The Autonomy and Integrity of Science

One indicator of the quality of scholarship is its shelf life. Findings from good science are often not immediately used or useful, but their value requires delayed gratification. My favorite example is the recommendation by the National Academy of Sciences, circa 1864, that the United States adopt the metric system: apparently we are still in the implementation phase.

In other cases, the intuitive worth of scientific knowledge can be abruptly undermined by exogenous events. Tight theories of the erosion of social capital, for example,however valid they seemed at the time, are unhinged by evidence from the COVID crisis: we may be physically distant, but we have found resourceful ways to stay socially close.

“The Changing Temptations of Science,” by Stephen P. Turner and Daryl E. Chubin (Issues, Spring 2020), a tour de force, eludes both those problems. (Disclosure: I’ve known Chubin for 35 years and have benefited mightily from his friendship, mentorship, and collegial criticism.) As it was written before the pandemic (the word doesn’t come up), one must marvel at its relevance in the light of current circumstances. And the essence of its arguments, even if not instrumentally applicable to short-term political or economic decisions, will guarantee its durability, especially if it becomes required reading, as I would heartily recommend, in courses on science policy and the policy sciences.

The timing of the article is, indeed, uncanny. As I write this, we have news of potential restructuring and increased funding of the National Science Foundation; reports of hundreds of millions of public and private dollars allocated worldwide toward finding a vaccine for COVID-19; demonstrations of ignorance, distrust, and disdain for factual evidence among key political leaders; and renewed calls for public accountability for investments in research and other public goods. Turner and Chubin’s key themes and arguments apply presciently on those issues. For example, they write that “big science … mean[s] big money, and big money mean[s] a need to justify the expenditure.” Finding a cure for COVID-19, a task likened to the Manhattan project for its magnitude and urgency, clearly will call attention to the benefits and risks of the size, source, and effects of the funding it will require.

Related is the question of incentives faced by scientists in an age of increased competition for public and private support: will pressure on scientists to please current and prospective backers fuel the “temptation to … overpromise … to sacrifice the pursuit of intellectually promising lines of work to those that can be funded, to produce work that is … scientifically trivial?” That’s a scary prospect, perhaps overstated but definitely worthy of attention especially as externalities of shoddy research on vaccines could be catastrophic.

In hopes that their article will sustain public interest, I offer three suggestions.

First, it would be good to facilitate a more comprehensive deliberation on the origins, roles, and potentially negative consequences of productivity and accountability—cherished linchpins of the nation’s economy and democracy—as determinants of the structure and governance of science.

Second, in a society that long ago rejected “philosopher kings” and, by extension, “scientist kings,” the future of the scientific enterprise should be informed by, but not left to, scientists. Designing organizational arrangements to facilitate such a process, which relies on mutual respect, contestation, and collaboration, should benefit from decades of experience with strategies for bringing evidence to bear on policy.

Finally, now may be the time to invest in a new science of science policy, one that more fully integrates disciplinary knowledge from the political, economic, and organizational sciences. Of course, where the funding for such a science will come from, what incentives it might create, and how to evaluate intended and unintended consequences, are questions for which we can be grateful to Turner and Chubin for providing an endurable framing resource.

Michael Feuer

In their very interesting article, Stephen Turner and Daryl Chubin ask whether we have reached “the inevitable conclusion to the story of science.”

Their article came before COVID-19 transformed the world. Suddenly, scientists are featured on the morning news; a best-selling bobble-head features Anthony Fauci, a leading expert on infectious diseases; and the US government has created the controversial Operation Warp Speed to coordinate and facilitate urgent development of a coronavirus vaccine. No longer is the place of scientists in question. The issues become how much money, global cooperation, and time will be required to develop a reliable vaccine, therapeutic pharmaceutical, or hopefully both.

Although a few people envisioned a global pandemic, none predicted this specific virus with its capability to upend daily lives, devastate economies, and reveal stark political, racial, and social inequities. COVID-19 also underlines the scientific paradigm shift from the twentieth century dominance of physics to the twenty-first century focus on biology. For example, Turner and Chubin emphasized the role of physics and the rise of Big Science and team work critical for the Manhattan Project, and they noted that the later controversy over solar neutrinos carried “no urgency, no political or economic stakes.” In contrast, urgency, political and economic stakes, and social impact drive biology to solve COVID-19.

