Introduction

X-Phi—the study of philosophical problems using methods familiar from empirical research, especially experiments—is a thriving and exciting discipline. During the last 20 years or so, X-Phi has been the source of several significant contributions (see, e.g., Alfano et al. 2022; Knobe and Nichols 2017; Li and Zhu 2023; Stich and Tobia 2016; Weinberg 2016). What I shall try to do here is to chart and assess central accounts of the relevance of X-Phi for theory acceptance in political philosophy and for the dominating method of reflective equilibrium within this field. X-Phi has affected political philosophy less than other areas in political philosophy (Lippert-Rasmussen 2024). Yet, as I shall show, the question of its relevance in this context connects in interesting ways with methodological debates within this field, for example regarding the extent to which principles should be fact-sensitive. And the various accounts of how X-Phi may be relevant to philosophy more generally are readily applicable to political philosophy. In setting out those accounts I draw on extant works (critically assessing their main claims). The contribution of the paper is especially to provide an architectonic of different ways of understanding the role of X-Phi within political philosophy. Specifically, I shall distinguish between the following four views: (i) X-Phi as a systematic method to avoid or reduce biases in our moral intuitions—The De-BiasingFootnote 1View; (ii) X-Phi as a tool for assessing the fruitfulness or consequences of various concepts—The Fruitfulness View; (iii) X-Phi as the best way to unearth the kind of moral principles we are interested in as political philosophers—The Unearthing Principles View; and (iv) X-Phi as a way of testing the coherence of principles with folk intuitions—The Defeasible Reason (or Squatters’ Rights) View.Footnote 2 The paper argues that the first two views and the last—subject to some clarifications—describe important X-Phi contributions to theory acceptance as this is traditionally understood within contemporary political philosophy. In contrast, the third view goes too far in suggesting that X-Phi can supplant more traditional, in part non-empirical, methods. My argument speaks especially to those who are not skeptical of standard political philosophical method, including its reliance on intuitions. It suggests that this method can benefit in several respects from engagement with experimental approaches. Further, such an extended research program can arguably embrace at least important aspects of officially alternative methodological programs such as normative behaviorism (see further, Floyd 2017).

The article is organized in the following way. In section “Preliminaries”, I address some preliminary matters, including what is meant by ‘experiments’, and what their key scientific value is. I also specify what X-Phi is and distinguish between central programs within it. The following four sections (i.e., sections “The De-Biasing View”, “The Fruitfulness View”, “The Unearthing Principles View”, and “The Defeasible Reason (or Squatters’ Rights) View”) consider in consecutive order the mentioned four accounts of the relevance of X-Phi for theory acceptance. Section “Conclusion” concludes.

Preliminaries

The key virtue of experiments is, in brief, their conduciveness to the aim of making valid causal inferences—an important scientific aim (Angrist and Pischke 2015; King, Keohane, and Verba 1994). The so-called randomized experiment or trial is crucial in this regard. What we are interested in is the causal effect of a certain factor, say a given intervention or policy. Recent fascinating and highly practically important works in developmental economics are cases in point. For example, we want to know the exact effect of interventions such as equipping classrooms with a white board or of attending to high teacher absenteeism rates (the Royal Swedish Academy of Sciences 2019). This is tantamount to us wanting to know how a group (e.g., a group of students) would have fared in the absence of the intervention in question—what difference (if any) that the intervention makes.

Now, we can estimate the average impact of an intervention on a group by comparing it to a group not exposed to the intervention. Yet when making this comparison we should be cautious. While potential differences in outcome reflect in part the effect of the intervention, such differences may also reflect so-called selection bias, that is, the differences in ‘counterfactual outcomes between the two groups in the absence of the program’ (the Royal Swedish Academy of Sciences 2019, p. 8). If we fail somehow to remove this bias, our comparison would not be an ‘other things equal’ comparison, meaning that we cannot ascribe estimated effects to the intervention (i.e., we cannot confidently say that the estimated effect is a causal effect of the intervention) (Angrist and Pischke 2015, pp. xii–xiii).

One way of achieving an ‘other things equal’ comparison is by randomly assigning individuals to a control and a treatment group. This is a way of ensuring that the only systematic differences across the groups are due to them being exposed to the treatment or not. That is, we isolate the component ascribable to the treatment and get rid of the selection bias. The basic idea of a randomized trial and its credentials in terms of causal inference can be realized in the form of various experiments, including field, natural, survey, and laboratory experiments (Alfano et al. 2022; Sect. 1.1; Petersen et al. 2010; Miller 1992; the Royal Swedish Academy of Sciences 2021; Ulas et al. 2017).

