This week, New 杏吧原创 begins its 50th year of reporting on scientific discoveries. But what will produce the stories behind the next 50 years of headlines? Is the process of scientific breakthrough a product of hard work, personality, inspiration, intelligent guessing or plain old-fashioned luck? Physicist and novelist Alan Lightman explores the recipe for scientific progress
IN ONE of the most remarkable narratives of scientific discovery, Otto Loewi, at age 87, recalled how the idea came to him for testing the way nerves communicate. The thought arrived in a dream:
鈥淭he night before Easter Sunday of [1921] I awoke, turned on the light, and jotted down a few notes on a tiny slip of paper. Then I fell asleep again. It occurred to me at 6 o鈥檆lock in the morning that during the night I had written down something most important, but I was unable to decipher the scrawl. The next night, at 3 o鈥檆lock, the idea returned. It was the design of an experiment to determine whether or not the hypothesis of chemical transmission [of the nervous impulse from nerves to their respective organs] that I had uttered 17 years ago was correct. I got up immediately, went to the laboratory, and performed a simple experiment on a frog heart according to the nocturnal design.鈥�
Advertisement
At the time of Loewi鈥檚 dream, it was well known that signals travel down the spindly filaments of a nerve in the form of electricity. What was not known was how nerves conveyed their electrical impulses to muscles, organs and other nerves. In short, how did nerves talk to the rest of the body? Most biologists believed that such communication was also electrical 鈥� that tiny currents flowed from nerves to heart muscle or thyroid gland or the waiting antennae of other nerves.
Loewi鈥檚 late-night experiment was not only simple but elegant. First, he isolated the live hearts of two frogs and removed the nerves from the second heart. Into both hearts Loewi inserted a tube filled with Ringer鈥檚 solution, a liquid that matches the concentration of salts in the body and keeps isolated organs alive. He then stimulated the vagus nerve of the first heart. The vagus slows down the functions of organs, and the frog heart鈥檚 rate of throbbing decreased as expected.
After a few minutes, Loewi poured the Ringer鈥檚 solution from the first heart into the second, nerveless heart. It also slowed. Its beating diminished just as if its own vagus nerve had been roused. Similarly, when Loewi stimulated the 鈥渁ccelerator鈥� nerve of the first heart and then transferred its liquid to the second heart, that heart speeded up. The results showed without doubt that a stimulated nerve releases some chemical substance that then activates organs. The transmission of nerve impulses to organs is chemical, not electrical. Loewi had discovered the existence of neurotransmitters such as acetylcholine and adrenaline. His finding would eventually have an impact on everything from our knowledge of brain function to the treatment of neurological disease.
Einstein had his thought experiments and his philosophical principles, Ernest Rutherford had his wildly intuitive 鈥渄amn fool鈥� ideas, and geneticist Barbara McClintock recognised meaning in the patterns on leaves of corn. But few scientists have reported receiving their great ideas in a dream.
How did Loewi arrive at his dream? More generally, how do scientists arrive at their discoveries? Several years ago I decided to explore some of the great discoveries of science in the 20th century 鈥� to embark on a discovery of discoveries. As case studies, I chose two dozen discoveries from physics, chemistry and biology. How did they happen? Which discoveries were accidents and which were purposeful? Were there common patterns of thought? How did styles of working and thinking vary from one science to the next, from one scientist to the next?
From my sample, I have developed what one might call a taxonomy of scientific discovery. Any such classification scheme is subjective, of course. There is a broad range of different kinds of discoveries, and they may be categorised in different ways. Furthermore, no one knows exactly what happens in the creative process or what chain of mental and external events leads to recognition and discovery. With these caveats, and with the hope of stimulating further discussion, here is a tentative taxonomy. It will undoubtedly be challenged and revised in the future.
1 The Accident,
in which the scientists discover something they were not looking for. This category breaks down into (a) discoveries whose significance the scientists did not understand at the time, and (b) discoveries whose significance was immediately appreciated.
