A group of us have been working on a (semi-unofficial) project to describe the narrative content within and across media objects (contextus.net/). Recently we have began looking at bootstrapping the system by playing with TEI marked up plays to identify basic character (and character-interaction) information in addition to our ongoing discussion about at ways to semi-automate or crowd-source the annotation process.
We will be presenting some of our work at DH but I am interested in discussing topics related to semantically describing the narratives within texts (from actual textual data to multimedia sources) such as the applications that data could be used for, to the issues of annotation, search/filtering, analysis, visualisation/presentation and how to make the data easily available/useful to researchers.
I’d also be very interested in talking to other people who had material/projects that they thought might benefit from this type of annotation as we are always looking for different use-cases to explore the possibilities of.
Hi Faith.
This reminds me of something I looked at years ago, piespy which was a tool to infer and visualize social networks based on IRC conversations. The more two users talked, the more the stronger there ‘connection’ and you can guess who is talking to whom through various simple algorithms. They did some movies of conversational relationships between characters in shakespearean plays.
www.jibble.org/piespy/
www.jibble.org/shakespeare/
Now of course I understand what you are talking about is semantic description of narrative contexts, and I’m assuming this is being done as open linked data with known standard ontologies and with a sparql endpoint for querying?
I actually am working with some data that, while non-narrative, might benefit from this as it encapsulates lots of social networked relations between named individuals. (Historical diary where author notes lots of meetings he had with A, B and C, and elsewhere B, C, D, and E. So lots of potential for finding out strength of network between B and C being stronger since they appear more frequently together, etc. etc.)
-jamesc
Heya,
Thanks for those links. They look rather cool and I will pass them on to the others. The strength issue is an interesting one. Currently I’m not sure if we are pulling the information that you would need for that from the bootstrap data (although I could be wrong as that isn’t one of the parts I worked directly on) but it would be feasible and we have previously looked at the relative level of involvement of characters (and groups) within a narrative to give us an idea of their importance within the story (the protagonist was involved in many more events than the secondary characters etc).
You are assuming correctly – we have a 4store for the data with a snorql interface for sparql queries. At some point we will probably add something prettier there for visualisation as well but sadly that hasn’t made it to the top of our to-do lists yet. We have our own upper level ontology to deal with some of the aspects that the existing ones we looked at had problems with but it was designed to link into other standard ontologies and we are also working on a ‘lite’ version which it totally integrated with FOAF and the Event and Timeline ontologies developed by Yves Raymond.
With regards to the diary – the data may be ‘non-narrative’ but it sounds like a series of events over time are being described and that is a narrative to me 🙂 It does sound like the type of thing that would be interesting to not only look at the relative strengths of the social links but also whether they shift over time especially if it linked in with other events either in the diary or know through other sources which might relate to the shifts in the social patterns. Fun, fun, fun.