From Fact-checking to Value-checking: Normative Reasoning in the New Public Sphere

This article suggests that fact checking is a useful but incomplete framework for delivering an epistemically healthy public sphere. Through a brief history of the fact/value distinction, it is argued that there is no secure justi ﬁ cation for limiting interventions aimed at improving the emergent digital public sphere only to factual claims. On this basis, the heuristic principle of ‘ value checking ’ is outlined, as a complement to fact checking in the epistemic regulation of democratic discourse. Value checking would accept that more sophisticated and deliberative communication is a vital requirement for a well-functioning public sphere, and that this can be promoted through new forms of epistemic regulation. However, it would reject the notion that fact checking is suf ﬁ cient to achieve this, suggesting that the promotion of healthy political communication should also extend to value-based reasoning. The principle of value checking could be added to the fact-checking paradigm as a means of further enriching the public sphere in the ‘ post-truth ’ age.


Introduction
THE DEBATE OVER political debate is in a muddle. Across the political spectrum we see acute concern over the declining quality of political discourse and its implications for the epistemic health of democracy. 1 There is no consensus over the legitimate scope or nature of regulatory interventions that might improve this situation, or even who the proper actors to perform these new regulatory functions might be. 2 The ad hoc nature of the permanent suspension of Donald Trump from Twitter in January 2021 has compounded this sense of confusion over the principles that should regulate the digital public sphere, and the extent of their application. There is, however, broad agreement on a key objective of these new epistemic interventions: they should seek to establish 'the facts'.
As explored in a recent special issue of The Political Quarterly, the fact-checking paradigm has emerged as a dominant response in media ethics and regulation to fears over fake news, the dissemination of misinformation and the broader rise of post-truth politics. 3 Before his permanent suspension, Twitter dramatically employed fact-checking principles during the 2020 US presidential election campaign to flag Trump's claims about electoral fraud and to direct users to official sources of information. Private third party fact-checking organisations have been championed by Facebook in response to demands by regulators and legislators that they take responsibility for political content circulated on their platforms. The growing use of fact-checking firms to flag and analyse content was highlighted by Mark Zuckerberg in his 2019 appearances before Congress as evidence that tougher external 1 regulation of its practices was unnecessary. Fact checking also features increasingly prominently in media coverage of political speeches and parliamentary debates. In the 2019 UK general election campaign, for instance, ITV News employed FullFact.org to analyse statements made by Boris Johnson and Jeremy Corbyn in their televised leadership debate, with the aim of better informing voters about the validity of opposing claims. Indeed, 'reality check' (a fact-checking service) now constitutes an entire section of the BBC's news website. 4 While other goals (notably preventing hate speech and the incitement of violence) are also central to this nascent regime of epistemic regulation, fact checking is thus firmly established as a dominant principle for improving the quality of political debate and deliberation. We can distinguish many different forms and roles of fact checking-based for instance, on whether it is actively used to remove content, flag content, or offer supporting contextual analysis of the evidence used within that content. 5 But all of these forms rest implicitly on the 'fact'/'value' distinction, and on a somewhat narrow account of the requirements of a healthy public sphere. In particular, in the regulatory model we are moving towards, the factual 'truth' content of online debate is understood as fair game for epistemic interventions intended to promote a more informed democratic debate. 6 Crucially, however, normative reasoning (reasoning about 'values') is left out of this framework. Facts are seen as objective (if complex, multifaceted) features of 'the world out there', and thus can be established and interpreted by independent actors. 7 Values, on the other hand, are understood as a matter of personal opinion, sentiment or ideology, and should be worked out in the free marketplace of political ideas without similar supporting interventions.
Here, I suggest that fact checking is a useful but incomplete framework for delivering an epistemically healthy public sphere. Through a brief history of the fact/value distinction, it is argued that there is no secure justification for limiting interventions aimed at improving the emergent digital public sphere only to factual claims. On this basis, the heuristic principle of 'value checking' is outlined as a complement to fact checking in the epistemic regulation of political discourse. Value checking would accept that more sophisticated and deliberative communication is a vital requirement for a functioning public sphere, and that this goal can be usefully promoted through new forms of epistemic regulation and scaffolding. However, it would reject the notion that fact checking is sufficient for this, suggesting that the improvement of political debate should also extend to value-based reasoning. The principle of value checking could be added to the fact-checking paradigm as an additional means of enriching political discourse, so that it more closely approximates the ideal of a functioning public sphere.