Experimental philosophy, we may say, is ‘the empirical study of moral intuitions,Footnote 3 judgements, and behaviors’ involving ‘gathering data using experimental methods and using these data to substantiate, undermine, or revise philosophical theories’ (Alfano et al., Intro.). The mentioned key virtue of experiments in empirical research—i.e., causal inference—plays important roles in experimental philosophy. For example, causal inference is useful for the task of detecting biases in our moral intuitions (see The De-Biasing View). Further, experiments allow us to pinpoint exactly the moral factors at play in our assessment of moral issues or dilemmas (see The Unearthing Principles View). Yet what is undertaken under the auspices of X-Phi is sometimes of a more descriptive nature. We are often interested simply in charting ‘folk intuitions’, and to compare them with standard views in the discipline of philosophy. Regarding X-Phi, strictly speaking it should be mentioned that despite its virtues as pointed to, some critics are concerned about the so-called external validity of some of the experiments I have mentioned (a worry that applies to those methods generally, not only to their use in addressing philosophical problems). That is, what we find in the laboratory might not reflect our social world. Accordingly, we might need other forms of empirical research that are likely to do better in this regard.

It is common to distinguish between so-called negative and positive programs within X-Phi. In brief the distinction is this. While the two programs are united in using an experimental method, they differ in terms of the use to which they put the method in question. In the negative program, it is used to debunk or undermine moral intuitions—the ‘data’ of traditional philosophical work within the method of reflective equilibrium (McMahan 2013; Rawls 1999, pp. 15–19, 40–46). In the positive program, in contrast, the experimental method is used in ‘addition to traditional philosophical appeals to intuition’ (Alfano et al. 2022, Sect. 2.1). For example, it is used to assist the philosopher in distinguishing between reliable and unreliable intuitions, to provide support for certain conceptions and related intuitions, to help to discover the principles underlying our intuitions, or to ascertain how closely a given philosophical view coheres with folk intuitions. In this way, the views I discuss here all fall within the latter, positive, program. It is to those views I turn now, addressing the above-mentioned four views in consecutive order.

The De-Biasing View

The first view regarding the relevance of X-Phi to theory acceptance is a staple in the X-Phi program. It is referred to using various names, including ‘buck detecting’, and ‘the-wheat-from-chaff’ project (Handby 2022, p. 1597; Weinberg 2016, p. 77). The argument is that our moral intuitions—folk intuitions and philosophers’ intuitions alike—are not free of errors or biases. Further, detecting those biases, and in part also successfully overcoming or mitigating them, requires the use of the experimental tools of X-Phi. For example, whereas it seems clear that our assessments of various cases should typically not be influenced by the order in which they are presented to us they might be thus influenced. While philosophers can in part self-correct for such sources of distortion, a third-party perspective is often an important systematic supplement to such attempts (cf. Andersson et al. 2014; Kahneman 2011; Mengarelli et al. 2014). This is in part confirmed by X-Phi studies that show that philosophers themselves are subject to sources of errors almost on a par with non-experts (Floyd 2017, p. 149; Stich and Tobia 2016, Sect. 1.4). The aim of such systematic analyses is to clear away, to the best of our abilities, possible sources of error pertaining to our moral judgments.

What are the relevant sources of error that political philosophers should be aware of and try to avoid? In a recent article Handby (2022) provides a useful overview. For those familiar with the works of Kahneman (2011) and Richard H. Thaler and Cass R. Sunstein (2009) the items on the list will not be surprising; so unless it is crucial to my argument, I shall not explain the decision-making foibles alluded to. The sources of error Handby points to include: availability heuristics; anchoring; nudges; ordering; and framing. He also emphasizes a broader way in which our intuitions may be biased, namely that they are importantly shaped by the institutional structures that we are subjected to.

How should we regard this view’s relevance for theory acceptance in political philosophy? One may initially think that the view would ground a skeptical conclusion: as the catalogue of biases brings out, our decisions and judgments (including our intuitions) are subject to many irrelevant influences; why should we trust them? Yet I think that a very different conclusion is warranted. While we have every reason to pay attention to the indicated sources of error, we should not, so to say, throw the baby out with the bathwater. That is, we should use the insights regarding our susceptibility to various sources of error to try to avoid those errors. Consider as a parallel, errors that pertain to our empirical observations. For example, we see the stick bending when we immerse it in water, although it does not. Likewise, there are plausible sources of error pertaining to our moral judgments. For example, our obligations to our fellow nationals seem to us to be of crucial moral importance. Yet such obligations may in fact be derivative from our obligations to a wider moral community (ultimately to humanity or sentient beings as such). Our intuition that fellow nationals are more important morally speaking than strangers may reflect the fact that coordination is of the essence in the discharge of such impartial obligations (cf. Goodin 1988).

