Contemporary public discourse is saturated with speech that vilifies and incites hatred or violence against vulnerable groups. The term “hate speech” has emerged in legal circles and in ordinary language to refer to these communicative acts. But legal theorists and philosophers disagree over how to define this term. This paper makes the case for, and subsequently develops, the first corpus-based analysis of the ordinary meaning of “hate speech.” We begin by demonstrating that key interpretive and moral disputes surrounding hate speech (...) laws—in particular, surrounding their compatibility with the rule of law, democracy, and free speech—depend crucially on the ordinary meaning of “hate speech.” Next, we argue, drawing on recent developments in legal philosophy, that corpus linguistics constitutes a distinctively promising tool for ascertaining the ordinary meaning of “hate speech.” Finally, we offer a proof of concept, by outlining, and analyzing the interpretive and moral implications of, the first such study. (shrink)

Faced with hate speech, oppressed groups can use their own speech to respond to their verbal oppressors. This “counterspeech,” however, sometimes itself takes on a hateful form. This paper explores the moral standing of such “hateful counterspeech.” Is there a fundamental moral asymmetry between hateful counterspeech, and the hateful utterances of dominant or oppressive groups? Or are claims that such an asymmetry exists indefensible? I argue for an intermediate position. There _is_ a key moral asymmetry between these two forms of (...) speech. But, this asymmetry notwithstanding, hateful counterspeech is capable of enacting serious harms—and so, contrary to what many legal theorists have argued, it is in principle an appropriate object of legal regulation. I begin by considering the central argument for thinking that hateful counterspeech is not seriously troubling. This argument holds that oppressed groups lack authority—and, by extension, “speaker power.” Yet this argument, I suggest, sits in tension with the fact that low-status members of dominant groups can, through their utterances, seriously harm members of oppressed groups. Philosophers of language have developed sophisticated arguments to explain this last phenomenon: they have argued that speaker power is relativised to particular jurisdictions; that it can be acquired dynamically in local settings; and that it is socially dispersed. I argue that, in light of these arguments, it appears that hateful counterspeech, too, can generate serious harms. Nevertheless, I show that this conclusion is compatible with recognising a crucial moral asymmetry between hateful counterspeech and the hate speech of oppressors. (shrink)

This article surveys the classic and contemporary literature on the nature and limits of freedom of expression (or free speech). It begins by surveying the main philosophical justifications for free speech, before moving to consider the two most discussed topics in the free speech literature: hate speech and pornography. The article offers some brief reflections on the large number of arguments which have been offered on these topics. Three newer battlegrounds for free speech are examined at the end: no platforming, (...) fake news and online shaming. (shrink)

Artificially generated content threatens to seriously disrupt the public sphere. Generative AI massively facilitates the production of convincing portrayals of fabricated events. We have already begun to witness the spread of synthetic misinformation, political propaganda, and non-consensual intimate deepfakes. Malicious uses of the new technologies can only be expected to proliferate over time. In the face of this threat, social media platforms must surely act. But how? While it is tempting to think they need new sui generis policies targeting synthetic (...) content, we argue that the challenge posed by generative AI should be met through the enforcement of general platform rules. We demonstrate that the threat posed to individuals and society by AI-generated content is no different in kind from that of ordinary harmful content—a threat which is already well recognised. Generative AI massively increases the problem but, ultimately, it requires the same approach. Therefore, platforms do best to double down on improving and enforcing their existing rules, regardless of whether the content they are dealing with was produced by humans or machines. (shrink)

The weaponizing of scientific expertise to oppose regulation has been extensively studied. However, the relevant studies, belonging to the emerging discipline of agnotology, remain focused on the analysis of empirical corruption: of misinformation, doubt mongering, and other practices that cynically deploy expertise to render audiences ignorant of empirical facts. This paper explores the wrongful deployment of expertise beyond empirical corruption. To do so, I develop a broader framework of morally subversive expertise, building on recent work in political philosophy (Howard, 2016). (...) Expertise is subversive if it sets up its audience to fail morally, either intentionally or negligently. I distinguish three modes of subversive expertise: empirical subversion (the focus of agnotology), normative subversion and motivational subversion. Drawing on these distinctions, I offer a revisionary account of the Trump Administration’s regulatory science as a case study. I show that the Trump Administration’s use of expertise to dismantle climate regulation, contra the standard charge, cannot be explained using the resources of agnotology alone: the Administration produced highly reliable climate assessments, detailing the risks of climate change, candidly admitting the harms of its proposed policies, and still successfully deployed these findings to justify massive climate deregulation. The lesson of the analysis is that dismissing the expertise that underpins climate deregulation as empirically corrupt ‘anti-science’ both obscures its actual role in the politics of climate change and understates its wrongfulness: it misses the breadth of the assault on moral agency that sustains climate injustice. (shrink)

