Digital Currencies and the Social Construction of Value
The advent and tumultuous rise of digital currencies such as Bitcoin has thrown into stark relief many fundamental questions about the social nature of both value and money. Using a combination of computational modeling, automated text analysis, and quantitative methods, my dissertation research considers a three separate but interrelated questions concerning the social processes involved in the definition and valuation of these new, monetary objects.
The first part of this research looks at how a reconceptualization of valuation as a process of learning under uncertainty allows us to account for how initially useless objects, such as the “string of bits” which underlie new digital currencies, can come to hold value via social interaction. Using agent-based modeling and Bayesian agents, this work shows how “something” can arise out of “nothing” in social valuation processes, demonstrates the sensitivity of social valuation processes to time, initial states, and early actors, and shows why assumptions concerning markets’ ability to find the “true” value of an asset are not likely to hold when actors rely upon mixtures of social and non-social sources of information in their valuation processes.
For the second part of this work, I use text scraped from 100,000s of posts generated in the online community forums that have been at the center of digital currencies’ invention and adoption. Via a mix of qualitative coding and automated content analysis, I focus on exploring and to identifying the different “talks” (Swidler 2001) of money and value that participants at the heart of this new monetary project have engaged in to make sense of whether Bitcoin qualifies as money and the processes through which it has come to hold value.
The final chapter combines automated content analysis and topic modeling of news reporting on Bitcoin, along with historical documentation of Bitcoin’s evolution to date and trends in quantitative metrics such as Google search volumes, Bitcoin exchange rate and market activity, and venture capital funding in digital currency businesses, to consider the different ways Bitcoin has been defined over the course of its development. The goal of this work is to show how Bitcoin’s multivalent identity has both facilitated its adoption across a diversity of groups while simultaneously, leaving it vulnerable to co-option by the very same powerful actors it was initially intended to subvert.
Mental Representation and the Emergence of Cultural Processes
Cultural sociology has offered some of the most compelling and convincing models of social life ever developed. At the heart of many of them lies a notion of man as a meaning-driven entity and the social world as a milieu of sense-making activity which fundamentally determines the structure of social life. As compelling as these models often are, however, their theorizations of how individual meaning-making aggregates into collective, cultural phenomena are often characterized by a level of opaqueness and vagueness that make their application to social research more intractable than ideal.
This line of work uses contemporary research into the cognitive mechanics of mental representation and the known dynamics of social learning systems to investigate 1) if a parsimonious, and empirically grounded model of a meaning-making actor can be specified and 2) if we can rigorously elaborate upon that individual-level microfoundation in order to systematically generate emergent cultural phenomena. Using agent-based modeling and established work in complex systems research, I am able to demonstrate the viability of such a cultural actor for general social theory and show how identifying mental representation as a basic mechanism in social construction processes enables a variety of new developments in both cultural theory and empirical research methods.
Quantitative and Computational Methodologies in the Study of Culture
Social researchers and theorists have long understood the role of culture in structuring not only perception, but social life more generally. By specifying more precisely the relationship of individual and collective sense-making to observable profiles of behavior and statement, I hope to help establish a methodological approach that aims to investigate cultural dynamics via trends in how variables come to “hang together” and “fall apart” across time and interactional spaces. Toward this end, I am especially interested in the natural affinities that exist between this perspective of cultural and statistical modeling and machine learning methods (i.e. principle component analyses, clustering, topic modeling) that are built to identify and trace this sort of structuring in large amounts of social data.
Lynette A. Shaw, Ph.D. 