Oxford University Press, open access
Around the world many people—parents and children, young and old, healthy and immunocompromised—remain frustrated by the refusal of some of their fellow citizens to get vaccinated or wear masks. The charge is not only ethical: that the naysayers are selfishly endangering the well-being of others, especially the most vulnerable people whose lives are still threatened by everyday activities. It is also epistemic: that the refuseniks are ignorant or irrational, and in any case failing to “follow the science.”
What do the refuseniks say in their defense? A very small number may be excused on the grounds that vaccination or mask-wearing is unsafe for them. But most likely, they will simply protest the accusation: as they see it, they are neither selfish nor anti-science. “To be sure,” their retort goes, “we are not following what you call ‘science.’ But that’s because the people you cite as authoritative are biased. Advice like theirs isn’t worth heeding.”
Also spricht Antivax. His riposte should not be dismissed out of hand. Should the fact that certain individuals have been anointed by government officials—perhaps belonging to an administration Antivax opposes—settle the question? Is it really unreasonable to wonder how research may be influenced, subtly or not so subtly, by the hand that feeds it? Perhaps the most compelling version of Antivax will not elevate some rival group of experts—Ivermectin hawkers, say—but simply contend that the situation is deeply uncertain, that nobody has a good strategy for coping with COVID-19, and that the scolds therefore have no basis for confidently proclaiming that theirs is the most reliable course of action.
Antivax raises deep questions about what science is and does—concerns that have long been debated by scientists and philosophers. Michela Massimi is among them. A philosopher at the University of Edinburgh, she has pioneered a distinctive form of “perspectivism” in the philosophy of science. Her magisterial new book, Perspectival Realism, is the culmination of two decades of work on this score. It stems, she tells us, from “worries of a concerned citizen in a society where trust in science was being eroded under the pressure of powerful lobbies,” anxieties reinforced as she worked on the final draft during the pandemic. Her aim is not to address Antivax directly, but to answer a more fundamental need: developing an accurate picture of scientific practice, in order to enable citizens and scholars alike to identify the sources of its triumphs and its limitations.
Painting such a picture is notoriously difficult—much more difficult than we tend to acknowledge. Contrary to popular depictions, science is an immensely complex institution, more like a Rube Goldberg machine than a smooth-running assembly line. Yet the dream of frictionless knowledge production persists in textbook caricatures, media misrepresentations, and even the reports of some scientists themselves. On this view, science proceeds by fanfare: just put some scientists to work, let them employ The Scientific Method, and out comes Truth with a capital t.
So simplistic and confident an attitude collapses under any serious study of the nature of scientific research, past and present. Seventeenth-century savants may have dreamt of a mechanical method for answering questions about nature, but successive centuries have driven home the point that no such method is to be had. The road to so many of the greatest achievements in the history of investigations of the natural world is full of twists and turns, false starts and dead ends. Even at their most ingenious and most industrious, highly talented scientists fail and fail again. If some are fortunate enough to crown their failures with some success, others, equally brilliant, only heed Samuel Beckett’s injunction: they fail better.
Two mid-twentieth-century thinkers found vivid ways of evoking the fallibility of science. Karl Popper characterized science as a bundle of conjectures, constantly in danger of being undermined by empirical evidence—that is, of being “falsified.” Perhaps perversely, the bolder the conjecture, the better, in Popper’s scheme; a spectacular fall is the glory of the Popperian game. Nevertheless, by a form of inference Popper never adequately explained, the conjectures that escape falsification are entitled to guide practical projects. Thomas Kuhn, trained in physics, understood the limitations of this conception—how many scientists really want to falsify their pet hypotheses?—and proposed a different structure for scientific practice. Mature sciences, he suggested, are pursued within communities, each dominated by some constraining structure—a “paradigm,” he called it (a term he came to regret). Individual scientists work on puzzles fixed by this structure; when they fail (as, like high school physics students, they often do), they are told to try again. Only when the puzzle defeats the ingenuity of numerous contestants does it acquire a new status, becoming an anomaly. At such moments the course of “normal science” breaks down and a sense of crisis arises, which may prompt the transition to a new structure for community-wide research: a scientific revolution.
