2025/06/17 by Penelope M. Allison, Daniël van Helden · 1 voice
Computer Science · #Image Processing and 3D Reconstruction
paper · doi:10.1086/735650
openalex publication_date 2025/06/17 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/31
<p dir="ltr">This article compares approaches to counting ceramic vessels among large datasets of al- <p dir="ltr">ready catalogued sherds and aims to demonstrate that relative proportions of vessel forms <p dir="ltr">can be used to address questions of sociocultural behavior, irrespective of approaches <p dir="ltr">to recording sherd counts. We aim to demonstrate that it is valid to carry out quantita- <p dir="ltr">tive, comparative, intersite analyses of sufficiently large ceramic datasets of sherds that <p dir="ltr">were recorded in different ways and for which specific, time-consuming methods for <p dir="ltr">calculating vessel quantities were not carried out at the time of recording. We use the χ2 <p dir="ltr">goodness-of-fit test to demonstrate that to be able to do comparative analyses of pottery <p dir="ltr">types among such datasets, it matters little how sherds are counted. As a case study we <p dir="ltr">use the Arch-I-Scan Project’s catalogue of terra sigillata sherds (often called “samian” in <p dir="ltr">the United Kingdom) from three different locations in Roman Britain. We demonstrate <p dir="ltr">that different ways of recording vessel forms have little impact on the proportional rela- <p dir="ltr">tionships of such forms. This method can thus document food-consumption practices <p dir="ltr">across different contexts.