Text Mining the Coffin Texts

Earlier Ancient Egyptian mortuary texts constitute a privileged showcase for the profound ideological and material changes that occurred during the Middle Kingdom, especially between the second half of the Eleventh Dynasty and the first half of the Twelfth Dynasty. A corpus-driven approach appears as the more efficient and reliable method to provide a comprehensive assessment of the complex situation of the period. In such an approach, the corpus is not a mere object under study but also a control group; this, most importantly, implies that the corpus’ assessing value is privileged over any external model. Fundamental to corpus-driven approaches is corpus size and access to it. Earlier Ancient Egyptian mortuary texts are an ideal candidate because they include two large corpora: the authoritative editions of the Pyramid Texts (PT) and Coffin Texts (CT), which extend over six and seven volumes, respectively, plus one volume for the copies of PT on Middle Kingdom coffins. TM-CT will focus on the CT, an important amount of which is now available from the MORTEXVAR database beta version. This dataset will be enriched with the image dataset from the OCR-PT-CT and TTAE projects using a pre-trained artificial neural network based on the YOLO v3 strategy and Natural Language Process techniques.

The resulting database is expected to allow, for the first time, to analyse variability in the complete corpus and on the original text in parallel with the transliteration and translation, plus the material philology metadata, much of them already in place.