.
No citation counting
Four citations may still be one tradition repeated four times.
No hidden certainty
When evidence is unresolved, the system is allowed to say so without filling the gap.
This is where product defensibility can accumulate.
The conversational model can change. The durable asset is the governed evidence structure, its version history, its failure cases and the rules that keep the answer bounded when the conversation changes.