How (not) to integrate scientific and moral realism

Abstract

In this essay, I seek to clarify and defend a unified account of realism, i.e. a conception of realism that does not only apply to philosophy of science, but also acknowledges how realism is understood in other philosophical disciplines—particularly, how moral realism is treated in metaethics. I will argue that integrating scientific and moral realism is less straightforward than is commonly assumed, due to several substantial, but often unnoticed disanalogies that obtain between both views. As a consequence, scientific realists should consider modifying their traditional understanding and move towards an alternative conception of realism—one that is much more in line with the conception that moral realists usually adopt. Realism (about science or ethics) is, in the final analysis, best characterised as an alethic view which restricts itself to an idea about the objectivity of truth, rather than an epistemic view which underwrites more extensive theses concerning the accessibility of this truth for human knowledge.

1 Introduction

The debate on scientific realism is one of the most prominent topics in general philosophy science. While it is difficult to determine exactly when discussions of the subject first started,¹ it is clear that the debate has been around for a significant period of time and has become exceedingly sophisticated over the past decades. In the literature, several versions of scientific realism (as well as anti-realism) have been discussed and an ever-growing number of arguments that either support or criticise a realist outlook on science has been introduced into the debate. As a consequence, understanding the debate on scientific realism requires close attention to its manifold details—so much attention, indeed, that one may easily lose sight of the fact that scientific realism is but one realist view among others, i.e. that philosophers of science share the concept of realism with philosophers working in other fields.

The debate on moral realism is, by analogy, one of the most prominent issues in modern metaethics. As is true of scientific realism, the debate on moral realism can be traced back at least several decades² and is likewise said to have gained complexity, nuances, and depth. Many different versions of moral realism (as well as anti-realism) exist and, even in the most recent contributions, novel arguments keep being introduced into the metaethical debate.³ Not surprisingly, scholars working in metaethics are rather preoccupied with their own controversies and usually remain within the conceptual boundaries of their specific discussions, without paying too much attention to conceptions of realism in other philosophical fields.

In this essay, I would like to establish a broader perspective on realism—one that allows us to situate the debates on scientific and moral realism within a more general conceptual framework and to restore the hidden connection between both views.

My main focus will be on the conceptualisation of realism, not on its justification. I am not going to argue for scientific and moral realism, or against their anti-realist alternatives. Instead, my main objective is to provide clarification and to elucidate what realism is (supposed to be) about, i.e. which claims we are required to accept in order to qualify as a realist, and how we should conceptualise realism as a philosophical view. The main question I wish to answer in this context may be phrased as follows: Can we identify a universal understanding of realism that is applicable to both philosophy of science and metaethics? Do philosophers from both disciplines share a common understanding of what realism genuinely says? Is it possible to integrate scientific and moral realism by drawing attention to some overarching themes that are present in both debates? In short, my answer to these questions is: yes. There is indeed a universal understanding of realism—one that can be applied to both domains in question and informs us about what we should regard as the core idea that characterises scientific and moral realism alike.

Nevertheless, as is commonly the case in philosophy, things are slightly more complicated than we would like to think. In particular, I would like to argue that integrating scientific and moral realism requires us to revoke a presentation of realism that is exceptionally popular in the literature. Although this presentation is prima facie plausible and has gained its popularity because it does characterise realism quite elegantly, it can be shown to be rather dysfunctional when we turn our attention to the relationship between scientific and moral realism—this is my descriptive claim. As a consequence, I would like to recommend scientific realists to consider a modification of their traditional understanding, because it turns out to be their current self-image that, at least, appears to be flawed, in need of correction, and ultimately responsible for their inability to preserve the analogy to moral realism—this is my normative claim.

2 The standard account

Let me begin my discussion by drawing attention to what I would like to call the standard account. The standard account, which I take to be the received view in the present context, provides us with a generic definition of realism, i.e. it addresses the question of what realism amounts to in various fields of application, including both science and ethics. In doing so, the standard account provides a first proposal for integrating scientific and moral realism into some more abstract scheme: It proposes a unified account of realism that philosophers from different sub-disciplines, such as philosophy of science and metaethics, are supposed to share, enabling them to point to conceptual analogies with respect to their fields of application.⁴

According to the standard account, both realist views exemplify, then, the same general idea, the same generic realism, with respect to two different domains. Scientific realism specifies the general idea of realism in the theoretical domain of science, while moral realism exemplifies this very same idea in the practical domain of ethics. Strictly speaking, the standard account identifies three components that both views appear to involve—a semantic, a metaphysical, and an epistemological component:

• First, scientific and moral realists take the discourses of their respective fields at face value, interpreting scientific statements or moral judgments as truth-apt in a literal way, i.e. as having a propositional content that is either literally true or literally false. For instance, in philosophy of science, realists typically reject instrumentalist approaches that, by their very definition, refrain from ascribing truth-values to theoretical statements about unobservable entities.⁵ Similarly, in metaethics, realists are opposed to non-cognitivist approaches that consider moral judgments to be mere expressions of the speaker’s attitudes that, again, do not qualify as either true or false.⁶

• Second, scientific and moral realists commit themselves to the idea of mind-independence. Accordingly, scientific realists believe that whether a scientific statement is true does not depend on anybody’s believing the statement. In a similar vein, moral realists insist that the truth of moral judgments obtains independently of anybody’s actual or hypothetical beliefs. Rather, scientific statements and moral judgments are made true by adequately representing objective facts of reality.⁷

• Third, it seems to be the case that both realisms share some epistemic optimism concerning the success of our attempts to gain scientific or ethical knowledge. On the one hand, scientific realists often hold that we have good reasons to believe that our most successful scientific theories are (at least approximately) true; as a result, scientific realists are committed to oppose agnostic accounts, such as constructive empiricism.⁸ On the other hand, moral realists are frequently said to endorse a success theory that seems to entail the claim that we are (at least sometimes) successful in recognising true moral judgments; consequently, moral realists typically commit themselves to reject an error theory.⁹

The standard account thus claims to ultimately provide us with an exhaustive and overarching definition of realism: First, the three components can be treated as a set of conditions, each necessary and together sufficient, for any philosophical view qualifying as a realist position, covering the most prominent philosophical issues of meaning (semantic), ontology (metaphysical) and knowledge (epistemological). A philosophical view that accepts all three theses listed above is by definition a realism, and in this sense the standard account gives us sufficient conditions for characterising realism; moreover, the three components are necessary in that any philosophical view that rejects just a single one of them already fully counts as an anti-realism. Second, since this set of conditions is so easily applicable to different areas, the standard account does not only facilitate the orientation within a particular debate on realism (not least by allowing for an easy and clear classification of the most important anti-realist positions, depending on which of the three components they negate), but also helps us to draw connections between different realism debates (by enabling us to compare distinct realist views with respect to the three components). This must be why the standard account is so popular in the literature.

