Dog-whistling and democracy

Abstract

Despite our fascination with dog-whistling, neither dog-whistling itself, nor the relationship between dog-whistling and democracy are well understood. This article separates the content from the technique of dog-whistling and develops a more precise conceptualisation of that phenomenon in order to untangle the ambiguity about it. Dog-whistling, it argues, should not be reduced to racism or verbal communication tactics but can be combined with various different worldviews, and can encompass a multiplicity of verbal and non-verbal communicative means which surreptitiously nudge or wink at a specific subgroup. Contra the prevalent conviction that dog-whistling is antithetical to democracy in toto, the article suggests that some manifestations of that phenomenon constitute a ‘lesser vice’ vis-à-vis a politics of zealous candour, and an ineluctable feature of democratic politics. This recognition has important implications for democratic theory, and recent lamentations that ours is an age of moral crisis, marked by the rise of post-truth politics.

‘When I use a word’, Humpty Dumpty said in rather a scornful tone, ‘it means just what I choose it to mean-neither more nor less’.

‘The question is’, said Alice, ‘whether you can make words mean so many different things’.

Lewis Carroll, Through the Looking Glass

Cynicism about politics is ubiquitous, and not without reason. Such attitudes, which are a persistent finding of numerous opinion polls, encapsulate the conventional wisdom that political discourse is rife with duplicity (Hay, 2007). Recently, however, a more subversive phenomenon has become mainstream in our political lexicon and has attracted considerable attention by pundits, politicians, and democratic citizens: ‘dog-whistling’ – the utilisation of words, which as Alice notes in her riposte to Humpty Dumpty, ‘mean many different things’. The contemporary fascination with that phenomenon – gleaned in the raft of headlines and pieces reporting and castigating instances of it (BBC, 2019; Haney-Lopez, 2019; New Statesman, 2019) – suggests that dog-whistling deserves far greater analytical attention than it has thus far received from political theorists and scientists. For, despite the ubiquity of denunciations of dog-whistling, it is not clear what we mean when we talk about dog-whistling in the first place.

While dog-whistling is generally construed as a species of racist propaganda, ¹ it is specifically invoked in at least two different ways. The first, narrow conception, conceives of dog-whistling as a ‘secret language’ (BBC, 2019), and reduces it to deplorable, clandestine rhetorical operations (Filimon, 2016; Lasch, 2016; Wetts and Willer, 2019). For instance, the terms “states’ rights”, “welfare” and “inner cities” (among others), popular among Republican politicians, constitute veiled references to African Americans, and – as numerous studies illustrate – effective, coded appeals to the racially resentful (Hurwitz and Peffley, 2005; Mendelberg, 2001). The second, seemingly more expansive, conception is utilised against narratives which communicate racist views more overtly (Haney-Lopez, 2019; Lasch, 2016; Steen, 2019; Wetts and Willer, 2019). For example, the cries of dog-whistling were heard loud and clear against Boris Johnson’s description of veiled women as ‘bank robbers’ and ‘letterboxes’ (New Statesman, 2019). Commenting on Trump’s election, Michael Hirsch, and Robbie Gramer (2018) note that:

‘Make America Great Again’ carried [Trump] . . . to the White House on a torrent of innuendo about the sorts of minority people – Mexican ‘rapists’, migrants from ‘shithole’ countries . . . who didn’t fit his vision of the future . . . a twisted vision of a racist past . . . shaped by vicious dog-whistles that spoke so clearly to white supremacists.

These examples reveal the conceptual poverty and vacuity of the rhetoric of dog-whistling and of accusations of it. This article seeks to articulate a more precise conceptualisation of dog-whistling in order to go some way towards untangling the ambiguity about the term. Contra the narrow and seemingly more expansive conceptions of that phenomenon, I suggest that dog-whistling should not be reduced to racism or verbal communication tactics. Rather, it can be combined with a range of different worldviews; and, it might encompass a multiplicity of verbal and non-verbal communicative means which serve to nudge or wink at a specific subgroup, and which are intended to remain inaudible to those who espouse different worldviews and/or who might deem more explicit appeals disagreeable or offensive. By separating the content from the technique of dog-whistling, and thereby rethinking dog-whistling along the aforementioned lines, I endeavour to shed new light on the question of what place dog-whistling should have in a democratic polity. Contra the prevalent conviction that dog-whistling is detrimental to democratic politics – that, a democratic polity should be inhospitable to dog-whistling tout court and that democratic politicians should steer clear of it (BBC, 2019; Stanley, 2017) – I wish to argue that dog-whistling qua technique, might not be as problematic as is often thought. I do not deny that dog-whistling poses problems to our politics, or that certain manifestations of it are dangerous for democracy. Rather, I wish to carve some room for dog-whistling in democratic politics. Dog-whistling, I argue, might be a ‘lesser vice’ vis-à-vis a politics of zealous, uncontaminated candour. More affirmatively, I suggest that democratic societies might be implicated in creating the impetus to dog-whistling: democratic politicians operate in a messy context which often renders dog-whistling necessary.

My argument has crucial implications for the ancient-old question of how to conceive of democratic politics and what it means to lead a virtuous political life in this context. Specifically, it casts doubt on a set of core aspirations of democratic theory, and of the conventional vogue of the deliberative turn which fuels scholarly dismissals of dog-whistling – the conviction that there exists an affinity between a liberal democratic culture, unconditional transparency, and the realisation of the common good. Correspondingly, it also poses problems for historically persistent, albeit recently more pronounced, lamentations that ours is an age of a deepening moral crisis in public life, marked by the rise of post-truth politics. This is not to say that anything goes in public life. My argument suggests that we should re-think the place of critique in democratic politics and sets the foundations for an alternative way of critiquing forms of discourse without perpetuating romanticised visions of democracy.

