Theoretical Foundations of Trust

Since the very beginning of research on trust (around 1980s), it has been made clear that trust is pervasive, elusive (hard to grasp and study), and involves many disciplines.

The earliest well cited collection of academic writing on trust can be found in the 1988 book, Trust: Making and breaking cooperative relations, edited by the Italian born Diego Gambetta [1]. During 1985-86, Gambetta and a series of social scientists met at King's College, Cambridge to discuss the notion of trust. In the introduction of the book [1], Gambetta traced the interest in studying trust to "my own struggle to make sense of the persistent and apparently insoluble political and economic problem Italy has faced over the century...:the underdevelopment of most of her southern regions".

Trust can make or break cooperation. This is best illustrated by using prisoner's dilemma (the definition and analysis can be found here). If both parties trust each other, the outcome would be more optimal than both being a purely rational thinker -- moving the result to the strong Nash equilibrium of mutual defection. In other words, the benefit of mutual trust is increased benefits for both parties involved in the exchange. We will see this underlying principle in many subsequent trust studies -- including the trust game by Berg [2], the modern peer-to-peer marketplaces, etc. This is also the fundamental reason that I study trust.

There is no doubt that trust is important, in addition to the understanding of economic actions. Gambetta pointed out that "the importance of trust pervades the most diverse situations where cooperation is at one and the same time a vital and a fragile commodity: from marriage to economic development, from buying a second-hand car to international affairs, from the minutiae of social life to the continuation of life on earth [1]. However, the pervasiveness creates challenges for studying trust. This "very pervasiveness seems to have generated less analysis than paralysis: in the social sciences the importance of trust is often acknowledged but seldom examined, and scholars tend to mention it in passing, to allude to it as a fundamental ingredient or lubricant, an unavoidable dimension of social interaction, only to move on to deal with less intractable matters." [1]. Also as a result of the pervasiveness of trust, the study of trust has been spread out in many disciplines, including in Gambetta's book, he counted "anthropology, economics, history, philosophy political science, socio-biology, sociology, and socio-psychology".

Almost forty years in, little has changed in that trust is still persuasive, elusive, and even more interdisciplinary. And as more and more social interactions become mediated by technology, the scale and complexity of social systems has exploded, making trust even harder to study. New phenomenons such as reputation systems, eCommerce websites, peer-to-peer economy emerge, giving birth to booming research areas such as social networks analysis. Many more recent studies have also taken the advantages of the advancement of technology, using computerized experiments to increase the scale of the studies, or developing new algorithms to analyze trust in more complex networks. My work on trust follows the same path, leveraging advancing computer science techniques to address the complex problem of trust.

Before moving on to reviewing trust literature, I want to briefly discuss another mechanism working hand-in-hand with trust -- power and control. In the conclusion section of the edited book, Gambetta wrote (a bit tongue-in-cheek), "Can We Trust Trust?". Gambetta provided a definition of trust, which is "a particular level of the subjective probability with which an agent assesses that another agent or group of agents will perform a particular action, both before he can monitor such action (or independently of his capacity ever to be able to monitor it) and in a context in which it affects his own action [1]. This definition of trust allows for the clarification between distrust, complete trust, blind trust, faith, etc. Almost paradoxically, for trust to be relevant, there must be the probability that the partner can betray or defect. Without risks, trust is irrelevant. In other words, trust is device for coping with the freedom of others. Power and control can reduce the freedom of others, through possibly coercion, contracts, even reputation, making trust irrelevant. In fact, Luhmann also wrote extensively on power theory, in the same book where he defined trust as the "mechanism for the reduction of social complexity" [3].

The social structures can be controlled and shaped by certain forces of power, increasing or decreasing risks -- then the trust decisions at individual and system level will need to react to the forces to further deal with the complexity. In the next chapters, I focus only on trust while being aware of the forces of control at work (e.g., reputation systems).

References