The ideas of Popper and Kuhn proved enormously influential. In fact, despite the flourishing of sophisticated studies of many aspects of scientific practice over the last few decades, their ideas remain more or less hegemonic among those who have advanced beyond the simple vision of Science as Guarantor of Truth. (Often, it must be admitted, a dash of Popper is now added to a soupçon of Kuhn—a blend that would aggravate the palates of both men.) In its range and depth, Perspectival Realism should begin to loosen their hold, offering an essential new touchstone in the analysis of science. Though much of the book assumes a fair degree of familiarity with the history and philosophy of science, its overarching argument holds important lessons for general debates about the nature of scientific knowledge.
Massimi is thoroughly aware of the wealth of insights amassed by half a century’s dedicated study of the history, philosophy, and sociology of science. If she has seen further, or from more angles, it is because she has stood on the shoulders of these predecessors, from Helen Longino’s study of the social norms of scientific communities to Bas van Fraassen’s analysis of scientific representation. To understand her position and to measure its significance, it is necessary to recognize the components delivered by the tradition.
During the 1970s and ’80s, a group led by Stanford polymath Patrick Suppes undermined the popular idea of science as a unified whole. In the wake of this work, talk of science in the singular gave way to appreciation of the diversity of the sciences. (By the 1990s the insight was encapsulated in influential books: John Dupré’s The Disorder of Things, Nancy Cartwright’s The Dappled World, and Peter Galison’s and David J. Stump’s edited collection The Disunity of Science: Boundaries, Contexts, and Power.) Contrary to eminent physicists, the dream of a final theory was exposed as an unattainable illusion. Studies of the special sciences, from most of chemistry through biology to studies of human beings and of their societies, revealed that, even in principle, not everything could be reduced to physics.
The thought that all science—or at least, real science—could be cashed out in strict laws and axiomatized theories was unmasked as an absurd over-generalization. Scholars came to understand the importance of models and idealizations in scientific practice, to appreciate the intricacies of experiments, to recognize the roles of instruments and of their development, and to view the science of any age as embedded in—and limited by—its ambient material culture. In light of all these developments, Popperian falsification began to seem at best a distant approximation to the actual dynamics of complex and diverse scientific practices.
Kuhnian ideas have fared better, but their original articulation would require considerable elaboration and refinement to cope with what have become commonplaces in interdisciplinary reflections on the sciences. During the 1960s and 1970s, Kuhn’s influence on sociologists and socially minded historians, initially in Edinburgh and later far more broadly, prompted major modifications of the image of scientists at work. Historians such as J. E. McGuire, P. M. Rattansi, and Betty Jo Teeter Dobbs demonstrated how the revolutionary thoughts and primary achievements of the most esteemed figures were entangled with ideas we find far more suspect, conceptions current in the societies around them. Meanwhile, sociologists such as Barry Barnes and Harry Collins exposed the decisions made in the course of experimental inquiry, arguing that they could not be understood in the terms favored by philosophers—especially not by those who took such decisions to instantiate formal, quasi-logical, rules.
These studies had a large impact. Emphasizing the role of people beneath the lab coats—beings with feelings as well as inferential skills, with ambitions and political affiliations as well as dispassionate love of truth, living in particular societies with customs and conventions and not in some protected sterile laboratory—scholars from both disciplines offered a picture of science as situated, inevitably influenced by the values of a time and place, perhaps even not so dissimilar from the “local knowledges” preserved and passed on by many social groups who have been left out of the triumphant narrative of science as the greatest achievement of humankind.
In short, Max Weber’s famous ideal of science as value-free began to waver; the idea of scientific rationality started to look suspect, and the prospects for any large-scale account of what science is and why it is so special came to look dim. Objectivity seemed lost. Small wonder, then, that these days Popper and Kuhn are short of emulators. Developing a new general picture of science, responsive to so many disparate themes and considerations, has come to seem not just unfashionable but impossible. The issue is a timely one—historian Naomi Oreskes’s recent book Why Trust Science? (2019) explored similar themes, for example—but it remains philosophically daunting.