The discussion proceeds as follows. In the first section, I develop a preliminary account of dog-whistling. In the second section, I consider the conventional thesis that democratic politics should be inhospitable to dog-whistling. I shall then upset that thesis. The alleged incompatibility between dog-whistling and democracy, I suggest, rests on an unsatisfactory idealisation of democratic politics. I then explain in more substantive terms why dog-whistles might be ineluctable and not altogether worthless modes of democratic politics. In the final section, I consider the wider implications of my argument, by locating it within the larger context of contemporary discussions on the rise of post-truth politics and the erosion of democracy.

The meaning of dog-whistling

Dog-whistling is an ambiguous concept. The ambiguity stems, in part, from the fact that the term is used as a political weapon, and partly because that phenomenon exhibits considerable variety in the way it manifests itself (Saul, 2018a). Although difficult to establish a comprehensive definition of dog-whistling and do justice to the various scholarly definitions of it, it is still possible to offer a more precise characterisation of it. I shall sketch a particular type of dog-whistling, some manifestations of which, I argue, are inescapable in democratic politics: the calculating dog-whistler.

The earliest meaning of dog-whistling which signifies a high-pitched whistle which, though inaudible to humans, produces behavioural arousal in dogs (Krantz, 2009), has important links to the figurative meaning of the term, insofar as the recognition that a single piece of communication can be interpreted differently by different groups remains central to our understanding of it. Dog-whistling in politics constitutes a way of nudging or winking at potential supporters in a way so as to render the substance of the nudge or wink inaudible to others whom it might alienate, or deniable for those who might deem more explicit appeals disagreeable or offensive (Goodin and Saward, 2005; Saul, 2018a).

That dog-whistling enables speakers to reach their target audience through implicit appeals – the utilisation of ‘language’ which ‘has a special meaning for a subset of the population’ (Albertson, 2015: 4) – captures some of the reasons why ‘code words’ – a term which originates in cryptography – is often used as a synonym for it (Khoo, 2017). Communications readable by one party and prima facie indecipherable to another, exhibit features of a code when the source of the message embeds a hidden, not clearly discernible to outsiders, meaning ‘inside’ a general meaning. However, unlike ‘code words’, traditionally construed, dog-whistles need not merely entail a wink – a piece of communication which surreptitiously pledges loyalty to one’s target audience and which is discernible to that audience. Rather, as noted, some dog-whistles might entail a nudge at a specific subgroup – seemingly innocuous statements which serve to ‘recruit’ potential supporters but which, nonetheless, work outside of the target audience’s realm of consciousness, and which would have been also rejected by one’s target audience had they been overtly expressed ² (Mendelberg, 2001; Saul, 2018a). Furthermore, that codes are frequently cracked, to push the metaphor, and that dog-whistling receives bad press, suggests that the message need not be inaudible to others beyond its intended audience (Goodin and Saward, 2005). Yet, while public unmasking and denunciation might diminish the success of some dog-whistles (Saul, 2018a), they need not diminish the political value or impact of dog-whistling as such (Hurwitz and Peffley, 2005). Decoding a dog-whistle does not just rest on the critic’s capacity to access the distinctive systems of knowledge, language, and references specific to the peculiar traditions or subcultures which constitute the target audience of that message; rather, the effectiveness of public denunciation in diminishing the impact of dog-whistling also depends on the electorate’s willingness to punish politicians for engaging in such practices (Goodin and Saward, 2005), and the speaker’s effectiveness to deflate the public’s attention away from that practice, and/or convince their audience that they were not engaging in it. What seems disturbing with dog-whistling is that it trades in studied ambiguity; public pretensions of innocence, even outrage, constitute part of its essence. The ‘point of the code’, Ian Haney-Lopez (2014: 130) notes, is that it enables its practitioners to ‘parry any resulting outrage by playing dumb’.

If, however, dog-whistling constitutes a species of coded communication, what, exactly, is being coded? The standard way of thinking about dog-whistling, which unites both narrow and seemingly more expansive accounts of it, offers a pithy answer to that question: dog-whistling communicates a set of insidious views. This conception of dog-whistling is not counterintuitive; it is traced to a rare public acknowledgement of the conscious usage of dog-whistles in political campaigning – the infamous remarks of Ronald Reagan’s consultant, Lee Atwater:

You start out in 1954 by saying, ‘Nigger, nigger, nigger’. By 1968, you can’t say ‘nigger’ – that hurts you. Backfires. So, you say stuff like forced busing, states’ rights . . . cutting taxes. And all these things you’re talking about are totally economic things and a by-product of them is [that] blacks get hurt worse than whites . . . ‘We want to cut this’, is much more abstract than even the busing thing, and a hell of a lot more abstract than ‘Nigger, nigger’. (Perlstein, 2012)

Atwater’s statement serves as the starting point of some of the most sophisticated accounts of dog-whistling. For instance, Jason Stanley (2017: 81, 151–152) conceives of dog-whistling as ‘nefarious’ propaganda – racist demagoguery which introduces ‘words that function in discourse as slurs but are not explicitly slurs’. Similarly, dog-whistling is construed as a species of ‘strategic’ racism (Filimon, 2016: 25); ‘coded racial narratives’ (Lasch, 2016: 162); ‘messages’ in which ‘race is cued through coded language’ (Wetts and Willer, 2019: 1); ‘words that are fundamentally non-racial’, but have, through ‘association, assumed a strong racial component’ (Hurwitz and Peffley, 2005: 101); and, ‘tactics’ camouflaging ‘racist rhetoric’ (Steen, 2019).

