For their study, Yudkin et al. compiled a dataset of nearly 370,000 AITA posts, along with over 11 million comments, posted between 2018 and 2021. They used machine learning to analyze the language used to sort all those posts into different categories. They relied on an existing taxonomy identifying six basic areas of moral concern: fairness/proportionality, feelings, harm/offense, honesty, relational obligation, and social norms.
Yudkin et al. identified 29 of the most common dilemmas in the AITA dataset and grouped them according to moral theme. Two of the most common were relational transgression and relational omission (failure to do what was expected), followed by behavioral over-reaction and unintended harm. Cheating and deliberate misrepresentation/dishonesty were the moral dilemmas rated most negatively in the dataset—even more so than intentional harm. Being judgmental was also evaluated very negatively, as it was often perceived as being self-righteous or hypocritical. The least negatively evaluated dilemmas were relational omissions.
As for relational context, cheating and broken promise dilemmas typically involved romantic partners like boyfriends rather than one’s mother, for example, while mother-related dilemmas more frequently fell under relational omission. Essentially, “people tend to disappoint their mothers but be disappointed by their boyfriends,” the authors wrote. Less close relationships, by contrast, tend to be governed by “norms of politeness and procedural fairness.” Hence, Yudkin et al. prefer to think of morality “less as a set of abstract principles and more as a ‘relational toolkit,’ guiding and constraining behavior according to the demands of the social situation.”
DOI: PsyArXiv, 2024. 10.31234/osf.io/5pcew (About DOIs).
Fractal scaling of trees in art
De grijze boom (Gray tree) by Piet Mondrian, 1911.
Credit:
Public domain
Leonardo da Vinci famously invented a so-called “rule of trees” as a guide to realistically depicting trees in artistic representations according to their geometric proportions. In essence, if you took all the branches of a given tree, folded them up and compressed them into something resembling a trunk, that trunk would have the same thickness from top to bottom. That rule in turn implies a fractal branching pattern, with a scaling exponent of about 2 describing the proportions between the diameters of nearby boughs and the number of boughs with a given diameter.
