Stefano De Giorgis

Values, as intended in ethics, are part of the broad and challenging area of research about Commonsense Knowledge. The attempt to untangle the complex structure of relations among human moral and social values requires investigating subjective human perception of the world as well as socio-cultural dynamics. The Values Ontology is a modular ontology representing the domain of moral and cultural values. It models several theories such as Schwartz's Basic Human Values, Graham and Haidt's Moral Foundations Theory, and Curry's Moral Molecules. Furthermore, each value is modeled as a Conceptual Frame, inteded as in Frame Semantics. The Values Ontology is based on reusable Ontology Design Patterns, it is aligned to the DOLCE foundational ontology, and it is a component of the Framester factual-linguistic knowledge graph.

Embodied cognition and the theory of cognitive metaphors ground our commonsense reasoning ability in language, linking subjective perception of the external world with cognitive inferential patterns. Furthermore, commonsense reasoning is linked to human sense-making, pattern recognition and knowledge framing abilities. In this work we developed ISAAC, the Image Schema Abstraction And Cognition modular ontology, a new resource that formalizes the cognitive theory of Image Schemas. Image Schemas are conceptual dynamic building blocks originated from recurring sensorimotor interactions with the physical world. These experiential patterns assign coherence and structure to entities, sequences of events and situations we experience everyday. ISAAC ontology provides a formalization of theoretical background and integration of different theories regarding linguistic and factual entities.

Human cognition, and particularly the sense-making ability, work on multiple higher layers of meaning at the same time. A serious investigation about commonsense knowledge and meaning attribution must therefore take into account at least the above three layers: embodied cognition, emotional layer, and value and social norms layer. This is the focus of my PhD thesis: Ethics in the Flesh: how moral values meet emotions and embodied cognition. Taking advantage of the FRED tool, we built a fully explainable detector that exploits knowledge graph patterns to extract emotions, values, and image schemas, and make inferences about their possible combinations considering their semantic dependencies on the graph. The knowledge graph in the image is generated automatically from the sentence The police finally stopped the awful killer. The detector retrieves the image schema Blockage, evoked by the verb "stop", the emotion Disgust from the adjective "awful", the moral value-violation Harm from "killer" and the social values Security and Authority, triggered by "police". A larger version of the image is available here.