This response builds on some of the issues that Jason Tyndal raises in ‘Public Reason Liberalism and the Certification of Scientific Claims’ (2019b), itself a reply to my paper ‘What Facts Should be Treated as ‘Fixed’ in Public Justification?’ (Reid 2019) … [please read below the rest of the article].

Image credit: MedGreen Journalists via Flickr / Creative Commons

This response builds on some of the issues that Jason Tyndal raises in ‘Public Reason Liberalism and the Certification of Scientific Claims’ (2019b), itself a reply to my paper ‘What Facts Should be Treated as ‘Fixed’ in Public Justification?’ (Reid 2019). Because I do not disagree with much that Tyndal writes I won’t offer a point-by-point response or rebuttal; instead, I intend to use this piece to offer some preliminary thoughts on the role of factual claims in ‘convergence’ public reason, something I did not cover.

In the original paper, I discussed the possibility of there being facts that are considered beyond reasonable disagreement for the purposes of Rawlsian, or ‘consensus’, public reason. I called these ‘fixed facts’. Rawls suggests that such facts emerge within his model of public reason, but does not clearly delineate the set of fixed facts, or justify his suggestions for facts that would go into this set, at any great length. I argued that the implied set of fixed facts was too large, and risked dismissing some people as unreasonable for not accepting facts as ‘fixed’ when they did not have adequate reason to.

As Tyndal notes, I only touch on the other predominant model of public reason, the ‘convergence’ view, in passing. He is right to suggest that this is worth discussing in greater detail; the convergence view can handle some of the problems that arise when dealing with factual claims better than the consensus view, but is still afflicted by some of the problems I describe.

In this short piece I will first set out the original argument, briefly. I will then argue that there is logical space for there to be fixed facts in a convergence theory of public reason as well. Finally, I set out some reasons that convergence public reason theorists would probably resist committing to the view that this set of fixed facts is very big—or indeed populated at all. I briefly sketch some of the moves open to convergence public reason theorists to make this argument. I conclude by noting that they cannot, in spite of this, escape from the central problem that all public reason theorists face when they invoke epistemic standards in public justification: that for these standards to be of any real consequence requires some people to accept as ‘true’ facts that they do not believe to be so in a deep, metaphysical sense. They also, as Tyndal notes, cannot escape from the problem of “certification” (2019b): that most people’s assessment of facts is heavily reliant on second-order assessments of others’ testimony.

The Initial Problem

The consensus view of public reason argues that, on issues where the norms of public reason apply, citizens must justify themselves to others drawing on reasons that all other reasonable citizens can accept. In Rawls’ paradigmatic account of this model of reasoning, the reasonable citizens described are idealised individuals, who amongst other things, are committed to certain epistemic norms, which are contained within what he calls the “burdens of judgment”. These include a commitment to attempt to reach cooperative agreements with other reasonable people, and an acknowledgement of the difficulty that arises in assessing the empirical evidence on certain issues (Rawls 2005, 54-8). He also, briefly, notes that there are certain facts that should be treated as true during the process of public justification. These include “presently accepted general beliefs and forms of reasoning found in common sense, and the methods and conclusions of science where these are not controversial” (224).

Though Rawls makes this claim in passing, I argued that the implication of this view is that those who deny some of these ‘fixed facts’ do so because they are being unreasonable. In such cases real individuals who disagree with such a consensus are guilty of failing to adhere to epistemic norms that the idealised constituency we appeal to in public reason would. Arguments that deny fixed facts are somehow unreasonable, and therefore cannot underpin arguments for coercive laws that would have normative legitimacy.

I argued that this was problematic for a number of reasons, at least when we broaden the set of fixed facts to include things upon which there is a scientific consensus. My objections stemmed from the fact that many individuals will struggle to verify scientific claims, and thus the argument that they are unreasonable to deny such facts hinges on an assessment of the credentials of experts. I feel instinctively uncomfortable with this, because acting reasonably here seemed to require ceding significant epistemic authority. On the other hand, the denial of certain facts where there is overwhelming evidence to the contrary based on others’ expertise also seems disrespectful.

In the original article I tried to strike a balance; I argued that whilst it is possible to conceive of a situation where all reasonable people agreed to defer to an expert or set of experts on an issue that most struggled to fully understand, it seemed unlikely that such a consensus could be reached in such sweeping terms as Rawls described, for example deference to the consensus amongst scientists. Instead, it would be possible on a case-by-case basis, where all people had at least partial access to relevant information, and could critically interrogate the expert consensus.