While most reported cases of dog-whistling conform to the standard conception of that phenomenon, to reduce dog-whistling to veiled, bigoted rhetoric would be a mistake. This is acknowledged en passant in Haney-Lopez’s (2014: 4) earlier work: dog-whistling ‘has no particular political valence’ and ‘could encompass clandestine solicitations on a number of bases’. Dog-whistling has certain parallels with Cas Mudde’s (2004: 544) conception of populism as a ‘thin-centred ideology’ which exhibits a ‘restricted core’ and can be attached to various host ideologies. For, it is more broadly characterised by a dual-register function: a nudge or wink at specific aspirations, interests, and attitudes, which form part of a particular tradition or subculture, but which is, at the very least, intended to be veiled from a larger audience – the totality of traditions or subcultures which comprise the fabric of public life and which espouse different, antagonistic substantive aspirations, and interests. On this perspective, dog-whistling can be combined with various substantive worldviews and conceptions of the good. Bethany Albertson’s analysis of religious dog-whistles captures a straightforward manifestation of this broader conception. ‘Wonderworking power’, Albertson notes, a term utilised in G.W. Bush’s 2003 State of the Union, might seem innocent to Bush’s larger audience, but those ‘exposed to a popular evangelical hymn recognized the line as a refrain in “There is Power in the Blood”’. Yet, she emphasises, dog-whistling is not ‘specific to Bush’s rhetoric, or Republicans; Bill Clinton used the phrase “send me”’, which references a passage from the Bible, ‘to structure his endorsement for John Kerry at the 2004 Democratic National Convention’ (Albertson, 2015: 4).

At this point, I should emphasise three issues. First, the recognition that dog-whistling can be elided with different worldviews entails that it need not only constitute an embodiment of the politics of exclusion, as is often thought. Rather, it might entail nudges, implicit appeals, or pledges of solidarity to the historically excluded. Obama’s effective utilisation of ‘race-specific, race-free language’ (Li, 2012) – implicit allusions to racial injustice amid a widely perceived, intentionally crafted ‘race-neutral’ campaign strategy (Stout, 2015) – captures this point neatly. For instance, in January 2009, Obama’s speech on the economy contained a reference to ‘“American dreams that are being deferred,” a phrase that black audiences understood without a citation as black poet Langston Hughes’ (Henderson, 2009). And, in 2010, while refuting e-mails falsely identifying him as a Muslim, he quoted Malcolm X’s words from Spike Lee’s movie to the elation of black audiences: “They try to bamboozle you, hoodwink you” (Li, 2012: 20).

Second, given that politics is the ‘world of appearances’ (Arendt, 1971: 19), dog-whistling need not merely entail coded verbal modes of communication, but can also encompass non-verbal communicative means which surreptitiously nudge or wink at specific audiences – means which have a symbolic meaning, and which form a language which allows audiences to consciously or unconsciously ‘hear’ coded words with their eyes. As Rebecca Klatch (1988: 139–141), in her analysis of symbolism, notes, political symbols constitute ‘sacred expressions of group solidarity’ – they ‘evoke feelings of identification’, ‘binding individuals into a unified whole’ and are ‘essential for political mobilisation’ – but also of differentiation; for, ‘given a plurality of social groups, symbols which unite one group inevitably separate that group from others’. This insight is crystallised in the utilisation of images, visuals, and symbolic objects. ³ A rather straightforward example here is the infamous Willie Horton ad, utilised by G.H.W. Bush in his campaign against Dukakis which criticised the latter’s prison programme by depicting Willie Horton, an African American who raped a woman and stabbed her husband, without mentioning his race and which effectively appealed to racially resentful voters (Mendelberg, 2001). Again, however, such means are not the exclusive property of those courting the racially resentful. Peter Oborne’s (2005: 102–104) analysis of Tony Blair’s ‘chameleon quality’ – his ‘ability to give opposite groups the impression that he is on both of their sides’, by ensuring that ‘the correct message went out to one target audience, and different message elsewhere’ – is instructive. In 1999, for instance, Blair, while being interviewed by Guardian journalists, conspicuously displayed ‘as if it was being read’ Bring Home the Revolution – a book denouncing British Monarchy written by Guardian columnist Jonathan Freedland – with the intention of ‘flattering Guardian journalists with the notion that he is one of them’, only to deny his association with republicanism when interviewed for the Daily Mail, the readers of which are predominantly supporters of the monarchy (Oborne, 2005: 103).

Yet, visuals and imagery, as suggested in the burgeoning literature on political performance, are just two specific communicative means out of the many – posture, gestures, accent, and clothing are also crucial vehicles of suasion (Mendonca et al., 2022; Rai, 2015). Though not all manifestations of these means should be understood as dog-whistling, and while elaborating on all of these is beyond my purposes, I wish to emphasise that such non-verbal communicative means – insofar as they surreptitiously nudge or wink at a specific subgroup – can be understood as dog-whistling. A case in point is Obama’s swagger, which, as argued, constitutes a ‘symbolic racial representation’, implicitly communicating ‘issues in black identity and racial pride . . . rooted in the history and tensions in American politics’ (Spencer, 2014: 168). As Cynthia McIntyre, president of ΔΣΘ sorority notes, Obama’s ‘wisdom, poise, eloquence, even his swagger . . . illustrate the greatness of a people who have been dismissed, discarded, discouraged, disregarded’ (Harris, 2012: 136).

Finally, dog-whistling, so construed, should be distinguished from the conception animating the seemingly more expansive account. The trouble with the latter is not just that it reduces dog-whistling to racist rhetoric and is insufficiently encompassing, but, rather, that it is expansive in the wrong way: it elides dog-whistling with rhetoric which valorises openly that which dog-whistles only implicitly communicate. For instance, Trump’s 2016 campaign rhetoric, seen as a paradigm case of dog-whistling, is quite ‘openly racist’ (Saul, 2018b). The word quite should be emphasised; such rhetoric features ‘utterances’ which provide ‘enough cover for the audience to believe that nodding along does not make them racist’. For example, ‘when Trump called Mexicans rapists, he added the caveat that “some, I assume, are good people”’. Caveats of this sort which typically take the form “I’m not racist but . . .”, and which are also observed in Johnson’s denunciation of veils (his article was presented as a defence of tolerance), serve what Saul (2018b) terms a ‘figleaf’: rhetoric which ‘barely cover(s) something which one isn’t supposed to show in public’. While dog-whistles and figleaves share certain attributes – for example, they both allow for plausible deniability – they should be thus distinguished from one another: ‘If you use a dog-whistle you don’t need a figleaf because the racism is well-concealed’ (Saul, 2018b; see also Saul, 2016).

