we develop the tools of discovery There is another side to the coin: researchers have focused on exciting serendipitous moments in discoveries, but unexpected scientific tools used to spur them have not been studied. Here we explore all 149 nobel-prize-winning method discoveries— all major methods and instruments awarded the prize. Interestingly, these major new tools are less likely to involve a serendipitous moment—about one in eight— compared to scientific discoveries they enable—about one in six. By examining these major tools, we find they almost always emerge from specific method constraints. Our tools are developed deliberately. This is an important distinction: one between intentional tool development and the serendipitous observations they enable. Only few nobel-prize-winning tools emerged somewhat unexpectedly—such as the x-ray method —or were invented without a specific application in mind like the maser (built just to amplify beams of microwaves). The first microscope and telescope were also invented with no scientific purpose. Yet others then harnessed these surprisingly effective tools. Such a tool can power a kind of metaserendipity: it opens up the possibility of multiple discoveries that no one could have predicted before the tool became available. Serendipity generally comes after the tool is developed and applied. What is key here is that some powerful tools are not originally designed for the specific breakthroughs they later enabled—they were never meant for that purpose. And yet, the tools still caused the new discoveries they were not originally intended for—and would not have been possible without them. Again, this disconnect between a tool's original purpose and its later impact provides a kind of quasi-experimental insight: because the tools' creators had no idea what breakthroughs they would eventually unlock, we can directly attribute the breakthroughs to the tools as the causal spark. It also helps reduce alternative explanations as the key trigger, like serendipity (Chapter 1). Yet some researchers think serendipity 'is one of the important factors in scientific discovery' . But what is clear is there is no chance of serendipitous discoveries before we generate the enabling tools—and once we create them, the likelihood of surprising findings rises exponentially (see Figure 1.8). This shows that chance plays more of a background role, while new discovery tools commonly spark chance moments. Science also reveals fascinating cases of multiple discoveries where the same phenomenon is uncovered by scientists working independently of each other. But we find that they often actually leverage the same transformative tool to uncover them. A classic example is discovering oxygen independently by both Scheele in Sweden and Priestley in England, using the same method: a pneumatic trough with mercuric oxide. Numerous examples of nobel-prize breakthroughs exist—showing that they do not just arise by a chance observation, but by powerful methods. Take giant magnetoresistance discovered independently in 1988 by Fert in France and Grünberg in Germany—both implementing the same technique invented in 1968: molecularbeam epitaxy. The breakthrough makes it possible to read data on computer hard disks. Take the structure of antibodies uncovered in 1959 by Edelman in the US and Porter in England—both applying methods like improved chromatography created in 1956. This breakthrough has enabled understanding how antibodies in our blood defend us against deadly infections—and helped found immunology. The chemistry of compounds called organometallic sandwich compounds was revealed in 1952 by Wilkinson in the US and Fischer in Germany. Both used methods like x-ray crystallography created in 1913. This revolutionised our understanding of chemical bonding and spurred new fields in chemistry. E Even the theory of the Higgs particle was developed independently in 1964 by Higgs in Scotland and Englert and Brout in Belgium. Both used mathematical gauge theory formulated in 1954, while also building on recent empirical breakthroughs in particle physics—and motivated by key insights into how particles interact. This field discovered many fundamental particles over the previous decades using particle detectors and accelerators (Chapter 1). The existence of the particle was then confirmed using the world's most powerful particle accelerator, the large hadron collider.