Limitations of the fact-checking paradigm
Fact checking has developed as a largely ad hoc response to the problems of post-truth politics and digital media. While it occupies an increasingly central place in our political culture, it has not been underpinned by a sophisticated account of the epistemic conditions for a healthy public sphere and the proper exercise of democratic citizenship. Moreover, once we start to try and flesh out what this might look like, we find there appear to be no secure grounds for cordoning off the purely 'factual' content of political claims from their normative or ethical content, or for accepting the former as a legitimate object for epistemic regulation and public education, but not the latter. 4 BBC News, Reality Check, various dates; https:// www.bbc.co.uk/news/reality_check (accessed 21 April 2021). 5 L. Graves and A. Mantarlis, 'Amid political spin and online misinformation, fact checking adapts ', The Political Quarterly, vol. 91, no. 3, 2020, pp. 585-591. 6 As Full Fact describes its mission: 'Full Fact fights for the right information to reach the people who need it most … As fact checkers we've seen firsthand how bad information promotes hate, damages people's health, and hurts democracy'; https:// fullfact.org/about/ (accessed 21 April 2021).

Regulating the digital public sphere
The rationale for any sort of interventions to regulate, oversee or support the nature of online political debate has its basis in theories of democratic citizenship. In particular, regulation to improve the nature of online debate has often been understood through the lens of Jürgen Habermas's notion of the public sphere. 8 According to Habermas, while exclusive and imperfect, the bourgeois public sphere of the Enlightenment and the ideal of rational public deliberation it was based onborn in the salons, coffee shops and the early print press-nevertheless contained an emancipatory kernel. It pointed the way towards a potentially radical and egalitarian basis for political authority, after the collapse of feudal absolutism and divine right. The goal was to universalise, expand access to and improve the functioning of this public sphere so that its participatory potential could be realised and it could approximate the standards of rational interpersonal communication.
However, according to Habermas, the fragile ideal of the public sphere opened up by the Enlightenment had been gradually trivialised and closed through the concentration, commodification and commercialisation of the mass media in the nineteenth and twentieth century, so that the press became a source of sensationalism, hysteria and propaganda. 9 Rather than fulfilling the original promise of the public sphere as a participatory space for egalitarian deliberation, the media became subsumed under the totalising systems of control of Fordist capitalism. Understood through this lens, the radical ease of access and expression that digital platforms such as Twitter represent offers the potential to overcome some of the limitations of the commercialised, traditional mass media that dominated late twentieth century political communication. This utopian vision underpinned optimistic early assessments of the potential of online spaces to improve democratic participation and discourse.
However, this is clearly not the reality that we see on online platforms today. In fact, the digital sphere has fallen short of the Habersmasian ideal of the public sphere in much the same way as other mass media before it. 10 Owing to its own creeping commercialisation and concentration (as key platforms exploit network effects to colonise an ever greater share of online communication), uneven power relations that determine access to and control of information and data, and the increasingly sophisticated uses of that data to target and fragment political messaging, the digital sphere has also been closed and privatised in much the same way. 11 Add to that the constraints put upon political expression by the format of platforms such as Twitter, which encourage sensationalised and simplified statements, and also the broader rise of valence politics which extols opinions and identity over technical policy detail, and we have the general sense of a public sphere in crisis once again.
These basic arguments provide justification for new forms of non-coercive epistemic regulation and scaffolding to improve the quality of digital debate and deliberation, and to correct for some of these tendencies in contemporary political communication. 12 They provide a basis for rejecting libertarian accounts that any such interventions represent an illegitimate or undemocratic intrusion into 'free speech'-because the practice of democracy is understood as having important epistemic pre-conditions. However, they do not tell us much about what these regulations should do, or their proper scope. 8 S. Chambers, 'Truth, deliberative democracy, and the virtues of accuracy: is fake news destroying the public sphere? ', Political Studies, vol. 69, no. 1, 2021, pp. 147-163 Beyond the fact/value distinction With the rise of the fact-checking paradigm, this gap has been (implicitly) filled through a problematic account grounded in the fact/ value distinction, with its origins in a positivist understanding of legitimate authoritative knowledge. The concept of fact checking frames the problems of post-truth politics as reducible mainly to a lack of information ('facts'), leading to sub-optimally rational decision making on the part of publics and electorates.
