What Needs to Be Explained

Science, in fact, gives us great powers of prediction and control. What more should we want from science? What is it, exactly, that the realist thinks science does that the pragmatist cannot explain? Scientists and philosophers of a realist bent have regularly insisted that we should want more, much more, from science than mere pragmatic success in prediction and control, and that science can or does deliver such things.

One common refrain is that what is central to scientific progress is basic or pure science, the pursuit of scientific knowledge wholly independent from our practical aims, interests, and activities. Such science pursues or arrives at Truth in the sense of accurate representation of Reality, or at least knowledge of the deep structures or unobservable features of our world. Science aims at, and its success is judged by, not only increasing success in prediction and control but also increasing our power of explanation and understanding of the world, judged according to a set of explanatory or superempirical virtues or epistemic values; these virtues or values guide inquiry as much as the pragmatic ones, says the realist. Ultimately, the goal is not just instrumental but also to arrive at a full scientific world-conception or worldview in which we can understand the universe and our place within it. Indeed, realists commonly argue that the high degree of predictive, experimental, and technological success science has achieved would be a miracle if its theories were not in fact tapping into deep truths about the nature of the world beyond our senses. It is generally thought that this set of values, aims, goals, and achievements cannot be accounted for by the pragmatist. Let us try to get clear about what, exactly, the lacunae are supposed to be, and then determine which pose genuine problems for the pragmatist.

As a preliminary point, while some pragmatists of the past may have been committed to philosophical views that would prevent them from acknowledging that science posits unobservable entities, the contemporary pragmatist has no such compunctions. If electrons, quarks, markets, mental states, laws, kinds, or what have you play significant roles in bodies of knowledge that enhance our abilities to predict and control, many contemporary pragmatists have no qualms about them as objects of knowledge realistically construed. There is no reason that the pragmatist need be a strict empiricist. Indeed, the classical pragmatists frequently criticized the traditional empiricists for their view of experience; rather than understanding experience as composed of atomic sense-data (a bundle of independent and simple sensations such as color and shape), the pragmatists saw experience as having depth, structure, and continuity. There’s no reason that a pragmatist cannot say, first, that our main contact with electrons concerns what they can help us predict and control, and second, that on that basis we understand them as real elements of the furniture of the world.

Preliminaries out of the way, there are three challenges to the pragmatist in accounting for these aspects of science that seem to go beyond mere prediction and control.

First, there is the question of “pure” or “basic” research with no obvious or immediate applicability. In such cases, scientists are surely doing scientific inquiry, and that inquiry is aimed at expanding our powers of prediction and control; but being able to predict and control those particular phenomena serves no particular use that we can foresee. That might be because the objects of that research are distant in time (paleontology and the biology of dinosaurs) and space (astronomy and the distant stars and galaxies). It might be because there is nothing inherently interesting about the subject to anyone but the scientists who study it, perhaps because it is too removed from our common experience and no technological application has been conceived (much recent high-energy particle physics). It might be because the nature of the phenomenon forbids fine-grained prediction or any intervention whatsoever, such that no utility is on offer (physical cosmology). How can the pragmatist account for the value of such inquiry?

Second, there is the question of virtues, values, standards, or criteria for science that go beyond the pragmatic concerns of prediction and control. We can think about this challenge in a few different ways. Such superempirical epistemic standards might include things such as a scientific theory’s explanatory power, simplicity, unifying ability, or fruitfulness for future research. Some have tried to give pragmatic justifications for these standards, arguing for example that they make a theory easier to use, and thus more testable. In this case we reduce the supposed nonpragmatic virtues to pragmatic ones. Some would argue that such standards come in only when the evidence has run out, when empirical and pragmatic factors underdetermine theory choice. If we hold that these superempirical standards are on a par with pragmatic and empirical criteria, then there may be contexts where we choose less accurate and less “useful” theories, because, say, they provide simpler, more unifying explanations. (If we would never do this, then those standards are not actually on a par, and the challenge to the pragmatist is minimized.) If choosing such theories is a reasonable way to proceed, the pragmatist must be able to account for it.