Biology is not the sole discipline critical to solving the pandemic. Computer scientists to sequence genomes, epidemiologists to model the predicted spread in the population, and social scientists to outline behavioral impacts will join bioscientists on interdisciplinary teams seeking solutions. Researchers from universities, industry, and government must join resources and work in effective partnership. The sole scientist—unfunded, self-policing, and talking with only a few qualified friends—who characterized the theory of liberal science, will not suffice.

Ironically, many of the aspects that the authors feared might foretell the death of science—such as competition, large teams, industry and governmental influence, and enormity of scale—characterize what is now needed. A May 8, 2020, editorial in Science called for a “COVID-19 Defense Research Committee to be empowered to coordinate and fund solutions to the pandemic.” This would parallel the National Defense Research Committee created in 1940 to pursue war-related innovations, which resulted in a new sense of social responsibility and the role of scientists. COVID-19 has already resulted in such a renewed sense for scientists, their role, and the impact of science.

Sue Rosser

In Stephen Turner and Daryl Chubin’s article, there are two things at issue. One is curiosity-driven research versus commercially oriented research. The other is unconstrained research versus  mission-oriented research. The lines are easily blurred, of course, but it is easy to see the difference.

Consider approaches to health—say, the problem of high blood pressure. Commercially oriented research looks for drug solutions, since that would lead to intellectual property rights and potentially big profits. Unconstrained research would also investigate the effects of diet and exercise, even though these would result in no royalties. Both approaches are mission-oriented in that they have the specific aim of tackling high blood pressure. Unconstrained research might start out investigating blood pressure but get sidetracked by uncovering interesting connections among, say, blood pressure, exercise, and cognition. Research might continue to the point where any interest in blood pressure drops completely out of the picture.

Turner and Chubin are worried about both these issues, though they do not put it in quite the same terms as I do. How can we overcome these problems? That is, how can we loosen the grip of commercially oriented research in favor of more that is curiosity driven, and how can we loosen the grip of mission-oriented research in favor of more that is unconstrained?

My preferred approach is to preach the doctrine of science for its own sake in tandem with science for the betterment of humanity. However, except for a few special cases this hasn’t had much success. Cosmology is a such a case. Most people are fascinated and happy to fund expensive telescopes while not expecting anything useful to come of it. But most research is of interest only to specialists. We need a Plan B.

Here is my tentative (and, I admit, not well thought out) alternative for funding research. Many corporations adopt so-called 20% time programs. Google, for instance, assigns projects to its employees but allows them to use 20% of their time (in effect, one day a week) to work on something of their own choosing. Here is where curiosity-driven, unconstrained research could take place. It seems to have paid off handsomely for Google.

This is a modest proposal—after all, it is only 20%—but at least it addresses the very legitimate concerns that Turner and Chubin raise. There is no mandate for a complete overhaul of the funding system, so in the interim this might be the best we can hope for. And who knows? Maybe it will creep up to 30%, then 40%, and so on.

James Robert Brown

Science is under strain. The prevalence of fraud, retractions, scandals, and scrambling for grant money suggests an institution in crisis. Stephen Turner and Daryl Chubin diagnose this crisis as the result of changing culture and incentives in science, from an emphasis on discovery to productivity, from autonomy to accountability, from fame to funding. Although they avoid calling for a nostalgic return to the post-World War II culture of science, they clearly think more has been lost than gained in this shift. But is the picture of the past too rosy? And can we chart a more promising future with a clearer picture of the past?

The science of the mid-twentieth century in the United States was well-funded, as public largesse poured into science for the first time. It gave scientists unprecedented autonomy to pursue what they found interesting, and allowed them to train students at unprecedented rates. The ranks of scientists swelled, and funding competition inevitably increased. But science during that period was also deeply sexist and excluded researchers who were not white or male, contrary to ideals offered in the sociologist Robert K. Merton’s universalism.

Science of that time also had its missteps (e.g., eugenics and racist theories of IQ) and unreliable results (e.g., DES as a miscarriage drug). It is not clear that the promises made by Vannevar Bush’s Science, the Endless Frontier came to fruition, at least by the mechanisms Bush presented. As many observers have argued, the great breakthroughs of the post-war era were not made by the free pursuit of science by free intellects, but rather by cross-institutional collaborations that kept their eyes on grounded goals, and held each other accountable to those goals. The success of such science was shaped very much by accountability to those collaborators outside science.

The problem with today’s science is less about a loss of autonomy and more about to whom scientists are expected to be accountable. Producing knowledge that pleases the funders is not the kind of accountability we want in science, as bias is the obvious result. We should think carefully about what we want scientists to be accountable for, and to whom, and shape the norms of science accordingly.

