Introduction

Who gets to decide what it means to live a healthy lifestyle, and how important a healthy lifestyle is to a good life? As more governments make preventing obesity and diet-related illness a priority, it has become more important to consider the ethics and acceptability of their efforts. When it comes to laws and policies that promote healthy eating—such as special taxes on sugary drinks and programs to encourage consumption of fruits and vegetables—critics argue that these policies are paternalistic, and that they limit individual autonomy over food choices.

In our book Healthy Eating Policy and Political Philosophy: A Public Reason Approach (Barnhill and Bonotti 2022), we argue that both paternalistic justifications for healthy eating efforts and anti-paternalistic arguments against them can be grounded in perfectionist views that overly prioritize some values over others. We therefore propose a more inclusive, public reason approach to healthy eating policy that will be appealing to those who take pluralism and cultural diversity seriously, by providing a framework through which different kinds of values, including but not limited to autonomy and health, can be factored into the public justification for healthy eating efforts. Additionally, the book adopts a ‘farm to fork’ approach to the ethics of healthy eating efforts: it engages with theories and debates in political philosophy, considers the implications of different theoretical positions for healthy eating efforts, and then develops a concrete tool for assessing policies that will be of interest to both scholars and policymakers. As well as offering a novel normative analysis of healthy eating policy, we offer a new theoretical framework that will be applicable to a wide range of public policy scenarios.

We are extremely grateful to the contributors to this symposium for their thoughtful commentaries on our book. In this article, we provide a critical reflection on the issues they raise with regard to some of the key aspects of our analysis.

Public Reason, Deliberation, and the Role of the Moderator

In his astute commentary, Federico Zuolo points out that the deliberative process that we sketch out based on our public reason framework does not have a moderator, yet one is needed. We are certainly not opposed to including a moderator (or facilitator) in the deliberative process that we propose. In fact, their presence can help realize the very goals of deliberation and public reason by helping participants in the deliberative process to think and relate to each other in a fair-minded way. However, not all public reasoning occurs in the form of actual deliberation among different people. In many cases, due to resource and time constraints, public health officials must make policy decisions on their own. In such cases, the figure of the moderator should be replaced by procedures and guidelines that can help the public official to make fair-minded decisions.

Zuolo further asks what the role of deliberation is in our account, i.e. whether it is just ‘a procedure to filter out what is publicly non-acceptable’ or whether, more in the spirit of deliberative democracy, it consists of an ‘open-ended…procedure…[and] an epistemic practice of looking for the best solution’ (8). Deliberation serves both roles in our account. On the one hand, it is meant to ensure that a healthy eating policy meets the criteria of public reason, e.g. that it is not grounded in sectarian values and interests, and that it advances the public good in some way. On the other hand, it is also a process that can help improve policymaking, e.g. by helping policymakers to acquire a better understanding of the actual or likely effects of policies, and design policies that minimize negative side effects, especially for members of already vulnerable groups. In fact, the two goals are related and can both be traced back to public reason since the latter requires that as well as excluding sectarian reasons from public justification, policies are also justified based on sound empirical evidence and do not impose excessive ‘strains of commitment’ (Rawls 1999a, 153) on citizens. The ‘strains of commitment’ can be understood in the following way:

I fail to reason impartially if I support a policy whose burdens and benefits are distributed in such a way that I wouldn’t agree to place myself in the position of those who are worst-off under the policy (Quong 2006, 60).

Relatedly, Zuolo also asks what the role of a moderator would be in our framework, should we be willing to include one (which, as we explained earlier, we certainly are). More specifically, he asks whether besides facilitating deliberation and ensuring that the deliberative process is fair, and that all participants have an equal opportunity to speak, the moderator’s role may also be to police the conversation and ensure that the latter remains within the boundaries of public reason. Should that be the case, Zuolo argues, this would be a concern since

this solution would come at the cost of transforming the possible activity of facilitation and improvement of the discussion into a policing of the discussion. If so, public reason, instead of being a set of principles that could inspire fair dialogic relations, would become a set of rules to be enforced accompanied by sanctions for any lack of compliance. I assume that this would not be a desirable option for Barnhill and Bonotti (9).

We do not believe that the moderator’s role, in our proposed deliberative forums, would be to police the discussion in order to keep it within the bounds of public reason exactly—it would be fine if the participants in the deliberation also discussed the relevant policies in other terms. And nor do we advocate any kinds of sanctions that the moderator (or the forum organizer) could apply should any participants fail to comply with public reason norms. However, the deliberative forum would have to perform a specific analytic task, which is to assess whether the relevant policy satisfies the criteria of public reason. Part of the role of the moderator would be to ensure that the group performs this task as adequately as possible given any existing constraints (e.g. time). In this sense, our approach aligns with Rawls’s ‘wide view’ of public reason, according to which non-public reasons can be introduced in deliberation as long as policies are ultimately justified based on public reasons (2005, 453). Furthermore, it is also plausible to argue—as we do in our book (198)—that although public reason offers the criteria based on which healthy eating policies should be evaluated, those policies could be even more justified—i.e. ‘fully’ justified rather than just ‘pro tanto’ justified (Rawls 2005, 386–7)—when they are also supported by non-public reasons. Moreover, an ‘overlapping consensus’ of comprehensive doctrines, in which citizens with different worldviews are able to endorse the same policies from their diverse perspectives, can also provide more stability in diverse societies (Rawls 2005, 38ff).

Zuolo also queries the inclusion of a ‘fair-minded group of people’ in our framework, pointing out that it is not clear what function this group is serving. On the one hand, he argues, the ‘fair-minded group of people’ could be considered as ‘the good will of all the participants, namely as the capacity to obtain the best outcome of the deliberation on a certain issue, given the available evidence. In this sense, the fair-minded group of people is a normative figure, namely the personification of truth or reasonability in the debate’ (6). On the other hand, it could be viewed ‘as the functional equivalent of a moderator because its role is particularly invoked when there is a disagreement’ (6).