In the remainder of this essay, however, I would like to argue that this presentation of realism, though seductive and common, is deeply misleading and ultimately false, because it neglects and obscures certain substantial disanalogies between scientific realism (as this view is understood in philosophy of science) and moral realism (as this view is treated in metaethics). These disanalogies are, as we are about to see, closely interconnected and result from what may be called an epistemic deformation of realism—i.e. a nearly (but not entirely) unnoticed tendency in the literature to exaggerate the epistemic commitments inherent to the realist thinking. Since this tendency is much more pronounced in philosophy of science than it is in metaethics, the respective positions of scientific and moral realists are, indeed, more alienated from each other than is commonly assumed (descriptive claim). Based on this observation, I would like to recommend philosophers of science to consider a modification of their usual notion of scientific realism and encourage them to explore an alternative to their typical self-conception (normative claim).

In the following sections, I will develop this line of thought by, first, drawing attention to those disanalogies that are ignored, neglected, and obscured by the standard account (Sect. 3–5). Second, I will seek to shed some light on how these disanalogies fit into a bigger picture, how they are connected to each other, and how they ultimately lead us to the idea of an epistemic deformation that realism is subjected to in the modern literature (Sect. 6). Finally, I would like to present what I take to be a fruitful alternative—an alternative understanding of realism that is much more suited to avoid the discussed disanalogies in the first place and to counteract the described tendency of exaggerating epistemic commitments, namely by focusing on the alethic core idea that is (in my opinion) truly characteristic of realism (Sect. 7).

3 The first disanalogy: neutrality

The first disanalogy between scientific and moral realism can be stressed in terms of neutrality.

In metaethics, it is common to consider moral realism a second-order view that is neutral regarding first-order normative-ethical stances. According to its metaphysical core thesis, moral realism claims that moral truth is objective, or that the truth of moral judgments obtains mind-independently, i.e. irrespectively of anyone—individual or group—choosing to accept the judgment or not. However, moral realism does not entail which particular moral judgments are true. For instance, moral realism is compatible with both Kantian ethics and utilitarianism,¹⁰ it does not take a stance on whether abortion is right or wrong, and it does not imply that liberal democracies are morally superior to totalitarian dictatorships. Moral realism is a purely second-order view that is compatible with each of these first-order views. Of course, every moral realist adopts certain normative-ethical views and holds certain opinions on, say, whether it is morally permissible to tell a harmless lie in a specific emergency. But these normative-ethical commitments are not part of their metaethical conception of realism. You may believe that it is sometimes permissible to tell a white lie; or you may believe that it is never permissible to do so. As long as you believe that the truth-values of each of these moral judgments—whatever they may be—are objectively determined by the moral facts of reality, you qualify as a moral realist.¹¹

In philosophy of science, by contrast, scientific realism is conceptualised in a way that is decidedly not neutral regarding first-order scientific theories. Scientific realists take our current theories to be approximately true: they believe in electrons, but not in phlogiston; they believe in gravitational waves, but not in the electromagnetic aether; they believe in evolution, but not in intelligent design (see Psillos, 1999, p. xvii; Devitt, 1997, pp. 23f, 46f, 109; Hacking, 1983, pp. 265, 274). Needless to say, these scientific statements are fairly plausible due to the available evidence and the overwhelming consensus among the scientific community. However, even if we assume that scientific realists’ stances on each of these scientific issues are plausible, these first-order implications of scientific realism nevertheless establish an undeniable disanalogy to moral realism. The reason for this is that these first-order commitments are treated as inherent parts of their conception of realism, not as additional supplements. Simply believing in the objectivity of scientific truth—whatever it turns out to be—is not enough for realists in philosophy of science. In order to qualify as a scientific realist properly, you must believe in the approximate truth of general relativity theory. And this begets the first disanalogy to moral realism.

This is a problem for the standard account. It raises questions like: Where does this disanalogy come from? How might we account for it? Is it possible to correct it somehow? The problem is that not only does the standard account not answer these questions; rather, by defining realism via the three components, the standard account simply presupposes that both realist views are conceptualised in a parallel way and thereby obscures the fact that there might be any disanalogy between scientific and moral realism at all. It treats scientific and moral realists as if they had an analogous understanding of what realism is supposed to be about. However, the disanalogy undoubtedly reveals that this is not the case and that the standard account thus cannot be quite right. (In the following sections, I will discuss in more detail what exactly the standard account gets wrong.)

Before we turn to the other disanalogies, let me briefly discuss why I think that though this first disanalogy may not immediately call for a modification of either scientific or moral realism, the comparison with moral realism, nevertheless, does indicate that there may be something wrong with the way scientific realism is commonly conceptualised. To put it straightforwardly, the non-neutrality of scientific realism is, when you think about it, definitely strange. Scientific realists’ stance towards scientific theories seems to promote a weird confusion of philosophical and scientific questions that tacitly undermines the status of scientific realism as a second-order view about science. Shouldn’t we expect scientific realism to be merely concerned with philosophical questions about the aim of scientific theorising or the concept of truth employed in scientific reasoning? Instead, we find scientific realists endlessly fighting over the question of whether we have good reasons to believe that Bohr’s atomic model is approximately true. These first-order implications of scientific realism regarding which theories you must—qua realism—accept as (at least approximately) true are so extensive that one may wonder whether the view is still best characterised as a philosophical view at all. Rather, scientific realism turns out to be a hybrid position created from both philosophical and scientific commitments.

