Relaxed Naturalism, as originally proposed in Hutto & Satne (2015), was offered as a corrective to what its authors regard as unwarranted restrictions imposed on naturalising projects by purer forms of scientific naturalism. In this respect, Relaxed Naturalism rejects and opposes familiar varieties of post-Quineian naturalism – specifically those variants of physicalism that operate with and only tolerate a narrow vision of nature that can be etched out by the hard natural sciences or, even more restrictively, a completed physics (see Hutto & Satne 2018a, Hutto 2000).

When making its debut on the philosophical scene, Relaxed Naturalism’s main ambition was to argue that if we are ever to properly understand certain natural phenomena of central importance then we will need to avail ourselves to resources beyond those of the hard, natural sciences. Accordingly, Relaxed Naturalism stressed the importance of being able to investigate target explananda by drawing on and seeking to harmoniously integrate the findings of a wide range of relevant scientific resources which might, in turn, require making use of the findings of a wide range of sciences, including – for example – anthropology, developmental psychology, comparative psychology, cognitive archaeology, social neuroscience, and so on. At its heart, Relaxed Naturalism seeks to show how philosophy and the sciences can connect productively and have cooperative dealings.

The value of using and synthesising these many and various resources, both scientific and philosophical, in the way Relaxed Naturalism recommends is defended in detail elsewhere. For example, this can be seen in Hutto & Myin’s (2013, 2017) and Hutto & Satne’s (2015) attempts to understand the natural origins of content. It can also be seen in the attempts Hutto (2008) and others have made to understand our everyday practices of making sense of ourselves and one another in terms of narrative practices.

In the light of its original statement and focus, Macarthur (2015) characterised Relaxed Naturalism as nothing more than a more expansive, broader version of orthodox scientific naturalism. Yet, attracted to some of the core ambitions of Relaxed Naturalism, Macarthur (2015) recommended that its advocates would be better served by embracing Liberal Naturalism.

Relaxed Naturalists have, to date, resisted that invitation. In subsequent work, they further articulated their philosophy of nature, inspired by Wittgenstein, showing that Relaxed Naturalism is not simply a broad or more expansive version of scientific naturalism (Hutto and Satne 2018a, b; Hutto 2022). At the same, they raise concerns that Liberal Naturalism – at least as it is characterised by Macarthur (2015, 2018, 2019, 2022) – gets its commitment to pluralism right but it struggles to articulate a clear stance on metaphysical and ontological issues. Moreover, they worry liberal naturalism, as Macarthur presents it, does not do enough to illuminate the distinct contributions of philosophy and science and how they work together positively in naturalistic projects. Most of all, Relaxed Naturalists worry that, without greater clarity and definition, Liberal Naturalism risks being, in a word, too liberal.

This paper continues the conversation, working to identify issues on which Relaxed Naturalist and Liberal Naturalists agree and disagree, and to determine how their joint cause of developing a tenable alternative to scientific naturalism might be successfully advanced. The action of this paper unfolds as follows: Section one, Broad Agreements and Specific Concerns, highlights the important point of agreement between Relaxed and Liberal Naturalism before identifying specific concerns about Macarthur’s particular rendering of the latter. Section two, Slicing Away the Supernatural, offers a proposal about how one of those specific concerns might be addressed – adding a key philosophical item to the naturalist’s toolbox when it comes to identifying and dealing with supernatural excesses. Section three, A Positive Contribution to Metaphysics, offers a further corrective to some of the more pessimistic assessments about the possibility that philosophers might do any positive metaphysical work in support of the naturalistic project. Such positive work can be done by removing confusions and clarifying our everyday ontological commitments. Crucially, the idea that a satisfactory naturalism must be wholly negative and anti-metaphysical is rejected.

1 Broad Agreements and Specific Concerns

What is the relation between Relaxed Naturalism (henceforth, RN), and Liberal Naturalism (henceforth, LN)? Are they competing or complementary cousin frameworks, or is the former simply a species or special variant of the other? We cannot answer these questions properly without getting clearer about the extension of the LN label.

Certainly, there is much on which RN and LN agree. Both maintain that scientific naturalism is not a tenable, coherent option for thinking about the natural world and our place in it (for assessments along these lines see, e.g., Hutto & Satne 2018a, 2018b, Macarthur 2019; De Caro 2022; Staiti 2022, Redding 2022). Both RN and LN hold that we must look in a different place than scientific naturalism to guide our thinking about what belongs in the category of the natural.

RN and LN also agree in acknowledging a world of things that are not posits of the sciences. Macarthur (2022), a chief spokesperson for LN, explicitly borrows from Sellars (1963) in taking the Manifest Image – that is to say, the world as it ordinarily appears to us – very seriously as a starting point. This homely starting point regards the mundane things we encounter in everyday life – such as people, much of what we do and feel, and the artefacts we produce – as shining examples of things that, though wholly natural, would not be visible in the Scientific Image (henceforth, SI) – which is the world depicted only by the sciences.

Liberal naturalists assume that the class of non-scientific is not wholly co-extensive with the class of the nonnatural. Some things that are invisible to the SI are not found anywhere in nature, but this is not true of everything that is invisible to the SI. As such, we cannot simply use science as our exclusive guide to the natural.Footnote 1

Macarthur (2022) provides a helpful clarification of the notion of the non-scientific-yet-non-supernatural with his example of artworks:

When I speak of artworks as nonscientific, I am not denying that they can be objects of scientific study. What I am denying is that their existence depends on whether or not they are ‘posits’ of a successful scientific theory, to use Quinean terminology. In other words, they have a life science knows nothing about outside of the context of scientific inquiry (Macarthur 2022, p. 275).