In response to this query, we would like to stress, first, that in our account the ‘fair-minded group of people’ only appears as a heuristic device in one of the questions in our framework; nowhere do we recommend that actual deliberation should always occur involving a ‘fair-minded group of people’ who are selected as such. Our framework consists of a set of questions, followed by some instructions about how the user should approach answering these questions. When considering the questions that comprise the framework, the user should consider how different groups of people would answer the questions: experts, advocates or other professionals; a typical member of the community targeted by or affected by the policy; and someone with a worldview different from theirs. And then, this question is posed for the user: ‘If there is disagreement between these three answers to the question, how [would] a fair-minded group of people who are listening to everyone’s point of view, but also trying to reach agreement…resolve this issue’ (195)? True, when the framework is used in actual deliberative forums involving several participants, these questions are potentially superfluous. Since these forums are normally randomly selected from the general population, their composition will display diversity of background, experience and thought, and the participants’ answers to the framework’s questions will arguably reflect this. Through the very process of deliberating together, listening to everyone’s point of view, and trying to resolve the different issues discussed by the group, the participants will hopefully practice fair-mindedness, in which case the requirement for them to consider how a fair-minded group of people would view the policy would be redundant.

Of course, not all the participants in a deliberative forum will necessarily be fair-minded. For example, some may be biased and prejudiced, and systematically dismiss other participants’ views; and/or they may believe that their view is the only reasonable one and not be prepared to change it; and/or they may not approach the deliberative process with good intentions but rather as a negotiation in which they are trying to win as much as possible for their side. In all these cases, the final question related to a ‘fair-minded group of people’ can play a corrective role by prompting them to deliberate in a different (i.e. fair-minded) way. Furthermore, changing the course and quality of deliberation will often also require more structural changes and cannot be left entirely in the hands of participants. The presence of a competent and effective moderator, the provision of balanced materials and of a diverse panel of experts, and mechanisms to ensure that all participants have a fair opportunity to participate in the discussion and be listened to (e.g. via allocated time slots) can help foster fair-minded deliberation (e.g. see Fishkin 2018).

However, it is when our framework is used by a single policymaker (e.g. a public health official) that, in our view, the question concerning the ‘fair-minded group of people’ will play the most important role. In this scenario, fair-mindedness cannot emerge from a process of deliberation because there is no such process—the policymaker cannot deliberate with themselves, so to speak. Instead, they are asked to engage in counterfactual thinking and consider how people with diverse perspectives would answer the framework questions; the last question—the one that evokes the ‘fair-minded group of people’—is meant to prompt the framework user to consider how this group would deliberate together about the question, starting from their initial disagreement. This thought process is important because simply considering how different people and groups view the policy is not sufficient to ensure that the latter complies with public reason. A policymaker could gather that knowledge, for example, by reading the manifestos of different political parties, social movements and advocacy groups, and by checking opinion polls. None of that information, however, will help them understand what public reason demands. Adopting the perspective of a hypothetical group of people, instead, will force them to consider what kinds of shared political values (rather than non-shared comprehensive doctrines) people with different perspectives might endorse, and to then try to identify reasonable balances of those values and the policies that could be justified by them. While certain guidelines and procedures, as we explained earlier, can encourage an individual policymaker to engage in this kind of counterfactual thinking, this is admittedly more difficult than in the case of deliberative forums involving many diverse participants.

Bias, Asymmetry and Epistemic Injustice in Public Reasoning

In her insightful commentary, Giulia Bistagnino observes that the role we assign to public health officials in our public reason approach is problematic. Although our framework, she argues, would seem to require that public health officials ‘can and should be considered disinterested, neutral actors who provide data and evidence that command unquestioning approval’ (5), in fact their presence risks undermining the very process of public reasoning they are expected to contribute to. This is due to two key reasons: first, public health officials are likely to be biased since they ‘have a political agenda’ (5) based on which they might ‘offer only the scientific and moral considerations that resonate best with their policy preference’ (5) (bias challenge); second, public health officials are in a position of authority, and there is a risk that ‘citizens will blindly defer to their suggestions and reasoning…[which]…is a problem because democratic citizens ought to retain control and should not be required or expected to submit their judgement to the decisions of others’ (5) (asymmetry challenge). As a potential solution to this problem, Bistagnino recommends Fishkin’s (2018) influential ‘Deliberative Polling’ model, in which participants ‘receive balanced briefing materials prior to deliberating, and their deliberation is interwoven with plenary sessions during which they can question panels of experts’ (5, original emphasis) rather than one single public official dominating the discussion.

While these are important criticisms, a number of considerations should be made. For a start, both the bias challenge and the asymmetry challenge depend on the political and social context, especially the culture of the health agencies in which these public health officials work. Changing that culture, if it is conducive to those challenges, would be especially demanding and certainly not feasible in the short term (although it would be interesting to develop proposals for reforms that could help address this problem). If and where bias and asymmetry challenges are present, Bistagnino’s proposal that public reasoning should involve a panel of experts would therefore seem to be the best solution. However, given the range of decisions related to healthy eating policy public health officials are required to make on a regular basis, and given resource constraints, it is unrealistic to expect this kind of robust deliberative activity involving a panel of experts to occur frequently. In most cases, policy decisions may be made only by officials within health departments considering the policy. While our public reason framework is certainly consistent with—and, in fact, openly encourages—the use of deliberative democracy mechanisms to help policymakers make these kinds of decisions, it also aims to provide a tool for public officials who are operating under resource and time constraints, and therefore often cannot avail themselves of the input of deliberative forums. The question that our public reason approach therefore faces is the following: when the use of resource- and time-intensive mechanisms is not an option, how/through what processes can we ensure that a single public health official, or small group of officials, adopts a more balanced and less biased approach, also considering perspectives that are not focused on public health? And, more importantly, how can these processes be institutionalized, so that they are not left to the whim and inclination of individual officials? These are empirical questions. Potential answers/strategies might include that before making a decision, the relevant public health official identifies and familiarizes themselves with a certain number of research papers (e.g. provided to them by their office in the form of a literature review) both in favour of and against the relevant policy, and, where possible, meets/discusses the policy with the authors of those papers. Other similar strategies could also be conceived.

