Sofia Baroncini: Artworks interpretations in Linked Open Data: an ongoing project on Panofsky’s iconological studies

Recently, researchers have shown an increased interest in the semantic modeling and description of artworks. Thanks to the thoroughness with which data and their relations can be expressed, the semantic web is an appropriate means to represent the complexity of the art history domain. Currently, several ontologies (e.g. VIR ontology[1], Simulation Ontology[2], vocabularies (Getty Vocabularies[3], Iconclass[4]), and knowledge graphs (e.g. ArtGraph[5], Hyperreal, Wikidata) describing art-related topics are available. Although the presence of such examples, some relevant aspects of the art historical research are poorly represented. Among them, we can cite the iconographical and iconological ones, which focus on artworks’ subjects (i.e. iconographies) and meanings (i.e. symbols, and cultural aspects).

The project presented in this lecture aims at filling this gap by creating a Linked Open Data dataset of described iconographical and iconological interpretations and by conducting quantitative analysis on it. To this aim, a selection of Panofsky’s iconological studies was chosen as a source for data creation because of the relevant role he held in the theorization of the discipline.
The modeling of the data was realized according to standard ontologies and ICON, an ontology newly created for expressing iconographical and iconological interpretations with a high level of granularity, based on Panofsky’s theory. The modeling decisions are motivated by the results of a preliminary typological study of iconological interpretations (Baroncini, Daquino, Tomasi, 2021) and by the identification of research questions relevant to the domain. The art historian’s claims were then described manually according to the model, converted to RDF, and aligned to existing sources.
Currently, the dataset contains interpretations about ca. 400 artworks mainly from Middle Ages and Renaissance Western art, extracted from three books plus one article by Panofsky. The interpretations are divided into three levels, from a more superficial understanding to a deeper one, as described by the art historian’s theory, and inter-level
inks among identified subjects are provided. The subject types recorded include natural elements, actions and emotions (level 1), characters, events, places, objects with a specific identity (e.g. the Bible), personifications, symbols, stories and allegories (level 2), concepts, and cultural phenomena (level 3). For each subject identification a provenance of the assertion can be provided, indicating the author, source, and cited evidence, so as to allow the coexistence of multiple (diverging) interpretations.
The availability of such described data allows for answering domain-specific research questions in a quantitative way. Therefore, the first phase of the data analysis, currently under development, consists of answering several research questions formulated during the modeling phase and verifying to what extent they can be answered with data.
The main aim of the data analysis is to contribute to the foundation of iconological analytics, by detecting patterns and tendencies in iconological studies (e.g. symbolic meanings reframed according to their context). Other benefits include 1) its reuse as ground truth for developing inferential (automatic reasoning) and computer vision methods, and 2) the exploration of distant artworks through an iconological narrative thread since the selected iconological studies concern cultural objects preserved by different collections.

Lecture within the Data for History Lectures series
Time: Wednesday, 11 January 2023, 4.15-5.45 pm
Place: via Zoom – Request access via email (digitalhistory@hu-berlin.de)

Notes:
[1] https://ncarboni.github.io/vir/
[2] https://www.w3id.org/simulation/ontology/
[3] https://www.getty.edu/research/tools/vocabularies/
[4] https://iconclass.org/
[5] https://doi.org/10.5281/zenodo.6337958

References
● Baroncini S., Daquino M., and Tomasi F. (2021). Modelling Art Interpretation and
Meaning. A Data Model for Describing Iconology and Iconography. arXiv:2106.12967
https://www.aidainformazioni.it/index.php/aidainformazioni/article/view/10

Contributed image:
Piero di Cosimo, Public domain, via Wikimedia Commons: https://commons.wikimedia.org/wiki/File:Piero_di_Cosimo_The_Discovery_of_Honey_by_Bacchus_about_1499.jpg

---

---