Science on the Verge
Science on the Verge by Andrea Saltelli exposes the cracks spreading through contemporary science and the mechanisms that fuel them, anchoring its account in ethical, methodological, and political dimensions. This text scrutinizes the forces shaping scientific output, the values underlying quantification, and the volatile relationship between evidence and policy. In tracing the dynamics from laboratory to legislature, the book constructs a map of crisis, reform, and possible futures for science at a crossroads.
The Roots of the Scientific Crisis
Scientific research experiences explosive growth in both volume and specialization, with new data and publications appearing at a rate unmatched in previous generations. This acceleration generates a paradox. As the number of discoveries multiplies, the reliability of claims erodes. Rigorous attempts to replicate high-profile studies often fail, suggesting that headline-grabbing results may dissolve under closer scrutiny. Scholars like John P. A. Ioannidis and C. Glenn Begley document these trends, revealing systemic flaws rooted in incentive structures and methodological shortcuts. Publication pressure distorts priorities, driving scientists to maximize output and secure funding rather than validate results. With incentives aligned to favor quantity, scientific integrity can falter.
Metrics, Incentives, and the Manufacturing of Knowledge
Journal Impact Factors and citation counts shape reputations, promotions, and funding decisions, but these metrics offer only a proxy for scientific value. When institutions judge researchers on the basis of such metrics, the process encourages strategies that game the system—salami slicing, citation cartels, and rushed publications. A declaration in San Francisco, supported by thousands of scientists and hundreds of organizations, calls for the rejection of journal-based metrics as surrogates for quality, urging a return to direct assessment of research merit. Despite the recognition of the problem, entrenched incentives persist. As the race for visibility intensifies, researchers increasingly lean on statistical tools that mask rather than reveal uncertainty.
Statistical Practices and the Mirage of Certainty
Statistical tools once intended to clarify, now often confound. Widespread misuse of p-values, rampant p-hacking, and the production of implausibly precise results undermine confidence in published science. These statistical manipulations yield apparent significance without substantive discovery. Fields that rely heavily on mathematical and statistical modeling—psychology, medicine, and environmental science—become especially vulnerable to such distortions. The American Statistical Association issues public warnings, and a wave of dissenting commentaries highlights the stakes: unreliable science erodes trust and weakens the foundation of policy decisions built upon quantitative claims.
Commodification and the Erosion of Quality Control
Science transforms into a commodity as universities, research centers, and private institutions compete for prestige and funding. This transformation undermines the shared standards and community-based quality control that once underpinned scientific practice. Jerome R. Ravetz and Jean-François Lyotard describe a process by which science, stripped of communal values and subject to market logic, risks losing its integrity. The pursuit of grant money and publication leads to shortcuts, as researchers deprioritize thoroughness and critical review. Derek J. de Solla Price foresaw the possibility of science reaching a saturation point, where exponential growth could generate senility—a weakening of purpose and process as the field outgrows its capacity for self-correction.
Ethics, Authority, and the Role of Science in Policy
Science’s dual legitimacy emerges from its role as a source of facts for policy and as a self-governing pursuit of truth. Trust in science persists among the public, often because technical complexities remain opaque. However, when policy relies on scientific advice to adjudicate disputes, this trust faces unprecedented challenges. Bureaucrats and officials seek the authority of numbers to legitimize decisions, especially when their own mandates lack democratic sanction. Quantification offers the appearance of impartiality and fairness, but the process depends on judgment and trust. Without these, objectivity devolves into mere mechanics, leaving systems open to manipulation and loss of credibility.
Uncertainty, Quantification, and the Search for Robust Methods
Scientists and policymakers require tools to navigate uncertainty. The NUSAP system introduces a method for managing and communicating the uncertainty of quantitative statements, categorizing information by numeral, unit, spread, assessment, and pedigree. This multidimensional approach aims to clarify where numbers originate and how much weight they deserve. Sensitivity auditing, practiced within the European Commission, interrogates the inferential chains underlying policy-relevant models. These innovations respond to the call for responsible quantification—a recognition that numbers, far from being self-explanatory, demand context, explanation, and critical examination.