The LN label is said to have been coined by McDowell in 1994 (see, e.g., Smith 2022, p. 30). It can and is sometimes used to denote any position that rejects scientific naturalism but still seeks to exclude the supernatural from its reckoning of what exists.Footnote 2 When defined as loosely as above, it appears that many philosophers down the ages, including Davidson, Putnam, Price, Sellars, Strawson, and Wittgenstein, can be said to have embraced LN or something near enough.Footnote 3 If LN, broadly construed, is used as a label for the sort of position described above, then Relaxed Naturalists can be said to embrace it too. Yet this is only because with great looseness comes great inclusivity.Footnote 4

However, when we compare particular versions of LN with RN, specific concerns and disagreements can arise. The analyses of this paper focus on Macarthur’s articulation of LN, highlighting particular issues that Relaxed Naturalists seek to address concerning his formulation. Taking Macarthur’s version of LN as a departure point seems entirely reasonable, given his invitation to Relaxed Naturalists to embrace LN.

Morag (2022) epitomises the core of Macarthur’s take on liberal naturalism, noting that he conceives of it “in positive terms as a philosophy of ‘the manifest image’” (Morag 2022, p. 384). She stresses, in this regard, that it is more than “a mere reaction to scientific naturalism” (Morag 2022, p. 384). Morag (2022) regards it “not as a doctrine, but as a method of inquiry … [and hence it is] not merely an inventory list of our ontological commitments” (p. 384).

What could it mean to offer a philosophy of the Manifest Image? As deVries (2022) reminds us,

The manifest image is a conceptual framework embedded in the practices and natural languages developed by humans in terms of which they make sense of and cope with the world and their place therein. It is an idealisation that abstracts away from the many differences of detail across languages and cultures. But it is not so idealised that it can be regarded as fixed and immutable ….

Sellars thinks that the fundamental structures of the manifest image have been best captured by such philosophers as Aristotle and P.F. Strawson. It portrays a world in which persons and things are the basic kinds of objects, though the emphasis is on persons (p. 113).Footnote 5

If we understand the Manifest Image (henceforth, MI), as an ‘idealisation’ which is not ‘fixed and immutable’, then it is not credible that everything that ‘we’ uncritically commit to in our everyday, folksy and familiar practices belongs on the final inventory list of ‘the’ natural, even if we, initially, take the MI seriously and assume that its counterpart, SI, cannot be our sole guide to what’s natural.

In other words, not everything in the domain of the non-scientific – that to which we accord everyday status – is sacrosanct or safe, ontologically speaking. Not everything in ‘the’ MI as we initially depict it will make the final list of ‘the natural’. As deVries (2022) informs us, “Though Sellars believes that the MI can (and ought to) be given a naturalistic analysis, he does not believe that the MI is ultimately a stable and enduring framework. It will have to be superseded” (p. 114).

On Sellars’ view, the sciences do the heavy lifting in modifying the MI. This is why Sellars holds that “science does not simply add to our ontology; ultimately, it challenges the ontology and the explanatory structures constitutive of the manifest image” (deVries 2022, p. 114). How much revision is necessary or even possible to ‘the’ MI? How deep might the cuts go when sculpting it? How much and which bits of the MI can be replaced by the SI? Here it is important to note that Sellars acknowledges “the SI is methodologically dependent on the MI; it could not get off the ground unless the MI were already in place. He denies that this entails any substantive priority of the MI over the SI” (deVries 2022, p. 114).Footnote 6

Reasons to doubt the coherency of scientific naturalism are also reasons for thinking that a complete replacement of the MI by the SI is not feasibly on the cards. At this juncture, it can be tempting to give our full attention to ‘the clash’ between the MI and the SI, and to wonder how we might adjudicate the conflict between the two images. After all, “the critical manifest image … is the result of subjecting the manifest image to critical scrutiny, which includes how well it hangs together with the scientific image of the world” (Macarthur 2019, 574).

These considerations prompt the following questions: When engaged in this critical work on the MI, what, if any, tools should we use other than those provided by the sciences? And, if there are additional tools that we should be using, exactly how and when should we use them? Those are questions of particular interest to RN.

Furthermore, there is something else of notable concern here. On the face of it, Morag’s characterisation of Macarthur’s LN as seeking to make positive contributions to the task of refining and providing a philosophy of MI does not tally well with some of Macarthur’s (2018) own statements about the character of his LN. Highlighting its anti-metaphysical ambitions, he bills LN as “a negative discipline that … acknowledge[s] that there is a plurality of different kinds of understanding or knowing both within and outside the sciences” (p. 103).

To be fair, in more recent work Macarthur stresses that his LN has “the aim of articulating a naturalism that makes available a ‘cooperation’ or ‘integration’ of scientific explanation and humanistic interpretation or understanding” (Macarthur 2022, p. 267).Footnote 7 However, despite making this statement, Macarthur provides scant details as to how LN will achieve such aspirations.

We may wonder how well these characterisations of LN hang together. How can LN, in any variant, provide a pluralistic, positive philosophy of the MI that connects and integrates scientific explanation and humanistic understanding, while at the same time remaining a negative, anti-metaphysical discipline or method of inquiry?

Answering this question is important, especially since, resisting Macarthur’s characterisation of LN, RN does not conceive of itself in purely, or even principally, negative terms. It proposes adopting a positive – if positively relaxed – view of nature and a methodologically multifaceted naturalistic enterprise – one that earmarks a variety of roles for philosophy and the sciences to sculpt what, in the final reckoning, will be our understanding of nature.