Third, there is the question of scientific worldviews. What I call “construction of a scientific worldview” is an important part of the creative and constructive activity of science that does not consist of empirical inquiry into specific phenomena. Scientific theorizing also involves synthesizing across a wide range of empirical inquiries, in order to provide a larger picture of the universe (and our place in it). This form of theory construction typically builds on past achievements of observational and experimental research, but it need not and often does not have much direct contact with empirical inquiry itself. The grand theorizing by figures such as Newton and Einstein sometimes have such a character; so does the work of synthetic popularization by figures such as Stephen Jay Gould, Richard Dawkins, Stephen Hawking, and Carl Sagan. Sometimes this work inspires, redirects, or guides future empirical inquiry, and so in retrospect its pragmatic value seems clear. In other cases, often the most ambitious examples of such work, the connection to particular empirical inquiries remains tenuous. The significance of the latter seems very difficult for the pragmatist to explain.

In what follows, corresponding to these three challenges, I seek to provide some basic evaluation from within a broadly pragmatist point of view of the following:

1. (1) Scientific inquiry that is “useless.”

2. (2) Scientific inquiry guided or judged by standards other than prediction and control.

3. (3) The construction of a “scientific worldview.”

In the following three sections, I explore these three topics in turn. As we will see, the first two of these challenges can be handled in a relatively straightforward way. The third challenge, however, will require us to explore in depth pragmatist views about religion and humanism in order to fully assimilate the positive significance of scientific worldviews into a pragmatist approach.

The Pragmatic Value of “Useless” Science?

In a way, ironically, those inquiries that are supposedly impractical are the easiest for the pragmatist to account for. The results of a scientific inquiry may not be “useful” in the narrow sense of immediate applicability in medicine, engineering, or policy; but nevertheless, as genuine inquiry, it might be governed by the broadly pragmatic criteria of prediction and control. Still, insofar as the pragmatist emphasizes the increase in our prediction and control of phenomena that interest us, there remains a problem of accounting for why we would be interested in such recherché phenomena as basic science often tackles, such as the biology of prehistoric creatures or the behavior of distant stars.

In “Genuine Problems and the Significance of Science” (Brown Reference Brown2010), I worried about this issue in the context of thinking about Philip Kitcher’s (2001) account of “scientific significance” and the aims of science. According to Kitcher, the “significance” of a scientific problem or scientific inquiry can be understood as its place in a network of interconnected aims and projects, and the grounding points in this network, from which all significance ultimately flows, are obvious practical uses, on the one hand, and questions of what he calls “natural curiosity,” on the other.

Consider the work of an auto mechanic. A mechanic is an inquirer, engaged in problem solving in relation to the diagnosis and repair of malfunctioning vehicles. The work is not quite scientific, although it is notoriously difficult to draw such lines, but it draws on some science and engineering knowledge, design specifications of the vehicle, manuals, heuristics and rules of thumb, intuition hard-won by experience, and a good bit of guess-and-check. It is typically pretty unsystematic, ad hoc, in response to the specific case in front of the inquirer.

Suppose the auto mechanic consistently runs up against a problem that cannot be solved with the resources available to them. After reflecting on the pattern of failures, they determine that the source is not in a lack of skill or knowledge on the part of the mechanic but with some principle of engineering that they regularly rely upon. In most circumstances, this principle helps the mechanic in the repair of diverse automobile engines. But in a certain number of cases, their inquiry fails, and no successful repair can be made. Most mechanics would just accept that some cars cannot be fixed, but our mechanic is particularly dogged and becomes so consumed with the solution to the problem that they go to school for an engineering degree, hoping to determine its source, and so becomes eventually a working researcher revising the very principles of engineering that they once used as a mechanic.

Suppose our newly minted engineering researcher consistently runs into trouble when dealing with particular principles learned from basic physics. Often those principles serve them well, but on certain occasions forming a pattern, they fail to aid the engineer in their inquiry. Our engineer reads more and more about the physics involved and realizes that the ultimate source of the problem is a gap in our knowledge of physics, finally pursuing another degree in that field so as to work on revising our understanding of the laws of physics.

In this fantasy story, we see how the work of a physicist might have traceable lines to the work of the engineer, even the auto mechanic – and such lines (understood conceptually rather than embodied in a single person) are part of the story of significance for Kitcher. But other work in physics does not seem to have such clearly traceable lines. Nevertheless, the physicists working on problems without such traceable lines of connection to practical concerns are drawn to those problems as much as our fantasy mechanic-turned-engineer-turned-physicist. According to Kitcher, there are some kinds of questions about the nature of the world, life, and human nature that we are all naturally curious about. A similar story could be told in terms, not of practical inquiries, but of these questions of natural curiosity. According to Kitcher, for even seemingly abstract and technical scientific inquiries, we can trace their significance back to a combination of all the practical problems and questions of natural curiosity that they bear on in some way. The amount and strength of such connections helps us compare the significance of different scientific projects.