While these strategies might help address the bias challenge, however, they may still fail to respond to the asymmetry challenge. The latter, it should be noted, is only (or most) likely to emerge when public health officials engage in consultative or deliberative processes involving other citizens, and therefore our discussion here only applies to those scenarios (which, as we have just explained, do not exhaust the range of situations in which policy making can be driven by public reasoning). In these contexts, even if panels of experts (as suggested by Bistagnino) were to replace an individual public official, citizens might still unthinkingly defer to those experts. Indeed the presence, during deliberative mini-publics, of ‘plenary sessions during which [participants] can question panels of experts’ (5, emphasis added) may not be sufficient to eliminate the asymmetry challenge: regardless of the number of experts involved, citizens may still feel obliged to defer to their expertise. What is required, in order to address this further challenge, is another aspect that Bistagnino does not mention, i.e. the role of facilitators (or moderators) in the deliberative process (Ackerman and Fishkin 2004; Kuhar et al. 2019), an aspect that we have already highlighted in our response to Zuolo. Only by ensuring that consultative and deliberative processes are conducted in certain ways—in addition to ensuring that they include certain kinds of participants—can the asymmetry challenge be addressed. Facilitators can help ensure, for example, that experts only speak for their allocated length of time, that they express themselves clearly and politely and, perhaps most importantly, that they listen to those non-expert citizens who question them.

An emphasis on the role of facilitators can also help address another important criticism advanced by Bistagnino, one concerning epistemic injustice, i.e. ‘a kind of injustice in which someone is wronged specifically in her capacity as a knower’ (Fricker 2007, 20). According to Bistagnino, ‘[t]he procedures of deliberation and consultation [we] propose are meant to include and give voice to minorities that may be harmed by the policy, but [we] do not say much about the environment that should be fostered to enable those minorities to reason about their condition and express their views’ (6). That is, even though we do acknowledge the disparate effects that certain healthy eating policies can have—e.g. imposing restrictions on the use of food assistance programs may be stigmatizing towards the people who participate in such programs, and interventions that encourage parents to make home-cooked meals may especially burden women, who normally do the majority of feeding work—and we recommend that (representatives from) such groups be included in policy deliberation, we do not consider the issues that may arise during the deliberative process itself. More specifically, the views advanced by members of these groups may be dismissed and, therefore, not taken into account in the design of healthy eating policies, thus de facto rendering their participation in the deliberative process inconsequential. This is because, as Bistagnino points out,

prejudices tend to be deeply entrenched in people’s minds, and they operate automatically, unconsciously, and more rapidly than conscious thought…They are difficult to control and may cause discriminatory attitudes in people who sincerely and wholeheartedly reject them…In this sense, without a tool in place to alleviate the problem, it may be difficult to not neglect the experiences of women with the preparation and organization of meals or to not underrate the problem of stigmatisation of minorities when discussing a healthy eating effort (6).

While this is an important criticism, we do not think that it fundamentally undermines our approach, and particularly our proposal to include members of minorities negatively affected by healthy eating policies in the kinds of deliberative and consultative processes we propose. Instead, it only demands that we qualify our proposal and identify tools and mechanisms which can help prevent the kinds of instances of epistemic injustice that Bistagnino highlights. Indeed, as Fishkin (2018) points out, empirical evidence shows that, contrary to what many suggest, deliberative processes are not always dominated by more powerful voices (e.g. more educated white men). More specifically,

[t]he vulnerability to distortions seems dependent on the design of the deliberative process. At least with Deliberative Polling, if one looks at the initial positions of the more advantaged (the more educated, the males, the rich, etc.) on policy indices on the issues deliberated about, there is no pattern of movement by the small groups in the direction of those positions. The more advantaged are not dominating the discussions by imposing their views. Further study in controlled experiments to disaggregate the precise mechanisms may help explain why, but it seems likely that elements of the design, such as the role of moderators, balanced materials, and ample opportunities for everyone to discuss the issue, play an important role (Fishkin 2018, 76).

These kinds of measures could help reduce or eliminate the risk of epistemic injustice highlighted by Bistagnino (see also Anderson 2012, 168–169) and, while we do not explicitly discuss them in our analysis, our approach is certainly not incompatible with them.

A third criticism advanced by Bistagnino concerns the role of the ‘fair-minded group of people trying to reach agreement’ (6) that we include in our approach. As we have already explained in our response to Zuolo, in our framework the ‘fair-minded group of people’ is mainly conceived as a heuristic device, i.e. the user of the framework is asked to imagine how a fair-minded group of people would decide. However, Bistagnino asks, when an actual ‘fair-minded group of people’ is involved in decision-making, how should members of this group be selected? If the idea is that this is ‘a group of individuals who know better being called upon to resolve disagreements’ (7), then there is a serious risk that the Rawlsian approach that we defend, grounded in the idea of citizens as free and equal, be replaced by ‘a form of epistocracy’ (7), where some citizens should be entrusted with policy decisions more than others. This is an important criticism that requires clarification. The idea of a ‘fair-minded group of people’ should be understood procedurally rather than substantially. That is, we do not claim that certain individuals are more fair-minded than others because of certain intellectual or other personal qualities they may possess. Instead, fair-mindedness is a function of certain rules, procedures and other design features that characterize deliberative and consultation processes. Thus, for instance, in the context of Deliberative Polling discussed earlier, it is ‘the role of moderators, balanced materials, and ample opportunities for everyone to discuss the issue’ (Fishkin 2018, 76) that can help participants act as a ‘fair-minded group of people’, regardless of their education, knowledge, and intellectual skills. This kind of design is more difficult in contexts where, as we showed earlier, policy decisions are in the hands of an individual public official. However, in those contexts too, guidelines like those that we illustrated earlier (e.g. public health officials being required to identify and familiarize themselves with a certain number of research papers both in favour and against the relevant policy) could be used to encourage individual policymakers to reflect on the policy as if they were a fair-minded group of people.

