Why There is No Definition of Trust: A Systems Approach With a Metamodel Representation

Trust is an essential component in HRI, yet it has been impossible to agree on a definition within the HRI context. Moreover, the volume of definitions with different conceptualizations has led trust research to be viewed by some as a quagmire. Instead of attempting to define trust, this paper takes a bottom-up approach, starting with the body of current literature and breaking it into conceptual components within a trust system. Applying concepts such as abstraction and encapsulation from computer science, this paper attempts to synthesize the current body of literature into a metamodel where the components can either represent models themselves or natural groupings of elements. By viewing trust as a system (represented by a metamodel), it is possible to hide some of the details and functionality in each component without losing semantic value. This helps to clarify the trust workspace and terminology, highlights areas of trust research that are actively being researched, shows how different areas of research are connected, and captures the current state of trust research in a single framework.From a practical perspective, using this model also provides trust researchers with a single reference point, simplifies scoping one’s research, and enhances homogeneity by making cross-referencing and comparing study methods easier. Finally, viewing trust as a system allows capturing nuances definitions cannot, and may help identify gaps in current research.