As a result, scientific realists appear tempted to usurp the authority to decide scientific discussions on their own. In the literature, van Fraassen has pointed out this problem lucidly:

“I find this very puzzling […]. Is it appropriate for a philosophical position to include answers to the sort of questions that scientists investigate? […] Could such a belief [in a specific scientific theory] really be a matter of philosophical debate, or of metaphysical argument?” (Chakravartty & van Fraassen, 2018, pp. 12f)

Van Fraassen’s answer leaves no doubt that he thinks it can never be appropriate for philosophers to decide these kinds of questions. However, since scientific realists deliberately do choose to include such answers to first-order scientific questions into their understanding of realism, he soon reaches a compelling verdict: “the typical scientific realist self-image is confused” (ibid., p. 17). The alternative conceptualisation van Fraassen suggests is aimed at putting an end to this latent confusion. According to van Fraassen, scientific realism should be understood as a purely second-order view that abstains from scientific debates completely. In particular, he proposes a conception of scientific realism which concedes that theoretical truth is the proper aim of scientific research but does not speculate on whether we have already reached this goal in certain areas.

“Could someone be a scientific realist and not have such beliefs to the effect that certain unobservable entities are real, or that certain theories […] are actually true? The answer is yes, on my understanding of scientific realism as the view that the aim, the criterion of success in science is to arrive at true theories, rather than merely empirically adequate ones. This has no implication for whether that criterion is met in any particular case, or whether even our best theories today are successful by that criterion.” (ibid., p. 17, emphasis in the original)

What this brief discussion is supposed to achieve is to prepare us for the (presumably) more controversial suggestion of this essay: namely, that this first disanalogy (as well as the disanalogies we are about to observe) may not just be a manifestation of merely different priorities that scientific and moral realists have, but could rather be taken to result from a deficient conception of scientific realism that is in need of correction. This is, again, the normative claim of the essay. In this context, then, we should, at least, consider admitting that van Fraassen may be right in this respect and that a modification of our prevalent conception of scientific realism is worth thinking about (for such an approach, see also Godfrey-Smith, 2003, pp. 173ff).

To be clear, van Fraassen himself does not argue for scientific realism. He is, of course, one of the most prominent opponents of realist philosophy of science and does not believe that theoretical truth is the proper aim of science (see van Fraassen, 1980, p. 12). This is why van Fraassen’s insight here merely concerns the conceptualisation of realism, i.e. the question of how we should think of it as a philosophical view. Scientific realists may (and, indeed, should) consider agreeing with him on that—most importantly, because such a revision could sharpen their notion of scientific realism as a philosophical view that desists from creating confusion with its various scientific commitments. Moreover, it might also reconcile scientific realism with its metaethical counterpart, moral realism. And although van Fraassen may not primarily call for such an adjustment because of its potential to re-establish the analogy to moral realism, the fact that it does allow for such an integration of scientific and moral realism should, at least, reassure us that it is worth considering. In order to be able to fully understand what such an integration of scientific and moral realism might look like, we must, however, first turn our attention to the two remaining disanalogies between both views.

4 The second disanalogy: the epistemic commitment

The second disanalogy unfolds around the fact that a success theory, in contrast to optimism, is not an epistemic view.

In metaethics, the concept ‘success theory’ has been introduced as an antonym to the term ‘error theory’.¹² But whereas an error theory has an epistemic implication, this is not true of a success theory. To see this, consider an error theory first. At its centre, an error theory involves a metaphysical nihilism, according to which there are no moral truths; rather, every moral judgment is false, because the values we talk about do not exist. As a consequence, an error theory incorporates an epistemic scepticism, according to which we do not have any ethical knowledge. This association of nihilism and scepticism under the term ‘error theory’ is natural, only insofar as the epistemic claim is a direct implication of the metaphysical one: given that every moral judgment is false and we are always mistaken when we utter a moral judgment, our ethical beliefs can never qualify as knowledge. By contrast, in rejecting an error theory, success theorists adopt a metaphysical anti-nihilism, according to which there are (at least some) moral truths. According to a success theory, (at least some) moral judgments are true, because moral values, facts, or properties do exist in the universe. This metaphysical anti-nihilism does not imply, however, an epistemic anti-scepticism, because our cognitive abilities might be insufficient for knowing anything about these truths at all, because such truths could be epistemically inaccessible to imperfect creatures like us. Rather, the anti-nihilist claim that there are moral truths is compatible with both an anti-sceptical and a sceptical stance concerning the question of whether we can know anything about these moral truths. Consequently, success theorists (in general) and moral realists (in particular) unanimously insist that their views are purely metaphysical accounts about moral truths that have no epistemological implications for how we come to recognise such true propositions:

“Since [the moral realist] believes in moral truth (in contrast to nihilists) and since he does not consider this truth the product of subjective attitudes (in contrast to constructivists), the danger of scepticism – the view that doubts that our epistemic capacities suffice to make these truths accessible – is imminent. However, even if such an epistemic scepticism could not be prevented in the area of morality, it would not question moral realism – for, realism is characterised by the fact that it keeps the question of truth separated from the epistemic accessibility of truth.“ (Halbig, 2007, p. 296, my translation)

In philosophy of science, scientific realism is conceptualised in a way that differs immensely from moral realism in this respect. Scientific realism does indeed involve an epistemic optimism and incorporates an epistemic commitment that is pretty robust and, therefore, does not correspond to anything inherent to the conception of moral realism. In fact, scientific realism is a decidedly anti-sceptical project that targets even agnostic approaches vigorously; suggesting that realism might be compatible with scepticism (as moral realists clearly do) is more than unorthodox in philosophy of science.¹³ This second disanalogy becomes even more apparent, when it is noticed that optimism is often identified as the core thesis of scientific realism deserving the highest amount of attention. For instance, Psillos explicitly emphasises that his “main aim […] is to defend the epistemic optimism associated with scientific realism.” (Psillos, 2009, p. xiv) So, not only do scientific realists commit themselves to an epistemic optimism that has no equivalent thesis in metaethics, they even present these commitments as the most important parts of their view. By contrast, moral realists focus on the metaphysical idea of mind-independence and widely agree that their view is compatible with the claim that “any actual agent might fail to recognize a moral truth” (Shafer-Landau, 2003, p. 17; see also Sturgeon, 1986, p. 127).