How then should we go about avoiding, to the best of our abilities, sources of error pertaining to our moral judgments? John Rawls—one of the leading representatives of contemporary political philosophy and the accompanying method of reflective equilibrium with its reliance on moral intuitions or considered moral judgmentsFootnote 4—was highly attentive to errors of the indicated sort, and his famous contractualist method was in part designed to neutralize them (Rawls 1999, pp. 15–19, 40–46). Most directly he pointed out that, for example, situations in which our self-interest looms large may distort our moral judgments; and so may, for example, circumstances in which we are upset or angry.Footnote 5 Further, the so-called veil of ignorance in Rawls’s theory is attuned to some of the broader problems that Handby points to. The veil deprives those assigned the job to choose principles for regulating society’s basic institutions from various kinds of particularized information about themselves and their place in society. In general, the veil sees to it that people’s assessments of principles of justice are not unduly influenced by the fact that they have lived under certain institutional structures, and by the positions they have occupied within those structures. Handby points to logically similar, but more empirical, methods to correct for parochialism (Handby 2022, pp. 1602–1603). Relatedly, Erik Gustavsson and Lars Lindblom have recently, in continuation of the work of Elizabeth Anderson and others, pointed out how including a broader set of voices or social perspectives can in part correct for parochialism, idiosyncrasies, and other such phenomena (Gustavsson and Lindblom 2023, p. 4). Distortions of the indicated sort constitute a real risk if philosophers rely only on their own assessments—assessments that are in part developed against the background of an often-privileged social position. I see these works as continuous with the guiding aim of avoiding distortions in how we form our judgments. Yet they also in part go beyond the traditional method of reflective equilibrium in that they include systematic information about people’s intuitions based on representative samples. A case in point is a recent X-Phi study that shows that while people’s intuitions in some respects track a mature luck egalitarian account of distributive justice (i.e., they reflect the stance that disadvantages that people can control are less problematic than disadvantages they cannot control) in other respects they do not (i.e., they do not reflect the view that it is problematic if people are worse off than others due to choices that reflect some of their important values and hence are costly for them not to make). Arguably, such discrepancies call at least for reflection on the part of philosophers regarding their cherished doctrines, suggesting for example that disadvantages of the latter form are not categorically wrong (Pedersen et al. 2024).

In sum, on the view sketched in this section X-Phi is plausibly seen as continuous with traditional endeavors of political philosophy: the former is instrumental in reducing errors or biases in our judgments.

The Fruitfulness View

In presenting the second view on the relevance of X-Phi to theory acceptance in political philosophy I draw on a recent article by Matthew Lindauer which I think represents the view well (Lindauer 2019). The general idea is this. Especially for conceptsFootnote 6 in political philosophy it is true that they play a practical role. That is, they are in part formulated against the background of some challenges that a community or a polity faces. John Locke’s notion of toleration and Rawls’s conception of ‘justice as fairness’ come to mind (Lindauer 2019, pp. 2131–2132, 2139; Locke 1689/1983, pp. 25–26; Rawls 1999, Part III; Rawls 1993, p. 4). The core concepts of those works are in part recommended on the ground that they are likely to have certain consequences in realistic circumstances, for example facilitate peaceful coexistence and to some extent generate their own support when regulating society’s fundamental institutions. Jonathan Floyd has recently suggested that when we assess normative principles and institutional structures regulated by such—and indeed, more generally, when we assess various answers to ‘how should we live’—we should focus on the extent to which citizens living under them will be inclined to engage in criminal conduct and ultimately in rebellion (with a comparatively high likelihood of being associated with such consequences being something that speaks against the principles and institutions in question) (Floyd 2017, esp. Chap. 3).

On this understanding of the nature of concepts, their success and acceptability hinge importantly on their foreseeable effects when adopted by individuals in societies and when they regulate basic institutions in society. Especially, their ability to generate citizens’ motivation is arguably key (see further, Lindauer 2019, p. 2133). Now, to ascertain such effects we need reliable methods to uncover relevant causal relations. In that regard the credentials of X-Phi narrowly conceived (i.e., as involving controlled experiments) are unsurpassed. Yet X-Phi more broadly conceived, e.g., surveys and other empirical methods of the social sciences, are clearly also relevant to the task of uncovering the effects of concepts. Together, various empirical methods may contribute directly to theory acceptance in political philosophy. In a nutshell, this is the argument for the relevance of X-Phi from the practical role of concepts in political philosophyFootnote 7—the argument from fruitfulness.

The first point I want to make about this view is that insofar as we are dealing with theories of the kind indicated above—i.e., theories that imply claims about consequences or causal relations—the relevance of X-Phi seems undeniable (and important). In some cases, proponents of theories of this kind have comprehensively embraced this feature of their theories. For example, they have conducted a range of surveys to determine what kind of principles are most conducive to eliciting agreement or consensus, have undertaken historically informed analysis of the relative performance of various institutional schemes, or they have proceeded more qualitatively, trying to unearth common grounds in ordinary morality across various cultures (Beitz 2009, Chap. 4; Floyd 2017, Chap. 3; Klosko 2000).Footnote 8 Others have been more reluctant in this regard. They have emphasized, for example, that the consensus they are interested in is of a highly idealized form. For example, they have claimed that it is only consensus between ‘reasonable’ citizens that matter (Cohen J. 1993; Quong 2010; Rawls 1993). In brief they have put forward a much more stylized setting as the relevant setting in which to assess for example the ability of a conception to generate the support of those subjected to institutions regulated by it. Yet as stylized as the setting may be, this does not render empirical testing irrelevant—indeed the acceptability of the theories in question depends importantly on their ability to generate certain consequences in the relevant (idealized) settings. Ascertaining such consequences bears some similarity to model construction in economics used for explanatory or predictive purposes (Hausman 2018, Sects. 2.4, 3.2).Footnote 9 So at least with respect to the important group of consensus-oriented theories alluded to here, Lindauer’s fruitfulness view on the relevance of X-Phi appears sound and informative.

