Carlo Scarpa in His Own Words: Encoding Challenges and Models of Representation in Recorded Lectures from the Franca Semi Fonds

This thesis investigates the digital representation of oral sources within the field of Digital Humanities, with a specific focus on text encoding practices applied to recorded scholarly speech. The research addresses the methodological problem of transforming an analogue recording into a machine-readable and philologically annotated resource, while preserving the paralinguistic phenomena, the material discontinuities, and the multimodal referents that characterise the source. The case study is drawn from the Franca Semi Fonds, held at the Archivio Progetti of the IUAV University. The fonds preserves a corpus of audiotape recordings of lectures delivered by Carlo Scarpa during the academic years 1974-75 and 1975-76, later transcribed and edited by Franca Semi in the volume A lezione con Carlo Scarpa. This coexistence of the audio witnesses and a printed editorial mediation defines a philological configuration whose treatment constitutes one of the interpretative challenges of the study. Within the corpus, the thesis analyses the lecture of 23 January 1976, titled by Semi: «Musei: Gipsoteca Canoviana a Possagno, Castelvecchio a Verona». The study is organised in two main parts. The first defines a theoretical and operational framework for the treatment of oral sources: it examines the material assessment of the carrier, the ethical and legal framework governing scholarly reuse, the potential and limitations of automated speech recognition, and existing encoding models for spoken language. On this basis, TEI-XML is adopted as the encoding standard and a manual transcription is produced according to the KIParla Corpus-Jefferson conventions, adapted to the register of the lecture. The second part applies this framework to the case study and produces a digital scholarly edition, it documents the modelling of the body of the text, the editorial treatment of Semi’s volume, the architecture of the teiHeader and the visualisation of the encoded material through Torchio, an open-source tool for the publication of static digital scholarly editions. In doing so, the research articulates a philologically grounded model for the digital representation of oral sources.