The concern I raised in my earlier paper in response to Kitcher was that when we start weighing the value of different projects, on Kitcher’s view, and we have to rank projects that might, for example, contribute to reducing worldwide deaths from malaria (or cancer or COVID-19) with projects that mainly satisfy our “curiosity,” the latter would be totally swamped. “Curiosity” seems inadequate to defend anything like a robust program of basic research whose significance is largely basic knowledge rather than practical results, given the wide range and depth of immediate practical needs that scientific inquiry might help us meet. In other words, Kitcher’s account of significance seemed to me unable to provide the defense of basic research that he seemed keen to provide. So, although Kitcher’s account seeks to defend basic research with little practical application on the basis of our natural curiosity, it seems like his account will systematically devalue it in favor of practically significant inquiries.

Perhaps this devaluation is the right approach, though. After all, when it comes to possibly saving human lives or satisfying our curiosity about whether megafauna from tens of millions of years ago had feathers, does it not seem inhumane to prefer the latter? When these trade off, should we not obviously prefer the former? The pragmatic point of view seems to be understood that way.

Scientists develop conceptual, material, explanatory, and methodological resources in systematic ways. They are often driven by the existence of a difficult puzzle that only a specialist can understand as a puzzle. Some of these have the sort of obvious lines of relevance to practical problems that I described earlier. Other puzzles may only have relevance later, when systematic generalization of the puzzle solution is achieved, and the practical payoff can be seen. Vannevar Bush made a strong claim that basic science would inevitably yield useful by-products. At the same time, he held that scientific progress depended on it being unconstrained by a focus on practical results: “Scientific progress on a broad front results from the free play of free intellects, working on subjects of their own choice, in the manner dictated by their curiosity for exploration of the unknown” (Bush Reference Bush1945: 12). In this strong form that guarantees progress and applicability from unconstrained basic research, this argument is untenable, because often basic research fails to translate to application, while mission-driven, applied research is much more fruitful on its own than Bush would admit (Sarewitz Reference Sarewitz2016). Still, this is one path by which seemingly useless science sometimes proves its use, through unanticipated future application, even when Kitcher’s “lines of significance” cannot be traced beforehand.

Perhaps part of the problem concerns the way that we think about what is or isn’t “useful.” There is a tendency, under the regime of neoliberal capitalism, to assimilate the “useful” to the economically productive; to see scientific knowledge as useful if it contributes to technology that sells or to technocratic solutions to policy problems. Here, turning to classical pragmatist philosophy and particularly Dewey’s reflections on the concept “useful” provides a crucial corrective. In Art as Experience, Dewey writes:

And again in Experience and Nature:

Here Dewey captures and responds to a common misconception of pragmatism and “instrumentalism,” and helps us resolve some worries about “useless” inquiries and the value of curiosity. Dewey points out that the definition of “useful” must be situated not in the demands of capitalist modes of production but within human experience, where what is useful concerns not only survival but flourishing. In this sense, delight, wonder, the enjoyment in finding a clever solution to a difficult puzzle, are all definitely useful, perhaps even more useful than those results that “contribute” in an economic sense. In any case, insofar as scientific practices of inquiry enable forms of prediction and control that contribute to the enrichment of life and experience, our “appreciation of the meaning of things,” then they are useful and significant.

Nonpragmatic Criteria in Science

Some philosophers of science insist that there are standards or criteria for hypothesis acceptance or theory appraisal that are independent from the criteria of successful prediction and control and that these nonempirical criteria may in some cases supplement or even outweigh empirical prediction and control. In some cases, it is taken as simply a brute fact about science that it answers to such nonempirical standards or “epistemic values” (Kuhn Reference Kuhn1977). Others see these “superempirical virtues” as independent grounds for belief in the truth of a scientific theory, and thus part of a robust case for scientific realism (Churchland Reference Churchland, Churchland and Hooker1985). My response here is relatively brief: These claims are largely confused; either the criteria in question reduce to or are instrumental to prediction and control; or they are not criteria that guide genuine scientific inquiry and belief.