In contrast, theories that seek not to rely on claims regarding consequences—including facts about motivation—appear clearly not to be amenable to the argument from fruitfulness as it has been set out here. A case in point is G.A. Cohen’s defense of what he calls fact-insensitive (ultimate) principles (Cohen 2008, Chap. 6). In this sense the scope of the fruitfulness view is restricted. Yet this does not show that the fruitfulness argument is irrelevant even as regards theories that assign ultimate principles the primary place. First, proponents of such theories, including Cohen, do not dismiss the relevance of so-called regulative principles. These are precisely principles that are designed to have certain consequences, which makes the relevance of X-Phi obvious. More generally, as proponents of fact-insensitive (ultimate) principles are keen to stress, facts, and a reliable grasp of such, are staple in all-things-considered assessments; and the practical import of any given theory rests on such assessments (Cohen 2008, Chap. 6). Accordingly, the relevance of reliable empirical methods for theory acceptance is clear even as regards normative theories that deny the relevance of such methods at the ultimate level, that is, as regards ultimate principles.

While consequences or facts are relevant to theory acceptance in the indicated respects, their relevance is sometimes overstated. One of Lindauer’s arguments is a case in point. He mentions the so-called self-effacingness objection to consequentialism and takes this to be an instance of political philosophers’ empirical interest in the motivational fruitfulness of normative concepts (Lindauer 2019, pp. 2133–2134). I shall show that appropriately conceived it is not: rather, the objection demonstrates that sometimes empirical analysis is not of relevance to important contributions within political philosophy.

In brief, the self-effacingness objection is this. Consequentialism tells us to act to make the outcome as good as possible (let us assume in well-being). Yet if people are motivated by this doctrine, chances are that consequences would not be as good as they could be, that is, compared to a situation in which people followed another doctrine closer to ordinary morality. This is in part due to various flaws in our decision-making skills mentioned above. Accordingly, it seems, consequentialism recommends that people should not believe and adopt itself—it ‘effaces’ itself in this way. Some think that this is a problem for a theory, for example because they think that acceptable theories should satisfy a publicity constraint. Lindauer’s point is that what X-Phi can tell us about motivation figures centrally in the self-effacingness objection. I disagree. This is why. Those objecting to consequentialism because it is self-effacing do not claim, and need not claim, that the doctrine has the indicated consequences—consequences that are non-optimal from its own perspective. Indeed, such a claim would be irrelevant to their objection. The objector does not even claim, or need to claim, etc., that the consequentialist doctrine realistically has such consequences. What she claims is that if circumstances and psychological facts about motivation, facts about people’s limited rationality, etc. were such that if people internalized the consequentialist doctrine this would have undesirable consequences from the point of view of the doctrine itself, then the doctrine would seem to suggest that people forgot about it and acted in accordance with another; and this implication of the doctrine in the indicated hypothetical circumstances is what makes consequentialism problematic (in light of plausible publicity constraints or similar such constraints).

The point I want to make here is that the self-effacingness objection does not rely on motivational facts of the kind Lindauer thinks it does—empirical claims of the kind that X-Phi is apt at analyzing. As noted, the objector does not claim that the doctrine in question has certain consequences. She asks what follows if it had certain consequences. That is, the argument in question is conditional in this way. Something similar is true of at least other important objections to consequentialism (see, e.g., Cohen 2008, pp. 263–268).

The upshot of this criticism of Lindauer’s interpretation of the self-effacingness objection to consequentialism is that sometimes X-Phi is to no avail with respect to important arguments and analyses of political philosophers. This does not of course detract from its relevance in other respects including those mentioned above. Yet the criticism suggests that X-Phi is one tool among others of the political philosopher. I shall make a similar point in my assessment of the third view (and for that matter also with respect to the fourth view) to which we turn now—a view that is fairly imperialistic regarding the role it envisages for X-Phi in political philosophy.

The Unearthing Principles View

As indicated, the view now to be examined is very ambitious in the sense that the role it sets out for X-Phi with respect to theory acceptance is pivotal. Indeed, according to this view X-Phi is what political philosophy is about or ought to be about, namely uncovering fundamental principles that organize our more concrete moral judgments and doing so in the most reliable and systematic way available to us (i.e., by conducting experiments). In presenting this view I take my cue from a perspicuous statement of it by Kahane (2013). This third view presupposes the first view I presented above in the sense that the former assumes that ‘our intuitions are largely trustworthy’ (Kahane 2013, p. 422. Cf. p. 428). That is, it presupposes that we can filter our intuitions in such a way that the set we are left with consists of intuitions that we have reasons to rely upon (I presume here that the view that largely trustworthy intuitions are readily available to us, that is, that filtering is not needed, is implausible.)

