Democracy and the Corruption of Speech

The Paradox of Discourses of Corruption

While the United States is currently the most prominent case, the discourses of corruption exist in virtually all of today’s democracies – and, paradoxically, in countries that are the least corrupt in the world. Of the cleanest 30 or so countries listed in the 2021 Corruption Perception Index (CPI), almost all are relatively wealthy democracies located in Europe, North America, and Australia-Asia (Transparency International, 2022). The United States, which is ranked 27nd out of 180 countries on the 2021 survey, is not the cleanest of democracies, but it roughly groups with other large economic powers, including Germany, France, Japan and the United Kingdom. Citizens of these countries will rarely if ever encounter a corrupt demand or transaction. Denmark, for example, has been ranked among the least corrupt countries in the world since the CPI first began their rankings in 1995. Consistent with their high rankings, a 2013 survey commissioned by the EU found that only 3% of Danes reported that they have experienced or witnessed a case of corruption in the last 12 months. And yet 47% nonetheless held the view ‘bribery and connections’ are the easiest ways of obtaining certain public services in their country! The Danish number is the lowest in Europe, where 73% of citizens hold this view (European Commission, 2014: 7). The opinions of Americans track those in other developed democracies. But ask Americans – or, indeed, citizens of most other developed democracies – whether politics in their countries are ‘corrupt’, and chances are that the answer will be ‘yes’. In a New York Times review article following the American 2010 midterm elections, pollster Stanley Greenberg noted that citizens in his focus groups were tuning out ‘politicians’ fine speeches and plans’, and expressing sentiments like these: ‘It’s just words. . . . We don’t have a representative government anymore’ (Greenberg, 2011). Trump later expressed the same dismissal of claims in responding to a sexual assault tape revealed during the 2016 presidential campaign: ‘It’s just word, folks. It’s just words’ (Real Clear Politics, 2016). Trump was exploiting a broadly shared sentiment that speech is not about truth, rightness or truthfulness; it is at best expressive. It follows that no accountability or answerability for speech can follow. Nor can political speech be trusted. The 2018 UK Ipsos Veracity Index (would you generally trust this kind of person ‘to tell the truth, or not?’ Ipsos-MORA, 2018), shows the UK people ranking government ministers and politicians generally as 23rd and 24th out of 25 professions!

I do not deny that corruption of the standard variety does not remain a problem, however small, a problem in the developed democracies. But here I want to examine a different proposition: that the everyday discourse of corruption is not so much the kind measured by Transparency International – though it can be about that – as it is a condemnation of the ways language is used in politics – that is, manipulatively and often deceitfully by public representatives to hide matters that ought to be of public concern. What are people saying when they accuse, generally, politicians of dissimulating or lying? What are they implying when they accuse politicians of speaking hypocritically or double-talking?

In attempting an answer, the strategy I shall pursue will be something like a political theory experiment, in which we take these kinds of popular obsessions at face value, and then ask what they reveal. When we do, we stumble upon a key intuition about the centrality of preserving speech-based commitments in democratic political systems. Democracies work when conflict can be conducted through speech, which means that speech must function as a medium of influence. It must enable negotiations, bargains and compromises; it must involve claims for which speakers are accountable; it must have cognitive content, the references for which are publicly knowable. Without these qualities, speech is ‘just words’. Speakers become locked within epistemic bubbles, unable to share positions, opinions, values and facts (Anderson, 2021). When speech is thus degraded, meanings that might be collectively created in response to conflict find their anchors outside of politics – in religion, tradition, honour or personality – short-circuiting commitments that follow from speech in favour of those based on extra-political authority. These are, of course, strategies of autocrats that find fertile fields within the discourses of corruption, including one of its most damaging incarnations, what Rosenblum and Muirhead (2020) call the new conspiracism.

A New Discourse of Corruption?