This is yet another problem for the standard account. The standard account suggests rather straightforwardly that there is a robust parallel between epistemic optimism (understood as the idea that we have good reasons to believe in scientific progress) and a success theory (understood as the view that we are at least sometimes successful in grasping moral truths). This presumed parallel is usually indicated by projecting the optimistic commitments which scientific realists identify with onto moral realism, claiming that moral realism must be an “optimistic position” (Boyd, 1988, p. 202; see also Tarkian, 2004, p. 303) too, because “realists are not sceptics” (Sterelny & Fraser, 2017, p. 981).

Relatedly, the standard account assumes a corresponding analogy between the anti-realist approaches which realists are confronted with in their debates. In this vein, agnostic approaches in philosophy of science (like van Fraassen’s constructive empiricism) are often likened to sceptical approaches in metaethics (like Mackie’s error theory), suggesting that both views might be understood as error theories, applied to different domains:

“An anti-realist can perfectly well acknowledge that the disputed claims have a truth-value, and even that these truth-values depend on something external, while going on to say that none of the claims is true. An anti-realist might, in other words, advance an error theory. […] Constructive empiricists in the philosophy of science take a similar line about all claims concerning unobservable entities, arguing that while such claims may have truth-values, we have no good reason for thinking of any of them that they have the truth-value true.” (Sayre-McCord, 1988, pp. 10f, my emphasis)

“What all error theorists [including error theorists in philosophy of science] recognize is that granting cognitivism to a disputed discourse doesn’t by itself secure the legitimacy of its claims.” (Ibid., p. 11; see also Enoch, 2011, p. 114)

Nevertheless, neither of these alleged analogies holds. Here, the standard account most evidently leads us astray. First, a success theory focuses on a metaphysical commitment to moral truths, values, and facts, but does not imply an epistemic theory regarding ethical knowledge: more precisely, it involves anti-nihilism, not anti-scepticism. Thus, an epistemic optimism and a success theory are in no way equivalent theses. (To be sure, the term ‘success’ insinuates an epistemic meaning, but do not let yourself be confused by this terminology: Moral realists clarify unequivocally what they mean.)

Second, ethical scepticism, as exposed by Mackie, relies on nihilist presuppositions; it concludes that we can never have any ethical knowledge, simply because there are no moral facts in the first place. The argument from queerness Mackie employs targets the existence of moral values primarily and attacks our prospects of acquiring ethical knowledge merely indirectly.¹⁴ By contrast, scientific agnosticism, as suggested by van Fraassen, does not subscribe to any metaphysical assumptions. Van Fraassen does not formulate a similar argument from queerness which raises metaphysical concerns about the existence of electrons; instead, his agnostic attitude results from epistemological considerations alone.¹⁵ Therefore, constructive empiricism must not be treated as an error theory that is merely applied to philosophy of science. (Again, it is true that both accounts share an epistemic diffidence; but do not let it fool you: On the one hand, we face a sceptical account that is entirely founded on metaphysical assumptions (Mackie); on the other hand, we have an agnostic approach that involves no nihilist presuppositions at all (van Fraassen).)

Russ Shafer-Landau has pointed out this distinction in the metaethical context astutely:

“Moral scepticism is the view that […] no moral claims can be known or justifiably believed. Such scepticism sometimes originates in moral nihilism: if a domain of enquiry contains no truths, then it can yield no knowledge. But one can be sceptical about an area of enquiry even if one acknowledges that there are truths within it.” (Shafer-Landau, 2003, p. 231)

An error theory most certainly is such a case of an ethical scepticism which “originates in moral nihilism”; constructive empiricism, however, “acknowledges that there are truths” about unobservable entities, but chooses to remain agnostic about such statements. Hence the disanalogy.¹⁶

Before we move on to the last disanalogy, allow me to note once more that it is—again—the conceptualisation of scientific realism that appears to be strange and in need of modification. What makes scientific realism so strange in this context is that is tends to conflate not only philosophical and scientific issues (see Sect. 3), but also metaphysical and epistemological questions. Scientific realists attempt to answer two different questions simultaneously—namely, a question about truth (metaphysical component) and a question about our epistemic access to the truth (epistemic component). They want to know (i) what truth is and (ii) whether our cognitive abilities suffice to know anything about what truths there are (see for this differentiation Alston, 1996, pp. 86f). Shouldn’t we expect philosophers of science to tackle these issues one after another, in debates that are cleanly separated from each other? Instead, we face a realism debate in philosophy of science that notoriously blurs this essential line and presents itself as a controversy in which both issues appear intrinsically conjoined with each other. This is already weird enough, but it becomes even more noticeable, given the fact that scientific realists (as well as moral realists, by the way) typically adopt a non-epistemic conception of truth, i.e. a theory that is most decidedly dedicated to understanding truth in entirely non-epistemic terms. Proponents of such a conception of truth should not be expected to conflate the non-epistemic issue of truth with the epistemic issue of knowledge. But this is exactly what scientific realists do. By answering metaphysical and epistemological questions inseparably, scientific realists thus undermine their own efforts with respect to a non-epistemic conception of truth. This is the reason why realism must not to be confused with anti-scepticism. John Bigelow emphasised this point convincingly:

“It is possible for a skeptic to be a [scientific] realist. In this context, what I mean by a skeptic is simply someone who suspends judgment about something […]. And yet such a skeptic can nevertheless be a scientific realist. […] Realism is a metaphysical stance, and this is not to be conflated with the epistemological stance adopted by the opponent of skepticism.” (Bigelow, 1994, p. 5)

Later, Bigelow explicitly draws the analogy to moral realism, and indicates that van Fraassen’s agnosticism does not conflict with scientific realism thus understood (1994, p. 18; see also Devitt, 1997, p. 303). As I see it, Bigelow is (mostly) right about this.¹⁷ To be sure, the idea that realism and scepticism are compatible may be considered an inconvenience, since scientific realism is commonly treated as a decidedly anti-sceptical project. Even Godfrey-Smith, who appears to be open to this idea, expresses his concern that “there is a limit to the pessimism that is compatible with scientific realism as I understand it” (2003, p. 177). However, giving up our anti-sceptical attitudes (or, at least, refraining from treating them as defining characteristics of realism) is precisely the sacrifice we might have to consider for a sensible—and, not least, unified—understanding of realism. Indeed, such a reformulation of scientific realism would not only have the potential to restore the analogy to moral realism, but also to draw our attention to a long overdue discussion that has been notoriously overshadowed by the epistemological commotion over van Fraassen’s constructive empiricism. This brings me to the third and final disanalogy between the two realist views.