On the face of it, Massimi’s title may seem oxymoronic. “Realism” connotes objectivity; “perspectival” suggests a definite point of view. But there is no contradiction, Massimi is out to show; she is engaged in a project of reconciliation. Appreciating the cogency of what past decades have taught, she affirms that all knowledge is situated; scientists cannot find any Archimedean point or achieve the “view from nowhere.” Their research is undertaken from perspectives. But she sees the plurality of perspectives as crucial to the production of reliable knowledge. Her task is to understand how this works, how some kinds of cognitive achievement—knowledge, understanding, an ability to steer intervention—are possible across perspectives.
Early in her discussion, Massimi distinguishes two notions of perspective. One is thoroughly familiar, defined by the standpoint of the viewer: what we see from a particular vantage point. The second inverts the direction, borrowing the idea of a “vanishing point” from the theory of perspective in pictorial art. For Massimi, this point where parallel lines converge at the horizon also individuates a perspective—even if there is no viewer.
Massimi illustrates these ideas by referring to two famous paintings from the history of group portraiture. In both Diego Velázquez’s Las Meninas and Jan van Eyck’s Arnolfini Wedding, a mirror is depicted near the vanishing point of the pictorial perspective. In the former painting, we have a direct view of the artist at his easel, painting some court scene, but, from the image reflected in the mirror, we can discover what he is painting (the Habsburg royal couple). The Arnolfini portrait is interesting in a subtler way, for the space the mirror reveals includes a window, opening on a wider space, in which objects whose presence we would never have suspected are to be found.
Think, then, of the vanishing point fixing a perspective as something that opens up new possibilities for those who look at it, enabling them to explore more broadly than they would otherwise have been able to do. A suggestive thought if you regard science as a process of fallible exploration, inevitably perspectival, capable of issuing in cognitive changes, at least some of which enable people to intervene more reliably in nature to achieve their goals. So far, however, the analogy is incomplete; it is unclear how appeals to vanishing points help restore objectivity. Where is the counterpart of the helpful mirror behind the newlywed Arnolfinis? How can scientific activity be compared to what artists do in painting their pictures? Massimi’s strategy is to identify the constructive or creative aspects of scientific practice, before trying to explain how those practices are checked—objectively checked—by something qualified to count as part of reality.
Scientists are human beings, animals with particular kinds of sensory and cognitive capacities. Once, they were young. During the course of their early development, as children, they acquired the ability to organize their experience, to distinguish physical objects, sort them into varieties, recognize events and processes in which they are involved, understand some of the causal connections among those events and processes—in short, they came to live in an organized world. The organization they could acquire was not only restricted by their biological characteristics. It also reflected the ways prior generations found it useful to solve their problems and to pursue their ends. All of us borrow in this way from the social world, the structured world of shared experience. In a sense, that world is a collective human creation. And the institution we label “science” has played a prominent part in that creative work, which we pass on to later generations.
Yet our powers of construction are far from unlimited. Our cognitive creations don’t always fit our experiences. Scientific communities are drawn to areas where such a gap emerges, sometimes in service of widely shared goals (say, to treat disease), other times simply because a particular group finds the challenge intriguing. Investigators build models to resolve the mismatch. Out of their efforts, the framework is reconstructed, not necessarily wholesale (as Kuhn supposed) but frequently by means of small adjustments. Much of the time their efforts are frustrated; models remain inadequate. Along the way, though, investigators come to better understanding of the places where something beyond their control pushes back. As they do so, they come to isolate stable elements in experience, pressures to which their reconstructive endeavors must conform. They have discovered phenomena.