When most think of major scientific advances, the moment of discovery comes to mind—the final output. Other aspects of the discovery process simply do not catch the eye like a flash of insight and final discovery. Behind breakthroughs, there is a process not only of developing methods but also of designing experiments, collecting and analysing data, and replicating results—all powered by the tools we create. These features are just not as appealing, but we can trace them in the discovery process. They make evidence—and discoveries—measurable, reliable and replicable. Our scientific toolbox is the bridge that connects us to the world and how we measure, experiment and generally experience those moments of surprise. By leveraging the tools of a scientific community, we—and science as a whole—transcend individual intuitions and chance moments.What is fascinating is that, independent of serendipity, other researchers often apply the same tools to arrive at the same findings or discovery. This deflates what seems like an individual's (or lone genius's) unique serendipitous moment to a common tool of discovery used by multiple researchers. Serendipitous view of the discovery process driven by a flash of insight Scientific toolbox view of the discovery process driven by systematically using new tools serendipitous moments, if they arise, are often sparked by methods and tools used to perceive, measure and study the world . Louis Pasteur said, 'chance favours only prepared minds' . And Nobel laureate Albert Szent-Gyorgyi expanded on the idea, ' A discovery is said to be an accident meeting a prepared mind' . The idea is that without training, a serendipitous discovery can go unrecognised. But when we examine science's major discoveries, what becomes clear is that: chance favours those applying new methods and tools . The scientific toolbox view explains how we discover by using new tools that open a new lens to the world and it integrates the possibility of chance surprising us by using these tools. Think of breakthroughs described as serendipitous—like discovering cells and bacteria after inventing the microscope, or uncovering Saturn's rings and Uranus after creating the telescope. With these remarkable new tools, other researchers could have made the surprising new observations (whether we call them unexpected discoveries or not).If we do not know what factor to look at, then the surprise seems random—but once we begin turning our attention to the tools that commonly spark surprise, much of the randomness of discovery fades. In many ways, those were basically inevitable discoveries once we came up with the needed tool and put them in the hands of researchers. This insight is important: it moves the spotlight from individual moments of serendipity and places it on the innovative tools that drive surprising findings.
Most people think serendipity is the most puzzling and the least predictable aspect of science and discovery. After all, it is about coming across something we did not expect. Discovery generally means uncovering a surprising finding, but surprise generally means we cannot anticipate it. So if serendipitous discoveries arise by chance, how could they possibly be predictable or replicable? What we have uncovered here is that surprising new findings commonly follow a hidden pattern. In fact, serendipitous discoveries across science and history—from revealing viruses to the first exoplanet— are sparked not by chance, but generally by applying a new tool to a problem: from recently created electron microscopes to spectrographs. So, they are best described not as serendipitous, but as methods-powered discoveries. They are hidden in plain sight—until we use the right tool to reveal the unexpected. Other researchers who would have gotten their hands on these new tools could have surprisingly observed them—and discover what we were not even looking for. Chance and serendipity alone cannot make any discovery possible, highly likely or replicable, but powerful new tools we develop to reveal the hidden world can. In fact, serendipitous discoveries are generally even made with limited or no theoretical understanding of what is revealed. The key overlooked insight we uncovered is powerful: because new tools commonly unlock surprise, we can actually design and engineer discovery much more than researchers have thought. By developing sharper instruments, new experimental techniques, smarter AI algorithms, we actively create the conditions for serendipity—and multiply the odds of surprise with ever more powerful tools. New tools are the levers that uncover the unknown—whether we are surprised by what we find or not. Take the breakthrough gene-editing method, CRISPR. The method was not the original focus but emerged serendipitously, using improved methods such as differential RNA sequencing. The method itself then opened up entirely new research areas that were not expected by the method's creators: from finding cures for genetic diseases to developing new crops. This extraordinary method did not just enable a single breakthrough, it creates the foundation for a series of breakthroughs. When we shift our attention from serendipitous moments in the discovery process to the new tools that enable them, we realise that it is new tools that ensure discoveries are replicable. Different researchers using the same tool can arrive at the same findings. At first glance, serendipity seems to offer the weakest evidence of a logic of discovery, but it actually provides the strongest evidence we have. The fact that discoveries can be replicated using the same