5 The third disanalogy: the primary opponent

The third disanalogy concerns the issue of what is to be regarded as the primary opponent of realism.

In metaethics, moral realism is primarily contrasted with constructivist and relativist accounts. Moral realists focus on the idea that moral truth is objective, obtains mind-independently, and does not lie in the eye of the beholder. In other words, their main emphasis is on what the standard account identifies as the metaphysical component: they believe that (true) moral judgments are made true by the moral facts of reality and reject views that in some way call the objectivity of moral truths into question. On the one hand, they identify ethical constructivism as a major anti-realist opponent, because constructivists employ an idea of mind-dependence, according to which morals truths are non-arbitrarily constituted, e.g., in a hypothetical social contract that rational agents would agree to, if they were in a fair original position (for ethical constructivism, see Rawls, 1980, pp. 518f, 567f). Therefore, constructivists directly repudiate the metaphysical core thesis of moral realism. On the other hand, moral realists regard ethical relativism as another important anti-realist opponent, because relativists take moral truths to obtain in relation to what is actually (and rather arbitrarily) accepted in certain societies, cultures, or traditions (for ethical relativism, see Harman, 1996, p. 5). Thus, they dispute the realist idea of mind-independence in an even more radical way. This is why realism in metaethics is conceived of as a profoundly anti-constructivist and anti-relativist project.

Of course, moral realists also attack non-cognitivist approaches; but it is obvious that this controversy which the standard account associates with the semantic component is merely an implication of the metaphysical core thesis. If you want to declare the mind-independence of moral truths, i.e. the idea that the truth-values of moral judgments are objectively determined by the facts of reality, then you must already presuppose that moral judgments are bearers of truth-values in the first place. This is why realism is a version of cognitivism, and those who want to be realists are ipso facto committed to defend cognitivism as well. Rejecting ethical non-cognitivism is thus mandatory for every proponent of moral realism; however, since non-cognitivism contradicts with the core thesis of moral realism merely indirectly, the main focus still remains on arguing against ethical constructivism and ethical relativism. These are the primary opponents of moral realism.

In philosophy of science, scientific realism is conceptualised in an entirely different way. To be sure, scientific realism is again opposed to constructivist or relativist accounts, but its primary challenge comes from agnostic, pessimistic, and sceptical approaches like, for instance, van Fraassen’s constructive empiricism. Scientific realists concentrate on the issue of optimism, i.e. their main focus is on what the standard account classifies as the epistemic component. They seek to defend epistemic optimism, or “historical progressivism” (see Goldman, 1986, p. 157), according to which we have good reasons to believe that our most successful theories are (at least approximately) true, that epistemic confidence in science is rationally justified, and that historical progress towards the truth is actually achieved by scientific theorising. As a result, scientific realists are (almost clinically) obsessed with defending themselves against epistemic positions that somehow attempt to shake and undermine this optimistic confidence. This is why realism in philosophy of science does not have the form of an anti-constructivist or anti-relativist, but rather, as I have already mentioned before, of a decidedly anti-sceptical project.

This is particularly apparent when we look at the central arguments that the debate revolves around: the pessimistic meta-induction (PI) and the no-miracles argument (NMA) (see Laudan, 181, pp. 32f; Putnam, 1975, p. 73; see also Psillos, 1999, pp. 71, 79f, 101ff). Neither of these arguments has any obvious bearing on whether scientific truths obtain mind-independently or not. Rather, what these arguments are primarily concerned with is the epistemic issue of optimism. Particularly the NMA, sometimes even praised for being the “ultimate argument” for scientific realism (Musgrave, 1988, pp. 229f; see also van Fraassen, 1980, p. 37), is specifically designed to silence agnostic and sceptical doubts, by claiming that optimism is the best explanation for the empirical and technological success of science. Though some scientific realists provide sophisticated analyses of how this argument might, additionally, support the metaphysical idea of realism (see Alai, 2023, p. 396; see also Sankey, 2008, p. 141), these considerations appear to be merely derivative.¹⁸ The metaphysical idea that scientific truths obtain mind-independently, that scientific statements and theories are objectively made true by the facts of reality, is not the primary focus of the NMA. Instead, it is the epistemological idea that the NMA is most distinctly dedicated to: the NMA tells us, first and foremost, why epistemic confidence in scientific theorising and historical progress is justified.

Of course, the fact that the NMA is primarily concerned with securing the epistemic optimism that scientific realism is associated with should not be taken to imply that scientific realism is a purely epistemic position that does not involve any metaphysical commitments at all. On the contrary, scientific realism does include metaphysical claims about the mind-independence of reality, and scientific realists usually do encounter versions of metaphysical anti-realism, which resemble the primary opponents of moral realism, insofar as they similarly promote an idea of mind-dependence and are referred to as ‘(social)constructivism’ or ‘relativism’. However, as far as the NMA is concerned, these metaphysical issues do not have the highest priority. Scientific realists certainly acknowledge these metaphysical controversies and may, from time to time, attempt to defend themselves against the anti-realist alternatives they face in this context; but often they disregard them in the end to focus almost entirely on the issue of optimism:

“It is implicit in its metaphysical stance that scientific realism is incompatible with much-in-fashion social constructivism (or constructivist anti-realism). Its defence against constructivism, though, is not part of this book and has to await a different project.” (Psillos, 1999, p. 301)

The reasoning behind excluding this metaphysical controversy is clear. Scientific realists have different priorities than moral realists: They do not regard metaphysical, but rather epistemological positions as their primary opponents.¹⁹

This last disanalogy raises another problem for the standard account. Scientific and moral realists identify completely different approaches as their primary opponents and, therefore, characterise their own views in profoundly different ways. Do we really want to rely on the standard account as a basis for integrating scientific and moral realism, when it fails to recognise these thoroughly distinct priorities? Once again, the standard account deceives us in erroneously suggesting that both realist views are parallel. In this sense, the standard account is prone to seriously misleading scientific and moral realists alike, particularly when they attempt to draw interesting connections between their views. On the one hand, moral realists might be inclined to apply the popular NMA to the metaethical debate, without realising that the NMA is originally conceived of as an epistemic argument for optimism—a component that moral realism lacks completely, which makes an application of the argument prima facie impossible.²⁰ On the other hand, scientific realists might feel tempted to impose their epistemic priorities onto moral realism, thus failing to acknowledge that moral realists do not share the anti-sceptical commitments that are so essential to them (see again Boyd, 1988, p. 202; Sterelny & Fraser, 2017, pp. 981f). The standard account must be rejected.