One version of the idea of nonpragmatic criteria for science is simply that science aims at more than prediction and control; it aims at truth. Another version is that certain superempirical epistemic standards or values are valuable because they are truth promoting. We could mean two things here by “truth.” One thing we might mean is true predictions of observation and experiment. This is just to restate the claim that science aims at prediction and control, rather than being an alternative to it.Footnote ³ Or we could mean truth in a broader sense, the truth of the whole theory in all its parts, not just the truth of the predictions it makes. Truth in this sense by definition exceeds prediction and control; but also truth in this sense cannot be considered an independent aim or standard for anything. Our only means of assessing whether we have come near to the truth in this sense is our assessment of the success of theories in facilitating successful prediction and control. We have no additional access to the truth, no way to aim at it, other than through our most successful science.Footnote ⁴

This is not to deny that science achieves the truth sometimes; that is a separate issue. Nor is it to deny that science might aim indirectly at the truth, by aiming at successful prediction and control. But the idea that science aims at truth directly and independently as such is a misunderstanding of what an aim is. To be able to aim at something, one must have some sense of how the aim might be achieved, how to recognize whether the aim has been achieved, or whether one is moving close toward achieving the aim, however indirectly. In the case of truth in a sense that goes beyond accurate prediction, we cannot meet any of these requirements (Laudan Reference Laudan1984: 137). If truth means that our theory accurately pictures the “underlying reality” beyond our concepts and observations, then we have no way of getting outside of our experience and conceptual frameworks to compare the picture with what is pictured. Likewise, if we want to know whether certain superempirical virtues of theories are truth promoting, not in the sense of enabling better prediction and control but in the broader sense, we have no independent grounds to answer this question. At best, we can say either that theories with such virtues tend to have characteristics that match our assumptions about what the world is like, or that they are instrumental to better prediction and control of phenomena.

An argument from the pragmatist philosopher Charles S. Peirce suggests a third option for thinking about truth as an aim (or, in his terms, “regulative ideal”) of science. According to Peirce, the truth is whatever belief scientific inquirers would tend to settle on in the long run of inquiry; a belief (result, hypothesis, theory) is true if it would withstand testing and evaluation in every test scientific inquirers might subject it to, without falling into doubt. This is sometimes taken to be a pragmatist definition or theory of truth, and often not considered a plausible one.Footnote ⁵ It is probably better understood as an elucidation of the relations between truth, belief, and inquiry (Misak Reference Misak2004). This is of a piece with various other attempts by pragmatists to redefine or elucidate truth in ways that make it a tractable aim: as successful belief (James), as unrevised in the ideal limit of inquiry (Peirce), as warranted assertibility (Dewey),Footnote ⁶ or as ideal rational acceptability (Putnam Reference Putnam1981). This pragmatist line intrinsically links truth to success in prediction and control, rather than treating it as an independent goal or standard.

One could also understand the long-run achievement of truth as a substantive claim rather than a definitional move: If there is a truth of the matter, then scientific inquiry, taken to indefinite lengths, would settle upon it. The long-run success of science is thus evidence that science aims at truth. This is a common claim of the scientific realist. However, truth in any of these senses is, again, not an aim that inquirers could have in view instead of successful prediction and control. Rather, the claim is that in thoroughly exploring improvements to successful prediction and control, inquirers get the truth, in the long run. In this way, “truth” is aimed at only indirectly, through the pragmatic aims of prediction and control. The former does not provide independent standards of evaluation.

Some philosophers of science have argued for certain nonpragmatic standards for scientific inquiry as being intrinsic scientific criteria in their own right. Thomas Kuhn, for example, articulates five criteria for good scientific theories: “accuracy, consistency [with other theories], scope [of phenomena encompassed by the theory], simplicity, and fruitfulness [for future research]” (Kuhn Reference Kuhn1977). The first, accuracy, is just another term for predictive success. The other four are ambiguous: They might be understood purely as features of the theory or relations between theories, or they might be understood as characterizing relations between theory and evidence (Douglas Reference Douglas2013). Simplicity understood as a feature of a theory might be characterized ontologically (number of theoretical posits required), mathematically (as a property of the equations or models constituting the theory), computationally (in terms of the difficulty of making calculations), or otherwise. Understood as a relation between theory and data, simplicity means that the theory has a lower degree of complexity than the evidence it covers (Douglas Reference Douglas2013: 799). If these nonpragmatic standards (also called epistemic standards, superempirical virtues, epistemic values, or cognitive values) are ways of characterizing the relation between theory and evidence, then meeting them is either instrumental to greater success in prediction and control, or they pick out a particular type of prediction or control as particularly valuable. They do not constitute a standard independent from success in prediction and control. On the other hand, if we conceive of standards such as simplicity as inherent properties of theories independent of their relation to evidence, then such standards cannot be criterial for science at all.