In short, model-building is constrained. But in becoming aware of phenomena, and attempting to provide integrated models of them, researchers working from different perspectives begin to construct a space akin to van Eyck’s. Combined scientific endeavors—initially parallel, but ultimately converging—delineate a structure for the world. Through chains of inferences, leading from data to characterizations of phenomena, a new organization of (part of) the world emerges. From what is initially a scattered collection of data—the tabulations of measurements and instrument readings taken on particular occasions or the qualitative observations investigators make—something stable is forged, a conclusion about what must or what may occur under specified circumstances. As in the Arnolfini portrait, this structure opens surprising new lines of sight onto the world, allowing us to see what we otherwise could not.
Another title for this book might thus have been Objectivity Regained. Massimi starts with the first lesson of recent history, philosophy, and sociology of science: scientific practice is shot through with contingency, the organization people have devised at a time and place. It is thoroughly and multiply perspectival. Yet far from being lost, objectivity emerges from the intersecting lines of coordinated model-building, as investigators learn to bow to the constraints of reality—not, as in Kuhn’s account, because a single hegemonic framework unearths puzzles that continue to prove recalcitrant, but through the interaction of different perspectives, each with their own partial successes and difficulties. Moreover, perspectival convergence does not, for Massimi, imply convergence to some forever-stable truth about the World Out There. As she puts it, “The vanishing point(s) towards which the lines of perspective converge in a drawing should not be understood as a proxy for a metaphysical reality to which we all converge at the end of inquiry. After all, they are only vanishing points.” No matter these achievements, the practice of science remains always and everywhere “ongoing and open-ended.”
The development of the perspectival analogy, with emphasis on the vanishing point, is the core of Massimi’s long book. She elaborates on the highly simplified sketch I have offered here in two main ways.
First, she demonstrates, with great care and thoroughness, how her ideas apply in a wide range of scientific instances. In their number, diversity, and depth of treatment, these discussions are truly remarkable. Besides her three most extensive examples—studies of the career of the shell model of the atomic nucleus, of model pluralism in climate science, and of attempts to understand the causes of children’s difficulties in learning to read—she reviews a very long list of cases with impressive attention to the scientific and historical details. Among them are J. J. Thomson’s research on the electron, the discovery of blueprinting, the construction of the fountains of the Villa d’Este, James Clerk Maxwell’s model of the ether, uses of computer science in organic chemistry, the discovery of the Higgs boson, difficulties with Antoine Lavoisier’s notion of caloric fluid, the classification of Alpine flowers, attempts to synthesize eight-base (hachimoji) DNA, investigations of carcinogenesis, the development of the magnetic compass, the history of Planck’s constant, and disputes about the principles of biological taxonomy. The range of examples is important: the fact that Massimi’s general theory of science can handle so many special cases testifies to its robustness.
To illustrate, consider her discussion of the way Planck’s constant, h, became recognized as a fundamental physical constant. By the late nineteenth century, very different perspectives on electrical phenomena had arisen. One tradition stemmed from Maxwell and Michael Faraday, viewing electromagnetism in terms of wavelike “fields.” Another tradition—most evident in work on electrochemistry by Hermann von Helmholtz—seemed to imply the existence of discrete “corpuscles” of electric charge. (Thomson’s work on cathode rays emerged from the field-theoretic tradition, Massimi notes, but it flirted with the corpuscular view: indeed “corpuscle” was his term for what we now call the electron.) Though these rival perspectives grew out of very different experimental contexts and research communities, by the 1880s they appeared to converge in pointing toward the existence of a minimal unit of electric charge. Efforts to measure this minimum yielded a range of empirical values, however. Max Planck brought order to this situation by showing how his work on blackbody radiation—which we now recognize as a crucial step to quantum mechanics—could be used to predict the minimum charge, facilitating connections across the rival perspectives and establishing a set of fundamental physical constants. The interplay of very different perspectives thus opened a significant new “window on reality” (an idea Massimi credits to art historian Erwin Panofsky)—giving us a new physical constant in the process.