Dog-whistling, I suggested, need not be reduced to verbal communication or racist rhetoric but can encompass a range of means which covertly nudge or wink at specific subgroups and can be combined with a range of different worldviews. While this broader conception of dog-whistling qua technique might alleviate some of the worries voiced by its opponents, its place in democratic politics remains ambiguous still. If the desired effect of dog-whistling is to evade the speaker’s wider audience and reach its target audience through implicit appeals, then that phenomenon is bound to involve aspects of manipulation and dissimulation. Indeed, one might well suggest that dog-whistling is more pernicious than lying, often seen as the antithesis of democracy (Cliffe et al., 2000). For, lying is typically thought to involve false statements uttered with the intention to deceive one’s audience as such (Mahon, 2016). Dog-whistling, in contrast, entails a multiplicity of rituals which surreptitiously nudge or wink at specific groups and rests on a prior segmentation of one’s audience into groups based on their distinctive worldviews and the communicative conventions inherent in these. What is peculiar of some forms of dog-whistling, then, is not just that the aggregation of the distinctive meanings contained within a specific piece of communication entail a melange of incompatible commitments which cast doubt on the speaker’s truthfulness. Rather, the performative aspect of some forms of dog-whistling is more persistent, more wide-ranging than the liar’s limited repertoire of deceptions; for, it turns on questions of character, identity, and conviction and facilitates the construction of a persona which serves to convey trustworthiness – it implicitly communicates to one’s audience that “I am one of you, because I talk and behave like we do” (Henderson and McCready, 2019). Yet, it is the aforementioned functions which often render dog-whistling ineluctable in democratic politics: dog-whistling enables politicians to cultivate the support necessary for satisfying certain political goods amid a messy, conflict-ridden domain. Before elaborating on this, I shall sketch the conventional way of thinking about the relationship between dog-whistling and democracy.

Dog-whistling as the antithesis of (deliberative) democracy

The contention that dog-whistling and democracy are uneasy bedfellows is straightforward: it rests on the conviction there should exist an affinity between democracy, truthfulness, and the pursuit of the common good. Put differently, democratic politics is attractive partly because it renders the moral vices in general, and the vices of manipulation, and secrecy in particular, undesirable, and unnecessary. This conception of democratic politics is not uncommon: it can be gleaned in various accounts of democracy (Cliffe et al., 2000; Dovi, 2007), and opinion polls exuding our obsessiveness with straight talkers (Allen and Birch, 2012); it also animates recent lamentations that democratic cultures are plagued by a deepening crisis, effected by rise of polarisation and post-truth politics, and the corresponding attempts to extirpate deception from, and revive the pursuit of the common good in, public life (Bunting et al., 2010; McIntyre, 2018). While my argument challenges this way of thinking about democratic politics, I shall focus on a specific manifestation of it. I elaborate a set of arguments positing a radical discontinuity between democracy and dog-whistling, rehearsed by the two most prominent critics of the latter – Jason Stanley, in How Propaganda Works, and Robert Goodin in Innovating Democracy – from the conventional vogue of the philosophical creed of deliberative democracy.

Notwithstanding important differences in the way Stanley and Goodin conceive of deliberative democracy, the general idea which sustains their accounts is that dog-whistling is incompatible with democracy in toto as it disrupts the identity between democracy and “the will of the people” or the majority. Dog-whistling thus destroys a core aspect of democratic politics, the application and enactment of the volonté générale, by compromising the process through which this could be discovered – certain aspects fundamental to political rumination and communication-proper in democratic cultures: the possibility of respectful, impartial intra-personal and inter-personal deliberation on the common good or certain substantive principles to which rational individuals should ascend. At the core of this point lie two distinct, though entwined, arguments, each of which identifies a more specific malaise which dog-whistles engender.

The first argument – call this the respect argument – focuses on the way dog-whistling hinders the possibility of consensus, by elaborating its effect on what Stanley (2017) terms, in a Habermasian or Rawlsian fashion, ‘the normative ideals governing public speech’ (p. 81) – ideals which are distinguished from strategic, manipulative communication or ‘mere’ rhetoric. Dog-whistling, Stanley notes, ‘is impermissible because it is illiberal’: the use of ‘innocent words that have the feature of slurs’ erodes ‘reasonableness’ and cuts off rational discussion – the commitment that arguments and policies must stem from a ‘fair and honest’ deliberative process which treats all viewpoints equally and impartially, and which can be justifiable to every member of the community on perspective-independent grounds (Stanley, 2017: 55, 94, 151–152). This objection does not merely take issue with the manipulative nature of dog-whistling; it also captures its effect on our ability to reason. Reasonableness is compromised because dog-whistling constitutes the sort of public speech which ‘“[divides] mankind into parties, [inflames] them with mutual animosity,” and erodes respect for certain targeted groups’ (Stanley, 2017: 94, 127).

The second argument – call this the policy mandate argument – appeals to our seemingly mundane democratic practices, the rituals of elections, and elaborates the implications which follow from the corrosive effect of dog-whistling on rational deliberation. As Goodin (2008: 200–203) notes, while the deliberative expectations of representative democracy might not be easily realisable, election campaigns should ‘display two deliberative virtues’: ‘openness’ and ‘concern for the common good’. Hence, political parties are vehicles for achieving ‘a shared understanding on the meaning of a vote’; their ‘role is to develop coherent policy arguments’, and ‘it is through them that the deliberations of the electorate are co-ordinated, so voters end-up talking about the same thing and coming to a shared understanding of what should be inferred from a vote’. The ‘perversity’ of dog-whistling, then, is not just that it ‘undermine[s] democratic deliberation’, but also parties’ raison d’être (Goodin, 2008: 6–7). The sending of contradictory messages to different groups entails that different voters reach different, incompatible understandings of what a politician or party stand for. While dog-whistling might not undermine parties ‘mandate to rule’, it does – by virtue of its affinity with manipulation – compromise the more substantive ‘right to implement a specific set of policies explicitly stipulated during the election and explicitly endorsed by the electorate at that election’ (Goodin and Saward, 2005: 472). In Oborne’s (2005: 120) words, politicians who manipulate voters ‘deprive them of the ability to come to a well-informed decision about how to cast their vote’ and ‘convert them into dupes’.