There are both endemic and contemporary reasons that the discourse of corruption is prominent in today’s democracies. The endemic reasons are owing to centrality of speech-based commitments to framing and resolving political conflicts. Wherever politics works through the medium of speech, with all of the normatively desirable qualities that speech-based politics has to offer, it comes with the hazard of corrupt political speech, and with this, discourses of corruption – a point to which I return below. Of course, discourses of corruption can reference actual corruption – today, less the petty street-level corruption evident in bribes for government services, and more the grander forms of corruption evident in campaign financing in the United States, legislation influenced by powerful interest groups, voices magnified by money and unequal access to legislature and government for those who are less connected or well-resourced (Teachout, 2008; Thompson, 2000; Warren, 2004). There is also evidence that when people claim that political systems are ‘corrupt’, many mean that the systems respond disproportionately to the wealthy, reflecting, in effect, well-documented resource biases in access (Gilens, 2012; Peters and Ensink, 2015). Or some may simply be uninformed: discourses of corruption may not be rooted in any factual diagnosis of standard corruption, but rather reflect a generalized disaffection from politics, democratic or otherwise. What I am interested in here, however, are discourses of corruption that may reflect a disappointment with politics and political speech, but also function to corrode the central medium of democracy itself, speech-based approaches to political conflict.

Of course, discourses of corruption are not in themselves new (Hofstadter, 2012). What is new is that they now combine with recent structural developments and new technologies, so that they penetrate further, with more speed, than ever before. The structural conditions have been evolving over the last several decades, and are widely recognized: large segments of populations are status insecure, threatened by immigration and demographic changes, stronger civil rights for women and minorities, and changes in religiosity (Pettigrew, 2017). Some places have been left behind by economic globalization. The rapid rise of social media over the last decade now makes it easier for malign actors to organize and exploit these real discontents that result. These new media make it easy to start rumours, offer conspiracies, provide misinformation and falsehoods, which gain their authority simply through rapid repetition to targeted audiences. These tactics and effects crowd out credible but more demanding sources of information, explanation, claims and discourse. Social media makes it possible for sophisticated political organizers to target news and ads, matching content to psychological profiles. Discourses of corruption come loose from any moorings they may have had in diagnosing real problems and pathologies, and function simply as a corruption of speech. Trump’s strategies are emblematic, linking grievances (and often real problems and discontents) to constant but unsubstantiated charges of corruption (‘corrupt Hillary’), and preying on the deep public sentiments that all politicians are corrupt. His rejections of claim-making (‘it’s just words, folks’), constant lying and shifting of positions, reduce speech to mere expression, and represent an assault on the core feature of democracy, speech-based politics.

Why Democracies Are Susceptible to Discourses of Corruption

These developments are well-known. Here I am interested in the question as to what insights into democracy might be inferred from the discourses of corruption. Let us start with the basics. The ethos of democracy is simple and compelling: those who are affected by collective decisions should be entitled to be included in making those decisions (e.g. Fung and Gray, 2023; Young, 2000). Democratic systems empower inclusions – always imperfectly – by distributing entitlements that work as empowerments for those with legitimate claims to inclusion. These include protections from domination, direct empowerments such votes, rights to organize, speak, pressure and lobby, as well as indirect empowerments that underwrite citizen capacities such as education or basic income. Ideally, this institutional infrastructure channels political conflict into talking, voting and other responsive means of collective decision, managed through ecologies of institutions, including public spheres, legislatures, professional administrations and court systems.

A related and fundamental claim of democratic theory (particularly deliberative democratic theory) is that if speech is doing political work – as it must once other resources of collective-decision-making are both limited and broadly distributed – then political decisions will draw their legitimacy from speech-based processes, including the commitments (agreements, bargains, compromises) that result from these processes. Speech generates influence only if commitments follow from the offering and receiving of claims in such a way that participants are influenced by the content of claims (even if decisions are backed by institutionalized powers such as votes) (Bächtiger et al., 2018).

These are all now basic and widely held ideas within most democratic theory. Less attention has been paid to the kinds of securities people need to invest trust in deliberative approaches to political problems, particularly as compared with the other resources they might deploy, such as coercion, economic inducement or reliance on personal or traditional authorities. Why would people invest in the apparent uncertainties of talk-based politics? It is within this domain of uncertainty, generated by political conflict, that the discourses of corruption and conspiracy gain their traction. The insecurity upon which these discourses feed is this: when people are motivated by partisanship – especially professional politicians – they are also motivated to use speech merely to express strategic positions or preferences, and, too often, also to manipulate, frame, distort and mislead. So what the discourses of corruption and conspiracy suggest is a default position – held, probably, by majorities even in the best of democracies – that all political speech is corrupt speech, that the integrity we associate with everyday talk is mostly missing from the domain of political discourse. Politicians like Trump flip this script: if it is all ‘just words’, then people should place their faith not in claims or promises and the institutions that embody them, but in the authority and protection of persons. This is, of course, the authoritarian moment, achieved by using the discourse of corruption to dismiss speech-based claims and commitments.