To sum up what has been said so far, we have considered two types of scientific inquiry that at first glance seem not to fit the pragmatists’ account of science: “useless” science that lacks immediate applicability, and science guided by nonpragmatic criteria. In both cases, we have found that the pragmatist can fully accommodate the value of the relevant science while clearing up certain misconceptions. But not all scientific activity can be understood as problem-solving inquiry directed at our capacity to predict and control. Much of the science that we find really inspiring, that informs both public understanding of science and science education, consists of attempts to synthesize and build on the results of pragmatic inquiry in order to understand how it all fits together. I group such attempts under the traditional heading of “the scientific worldview” or “scientific worldviews” in the plural; it is these that constitute the greatest lacuna for the type of pragmatism under discussion here.

Scientific Worldviews

Religionists are often inclined to see the tension between science and religion in terms of a clash between very different worldviews. “The scientific worldview” is variously depicted in terms such as materialism, mechanism, determinism, and reductionism, and as opposed to ideas such as spirituality, idealism, magic, and miracles. An early exemplar of scientific worldview building is René Descartes’ treatise The World, written between 1629 and 1633 and published posthumously (1998/1677). The book combines epistemology, physics, biology, and metaphysics to paint a picture of an entirely mechanical understanding of the physical world encompassing the nature of matter and light, astrophysics, living organisms, and the mechanics of perception (with room, however, for God and the rational soul). Descartes sought to provide a complete and systematic alternative to the worldview late medieval philosophers had created in synthesizing Aristotle and Christianity; in the process he synthesizes new scientific research and ideas with his own creative speculations. Many scientists, philosophers, educators, and popularizers have followed in Descartes’ footsteps, attempting to build on the latest science to create a comprehensive account of the nature of the world.

How are the various results of science synthesized into a scientific worldview? It is a more difficult and complex matter than it may seem. On the ground, we see diversity and disunity in science. Science involves a hodgepodge of approaches, theories, concepts, and conflicting results. A frank survey of everything actually going on in science shows that it provides no single map of reality, that it has little or no overall organization to its theories and methods, and that inconsistencies abound (see Feyerabend Reference Feyerabend and Terpstra1999; Dupré Reference Dupré1993; Galison and Stump Reference Galison and Stump1996; Kellert, Longino, and Waters Reference Longino and Galavotti2006). Even fundamental physics contains different approaches that are apparently inconsistent and so far resist combination into a successful, testable theory. Beyond the realm of theoretical physics, we’re in even more of a mess. On a pragmatic and contextual account of science as problem-solving inquiry, aimed at prediction and control of phenomena that interest us, this is not much of a surprise. We might well expect science to be as diverse as our interests. In everyday science, there is no need to take it all together – specialization and contextualism help us keep the mess in hand, and localized conflicts are a spur to further inquiry.

But all this plurality, as Paul Feyerabend aptly pointed out in his later work (1999), problematizes the notion of the scientific worldview. One way we might articulate “the scientific worldview” would be a thorough survey of this mess, a list of achievements, gaps, and internal clashes. This would be pretty convoluted, to put it mildly, not to mention self-contradictory, and it would not do the job the scientific worldview is thought to do; when we ask how it all fits together, we expect something more than concatenation in response. Instead, we need to ask what happens when we try to create, from a survey of this mess, a single, coherent worldview. Thanks to the messy reality of scientific practice, any attempt to craft a coherent worldview based on science is not a straightforward matter.

The construction of a scientific worldview has three features. (1) it is selective; it leaves a lot out, and emphasizes certain aspects of science over others. (2) it is constructive and creative; it stitches the remaining pieces together into a coherent and compelling story. Finally, (3) it is philosophical; it is part of metaphysics or ontology, not on the same footing as ordinary scientific inquiry, whose warrant is largely connected to situations of practical problem solving and successes in prediction and control. It goes beyond particular empirical problems to paint a grand picture but not a uniquely compelling one. In other words, the scientific worldview cannot be read off of science directly but must instead be constructed by creative, philosophical interpretations that select certain elements of science for emphasis. Because there is a degree of free choice in deciding how to construct a worldview from the materials of science, it is probably best to think about this in terms of multiple, potentially competing scientific worldviews in the plural.