The Epistemology of Science for Policy
The demarcation model—separating facts from values, scientists from policymakers—shapes the interface between research and governance. This model prescribes a boundary that, in theory, shields science from political interference and prevents scientific information from serving private agendas. Anne Glover and others stress the need to safeguard the evidence-gathering process from the pressures of the political imperative. However, complex, contested policy issues resist such neat boundaries. The model’s prescriptions can fray, as competing interests and uncertainties expose the limitations of rigid separation.
Post-Normal Science and Extended Participation
A new epistemological model emerges: post-normal science. This framework, developed by Silvio Funtowicz and Jerome Ravetz, asserts that when facts are uncertain, values are in dispute, stakes are high, and decisions are urgent, traditional scientific methods alone cannot deliver clear guidance. Post-normal science proposes extended participation, drawing in multiple disciplines and engaging diverse communities, including stakeholders and citizens. This approach recognizes that knowledge relevant to policy does not flow solely from scientific authority, but arises through negotiation, deliberation, and the synthesis of perspectives. Facts become “extended facts,” woven from both scientific data and experiential knowledge.
Historical Roots and Ideals of Control
The dream of science as a predictive, controlling force traces to the works of René Descartes, Francis Bacon, and Nicolas de Condorcet. Bacon’s vision of harnessing the wonders of nature for human use, Descartes’s rational mastery, and Condorcet’s mathematical vision for society shaped generations of scientists and policymakers. These ideals train scientists to see themselves as both explorers and adjudicators of nature’s mysteries, responsible for delivering knowledge that prolongs life, cures disease, and expands human capability. The contemporary crisis signals a need to revisit these philosophical foundations, questioning which aspirations serve the public good and which risk distorting the mission of science.
Narratives of Innovation and Institutional Tensions
Innovation follows distinct trajectories as science navigates its relationship with society. Alice Benessia and Silvio Funtowicz describe modes of demarcation: separation, hybridization, and substitution. These modes generate contradictions, as institutions oscillate between maintaining independence, seeking integration, or surrendering expertise to external actors. The European Commission’s Joint Research Centre serves as a case study in the complexity of working at the science-policy interface. Boundary institutions face challenges in upholding rigor, transparency, and responsiveness while serving as translators and arbiters of expertise.
Quantitative Storytelling and the Power of Numbers
Numbers possess compelling appeal for officials tasked with making consequential decisions. Quantification promises fairness, clarity, and the ability to avoid accusations of arbitrariness. Theodore M. Porter, in Trust in Numbers, illuminates how scientific objectivity fulfills a moral demand for impartiality, making decisions appear less personal and more authoritative. However, the power of numbers depends on the trustworthiness of the processes that generate them. If those processes falter, quantification can become a tool for obfuscation rather than enlightenment.
Financial Implications and Societal Stakes
Science’s quality crisis incurs tangible costs. The Lancet estimates that in 2010 alone, $200 billion in the United States supported research that failed to produce reproducible or reliable results. This waste signals inefficiency and squanders opportunities to advance knowledge or improve lives. As the stakes rise in fields like climate science, medicine, and environmental management, the pressure to reform intensifies. Decisions made on the basis of weak or manipulated science can misguide public policy, divert resources, and erode public trust.
Reforming Science: Toward New Standards and Practices
International conferences, declarations, and checklists articulate a pathway toward reform. Recommendations include fostering large-scale collaborative research, cultivating a culture of replication, mandating registration and data sharing, and prioritizing reproducibility and transparency in study design. John P. A. Ioannidis offers concrete remedies—expanding peer review, improving workforce training, and embracing robust statistical methods. The adoption of these practices depends on institutional will and the capacity of the scientific community to realign incentives with values of rigor and public responsibility.
The Future of Science in Society
Science on the Verge culminates in a call for humility, reflection, and innovation in the practice and governance of science. The text urges a shift from speaking “truth to power” toward working deliberatively within imperfections, recognizing science as one knowledge system among many. Integrating extended participation, reinforcing quality controls, and accepting the limits of quantification lay the groundwork for a more resilient and responsive science. Where can science go from here? The book affirms that the future lies in ongoing negotiation—between disciplines, between science and policy, and among communities whose fates converge on the integrity of knowledge.