2 Slicing Away the Supernatural

Any naturalism worth its salt needs to assume that there are non-arbitrary, concrete procedures for deciding which bits of the MI should stay and which should go. A worry that has been raised about Liberal Naturalism is precisely that “if nothing is ruled out, LN’s epistemological openness threatens to collapse into an ‘anything goes’ non-naturalism” (Reynolds 2022, p. 71). Naturalists need the right resources for making such calls, otherwise the resultant naturalism will be “too tolerant” (Rouse 2022, p. 178).

Unsurprisingly, then, the pivotal concern of naturalists is to say how we can keep the supernatural at bay. Liberal Naturalists are fully and openly aware of the importance of meeting this challenge.Footnote 8

Liberal naturalism … is … a form of naturalism on the grounds that it refuses to admit the supernatural. From this perspective, ‘natural’ is … equivalent to the anti-supernatural (De Caro and Macarthur 2022, p. 2).

The ‘naturalism’ in liberal naturalism is … its commitment to anti-supernaturalism (De Caro and Macarthur 2022, p. 2).

The natural order is simply the non-supernatural order (Macarthur 2022, p. 275).

As far as official statements go, these are all well and good, but – as Morganti (2022) points out – lacking further details: “the domain of the ‘non-scientific-yet-natural’ envisaged by liberal naturalists remains not entirely clear” (p. 245). In particular, he complains that liberal naturalists tend only to provide “putatively paradigmatic examples of natural, supernatural and non-scientific-yet-natural entities” (Morganti 2022, p. 246). He goes on to note that the problem with the standard LN proposal is that “while it gives the impression that the relevant distinctions are based on clear-cut and fixed boundaries, this is not the case.” (Morganti 2022, p. 246).

Clearly, merely making lists of what is non-scientific-but-natural won’t cut it. How, then, should we proceed? Morganti (2022) proposes that “the truly interesting dichotomy is not so much that between the natural and the supernatural but, rather, that between supernatural (i.e. metaphysical) entities, properties, processes, etc. that play a genuine explanatory role in connection to current science and those that do not” (p. 246).

Employing this sort of genuine-explanatory-role scientific test can and should certainly play a key part in helping naturalists determine what to keep and what to throw away. However, in many cases such a test won’t suffice. Supernatural commitments are not always bound up with explanatory commitments. In the next section, I propose that the relaxed solution is for naturalists to call on philosophical considerations, beyond the resources of the sciences, to help achieve their end of satisfactorily determining which categories of things deserve to be counted among the non-scientific-yet-non-supernatural.

Exactly how open should a naturalism be when it comes to acknowledging plural kinds of entities, forms of understanding and knowing that exist beyond the sciences? Where should naturalists draw their line?

Liberal naturalists want to be liberal but do not wish to be utterly indiscriminate. Their strategy is to appeal to the sciences to demarcate the boundary between what’s natural and what’s supernatural. For example, De Caro and Voltolini (2010) tell us that the supernatural should be identified as anything which is, in principle, contradictory to or incompatible with science.Footnote 9 As such, at the top of their list of supernatural posits are causally inert entities or forces that are accessible to or knowable by those with special cognitive or mystical powers – powers that cannot be reconciled with what we take to be natural forms of human understanding.Footnote 10 A bit more sweepingly, Macarthur (2018) defines a supernatural being as “an unobservable super-human being posited for causal explanatory purposes but which lacks the requisite scientific support to be credible” (p. 102). Indeed, he proposes that a crucial role of the sciences is to discredit supernatural phenomena by showing them to lack “sufficient or strong enough connections to the empirical” (p. 104).

As far as they go, these statements give some direction for dealing with the supernatural. Do they suffice to deal with all relevant cases? Do they go far enough to be useful?

When we try to apply the recommended principles in practice it becomes evident the approach generates additional questions, and that more is needed. Is being incompatible with or contradictory to science the only mark of the supernatural? Are we to run these tests with reference to current science or some future, finished science? How strong need a connection to the empirical be for something to qualify as natural as opposed to supernatural? Without solid answers to these questions, the approach the liberal naturalists propose for deciding what is supernatural and dealing with it remains too vague and vacuous – and, hence, too liberal – to rule out everything that needs ruling out.

To illustrate the point, let’s consider a concrete case. It is a longstanding philosophical myth that science debunks all supernatural entities by showing them to be inferior theoretical posits to those of some rival explanatory theory (Quine 1976; Churchland 1981). Witches are the favourite, go-to example of such debunked entities in the philosophical literature, as this passage from Nericcio (2017) reveals:

Consider two theories about the causes of illness: witch theory and germ theory. Witch theory quantifies over witches. It posits that for any given illness there must have been a witch who caused it. Using this theory, one might make certain predictions about the world, but it is unsurprising that [such] predictions will turn out wildly inaccurate. The germ theorist applies her theory and finds that it allows for tremendous predictive success, and so it is very likely that the entities, germs, quantified over by the theory actually exist – and that witches do not (Nericcio 2017, p. 3).

When it comes to telling this standard story, the fly in the ointment is that it is not credible that commitment to the existence of witches – which was and remains a widespread commitment for many – stems from or is primarily based in efforts to produce proto-scientific explanations.Footnote 11 The idea that witches are theoretical posits is not credible, even though witches are generally ascribed special, often maleficent, powers. The idea of a witch, found in most human societies, is complicated. It is inspired by multiple sources, it takes many forms, and it precedes and survives much adjustment: many different sorts of properties have been attributed to witches across time.