The corruption of speech is a particular hazard for democracy, just because it is the form of politics that places speech at the centre of political conflict resolution. For those who are attracted to the discourse of corruption (in contrast to politicians, like Trump, who use the discourse strategically), the disappointment with political talk in a democracy probably comes from a sensibility – maybe even a disappointed moral sensibility – that develops within democratic polities that seek to accomplish political work through speech. Without this sensibility, there would be little point in speaking of the corruption of language, or, indeed, corruption more generally.

Framed in this way, the hazards of speech-based politics manifest as deceit in its various forms – not just lying, but also manipulation, dissimulation, conspiracy theories and uses of speech that undermine deliberation, bargaining, negotiation and similar speech-based approaches to political conflict. The deep damages to democracy are twofold. First, corrupted speech undermines self-rule with a species of coercion. ‘To the extent that knowledge gives power’, writes Sisela Bok (1978: 20), ‘to that extent do lies affect the distribution of power; they add to that of the liar, and diminish that of the deceived, altering his choices at different levels’. Deception can, then, ‘be coercive. When it succeeds, it can give power to the deceiver – power that all who suffered the consequences of lies would not wish to abdicate’ (1978: 23). Second, corrupted speech undermines collective self-rule because it erodes speech-based social order. Deceit is one way of replacing claim-based social cooperation motivated by agreements with a kind of coercion that undermines both individual and collective self-rule. It does so both by corroding the commitments that follow from statements, and by undermining the recognitions among speakers that enable speech to generate social cooperation. Social disorder follows not from just from questions of the truth or falseness of propositions – although this can one source of disorder – but more generally from the ways deceit breaks the social bonds between speaker and listener that enable social cooperation (Bonotti and Seglow, 2022).

Corruption of Speech as Social Disorder

The central feat of democracy, channelling conflict into speech, thus generates the hazard is that speakers who trust that commitments follow from claims and promises may be suckered by deception, and those who are suckered are vulnerable to coercion. It is this hazard, according to the theoretical hypothesis I am suggesting here, that keeps the discourses of corruption alive within the heart of democracy – including its distinctive language of betrayals, distrust, manipulation, and conspiracies – as well as authoritarian dangers from those who exploit these discourses.

It follows that what is ‘corrupted’ is not the relationship between words and referents that make them true or false, but rather between words and the social commitments implied by statements. The now common idea that shared facts are under attack in a post-truth era underspecifies the problem. What it at stake are the social relationships established by uses of language (Brandom, 2000). Speech acts both perform and disclose a social world of actors who are, in principle, solid enough that they can trust one another (Warren, 1999). If language has the power to coerce through deceit, this power is parasitic on its more general and essential power to generate social order built on relationships of trust, established by the world of speech. Language use provides social order, not because it references truths external to that order, but because it generates that order as a consequence of what is accomplished through language use. It follows that a democracy must protect from this kind of ordering from corruption of language in this sense, as this kind of corruption destroys a common resource – social interdependence expressed and ordered by language – upon which speech-based politics depends.

Democratic theorists have not fully or adequately conceptualized the potential for language use itself to become a site for corruption. Indeed, this is why Rosenblum and Muirhead’s A Lot of People Are Saying (2020) is important and timely. The oversight in democratic theory – particularly deliberative democratic theory – is owing to the fact that we have focussed mostly on the cognitive work that is accomplished by deliberation in creating, settling, or negotiating moral and factual claims and assertions (cf. Bonotti and Seglow, 2022). But the political work that can be accomplished by speech is only in part about what is deliberated – the conflicts, claims, values, information and matters of substance communicated through language. Speech, especially deliberative speech, is also about the relationships that are established as a consequence of speaking and listening – relationships that constitute speakers as agents who have the kind of solidity others can trust. The exceptions are those who, like Habermas (1984, 1987), ground their approaches in the pragmatist speech act theories originating with the later Wittgenstein. Robert Brandom (2000) – a philosopher of language within this tradition whose work is rich in implications for democratic theory – shows that ability of speakers to convey meaning through statements is enabled by the social relationships they establish as a consequence of speaking. When social actors make claims, they are able to convey content just insofar as each participant in a conversation can assume that every other participant knows how to continue from the commitments (e.g. to descriptions of states of affairs or facts, to norms or to preferences) that are implied in claims or actions. Others then count on the inferences they draw, which include the inference that the speaker takes responsibility for the claim. Over time, commitments form a web of inferential relationships in such a way that actors know how to ‘go on’ from any particular claim or action by another, and they can trust that others will do likewise.