Healthy Eating Policy, Racial Justice, and Global Food Systems

In his commentary, Christopher Mayes suggests that ‘we need to think about racial justice in every policy and consider what that would look like in relation to food, health, housing, and educational policies’ (4). This is because ‘[e]ach of these areas intersect with each other and have long and troubled histories of contributing to racial injustices in middle- and high- income liberal democracies’ (4). We agree that we should think about racial justice in every policy, and we think that our view allows this. Our view—at its simplest—is that healthy eating policies must strike a reasonable balance of shared political values. Racial justice is one of the shared political values that should be attended to, when assessing proposed policies. Therefore assessment of policies should take into account how these policies could introduce new forms of racial injustice, exacerbate existing forms of racial injustice, or alternatively help to ameliorate racial injustice; this may require hearing from members of groups that experience racial injustice how policies affect them. Thus our account is compatible with thinking about racial justice in every policy.

However, there are real questions about how, on our account, consideration of racial justice would play out in practice, and whether racial justice would get due consideration. Racial justice is a shared political value; however, as with other shared political values, people may have different conceptions of it. For instance, some people may believe that racial justice requires roughly equal outcomes across all racial/ethnic groups in a society, for example equal health outcomes such as roughly equal rates across racial/ethnic groups of diet-related illnesses like diabetes or hypertension. Others might believe that racial justice does not require equal outcomes across all racial/ethnic groups, but only requires equal access to key resources and services such as healthy food and health care, or only requires adequate access.

Along with holding different conceptions of racial justice, people may have different views about what constitutes a reasonable balance between racial justice and other shared political values. Therefore, when they consider whether a particular policy strikes a reasonable balance between racial justice and other political values, they may reach different conclusions. For example, some public health advocates and researchers have noted that Black people in the United States experience more food marketing than other groups, the vast majority of which is marketing for foods high in fat, sugar and salt, and that this marketing is likely a contributor to the higher rates of diet-related illness among Black people in that country (Barnhill et al. 2022). In response, one might propose a law limiting the kinds of food marketing that companies may engage in, specifically targeting modes of marketing of unhealthy food that disproportionately affect Black people. While our view requires that a policy is justified by a reasonable balance of shared political values, it does not require that any particular value (e.g. racial justice) be given priority over other values. Therefore, when deliberating about this specific law, policymakers and other deliberators could conclude that in this case it is reasonable to prioritize other values (e.g. the free speech of commercial actors) over the dimension of racial justice at issue (e.g. equal health outcomes). On our view, as articulated in our book, that decision and justification could satisfy the requirements of public reason.

However, there are two risks. First, this decision—prioritizing another shared political value over racial justice—could occur not only with regard to this specific healthy eating policy but also with regard to all or most healthy eating policies, or even all or most policies more generally. Second, deprioritizing racial justice relative to another shared political value could occur repeatedly and persistently over time. In these ways, (healthy eating) policy as a whole and/or (healthy eating) policy over time could fail to address racial justice, even if the evaluation of single policies follows our account.

Thus we need to modify our account—and the framework that operationalizes it—so that it focuses not just on the evaluation of single policies at one moment in time, but on the evaluation of sets of policies over time. Shared political values (such as racial justice) require that certain outcomes be attained and that governments take corrective action as needed—that is, governments must enact policies that address racial injustice in a continuing way. Therefore, along with analyzing specific policies to make sure that these policies strike a reasonable balance between shared political values, there must also be a critical analysis of a) policy as a whole and b) policy over time to make sure that the aims pursued by policy as a whole over time satisfy a reasonable balance of shared political values. If a government’s policy as a whole over time fails to address racial justice, because other policy aims are (almost) always given priority, this amounts to a failure to strike a reasonable balance between shared political values.

Another thread of Mayes’s commentary focuses on the global nature of the food system and food policies. Mayes notes that our analysis is focused on the nation-state, the obligations of states towards their own citizens, and which public policies are justifiable in light of these obligations. However, he observes, there is a global food system. Many of the food policies adopted by one country will affect residents in other countries, and there are multilateral and international agreements shaping the global food system (for example, trade agreements). How should we analyze the justifiability of these policies and agreements? Mayes prompts us to extend our analysis to the global level, ‘[m]irroring Rawls’ extension of his theory of justice from domestic to international politics (Rawls 1999b)’. Providing such an account is beyond the scope of this response, but we can identify a key issue for such an account. Should we extend our public reason account to the global level in a straightforward way, giving a ‘global public reason’ account that mirrors our public reason account? On such an account, policies— including international agreements and policies adopted by nation-states that significantly affect other nation-states—should be justified by a reasonable balance of shared political values, but in this case the political values would be ones that are shared by all involved states. As Quong notes, there is a compelling case that public reason does not straightforwardly extend to the global level in this way, given that there may be no political values that are shared globally and that the ‘relevant agents’ in international relations may be states rather than persons (2022). It stands to reason that the obligations of states towards other states are fundamentally different from the obligations of a state towards its citizens.