Any worldview has consequences for our lives, hopes, and sense of purpose. In Science and Moral Imagination, I describe worldviews as “complex evaluative standpoints where particular valuations are tied up with more general ideals, principles, and institutions, as well as factual beliefs, theoretical claims, and metaphysical commitments” (M. J. Brown Reference Brown2020: 142). This emphasizes the fact that our sense of how it all fits together is never a neutral, disinterested matter, but one that touches on questions about the meaning of human life and our place in the world. This connection should come as no surprise to those familiar with constructions of scientific worldviews. Many of the most creative articulators and passionate defenders of versions of the scientific worldview – such as Jacques Monod, Carl Sagan, Richard Dawkins, Neil deGrasse Tyson, and Jerry Coyne – have explicitly drawn out moral or political lessons, or have argued that the scientific worldview challenges not only earlier ideas but traditional ways of life.Footnote ⁷

As such, scientific worldviews closely connect with the traditional function of myth or religion, and a scientific worldview will have to compete with other worldviews – traditional religious, speculative philosophical, and alternative interpretations of science. These worldviews, because they are so loosely connected to the pragmatic dimensions of science, will be judged less by typically scientific standards than by philosophical, aesthetic, and ethical ones.

Why do we need a scientific worldview at all? What is the pragmatic value of such a thing? Why should we not simply rest content with the various results of particular scientific inquiries and leave these further questions to theologians and metaphysicians (or perhaps just leave them alone)? I think there is a story for the naturalistic pragmatist to tell about why we should want a well-crafted scientific worldview. To tell it, I turn to what might seem like an odd source: pragmatist analyses of religion and religious experience.

Pragmatist Accounts of the Religious

Recall a part of the earlier quote from Dewey’s Experience and Nature: “The characteristic human need is for possession and appreciation of the meaning of things” (1988/1925: 272). We can join this to a statement from Dewey’s pragmatist forerunner James that “the life of religion … consists of the belief that there is an unseen order, and that our supreme good lies in harmoniously adjusting ourselves thereto” (James Reference James1902: 53).Footnote ⁸ This starts to give a flavor of the classical pragmatist philosophy of religion. According to James and Dewey, religion speaks to deep needs in human life and experience, for meaning on the one hand and guidance on the other. Insofar as religion allows us to understand and appreciate the meaning of things and to adjust ourselves to the world in a way that promotes the good, it thus performs a pragmatic function.

These pragmatist thinkers draw a distinction between two aspects of religion as ordinarily understood. On the one hand, there is institutional religion. Both thinkers exclude the institutional side from their positive account of the value and function of religion. They considered organized religion a secondary development that at best did not get at what was centrally important about religion; at worst, they saw religious institutions as tied to “creeds and cults” that interfered with the expression of religious experiences and values (Dewey Reference Dewey and Boydston1989/1934: 21). On the other hand, there is what James calls “personal religion” and what Dewey calls “the religious,” which they see as playing a positive role in helping us understand the world and our place in it, thereby providing a solution to our uneasiness about ourselves and our standing in the world, as James puts it (1902: 508).

Where James and Dewey differ is on the question of whether the supernatural is an essential feature of religion. For James, religion requires not only some “unseen order” but also a mystical or supernatural order that resonates with the “higher part” of our own being (1902: 508). On a Jamesian account, then, it seems doubtful that a scientific worldview could perform the religious function. Dewey believes, on the contrary, that supernaturalism is an addition to the religious from the institutional side of religion, and that we can liberate the religious to better function in our lives by removing the supernatural accretion.

Supernatural belief hardens into dogma as a result of institutional forces; Dewey explores this claim at length in A Common Faith (Reference Dewey and Boydston1989/1934). But supernaturalism has its origins, not as a primitive attempt at science or philosophy, but in an aesthetic urge, Dewey argues in Art as Experience:

In sum, religion functions as much or more so on the aesthetic plane than on the intellectual, scientific, or philosophical. This is not to dismiss the value of religion in the least; according to Dewey, the aesthetic is also the realm of experience where meanings are at their fullest. Insofar as the role of religion is to give us and help us appreciate the meaning of things, we must ultimately operate on the plane of artistic expression as much as if not more so than the merely cognitive or intellectual.

The function of religion is not only aesthetic, however, but also moral. In A Common Faith, Dewey emphasizes both the continuity of humanity with nature and the continuity of the human community past, present, and future, calling this the “community of causes and consequences” that is “the widest and deepest symbol of the mysterious totality of being the imagination calls the universe” (Reference Dewey and Boydston1989/1934: 56). The interacting network of human beings with nature throughout time is the environment in which our ideals, aspirations, and values are formed by acts of moral imagination, which unifies the aesthetic and moral functions of religion with its attempt to wrestle with the universe as a totality. As we use our moral imagination to shape the purposes of our lives, we make our character as we make a work of art (Fesmire Reference Fesmire2003: 107).