The unearthing principles view under consideration now is interesting in part because it can be seen as continuous with influential philosophical accounts of our key endeavor in political philosophy. In Rawls’s statement of the method of reflective equilibrium he emphasizes that although we are in the first instance focusing on ‘one (educated) person’s sense of justice’ (on the supposition that ‘everyone has in himself the whole form of a moral conception’) we do so while assuming that when this person works out his conception of justice, it will be approximately similar to the conception other individuals would reach if they followed the same procedure. Should people’s conceptions turn out to differ in the end, this, and the way in which they do, would, as Rawls emphasizes, be ‘a matter of first importance’ (Rawls 1999, p. 44). Accordingly, it would seem that on this understanding, our prominent aim in moral and political philosophy is to achieve a systematic grip of the principles underlying our more concrete judgments in a number of contexts. And—one may rhetorically ask—what better way (or what alternative way is there) to achieve such an understanding than by conducting experiments. X-Phi, one might say, is not as such an alternative to the method of reflective equilibrium, it is the machinery that allows us to achieve or approximate such an equilibrium.

Similarly to Rawls’s interest in principles that underlie our ordinary judgments, Immanuel Kant, in his account for his key moral notion of a will that is good in itself, says: ‘This concept [the concept of a will estimable in itself and good without regard to any further end] already dwells in the natural sound understanding and needs not so much to be taught as merely to be elucidated’ (Kant, 1785/1993, p. 9 [397]). Recently, scholars such as Francis Kamm—Kahane’s paradigmatic example—has formulated views in continuation of Rawls’s and Kant’s in this regard (Kahane 2013, p. 423), that is, she has implied that the task of the moral and political philosopher is to elucidate or unearth deep structures in our moral intuitions. Against this backdrop, and the obvious role for experimental method it leaves, Kahane sets out his central argument (Kahane 2013, pp. 428–429):

  1. (1)

    ‘Our moral intuitions about particular cases give us defeasible reason to believe in their contents.’

This is the presumption of trustworthiness which I mentioned above and the implied conditions for such. Yet it is clear that many, including proponents of the negative program in X-Phi, would have much to say concerning this premise. However, I follow Kahane in setting such concerns aside here.

  1. (2)

    ‘Our moral intuitions about particular cases track certain moral principles.’Footnote 10

According to our moral intuitions it is, for example, wrongfully discriminatory if an owner of a bar puts up a sign saying that ‘we do not serve Roma’ (Altman 2020, Sect. 2.1). Yet it must be wrong by virtue of something. It is wrong by virtue of certain underlying moral principles and the properties they cite (Kahane 2013, p. 428; McMahan 2013, pp. 111–112). If our intuitions are broadly reliable, they track the properties in question. If we identify those properties, we can unearth the principles pertaining to them.

From 1 to 2:

  1. (3)

    ‘Evidence about what moral principles our intuitions track gives us defeasible reason to believe in these moral principles.’

But:

  1. (4)

    ‘Facts about what principles our intuitions track are empirical facts, and are therefore discoverable using the methods of empirical psychology.’

Therefore:

  1. (5)

    ‘Psychological evidence about the principles our intuitions track gives us defeasible reasons to endorse these moral principles. [From 3,4].’

As Kahane points out, the properties or factors that our intuitions track (and the underlying principles) are often initially opaque to us. We have firm intuitions, for example, that the mentioned bar owner acts in a way that is wrongfully discriminatory. Yet exactly why or by virtue of what might not at first be transparent to us (Kahane 2013, p. 432). When trying to locate the factors at play, philosophers can, Kahane submits, be seen to make testable ‘empirical predictions’ about what factors are at play in our intuitions, using various cases to distinguish potential candidate factors and ascertain their relative strength. X-Phi does this in the most systematic and reliable way available to us (Kahane 2013, pp. 432–433). A nice and compelling example of this is provided in Joshua Greene’s Moral Tribes where he recounts the search for the factor(s) underlying our moral intuitions in key so-called Trolley cases (Greene 2015, pp. 211–224). One of the interesting findings that have played an important role in discussions within especially the negative program mentioned above, is that apparently our strong resistance against using an innocent person as a means to save, for example, five other innocent persons has more to do with an aversion to inflicting harm on another person in a face-to-face contact situation than it reflects the kind of deontological constraints philosophers discuss (e.g., the prohibition against using other human beings as mere means).

When one suggests, as Kahane does, that ethical inquiry at its best is essentially empirical and experimental it may seem as if one proposes that we base our moral or ‘ought’ claims on empirical or ‘is’ claims (Kahane 2013, p. 437). As Kahane explains this is not the case. What we do is this. We begin with intuitions that we endow with some defeasible credence—with particular ‘oughts’ as Kahane says (Kahane 2013, p. 437). We unearth the principles—the general ‘oughts’—in these particular ‘oughts’. The credence of the latter carries over to the former. Accordingly, what we do using empirical methods is not to derive normative claims from empirical claims, instead we discover systematic connections between our particular and general normative judgments.

While Kahane’s argument is ingenious, he seems to me to overstate the role of experimental empirical research for theory acceptance in political theory. That is, I deny that his argument shows that more traditional philosophical inquiry should be ‘pushed to the corner’ (Kahane 2013, p. 422). What ‘pushed to the corner’ means is of course opaque and makes it difficult to say whether one agrees with Kahane’s view or not. Perhaps we may here distinguish between the following three views: (i) X-Phi as a sub-discipline of political philosophy; (ii) X-Phi as a supplement to more traditional ways of doing political philosophy (Weinberg 2016, p. 72); and (iii) X-Phi as supplanting non-empirical approaches to political philosophy. I take it that of these three views Kahane would embrace (iii), whereas I, for the reasons to follow (and for those given above), lean toward (ii).