If you have made it thus far through this paper, I hope you are not surprised by what is coming next. Allow me, one last time, to hammer this point with brute force into your head: Scientific realists seem to be responsible for this muddled situation; it is therefore, if at all, up to them to fix the analogy to moral realism—not vice versa. The reason for this is, once again, that their priorities appear to be strange and call for correction. Of course, the focus on rejecting constructive empiricism would be (at least somewhat) understandable, if scientific realism were not challenged by constructivist and relativist approaches at all, or if these approaches were not that popular in philosophy of science. But unfortunately, they enjoy immense popularity—in philosophy of science and even in the humanities in general. Michael Devitt calls constructivism “the most influential bad idea in philosophy” (1997, p. 236) and even “a veritable epidemic” which “attacks the immune system that saves us from silliness” (ibid., ix). The list of the “usual suspects” (Sokal & Bricmont, 2004, p. 19) he identifies, i.e. influential philosophers who use constructivist and relativist rhetoric that is incompatible with scientific realism, is quite long and includes eminent names like Kuhn, Feyerabend, Putnam, Rorty, and Latour (Devitt, 1997, pp. 155ff, 203ff, 236f, 256 (fn 4); see also Niiniluoto, 1999, p. 227ff, 252ff). Alan Sokal and Jean Bricmont similarly identify a “relativist zeitgeist” which “originates […] from contemporary works in the philosophy of science, such as Thomas Kuhn’s The Structure of Scientific Revolutions and Paul Feyerabend’s Against Method” (1998, p. 51), but which has also had “an impact on the humanities and the social sciences” in general (ibid., p. 183). According to Paul Boghossian, this version of “’postmodernist relativism’ about knowledge” is not just one view among others, but has—at least in “vast stretches of the humanities and social sciences”—“achieved the status of orthodoxy” (2006, p. 2). The popularity of these profoundly anti-realist ideas is so unprecedented that some of them have even transcended the academy, influencing (and to a concerning degree, in fact, polarising) the public discourse in general.²¹ This is precisely what makes scientific realists’ priority on epistemic issues so utterly incomprehensible. Why do they focus on this secondary theatre surrounding constructive empiricism, while letting their most powerful opponents promote their anti-realist rhetoric almost unchallenged? Why do scientific realists let them take over not only philosophy and the academy, but also the public discourse without even seeing them for what they really are?—Their most decided adversaries. Not only does it make no sense; it also alienates them, as I was hopefully able to show, from their fellow realists in metaethics.

6 The epistemic deformation of realism

Let us take a step back now and look at the bigger picture. We started this discussion with a simple question: Can we identify a universal understanding of realism that philosophers from both philosophy of science and metaethics share, a core idea that shines through in both realism debates and characterises scientific and moral realism alike? Though I already indicated that I would like to answer this question affirmatively, suggesting that there is indeed such a universal conception of realism that may be applied to both science and ethics, so far I have merely been criticising the standard account for not getting this job done. The standard account obscures (at least) three substantial disanalogies between scientific and moral realism, i.e. it fails to account for the fact that the conceptions of realism we find in both debates differ significantly from each other. None of these disanalogies is even hinted at by the standard account. So if we wish to re-establish the analogy between both views, we must identify what the standard account gets wrong and what a suitable alternative might look like.

In this section, I would like to condense the results of the previous sections, by disclosing the connection between the discussed disanalogies. As I have already mentioned before, it seems to me that the disanalogies result from an epistemic deformation of realism, i.e. a tendency in the literature to contort the central message of realism, to reinterpret realism as an epistemic view, and to deemphasise its metaphysical core idea.

The disanalogies emerge because scientific realism happens—for some reason²²—to be much more epistemically deformed than moral realism. Scientific realism is not neutral regarding scientific theories, it is so decidedly anti-sceptical, and it deliberately chooses to identify epistemic positions as its primary opponents, precisely because scientific realism is epistemically deformed, because its (alethic) core idea about truth has been turned into an (epistemic) view about our cognitive access to the truth. Moral realism, however, remains neutral regarding normative-ethical judgments, it does not involve any anti-sceptical commitments, and it rather identifies metaphysical, instead of epistemic views as its primary opponents, because it is not epistemically deformed in the same way as scientific realism is, because it maintains the form of a view about truth and does not develop into a view about our epistemic access to it. This is where the disanalogies come from; this is what the standard account fails to notice.

The concept of a ‘convergent realism’, which has already been introduced into the debate on scientific realism (see Laudan, 1981, p. 21), might help us to encapsulate this idea even further.

In philosophy of science, scientific realism is typically conceived of as a convergent realism, i.e. a realism that is inseparably entrenched in epistemological controversies and has become virtually indivisible from the optimistic commitments that it is associated with. As a consequence, philosophers of science usually deem it incoherent to defend (scientific) realism without optimism.²³ Realists in philosophy of science do not merely subscribe to the idea of the objectivity of truth, declare truth as the proper aim of scientific theorising, and reject constructivist and relativist approaches; rather, scientific realism revolves around the notion that we have good reasons to believe that science has already made considerable progress towards this goal. In this sense, scientific realism may be characterised as a convergent realism, insofar as it incorporates an idea about an approximation of or a convergence to the truth.