What respect? Which mandate? Democracy and conflict

Though appealing, the dissociation between democracy and dog-whistling posited by the respect and policy mandate arguments appears less plausible once the assumptions on which such arguments rest are confronted by the messiness of democratic politics. Both arguments are part of a long-standing tradition of democratic theory which, following the publication of Christopher Achen and Larry Bartels’ Democracy for Realists, has come under intense scrutiny. Conventional accounts of democracy, Achen and Bartels contend, conform to the ‘folk theory of democracy’, and efface the recognition that, rather than resembling a (serene) philosophy seminar, democratic politics comprises of a zealous, fragmented, biased multitude, incapable of any constant opinion or of converging to a shared understanding on matters of shared concern; they discount the insights of the ‘critical tradition’ in political science emphasising ‘the powerful tendency of people to form groups, the ensuing construction of “us” and “them” and the powerful role of emotion, rather than reason in directing group activity’ (Achen and Bartels, 2016: 1, 214–215, 299–306). Achen and Bartels’ controversial account has invited criticism, some of it warranted. For instance, their argument is sustained by mystifying dichotomies between ideology/identity and reason/emotion (Chambers, 2018); an oversimplification of the deliberative tradition, some theories of which need not be incompatible with their emphasis on group-identification (Chambers, 2018; Kolodny, 2017); and, a failure to acknowledge how well-designed participatory institutions historically had a crucial role in empowering democratic citizens (Frega, 2020). Though important, these critiques need not render Achen and Bartels’ core argument mute. I contend that the barebones of their critique of ‘the folk theory of democracy’, enriched with insights from critics of the canonical tradition of deliberative democracy (Johnson, 1998; Sanders, 1997), and scholars who paved the way for the philosophical tradition of value pluralism and realism in political thought (e.g. Berlin, 1990; Hampshire, 2000; Mouffe, 2005; Shklar, 1984; see Galston, 2010), cast doubt on the picture of democracy animating dismissals of dog-whistling. Offering a comprehensive review of political realism and deliberative democracy and exploring the relationship between realism in political science and political theory are – though fruitful enterprises – beyond my purposes. I wish to illustrate, however, that such approaches converge towards a rejection of the vision of consensus on which the respect and policy mandate arguments rest.

The language of shared understandings, of a rational consensus on the common good or certain substantive principles, and the corresponding demands that individuals should adopt ‘the standpoint of the impartial observer’ (Stanley, 2017: 94), and positively respect others and their views (Goodin, 2008: 96–121; Stanley, 2017: 186–188), circumvents the recognition that democratic societies, are not merely marked by diversity and difference, but also by ineradicable antagonisms and contempt between different groups which espouse different, incompatible principles and aspirations – groups which, though often appearing to be respectful for strategic reasons, contest one another at ‘a fundamental, even existential level’ (Johnson, 1998: 165; Sanders, 1997). As Judith Shklar (1984) notes:

we do not even see the same social scene before us. We do not agree on the facts or figures of social life, and we heartily dislike one another’s religious, sexual, intellectual, and political commitments – not to mention one another’s ethnic, racial, and class character. (p. 78)

It might be tempting to suggest that the prejudices, animosities and antagonisms which characterise pluralistic societies reveal the pervasive effects of dog-whistling or, more broadly, of manipulation in public life – the erosion of ‘rationality’, and ‘respect’ (Stanley, 2017). But the empirical literature does not suggest that such prejudices and animosities are ‘caused’ by dog-whistling (pace Stanley, 2017: 217), but, rather, that they are built in (Achen and Bartels, 2016; Brennan, 2017). This much follows from Mendelberg’s analysis on the effects of code words on public discourse to which Stanley appeals in support of his thesis but which highlights a problematic inversion of causal structures in his argument. Mendelberg’s (2001) findings, Stanley (2017: 320–321) concedes, do not illustrate that coded-language engenders racial resentment, but, rather, that it activates already-held attitudes of racial resentment.

The realist suggestion that conflict and contempt are part of the fabric of democratic cultures and, correspondingly, that ‘the concept of the common good rests on a cardinal mistake’ (Berlin, 1990: 43), is not a mere rehearsal of the non-ideal theory thesis that the vision of democracy animating the respect and policy mandate arguments, though difficult to achieve, is theoretically plausible (Hamlin and Stemplowska, 2012); or, that some of its constituent components might never obtain in certain contexts, and that ‘the most to which we can realistically aspire’ (Goodin, 2008: 202–203) is a distribution of deliberative desiderata across different stages of the political process. Though dog-whistling, as I argue, might be necessary in democratic politics, even if one endorses the vision of democracy stipulated by such arguments, the realist objection does not merely pertain to problems of ‘fact-sensitivity’, or ‘feasibility’, but rather with what should be seen as plausible in theory, even under the most ideal of circumstances. To proceed by a priori envisioning, the possibility of rational consensus on the common good or certain substantive principles (whatever these might be) is to propound a vision that is an innocent fairy-tale (Achen and Bartels, 2016; Hampshire, 2000).