For example, as Merton noted, scientists should have essential accountabilities to each other, as manifested through a healthy culture of criticism and response to criticism. Yet when scientists are accountable only to each other, they can get lost in rabbit holes that do not effectively answer questions (either for the public or for science), and just perpetually call for more research. Additional accountabilities should be properly external-facing, but tied to the public good rather than private interests. And it needs to be more than window dressing.

We should also keep in mind that responsibility and accountability are not the same. Responsible science demands more, and different, things than accountable science. Crucial are responsibilities to the broader society, to honor key moral constraints (e.g., in human and animal subject research) and to aid society in clarifying and pursuing its goals. The ideal of the autonomous scientists pursuing science for its own sake never did serve society or science very well. We need to forge a better ideal.

Heather Douglas

Stephen Turner and Daryl Chubin describe an evolution of science between circa 1930 and the present as starting out being led by lone geniuses who functioned autonomously and engaged in self-policing to now being beholden to productivity, impact, and funding agencies. They claim that the pace of discovery has slowed under these new constraints. They emphasize a few major discoveries from the former halcyon days, but with a few exceptions ignore the many breakthroughs in the past several decades. The authors scoff at the value of advances in mathematical modeling, which in the COVID-19 era is proving itself to be lifesaving.

Although the authors’ argument is elegantly written and thought provoking, their romantic view of the past and cynical view of the present derives from selectively emphasizing elements of both. They ignore how the previous norms of science excluded women, males of color, and individuals with disabilities, and even used scientific arguments to justify this exclusion. They describe how the current system has led to the “surrender of individual autonomy” by scientists whose “special status as experts is compromised.” Absent from this description is how the previous era of science enabled both scientists and society to ignore the contributions of female scientists, such as Rosalind Franklin in the discovery of DNA (which the authors highlight as an example of the benefits of the past approach to science), and scientists of color, such as Percy Lavon Julian, who synthesized physostigmine, used to treat glaucoma, among other conditions.

Although the authors praise previous generations of scientists for being able to police themselves with their own social norms, they ignore the egregious deficiencies in this practice. Among notable examples of past failures in self-monitoring, in the name of scientific discovery scientists from Harvard University and the Massachusetts Institute of Technology fed radioactive iron in oatmeal to developmentally disabled children to study the absorption of calcium; and in a study that was deemed scientifically valuable black men with syphilis in Tuskegee, Alabama, were left untreated to study the natural course of the disease.

The authors state that “progress on fundamental issues has stalled,” but who defines fundamental? One positive outcome of the current approach to science is its invitation to include a broader array of scientific disciplines than in the past. This has opened the door for scientific scrutiny of issues fundamental to creativity and innovation, such as research on study design, data analysis, peer review, team dynamics, and translation of discovery into practice.

I applaud the authors for writing such an enjoyable and provocative paper. They remind us that science exists in a larger society, and that there are benefits and harms to past, present, and most assuredly future approaches to conducting, evaluating, funding, and valuing science and scientists.

Molly Carnes

In their deep reservations about the “impact agenda” in contemporary science, Stephen Turner and Daryl Chubin reveal a disconcerting feature about the ideal of scientific autonomy that they defend. As their own repeated references to Vannevar Bush and James Bryant Conant show, this ideal has been historically promoted by the academic establishment, typically in aid of elite science.

Yet it is far from clear that academia has been the home of scientific autonomy in its broadest sense; namely, the free selection of the ends and means of research. Instead, what has been upheld has been the much more narrow and self-serving idea of academically oriented scientists being autonomous from any constraints imposed by the rest of society. This effectively grants the academic establishment exclusive rights to impose all the constraints it wants on those who would claim to do “science.” Thomas Kuhn’s totalitarian sense of paradigm, inspired by Conant, concedes this point by making it a condition of being considered a scientist that one self-presents as having undergone the appropriation forms of acculturation (aka indoctrination) into science; hence the significance attached to peer review.

Put another way: the defense of scientific autonomy has been really about the entitlement of a group of self-certifying academics to monopoly ownership over science, understood as the most highly valued form of knowledge in society. The terms on which the US National Science Foundation was established, masterminded by Bush and Conant, was a victory for this vision. Thus, Conant’s promotion of “no science before its time” should be understood as a form of soft power, whereby science domesticates the passions of a populace increasingly impressed by its achievements.

