A Lesson about Taking Chances

Richard Wiener is a Program Director at the Research Corporation for Science Advancement and a member of the Editorial Board of Physics and Society.

As Washington remains divided and highly partisan, it is clear we are in an era of little or no increase in federal support for fundamental research in the sciences. According to Michael Lucibella, writing in the November 2013 issue of APS News, “Federal spending on science has been nearly flat since 2010 as a percent of total discretionary spending and down in actual adjusted dollars.” In this climate, an obvious concern is that federal agencies will become even more conservative and solely commit limited resources to research likely to be successful even if resulting advances are only incremental.

As a Program Director for Research Corporation, a private foundation which supports fundamental research in the physical sciences, I struggle with the issue of how much risk my foundation should take. Everyone wants to support great science; no one wants to waste precious dollars. However, sometimes it is the willingness of funders to take a chance on truly risky projects that allows researchers to make the biggest breakthroughs. To remind myself of the importance of risk, I find it instructive to consider examples in which a funder didn’t play it safe and the resulting payoff was huge.

Here’s a case in point from the files of Research Corporation: the discovery that the expansion of the Universe is accelerating.

Back in 1980, Richard Muller, a physicist at the University of California, Berkeley, submitted a proposal for high-risk research to the National Science Foundation.

Ahead of the curve Muller recognized the potential for astronomical sky surveys using a CCD, at the time only recently developed for digital imaging, when operated with a small computer. Muller believed the CCD-computer combination would enable the automated search and discovery of supernovas in the early stages of their violent expansions.

When the NSF declined funding, Muller turned to Research Corporation. Muller wrote, “We propose an automated survey of several thousand galaxies every night, making full use of recent advances in automatic imaging techniques and small computers … A careful study of the supernova light curve and spectrum could allow the supernovas to be used as a “standard candle” for the measurement of Hubble’s Law and estimation of the average mass density of the Universe (i.e., to answer the question of whether the Universe is finite or infinite).”

Research Corporation’s records show the decision makers recognized how important, though risky, this project was. A review panelist wrote:
Grant. 1) Science is extremely important. 2) R.A. Muller is a prize-winning physicist – he’s a good experimentalist who gets things done. 3) The fact that even he has to come to us for funds shows again that the system is totally closed to new departures. 4) Muller’s previous ventures (observing 30K radiation from NASA airplanes and optical feedback control) show the same originality... He will vault over corpses...
And the program officer’s comment was prescient:
This is one of the more exciting proposals to come along in quite awhile. If Muller can pull this off, the ramifications could be enormous.

Research Corporation funded the project and Muller’s vision proved farsighted indeed; but it took several years to develop the technology to automatically locate supernovas. By that point Muller had founded the Supernova Cosmology Project (SCP) based at Lawrence Berkeley National Laboratory. Carl Pennypacker and Saul Perlmutter were key researchers participating in the project.

“In the early days, people thought measuring expansion with supernovas would be too hard,” reflected Perlmutter, as quoted by Paul Pruess, writing on the Lawrence Berkeley National Laboratory website. However, Pruess adds,
The SCP went on to show that distant supernovas, short-lived and unpredictable as they are, can nevertheless be collected “on demand,” allowing observers to schedule telescope time in advance and accumulate enough data to make confident estimates of expansion.
“In retrospect it seems obvious, but we realized that the whole process could be systematized [Muller’s fundamental high-risk idea],” Perlmutter explains. “By searching the same group of galaxies three weeks apart, we could find supernova candidates that had appeared in the meantime. We could guarantee four to eight supernovas each time, and all of them would be on the way up, growing brighter instead of already fading.”1

Eventually the SCP discovered the remarkable fact that the expansion of the universe, first announced by Édouard Lemaître in 1927, was actually accelerating. Perlmutter shared the 2011 Nobel Prize in Physics with Brian Schmidt and Adam Riess for providing evidence for that acceleration. By then cosmologist Michael Turner had coined the term “dark energy” as he and others struggled to explain the cause of cosmic acceleration.

Muller, although no longer searching for supernovas, continues to produce groundbreaking research and important new ideas, as evidenced by his leadership of the Berkeley Earth project, which produced an exhaustive analysis of historical global temperature change that shows a strong correlation with atmospheric carbon dioxide levels.

As Carl Pennypacker told me recently:
The fact that the Research Corporation was willing to take a chance on young investigators on potentially paradigm-shifting research was unique then, and is still largely unique and rare in modern science funding. As federal funds continue their relentless and steady decline, there is a tendency in many science communities to create experiments that are very safe and risk-averse and are designed by committee. Such safe experiments are important and needed, but they often cannot help develop the big breakthroughs we made.

1. http://newscenter.lbl.gov/feature-stories/2009/10/27/evolving-dark-energy/

These contributions have not been peer-refereed. They represent solely the view(s) of the author(s) and not necessarily the view of APS.