The objection to visions of societal harmony under the aegis of reason, is not necessarily connected to the relationship between reason and emotion or citizens’ irrationality per se but cuts deeper – it pertains to an issue unacknowledged by Achen and Bartels (2016): a troubling, or, at least, speculative picture of the faculty of reason itself. The idea animating dismissals of dog-whistling is that reason has a convergent property: it enables those who possess and can exercise that capacity, through interpersonal or intrapersonal rumination, to naturally converge towards the common good or certain substantive principles and values. Yet, Stuart Hampshire notes, ‘we do not know anything about reason as a faculty, apart from what philosophers . . . have chosen to put into the concept. Parts of the soul, unlike arms and legs, are a philosophical invention’, and ‘philosophers have been free to construct models of the soul as they please’ and present these as ‘natural to serve their moral and political advocacy’. A breezy survey of the fabric of history and public life, not only reveals that there is nowhere evidence that the exercise of reason converges in that manner, but that exactly the opposite seems to be the case: ‘all determination is negation’ (Hampshire, 2000: 34–35). Groups have defined themselves – their substantive conception of the good – in oppositional terms: not merely in terms of who they are and what they espouse, but also in terms of who they are not and what they reject. For instance, ‘some of the ideals of monasticism were a rejection of the splendours and hierarchies of the Church, and this rejection was the original sense and purpose of the monastic ideal’; ‘some forms of fundamentalism’, ‘define themselves as a principled rejection of secular, liberal, and permissive moralities’ – as ‘the negation of any deviance in moral opinion’ (Hampshire, 2000: 34–35). The idea of ‘self-definition by opposition’ can be also gleaned in Katherine Walsh’s (2012: 517) analysis of ‘rural consciousness’: a specific way of life, marked by distinct values and principles, and ‘a social identification with rural residents’ formed by ‘a sense of injustice’ and ‘alienation’ – ‘a perception of deprivation relative to residents of metropolitan areas’ and the principled rejection of widely held policies, or attitudes of urbanity; or, perhaps, in Black Lives Matter (2020), the core ideals of which are formed as a rejection of racism. This is not a comprehensive list of examples, but it does capture the realist recognition that ‘every identity is relational’: it implies and affirms ‘the establishment of a difference’; that, the creation of a ‘we’ can only exist by its demarcation from a ‘them’ (Mouffe, 2005: 15–16). On this account, our capacity to find substantive meaning and worth in our lives, to pursue what we find meaningful and worthwhile, is conditioned, in part, by the rejection of what we find meaningless and worthless. Inasmuch as political practitioners are aligned with conflicting traditions, have different life stories, memories, and imaginations, we should not expect agreement on a common conception of the good or certain substantive principles. Conflict – within and between communities or groups – does not constitute ‘a pathology’ (pace Chambers, 2018), an indication that something has gone wrong, but should be seen as the ‘normal’ condition of mankind and should not be expected to cease in theory or in practice ⁴ (Berlin, 1980; Hampshire, 2000: 33–36).

If correct, this alternative picture of reason and, correspondingly, of a democratic polity in perpetual flux does not just cast doubt on the plausibility of the respect argument, but also of the policy mandate argument. While we should not expect election victors to cease to appeal to the notion of a mandate, elections ‘do not produce genuine policy mandates, even when they are landslides’ (Achen and Bartels, 2016: 302). Even if one discounted findings which illustrate that democratic citizens, irrespective of education or intelligence, are uninformed about politics, swayed by campaign slogans, susceptible to framing and cognitive biases (Achen and Bartels, 2016; see Chambers, 2018), the mandate argument would still ascribe to us psychological, and intellectual capacities we do not possess (Saul, 2018a). The primacy of group-identification – a point embraced even by Achen and Bartels’s critics (Chambers, 2018; Kolodny, 2017) – entails that ‘group memberships – being Protestant rather than Catholic . . . a white person rather than an African American or other minority – powerfully shape vote choices’ (Achen and Bartels, 2016: 222–229). Democratic citizens tend to support a party for reasons discarded as a dereliction of duty by the policy mandate argument: rather than transcending their ‘own positions’ and adopting a common ‘Archimedean point of view’ (Goodin, 2008: 83), they ‘gravitate toward leaders of their own stripe’ and base their support on their perceived ties between parties and their leaders with prominent social groups – on whether ‘“their kind” of person belongs to that party’ (Achen and Bartels, 2016: 308–309).

This is not to say that elections say nothing about citizens’ preferences, or to deny that politicians have a responsibility to their followers to increase the chances of achieving what they stand for and materialise some of their commitments (Gutmann and Thompson, 2012). Yet, the idea animating the policy mandate argument – that, elections reveal the “will” of the majority on a set of issues – is problematic. This conclusion would be difficult to resist, even if one were to recast a modified version of the mandate argument which grounds individual preferences on membership in social groups and identity (Kolodny, 2017), as opposed to an abstract vision of rational deliberation. The affinity between identity or social group and political leader or party aside, neither social groups, individual identity or party manifestos are, to use Hampshire’s (1993) words, ‘blocks of marble’ which neatly correspond to coherent ideological or political frameworks and which can be combined in a harmonious whole; rather, they are akin to ‘swarms of bees’, characterised by ‘contrary desires, conflicting beliefs’ (Hampshire, 1993), ‘friction between politically inconsistent identities’ (Achen and Bartels, 2016: 308–309). To suggest that a constricted menu of alternative party platforms, composed of an assemblage of jointly inconsistent principles, can express even the majority of the electorate’s opinions, themselves composed by a plethora of incompatible aspirations, is to disregard the realities of diversity and conflict. This much is conceded by Goodin (2008: 230–231): ‘politics-as-usual involves mixed messages. Ordinary campaign manifestos are a grab-bag combining a variety of diverse and disjoint programmes’, ‘with no single proposition . . . commanding the support of a majority of voters’.

Dog-whistling and democracy reconsidered

That the picture of democracy animating denunciations of dog-whistling is doubtful need not deny that dog-whistling is problematic; for, there exists an intricate relationship between dog-whistling, manipulation, and deception. Yet, this is not to say that dog-whistling should not be practised, or that it is incompatible with democratic politics in toto. As Bernard Williams (1978) reminds us, ‘it is a predictable and probable hazard of public life that there will be situations in which something morally disagreeable is clearly required’ (p. 62).