The Convergence View

The convergence view of public reason is distinct from the consensus view, because it dispenses with the “public reasons only” requirement of the consensus view (Tyndal 2019b, 13). This means that, rather than suggest that citizens draw on a common set of reasons to justify their view to others, agreement for different reasons is sufficient. Notably, this brings religious reasons back on to the table as potentially justifying, because a law might be justified where person A accepts it for religious reasons, and person B accepts it for different, non-religious reasons.[1]

At first glance, it appears that a move to the convergence model would diffuse the issue of which facts to treat as fixed. Citizens who disagreed on the facts of a matter might still coincide, or not, on their views about adjacent policy areas. The convergence model of public reason does not depend on a shared factual base from which shared reason might be drawn, so a lack of fixed facts would not undercut potential agreement about laws.

However, most versions of the convergence view still set out some idealised constituency that we should appeal to in public justification, even if they eschew the term reasonableness. Gaus’s paradigmatic account uses the term ‘Members of the Public’ (MoP) to describe these individuals (2011, 26-7). Rather than the idealised constituency of reasonable people Rawls imagines, who do not relate to individuals, but simply know broad facts about society, Gaus’ MoPs are proxies for individuals (26). They are different in being ‘moderately idealised’ in a couple of ways: that they “deliberate well” (26), in the sense that they do not seek “bluff, bargain, or seek to engage in strategic behaviour” (276); and, more interestingly for the purposes of this piece, they respond appropriately to relevant reasons according to their own scheme of evaluative reasoning (26).

This “moderate” idealisation is less demanding than the standard suggested in the consensus view, in part because Gaus is particularly concerned to accommodate different levels of sophistication in moral and practical reasoning (Gaus 2011, 277-8). However, it retains some interpersonal epistemic standard, and convergence theorists are still concerned as to whether citizens are entitled to hold certain views given their own evaluative standards and metaphysical beliefs. In a separate discussion, Vallier suggests that in public reason, citizens should (and therefore, we can assume, MoPs will) still adhere to norms of “justificatory internalism”, and avoid “gross epistemic errors” within their own systems of reasoning (Vallier 2014, 104; 106).

It is this stipulation that opens up the conceptual space for the possibility for fixed facts. As I defined them, fixed facts were facts that all members of the idealised justificatory constituency agreed upon, but not all real individuals. Where a clause like this exists in convergence models, it means that individuals might be held to be making arguments incompatible with public reason because they are guilty of an epistemic failing that their idealised proxy would not be. Fixed facts could therefore emerge where there were factual claims that it would be impossible to deny absent some “gross epistemic error” of the kind MoPs do not make by definition.

The fact that there can be fixed facts according to the convergence view does not meant that the set of fixed facts will have anything in it. It is quite possible that the set of fixed facts by the convergence view will be empty, and indeed I think that the way that predominant versions of the view are constructed this is quite likely. It is for this reason that I downplayed the possibility of fixed facts in the convergence view in the original piece, and not because I thought it was logically impossible for fixed facts to emerge given the structure of the convergence view and the way I defined them.

There are good reasons, though, to think that the set of fixed facts will be empty according to this construction of the convergence view, rendering the point moot. Whilst it is possible to debate what constitutes a “gross epistemic error” in this context, and what the more “moderate” epistemic idealization implied by the convergence view entails—as Tyndal himself has done in greater detail elsewhere (2016)—this account of epistemic idealisation is thinner than that offered by the consensus view, and eschews things like the ‘burdens of judgment’.[2] As such, a greater degree of the profound disagreements that exist in everyday life on factual matters are retained in the idealised account of justification.

The fact that many convergence theorists are wary of over-idealisation does not rule out the possibility of defining the epistemic dispositions of MoPs in a way that is more idealised, yielding fixed facts. There are, I believe, additional features of the way that convergence public reason is structured which suggests to me that even a move to greater epistemic idealisation in how MoPs are defined is unlikely to yield a potentially large set of fixed facts in the same way the consensus view does.

As I have presented the convergence view, the types of epistemic failings that real people suffer from but MoPs do not are occasional egregious failings of coherence in reasoning or information acquisition within their own epistemic system. The epistemic errors MoPs do not commit are failures to maintain a minimal level of coherence between the beliefs they hold, facts available to them, and values they hold, and to consult the facts that their doxastic standards demand.[3]

Again, we might ask why all people reasoning coherently cannot all agree upon certain facts that others, reasoning incoherently, will not. The reason I think that convergence theorists (including Gaus and Vallier) this would not happen very often is that they assume people (and MoPs) exist within fairly discrete epistemic systems where these are “sets of norms, standards, or principles for forming beliefs and other doxastic states” (Goldman 2010, 187). Different citizens exist in different epistemic systems in part as a result of the varying metaphysical beliefs that they hold (Jønch-Clausen and Kappel 2015, 377-9). Adapting this architecture to the question of public reason the convergence view, as I read it, emphasises the incompatibility and variety of different people’ epistemic systems, at least compared to the assumptions underpinning the consensus view. If we assume that different citizens exist within different epistemic systems, none of which are uniquely reasonable or unreasonable, then it is quite possible for citizens to reach deeply divergent conclusions on simple factual questions without drawing incorrect inferences based on their own doxastic standards and systems of evaluative reasoning. They could therefore disagree on most factual matters of consequence without committing the kind of error that convergence theorists see as disqualifying.