The fact/value distinction can be traced back to the philosophy of David Hume in the eighteenth century. 13 Hume famously distinguished propositions about what 'is and is not' from propositions about 'ought and ought not'. G. E Moore later developed Hume's ideas into a critique of the 'naturalistic fallacy', the belief that you cannot infer an 'is' from an 'ought'. 14 With this line of thinking, we have the development of the notion that there are two fundamentally different types of human knowledge: 1) positive, factual knowledge of the type obtained through empirical scientific enquiry, observation and experiment; and 2) normative claims about ethical and moral principles or the good life. Moreover, normative knowledge is its own distinct form of enquiry, that bears no relation to the facts about the physical world established by empirical science.
Crucially, nothing in Hume or Moore suggests that normative knowledge is somehow illegitimate, or not the proper stuff of rational political analysis or democratic deliberation. Nor was there the idea that we cannot come to any conclusions or judgements about normative issues, and that they are simply matters of personal opinion. Indeed, the analytical tradition of normative enquiry Moore represents was dedicated precisely to trying to think through and securely ground the normative principles that should guide social and political life. Moore even subscribed to the idea of moral realism, which suggested that normative and ethical propositions represent objective features of the world, no less 'real' or true than empirical facts.
It was, rather, from logical positivism that we inherit the general contempt for normative forms of reasoning that continues to saturate our political and intellectual life, and that implicitly sets the limits of the contemporary fact-checking agenda. According to the logical positivist philosophers of the interwar years, there is not just a clear distinction between positive and normative knowledge. There is also a clear hierarchy. For the logical positivists there were only two valid ways of disproving a statement: firstly, empirical observations of 'the world out there'; secondly, mathematical reasoning and sound logical inference. This again had its origins in Hume, with his statement that: If we take in our hand any volume; of divinity or school metaphysics, for instance; let us ask, Does it contain any abstract reasoning concerning quantity or number? No. Does it contain any experimental reasoning concerning matter of fact and existence? No. Commit it then to the flames: for it can contain nothing but sophistry and illusion. 15 While this does not in principle exclude normative reasoning, in practice it led to the valorisation of mathematical and scientific enquiry, and the general assumption that all other forms of ethical, social and political deliberation were meaningless metaphysics.
This positivist suspicion toward normative reasoning was perhaps exemplified by the developments within economics over the twentieth century. 16 Economists increasingly argued that their discipline was a strictly positive and value-free science, which according to Milton Friedman was famously 'independent of any particular ethical position or normative judgements'. 17 It was this, moreover, that grounded its unique authority and legitimacy within political debate and the public sphere.
Over the course of the twentieth century, there have been innumerable critiques both of the privileging of positivist forms of expertise over normative reasoning, and also of the conceptual robustness of the fact/value distinction itself. 18 Turning to the first, the whole tradition of analytical political theory since at least John Rawls, represents a sustained attempt to logically clarify, distinguish and develop alternative normative principles and to apply them to the problems of political society. While, like any scientific discipline, consensus or a single orthodoxy has not emerged from these attempts, they are just as much a part of the body of knowledge and expertise that is relevant to considering and appraising the different statements made by different actors within the public sphere as the judgements of more positive forms of expertise.
Regarding the latter, numerous philosophical projects have called into question the very existence of the fact/value distinction. These include (inter alia) logicians such as William Quine and Hilary Putnam, philosophers of science such as Thomas Kuhn, Paul Feyerabend and Norwood Hanson, and pragmatist theories of language such as the late Wittgenstein and J. L. Austin. 19 These various schools of thought have highlighted the theoryladenness of all empirical observation, the social embeddedness of truth claims and knowledge production, and the implicit normativity contained within the structuring assumptions of many apparent value-free forms of enquiry (for instance, the normative conception of rationality that underpins neoclassical economic analysis). They have also shifted our understanding of truth from a correspondence theory (statements are true when they correspond to the objective world out there) to a coherence theory (statements are true to the extent that they internally cohere within a socially established set of norms for validating them). All of them suggest that any attempt to draw a clear line around 'facts' and to single them out as the sole site for remedying the ills of contemporary political discourse is precarious.