That dog-whistling constitutes an ineliminable feature of democratic politics rests on the recognition that, to achieve anything at all, democratic politicians need the cooperation of a plethora of others; unable to take their support for granted, subject to the frequent rituals of elections, they must continuously ‘mobilise the base’, ‘signal their fidelity to their tribe’, and court the support of potential coalition partners and the majority of the demos, through the difficult arts of persuasion and rhetoric, as opposed to open, brute coercion (Brennan, 2017: 46). But, as argued, neither one’s tribe nor the demos, are fixed, homogeneous entities; they are cultures of subcultures and contain a multiplicity of antagonistic groups. As Achen and Bartels (2016) note:

When political candidates court the support of groups, they are judged in part on whether they can ‘speak our language’. Small-business owners, union members, evangelical Christians, international corporations – each of these has a set of concerns and challenges, and a vocabulary for discussing them. Knowing those concerns, using that vocabulary, and making commitments to take them seriously is likely to be crucial for a politician to win their support. (p. 309)

Democratic politicians are required to manoeuvre dexterously amid a domain of conflict and contestation, by cultivating the support of various groups, whose interests, aspirations and worldviews conflict, and which are incompatible with one another (and, often, with those of the politician). Herein emerges the impetus to dog-whistling: securing even the basic goods of politics, requires politicians to ‘speak the language’ of different groups on whom they are dependent, to adapt their rhetoric to their audience’s prejudices and opinions, and to persuade them of their ex ante impossible loyalty, and trustworthiness, without inviting accusations of hollowness and hypocrisy, and without alienating those whose worldviews might compel them to reject such rhetoric or pledges of loyalty, if communicated more transparently. Despite the public’s obsession with truth-tellers, few campaigns would be successful if democratic politicians failed to appeal to a vision of the common good which, though fictitious and hollow, would serve to unite their diverse base (Waldron, 2011) and, instead, exposed the latter’s deep, insurmountable fragmentations by openly pandering to different, antagonistic groups; or, by earnestly conceding that expedient campaigning requires the creation of uneasy coalitions – the partial disappointment of some of their supporters’ worldviews, or, as per sections of the media and their opponents, the betrayal of some of their most cherished aspirations. ⁵

It is not a surprise then, that the efficaciousness of dog-whistling is demonstrated by numerous studies (Albertson, 2015; Hurwitz and Peffley, 2005; Mendelberg, 2001; Wetts and Willer, 2019), or, that even its most fervent critics concede that dog-whistling may deliver significant ‘electoral advantages’ to its practitioners by ‘increasing their share of votes’ (Goodin and Saward, 2005: 476). This is not to say that dog-whistling should be unconditionally practised or that it constitutes the only way in which democratic politicians can create coalitions, but, rather, that it forms an integral part of a larger arsenal of morally disagreeable, though often politically necessary, campaign tactics – for example, lying, hypocrisy, even negative advertising (Galston, 1991; Waldron, 2011). For dog-whistling functions in similar, though more economic, a manner to the well-remarked practice of ‘narrowcasting’ whereby politicians send different, incompatible messages and pledges of fidelity to opposing groups in different speeches, delivered in more restricted settings, and disseminated through selected media (Sunstein, 2001). This insight, again, is not restricted to verbal modes of communication or right-wing politicians who seek to appeal at the racially resentful without alienating more moderate supporters. Rather, as noted, it also encompasses a range of non-verbal communicative means, and applies to politicians of all stripes: those, for instance, for whom success in a campaign requires the backing of the religious and the secular, the republicans and the royalists, or the piecing together of the fragments of the centre-left – for example, pledges of fidelity to the plight of the historically excluded, without offending the sensibilities of those whose commitment to the ideal of neutrality might compel them to reject such appeals as excessively radical, or insufficiently liberal and just. ⁶ To borrow Williams’ (1978: 62) words then, a stiff-neck refusal to employ dog-whistling might entail that ‘one cannot pursue even’ some of ‘the moral ends of politics’.

This conclusion would be difficult to avoid even if one remained enamoured with the deliberative, liberal-egalitarian ideals animating denunciations of dog-whistling. The recognition that politicians operate in a non-ideal context of conflict and dependence entails that ‘openness’, ‘the normative ideals governing public speech’, and the aspiration to attain a policy mandate (Goodin, 2008; Stanley, 2017), might have to be suspended for the sake of attaining a mandate to rule. An obsessiveness with ideological purity and truthfulness, might, for defenders of liberal-egalitarian politics, come at the cost of failing to cultivate the support necessary to rise to power and realise some of their aspirations. Such an obsessiveness, which bespeaks of the morally admirable but politically inexpedient quality of ‘innocence’ shuns the demands of political life – the recognition that in running for office, politicians undertake ‘a responsibility to be as effective as possible in the pursuit of power’; and, that in ‘carrying out that responsibility’, they are obliged ‘to shape their tactics in relation to the world’ as it is, not as they might wish it to be (Galston, 1991: 186). Indeed, maximising one’s chances of electoral success – a sine qua non of such responsibility (Galston, 1991) – might not merely require the creation of uneasy, centre-left coalitions but also covert appeals to groups residing at the other end of the political spectrum. For instance, despite his long-standing popularity with black voters, Bill Clinton ‘won the presidency partly because he was too willing to campaign in dog-whistle terms’: his image as a ‘New Democrat’ – ‘one resistant on black concerns, “tough on crime,” and “hostile to welfare”’ – was crafted through surreptitious nudges or winks at white conservative, racially resentful voters (Haney-Lopez, 2014: 110–114). The paradox of dog-whistling then is that while it entails the betrayal of cherished liberal-egalitarian principles, a dogmatic refusal to employ it might jeopardise the hope of seeing the liberal-egalitarian vision realised.

This is not to suggest that there should be no limits to dog-whistling, but, rather, that it is impossible to determine a priori what these limits should be – for, ‘[i]t is difficult to exaggerate the width of the gap between the virtues of a political commentator and theorist on one side and the virtues of someone actually exercising power as ruler and leader on the other’ (Hampshire, 1989: 71). Nor need this deny the existence of an evaluative distinction between racist and more innocuous dog-whistles. Pursuing this argument, however, would require us, to use Shklar’s (1990: 9) terms, to shift our attention from conventional accounts of democracy, and the abstract visions of the summum bonum animating these, to an exploration of the concrete realities of the summum malum – the ‘full, complex, and enduring character of injustice as a social phenomenon’. Unlike innocuous dog-whistles, the enactment of which might entail the disappointment of one’s vision of justice, the realisation of some of the values animating racist dog-whistles entail the perpetuation of injustice, or failure to address pre-existing structures of oppression and the plight of the dispossessed. This issue is further compounded by racist nudges or ‘covert dog-whistles’. The recognition that such dog-whistles appeal, through rhetorical trickery, to racially prejudiced, but not necessarily racist, audiences – audiences who would have disapproved of more overt racial appeals (Saul, 2018a) – does not merely pile up manipulation on top of manipulation, but also renders such audiences involuntarily complicit in the perpetuation of injustice.