If witches are not explanatory posits, defeated or otherwise, then what is their status? In setting out the stall for her ‘easy approach’ to ontology, Thomasson (2015) importantly distinguishes between simple and explanatory realism – highlighting the differences in their motivations and the status of their commitments. Explanatory realists propose that certain entities exist based wholly on the putative power said entities have for explaining other things. By contrast, simple realists accept that certain entities exist based on nothing more than first-order existential commitments or entailments that they deem to be trivial or uncontroversial. As such, the entities to which a simple realist commits “are not ‘posits’ that are parts of ‘theories’, the inclusion of which is justified by their explanatory power” (Thomasson 2015, p. 157).

Pertinently, Thomasson (2015) observes that her easy approach to ontology and its simple realism need not leave us “accepting the existence of purported objects of absolutely any kind” (p. 155). This is because the application conditions associated with ‘witch’ are not met – on the assumption that “the application of ‘witch’ to a woman requires that she be endowed with supernatural powers in virtue of making a pact with the devil” (p. 155). Obviously, if there are no supernatural powers or entities, such as the devil, then it follows that witches do not exist.

Although the simple realist answer is useful in helping us to understand the status of what is involved in commitment to the existence of witches, it is of little help when it comes to advancing the naturalist’s cause of saying how we decide which entities to admit and to exclude from our ontology in a non-question begging manner. This should be evident from the fact that Thomasson’s grounds for thinking that application conditions are not met in the case of witches depends on our already having identified the relevant powers and entities associated with witches as supernatural, and hence non-existent.

For their purposes, naturalists appear to be on safer ground in agreeing with Thomasson (2015) that “there is no uncontroversial claim from which we may trivially infer the existence of witches—it’s not as if we all agree that there are particles arranged witchwise.” (p. 155, emphasis added).

It helps to see why the latter answer holds some promise by considering another concrete case. In a recent paper, Evan Thompson (2021) highlights the tensions that exist between Buddhist viewpoints and the assumptions of scientific naturalism in an effort “to keep the space open for a genuinely two-way and mutually critical dialogue” (Thompson 2021). In her foreword to the revised edition of The Embodied Mind, Eleanor Rosch (2016) makes indelibly clear how dramatically we may need to change our current thinking about what is possible in nature if we are to accommodate Buddhism fully as a living practice.

In particular, she holds that if we acknowledge the mental feats of Tibetan lamas properly then we must revise some quite familiar everyday and mainstream scientific assumptions about what goes on in nature. In this vein, she directs us to consider:

evidence that there could be aspects of mind that are separable from the brain and perhaps even the body. Tibetan lamas give mind-to-mind transmission of various kinds of wisdom states. Unlike the design of multitudes of failed extra sensory perception experiments in the West, such transmissions are not of mental contents but of what are considered deeper aspects of mind. Although the transmissions are not conveyed by ordinary sensory or intellectual means, they can be experienced — as is attested by many Western students of Tibetan teachers. ….

Evidence for separation of the deep mind from the brain occurs in even more paradigm-challenging circumstances. At death Tibetan high lamas enter into what is called the death samadhi. The lama is medically dead: no brain activity, no organ activity, but his heart center remains warm, and transmissions of enlightened mind states can emanate from him even more strongly and clearly than in life. This may continue for days, even weeks or longer. The Vajrayana yogic explanation is that the subtlest energies of the nondual mind have withdrawn from the outer body into the central channel, have then united in the heart center, and are now radiating to the world. Typically when the lama’s mind, in its most subtle yogic sense, is judged to have merged with the dharmakaya (the fundamental ground of being), and his body is cremated, rainbows appear. I have witnessed all of this twice; it definitely shakes one’s scientific preconceptions (Rosch 2016, xlvii).

Rosch (2016) anticipates that those in the thrall of Western thinking and science will resist when it comes to accepting the reality of the putative non-dual phenomena described above:

Where science, as it is done now with its mechanistic and materialist assumptions, meets experience, Buddhism, or anything else, the science simply takes over like a colonial ruler. This is body imperialism, not dialogue … The key to progress is to keep an open mind—and while we are at it, it would not hurt to also have an open heart (p. lii).

Yet, even without ruling out the possibility of such phenomena by appeal to mechanistic and material assumptions, it behoves us to ask just how liberal and open-minded ought we to be when evaluating such cases. Are they to be ruled in as natural-but-yet-to-be-explained or ruled out as supernatural in character? It is instructive, at this point, to consider the following comment from De Caro and Macarthur’s (2022) when thinking about this case:

[W]hat do we mean by supernatural? This is not easy to explain since it is widely admitted that science does not explain everything and some new previously unexplained phenomenon rather than automatically being banished as supernatural might rather show gaps or shortcomings in our current scientific understanding of the world that needs correction. What is “supernatural” today (in the sense of what cannot be explained by current science) may be part of tomorrow’s new science (p. 2).

Ultimately, Rosch’s (2016) whole point in bringing up this case is to show that if we are to take certain sorts of Buddhism seriously, then we must be open to rethinking ground-floor scientific convictions about what can and does occur in nature. How far should we be prepared to go? By comparison, Gallagher (2017) also maintains that, in key respects, enactivism’s philosophy of nature presents challenges for existing sciences in the here and now. He anticipates that the sciences of the mind will need to transform and expand, at least methodologically, in response to what he takes to be legitimate challenges arising from enactivist philosophy of mind and cognition.

Motivating this conclusion, he tells us that though “a philosophy of nature is not to do science, it can still offer clarifications relevant to doing science, and it can inform empirical investigations … It offers critical distance and practical suggestions at the same time. In some cases, it may make doing science more difficult” (Gallagher 2017, p. 23).