From this perspective, the “finalizers” of science discussed by Turner and Chubin are best seen as constructive critics. They were observing that scientists left to their own devices were likely to squander their entitlement by becoming increasingly self-involved in problems that removed them from larger societal concerns, even though they had already secured a body of usable knowledge. Today’s impact agenda in science policy is the heir to such constructive criticism.

However, the conditions under which the NSF was established—the success of the Manhattan Project—by no means guaranteed this academic power grab. Notwithstanding the elite scientists involved, the creation of the first atomic bomb was a task set by the government in the context of national security. Within those constraints, the scientists were given free rein to solve the problem, and their solution was validated outside academic peer review. In short, it was closer to the broad sense of scientific autonomy: no one really knew what an atomic bomb was until one had been successfully detonated.

In effect, the Manhattan Project scientists had more discretion over the product they delivered than academic scientists ever do. Indeed, the Defense Advanced Research Projects Agency would have been a more appropriate long-term US response to the Manhattan Project than the NSF to promote the potential power and benefit of scientific autonomy for the larger society. But creation of that agency required nearly another decade—and the launch of Sputnik.

Steve Fuller

Stephen Turner and Daryl Chubin question whether our values have changed along with the movement from individual to groups and with the changing mandates and mechanisms used to support discovery. The authors highlight examples and patterns that point to a shift in the ethical underpinning of the academic enterprise as a whole.

I appreciate their reminders of scholarly philosophical definitions of science and of essays from leadership that influenced the change in how science is supported. Although I am not a scholar in either philosophy or the history of science, I do dwell on causality. In that role, I’m not ready to accept many of the causal connections that they propose.

Turner and Chubin ask whether “the disappearance of exploratory science (is) a result of science’s natural growth and evolution … or the result of the structures of institutional science themselves.” I favor the former explanation. A natural progression of our ability to record observations and ideas is that our collective understanding of the world, and even the universe, has grown with time. New methods and new strategies replace the old. Deeper problems require more than single disciplines; they require collaborations. In some cases, huge collaborations are needed to advance the cutting edge.

Individuals and institutions adapted to take advantage of opportunities when presented. Applied science, for-profit science, is now as respectable as basic science, and in some quarters is seen as even more desirable. As the knowledge became more complex, the ability for nonexperts to judge the quality or contribution became increasingly difficult, thus making hiring, promotion, and funding decisions more challenging.

An unfortunate workaround has been the use of surrogates, such as numbers of papers or grants, or numbers of citations. Rather than examine the original work in depth, one might assume the respectability of the work merely from whether it was cited by others. The price we pay for not making the effort to dig deeper ourselves is that the use of surrogates leaves room for manipulation.

For most of us who joined the ranks of science professionals in the latter third of the twentieth century, the exciting areas to explore had been opened by the new tools developed by our predecessors. We led or followed the crowds, not because that was where we would find funding (as suggested by the authors), but because that was where the new methods and the revised cutting edge would be.

Turner and Chubin note that along with the emergence of large projects conducted by teams, the pattern of funding science changed, from people-oriented to project- and impact-oriented. The authors leave to our imagination the challenges and successes of the agency staff who were successful in their arguments to Congress for public funding of broad swaths of scientific inquiry. We imagine that their arguments were pragmatic and focused on projects and outcomes.

Indeed, the change in how projects are judged evolved from how the funding was justified. The agency staff was also concerned with bias in the review process, the “old boy network.” A resulting approach was to focus on the feasibility and potential value of the project rather than on the esteem of the person. In my opinion, serving the public good and reporting to Congress shaped the review process more than any shift in values of science as a method.

I believe that science, a broadly accepted method for understanding the what and the how, operates under different values as we become more enlightened. We can no longer rely on common sense and individual approaches to carry out our shared values of how research is conducted. As tools and methods of investigation evolved, we recognized other issues that went beyond the domain of individuals. Unfortunately, the bad behavior of individuals can influence the public impression of the methods and ethical values of scientists as a whole. We needed to assure the welfare of research subjects, whether animal or human. We needed to prevent exposure of lab workers or the public to hazardous materials. We needed to prevent or detect fraud.

In their thoughtful essay the authors have raised critical questions for those responsible for developing the next generation of scientists. I’d be delighted if this is chapter 1, recalling the good old days when individuals had more autonomy. I await a chapter 2 that might take up the challenge of telling us how to retain the noble nature of discovery and adapt it to the needs of our changing society.

Clifton A. Poodry