It is in this relation between our values and the “mysterious totality” of the universe that Dewey, contra James, gives a naturalistic analysis of the role of faith and of the religious in experience. That is, he attempts to give an account of faith and religious experience without any reference to supernatural entities such as spirits, souls, immaterial substances, and so on. Dewey’s hope is that a naturalistic picture can help provide what traditional religion once provided in terms of social cohesion and personal meaning, but for a secular, humanistic, democratic world. One of the core concepts in Dewey’s account is the concept of natural piety:

To practice natural piety is to approach the world with humility and reverence but not passive fatalism. In a secular age, it may be difficult to see the value in talk of “piety.” But for Dewey’s part, natural piety is a much-needed perspective attuning us to our dependence on the world for the success of our endeavors. We are in the world, Dewey is saying, not in the way that a button is in a box, but as part of a complex network of interacting dependencies. This should be a cause for humility and reverence. We are parts of the world that can act toward a reflectively chosen end to improve our lot. We form purposes and ideals reflecting not only momentary desire but our attempt to understand what is ultimately desirable.

It is in this capacity for reflective or intelligent action that Dewey finds room for naturalistic interpretations of faith and the divine. Dewey defines faith as “the unification of the self through allegiance to inclusive ideal ends, which imagination presents to us and to which the human will responds as worthy of controlling our desires and choices” (Reference Dewey and Boydston1989/1934: 23). To have faith is to believe in one’s ability to bring such ideals to realization through our desires and choices, even if only in the long run. Similarly, Dewey defines God or divinity in naturalistic terms as the unity of our ideal ends, “the values to which one is supremely devoted,” in our imagination (Reference Dewey and Boydston1989/1934: 29). It is important to Dewey that such ideals are neither (yet) actualized, nor “mere rootless ideals, fantasies, utopias” (Reference Dewey and Boydston1989/1934: 34). Rather, these ideal ends are possibilities made coherent through action in connection with conditions in nature that promote their realization. As Dewey says, “It is this active relation between ideal and actual to which I would give the name ‘God’” (Reference Dewey and Boydston1989/1934: 34).

Here, aesthetic meaning, moral values and ideals, human ingenuity and intelligence, and social cooperation come together to form secular concepts of faith and the divine (if words such as “God” and “the divine” are too inseparable for you from supernaturalism, consider instead using a term such as “the sacred” or “the spiritual”). Central, again, is the notion of natural piety, the humility and reverence for nature that makes our lives possible, as well as a faith in the human community to cooperate toward realizing those ideals.

According to Dewey, militant atheism and modern supernaturalism are allied in presenting an image of humanity in isolation from nature, and thus the negation of natural piety.Footnote ⁹ Against both, Dewey holds that the religious quality of our experience, our worldview, and our way of life is ineliminable but must be situated within the natural world of our experience. In the present day, it seems that natural piety and faith in our ideals and our community are precisely what we need more of. We need a common, secular, democratic faith that can support and encourage them.

The Religious Function of Science

Here I think we can begin to address the religious function of science. The synthetic and visionary parts of science associated with the articulation of a scientific worldview can, I believe, help fulfill the role of cultivating natural piety and providing the basis of a faith in our ability to realize our ideal values. There seems to be a deep human need to understand the nature of our world and our place in it, which has long been fulfilled by mythology and religion. This understanding is often linked with the grounding of the values of a culture. We can also point to the value of the experiences of wonder and belonging created by such an understanding, and, by extension, we can see worldview-making as an imaginative and inspirational attempt to use science to help us appreciate the wonder of the universe.

Dewey found just such a cultural role for science: “The flights of physicists and astronomers today answer to the esthetic need for satisfaction of the imagination rather than to any strict demand of unemotional evidence for rational interpretation” (Dewey Reference Dewey and Boydston1987/1934: 37). This is the aesthetic role that was previously played by supernatural “theologies and cosmogonies.” Perhaps on the somewhat darker side, in certain ways, science is also hardening into dogma in just the way that supernatural belief had done before:

From our perspective today, this may seem like an exaggeration. After all, are we not inundated, especially in the United States, by those who deny the authority of science? First, no; despite a few exceptions of politicized issues, the US public still strongly trusts scientists and scientific knowledge (National Science Board 2020). Second, Dewey, as both a preeminent philosopher of education and a pioneer in empirical education research, was long interested in science education and concerned that the instructional methods common to science teaching focused exclusively on content, taught as timeless truths, rather than on scientific methods of inquiry. We can and should question whether treating science as something that should be accepted unfailingly by a passive public is what we want, though it certainly makes the religionist analogy more persuasive.