According to Gallagher (2017), the major reason why enactivism poses such challenges is because it assumes that “nature cannot be understood apart from the cognitive capacity that we have to investigate it” (Gallagher 2017, p. 23). For this reason, he maintains that enactivism should be and is making a positive difference to the shape of experimental science in very particular ways. If Gallagher is right about the need for these changes, then this would be a case in which philosophy is taking the lead in shaping the practice of science. This is interesting because it establishes that the sciences might need to take important cues from philosophy on occasion. This point is well made, but surely there are limits.

Any substantive naturalism must be in position to give us guidance on how to adjudicate cases such as these, even if they are treated in retail rather than wholesale fashion. At least, we need to be able to say which resources we should call on, and in what measure, to decide whether the sorts of occurrences Rosch (2016) claims to have witnessed – and others report as occurring – have sufficiently robust empirical support to warrant making a fundamental shift in standard scientific account of what happens in nature or whether they belong on the list of otiose supernatural posits.

In responding to this challenge, it is important to highlight that the putative occurrences Rosch (2016) claims to have witnessed – dead high lamas transforming into rainbows – are truly special and extraordinary: they certainly do not form part of everyone’s familiar experience.Footnote 12 We are entitled to view claims that such occurrences happen as controversial and thus to treat with suspicion whether they feature in the manifest image since they are out of step with what the bulk of people ordinarily encounter and experience in their mundane daily lives. Borrowing Thomasson’s words, it’s not as if we all agree that such things happen.

The important point here is that due to their extraordinary character, we are entitled to place the phenomenon Rosch (2016) and others claim to have witnessed under suspicion – even in advance of testing whether, in the end, such purported happenings are compatible with scientific findings and prior to conducting serious scientific investigations into said phenomenon.

What’s the upshot? It is certainly true that the sciences can provide us with a means of demarcating the natural from the supernatural in some cases. But the sciences provide neither our only nor always the best resources at our disposal for drawing that line.

At this juncture, consider again Macarthur’s (2018) criterion for identifying the supernatural as that which is “posited for causal-explanatory purposes yet without credible scientific support”. It should now be clear why this criterion is simply too narrow to enable us to deal with the case under scrutiny.

If we take Rosch at her word, the remains of Tibetan high lamas converting to rainbow bodies, however rare, is a perfectly observable phenomenon. Crucially, it is not something posited for causal explanatory purposes. If such transformations truly occur then they scream out for explanation but the transformations, as such, are not, in themselves, explanatory posits.

Similarly, the claim that contentless deep wisdom states can be transmitted from the warm hearts of dead Tibetan lamas is, presumably, not itself advanced as an explanatory posit but something to be explained. The idea that wisdom is something that can be transferred from one person to another looks to have all the hallmarks of picture-driven commitment that will lead only to philosophical trouble. For this reason, the claims made about these alleged processes should set off philosophical alarm bells and, as such, they will be as ripe for philosophical debunking as, say, proposals embraced by mainstream cognitivism that our senses pick up and minds manipulate causally efficacious content-bearing information (see Hutto & Myin 2013, 2017; Hutto 2013a).Footnote 13

What philosophical tools might be in our arsenal for dealing with such cases? Moyal-Sharrock (2013, 2021) usefully introduces us to the notion of Wittgenstein’s Razor – a special device that can be used to cut away our ill-motivated picture-driven commitments that go beyond and distort our understanding of the everyday and the mundane.

To understand the nature and use of Wittgenstein’s razor, it proves useful to recall that, for him, descriptive philosophy is clarification and precisely not a matter of advancing hypotheses. It is, instead, the primary activity of the philosopher to provide reminders or perspicuous presentations of our use of concepts concerning everyday subject matters. Yet, Wittgenstein is clear that such reminders can only be appreciated by those who are not under the sway of or in the grip of pictures. Thus, his razor is designed to cut way away distortion so as to sharpen up the philosopher’s perspicuous presentations.

Shaving away such unhelpful features is a fundamental part of Wittgenstein’s philosophical rescue operation – that of bringing “words back from their metaphysical to their ordinary use” (PI § 116). The task is anything but easy: it requires breaking free – often, repeatedly – of certain compelling but distorting pictures or ways of thinking about various subject matters – ways of thinking that irresistibly attract us and “bewitch our intelligence” (PI § 109). As Hyman (1991) observes:

Pictures … wreak havoc if … we take them at face value … Their grip on the imagination is not the grip of a tremendous hypothesis, like the big bang, but more like the grip of an entrancing metaphor or myth; and their influence … is as permanent as the language in which they are lodged (p. 7).

One of the most pernicious and perverse effects of pictures, Wittgenstein warned, is that “when we have got a picture of our ordinary way of speaking we are tempted to say that our way of speaking does not describe the facts as they really are” (PI § 402).

Painstaking philosophical work is needed to liberate us from picture-driven tendencies of thought, and such work is primarily a matter of “working on oneself” (C&V, p. 16). It can be a prolonged and, potentially, unending process. Pictures can take hold of us and foster unshakeable convictions about how we must conceive of various subject matters. Picture-driven thinking can produce a “superstitious conviction (PI § 110) … [holding] in place a picture of how things must be (PI § 115), despite contradiction and anomaly” (Williams 2010, p. 8).