From fact checking to value checking
If we acknowledge this, limiting the regulation and appraisal of statements made within the emerging digital public sphere to 'fact checking' appears problematic. It means that, by itself, it cannot bear the burden our society is placing on it for improving the quality of political discourse. We can distinguish weaker and stronger versions of this claim. Firstly, even if we do accept the fact/value distinction, but reject the logical positivist dismissal of normative knowledge as pseudoscience, then there is no basis for restricting the critical appraisal of political statements and discourse to factual/ empirical knowledge. Secondly, if we take seriously the philosophical criticisms of the fact/value distinction, then improving political discourse also requires promoting a richer understanding of the values and assumptions that underpin even supposedly 'factual' or value-free claims and expertise.

Value checking: strong and weak versions
Turning to the first, weaker version of this claim, fact checking currently focusses only on clarifying and contextualising the empirical content of political claims. For instance, in response to a political claim about poverty or inequality in the UK, a fact-checking organisation might analyse the different methods for measuring poverty/inequality and its evolution over time. They might assess how these are used selectively or cherry-picked by public figures to tell different stories about how these have evolved, and thereby paint an incomplete or uncontexualised picture of the impact certain policies or governments have had upon this issue. This has been especially prominent in UK politics over the last decade, with the controversial change to the way in which child poverty was measured by the coalition government under then Work and Pensions Secretary, Iain Duncan Smith, and the more general debate between 18 P. Gorski, 'Beyond the fact/value distinction: ethical naturalism and the social sciences ', Sociology, vol. 50, 2013, pp. 543-553. 19 H. Putnam, 'Beyond the fact/value dichotomy', in A. Tauber, ed., Science and the Quest for Reality, New York, New York University Press, 1997, pp. 363-369.
absolute versus relative measures of income in the measurement of poverty.
However, under their current mandate, based around establishing only the 'facts', fact checkers do not similarly act to contextualise and clarify the distinctive normative content of political claims. This could include, for instance, the theories of distributive justice that they implicitly support and objections which could be raised against them. This risks giving the misleading impression (to a viewer or user) that, if they have understood whether a political figure has got the 'facts' right, their broader claim is necessarily sound and valid. Taking the above example once again, claims about inequality levels or poverty generally rest on some (often unstated) principle of distributional justice. The Rawlsian difference principle-which states that rising levels of inequality are just and morally acceptable, so long as the worst off are getting materially better-might be one. This might be contrasted with 'sufficientarian' accounts, which set a certain absolute minimum threshold standard of living beneath which no member of a given political society should be allowed to fall, egalitarian accounts of distribution fairness of the type developed by G. A. Cohen, or various other prominent normative positions on distributional equity. Interpreting these positions might be just as important to understanding and assessing whether a political claim is valid, or valid in what sense, as accurately interpreting the empirical facts about poverty and inequality.
Turning to the second, stronger critique of the limitations of fact checking: by focussing only on establishing the 'facts' and on their independence and 'objectivity', they tend to operationalise a hard fact/value distinction which, in the process, tends to validate certain forms of expertise as 'value-free'. This risks under-playing the political and normative assumptions embedded in scientific and social-scientific disciplines and theoretical traditions. As an example, mainstream macroeconomics is generally appealed to as the appropriate body of expertise to assess political claims made about 'the economy', usually without unpacking the normatively contestable assumptions this tradition of thought is founded upon. As an example, the Lucas critique that now underpins all mainstream macroeconomic modelling and forecasting (often held up as a source of value-free facts, for instance in debates over the economic impact of Brexit or the merits of austerity policies), has been shown to rest upon theoretical assumptions (an equilibrium account of the business cycle) which assume in advance a limited role for government intervention. 20 In other words, these models have political and normative assumptions built into them a priori (they are not a 'finding' of the model), and cannot be appealed to as a source of factual information without this context. Consequently, the sorts of political positions and claims that are bolstered by these forms of expertise and analysis are invested with a falsely secure authority.
