Persuasive language is an important tool used for changing people's beliefs and influencing their behaviour. It is widely applied in areas such as advertising, political speeches, and crowd funding. Framing is one technique adopted in persuasive language where certain aspects of an issue are highlighted more than others so as to elicit the desired response from a reader or listener. In this project, we aim to first identify the features that characterize persuasive language and framing using modern techniques from Natural Language Processing (NLP). We would then explore the effectiveness of using such language, for instance in the adoption of laws or amendments.
First, we will explore the use of such language in the speeches of the plenary debates of the European Parliament and their effect on the adoption of laws. There exists datasets of parliamentary speeches, and more recent data should be available on the website of the parliament. You will need to explore available datasets and scrape additional data from the Web if necessary. You will then use techniques from NLP, such as those developed for text classification, to identify the elements of persuasion and framing in the speech text and study their correlation with the adoption of the laws. Finally, we will also study other areas of persuasive language such as speeches within committees of the Parliament, persuasive techniques of lobbying groups, campaign speeches, or even those in other domains, such as advertising.
If interested, please send your CV and a transcript of your grades to aswin.suresh@epfl.ch and victor.kristof@epfl.ch.