Importantly, as Williams (2010) also highlights, “the superstition or picture that holds the philosopher captive is not itself a mistake” (p. 7). Pictures are not empirically false theories. Rather, at root, to be in the grip of a picture is to be attracted to certain ways of talking – to particular analogies – which, when coupled with certain other commitments, can foster confusion and drive misguided attempts to explain and theorize. Breaking free of pictures, and the picture-driven theorizing they can inspire, requires recognizing that the picture in question is not a source of special insight into a given subject matter but is, in fact, an imposition on our ordinary thinking. Hence, what is needed for dealing with and overcoming pictures is “therapy, rather than a better theory” (Williams 2010, p. 8).

Let us consider a familiar case within the confines of philosophy. The existence of mental objects of our private acquaintance seems intuitively compelling and, indeed, common-sensical to many. Wittgenstein’s corpus is, however, replete with efforts to show that the commitment to this idea of mental objects is an imposition on, and in no way an uncontaminated part, of our ordinary thought and talk.

What is the source of the conviction that such mental objects exist? Fischer (2011) argues that the concepts of ‘mind’ and ‘idea’ first made their debut in the West along with the rise of British Empiricism, during the early modern era – from the mid-17th to the mid-18th century. He holds that, at this point in history, ‘a new concept of mind’ as a kind of container of mental happenings was first forged (p. 35). On Fischer’s analysis, the new notion of mind – which was built on an analogy between perceiving and thinking – had two key aspects: it implied “a perceptual space in which objects of perception are located and a perceptual organ with which they are viewed” (Fischer 2011, p. 36). Accordingly, on this way of thinking about minds, “perception only occurs when some sort of idea is in the mind” (Fischer 2011, p. 36).

Once the perceiving-thinking analogy was firmly entrenched, the characteristic ideas inspired by this picture came to be treated as obviously true by many philosophers at the time – and yet, as Fischer emphasises, they are clearly “not part of common sense” (Fischer 2011, p. 37). Instead, they are “distinctively philosophical, and at the time, [this picture sponsored] fresh intuitions … shared, without explicit argument, by many early modern thinkers” (Fischer 2011, p. 37).

What might explain this? Fischer calls on psychology to assist philosophy at this juncture in his attempt to give a scientifically based explanation of how and why we are prone to certain philosophical pictures. In this vein, he claims that “a thinker is under the spell of, or adheres to, a philosophical picture iff he systematically makes non-intentional analogical inferences which assimilate targets to models of a conceptual metaphor, in ways the thinker knows them to be different” (Fischer 2011, p. 32).

Furthermore, Fischer (2011) maintains that “non-intentional analogical inferences lead to conclusions that the thinkers who make them are prone to find intuitively compelling” (p. 33). Hence, if Fischer is correct, once a thinker falls ‘under the spell’ of an analogy-sponsored picture, we can expect ‘musty’ ways of thinking to take over; we can expect thinkers to find certain things which would otherwise be evidently extravagant and controversial to be ‘obvious’ – and even deemed to be necessarily and unshakeably so. Furthermore, picture-sponsored convictions are regarded by those in their grip as legitimately established without the need for evidence or argument – and, indeed, the pictures that drive such thinking are without the possibility of evidence or argument.

Picture-driven thinking can thus explain the source of “common but otherwise apparently inexplicable” intuitions and how they come to be generated and held in place. According to Fischer’s explanation, “a conceptual metaphor can suggest suitable models to speakers of a language” (Fischer 2011, p. 35). Yet, it should be clear that the source of conviction sponsored by a picture does not stem from the uses of language when they are performing their everyday offices. Drawing on psychological theorising once again, Fischer (2011) proposes that “once picture-driven reasoning has led us to [a] philosophical conception, belief-bias effects may have us read this conception into ordinary talk” (p. 45).

Whatever we make of the details of Fischer’s proposed psychological explanations of what lies behind and holds picture-driven thinking in place, it should be clear that picture-driven thinking can inspire and sustain recalcitrant, evidence-resistant attitudes and commitments that cannot be overcome by confronting them with scientific evidence and explanations. Such attitudes and convictions only dissipate by means of special treatment.

Wittgenstein’s Razor is a fanciful term invented precisely to designate the special methods by which he proposed to deal with picture-driven convictions that are neither products of scientific theorizing nor clear-eyed observations about what presents itself to us in the manifest image.

It should be clear, given its function, that Wittgenstein’s Razor should not be confused with Occam’s Razor – at least in the latter’s most popular guise, according to which it is a parsimony principle that bids us to favour simplicity in theory choice when choosing between competing explanations.Footnote 14 That said, it is easy to see how Wittgenstein’s Razor can be used, side by side, with Occam’s Razor. Consider the proposed Buddhist explanation of how deep wisdom states are meant to be transmitted. This, so it is claimed, is achieved by means of subtle energies of a specific kind – an explanation that, as Rosch (2016) notes, challenges mechanist and materialist assumptions.

It might be thought that the proposed explanation, which points to subtle energies, is not necessarily incompatible, even with today’s science. After all, “We no longer think of physics as all about matter – there are fields and forces as well, at very least. Quantum mechanics has made even matter seem less material than it used to be” (deVries 2022, p. 111). However, this general observation provides little succour for those seeking to explain how wisdom states are transmitted by means of subtle energies unless it can be rigorously shown, not merely suggested, that the subtle energies in question can be incorporated into – or otherwise reconciled with – the posits of existing physics. Otherwise, the cost of taking such explanations seriously would push us to abandon, completely rewrite, or make inelegant ad hoc additions to hard-won physical theory.Footnote 15

So, all in all, this is surely a case in which something must give. ‘Never say ‘never’’, they say. After all, science is a turbulent business, full of adjustment and revision. We know the sciences are apt to change their accounts of what we find in nature over time. All the same, the devil is in the details, and in this case, the details would require us to reject physical theory or massively revise it to accommodate putative phenomena of dubious intelligibility. In this case, even to entertain such fundamental revisions to physics would be surely an instance of wrongly letting a mystical tail wag the scientific dog.

