← All conjectures · African book cultures
The dead-language clock
Status is derived only from the shepherd-authored triage/prediction data above -- community submissions and claims are a separate overlay and can never change it (see the participation panel below).
“Anticipated in print” means the direction of this claim already appears in scholarship, while this exact test has never been run — it records prior art, not proof; nothing here counts as supported or falsified until a registered resolution says so.
Claim (verbatim)
Geʽez ceased to be anyone's mother tongue by roughly 1000 CE, while the spoken successors, Amharic and Tigrinya, merged several laryngeal consonants and the vowels around them. From that point on, scribes copied sounds they could no longer hear, so spellings among the affected letter sets (the h-series, the s-series, the two alefs, the two tsadai) should drift by random error, generation after generation — a molecular clock written in orthography rather than DNA. The conjecture is that the confusion rate per thousand characters rises monotonically with copy date, and rises tightly enough that an undated Geʽez copy can be placed within about a century from its error rate alone. Ethiopia's uniquely stable script and dense colophon dating make it the only African corpus — perhaps the only corpus anywhere — where such a clock can be calibrated over 800 continuous years, and if it holds, thousands of undated manuscripts become datable at a stroke.
Prediction clause (verbatim)
Primary clause: in colophon-dated Geʽez manuscripts with transcriptions in Beta maṣāḥǝft, the rate of etymologically incorrect choices among the merged letter pairs increases monotonically across century bins from 1200 to 1900 (Spearman correlation above 0.6 on binned rates); secondary clause: the within-century spread is narrow enough that the fitted curve dates held-out manuscripts to within 100 years for at least two-thirds of cases. Killed if fifteenth-century copies confuse the merged letters at the same rate as nineteenth-century ones, or the trend is flat or non-monotonic.
Kill-dataset (verbatim)
Beta maṣāḥǝft — dated transcriptions.
In the kill-dataset registry:
Nobody has run this test. The kill-data is named above. If you can run it — or you know the paper that already settles it — claim the kill or submit the prior scholarship. Kills and prior scholarship are credited here, by name, as they come in.
On Inferpedia
This conjecture is linked to the following pages on Inferpedia, an encyclopedia of the missing — working atlas pages, some still early scaffolding.
Provenance
Run: Fresh agent generation · model: claude-fable-5
Generated blind by a fresh claude-fable-5 instance in a single Write with no reads, web access, database queries, or other tool calls.
Novelty / leakage triage
anticipated in the literature — this exact test has never been run
That Amharic/Tigrinya mergers of the laryngeal and sibilant series produced systematic scribal confusion in later Ge'ez copies is a commonplace of Ge'ez philology and editing. A calibrated monotonic error-rate clock for dating undated copies was not located — but the quantitative side of Ge'ez philology is thin, so unlocated does not mean unanticipated in spirit.
Sources cited by the triage
Predictions
No prediction registered yet.
Weigh in
No community feedback yet.
New here? Create an account first
Create an account or sign in and your feedback is tied to you — you can track it, get replies, and claim this conjecture so others know you’re working on it. Prefer not to? Just leave your take below as a guest — only the name you type is shown.
Add your take
Posted immediately (spam is removed). Community feedback is never an adjudicated verdict and never changes this conjecture's triage label or status above.