Euporia Rituals in ancient Greek Tragedy (RAGT) is an ongoing project for the creation of a browsable database of rituals and religious facts in ancient Greek tragedy. The project is the result of the cooperation between the Laboratorio di Antropologia del Mondo Antico (LAMA) at the University of Pisa and the Collaborative Philology Lab of the Institute of Computational Linguistics "A. Zampolli" at the CNR of Pisa. The annotation and the querying system were developed in order to offer digital support to Gloria Mugelli’s PHD Thesis, an historical anthropological research on the form and function of rituals in ancient Greek tragedy. The results (database, tagset, ontology and search interface) are aimed to be published in open access. The project involved two stages:
the annotation of ritual and religious facts in the corpus of the 33 surviving ancient Greek tragedies, carried out with an annotation system based on the annotation practices of classicists, that allows the user to quote continuous and discontinuous passages of various length and deals with textual and interpretive variants. The quoted passages were annotated with keywords expressed as hashtags [1].
the organisation of the tagset in an ontology and the development of a system to perform complex and expressive queries on the database, carried out adopting a bottom-up approach: a search engine (EuporiaSearch) was developed in order to perform queries on the database of the annotated passages [2]. The tagset has been subsequently organized in an ontology, that establishes relations between the different keywords used in the annotation. A system for querying the database, using the ontology in order to perform more complex and expressive queries, is under construction [3].
EuporiaSearch allows to query the database of the tragic passages dealing with ritual and religion. It is possible to perform queries combining up to three of the keywords used in the annotation. A list of the searchable keywords will be available soon.
To perform demo queries, users have to enter:
a keyword that represents a superclass, prefixed by t:, in order to find all the related subclasses (e.g. t:animal will identify bos, pecus, etc.);
a second keyword (e.g. sacrificium);
a third keyword, possibly prefixed or suffixed by the wild character % (e.g. victima or vict%). With the aforementioned keywords the search engine will search all the annotated passages where the three keywords marking subclasses of animal, sacrifice and victim co-occur);
the range of words within which the keywords have to co-occur (e.g. 0; 0 if the keywords have to insist on the same textual passage);
the range of words on which the keywords visualized in the results have to insist (e.g. 0, 0 to browse all the keywords insisting on the retrieved passages or 10, 10 to browse all the keywords insisting on a range of +/-10 words from the retrieved passage).
A query on the three keywords sacrificium, victima and pecus, with all the other parameters set at 0, retrieves all the annotated passages in which a sheep is mentioned as a sacrificial victim. The results visualize the annotated passages and the keywords used in the annotation. It is also possible to browse the passage in the context of the entire tragedy, by clicking on the greek words that mark the beginning and the end of the annotated passage
Further readings:
[1] The annotation system is described in G. Mugelli, F. Boschetti et al., ‘A User-Centred Design to Annotate Ritual Facts in Ancient Greek Tragedies’. BICS 59 (2).
[2] The bottom-up approach and the search engine were described at the Göttingen Dialog in Digital Humanities 2016 (http://www.etrap.eu/activities/gddh-2016/), the proceedings of which are forthcoming.
[3] The ontology and the querying system were presented at the workshop on Language, Ontology, Terminology and Knowledge Structures at the IWCS in Montpellier