Moral Foundations of Spirituality and Religion Through Natural Language Processing

Abstract

To explore the similarities/differences between the moral contexts in which scholars use the terms religion and spirituality, we use Moral Foundations Dictionary for Linguistic Analyses 2.0 (MFD), a dictionary developed to assess the moral content of text, and a Natural Language Processing algorithm (Word2Vec) that learns the semantic relationships in a corpus. The findings suggest that, except in the virtue words category of the Care foundation dictionary, religion semantically overlaps with a greater percentage of MFD words than does spirituality. Both religion and spirituality have greater semantic overlaps with virtue words than vice words; compared to religion, spirituality’s semantic overlaps with vice words are smaller. Spirituality has greater overlaps only with the MFD words for Care and Sanctity; religion has greater semantic overlaps with words for all foundations, particularly the “binding” foundations: Loyalty, Authority, and Sanctity. Similarities notwithstanding, the moral contexts of religion and spirituality feature different aspects of morality.