Public Reason Beyond Healthy Eating Policy: Sustainability and Non-Human Animals

In the western liberal democracies that are the focus of our book, there are a range of food system problems in addition to unhealthy dietary patterns, including significant rates of food insecurity, average dietary patterns that have a very large environmental footprint, food production practices with unsustainable environmental impact, poor working conditions and wages for many food system workers, and low levels of welfare for most of the animals raised and killed for food. Critical comments from Josh Milburn bring our attention to these kinds of problems. Milburn notes that there is little or no discussion of these other problems in our book, beyond an acknowledgement that some of them exist. He questions whether, given all these problems, healthy eating policy should be a priority for food policymakers, writing that ‘it’s possible that Barnhill and Bonotti are right about healthy eating policy, but that it shouldn’t be too high a priority in comparison to other kinds of food-related initiatives that policymakers could be pursuing’ (5). He asks: ‘Might Barnhill and Bonotti want to resist this possibility?’ (5).

We do not want to resist this possibility. Our book is focused narrowly on healthy eating policy and is silent about the relative priority of addressing other food system problems. This silence is not meant to imply that these problems are less important than the problem of unhealthy dietary patterns. Rather, our book concerns an imagined policy context in which making dietary patterns healthier or addressing diet-related illness has already been established as a policy aim. An example of such a policy context is when a health department official, for instance one who works in the bureau of non-communicable disease prevention and control, has been tasked with developing programs and policies to address diet-related disease. The framework we provide is meant to be used by that official to assess potential programs and policies aiming to make diets healthier. It is interesting to consider how other food policy aims could be incorporated into this policy context; that is, it is interesting to consider how the health official with a narrow mandate might nonetheless integrate other food policy aims and considerations into her work. Following a suggestion from Milburn, our framework could explicitly mention a wider range of side effects or unintended consequences of healthy eating policies. For example, healthy eating policies can make diets less sustainable, if they shift diets towards foods that are healthier but also have a larger environmental footprint. To address this, we could add a question to our framework asking the user to consider whether programs and policies could be designed to make diets healthier and more sustainable. An example of such a policy is one that promotes consumption of a set of fruits and vegetables that have a relatively low environmental footprint.

An interesting question is whether our account and framework, which were developed for a narrow policymaking context focused just on healthy eating policy, can be adapted to food policymaking that has a broader focus—that is, food policymaking that aims to address a range of problems with food systems. We think that it can. As Milburn suggests, a key question for food policymaking is how much priority to give to different food system problems. This challenge arises in part because there can be trade-offs between different food policy objectives, and therefore the question arises of which objective should have greater priority. For example, aligning food production with health-based dietary recommendations may have negative environmental effects, e.g. following US dietary recommendations to consume more fish could, if followed, deplete global fish supplies and threaten fishers’ livelihoods (Institute of Medicine, National Research Council 2015). Another example is that one way to reduce the environmental footprint of diets is to reduce consumption of red meat by replacing red meat in the diet with another food, but this could increase consumption of chicken which arguably is bad from the perspective of animal welfare (Mason-D’Croz et al. 2022). When there are these kinds of trade-offs between worthy policy aims, a key question is which food policy aim should have priority. Can our public reason account and our framework be modified to help users address this question? We think it can. Just as individual policies must strike a reasonable balance of shared political values, so too the set of overall policy priorities must be justified by a reasonable balance of shared political values (along with meeting our other criteria).

If we broaden our focus from healthy eating policy to food policy as a whole, we must also broaden our conception of who count as relevant stakeholders. Here another one of Milburn’s comments comes into play. Milburn notes that we argue that it is ‘especially important that groups which are (or risk being) overly burdened by a policy are included in [decision-making, deliberative, and consultative] processes and given voice’ (198). We had in mind groups such as women, the poor, and racial and ethnic minorities. But Milburn presses us to consider other groups that are especially burdened by food policies—for example, non-human animals and future generations of humans. We agree that when considering food policy as a whole, future generations of humans are a key group that may be overly burdened by policy (in particular, overly burdened by policies’ failure to limit the greenhouse gas emissions related to food systems and to promote environmentally sustainable food systems). It is a good idea—and consistent with our public reason approach—to include the interests of future generations of humans when designing and assessing food policies. There are small ways in which our account and framework can be modified to do this. First, statements can be added to our framework to prompt the user to explicitly consider the interests and potential perspectives of future generations of humans. Specifically, the question: ‘Is the policy likely to have any unintended positive or negative side effects?’ can be followed by the statement: ‘Consider effects on both current and future generations of people’. In addition, the framework asks the user to consider how various people would answer the framework’s questions, including advocates, members of the community targeted by the program, and someone with a worldview different from their own. ‘Members of future generations of people’ can be added to this list.

We would like to stress that what we are proposing here, in line with our overall account, is a heuristic device: i.e. the user of the framework is asked to imagine how future generations could be taken into account when designing and evaluating food policies. This means that it is not necessary, for this future-oriented public reasoning to occur, that people with (more or less direct) links with future generations (e.g. those who are parents or who can and would like to become parents; futurologists; etc.) be actually included in the process of public reasoning. Importantly, the presence of people who have links with future generations may be unnecessary even when our framework is employed in actual deliberative forums involving several participants. As Graham Smith explains in his analysis of deliberative mini-publics as a tool to advance the interests of future generations, the random selection of participants that characterizes mini-publics ‘is a protection against strategic action from those with structural power who benefit from current social and economic arrangements that privilege the short-term’ (Smith 2020, 1008–9). More importantly, Smith argues,

facilitated deliberation…orientates participants toward consideration of the long term. Within DMPs [deliberative mini-publics], participants have the opportunity to learn about the issue at hand, hear from and question a cross-section of experts and advocates, reflect on what they have heard, and listen to the views of other participants. Facilitation promotes equality of voice across a diverse group that differs significantly in terms of confidence and experience and engenders respectful interactions among participants with very different interests and perspectives. DMPs approximate the type of communicative rather than strategic motivation celebrated by deliberative democrats. Such conditions are particularly apposite for encouraging considered judgment…or collective intelligence…that is sensitive to the interests of future generations. Where DMPs are tasked to consider aspects of long-term policy making, questions of intergenerational equity are made salient to participants and they have time and space to reflect on the long-term consequences of social choices, informed by the variety of perspectives offered by fellow participants (Smith 2020, 1009).