A good example of (a) was the discovery of the cosmic background radiation by Arno Penzias and Robert Wilson in 1965. Those scientists, both excellent experimentalists, had no explanation for the residual hiss in their radio antenna, corresponding to a background cosmic temperature of about 3 kelvin. An example of (b) was Alexander Fleming鈥檚 discovery of penicillin in 1928. Although Fleming was completely surprised to discover a white fluff growing on his culture of staphylococci and, furthermore, that this white mould dissolved the staphylococci nearest it, he instantly realised that he had stumbled upon an antibacterial agent.
2 Principles First,
in which the scientist begins with a philosophical principle and explores the consequences of that principle, sometimes initially unaware of the precise problem to be solved.
The application of the foundational principle can, in fact, define both the problem and its solution. The premier example of this rarefied category was Einstein鈥檚 work on special relativity, in which the physicist began with the 鈥渟ymmetry鈥� principle that all frames of reference moving at constant velocity relative to each other are equivalent. In the process of working out the consequences of his starting principle, Einstein discovered that our notion of time had to be reconceived.
3 Principles Last,
in which the scientist engages in concentrated work to explain a particular experimental result and ultimately recognises that a new fundamental theoretical idea is required.
Max Planck鈥檚 discovery of the quantum in 1900 illustrates this category. The German physicist was trying to use the methods of thermodynamics and statistical physics to justify his own ad hoc formula for black-body radiation (itself a discovery). To do so, he had to assume that the energies of his 鈥渧ibrating resonators鈥� were not continuous and infinitely divisible but came in whole lumps 鈥� the quanta.
It should be pointed out that both categories (2) and (3) concern a new principle and apply only to theoretical discoveries.
4 The Timely Clue,
in which the scientist is confronted with an important clue while struggling with a recognised problem.
An example of this category was Niels Bohr鈥檚 creation of the first quantum model of the atom in 1913. While trying out different models of the atom, in which various parameters of the electron鈥檚 orbit were quantised using Planck鈥檚 new quantum constant, Bohr was shown Johann Balmer鈥檚 1885 empirical formula for the frequencies of light emitted by hydrogen atoms. The key clue was that each emitted frequency depended on one quantised number subtracted from another. Bohr had never before seen that formula.
鈥淎s soon as I saw Balmer鈥檚 formula,鈥� Bohr later recalled, 鈥渢he whole thing was immediately clear to me.鈥� Bohr then hypothesised that electrons emitted photons in passing from one quantised energy level to another, the energy of the photon being the difference in energies of the two levels.
Another illustration of the timely clue was Barbara McClintock鈥檚 discovery in the late 1940s that genes could move around on chromosomes, and in doing so alter controls, commands, and the range of stored information. For some years, McClintock had been pondering how genes, including pigment-controlling genes responsible for yellow leaf streaks, were turning on and off during the growth of a single corn plant. Variations in streakiness appeared to occur in some regular way, not in the random fashion you would expect from mutations.
Carefully examining her plants one day in 1946, McClintock noticed that these 鈥渃ontrolled鈥� mutations came in pairs. For example, if one section of a corn leaf had a higher than average number of streaks, a neighbouring section had a lower than average number. Here was a big clue! Since each of the sections of the leaf originated from a different progenitor cell, it appeared as if one of the progenitor cells had given something to the other during cell division.
McClintock recalled that when she first observed the 鈥渢win-sector phenomenon鈥�, it 鈥渨as so striking that I dropped everything, without knowing 鈥� but I felt sure that I would be able to find out what it was that one cell gained and the other cell lost, because that was what it looked like鈥�. This observation led first to the idea that control elements (genes) were passed between sister chromatids, the products of a chromosome鈥檚 duplication before cell division, and later that control elements could change position on individual chromosomes.
In both of these examples, the scientist needed a good understanding and vision to recognise the importance of the clue. A clue that comes before its time is of no use. The remarkable fit of the opposite coasts of South America and Africa, like adjacent pieces of a jigsaw puzzle, was known for centuries but never recognised as a clue to the ancient geography of the Earth until Alfred Wegener proposed his theory of continental drift in 1912, along with the radical idea that land masses could move horizontally across the surface of the Earth.