Relatedly, current fact-checking practices tend implicitly to validate certain forms of evidence over others. For instance, they reproduce a bias for quantitative data over qualitative data, even though the epistemic justifications for this preference may be shaky when applied to many political questions or claims. It is very rare to see fact checkers scrutinise claims that could be proven or disproven only by qualitative data (for example, 'I think the people of this country have had enough of experts'). This means, for one thing, that our political discourse is not being enriched by the wealth of qualitative research that could enhance our understanding and appraisal of political claims.
Fact checking also risks giving the impression that the proper standards of evidence for establishing the validity of a claim are internal to the claim itself. As such, if audiences only consult fact checkers in seeking to inform their political opinions and choices, they will have little basis on which to better understand or contest the background assumptions and 'problem frames' that political rhetoric mobilises. Consider a fact check of a BBC phone-in during the 2019 general election campaign, during which Boris Johnson responded to viewers' questions on the Conservative government's record and policy stances. 21 It included an analysis of childcare policy. But, in keeping with the general principles of fact checking, it limited itself narrowly to unpacking some further details about the government's scheme to subsidise paid childcare and its coverage. Thus, it did not interrogate the background assumptions that frame this way of analysing the problem: for instance, that childcare must be provided as a commodity, and that it cannot be addressed through more fundamental labour market reforms such as job sharing, working time regulations, or basic income provision. This is perfectly in keeping with fact checkers' current interpretation of their mandate and their hesitancy to conduct any analysis which seen as 'biased' or 'value laden'. But, as an aid to a voter trying to understand the full range of issues that are at stake in social policy on childcare and social reproduction, and the full range of possible viable positions that can be taken on these issues, it is of limited use.

Value checking as a supplement to fact checking
How might we extend and supplement the fact-checking paradigm so that it can address some of these limitations? One way would be to add another form of epistemic regulation for the new public sphere, which we might call value checking. Value checking could run alongside political news coverage and social media content, just as fact checking does already. However, it would have a distinctive mandate and function that would focus on correcting the three problems identified above.
First, it could seek to improve public understanding of the normative content underpinning political adverts and claims that we are exposed to on a daily basis (as well as simply their factual validity). This could include brief and easy-to-understand summaries of the (explicit or implicit) normative positions that politicians' rhetoric draws upon and mobilises. It could also include assessment of how coherent these are with existing positions in the literature and contrast them with viable alternative perspectives found in normative theory. Second, it could seek to identify the assumptions that different bodies or expertise are founded on, and the potential political and normative baggage of these-providing users, for instance, with links to alternative research founded on different epistemological commitments and starting premises. It would also seek to identify clearly these assumptions and premises so as to make them more transparent. Third, it could provide supporting analysis that briefly unpacks the assumptions embedded in certain ways of framing a problem, helping citizens to grasp how the universe of viable 'solutions' to a political issue is often constrained by the way in which that problem is framed, and the alternative problem frames that exist on the issue under discussion.
This sort of a mandate and mission would be quite different from the current framing of fact checkers, but it could provide an important supplement to the principles and practices they have pioneered. In other words, value checking would provide an additional broad principle that could sit alongside fact checking, aiding its mission to improve the epistemic functioning of the public sphere. Along with fact checking, it accepts that supporting epistemic scaffolding of public debate and discourse can assist in the task of building better forms of deliberation and more 'rational' (in the Habermasian sense) public discourse and communication. But, reflecting dominant trends in philosophy and epistemology over the past century, it would reject the fact/value distinction on which the fact-checking paradigm has been constructed and the problematic baggage that accompanies it.

Summary
Faulty ethical or normative reasoning damages the quality of democratic debate just as much as faulty information or 'facts'. Moreover, currently dominant schools of thought in philosophy and epistemology generally reject the notion of a clear distinction between the two. There is, therefore, no prima facie reason for justifying interventions into public discourse that improve the latter, while taking no action to improve the former. Consequently, while the rise of fact checking has been an important step in the right direction towards a richer democratic discourse, it seems imperative to supplement fact checking with value checking. This form of epistemic scaffolding would explicitly acknowledge normative reasoning as equally essential as empirical facts to the ability to scrutinise fully, understand and respond to political discourse-and thus