Second, Massimi engages at length with a vast range of philosophical scholarship, in an effort to show how her preferred language and her central theses are defensible. These sections may not prove as successful as those in which she deploys scientific examples—not because of any evident flaw in her discussions, but because of the propensity of many contemporary Anglophone philosophers to find grounds for carping. It would be sad to see so bold and visionary a book drown in a sea of quibbles. These discussions are also likely to frustrate general readers or scholars from other disciplines who would benefit from her innovative proposals. Readers whose analytic consciences have not yet reached the height of professional philosophical fastidiousness could simply skip them.
Through it all, the plurality of perspectives is crucial to Massimi’s story. Her closing chapter fulfills this commitment in a novel way by attending to perspectives typically excluded in philosophical accounts of the growth of science. The glass needed for the kinds of experiments Thomson undertook, she notes, was only available because of the labor of Scottish peasants, collecting kelp on remote islands. Yucatán beekeepers have enlarged our understanding of melliferous plants, she shows, and medical science has been enriched by Malagasy knowledge of the rosy periwinkle. Massimi eloquently protests the way marginalized groups are excluded from the consumption and production of knowledge, their perspectives ignored to the detriment of scientific progress. She sees a general phenomenon, “trademarking,” that restricts some items of “scientific knowledge to the exclusive socioeconomic benefit of one epistemic community at the expense of others.” The result is an instance of what she calls, following philosopher Miranda Fricker, “epistemic injustice.” Perspectival Realism concludes with a well-defended expansion of the scope of scientific perspectives to embrace a vision that cuts across geography, ethnicity, and class.
At this moment of intensifying distrust of science, Massimi’s focus on the exploratory character of scientific practice is very welcome. She is out to challenge prevalent oversimplifications. Scientists are not alchemists: they possess no magical method for turning data into truth. Phenomena must be carefully, painstakingly extracted from data, often by diverse methods applied to different domains. Models must be constructed to integrate whatever phenomena are retrieved. In her telling, science-in-process resembles a puzzle in which would-be solvers must first hunt out the pieces in different places and fit them together without any great confidence that their haul is complete—in fact, often with good grounds for belief that crucial parts are missing.
The scientific work undertaken during the COVID-19 pandemic is a case in point. Since the novel coronavirus was first detected in late 2019, epidemiologists, immunologists, and molecular biologists have analyzed a flood of information from many parts of the world. They have identified a significant store of phenomena, issuing claims about which people are likely to be affected to what degree by this or that variant, about the molecular structure of different coronavirus mutants, about the mechanisms through which it attacks the human body, and about how it can (and cannot) be transferred from organism to organism. On this basis, they have devised models to predict the spread of disease, in its recurrent waves of invasion, as well as proposing defenses—vaccines and modifications of human behavior—designed to limit the surges. All this is well worth having. Yet, as is abundantly clear, it is not a panacea.
Perspectival Realism reminds us of the work required to obtain even this much. The molecular understandings draw on concepts and techniques invented and refined over considerably longer than half a century in the long tradition of molecular biology. The contemporary world is fortunate also to inherit from earlier generations significant knowledge of the immune system and of epidemiological modeling. Yet, as Massimi’s work makes clear, new circumstances and new information often require creative extensions of earlier perspectives. Devising an mRNA vaccine to combat a new virus is far from a trivial exercise, and it is both remarkable and unprecedented that success came so quickly. Combining a sprawling dataset, reporting on people of different ages, living in different environments, engaging in widely varying patterns of social interaction with one another—all to yield the parameter values for an epidemiological model—inevitably involves weighing the relative importance of different factors. Judgments on many matters can only be tentative, based on similarities with cases that have previously been studied, and liable to be inundated by the next tsunami of data.
Does Perspectival Realism provide the antidote to Antivax? Like science itself, it is no magic bullet, but it certainly provides immediate relief for the simplest arguments—those that see scientific success as an all-or-nothing affair. Science isn’t infallible, Massimi shows, but it isn’t incorrigible, either. The best it can offer are hard-won, if incomplete and tentative, successes. Science grows through the interaction of diverse perspectives. What we do with the knowledge it is able to deliver is up to us.