And Robert Dicke鈥檚 observation in 1969 that the gravitational and kinetic energies of the universe are approximately balanced was never taken as a serious clue to the revisions needed in the standard big bang model of cosmology until Alan Guth鈥檚 theory of the inflationary universe in late 1979.
5 Analogy,
in which the scientist applies a concept or pattern from a previous problem.
An example of this kind of process was Hans Krebs鈥檚 1937 discovery of the citric acid cycle, also known as the Krebs cycle. Krebs and other biochemists were trying to discover what chain of chemical reactions was responsible for respiration 鈥� that is, the combination of oxygen with carbohydrates and fats to release energy in living organisms. Other scientists had found pieces of the chain. Several years earlier, Krebs had made the first discovery of a cyclic process in biochemistry, the ornithine cycle, in which ornithine is changed to citrulline, which is changed to arginine, and then back to ornithine, ready to begin the cycle again. Along the cyclical path, the toxic molecule ammonia is converted into urea and removed from the body. The citric acid cycle converts citric acid into a sequence of other substances, eventually returning to citric acid, while hydrogen atoms are pulled off the intermediate molecules to combine with oxygen to form water and release energy.
Krebs had cycles on his mind. The scientist searched for the missing chemical reactions and substances that would regenerate citric acid, thus allowing the sequence of steps to occur in a continuous loop. As he wrote in his memoir: 鈥淚n visualising the cycle mechanism it was of major relevance that five years earlier I had been concerned with the first metabolic cycle to be discovered, the ornithine cycle of urea synthesis.鈥�
6 The Mathematical Imperative,
in which a theoretical scientist, in exploring the mathematical world, is led to a discovery about the physical world.
A prominent example of this type of discovery was Paul Dirac鈥檚 discovery in 1928 of the equation describing the electron. The requirement that such an equation embrace both relativity and quantum mechanics in turn necessitated a particular mathematical structure. In following the narrow path of this mathematical landscape and its internal logical consistency, Dirac was directed to his discovery.
7 New Tools,
in which the availability of new instruments or new theories opens up opportunities for discovery. On the experimental side, this category might be further divided into (a) an inspired idea of how to use a new technology and (b) privileged access to new technology.
An example of (a) is Max von Laue鈥檚 realisation in 1912 that the regular spacing of atoms in a solid crystal would serve as a three-dimensional diffraction grating whose structure could be probed with a collimated beam of X-rays, just as a one-dimensional diffraction grating spreads a monochromatic beam of visible light. X-rays were new at the time. Furthermore, they were known to have wavelengths comparable to the spacing between atoms in solids, which was just what was needed for the job.
An example of (b) was Edwin Hubble鈥檚 discovery in 1929 that the distances to galaxies are approximately proportional to their recessional velocities, a fact later used to support the notion that the universe is expanding. Other scientists were also attempting to measure the distances to a group of exceptionally fast-receding galaxies discovered by Vesto Melvin Slipher. Hubble鈥檚 advantage lay in his exclusive access to the relatively new 100-inch Hooker telescope on Mount Wilson in California, at that time the largest telescope in the world.
An example of discovery following a new theory is Linus Pauling鈥檚 development of the modern theory of the chemical bond in the late 1920s and early 1930s. When the young chemist travelled to Arnold Sommerfeld鈥檚 Institute of Theoretical Physics in Munich in 1926 to learn the brand-new quantum physics, Pauling found himself in the enviable position of being the only chemist at the institute. Pauling was one of the first chemists to apply quantum mechanics to the nature of the chemical bond, using his particular genius to find solutions that minimised the energy for electrons shared between two atoms.
8 The Long Haul,
in which steady, incremental work on a recognised problem over a long period of time leads to discovery. An example is Max Perutz鈥檚 discovery of the three-dimensional structure of haemoglobin, one of the first protein structures to be worked out. Perutz and his team worked on the problem for 22 years, from 1938 to 1960, painstakingly producing and analysing hundreds of X-ray diffraction photographs and refining their experimental technique along the way.