While, of course, encouraging long-term thinking among participants in deliberative mini-publics may be further enhanced via certain design features, such as quotas for young people (Smith 2020, 1008), it is the very way in which participants are expected to deliberate that can help foster consideration of the interests of future generations. Our public reason framework, enriched with questions and statements concerning future generations, can play a key role in guiding this kind of deliberative process.

A further observation made by Milburn concerns the status of non-human animals in our framework. Healthy eating policy, and food policy more generally, have significant direct and indirect effects on animals, ranging from ‘the effects of our food systems on the global climate, on plastic pollution, on ecosystem collapse’ to ‘the fact that our food policies permit (or encourage) the slaughter of trillions of thinking, feeling beings annually’ (5). It is not clear, Milburn argues, what role (if any) our framework assigns to these considerations, and to the rights and interests of animals more generally. And it is also unclear, in his view, why healthy eating policies should be prioritized over policies that protect animals’ interests. Our response to Milburn’s concerns is twofold. First, and perhaps unsatisfactorily from his perspective, we agree with Zuolo (2020) that there is reasonable disagreement about the status of animals in contemporary liberal democracies, and that therefore it would not be legitimate, from a public reason perspective, to implement policies—including healthy eating policies—grounded in the view that animal imprisonment and killing should never be allowed. However, we also agree with Zuolo that, based on public reason, ‘[w]e ought to minimize animal suffering in interactions with human beings as much as reasonably possible’ (Zuolo 2020, 207) since this is a point on which reasonable persons agree despite their disagreements on other aspects of animals’ status. Our framework can be extended to include this consideration, for example when we examine the potential side effects of a healthy eating policy. Second, we are also not opposed to the idea that reasonable disagreement regarding the status of animals may decrease over time, and that beyond current agreement on the importance of minimizing animal suffering a broader and deeper agreement may emerge in the future which may result in animal killing and imprisonment also being considered inconsistent with public reason (cf. Flanders 2014). Our framework can and should be responsive to these kinds of changes. And, although this may be well beyond the scope of our book, it would be interesting to consider whether rather than simply waiting for societal agreement to emerge over time, specific deliberative forums could be designed precisely to encourage greater agreement on certain shared principles, e.g. by presenting participants with moral arguments and empirical evidence that can help shift their views on the status of animals. These forums’ findings could then be conveyed to the broader population and used to initiate broader debates about the status of animals, with potential implications for policy and constitutional change.

Understanding Public Health Practice: Population Health and Evidence-Based Policy

In his commentary, Donald T. Thompson identifies aspects of actual public health practice that he finds problematic. For example, he raises the concern that ‘public communication of the scientific information relating food, diet, and health often gives a mistaken impression about the nature of the scientific evidence’ (13). Most of the scientific evidence about food, diet and health is information about the incidence of an outcome (e.g. cardiovascular disease) among a population of people who experienced a certain ‘exposure’ (e.g. had a diet high in vegetables, fruits, and whole grains). This evidence shows us correlations between dietary patterns and health outcomes. But this evidence is often described in ways that ‘suggests a causative relationship’ (13) that may or may not exist, and which personalizes risk. Thompson writes:

In marketing of either dietary recommendations or food products it is not uncommon that putative population-level risk is personalized to the reader/consumer, as ‘your risk of CHD would be lower if you changed your diet’, or ‘eating this breakfast cereal will lower your risk’ (13).

Thompson objects to this personalization of risk on multiple grounds. One reason is that it can mislead individuals into thinking that they will experience individual benefit if they change their diet: if ‘the rhetoric of prevention is presented to individuals as a reason for them to expect an individual benefit, that will be misleading’ (14). Another reason is that the personalization of risk frames the conversation about dietary change and healthy eating policies around the issues of individual risk and individual benefit. This ‘leads to a sense of individual responsibility for individual risk’ (p14). It also leads to ‘asking whether government should intervene in a paternalistic manner to help individuals change their behavior to reduce their risk’ (14). In other words, when a change in dietary patterns is presented to the public as a way for an individual to reduce their risk of bad health outcomes, then policies that aim to change dietary patterns will also be seen as ways of shaping individual behavior in order to reduce individuals’ risk of bad health outcomes. Framed in this way, a salient question about these policies is whether the government is justified in interfering with individuals’ freedom in order to reduce their risk—that is, whether government paternalism is justified.

However, Thompson notes that a distinct and better way in which public health policies have been framed is that they aim to secure ‘population-level benefits’ (9), not (or not just) individual benefits. This ‘population-level’ frame aligns better with the actual nature of the scientific evidence, which concerns populations. The ‘population-level’ frame also aligns with the actual nature of prevention in public health, which is ‘reduced future population incidence’ of a disease, not a reduction in a given individual’s risk (14). But what are the ‘population-level benefits’ of public health policies? Thompson cites Latham (2016), who ‘claims that an individual’s improved health creates social benefits’ (12). We agree. The social benefits of an individual’s health, we could argue, might accrue from increased participation in a range of activities that benefit not only the individual themselves but also others, such as care-taking, paid work that benefits others, or civic engagement. Thompson also suggests that in addition to an individual’s health having social benefits, one could also argue that ‘the overall social benefits of population health would be qualitatively distinct from the aggregate of individual benefits of individual healths [sic]’ (12). He suggests that if the health of a large number of individuals is improved, this ‘would produce multiply interactive social-level benefits’ (12), such as ‘families having more years to enjoy each others’ company, longer and deeper friendships, or organizations of all types benefitting from the wisdom that accrues from members interacting while living longer lives’ (12). We appreciate the point that a healthier population could enable a society to achieve certain social-level benefits; these benefits should be considered when evaluating public health policies, including healthy eating policies, through our public reason framework. It is also worth considering Thompson’s suggestion that population health ‘might be considered a “primary good” of the civic body, going beyond the idea that primary goods only pertain to individual citizens to enable them to fully function as citizens’ (12). While this analysis cannot be accommodated here, it would be useful in future research to expand our public reason framework to also include these social primary goods.