First, as I pointed out in the discussion of the fruitfulness view, there are important philosophical arguments and tests that do not require the use of empirical methods—indeed such methods would not be pertinent to the (important) matter at hand.

Second—and more directly in response to the unearthing view as set out in this section—in my view unearthing the principles underlying folk intuitions only gets us so far. In various valuable ways philosophers go beyond what even the most precise and systematic description of those intuitions may reveal. They do so by, for example, displaying to us what we apparently are committed to believing given certain beliefs that we would be hard pressed to give up. ‘We’ refers to all of us, or members of the relevant moral community. Moral and political philosophy is, as Joel Feinberg puts it, ad hominem in that it begins with the convictions we already hold (Feinberg 1984, p. 18). Part of the philosophical discipline consists then in clarifying the beliefs and moral sentiments in question (drawing key distinctions etc.) and working out their implications in a consistent and bold manner. Further, philosophers uncover potential tensions in our beliefs and sentiments and explore various (sometimes surprising and provocative) paths for making them coherent. And it is contributions of this kind that I question would be forthcoming from the application of the tools of X-Phi to our commonly shared moral judgments. Political philosophy and X-Phi are indeed both ad hominem in a sense. Yet what the former is capable of drawing out from this shared basis differs from what the latter brings out. Further, the philosophical contribution alluded to (and exemplified below) may in part work on the basis of X-Phi analyses (i.e., on the basis of a systematic picture of what we presently believe in the moral department).

Here are a few prominent examples of the philosophical work I have in mind. Peter Singer has compellingly shown that a very plausible commitment to a sort of Samaritan duty—i.e., a duty to help others in need as long as this does not involve a significant risk to us (or a harm comparable to the harm the person in need would incur in the absence of our assistance)—would on realistic assumptions commit us to believing that we have obligations to assist distant strangers in need (Cullity 2004; Singer 1972). G.A. Cohen’s incentives argument is another apt example. As Cohen notes, initially a commitment to helping those who are disadvantaged in society seems to support a policy that allows material incentives (i.e., superior remuneration) to highly productive individuals, that is, to ‘the talented’. Given such incentives, the talented may be maximally productive, and part of that productivity, or the fruits of it, could be transferred to members of the worst-off group in society. However, if the talented are—as they are assumed to be in a Rawlsian ‘well-ordered’ society—genuinely committed to the plight of the worst off, they can reasonably be taken to task for making superior material remuneration a precondition for working effectively. The situation of the worst off could be improved more if the talented did not demand such incentives. Thus, the talented ought to relinquish incentives. In this way, an important and initially compelling argument for inequality is refuted (Cohen 1991, 2008, Chap. 1).

These examples are pertinent here in that they seem to support my supposition that such progress in our moral sentiments—or at least such challenges to our present moral horizon—would not have been forthcoming had political philosophers solely applied X-Phi tools. Of course, progress of this sort could in principle have materialized as a result of empirically oriented philosophers making bold conjectures or predictions of the kind Kahane allows and via shrewdly designed experiments (cf. Handby 2022, pp. 1602–1603). Yet I doubt it. I do so in part because respondents are less likely than professional philosophers are to be susceptible to entertaining the possibility that convictions we hold dear might commit us to beliefs that seem very radical, even odd, to us, and in light of such dis-equilibrium in our moral beliefs to be willing to explore various paths toward making them coherent. Philosophers have by training an openness to such considerations—indeed they seek out puzzles and tensions of this sort—and a related audacity to contemplate widening our moral horizon. So the present reason for being skeptical concerning Kahane’s suggestion that traditional philosophical methodology should be supplanted by X-Phi is that aspects of the former are uniquely suited to reshuffle our moral intuitions in a coherent and edifying manner.

Third, I think that there are some inquiries, including specific experimental set-ups that seem to depend on auxiliary assumptions and a theoretical background the application of which requires philosophical training. A case in point is the so-called ‘Subtraction Test’ recently proposed by Kolodny (2023, p. 37). It is used for ascertaining whether a certain ‘relation of rule’ that characterizes the state and its relation to citizens is what generates our complaint against the state—a feature that calls for justification. If we, in imagination, remove the candidate relation yet find that our complaint is intact we need to look for another candidate for our complaint. For example, the state’s use of force to impose deterrents on citizens is a popular candidate for our key complaint. Yet consider Kolodny’s ‘Myth of the Omittites’ in which a ruler might omit lowering a rope into pits that each citizen out of necessity must climb into every day for some time and need to get up from again (Kolodny 2023, p. 41). No force, yet it seems that a claim against the state remains and hence that ‘force’ is not the feature of the state that can account for our complaint against the state. Applying this test would seem to me to presuppose a prior understanding of the nature of various relations of rule and the possession of a clear grasp of the relevant complaints we may have against the state—notions that lay persons for good reasons cannot be expected to have. At the same time, the inquiry or test in question seems clearly relevant to locate the relation of rule that triggers our complaint against the state. Again ‘our’ here refers to our commonly shared moral judgments or how they look when they have been worked out in the most consistent and plausible manner.

