Journal of Free Speech Law: "Anonymity, Identity, and Lies," by Prof. Artur Pericles L. Monteiro

The sixth of twelve articles from the Knight Institute’s Lies, Free Speech, and the Law symposium.

The article is here; the Introduction:

Anonymity has emerged in recent years as an important focus of debates about the digital public sphere. An opinion piece in The Wall Street Journal argued that a solution to the problems besetting social media was to "end anonymity." Soon after, Senator John Kennedy announced he would introduce a bill to ban anonymity online. In the United Kingdom, anonymity also featured in the discussions about the Online Safety Act. Bills designed to curb anonymity are also frequent in Brazil, including more recently with one introduced by the select committee investigating the Bolsonaro administration's handling of the pandemic as part of the committee's recommendations included in the final report to punish those who engage in disinformation.

Supporters of proposals targeting anonymity sometimes argue that requiring users to make themselves known will remedy many of the pathologies afflicting the digital public sphere, including misinformation. Identification is seen as a tool for creating a more truth-based discourse, by inducing speakers to behave more responsibly, as well as providing listeners with information to assess the credibility of the speaker. The assumption often is that anonymity promotes lies and incivility, while identification induces truth and civility. Nathaniel Persily sums it up:‌ "If online anonymity is the cause of many of the democracy-related ills of social media, then disclosure might be the best disinfectant."

In fact, in an environment beset by political polarization, instead of serving as a disinfectant, identification can add fuel to the fire of mis- and disinformation. Not only that, anonymity can have a role also in enabling public political deliberation that has been underappreciated. This paper surveys literature from multiple disciplines and challenges assumptions behind the prevailing stances towards anonymity and mis- and disinformation. It argues that anonymity and identification do not have a fixed function; it instead refers to the plurality of identification and the plurality of anonymity. "Plurality" is meant to emphasize that both anonymity and identification shape and are shaped by factors such as social norms and platform affordances. As such, whether identification will contribute to a more truth-based public discourse and to a more civic-minded digital sphere is a question that can only be answered if we account for those factors. Considering the identity-based components of the spread of disinformation in polarized contexts, anonymity can serve as a device to create opportunities for conversation and avoid some of the mechanisms triggering those components.

A few notes on terminology and scope should be helpful. Anonymity stands for namelessness in the vernacular, yet conceptually it must be appreciated as going beyond names. In fact, names are less effective as unique identifiers, as they often can shared by more than one person. Identification correspondingly is not constrained to names. Identification and anonymity can be seen as "different poles of a continuum." Anonymity is relational:‌ Someone might have knowledge that allows them to identify a speaker, while another person might not.

This relational aspect can be relevant particularly when we are considering illegal content, where it is not just listeners who pass judgment on anonymous speech, but also authorities seeking to hold speakers accountable. That is, the audience not having knowledge that identifies a speaker (because their name is not unique or the speaker uses a pen name) can be a different issue than law enforcement and other officials being able to trace the speech. Although the questions are connected, this paper will not discuss traceability. It will focus on identification with one kind of identifier, real names, as a lever that commentators and policymakers have turned to with the aspiration of governing legal speech. Combatting mis- and disinformation is one reason why commentators want to expand identification. One shared hope is that both speakers and listeners will be closer to the truth through real-name identification. That is the central concern of this paper.

Part I introduces the concept of the plurality of identification, which the paper uses to call attention to how real names have a different operation on social media. Names, which were not ubiquitously employed to the same extent they are now (e.g., full names in Facebook profiles), work in markedly transformed ways when they offer an index to massively aggregated, permanent information on every one of us that is accessible through social media and search engines. Calls for identification often rest on an assumption that real names instantiate the same identity regardless of the context they are displayed. This ignores the impact of context collapse—the flattening of different social contexts—in impelling individuals to perform their identity to an imagined, unspecified audience with which they engage much like micro-celebrities.

At the same time, anonymity is thought to prevent accountability by disconnecting us from drivers of norm-abiding behavior. Part II shows that this is only sometimes true—and introduces the plurality of anonymity. It surveys research establishing that anonymous settings may produce greater conformity to local, i.e., group-related, social norms (which may or may not be democratically desirable). The paper then argues that the impact of anonymity on user behavior depends on content moderation practices and community norms.

Part III consolidates those points and discusses the role of political polarization and the sharing of false information. Although it is commonly assumed that identification is a means of fostering veracity, as well as civility, this is often not the case. The paper explores findings from psychology and computational social science to argue that real names are part of mechanisms that drive misinformation in settings marked by affective polarization (negative attitudes toward the other party). Anonymity, conversely, has potential as a device for reducing polarization as well as creating opportunities for conversations not infected by those mechanisms.

This paper aims to add to a years-long debate about the place of anonymity in a healthy digital public sphere. Much work has been done about the disproportionate effects flowing from real-name policies to marginalized communities, to individuals who have legitimate reason to fear for their safety in disclosing their real names, or to those whose names do not match their official government identification. Indeed, in 2011, the announcement of now-defunct Google Plus's real-name policy prompted considerable backlash along those lines, leading in what were described as the Nymwars; in 2015, a new battlefront turned to changes in Facebook's enforcement of its policies, which was met with opposition by a collection of civil society organizations gathered around the Nameless Coalition. Scholars have suggested that such concerns can be addressed in specific cases and exceptionally, only "where anonymity is needed to avoid 'threats, harassment, or reprisals,'" as Justice Scalia argued in McIntyre, a landmark case on the topic. My hope with this paper is to explore the role of anonymity and identification even beyond the risk of speech suppression and disproportionate effects.