In light of all of these considerations, making room for the transmission of deep wisdom states emanating from dead high lamas by subtle energies looks like a parade case in which “one should not accept philosophical conceptions that openly contradict our best scientific theories but also that one should not accept conceptions that expand our ontology in areas that are already well accounted for by the explanations and entities postulated by science alone” (De Caro 2022, p. 213).

De Caro (2022) here gives some firm direction to the liberal naturalists when he tells us that “mystical intuition is not legitimate – insofar as it is ‘broadly incompatible’ with the scientific worldview” (p. 214).

Though the latter is true, the proceeding discussion has been at pains to highlight that, sometimes, philosophical considerations alone can supply a sufficient reason for ruling out some things from the domain of the natural. Wittgenstein’s razors and Occam’s razor do different sorts of work: this is why they can be used in ways that complement one another. Using Wittgenstein’s and Occam’s Razors together delivers the closest shave a naturalist can get.

3 A Positive Contribution to Metaphysics

Relaxed naturalists assume that both philosophy and the sciences can be called on to do more than merely providing a non-arbitrary means to identify and slice away the supernatural. They also hold that both philosophy and the sciences can do positive metaphysical work too – work which includes, in some cases, enlarging our initial, everyday MI account of what exists in nature. The sciences make positive contributions of this kind as they actively earn their explanatory keep. Relaxed naturalists allow that philosophy can also positively contribute to metaphysics, even assuming its aims are wholly clarificatory and descriptive.

What is LN’s stance on these matters? Going by Macarthur’s presentation, LN is somewhat conflicted in its official statements about its stance on metaphysics. On the one hand, he tells us that:

Naturalism is, or should be, a non-metaphysical program of philosophical explanation given that naturalism of any stripe is committed to anti-supernaturalism and a broadly empirical attitude to inquiry. (Macarthur 2015, p. 567).

Yet, on the other hand, he also tells us:

Liberal Naturalism can admit … non-scientific, non-supernatural realities into the catalogue of the things it acknowledges – or what we might (with some reservations) call its ‘ontology’ in a deflationary sense of that term … (Macarthur 2015, p. 574).

And, in an accompanying footnote, he adds:

Of course, liberal naturalists respect the results of the natural and human sciences and, like any good naturalist, admit all of the scientific entities posited by successful scientific explanations into its ontology. (Macarthur 2015, p. 574).

Why fore the apparent conflict? Those, like Macarthur, who follow Wittgenstein’s lead, tend to be suspect and shy of metaphysics across the board. They assume that it is not the job of philosophy to add anything to our inventory of what there is in nature beyond helping us to recognise what is already there, as revealed by an appropriately refined MI. In this regard, by Wittgenstein’s lights, philosophy is not like the sciences: its sole job is to clarify, reveal, and describe what we are committed to, ontologically speaking, in our everyday practices. Curing us of the temptation to add superfluous entities into our ontology when we are under the sway of ultimately superstitious pictures is a means to that elucidatory end. On this conception of what proper work in philosophy looks like, the biggest mistake we can make is to think that philosophy operates in anything like the way the sciences operate. This, Wittgenstein held, “is the real source of metaphysics, and leads the philosopher into complete darkness” (1958, p. 18). If this is so, it is crucial to avoid thinking that philosophy is in the business of offering metaphysical, ontology-expanding explanations in the way that the sciences do.

Bearing this in mind, we can make sense of the divide that Moyal-Sharrock (2021) sees between Wittgenstein’s commitment to what she describes as a ‘soft realism’ and his utter rejection of any kind of metaphysics.Footnote 16 That distinction lies at the heart of her claim that Wittgenstein advances a “realist or realistic philosophy, as opposed to … metaphysical philosophy” (Moyal-Sharrock 2021, p. 10). These considerations also explain why Macarthur (2018) insists that LN must steer clear of explanations fuelled by commitments to misleading metaphysical pictures, and that, for that reason, all good, naturalists must work actively to ‘overcome metaphysics’.Footnote 17

Wittgenstein’s diagnosis of the general source of supernatural thinking highlights our tendency to succumb to such thinking because of explanatory urges. These observations provide the backdrop for his famous injunction that when doing philosophy, one must only describe and not explain: “we may not advance any kind of theory. There must not be anything hypothetical in our considerations. We must do away with all explanation, and description alone must take its place” (PI § 109).

Methodologically speaking, once we do the philosophical work needed to free us from distorting pictures, Wittgenstein tells us, “Philosophy just puts everything before us, and neither explains nor deduces anything. —Since everything lies open to view, there is nothing to explain” (PI § 126). Some are inclined to encapsulate the above lessons in the form of a rule: “It is the task of the scientist to explain; the philosopher must only describe” (Moyal-Sharrock 2021, p. 4).

The trouble is that if these Wittgensteinian considerations leads one to hold, as Macarthur does, that LN is ultimately non-metaphysical or, indeed, anti-metaphysical, then it rightly provokes the following response: “isn’t liberal naturalism itself ultimately a metaphysical thesis, based on specific metaphysical presuppositions?” (Morganti 2022, p. 251).

The question is reasonable. What can be said by way of reply? First and foremost, the Wittgensteinians can consistently hold that clarifying and refining our understanding of what we ordinarily take to exist – namely, by bringing the MI into sharper focus by removing distortions and making it clearer – need not be a matter of advancing and testing theories about what exists in nature. This is so even though one might do a better or worse job in one’s elucidatory efforts to make the MI properly perspicuous. As such, the results of such philosophical work might reasonably be rejected by another as inadequate. Even so, it does not follow that such attempts at clarification involving offering up hypotheses, theses, or theories (for a similar analysis along these lines, see Moyal-Sharrock 2021, pp. 198–226).