In metaethics, however, moral realism does not assume the form of a convergent realism at all and is rather conceptualised as a (potentially) divergent realism. Of course, moral realists do not necessarily adopt a pessimistic attitude with respect to moral epistemology; they do not typically believe in an inevitable moral decline, in the futility of our moral aspirations, or in the vanity of our prospects to acquire and deepen ethical knowledge or to refine and improve our moral character. Nevertheless, moral realists do not treat these epistemic attitudes as defining characteristics of their conception of moral realism. They do not deem it incoherent to defend (moral) realism without optimism. Moreover, moral realists adopt an objectivist conception of (moral) truth, precisely because they want to account for the fact that it is always possible for our subjective opinions to divert from the objective truth. They think that moral truth does not lie in the eye of the beholder, because if it did, we would have “far less room for error […], and so far less ground for modesty” (Shafer-Landau, 2003, p. 1). In this sense, moral realism may be conceptualised as a divergent realism, insofar as it is motivated by the possibility of a divergence between (objective) truth and (subjective) opinion.²⁴

This is why the standard account must be rejected: It fails to account for the difference between these two conceptions of realism and the disanalogies they manifest themselves in.

Of course, we are not allowed to infer from the fact that the standard account neglects the disanalogies in question (descriptive claim) that it is, therefore, the conception of scientific realism that has to be altered (normative claim). Scientific realists cannot be compelled to realign their conception of realism with that of moral realists simply because this would allow for a unified account; for we may have other options at our disposal to advance the debate and rise above the standard account.

On the one hand, it might be argued that the disanalogies could be accounted for by the first-order differences between science and ethics, that they are merely expressions of the different interests or even epistemic needs that philosophers from the respective fields display, and that they are, in the end, understandable deviations that nobody has to be blamed for. According to such a view, realists from different fields adapt their conceptions of realism to the intellectual environment they find themselves in; and since science is fundamentally an epistemic enterprise, it is only natural that scientific realists develop realism as a philosophical view that focuses on epistemic issues like optimistic and anti-sceptical commitments. What the standard account gets wrong, then, is merely that it erroneously assumes that the two conceptions of realism are parallel when they are, in fact, profoundly disanalogous. The proper response, then, is not to repeat the mistake of the standard account—i.e. not to force both realist views into an abstract scheme that satisfies the epistemic needs of neither scientific nor moral realists. Instead, it would seem to be wise to simply accept the fact that both realist views are conceptualised disanalogously—and that an integration of scientific and moral realism does, consequently, not appear to be feasible or perhaps even worthwhile to begin with.

On the other hand, some philosophers might object that even if we still deem an integration of scientific and moral realism possible and wish to restore the analogy between both views, it may, nevertheless, not be the conception of scientific realism that is to be blamed for the estrangement we observed. Although the analogy between scientific and moral realism can be reestablished by adopting a divergent conception of scientific realism that abstains from its common preoccupation with epistemic optimism, it could also be saved the other way around—namely, by adopting a convergent conception of moral realism that deliberately starts to include epistemic commitments about ethical knowledge. According to such a view, the standard account may have been wrong in assuming that moral realism already involves such epistemic commitments, but it might have given us a correct impression on what realism (about both science and ethics) is supposed to look like. The proper response, then, would be to fully embrace the epistemic deformation and to encourage moral realists to be less epistemically cautious, to include anti-sceptical commitments into their understanding of moral realism, and to transform realism into an epistemic view about (scientific or ethical) knowledge altogether. In this way, an integration of scientific and moral realism might still be achieved—but without requiring scientific realists to correct the epistemic deformation of their view.

Though I am willing to concede that both of these alternative views might be worth exploring, I think they share a common problem. Both of these views attempt to persuade us—in one way or another—that the epistemic deformation of realism is either harmless and can be interpreted as an expression of scientific realists’ interests and needs, or that it is even to be regarded as an achievement, that we should outright embrace the notion of an epistemically deformed realism, and that it is not just permissible for scientific realists to turn realism into an epistemic view about knowledge, but that moral realists are required (or, at least, recommended) to do so, as well. While such a conviction might very well turn out to be true, allow me to, nevertheless, express my concern about it: the epistemic deformation of realism is prone to seriously causing confusion among philosophers. It is, for this reason, at least doubtful whether it should be perpetuated in philosophy of science—not to mention to be exported to metaethics. As we have seen, it tends to conflate philosophical and scientific commitments (see Sect. 3), it blurs the crucial distinction between metaphysical and epistemological questions (see Sect. 4), and it entices realists into misidentifying their most determined opponents (see Sect. 5). The epistemically deformed conception of realism that philosophers of science have grown accustomed to, most importantly, bears the risk of losing sight of the metaphysical core idea that realism used to be about—namely, the commitment to mind-independence and a non-epistemic conception of truth.

This is why I have spent the preceding sections seeking to call upon scientific realists to change their mind on how to conceptualise realism, to modify their understanding of what realism genuinely says, and to realign scientific realism with the conception of moral realism. It is true that scientific realists cannot be compelled to realign their conception of realism with that of moral realism per se, but we may criticise scientific realists for the confusion their convergent conception of realism is prone to create—and we may think about an alternative conception of realism that philosophy of science could benefit from. Allow me, then, to conclude this essay by stating this alternative to the epistemic deformation more explicitly and thus providing a positive outlook on how we may reconcile the realism debates in the future.

7 The return to alethic realism

I would like to end this essay on a positive note, indicating that scientific and moral realism have, indeed, more in common than the preceding sections may have suggested. In this concluding section, I would like to explore, then, what I take to be the robust parallel between both views—a parallel that merely gets obscured by the disanalogous deformation that turns these originally alethic views (with varying degrees of success) into mostly epistemic ones. According to this suggestion, we may still think of scientific and moral realism as philosophical views that share the same general idea that is merely applied to different domains. This general idea, however, may no longer be augmented with epistemic claims about knowledge, but, instead, focuses entirely on alethic claims about truth. In particular, it seems plausible to rely on what is commonly referred to as a realist conception of truth or an alethic realism as a basis for such an integration of scientific and moral realism.²⁵