Thompson also questions the use of the term ‘common good’ in our analysis. He observes: ‘On nine of the pages the text refers explicitly to the common good in terms of health care costs, efficiency, or economic prosperity. This usage is not consistent with Rawls’ non-utilitarian theory. Other usages tend to refer to a broadly political sense, one which at times might well be consistent with Rawls’ political liberalism’ (15). We agree with Thompson that we could have articulated the meaning of the ‘common good’ more clearly in our analysis. However, we would like to reject Thompson’s suggestion that our understanding of the term is a utilitarian one. Instead, we believe that the economic benefits of healthy eating policy that we refer to should be understood in the same vein as the social-level benefits highlighted by Thompson in connection with population health. For example, economic prosperity may enable the state to raise more revenue through taxation and use those additional funds to advance both individual and social primary goods, e.g. by investing in education, public transport, and public health itself.

Another criticism raised by Thompson concerns the scope of public reason, i.e. the range of issues and policies to which the constraints of public reason apply. More specifically, Thompson challenges our decision to ‘expand Rawls’ narrow application of public reason beyond constitutional essentials and the basic structure’ (5) to also include healthy eating policy. He writes:

Although in PL [Political Liberalism] Rawls was careful to limit the required application of public reason in this way, in fact he said that deliberations of constitutional essentials and basic justice were restricted only to public reason; he suggested that elements of public reason might be applied more broadly, but in those situations he suggested that considerations based on other than public reason might be applicable as well. Expanding the application of public reason might mean that only public reason applied to ordinary legislation as well, but this seems not what Rawls argued (6, note 5).

We believe that here Thompson is conflating the scope of public reason with its content. When it comes to the former, we remain convinced that it is not unwarranted to extend the scope of public reason to issues of policy and legislation that are beyond the realm of constitutional essentials and matters of basic justice. As we stress in our book (125), Rawls himself seems to endorse (or at least not to oppose) this approach (2005, 215). Furthermore, one could also argue that healthy eating policy does concern—though perhaps only indirectly—matters of basic justice, insofar as promoting people’s health can help protect their rights and opportunities (Rawls 2005, 228; see also Daniels 2008; Bonotti 2015, 404). Defining the scope of public reason, though, is not sufficient per se to establish what complying with those constraints entails. Indeed, we agree with Thompson that, for Rawls, ‘considerations based on other than public reason might be applicable as well’ (6, note 5). This is in line with the ‘wide view’ of public reason, defended by Rawls (2005, 453), which we discussed in our response to Zuolo. However, we do not believe (and we never suggest in our book) that only public reasons may be invoked when healthy eating policies are debated, but just that the constraints of public reason do apply to this policy area. Those participating in these deliberations may also invoke non-public reasons, as long as they eventually provide public reasons too to justify their positions.

A final aspect of Thompson’s commentary that we would like to engage with concerns the relationship between the level of evidence for a policy and the kind of policy that the evidence may help justify. Thompson notes that the US federal government, specifically through the DHHS [Department of Health and Human Services] and USDA [United States Department of Agriculture], releases dietary guidelines based upon scientific evidence relating diet and health. He then writes:

By using the language of guidelines it would seem that DHHS and USDA believe that the evidence relating diet and health is best suited to policy recommendations rather than serving as the basis of coercive policy: particular policy proposals that rely on coercive state actions would seem to require stronger evidence (10).

Thompson is here attributing to DHHS and USDA the view that existing evidence about diet and health is not strong enough to support coercive policy; Thompson may share that view. We agree with Thompson that the coerciveness of a policy is certainly an important aspect of it that ought to be factored into its public justification, but it is not the only (or even the main) one. As we point out in our book (169–170), establishing whether the evidence base for a healthy eating policy is sufficient also entails considering other factors, such as the potential consequences of no (or delayed) implementation and the effects of a policy (or lack thereof) on different groups of people. As Nancy Kass, whom we cite in our analysis, explains,

[w]hile all programs must be based on sound data rather than informed speculation, the quality and volume of existing data will vary. The question for policy and ethics analysis, then, is what quantity of data is enough to justify a program’s implementation? As a rule of thumb, the greater the burdens posed by a program—for example, in terms of cost, constraints on liberty, or targeting particular, already vulnerable segments of the population—the stronger the evidence must be to demonstrate that the program will achieve its goals (2001, 1778-9).

Given that our public reason framework requires justifying healthy eating policies based on a reasonable balance of shared political values, and given that such values include both individual liberty and the fair distributions of the burdens resulting from the policy or its non-implementation (among others), its coercive nature alone cannot determine the amount and type of evidence necessary for it to be publicly justified.

