Skip to main content
KBS_Icon_questionmark link-ico
sunset in Athens ;

Day 9: Recogito Annotation and Benaki Museum

Sarah Middle and Rebecca Levitan

11 April 2019

On day 9, we gained some practical experience of an existing Linked Ancient World Data tool, with much of our work focusing on annotation using the Recogito platform.

Recogito is part of Pelagios Commons, and facilitates annotation of texts and images by identifying people, places, events, and other entities, tagging them with keywords, and linking to external authorities. These annotations can then be rendered as Linked Data. An advantage of the program is that it does not rely on any prior coding experience on the part of the user.

The morning started with an introduction to Recogito from PelagiosCommunity Director Elton Barker, who demonstrated its various features, including named entity recognition, creating relationships between entities, image annotation, and visualising annotations using a map view.


Recogito focuses on annotation of three main entity types – places, people and events – although any string of text can be tagged with keywords of the user’s choosing. While in many cases, places, people and events are easy to identify, some are less clear cut; for example, in ancient mythology, there is some blurring between places and people. Of particular interest to the Objects hub was a brief discussion on objects having agency; for example, in Homeric texts it is often the weapon that is described as having the agency rather than the person wielding it.

After Elton’s demonstration, we split back into our four ‘hub’ groups (Ideas, Objects, People and Styles) to experiment with using Recogito in relation to our hub topics. Here, we used a section of Pausanias’ Description of Greece that refers to Corinth, which we visited last week.


In the Objects hub, the first issue we encountered was that Object is not a specified entity type in Recogito, in the same way as Place, Person or Event. Instead we decided to systematically tag all text strings referring to objects with the keyword ‘object’, along with characteristics such as ‘material’ and ‘colour’. We found that objects were often described in relation to places and people, and found Recogito’s ‘Relations’ feature extremely useful for mapping these relationships. However, we also found that texts often do not describe objects neatly. For example, it can be problematic to annotate objects in sentences such as “The images of Asclepius and of Health are of white marble, that of Zeus is of bronze” because three images are being referred to, made of two different materials, each of which relates to a different person. We did, however, find this exercise extremely useful in thinking about how existing frameworks might be applied to more effectively describe objects in a semantic way, as well as how Recogito itself might be improved to facilitate object-related annotations.

The people team was interested in how a researcher might easily make the distinction between mythological and ‘real’ figures in a program like Recogito. The team discussed the possibility of visualizing figure’s relative timelines/chronologies using ‘real’/linear time vs. relative/mythological time, and how those different timelines might be able to intersect.

Later in the afternoon, some of us visited the Benaki Museum of Greek Culture, one of several museums holding the collections of Antonis Benakis. This particular museum is arranged chronologically, displaying objects from prehistory through to the 20th century, including archaeological artefacts, religious art, paintings, and costume.


As the museum is situated in the house where Benakis lived, we found it interesting to compare with the Soane Museum and Leighton House, which we visited during the first Institute in London. Reflecting on the different uses of space, it was noticeable that Benakis’ house had been completely repurposed as a museum, whereas the Soane Museum is preserved exactly as Soane had left it, and Leighton House is a recreation of Leighton’s original décor and collections.

One section of the Benaki Museum that we found particularly striking, and reminiscent of Leighton House, contained two richly-decorated rooms that had been transplanted from their original locations in people’s houses in Northern Greece to the museum. We discussed how each museum had situated objects in relation to their surroundings, as well as the importance of context – in Benakis’ case, transporting objects from one context to another, and in Leighton’s, recreating a context by assembling objects which either formed part of the original collection or are intended to represent them. One element we found particularly striking about the rooms in the Benaki Museum was the presence of costume; elsewhere in the museum, costumes were displayed in glass cases, but we found that including a costume in its original surroundings proved to be far more informative and evocative. Getty institute participant Ana Cabrera, who has previously worked with the textiles of the Benaki museum, provided valuable interpretation of these costumes.


A number of group members also attended a lecture at the Benaki museum as part of our visit. The talk , delivered by Massimo Vitti and Matthias Bruno, was part of the ongoing Roman Seminar series which bring together Athens-based scholars interested in the topic of Roman Greece. The talk, titled “The Odeion of Agrippa and the orchestra pavement in opus sectile,” was delivered in Greek and English, and provided a historical overview of the 1st century BC Augustan building located in the Athenian agora, and later changes to the elaborate marble pavement from the 4th and 5th centuries A.D.

Visiting the Benaki Museum helped tie together some of the threads that have woven their way through both the London and Athens Institutes and has provided much food for thought as we approach our final day.

Related links

Latest news