Alethic realism is, as occasionally indicated before, a theory of truth that seeks to account for the non-epistemic nature of the concept of truth, by introducing the facts of reality as the truth-makers of true statements (see Alston, 1996, p. 5; Künne, 2003, p. 20). Whether a statement is true or false does, for the alethic realist, not depend on whether we choose to believe in its propositional content, whether we have good reasons to believe in its truth, or whether we are in a position to know that it is true; instead, whether a statement is true or false does only depend on the way reality is—and is, therefore, independent of the epistemic attitudes or beliefs of any individual or group. In his book A Realist Conception of Truth (1996), William Alston provides a concise presentation of this core idea of alethic realism:

“What it takes to make a statement true on the realist conception is the actual obtaining of what is claimed to obtain in making that statement. […] Nothing else is relevant to its truth-value. This is a realist way of thinking of truth in that the truth maker is something that is objective vis-à-vis the truth bearer. […] This is a fundamental sense in which truth has to do with the relation of a potential truth bearer to a reality beyond itself.” (Alston, 1996, p. 7f (emphasis in the original); see also Goldman, 1986, p. 17)

In this sense, alethic realism is a decidedly objectivist view, i.e. a view that seeks to secure the objectivity of truth, and is prepared to reject any alternative view that somehow attempts to call this objectivity into question. On the one hand, alethic realists reject epistemic conceptions of truth that try to establish a connection between what is true and what is held to be true, thus undermining the status of the concept of truth as a non-epistemic concept (for such an approach, see Putnam, 1981, p. 55; 1990: vii; see also Alston, 1996, pp. 188ff). On the other hand, alethic realists identify—which is particularly interesting in our context—constructivist and relativist accounts that are even more outspoken about their radical rejection of the objectivity of truth (for such an approach, see Rorty, 1998, pp. 67, 82f, 86; see also Boghossian, 2006, pp. 25ff, 42ff; Searle, 1993, pp. 57, 62f).

What makes alethic realism so attractive in this context is that both scientific and moral realism already involve a commitment to a non-epistemic conception of truth, and thus seem to be related to alethic realism by default. This becomes particularly apparent when we look at the metaphysical components of scientific and moral realism. Scientific realists make it fairly clear that what they attempt to express with the concept of mind-independence “cannot be properly stated without reference to a non-epistemic conception of truth” (Psillos, 2005, p. 393); additionally, they explicitly endorse the idea that scientific realism is closely associated with a commitment to the objectivity of truth (see, for example, Sankey, 2021, p. 7). Moral realists, likewise, emphasise that their view has, essentially, to do with a commitment to the objectivity of truth (see DeLapp, 2013, p. 18; Shafer-Landau, 2003, p. 2), and that this commitment is, again, best encapsulated by a non-epistemic conception of truth (see DeLapp, 2013, p. 23; Shafer-Landau, 2003, p. 15). In other words, both realist views share the conviction that the truth of either scientific statements or moral judgments obtains mind-independently—insofar as the (objective) truth is deemed to be independent from (subjective) opinions.²⁶

This is what a return to alethic realism might achieve: it may offer an alternative to the epistemic deformation of realism, by drawing our attention to an alethic core idea that seems to be much more characteristic of realism than any of the fashionable speculations on historical progress that philosophers are all too often mesmerised by.²⁷

Some philosophers of science might still have some concerns about such an alethic conceptualisation of realism. They might, for instance, fear that the focus on truth bears the risk of obscuring interesting and much needed discussions on epistemic issues—discussions that are able to inform our understanding of scientific theorising in various helpful ways and that, consequently, must not be expelled from our philosophical outlooks on science. As a consequence, they might either advocate for an epistemic understanding of scientific realism or, at least, opt for a pluralist account on scientific realism that allows for the possibility of both alethic and epistemic conceptions.²⁸ Others might argue that they do not even see the point in adopting an alethic understanding of realism. They might have the impression that it seems rather strange to postulate and defend the objectivity of scientific truths when we afterwards remain indifferent as to whether these truths are epistemically accessible to us.

However, a return to an alethic understanding of realism can arguably be defended against such objections.

First, a focus on truth does not necessarily imply that epistemic debates have no place in philosophy of science anymore. Instead, such a return should merely serve as a reminder to keep these epistemological debates separated from alethic ones. It should remind us that before we turn our attention all too quickly to the epistemic issue of knowledge, it might be worthwhile to reflect a little longer on the non-epistemic issue of truth. This is why philosophers of science may still have their epistemic debates on scientific knowledge and historical progress; but we should also allow for a separate debate on the non-epistemic issue of truth and mind-independence—a debate that, particularly, takes the challenge of constructivist and relativist approaches seriously, without being too easily distracted by sceptical worries. In this sense, a return to alethic realism might, indeed, help us to separate these discussions more thoroughly from each other and could even contribute to minimise the risk that the one debate casts a pall over the other.²⁹

Second, it should be noted that while alethic realists are not actually indifferent towards sceptical worries, but merely strive to keep their conception of realism free from such epistemic commitments, their non-epistemic conception of truth should by no means be taken to be irrelevant for epistemological questions. Most importantly, realism matters because it is a presupposition of fallibilism: if truth is objective and determined by the facts of reality, then it is particularly difficult to access the truth and perhaps even impossible to achieve certainty in doing so. On the contrary, if truth lies in the eye of the beholder, then every beholder is entitled to create his own truth and, thus, can—by definition—evade any possible error. There is, in other words, an alliance between (alethic) realism and (epistemic) fallibilism—an alliance which is quite frequently invoked in metaethics (see, for example, Shafer-Landau, 2003, p. 26) and could achieve some plausibility in philosophy of science, as well (see already Popper, 1961, p. 491). This is why alethic realism is apt to contribute immensely to the advancement of our understanding of epistemology: its insistence on the objectivity of truth protects us from epistemic presumption and overconfidence.

To be sure, these brief responses are only indications of how an alethic conception of realism could be defended against some of the concerns philosophers of science might have, and the details of such an approach must surely be fleshed out somewhere else; but I hope that this discussion, at least, makes us ponder on the idea that scientific realism may be re-formulated along these lines and separated from epistemic optimism, that the analogy to moral realism may be restored that way, and that the most vigorous antagonist realists face, in philosophy of science and metaethics, may not be the (moderate) sceptic who has doubts in our epistemic capacities, but the (radical) constructivist and relativist who calls the notions of truth and reality itself into question.