Patterns of discovery
A taxonomy of eight classes is relatively large from a sample of only two dozen case studies. Its real test is to see what proportion of other discoveries fit or do not fit. Of course, many discoveries are messy affairs and spill over into several categories.
Although there is clearly a wide range of processes in scientific discovery, some common patterns emerge. Most discoveries involve a synthesis, bringing together different strands of information or ideas and connecting them. To formulate his quantum model of the atom, Bohr combined Rutherford鈥檚 nuclear model of the atom, Planck鈥檚 idea of quantised energy levels and Einstein鈥檚 idea of the photon.
Another pattern that happens in many, but not all, scientific discoveries is the following sequence of events: research and hard work, leading to the 鈥減repared mind鈥�; being stuck on the problem; finally, a shift in thinking or perception. German physicist Lise Meitner鈥檚 understanding of nuclear fission followed this pattern. So did McClintock鈥檚 work on the mutations in maize, as did James Watson and Francis Crick鈥檚 work on the structure of DNA. In the latter case, for example, Jerry Donahue鈥檚 hint that the dangling hydrogen atoms on the nitrogen bases should be in different locations provided a change in thinking about how the bases could fit together.
The prepared mind is critical. I know of no examples of major scientific discoveries in the 20th century made by untrained amateurs. Even when the discovery is accidental, it requires a prepared mind, as in Fleming鈥檚 discovery of penicillin. The Scottish scientist had been working on antibacterial agents since his 1908 medical school thesis. Being stuck also seems important in many discoveries. This frustrated mental condition, combined with purpose and a prepared mind, can catalyse the creative imagination.
These patterns of discovery are probably universal to the creative process in general and occur in the arts as well as in the sciences (New 杏吧原创, 29 October, p 39). Writers have given similar accounts of their creative process in dozens of interviews in The Paris Review over the last two or three decades. In an interview in 1990, for example, the writer Wallace Stegner commented: 鈥淚 don鈥檛 go in search of projects. Sometimes they appear before my eyes, and sometimes they grow over a long period of time as I brood鈥n the case of Crossing to Safety, the book just grew, more or less, through personal experience in Vermont, Wisconsin, and to a small extent in Italy鈥 knew from the beginning it was going to be a book. You have that feeling. It鈥檚 like having a fish on the line. You know when it鈥檚 an old boot and when it鈥檚 a fish. But I didn鈥檛 know what the book was鈥 had to do it by trial and error.鈥�
In Janet Sonenberg鈥檚 book The Actor Speaks, leading actors describe their acting process. From John Turturro: 鈥淥nce the scene鈥檚 dynamic is starting to occur, I鈥檒l go with it and then try to shift it, too, just like you would in life. The shifting is important. Then, if I can get to the point when that鈥檚 happening and I don鈥檛 know what I鈥檓 doing, that鈥檚 inspiration. I鈥檝e done all my work and then I try to achieve this other, living dimension.鈥�
As is well known among scientists but not among the general public, there is no single scientific personality type. Great scientists can be bold and self-confident revolutionaries, like Rutherford or Einstein or Watson. Great scientists can also be modest and diffident, like Krebs or Fleming or Meitner. William Bayliss, who discovered the first hormone in 1902, was cautious, meticulous and in love with the details, while Ernest Starling, his collaborator, was brisk, impatient, engaged mainly by the broad sweep of things.
What all of these men and women share is a passion to know, a sheer pleasure in solving puzzles, an independence of mind. Barbara McClintock, for instance, recalled that in high-school science classes she 鈥渨ould solve some of the problems in ways that weren鈥檛 the answers the instructor expected鈥t was a tremendous joy, the whole process of finding that answer, just pure joy.鈥�
When Lise Meitner was a child, her grandmother cautioned her that she should not sew on the Sabbath or else the heavens would come tumbling down. The little girl decided to do an experiment. She lightly touched her knitting needle to some embroidery and looked up. Nothing happened. Then she took a stitch, waited, looked up. Again, nothing. Finally, Lise was satisfied that her grandmother had been mistaken and went happily about her sewing.
