Innovation Is Not a Linear Race, It’s a Dance Between Discovery and Use
Investing more money in science is not enough to meet the challenges of the twenty-first century. We also must recognize a diverse set of approaches to scientific advancement.
One simple story that we tell about science is that basic research, driven simply by human curiosity, leads to discoveries, which enable applied research, which fuels the development of new technologies that then transform our lives and our economy. The operative metaphor of this simple story is that of a relay race, in which each runner hands off the baton to the next.
Another related, but equally simple, story starts with a lone scientist toiling for years to produce a profound insight that alters our understanding of the universe. The metaphor there is that of the long-distance runner.
This metaphor of a race, based on linear progression from insight to technology, is central to the way we talk about innovation in the United States. In no small part, that is because the compelling version of this story that Vannevar Bush told in 1945 led to the creation of the National Science Foundation, which has been a legendary success through the decades since. Still, a closer look reveals a history of advances in science and technology that is much less straightforward than these simple stories and metaphors would imply. The meandering path to success sometimes starts with a technological breakthrough and at other times with a fundamental breakthrough. The course of scientific progress is difficult to predict, and advances sometimes require insights from outside fields and inspiration from new challenges.
Now, with the perspective afforded by the COVID-19 pandemic, we can see unprecedented public, bipartisan, political support for ambitious new investments in science and innovation. Significantly, both the US Senate’s proposed Endless Frontier Act and the House’s proposed National Science Foundation for the Future Act embody a willingness—a desire, even—to boost innovation.
The proposed enhancements will be most effective if they both build on and preserve Vannevar Bush’s vision and simultaneously nurture and support these more complex pathways to innovation. The most effective strategies for advancing science and promoting innovation will alter the metaphor: not strictly a relay race moving linearly from basic to applied science, but rather a complex dance in which science and technology are partners at every stage. Both the Senate and House versions of legislation increase our national investment in basic discovery-driven and curiosity-driven science. Both also put significant new investment into “use-inspired” science, in which uses include technologies with commercial applications as well as new tools for further basic scientific advances. The proposed legislation also includes investments in people, programs, and the building of new collaborations and institutions. In addition, the legislation requires attention to how these investments are allocated to increase diversity in the STEM—science, technology, engineering, and mathematics—talent pool and enable a broader geographical distribution of technology-driven economic growth.
Maximizing the long-term return on investments in science requires that we understand the rich history of the way ideas and solutions develop. Basic research can be curiosity-driven or use-inspired—or a combination of the two. And research that is not use-inspired may become so later. When François Jacob, Sydney Brenner, and Matthew Meselson discovered mRNA in 1961, they were working to understand the fundamental processes that are at the basis of life. Building on this basic discovery, in the 1990s Katalin Karikó had the vision that mRNA could be used to fight disease. Today, mRNA vaccines are protecting us against COVID-19 and enabling our society to begin to return to a new normal.
These unexpected stories of science and technology dancing together can be very clearly seen in the development of instruments, which is usually seen as applied research. Experimentalists are constantly pushing the capabilities of exquisitely sensitive instruments as they explore the very small, the very large, or the very rare. But these are not one-way innovations: these and other use-inspired advances have applications “back” to fundamental research, and also “forward” to the development of economically important new products.
Consider the story of the development of the internet. In 1989, Tim Berners-Lee, a British scientist working at the CERN particle physics laboratory, conceived and developed the World Wide Web to meet the demand for automated information-sharing between physicists in universities and institutes around the world. John O’Sullivan, an Australian engineer developing novel approaches for detecting radio pulses from neutron stars, had started inventing new ways to detect weak signals in the 1970s. Many years later, his Wi-Fi patents became the basis for technology that is pervasive in our daily lives.
And, as these stories show, basic science is not always leading the dance. Advances in technology can work “backwards” to stimulate the development of fundamental science and mathematics. Practical steam power, the revolutionary technology that launched the Industrial Revolution, came before—and stimulated—development of the theory of thermodynamics. Subsequently, this deeper theoretical understanding made possible more powerful and more efficient steam engines, and the later development of internal combustion engines and turbines.
Today, we can see a similar trend in industry’s development of practical artificial intelligence—for speech and handwriting recognition, fraud detection, and commerce—that is now driving a more rigorous understanding of machine learning in academic settings. Meanwhile, the drive to understand the unexpected power of AI and machine learning is motivating theoretical computer scientists and mathematicians to explore the behavior of functions in high-dimensional spaces with new modalities. Standing in the midst of the rapidly changing field of machine learning today, one finds it difficult to predict where the field will have its greatest impacts.
Technology development can itself lead to serendipitous fundamental discoveries. When Arno Penzias and Robert Wilson discovered the cosmic microwave background, the leftover heat from the Big Bang, they were working to find something much more prosaic: the source of “sky noise” in AT&T’s work on microwave communications. It’s not a coincidence that they were working at Bell Laboratories, which to this day provides the historical prototype of the use-driven laboratory that explored scientific areas deemed relevant to AT&T’s role in communication technologies. By giving its scientists and engineers the freedom to explore, Bell Laboratories is now credited with discoveries in radio astronomy, the laser, the transistor, the charge-coupled device image sensor, and the fundamentals of information theory—work that earned its alumni nine Nobel Prizes and five Turing Awards.
Just as there are multiple paths to discovery and innovation, there are multiple ways to support the science enterprise in its advance along these paths. The proposed legislation recognizes the importance of pre- and postdoctoral fellowships and traineeships that advance young scientists to careers of creative independence. It also recognizes the complementary roles of “big” and “small” science, and the importance of the complementary institutional jewels of our science enterprise, the research universities and national laboratories.
This deepened investment in science is essential for facing the challenges of the twenty-first century. However, increased spending is not enough. We must recognize that there are multiple paths of discovery and innovation, and multiple means for supporting such paths. This complex dance of science and technology motivates a national science strategy that supports basic research in both its curiosity-driven and its use-inspired forms, while also supporting applied research and, in partnership with private sector industry, translational research and development. To effectively advance innovation for society, we must support and enable a diverse set of approaches to scientific advancement.