Towards an Automatic Easy-to-Read Adaptation of Dialogues in Narrative Texts in Spanish

Isam Diab, Mari Carmen Suárez-Figueroa, Roberto Peris

Abstract: People with cognitive disabilities have the right to actively participate in various aspects of society, including culture, on an equal basis with others. Related with this social need, a methodology called Easy-to-Read (E2R) was created to improve cognitive accessibility. This methodology provides a set of guidelines and recommendations to present clear and easily understood contents to different sectors of the population that include people with cognitive disabilities. One of these guidelines refers to structuring dialogues in a theatrical style, i.e. placing the name of each character ahead of their intervention, to make it explicit who is the speaker of each intervention and thus avoid having to infer this information. Since the E2R methodology is currently applied in a manual fashion, we have developed an initial method for automatically adapting dialogues in narrative texts in Spanish to theatrical style. This method is based on (a) Artificial Intelligence (AI) methods and techniques and (b) the E2R guidelines. We have implemented such a method as a proof of concept and we have validated its functionality trough a user-based study involving people with cognitive disabilities.