Let me instead emphasize the positive. Imaginatively constructing a scientific worldview can serve the positive religious function for the public identified by James and Dewey. It can give us an understanding of our place in the world, the meaning of it all, a pious relation to nature, and the faith in the ideals we seek to realize. “The scientific worldview” can help not only the secular public but also the scientific community; it can act as a motivation for scientists, something grand to work toward, a flag to rally to. For Dewey, “Faith in the continued disclosing of truth through directed cooperative human endeavor is more religious in quality than is any faith in a completed revelation” (Dewey Reference Dewey and Boydston1989/1934: 18). This faith is on display in the everyday inquiry of scientists that in turn is consolidated by the worldview-builder.

Ultimately, this part of science, which is an activity genuinely connected with scientific inquiry, science education, and science communication, has more in common with religion than with experiment, inquiry, or technical application. This similarity should affect the way we approach scientific worldviews. Understanding the religious function of scientific worldviews suggests complex criteria for responsible worldview construction that pulls against many current tendencies in contemporary naturalism, humanism, and scientism. In particular, we need to think carefully about the relation between our worldviews and our values and traditional ways of life.

In his very late works, Paul Feyerabend defended a thesis that he called “Aristotle’s principle,” or sometimes the “Existential Criterion of Reality.” The reason for the former name has to do with the way that he interprets Aristotle’s response to Parmenides’ monism, the philosophical theory according to which the World is one and unchanging: “Aristotle criticized Parmenides in two ways. He tried to show the mistakes in Parmenides’ reasoning[,] and he pointed out that change, which Parmenides had called unreal, is important in human life” (Feyerabend Reference Feyerabend and Terpstra1999: 200). Feyerabend extracts a principle behind the second strategy: “real is what plays a central role in the kind of life we identify with” (1999: 201).Footnote ¹⁰ This principle is already tacitly at work in Feyerabend’s work starting in 1975, where he attempts to combat overconfidence in science. Feyerabend came to see scientific realism as supporting a kind of dogmatism about science and an undeserved special role for science in society, which hurt the freedom of people in our society to pursue their own values and traditional forms of life. This was the sort of work for which Feyerabend was labeled “the worst enemy of science” (Theocharis and Psimopoulos Reference Theocharis and Psimopoulos1987). His work in that period exemplifies the negative version of Aristotle’s principle – don’t treat something as real if it conflicts with the life you want to live, and don’t accept pictures of reality that make that life impossible or burdensome. He thus became concerned with, as the title of one provocative essay put it, “How to Defend Society from Science” (Feyerabend 1975).Footnote ¹¹

The positive version of Aristotle’s principle treats “real” as an honorific appended to those results of inquiry that we are willing to incorporate into our worldview, as a result of endorsing their role in our practices. In other words, we are willing to treat something as real insofar as it plays a role in our valued practices and forms of life: what we care about and identify with. This is a value-laden judgment. The principle does not license an “anything goes” attitude toward what we should regard as real, but rather links it to our cherished values and practices. What “plays a role” in our practices should be understood pragmatically, as what actually plays a role in practices that we value, that are successful and unproblematic. What’s more, while our decisions about what we give the honorific “real” to makes a real difference to our practices, it is a matter of the philosophical interpretation of science, not a matter of acceptance or rejection of the science itself.Footnote ¹²

Between scientific worldviews and our values (and the practices and ways of life they are connected to) is a complex, two-way street. Of course, in various ways, values inform the results of scientific inquiry. (For studies of values in science, see Part III.) What’s more, as Feyerabend argues, our values should play a role in the selective activities of worldview building. The worldviews we adopt are or imply complex evaluative standpoints, informing the ideals we pursue and the values we hold.

As I put the finishing touches on this chapter, we continue to deal with the legacy of the COVID-19 pandemic crisis, and there remains major public dissensus over the severity of the problem, the efficacy of the scientifically validated preventatives for the disease and public health policy decisions made during the crisis, and questions concerning the seriousness of “long COVID.” We face increasingly severe and irreversible fallout from the climate crisis, but our elected representatives seem focused at best on half measures, where they acknowledge its reality at all. We have faced significant challenges to the institutions and the very values of our secular, pluralistic, democratic society. If ever there were a need for a secular worldview that could inspire natural piety and a faith in our ability to realize our ideals, now is the time. It is the religious function of science to provide such a worldview. Those of us committed to science and to humanism should bring our moral imagination to bear in order to better meet that need.