Enriching Public Reasoning on Healthy Eating Policy: Metaphysics, Ontology and Epistemology

In the first part of his insightful commentary, Andrea Borghini critically engages with the broad topic of our book. He asks: ‘is the book after policy and political philosophy issues related to healthy eating, a healthy diet, or certain healthy food items?’ (2). The book is about healthy eating policies, and these policies have different foci. For example, some policies—such as education about health-based dietary recommendations—aim to promote overall dietary patterns that are healthy (that is, dietary patterns linked to lower rates of diet-related disease), whereas other policies target specific food items the consumption of which is associated with higher rates of diet-related illness (for example, sugary drinks). There could also be interventions that aim to promote a ‘healthy’ approach to eating, for example mindful eating (The Nutrition Source 2020). Borghini correctly argues that it is important to identify the focus of a healthy eating policy, because that will affect what framework is needed to properly analyze it. Depending on the focus of a policy, different stakeholders should be included and different kinds of scientific expertise will be needed to evaluate the policy. Borghini suggests that the framework we offer, which is meant to be used to evaluate policies, should be revised so that it is sensitive to the different foci of policies. He provides the following example:

Over-simplifying for the sake of exemplification: a conversation on healthy foods and food products will be highly focused on nutrients, production processes, and food systems; it will involve actors such as food producers, distributors, and market regulators; and it will require an expertise in producing food, rather than consuming it. Instead, a conversation on healthy eating will focus on modes and practices of consumption; it will involve psychologists, marketers, and assorted activist groups (e.g. linked to fat activism, eating disorders, etc.); and it will call not only for experts from the scientific community but also for the perspectives of actors that can represent communities and groups affected by healthy eating (3).

Borghini makes a good point. We agree that our framework could be improved by providing more specific guidance about the kinds of stakeholders, experts, and evidence that should be included when evaluating different kinds of policies.

We also appreciate the broader point that Borghini makes, which is that our work—which is focused on the ethics and legitimacy of food policies—would be enriched if it were linked to additional work in the metaphysics and ontology of food and eating. As Borghini puts it:

The call is for experts in food metaphysics and ontology, to lay out a framework that can accompany specific policy effort, e.g., efforts about healthy food production, about healthy dietary habits, about healthy meals, about healthy food products, or about healthy eating. Theoretical tools should be aligned to match policies, or else we would run the risk of crisscrossing and conflating—say—a bunch of healthy food for what constitutes a healthy diet or being served a healthy dish with what makes for an occasion of healthy eating.

But the implications of Borghini’s criticism, which we welcome, are also broader. Beyond the specific domain of healthy eating policy, it is necessary that public reason more generally includes metaphysical and ontological standards. True, scholars of ontology and metaphysics also disagree with each other, and it is for this very reason that Rawls famously excludes metaphysics from the domain of political liberalism. Yet, these scholars also share common standards of inquiry (e.g. see Paul 2012; Hawley 2018), including inference to the best explanation, conceptual coherence, semantic transparency, and explicit disclosure (see Bonotti et al. 2022; 2023). Ensuring that public reasoning is grounded in these kinds of standards, as well as in shared political values and empirical evidence, could both strengthen its epistemic credentials and facilitate the tailoring of public reasoning to specific policy areas, as Borghini suggests.

An additional point highlighted by Borghini concerns the ‘shared evaluative standards’ that are central to our account and which ‘may include, e.g., the methods and standards of the natural sciences (139)’. Borghini argues that much more could be said about what specifically counts and what does not count as ‘shared evaluative standards’ in this domain. He notes the ‘messy structure of contemporary scientific research on nutrition’ which includes ‘processes and methods such as citizen science, privately funded research, and multiple contrasting approaches’ (6). He questions: ‘Are there really core methods that are shared?’ (6). We agree that more could be said about the different kinds of scientific research on nutrition, the different communities doing this research, and which methods they share. This, however, would require an approach grounded in the philosophy and sociology of science, and potentially require a book-length analysis in itself. While we could not systematically address this problem within the limits of our book, we agree that this is an important and fruitful area for future research and hope that our book will help to instigate this kind of work.

We trust that our framework, while currently limited in the ways highlighted by Borghini, does not preclude future expansion to incorporate the kind of complexity that he highlights. In fact, some of that complexity is already taken into account in our book. As we argue in our analysis, even when members of the scientific community employ the same evaluative standards, they often reach different conclusions, some endorsed by most scientists (‘majority science’), others by a minority of them (‘minority science’). Furthermore, we also identify ‘bad science’—which employs shared standards but commits gross epistemic errors when employing them—and ‘pseudo-science’, which instead relies on standards/methods not shared by the other three groups (156). Our framework could be further complicated by acknowledging a) that shared standards/methods may be different within different scientific domains (e.g. nutrition science vs. immunization science) and that b) there may also be disagreement about standards/methods within the same domain (e.g. nutrition science), without any of the different groups emerging from this analysis being like those engaged in pseudo-science. When different (non-pseudo-scientific) methods/standards are present, the challenge for us would be to understand whether such different methods are still grounded in shared standards of some kind, what those shared standards might be, and whether—if they are indeed shared—they can still provide the foundations for accessible public reasons. This is a challenging task that we hope future research on public reason will address.

Borghini also challenges our endorsement of an indirect model of public reasoning ‘where only certain actors (e.g. scientists and regulators) would join the public debate on behalf of all other actors too’ (6). ‘This solution’, he argues, ‘risks not only of being seen [sic] as elitist: it again undermines the complexity of scientific discourse’ (6). However, as we explain in our book (149), what the indirect model of public reasoning requires is that politicians and policymakers comply with public reason (and that political institutions be designed in ways conducive to this outcome), but this model does not say anything about scientists. The latter are normally not entrusted with decision-making qua scientists and therefore are not constrained by public reason. It is the use by politicians and policymakers of the evidence provided by scientists that is subject to public reason standards. Furthermore, when politicians and policymakers use consultative and deliberative forums to help them design policies consistent with public reason, those could be considered extensions of their role and hence also subject to public reason constraints. The purpose of the indirect model is to relieve ordinary citizens of the constraints of public reason and thus facilitate a more inclusive public debate on healthy eating policy, not to advance an elitist agenda.