Charles S. Peirce is widely acknowledged as one of the founding fathers both of philosophical pragmatism and semiotics. Following his death in 1914, Harvard University acquired Pierce’s writings to preserve them for future study. Peirce’s archives contain approximately 1,650 unpublished manuscripts for a total of over 100,000 pages. Carefully cataloged by his pupil Richard Robin, they were and microfilmed in the 1960s. The microfilms were subsequently digitized by the Houghton Library but their limited legibility recently persuaded Houghton to undertake a renewed digitization effort working with the original manuscripts. This new effort offers an exceptional opportunity to study Pierce’s sometimes difficult writings through the application of AI-based computational techniques that allow not only for automated transcription, but also the study of idiosyncratic graphical markings, color codes, revisions, insertions, and annotations.
Alongside traditional editing practices, the Peirce Interprets Peirce project aims to leverage the power of these new techniques, from machine learning to NLP processing/analytics to data visualization, for the development of a prototype for an interactive digital edition of the entire Pierce archive. A small team of undergraduate and graduate students with varying skillsets, from literary and paleographic analysis to coding to graphic design, will be assembled to this end and will collaborate with colleagues at the University of Lausanne (Davide Picca) in Switzerland, the University of Groningen (Dario Rodighiero) in the Netherlands, and the Bibliotheca Hertziana (Alessandro Adamou) in Rome, with metaLAB (at) Harvard serving as the project hub.
The project goals are articulated in 3 phases.
The textual analysis of Pierce’s work aims to demonstrate how the manuscripts open up a broad interpretive space. The first goal is connecting to its broader context through techniques of wikification and entity linking provided by LinkedOpenData. These connections will rest on an ontological framework based upon Peirce’s own semiotic theory. This process aims to track semantic slippages in his writings over the years.
The modern-day building of interoperable catalogs of cultural significance employs methods heavily influenced by Peirce’s work. Particularly, knowledge graphs and Semantic Web offer standards for constructing semantically enhanced digital scholarly editions. While TEI and IIIF can be adopted for publishing textual and visual data, the editions themselves will be available as RDF-based knowledge graphs and paired with semantic annotations of the archived content. This process will pinpoint Peirce’s notes in the landscape of computational linguistics and cognitive computing, as well as represent his contributions and theories, using widely-adopted vocabularies and ontologies for scholarly publishing and annotation.
From the standpoint of data visualization and interface design, our goal is to provide end users with a series of data visualizations capable of working as indices to specific facets of the archive. In addition to standard tools like text search and classification, comprehensive visualizations will allow for the overall exploration of the archives as well as for zooming down into particulars.
Harvard University. n.d. “Charles S. Peirce Papers.” Harvard Library. Accessed October 7, 2022. library.harvard.edu/collection….
Stjernfelt, Frederik. 2007. Diagrammatology: An Investigation on the Borderlines of Phenomenology, Ontology, and Semiotics. Synthese Library 336. Dordrecht: Springer.