Yet, the affinities between the enactment of the values which racist dog-whistling implicitly communicates and injustice aside, one might wonder whether dog-whistles of that sort, if merely employed as a campaign technique or tool, are as problematic as is often thought – especially, if winning an election is impossible without nudging or winking at the racially prejudiced. ⁷ This is not just because such practices might enable mainstream parties to prevent voters from turning to the far-right, the sole purpose of which is to enact the values to which such dog-whistles appeal (Norris, 2005), but rather because they might constitute a ‘lesser evil’ vis-à-vis their alternative: a type of zealous candour, which communicates those attitudes and values in an unfiltered manner. This point is gestured en passant by no less a proponent of liberal democracy, Shklar (1984):

It is . . . no longer acceptable in the United States to make racist and anti-Semitic remarks in public . . . Southerners and Northerners alike are down to a few code words at election time. Would an egalitarian prefer more public frankness? . . . one might well argue that liberal democracy cannot afford public sincerity. Honesties that humiliate and a stiff-neck refusal to compromise would ruin democratic civility in a political society in which people have many serious differences of belief and interest. (p. 78)

Shklar’s point forms part of her defence of hypocrisy in democratic cultures and unearths a neglected affinity between hypocrisy’s ‘civilising force’ – construed in La Rochefoucauldian terms as ‘the tribute that vice pays to virtue’ – and racist dog-whistles on the one hand, and a juxtaposition between hypocrisy and dog-whistling and a single-minded pursuit of sincerity on the other. While this type of dog-whistling entails a pledge of loyalty to resentful attitudes, it also contains, within itself, a recognition of the limits and undesirability of these. Dog-whistles do not just constitute an act of (private) affirmation of the aforementioned views; they also constitute an act of (public) negation of these: the dog-whistler recognises that such pledges should only be covertly communicated, not publicly proclaimed and celebrated, and in so doing, expresses a pledge of overriding fidelity to the core ends which democratic politics should serve – a modicum of order, civility, and ‘the pretense . . . that social standings are a matter of indifference in our views of each other’ (Shklar, 1984: 77). Such fidelity to these political ends is not observable in politicians who “tell it like it is”, whose unflinching commitment to the dictates of conscience and opposition to “political correctness” compels them to employ overtly racist rhetoric, and thereby grant such views a public standing. This much does not just apply to fully unfiltered hateful rhetoric, but also to racial figleaves. ‘What is most worrying about figleaves’, Saul (2019) notes, ‘is their ability to make otherwise racist utterances seem acceptable, shifting our norms in such a way that increasingly explicit expressions of racial hatred become permissible’ (p. 21). Whether fig-leaved or fully overt then, the public pronouncement of “whites only” attitudes, references to “shithole countries”, “bad hombres”, and the association of minorities with bank robbing, rape, and other sorts of egregious behaviour (just to name a few manifestations of zealous candour), renders these attitudes a ‘visible fabric of society’ by openly celebrating and legitimising these (Waldron, 2012: 3). The glorification of such attitudes might thus prompt descent into a state of cruelty, and fear – some of the evils which democratic politics should act as a bulwark against (Hampshire, 2000; Shklar, 1984); for, it openly communicates to the targeted minorities that they should ‘be afraid’, and to those who espouse such beliefs of their public acceptability, encouraging them to express and enact these in public life (Maitra and McGowan, 2012; Waldron, 2012).

Conclusion: Post-truth politics and democracy’s deepening crisis

The ubiquity and vacuity of accusations of dog-whistling compels political theorists and scientists to consider more carefully what that phenomenon entails, and what, exactly, its relationship with democratic politics might be. In this article, I have attempted to contribute to that enterprise. Dog-whistling, I suggested, can be combined with a range of different worldviews, and might encompass a multiplicity of verbal and non-verbal communicative means – means which serve to nudge or wink at a specific group, and which are intended to remain inaudible to those who might reject more explicit appeals. In doing so, I challenged the prevalent thesis which posits a radical discontinuity between dog-whistling and democracy. Though we should not expect accusations of dog-whistling to wither away, dog-whistling might well constitute an inextricable aspect of our democratic rituals, and a ‘lesser evil’ vis-à-vis a politics of uncontaminated candour.

My argument does not merely shed new light on the way in which we should think about dog-whistling and democratic politics. It also poses problems for certain long-standing, recently more loudly rehearsed, suggestions that the rise of post-truth politics entails the erosion of democratic politics (Bunting et al., 2010; Kalpokas, 2018; McIntyre, 2018; Oborne, 2005). This is not to say that we should not be alarmed by politicians whose rhetoric expresses disdain for and collapses the distinction between fact and fiction. However, what follows from my critique of dismissals of dog-whistling is that our sense of crisis might be misplaced if our conception of post-truth politics entails ‘circumstances in which objective facts are less influential in shaping public opinion than appeals to emotion and personal belief’ (McIntyre, 2018: 5), and if our disillusionment is sustained by the romantic supposition that there should exist an identity between democracy, truthfulness and the realisation of the common good. If recent suggestions that we are witnessing a shift away from dog-whistling to figleaves are accurate (Saul, 2018b), my argument uncovers a neglected possibility: the premise of a post-truth politics misdiagnoses the nature of the crisis with which we are faced – the disquieting recognition that democracy might be under threat not by the loss of truthfulness, but by the spread of particular forms of it.