Knowledge Graphs meet Moral Values

Proceedings of the Ninth Joint Conference on Lexical and Computational Semantics (*SEM), pages 71–80 Barcelona, Spain (Online), December 12–13, 2020 71 Knowledge Graphs meet Moral Values Ioana Hulpus, 1, Jonathan Kobbe1, Heiner Stuckenschmidt1, Graeme Hirst2 1University of Mannheim 2Department of Computer Science, University of Toronto {ioana,jonathan,heiner}@informatik.uni-mannheim.de, gh@cs.toronto.edu Abstract Operationalizing morality is crucial for under- standing multiple aspects of society that have moral values at their core – such as riots, mobi- lizing movements, public debates, etc. Moral Foundations Theory (MFT) has become one of the most adopted theories of morality partly due to its accompanying lexicon, the Moral Foundation Dictionary (MFD), which offers a base for computationally dealing with moral- ity. In this work, we exploit the MFD in a novel direction by investigating how well moral values are captured by KGs. We explore three widely used KGs, and provide concept- level analogues for the MFD. Furthermore, we propose several Personalized PageRank varia- tions in order to score all the concepts and enti- ties in the KGs with respect to their relevance to the different moral values. Our promising results help to progress the operationalization of morality in both NLP and KG communities. 1 Introduction Many of the choices that we make in daily life, such as political stance or position in debates on ideo- logical topics, are influenced by our moral values (Sagi and Dehghani, 2014; Wolsko et al., 2016; Amin et al., 2017). Besides, moral values and moral judgments are central to decision making and cultural cohesion (Dehghani et al., 2016). The last years have seen an increasing interest in op- erationalizing the concept of morality as defined by psychologists, particularly from the NLP, social media, and communication communities, into an effort of extracting the latent moral dimension of texts. Tweets (Garten et al., 2016; Araque et al., 2020), newspaper articles (Bowman et al., 2014), as well as scientific articles (Clifford and Jerit, 2013) or This work is licensed under a Creative Commons Attri- bution 4.0 International License. License details: http: //creativecommons.org/licenses/by/4.0/. religious sermons (Graham et al., 2009) have been targeted for moral analysis. Lately, this line of research has been widely motivated by the rise of social media campaigns such as #BlackLivesMatter and #MeToo, which have a very strong moral load. People take different stances with respect to such matters depending on their understanding and hi- erarchy of moral values, and can lead to clashes of visions even within the same culture. The general assumption is that the words used in discussions on such topics reveal the moral values of the discus- sants. One of the most widely adopted theories of morality is the Moral Foundations Theory (MFT) (Haidt and Joseph, 2004; Graham et al., 2013). MFT proposes at least five moral foun- dations, each one consisting on the one side of virtues and on the other side of vices: (1) the care/harm foundation which deals with the sensi- tivity towards the suffering of others; (2) the fair- ness/cheating foundation covering aspects of reci- procity and motivations …

Content truncated to 500 words