To use a concrete example, religious believers might have good first-person reasons to treat the epistemic authority of scientists as less robust, because their metaphysical beliefs lead them to put less credence in the scientific method. Doing so does not entail a gross epistemic error of the kind Vallier, for example, envisages, but is incompatible with the burdens of judgment as Rawls sets them out. This lesser view of the credence of scientific judgment as a result of metaphysical beliefs is a conclusion that an MoP might reach, but a reasonable person (by Rawls’ definition) will not. As a result, the possibility of fixed facts emerging via the scientific method is diminished in the convergence view, because there is greater disagreement about the merits of scientific inquiry as an enterprise.[4]

The relevant difference is that advocates of the consensus view expect individuals to exist within two epistemic systems simultaneously: their personal one, based in part on their metaphysical commitments and own levels of evaluative reasoning, and the one they draw on in public justification, that ignores any controversial metaphysical commitments and assumes a certain level of reasoning on their part. In the case of religious believers above, consensus theorists expect citizens to recognise interpersonal reasons to respect expertise even where they have first-person reasons not to. The convergence view is set out in part to avoid forcing individuals to reason in such a way.

Concluding Remarks

If it is true that the set of fixed facts is empty by an orthodox reading of the convergence view, does this mean that the problems I raised when I introduced the concept to the debate are dissolved? I do not think so. First, the certification problem remains either way. Second, if convergence public reason theorists want to appeal to epistemic standards in defining MoPs, they still mediate the same fundamental tension as consensus theorists: between applying epistemic standards that are so weak as to provide little actual guidance; and laying out thicker epistemic standards that end up producing a large set of fixed facts, or require that they accept claims that most intuitively feel are subject to reasonable disagreement.

Whilst the absence of fixed facts is less of a concern for convergence theorists, because their theory is less reliant on shared bases of understanding or prescriptive forms of deliberation, it does raise the question of whether the epistemic standard they set is too lenient to do the work they want it to. A useful avenue for future inquiry on this topic might be to use ‘fixed fact’ framework as a way of clarifying discussions around about epistemic idealisation in convergence models, and the ways different convergence theorists mediate the tension I outlined in the previous paragraph.

Contact details: Andrew Reid, University of Birmingham, A.Reid@bham.ac.uk

References

Gaus, Gerald. 2011. The Order of Public Reason. New York: Cambridge University Press.

Goldman, Alvin. 2010. “Epistemic Relativism and Reasonable Disagreement.” In Disagreement edited by Richard Feldman and Ted A. Warfield, 187-215. Oxford: Oxford University Press.

Jønch-Clausen, Karin, and Klemens Kappel. 2015. “Social Epistemic Liberalism and the Problem of Deep Epistemic Disagreements.”’ Ethical Theory and Moral Practice 18 (2): 371-384.

Quong, Jonathan. 2018. “Public Reason.” In The Stanford Encyclopedia of Philosophy edited by Edward N. Zalta. https://plato.stanford.edu/entries/public-reason. Accessed 18 February 2020.

Rawls, John. 2005. Political Liberalism (Expanded Edition). New York: Colombia University Press.

Reid, Andrew. 2019. “What Facts Should be Treated as ‘Fixed’ in Public Justification?” Social Epistemology 33 (6): 491-502.

Tyndal, Jason. 2016. “Moderate Idealization and Information Acquisition Responsibilities.” Res Publica 22: 445–462.

Tyndal, Jason. 2019a. “Public Reason, Non-Public Reasons, and the Accessibility Requirement.” Canadian Journal of Philosophy 49 (8): 1062-1082.

Tyndal, Jason. 2019b. “Public Reason Liberalism and the Certification of Scientific Claims.” Social Epistemology Review and Reply Collective 8 (11): 8-14.

Vallier, Kevin. 2014.  Liberal Politics and Public Faith: Beyond Separation. New York: Routledge.

[1] For a summary of the distinction see Quong (2018, Section 5).

[2] There are various reasons that convergence theorists are wary of over-idealisation, some of which Tyndal touches on (Tyndal 2019b, 13).

[3] An argument based on the contentious claim that standards of inference are agent-neutral, but standards of evidence are agent-relative (Tyndal 2019a).

[4] This is the case whether you accept my more cautious view about the number of scientific fixed facts in the consensus view, or Rawls’ more expansionist view that fixed facts emerge where there is a consensus of scientific experts.

Categories: Critical Replies

Tags: Andrew Reid, convergence view of public reasoning, Jason Tyndal, John Rawls, liberalism, public justification, public reason, reasonable disagreement, scientific claims