The two last reasons that I have set out face the following objection.Footnote 11 They both propose that political philosophers have some unique capacities acquired through their training—capacities that would seem to make them capable of contributing to moral inquiries in a way that is independent of what is achievable via the methods of X-Phi. Yet, the objector points out, this need not offer a reason to avoid X-Phi, rather it would seem to be a reason to use only highly trained philosophers as respondents. True, as mentioned above, such respondents are also subject to various sources of errors. However, there are, as we have seen, ways to avoid such errors.

In response, what might not be captured by even such a sophisticated form of X-Phi is philosophical dialogue or debate, that is, the careful canvassing of concepts, principles, and intuitions undertaken by highly skilled practitioners in the ‘normal science’ of contemporary political philosophy (cf. Kauppinen 2014). Now such dialogue may of course in part be simulated by the use of X-Phi methods, for example recurrent experimental testing of expert reactions to variation on the trolley dilemma akin to those undertaken by Joshua Greene where the respondents were non-experts. Still there would seem to be something categorically distinct about philosophical dialogue vis-à-vis empirical studies of this sort. For example, whereas X-Phi studies approach respondents (whether folks or philosophers) from the ‘objective’ perspective trying to decipher from the outside their underlying principles, a dialogical approach is an I–thou exchange between interlocutors who can argue with each other, seek to persuade each other, and so on and so forth (cf. Strawson 2020). It is at best an open question whether studies of the latter sort can be jettisoned without serious epistemic (and related practical) costs. I persist, therefore, in the judgment that the third view scrutinized in this section goes too far in its case for X-Phi and its role vis-à-vis more traditional philosophical methods.

Before moving on to the fourth account I shall comment in brief on a view related to Kahane’s stance that Lippert-Rasmussen puts forward in this special issue (Lippert-Rasmussen 2024). Lippert-Rasmussen usefully distinguishes between two ways in which authors working within the traditional method of reflective equilibrium may react to findings in X-Phi: First, ‘the supplementary stance’, according to which ‘political philosophers should supplement their reflective equilibrium approach with an experimental approach’, and second, ‘the radical stance’, suggesting, in contrast, that ‘they [political philosophers] should reject reflective equilibrium, at least in the form we know it, and adopt an approach much more informed by experimental philosophy in its stead’ (Lippert-Rasmussen 2024, p. 4). From what I have said it should be clear that my view is of the former sort, whereas Kahane’s falls within the latter category; and Lippert-Rasmussen’s, as indicated, is like Kahane’s in this regard (although their arguments differ). Taking my cue from one of Lippert-Rasmussen’s own points about social mechanisms that might affect how philosophers position themselves (such as a mechanism of polarization because a premium is put on novelty) I might point out that there is a pull toward taking radical stances such as the one Kahane and Lippert-Rasmussen adopt (i.e., ‘reject the traditional way of doing philosophy’) instead of a more moderate view such as mine (i.e., ‘reflective equilibrium is more or less right yet can benefit from the use of X-Phi tools and results’); and a pull that does not necessarily reflect that the radical view is more correct than less radical views. Yet I shall refrain from drawing any advantage from this point in that I do not share Lippert-Rasmussen’s view that attention to mechanisms of this sort should take center stage in the work of political philosophers. That is, while I think that it is clearly relevant for political philosophers to be attentive to mechanisms of the indicated sort, I do not think that such attention should take the form of overturning the traditional reflective equilibrium style way of doing political philosophy. Instead, I suggest that it should take the form of what I associated with the De-Biasing View mentioned above.

Let us consider briefly Lippert-Rasmussen’s case for the radical stance based on his (very illuminating) observations regarding the philosophical discussion of the question of equality’s putative intrinsic value and the related leveling-down objection to equality. Lippert-Rasmussen points out that, as per X-Phi insights regarding the psychological mechanism of loss aversion, we are less inclined to regard equality as intrinsically good when equality is framed as a matter of imposing a loss on better-off people than we are if equality is framed as a matter of missing out on an opportunity to gain. Relatedly, Lippert-Rasmussen submits, philosophers on either side of the debate tend to frame the question in the way that is most conducive to their favored view. Yet instead of this adding genuine support to the view that they respectively favor, the apparent support hinges upon the psychological mechanism of loss aversion. Therefore, Lippert-Rasmussen suggests that philosophers debating the intrinsic value of equality ought to pay more attention to empirical work regarding the mentioned mechanism (and to potentially other relevant mechanisms). I agree. Yet I fail to see how this motivates jettisoning the traditional method of reflective equilibrium. To illustrate. Those skeptical of the intrinsic value of equality (e.g., Parfit) could say to proponents (e.g., Cohen) that they ought to consider ‘loss’ cases too (and vice versa of course). Further they could maintain (as, for example, Rawls did) that even when we consider ‘gain’ cases, the intrinsic value of equality remains doubtful (Rawls 1999, pp. 130–131). Such moves in the debate can be crucial. For example, Casal’s point against sufficientarians that they ought to consider cases of loss above a given sufficiency threshold and not only cases in which one or the other group gains (or for that matter cases in which both groups gain) seems to me effective in suggesting that the ‘no care’ attitude to inequalities above the threshold is untenable (Casal 2007, p. 311). My general point is that it is difficult to see why insights regarding psychological (and other) mechanisms should motivate a fundamental change of methodology rather than a reflective equilibrium process informed by the information in question.