tools shows that science is far less random than we think. It often does not make much difference if one labels discoveries serendipitous or not: we do not role dice to generate discoveries but we pick up and apply new tools that make our major breakthroughs possible. What has been more random is who gets their hands on the tool first. So with powerful tools, we move from a random, unpredictable approach to discovery to one where breakthroughs are structured, engineered and repeatable. Let us picture the landscape of scientific discovery, where about 750 major discoveries make up the foundation of science, about nine million scientists are actively doing research worldwide and roughly '80 to 90 per cent of all scientists that have ever been, are alive now'. The odds of uncovering a major discovery in one's lifetime are thus very low. This reflects a lower-bound estimate of about a 1-in-11000 chance. But there is an unnecessary role of chance that still plays out in many discoveries—and in the speed at which they emerge. It is the chance of someone coming across the right tool, or improving one, without yet a theoretical understanding of how new tools trigger new breakthroughs. While chance may arise in discovery, giving the impression it is unexpected and unplanned, it takes a new tool to realise that chance moment—and spark any major discovery. In fact, we can accelerate—and anticipate—unexpected moments by making new tool innovations and applying them in new terrain. In short, fewer things are left to serendipity and chance if we create new tools in a targeted way: more new tools systematically foster more new tool-powered discoveries—whether we label them surprising or not. Here we reframe discovery as an active, design-driven process, offering a productive alternative to common misunderstandings of passive serendipity. Serendipity is especially surprising if not aware of this tool-driven principle of discovery. We can more accurately label most cases of serendipity as methods-driven—or even methodipity, a fortunate and unexpected new discovery sparked by a new method. Just as science uncovers patterns and stability in the world, we uncover patterns and stability in the discovery process itself. Unlike common belief, there is a nature and logic of discovery and it is driven by our new tools. With this new methods-driven discovery rule, we can demystify much of the discovery process. In 1873, the Polish Boleslaw Prus predicted that 'Someday a science of making discoveries and inventions will exist'. But systematic data were lacking at the time. Over 150 years later, we now have systematic data on science's major discoveries and how they are triggered with a new method or instrument. So we can lay out here a science of discovery: one that explains how developing new methods and tools is a necessary condition for major discoveries, and commonly the key trigger. The more we innovate, the faster discoveries will come—and the more we can predict and guide future breakthroughs. It is time we replace the classic image of the romanticised discoverer—someone who has a sudden eureka-like insight of luck—with a new image: the tool-driven discoverer, who builds on powerful methods and existing research to unlock the unexpected. We need a fundamental rethinking about scientific breakthroughs. While discoveries receive the most attention in science—by journals, university departments, hiring committees, awards and funding bodies—the tools that spark those discoveries are not celebrated in the same way. In fact, research into developing new tools remains heavily underdeveloped, and we will dig deeper into how this limits scientific progress in Chapter 6. The fact that we break new ground with recently invented tools significantly decreases the role of chance in discovery. The more we rely on new tools, the less room there is for chance. By strategically developing new tools—and testing them in unexplored areas—we reduce randomness and serendipity in both expected and unexpected discoveries. The future of science is in our hands—quite literally, through the tools we create. Expanding our toolbox allows us to better predict and speed up the pace of new breakthroughs following a new tool advance. In fact, shifting greater focus of research, labs and institutions towards devising new tools is commonly the most efficient path to discovery we have. Researchers with an open eye (for unexpected techniques) and with an exploratory mind (for pursuing and applying them) are commonly the ones who spark our future discoveries. If we want to make the next breakthrough, the best strategy we generally have is straightforward: develop and apply a new tool—or adopt one from another field—that gives us an entirely new perspective to a problem. To wrap up, we return to two prominent researchers and the stories they tell about scientific progress. We test here two of the most influential explanations of science to date: Karl Popper's explanation that discovery has no logical structure but is shaped by chance observations and intuitions (Chapter 2) and Thomas Kuhn's explanation that science does not progress cumulatively (Chapter 3). These different views sparked intense debate and intellectual rivalry. But by systematically examining how discoveries emerge, we uncover a very different picture of science: discovery does in fact follow a predictable logic, and scientific progress is cumulative—with both driven by our powerful and expanding toolbox. In the next chapter, we turn to how our discoveries—expected and unexpected—fit into the overall picture of scientific progress.