ULM-3: Stories and world views as a key to understanding language
Subproject ULM-3 will consider the interpretation of text built up from words as a function of our ways of interacting with the changing world around us. We interpret changes from our world-views on the here and now and the future. Furthermore, we structure these changes as stories along explanatory motivations.
|Title||Stories and world views as a key to understanding language|
This project builds on the results of the European project NewsReader. NewsReader develops technology to automatically read daily streams of news to determine what happened, where, when and who was involved. It determines what sources write about the same events, how they overlap, differ and complement each other. It also relates the news of today to the news of yesterday and before. From these massive news streams, NewsReader reconstructs changes in our world over longer stretches of time as told in the news.
In this project, we will consider the interpretation of news text as a function of our dealing with the changing world around us. We start from the assumption that we consider these changes with a world-view in mind. This can be positive or negative towards the here and now, and/or positive or negative towards the future. Changes in the world are considered as changes to the good (from a negative world-view towards a more positive future) or to the bad (from a positive world-view towards a more negative future). We also consider these changes as long-term stories along human intentions and goals, with actors that are divided along a dimension of good and bad, i.e. contributing to changes to the good or the worse. Stories can be considered as explanatory lines of protoganists acting in our changing world.
Two research lines are investigated by 2 PhDs. In the first, the data produced by the NewsReader project is used to acquire the world-views of the primary and secondary sources in the news. Opinion-mining technology and factuality-detection modules developed in NewsReader will be used to find the positive and negative world-views on the here and now (Greece being part of the EU) and the future (Grexit). These views will be connected to the sources and contrasted to each other and over time. The PhD will formulate hypotheses of perspectivation and narrative choices that are predicted constrasting one world-view to the other. Lexical choice and lexical-syntactic framing in texts of sources with opposing views will be measured against these predictions. Actions and events underlying these changes will be related to the world-views in which they are embedded. In this way, we try to determine for example the position with respect to specific measures taken within the European Union with respect to the banking crisis in for example Greece or Spain. World-views are mined from different types of sources: written news, social media, interviews.
The second project considers news as stories along abstract plot structures. We expect events stretching from the past to the present to be modeled along explanatory lines of causal relations, intentions, motivations, strategies, involving intentional actors: those responsible, criminals, victims, etc. From textual sources (both news and social media), the PhD will mine speculations on (hidden) and declarations of overt motivations of actors in the world. Such data can be acquired through learned patterns and machine-learning. After acquiring these models, we will apply them to analyse the database of news built in the NewsReader project. Sequences of events that resemble the sequences in the model in terms types of events and actors, order of events and labelling of events and actors will be selected according to the model to build up a storyline according to these motivational associations. Intentional-explanatory storymodels can be used to guess intentions behind new changes in the world as they are reported in the news. The results of both projects will be integrated in an attempt to flesh out provenance models of news and the sources reporting the news, to come to a better understanding of the way the news is brought to us.