Brandom (2000: 81) illustrates the idea that language use leaves a trail of normative commitments with the evocative image of ‘discursive practice as deontic scorekeeping’. When I speak or act, I take on an obligation with respect to you – namely, that, should you require it, I can take responsibility for the claims I make and explain the reasons behind them. If you respond to what I say, carrying the conversation forward, you take on the same kind of obligation with respect to me.

The key idea here is that speech acts are doubly constitutive of social relationships and individual agency. On one hand, knowing how to use language is to know how to ‘go on’ from the rules, expectations, and norms expressed in speech acts. On the other hand, in so doing speakers take on the normative characteristics of trustworthy agents – in particular, agents responsible for the content of their claims or intentional actions – in relation to those they seek to move or motivate with their claims. Language use is linked intrinsically to trust and trustworthiness of a normatively thick kind: in communicating, each individual becomes an author of claims in such a way that others can infer from them agent-like capacities to commit, and to take responsibility for commitments. Individuals build these fabrics of commitments in such a way that they can move through society with a trust that others are not only non-arbitrary in their actions, but that the rules of social engagement can, in principle, be figured out and negotiated through language where necessary, and then trusted.

We can now identify more precisely what the corruption of speech entails. We know, of course, that it entails deceit and its variations: lying, dissimulation, purposeful omission, evasive language or language that is intentionally obscure. But when we look at these senses of corruption through the pragmatics of language use, it is clear that the problem is not that words come loose from their referents. Rather, language is corrupted when actors violate, usually for the sake of some kind of strategic gain, the inferential structure of speech upon which actors depend to regulate social life and generate social cooperation. Corruption occurs when actors violate the norms that are intrinsic to language use. Among the most troubling examples today are conspiracy theories that are purely expressive and in principle unanswerable, as Rosenblum and Muirhead (2020) extensively show. Where speech is corrupted, so are the collective goods upon which democracy depends: in particular, the rules and norms that give confidence that political conflict can be conducted through the medium of speech. Democracy depends upon a trust that the conflicts, positions and principles that are expressed in words commit an actor to, say, a vote for or against, to a bargain or compromise or to an agreed consensus (Warren, 2006a).

Hedges against the Corruption of Speech

Following from this analysis, we can describe one of the key functions of democratic systems as hedging against the corruption of speech. With a few exceptions, systems that depend upon language should not regulate speech directly, which would damage the very medium upon which it depends. Exceptions include forms of speech that attack the status of persons as speakers, undermining the very possibility of deliberative public spheres (Bonotti and Seglow, 2022; Warren, 2006b). Hate speech, regulated in countries such as Canada and Germany, fall into this category. Likewise, speech that incites actions that harm others should be regulated, as in American First Amendment jurisprudence. But most direct regulation of speech is difficult because the powers that states can use are clumsy, and must meet a high burden of proof to be actionable. Most of the kinds of corruption of speech discussed here are difficult to regulate in these ways.

But institutions can hedge against corrupt speech indirectly, by providing the kinds of incentives and experiences that induce people to approach political conflict through persuasion, and to rely on language-based commitments that follow. We already do so when the stakes are very high and conflicting interests raise the incentives to deceive, as in the cases of sworn testimony in judicial proceedings and before legislative oversight committees. These are not models for political deliberation and discourse; but they do suggest that existing institutions pay tribute to the dangers of corrupt speech. While there are no silver bullets for the corruption of political speech, this analysis underscores the importance of assessing the incentives within institutional ecologies of democracies for strategic actors to use credible speech, and for citizens to exercise good judgement with respect to claims. While far beyond the scope of this short article, many familiar and some innovative institutions can be assessed from this perspective, including, for example, standard anti-corruption measures (e.g. Rose-Ackerman, 1999; Warren, 2006a), proportional representation electoral systems that increase deliberative incentives (Steiner et al., 2004), reforming legislatures to reduce strategic incentives for corrupt speech (Warren and Mansbridge, 2015), fact-checking and high journalistic standards, social media algorithms that disincentivize misinformation and disinformation and deliberative minipublics that engage citizen judgement and serve as popular trusted information proxies (Warren and Gastil, 2015). Democracies must continue to imagine, build and support institutions that treat credible language use and its cumulative effects as a common pool resource – one which, like any commons, can be squandered, despoiled and corrupted.