Santagiustina C., Wargalie M., Bernasconi M. (2018).

Talking About Uncertainty. Università Ca’ Foscari di Venezia



In the first article we review existing theories of uncertainty. We devote particular attention to the relation between metacognition, uncertainty and probabilistic expectations. We also analyse the role of natural language and communication for the emergence and resolution of states of uncertainty. We hypothesize that agents feel uncertainty in relation to their levels of expected surprise, which depends on probabilistic expectations-gaps elicited during communication processes. Under this framework above tolerance levels of expected surprise can be considered informative signals. These signals can be used to coordinate, at the group and social level, processes of revision of probabilistic expectations. When above tolerance levels of uncertainty are explicated by agents through natural language, in communication networks and public information arenas, uncertainty acquires a systemic role of coordinating device for the revision of probabilistic expectations. The second article of this research seeks to empirically demonstrate that we can crowd source and aggregate decentralized signals of uncertainty, i.e. expected surprise, coming from market agents and civil society by using the web and more specifically Twitter as an information source that contains the wisdom of the crowds concerning the degree of uncertainty of targeted communities/groups of agents at a given moment in time. We extract and aggregate these signals to construct a set of civil society uncertainty proxies by country. We model the dependence among our civil society uncertainty indexes and existing policy and market uncertainty proxies, highlighting contagion channels and differences in their reactiveness to real-world events that occurred in the year 2016, like the EU-referendum vote and the US presidential elections. In the third article, we propose a new instrument, called Worldwide Uncertainty Network, to analyse the uncertainty contagion dynamics across time and areas of the world. Such an instrument can be used to identify the systemic importance of countries in terms of their civil society uncertainty social percolation role. Our results show that civil society uncertainty signals coming from the web may be fruitfully used to improve our understanding of uncertainty contagion and amplification mechanisms among countries and between markets, civil society and political systems.