A correction of a correction
The Date Made From One Word Twice
October 24, 79 CE is often presented as Pompeii's corrected destruction date, but Pliny's manuscripts do not contain that complete reading. Rebuild the phrase and watch the date appear only when one ambiguous token is used twice, then scrub the Pompeii Park's ten-month charcoal replication. The experiment leaves October 17, 78 CE possible for the yearless inscription, so neither October 24 nor August 24 is established as the eruption's exact day.
There is one cardboard scrap on the table. It can be read as non., a day word, or nov., a month word. It cannot be two scraps.
Give the token one job
no complete date
not testable
Choose a slot for the token.
Roman calendar arithmetic is inclusive. The ninth day before the Kalends of November counts November 1 as day one, so it lands on October 24. Ordinary subtraction by nine would land a day early. But arithmetic cannot repair a phrase that no witness contains.
| Foss siglum or family | physical reading | day token | month token | modern result |
|---|
This is a three-line diagnostic excerpt from Foss's collation apparatus, not a count of all surviving manuscripts. M supplies a complete September phrase. The gamma family preserves a missing month; one subfamily's novum is an alternative to nonum, not a second token. Foss's broader published conclusion is the anchor: October 24 has zero evidentiary basis in the manuscript tradition.
The sophisticated objection is archaeological: perhaps the manuscripts are beside the point. A charcoal note says October 17, and Pompeii contained things called autumnal. Surely the eruption came later.
A date without a year
The ancient charcoal inscription reads XVI K NOV, October 17 by inclusive counting. It does not state a year. Its force as a lower bound for the eruption depended on a physical premise: an exposed charcoal line could not have remained legible from October 78 CE to the eruption traditionally dated in August 79 CE.
The Archaeological Park of Pompeii tested that premise. On October 17, 2023, its team wrote a replica in oak charcoal on the same ancient wall, at the same height, under a temporary roof. It photographed the writing through August 24, 2024. The authors reported that it remained perfectly legible.
Scrub the wall
The February and June sessions are listed in the paper but do not appear as unique frames in the two published plates.
The contrast number is this page's diagnostic: 100 × (P90 luma − P10 luma) / (P90 luma + P10 luma) over the displayed registered crop. It is recomputed from the embedded pixels. It is sensitive to lighting, crop, compression, wall texture, and camera processing. It is not the Park authors' measurement, not a calibrated decay rate, and not a test of legibility.
From October 17, 2023 to August 24, 2024 is 312 days: ten completed calendar months and seven days. The replica does not prove that the ancient mark was made in 78 CE. It removes the claim that survival for that interval was physically implausible. The year 78 therefore cannot be excluded by charcoal survival alone.
What does each autumn clue actually bound?
Choose a clue. The score is about calendar force, not importance: 0 means no 79 CE date bound; 1 means a conditional or season-sized bound; 2 would mean a secure lower bound within 79 CE. These are declared categories, not probabilities.
no 79 CE calendar bound
conditional or seasonal
secure lower bound in 79 CE
Select a clue
Its narrowest defensible bound will appear here.
The check
Every line below is recomputed in this browser. The offline verifier takes separate routes and also checks the page's public wording.
- Loading independent checks.
Free choices: the three-level clue score, the displayed collation excerpt, the ten unique printed photo frames, the crop coordinates, and the P10/P90 contrast diagnostic. Uncertainties: Pliny's text may not preserve the historical day; the charcoal inscription has no year; the photographs do not calibrate long-term charcoal decay; the coin reading and the seasonal interpretation of finds are contested. None of this establishes August 24 as the eruption date.
Run: node research/the-date-made-from-one-word-twice/verify-the-date-made-from-one-word-twice.mjs
Sources and the institutional reversal
Text and replication. Gabriel Zuchtriegel, Chiara Comegna, Fabrizio Conte, and Alessandro Russo, La data della distruzione di Pompei: premesse per un dibattito aperto, Pompeii E-Journal 27 (2024), especially pages 2 to 8 and 23 to 24. The embedded photographs are compressed crops of its figures 4 and 5.
Collation. Pedar W. Foss, Pliny and the Eruption of Vesuvius (Routledge, 2022), pages 128 to 132 and the author's online collation resources. The Park paper explicitly adopts Foss's correction.
Earlier Park claim. The Park's 2019 Regio V Excavations report presented the charcoal inscription as supporting October rather than August. The 2024 Park paper reports the same-wall experiment and says that October 78 CE cannot be excluded. Both are shown because the reversal is part of the evidence.