Nevertheless, Morganti (2022) would be correct to say that the kind of clarificatory work just described positively contributes to metaphysics – at least where the notion of metaphysics at play is the innocent one of saying what does and does not exist in the world. Here we might follow Thomasson (2015) in thinking that advancing a simple – and appropriately deflationary – realism is not at odds with metaphysics per se but provides “an alternative way of looking at that particular corner of metaphysics which deals with existence questions” (p. 11, see also p. 21).

Where does this leave us? If we concede that elucidatory philosophical work can reveal some of what there is, does it follow that Morganti (2022) is also right in thinking that “at least some of the concepts and categories employed by analytic metaphysicians need not be rejected as supernatural” (p. 246)? Certainly, the very idea that we must adopt a critical stance to refining the MI implies that “not all items belonging to the manifest image are on a par, and it is much wiser to proceed on a case-by-case basis” (Morganti 2022, p. 248).

Let’s examine a case in point. Consider Morganti’s (2022) suggestion that “there is no reason for not including [numbers] in the conception of nature endorsed by liberal naturalists” (p. 250).

We readily refer to numbers and make use of them in our everyday mathematical practices, indeed we have built a science around them, and their use is indispensable in other scientific practices. In our everyday dealings with numbers, we treat them as if they had objective properties that do not depend arbitrarily on how we think of them. On the face of it, then, naturalists need to take numbers seriously. But, as everyone knows, there are several developed philosophical options for doing so – ranging from full-blown Platonism to fictionalist, rather than eliminativist options.

Morganti (2022) wonders whether Platonism about numbers could be a live option for liberal-minded naturalists. He openly admits that Platonism is neither required by mathematics nor our everyday practices and that, as such, “there is a sense in which Platonic entities are not explanatorily necessary from a naturalistic viewpoint” (p. 250). Even so, he suggests that:

… additional philosophical considerations (in this case, about indispensability in scientific practice) leads to explanations that are overall more comprehensive. That is, to explanations that may add little or nothing to science per se and are not grounded in common sense, yet contribute to defining a larger picture of reality. … [As a result,] Platonism, which certainly is not part of the manifest image, could plausibly be regarded as part of our explanation and understanding of Nature broadly understood (Morganti 2022, p. 250).

The thing is, if we are talking about textbook Platonism, then naturalists will have every reason to look for other ways of making sense of how numbers should be accommodated into our worldview – if at all. And here, following Thomasson (2015), it proves useful to distinguish between the approach and motivations that drive one to adopt simple realism as opposed to explanatory realism about numbers. As she notes, both agree in the first-order existence claims they make about numbers; but unlike explanatory realists, simple realists do not posit or attempt to justify the existence of numbers in order “to explain our number talk, its objectivity, its usefulness in science” (Thomasson 2015, p. 155). As such, though simple realists recognise that numbers exhibit certain features – for example, that they are abstract – they are not committed to any explanation of why they are abstract or any particular understanding of what such abstractness entails.

Certainly, to think of numbers as abstract entities that stand outside the causal order generates standard puzzles about how we could ever gain access to things with that sort of imagined metaphysical status (see, e.g., Douglas 2022). For this reason, including such things in our naturalistic metaphysics wouldn’t give us a better overall account of how things hang together. This is especially so once we realise that our “Cognitive activities do not require commitment to the existence of non-natural entities such as abstracta” (deVries 2022, p. 111).

Given this, even if we recognise that numbers exist and that they are abstract, following, say, Thomasson’s (2015) simple realism, naturalists have every reason to actively develop alternatives to Platonism when it comes to making sense of the ontological status of numbers. Glock (2002) identifies Strawson’s descriptive metaphysics as a possible way of taking numbers seriously “without committing itself to the existence of entities which are part of a quasi-causal supernatural order” (Glock 2022, p. 452). Strawson allows for the existence of abstracta but thinks of such entities as being aspatial and atemporal instead of their standing outside of space and time.

Moreover, Strawson accepts the burden of clarifying what it means for such entities to exist. As Glock (2022) reports, Strawson attempts to do this “by referring to cognitive capacities of the kind highlighted by anthropological naturalism – capacities for identifying and classifying particulars in the case of universals, capacities for linguistic understanding in the case of meanings” (p. 452).

Notably, what is on offer from Strawson is not any kind of picture-driven speculative metaphysics of the sort to which Wittgenstein objects. Perhaps Strawson’s effort to clarify how numbers fit into our ontology won’t work out in the end. Perhaps there is a better way to understand how numbers are real. Nevertheless, and this is the point, whatever we conclude about Strawson’s proposal or others like it, as philosophers, in working it out, we will positively contribute to the metaphysical project of naturalism in ways quite different to that of scientists.

4 Conclusion

There is much more to be said about how philosophy can contribute to our understanding of the natural world by adopting a positively relaxed naturalism. For example, beyond bringing the refined manifest image into clearer focus, philosophy, along with the sciences, also does important synthesising work in helping us understand how aspects of the natural world connect. Arguably, such work is not best understood either as purely descriptive or as purely speculatively theoretical in character.

Yet, even if we only focus on the cases discussed in this paper, we have sufficient reasons to conclude that a deeper understanding of the positive contributions that philosophy and the sciences make to our understanding of nature requires embracing a relaxed, though not too liberal, naturalism.