In Defense of Democratic Noise: Crisis Management, Big Data and Social Feedback

• In Event: The Limitations of Modern Democracy

Abstract

Does democracy provide the appropriate framework for effective crisis management? In cases of crisis, such as climate change and recently the COVID-19 pandemic, there is an overestimated appreciation of what expert analysis can do and an undervalued place for the contribution of an open and inclusive process of democratic participation in crisis management. In this paper, I explain why democracy through acts of public participation and deliberation, is a necessary epistemic complement to expert analysis that relies on data analysis.

It is tempting to consider democracy as a system that generates too much noise and is likely to block the implementation of an effective strategy for crisis prevention, mitigation and management. How is this effective strategy defined? Governments turn to experts for sound analysis and policy prescriptions. Experts use empirical observations and the processing of large data to understand ongoing developments and anticipate future outcomes. For instance, statistical inference from past data supports probability statements about the future. Advances in data analytics, aided by faster data mining software and machine learning technology, allow experts to track and process large amounts of data as they evolve and generate diagnostic predictive and prescriptive analysis. In this paper, I point to the limits of data analysis of human behavior and policy effect, and I emphasize the indispensable need for democratic voice to enable governments to act more effectively.

I defend democratic voice on epistemic grounds. First, I challenge the idea that, in human systems, we can forecast the projected behavior of a system with probability statements inferred from past data. Human systems are not ergodic systems, meaning that causal and probability statements inferred from past samples will not hold in the future, because human agency continuously changes the properties of the social environment by engaging in learning and reflection. Agential interactions come to generate unique and novel developments compared to what has happened before. Each context is unique and evolving in distinct ways. Second, tracking and processing big data may give us evolving snapshots of some developments on the ground, but these are partial observations of what is happening, and cannot support predictions of policy effect. In that regard, all policies have some unintended and largely unforeseen outcomes with regard to the kind of data considered. Moreover, policies trigger differentiated reactions in society, including new disagreements and protests. This information is dispersed, generated from interactions in society, and has to be continually discovered by policymakers. Policymakers must track uncensored feedback directly from within society itself. This discovery process depends on citizens being able to openly talk about their experiences and raise any new concerns and complaints. Rather than just tracking given datasets, crisis management requires the participation and initiative of citizens themselves to express their evolving needs, fears and frustrations in society, what I call ‘social feedback discovery’.

Author

• Aris Trantidis, University of Lincoln