Work
Evidence-Guided Decipherment
What could this broken line have said—and what would distinguish the possibilities?
Early research prototype · constructed examples only
Keep the reasoning beside the reading.
A surviving fragment gives a researcher both observations and gaps. Evidence-Guided Decipherment explores how several possible readings can be compared without allowing a plausible reconstruction to take the place of the surviving evidence.
The intended audience is researchers examining incomplete or uncertain language evidence. The aim is an inspectable comparison: what each reading explains, what it assumes, where it struggles and which new observation would help discriminate.
Illustrative comparison · no real-language result
- Observation
- Only part of a line survives.
- Alternative A
- One proposed reading fits the visible signs but requires a missing segment.
- Alternative B
- Another reading needs a different grammatical assumption.
- Next check
- Identify a dated parallel or a clearer image that could separate the alternatives.
What is established.
The public project is an early prototype tested on constructed examples. No real-language decipherment is claimed. A useful next step is an independently inspectable, bounded example with explicit inputs, alternatives and evaluation criteria.
Reviewed 9 October 2026
Sources and scope: The public scope and limitations are recorded in Anavi’s inquiry directory.