People sometimes use speech to interfere with other people’s speech, as in the case of a heckler sabotaging a lecture with constant interjections. Some people claim that such interference infringes upon free speech. Against this view, we argue that where competing speakers in a public forum both have an interest in speaking, free speech principles should not automatically give priority to the ‘official’ speaker. Given the ideals underlying free speech, heckling speech sometimes deserves priority. But what can we say, then, (...) about situations in which heckling clearly seems to infringe upon people’s civil liberties, in a way that intuitively justifies intervention? In such cases, we argue, heckling infringes upon people’s associative freedom. We present and defend an ethical framework for the institutional management of ‘Speech Fights’, geared around this insight. (shrink)

Our time is marked by a resurgence of hate that threatens to increase oppression. Social media has contributed to this by acting as a medium through which hate speech is spread. How should we model the spread of hate? This article considers two models. First, I consider a simple contagion model. In this model, hate spreads like a virus through a social network. This model, however, fails to capture the fact that people do not acquire hatred from a single infectious (...) contact. Instead, it builds up in a person's beliefs and attitudes through time until the infection reaches a level where the subject themselves becomes a generator of hate speech. Second, to accommodate this, I consider an alternative model known as complex contagion. I argue that not only is a complex contagion model more explanatory and predictive, but it can be used to explain why certain features of social media cause it to be a promoter of hate. I conclude by sketching some mitigation strategies. (shrink)

University administrators should not enforce speech codes because speech codes are generally counterproductive to a university’s educational mission. In making the case against campus speech codes, we consider and reply to four of the most prominent arguments in favor of restricting student speech. These arguments appeal to the values of harm prevention, inclusive education, relational equality, and the overall promotion of free speech. We show that speech restrictions do not effectively promote these values. We conclude that campus administrators should uphold (...) protections for freedom of expression on college campuses in order to promote a more egalitarian and inclusive learning environment. (shrink)

In this article, we argue that YouTube’s algorithm should be programmed to make a modest but significant percentage (e.g. 2%) of recommendations for the climate. Just as a librarian has a (meta-editorial) responsibility to highlight certain titles and not others, we believe that so should YouTube’s algorithm. The company, we argue, has duties of content moderation, reparation and meta-editing, as well as strong consequentialist reasons to program its algorithm to do so. With 2 billion users, our proposed intervention could be (...) an effective contribution to mitigating the climate crisis in a transparent and accountable way. We consider different setups, with varying degrees of transparency and centralization. We then address the worries that such a project may raise: the risk of manipulation, the threat of a slippery slope, and the concerns for freedom of expression. We conclude that none of these elements seriously undermine the desirability of our proposal. (shrink)

Journal of Social Philosophy, EarlyView.

Freedom of thought means freedom from social tyranny, the capacity to think for oneself, to encounter even shocking ideas without shrinking away from them. That aspiration is a core concern of the free speech tradition. It is not specifically concerned with law, but it explains some familiar aspects of the First Amendment law we actually have—aspects that the most prevalent theories of free speech fail to capture. It explains the prohibition of compelled speech, and can clarify the perennial puzzle of (...) why freedom of speech extends to art and literature. It also tells us something about the limits of legal regulation, and about the ethical obligations of private actors. (shrink)

Harmful and dangerous language is frequent in social media, in particular in spaces which are considered anonymous and/or allow free participation. In this paper, we analyse the language in a Telegram channel populated by followers of Donald Trump, in order to identify the ways in which harmful language is used to create a specific narrative in a group of mostly like-minded discussants. Our research has several aims. First, we create an extended taxonomy of potentially harmful language that includes not only (...) hate speech and direct insults, but also more indirect ways of poisoning online discourse, such as divisive speech and the glorification of violence. We apply this taxonomy to a large portion of the corpus. Our data gives empirical evidence for harmful speech such as in/out-group divisive language and the use of codes within certain communities which have not often been investigated before. Second, we compare our manual annotations to several automatic methods of classifying hate speech and offensive language, namely list based and machine learning based approaches. We find that the Telegram data set still poses particular challenges for these automatic methods. Finally, we argue for the value of studying such naturally occurring, coherent data sets for research on online harm and how to address it in linguistics and philosophy. (shrink)