Polarization in groups of Bayesian agents

Josefine Pallavicini; Bjørn Hallsson; Klemens Kappel

doi:10.1007/s11229-018-01978-w

Polarization in groups of Bayesian agents

Josefine Pallavicini^*, Bjørn Hallsson, Klemens Kappel

^*Corresponding author for this work

Department of Communication

Abstract

In this paper we present the results of a simulation study of credence developments in groups of communicating Bayesian agents, as they update their beliefs about a given proposition p. Based on the empirical literature, one would assume that these groups of rational agents would converge on a view over time, or at least that they would not polarize. This paper presents and discusses surprising evidence that this is not true. Our simulation study shows that these groups of Bayesian agents show group polarization behavior under a broad range of circumstances. This is, we think, an unexpected result, that raises deeper questions about whether the kind of polarization in question is irrational. If one accepts Bayesian agency as the hallmark of epistemic rationality, then one should infer that the polarization we find is also rational. On the other hand, if we are inclined to think that there is something epistemically irrational about group polarization, then something must be off in the model employed in our simulation study. We discuss several possible interfering factors, including how epistemic trust is defined in the model. Ultimately, we propose that the notion of Bayesian agency is missing something in general, namely the ability to respond to higher-order evidence.

Original language	English
Journal	Synthese
ISSN	0039-7857
DOIs	https://doi.org/10.1007/s11229-018-01978-w
Publication status	Published - Jan 2021

Keywords

Bayesian updating
Epistemic rationality
Epistemic trust
Group polarization
Higher-order evidence

Access to Document

10.1007/s11229-018-01978-w

Cite this

@article{374fdb1cedad46b599d5a9a017181630,

title = "Polarization in groups of Bayesian agents",

abstract = "In this paper we present the results of a simulation study of credence developments in groups of communicating Bayesian agents, as they update their beliefs about a given proposition p. Based on the empirical literature, one would assume that these groups of rational agents would converge on a view over time, or at least that they would not polarize. This paper presents and discusses surprising evidence that this is not true. Our simulation study shows that these groups of Bayesian agents show group polarization behavior under a broad range of circumstances. This is, we think, an unexpected result, that raises deeper questions about whether the kind of polarization in question is irrational. If one accepts Bayesian agency as the hallmark of epistemic rationality, then one should infer that the polarization we find is also rational. On the other hand, if we are inclined to think that there is something epistemically irrational about group polarization, then something must be off in the model employed in our simulation study. We discuss several possible interfering factors, including how epistemic trust is defined in the model. Ultimately, we propose that the notion of Bayesian agency is missing something in general, namely the ability to respond to higher-order evidence.",

keywords = "Bayesian updating, Epistemic rationality, Epistemic trust, Group polarization, Higher-order evidence",

author = "Josefine Pallavicini and Bj{\o}rn Hallsson and Klemens Kappel",

year = "2021",

month = jan,

doi = "10.1007/s11229-018-01978-w",

language = "English",

journal = "Synthese",

issn = "0039-7857",

publisher = "Springer",

}

TY - JOUR

T1 - Polarization in groups of Bayesian agents

AU - Pallavicini, Josefine

AU - Hallsson, Bjørn

AU - Kappel, Klemens

PY - 2021/1

Y1 - 2021/1

N2 - In this paper we present the results of a simulation study of credence developments in groups of communicating Bayesian agents, as they update their beliefs about a given proposition p. Based on the empirical literature, one would assume that these groups of rational agents would converge on a view over time, or at least that they would not polarize. This paper presents and discusses surprising evidence that this is not true. Our simulation study shows that these groups of Bayesian agents show group polarization behavior under a broad range of circumstances. This is, we think, an unexpected result, that raises deeper questions about whether the kind of polarization in question is irrational. If one accepts Bayesian agency as the hallmark of epistemic rationality, then one should infer that the polarization we find is also rational. On the other hand, if we are inclined to think that there is something epistemically irrational about group polarization, then something must be off in the model employed in our simulation study. We discuss several possible interfering factors, including how epistemic trust is defined in the model. Ultimately, we propose that the notion of Bayesian agency is missing something in general, namely the ability to respond to higher-order evidence.

AB - In this paper we present the results of a simulation study of credence developments in groups of communicating Bayesian agents, as they update their beliefs about a given proposition p. Based on the empirical literature, one would assume that these groups of rational agents would converge on a view over time, or at least that they would not polarize. This paper presents and discusses surprising evidence that this is not true. Our simulation study shows that these groups of Bayesian agents show group polarization behavior under a broad range of circumstances. This is, we think, an unexpected result, that raises deeper questions about whether the kind of polarization in question is irrational. If one accepts Bayesian agency as the hallmark of epistemic rationality, then one should infer that the polarization we find is also rational. On the other hand, if we are inclined to think that there is something epistemically irrational about group polarization, then something must be off in the model employed in our simulation study. We discuss several possible interfering factors, including how epistemic trust is defined in the model. Ultimately, we propose that the notion of Bayesian agency is missing something in general, namely the ability to respond to higher-order evidence.

KW - Bayesian updating

KW - Epistemic rationality

KW - Epistemic trust

KW - Group polarization

KW - Higher-order evidence

UR - http://www.scopus.com/inward/record.url?scp=85055968291&partnerID=8YFLogxK

U2 - 10.1007/s11229-018-01978-w

DO - 10.1007/s11229-018-01978-w

M3 - Journal article

AN - SCOPUS:85055968291

SN - 0039-7857

JO - Synthese

JF - Synthese

ER -

Polarization in groups of Bayesian agents

Abstract

Keywords

Access to Document

Other files and links

Fingerprint

Cite this