A last point regarding Lippert-Rasmussen’s argument (many aspects of which I cannot do justice to here) is this. I think that there is a sense in which it risks putting the cart before the horse. It may do so because it is unclear that debates such as the one about equality’s intrinsic value that Lippert-Rasmussen focuses on—debates that he proposes should be addressed in a much more empirical fashion than practitioners of reflective equilibrium do—would even have gotten off the ground without philosophical explorations of the bold and in part non-empirical form canvassed above conducted under the auspices of the traditional reflective equilibrium method.

The three responses to Kahane’s view I made above, together with my response to Lippert-Rasmussen’s related view add, I think, at least indirect support to my favored view that X-Phi is best seen as a supplement to the traditional reflective equilibrium method.

The Defeasible Reason (or Squatters’ Rights) View

A considerably more modest view than the Unearthing Principles View is this. Many philosophers appeal to coherence with common sense or ordinary morality as constituting a reason in favor of a certain philosophical theory or concept. This is familiar from, for example, debates between proponents of deontological theories and proponents of consequentialist theories, where the former emphasize the virtue of their view in this regard. The present view on the relevance of X-Phi to theory acceptance takes as its point of departure the indicated desideratum for philosophical theories. It does so in the form of regarding it as a defeasible reason in favor of a given theory if it coheres better with folk intuitions than relevant alternative theories do. Consider:

‘On this view, philosophical theories that most closely accord with and account for ordinary beliefs and practices should enjoy “squatters’ rights” until they are shown to be defective for other reasons.’ (Nadelhoffer and Nahmias 2007, p. 126) X-Phi is then an apt way of testing in a systematic way claims of (in)coherence of the indicated sort—claims that have traditionally been made in a more hand-wavering manner. An example of a study that falls within this view is Albertsen et al. (2023). Against the background of the indicated presumptions, they compare two leading theories of the wrongness of discrimination, to wit a respect-based view and a harm-based view (finding some support for a pluralist view on the wrongness of discrimination).

Three points about this view. First, while it is well-documented that philosophers make claims of the indicated sort—i.e., claims about the relevance of their theories being coherent with folk intuitions—the claims in question strike me as more vague and open to interpretation than the ‘defeasible reason’ interpretation would seem to suggest. Second, even given clear statements of this sort we would need a ground for thinking that folk intuitions should play the role suggested by the view under consideration. Is this to be found in one of the other accounts mentioned above or might one point to a fourth grounding of the relevance of folk intuitions for theory acceptance? Third, it is important to remember that compatible with the view in question there may be reasons in concrete contexts to discount folk intuitions altogether. For example, philosophers might put forward a persuasive error theory for folk intuitions being as they are (cf. Enoch 2016, pp. 40–43). I make this point not as an objection to the view, just as a reminder about its modesty.

Conclusion

My question in this paper concerned the relevance of X-Phi to theory acceptance in political philosophy. I addressed this question by canvassing four views on this relevance, to wit: (i) The De-Biasing View; (ii) The Fruitfulness View; (iii) The Unearthing Principles View; and (iv) The Defeasible Reason (or Squatters’ Rights) View. View (i) implied that X-Phi is useful with regard to sifting intuitions such that we avoid a range of errors or distortions. I wholeheartedly enjoined this view and showed that it was continuous with paradigmatic examples of traditional philosophical views. View (ii) pointed to X-Phi’s merits in terms of facilitating a thorough examination of the fruitfulness of our concepts. I argued that with respect to some, consensus-oriented approaches to political philosophy the view in question emphasizes tools in X-Phi that are of obvious relevance for theory acceptance. With respect to a different group of theories that emphasize ultimate principles, the tools in question are not of relevance to principles of this sort. Yet even such theories include so-called regulative principles to which the tools in question are of relevance. View (iii) hailed X-Phi on the ground that empirical experimental method is an apt tool—indeed the best available tool—for connecting concrete moral judgments with underlying moral principles which, according to this view, is or ought to be philosophers’ central aim. I expressed skepticism regarding the suggestion that X-Phi exhausts the relevant toolbox of political philosophers. For example, expert dialogue appears to be an invaluable source of important insights. Finally, view (iv) presented a more modest and plausible account of the relevance of X-Phi, to wit, that it may serve to test the (in)coherence of certain theories with folk intuitions. All in all, the message of the paper is that X-Phi is a valuable addition to the existing tools of political philosophy and not a threat to standard methods. To the extent that it is, as in the case of the third view I discussed, implying that traditional methods must give way, I proposed that this overstates the relevance of X-Phi and would mean a loss of important (non-empirical) tools of political philosophy and related occasions for widening our moral horizon.