Linked Pasts 7 Working Group

Recogito and TextLinker: pathways to semantic annotation in Latin Texts

Organizers

• Valeria Vitale, The Alan Turing Institute.
• Elton Barker, The Open University.
• Francesco Mambrini, Università Cattolica del Sacro Cuore.
• Rainer Simon, Austrian Institute of Technology.

Linked Pasts 7

The activities of this Working Group will be held during the seventh edition of the Linked Pasts symposium, Linked Pasts 7. The symposium, organized by Ghent Centre for Digital Humanities, will run entirely online between 13-20 December 2021.

Date and time of the activity

Monday, 20 December 2021, 15:00 – 17:00 CET.

Location

Online (Zoom): link shared with registered participants only.

Objective

This working group aims to create a shared space to discuss the potential of linked open data in linguistic and philological analysis of historical texts. In particular, we will look at two annotation platforms well known within the Linked Pasts community: Recogito and the LiLa TextLinker. The first one is an online platform for semantic annotation, georesolution and map-based visualisation developed within the Pelagios project. The second is a LOD-powered online tool for the lemmatization and part-of-speech tagging of Latin texts. 

Through demos, hands-on exploration, feedback, and user-oriented discussion we want to take a closer look at these tools, testing them against specific case studies and evaluating their performance. The insights we will gain through this working group will then inform future developments of both tools, identify synergies and potentials for interoperability, and contribute to the design of useful and repeatable workflows.

Description: annotation with TextLinker and Recogito.

During the working group activity, we will make available to the participants a small number of sample Latin texts, representatives of different kinds of corpora. We will then demo the annotation of the same text in both TextLinker, producing linguistic annotations, and Recogito, creating more spatially-oriented annotations. We will then invite the attendees to try their hands at the annotation processes, either with the selected texts, or with examples from their own corpora.

After a brief round of annotations, we will analyse the added information we have just created, and discuss the pros and cons of each approach. Can these two annotation paths intersect in a meaningful way? Can one tool benefit from functionality available in the other? What would the ideal interaction look like in a system that combines linguistic annotation with entity linking, and what technological changes and implementation steps would be needed to make it possible? What remains outside the scope of both Recogito and TextLinker, and how can we build one or more ideal workflows?

The discussion will be documented, through notes and questionnaires, and will become the foundation to design more effective user profiles that mirror actual needs and interests of our user communities.

Target audience

This activity is intended for those who work (or plan to work) on the digital annotation of ancient texts using web tools, and who are interested in the interactions between general linguistic annotation, semantic web technologies, and named-entity tagging and resolution. The discussion is focused on Latin, but it involves questions that are relevant for professionals working with other languages and all kinds of historical documents as well.

No prior experience of Natural Language Processing and Linked Data technologies is required. While knowledge of Latin is preferable to follow the annotation task, participants who don’t know Latin but are nevertheless interested in the subject are welcome to join.

Students and teachers of Latin, who are curious about how the annotation of named entities and other linguistic phenomena could help their learning experience, are especially invited!

Materials and technical requirements

The annotation activity will be focused on a single Latin text: Pliny, NH, 3.123-126 [The Latin Libary: Latin text; Perseus: English translation].

A txt file with the Latin text (taken from The Latin Library) can be found in GitHub, where we will also upload all the annotations, the minutes of the meetings and any other materials:

https://github.com/CIRCSE/Tutorials/tree/main/LinkedPast7

Registration

Linked Past is a free event, but advanced registration is required.

Use this form to the register for the activity.

Please, register BOTH to the Linked Past Symposium (here) AND to the Working Group activity.

Protection of personal data

Personal data gathered through the registration form is intended for the exclusive use by the organisers, and is collected only in order to help them prepare for the activity. These data will not be passed on to third parties nor will any data be stored for future use; they will be deleted no later than one month after the end of the activity. The email addresses may be used to inform the participants about the activities and to share materials (slides, texts, feedback forms), before and shortly after the event.