Dr. Arden L. Bement, Jr.
National Science Foundation
"The Conduct of Science Is Not What It Used to Be"
Address to Philosophical Society of Washington
October 22, 2004
Good evening, everyone. I'd like to thank Bob Hershey for this invitation to address the Philosophical Society of Washington, and I'm honored to be your speaker tonight.
From the standpoint of acting director of the National Science Foundation, I plan to offer some observations on the changing conduct of science in our times.
I'll begin, however, by turning to a famous detective who was known for his observations--Sherlock Holmes. I recall reading that one time, in the course of an especially taxing investigation, Holmes and Dr. Watson were obliged to camp out for the night.
As they bedded down, Holmes asked his colleague, "Behold the sky, Dr. Watson: What do you see?"
"I see millions and millions of stars," the doctor replied, "and I infer that they represent millions of galaxies. What do you infer, Holmes?"
"Elementary, my dear Watson," said Holmes; "Somebody stole our tent!"
This tale reminds us that the real meaning of observations can be a challenge to perceive and to understand. It's been said that "The eye cannot see what the mind can't perceive." Today, however, new tools for research and computation are turning that old saw on its head. The process of observing and understanding in science is ever-evolving.
As philosopher Pierre Teilhard de Chardin said, "The history of the living world can be summarized as an elaboration of ever-more-perfect eyes within a cosmos in which there is always something more to be seen."
Indeed, the conduct of science is not what it used to be--and that's what I've titled my talk. The transformation of scientific tools has brought us observations unprecedented in quality, detail and scope--observations that sometimes hover on the edge of comprehensibility.
Evolving in concert with the new tools are different ways of working within science, such as collaboration across large, multidisciplinary, often international teams. These new modes of working are essential to meeting the grand scientific challenges of our era--those overarching goals that require a concerted, large-scale effort. Such grand challenges can only be met by changing the conduct of science to encompass the social sciences in a truly organic way.
I'd like to explore with you, now, this interplay of tools, observations, and human dynamics that is transforming how science takes place.
The new modes of conducting science employ embedded sensors in large grids, synthesis of massive databases, computational models of complex behavior, and international teams. We see these patterns whether the topic of investigation is earthquakes, ecological systems, oceans, or even gravitational waves.
Computing power is critical to refining the "ever-more-perfect" eyes that bring unparalleled acuity to modern scientific vision. In this vein, I would like to recall an aircraft incident from 1992. Investigation of atmospheric conditions at the time illustrates how computing can help to "see" what was heretofore invisible.
On December 9, 1992, a DC-8 cargo jet was flying over the Colorado Front Range at 9.7 kilometers altitude. Completely unexpectedly, it encountered extreme turbulence--clear-air turbulence--which ripped off an engine and nineteen feet of wing.
Miraculously, the plane did land, and researchers at the National Center for Atmospheric Research began to investigate the case. Only by transforming the weather data into computer images, however, were they able to see the cause. They were able to discern, in the atmosphere, horizontal vortex tubes--turbulence--at the aircraft's elevation. 1
"More and more scientists are having trouble understanding their data," NCAR scientist Don Middleton, who worked on the study, told Science Magazine. "Now we have the computing power to use visualization as an important tool."
Although we marvel at the ability to see the once-invisible, we need to incorporate the social sciences to fully realize the benefit of such vision.
Recall, for example, the space-shuttle Challenger disaster. It became evident afterward that the engineers had foreseen the threat to the launch, but communication failures had prevented their analysis from halting the liftoff. Understanding how organizations can evolve to improve sensitivity to safety concerns is a compelling area for social scientists to examine.2
I'd like to turn now to some illustrations of how new tools are transforming a range of disciplines. My first example is from the geosciences and it is one of the major platforms within a comprehensive program called Earthscope.
This observatory, newly installed three kilometers down in California's San Andreas Fault, is now probing one of the world's most active faults. The animation shows how the drill has burrowed down through the granite beneath Parkfield, California, puncturing the fault like a soda straw.
Sensors lining the tunnel will be able to search--for the first time--for signals that could alert us to a major earthquake. Earthscope in its overall configuration will ultimately give a three-dimensional view of the North American continent, from surface to core.
Earthscope exemplifies a new way of learning about the Earth's structure. Up to now, we've had to wait until an earthquake happens, and then study the resulting seismic records--or turn to rock outcrops to piece together earth history. Now we're moving beyond static, limited views, to sweeping spatial scales and continuous observations over time.
As NSF's chief geoscientist, Margaret Leinen, puts it, "It's not postage stamps and snapshots in geoscience anymore."
A kind of flipside to Earthscope is NEES--the Network for Earthquake Engineering Simulation, dedicated to the grand challenge of preventing earthquake disasters. NEES facilities will simulate earthquakes and study how infrastructure and materials perform during seismic events.
Why focus on earthquakes? According to a National Academy of Sciences study, a large earthquake in an urban area could cause economic loss on the same magnitude as the 9/11 attacks.
Here is one component of NEES: the Multi-Axial Subassemblage Testing Laboratory at the University of Minnesota. This lab does three-dimensional testing of large-scale structural specimens. It is part of the NEES virtual laboratory encompassing facilities at 15 universities across the nation. The network will be inaugurated on November 15.
Up to now, engineers were tied to a lab that had physical equipment like a shake table. Now, a community of networked researchers and students in remote locations will be able to observe and participate in experiments at the university sites, and will access data in a central repository. NEES should extend the ability to do earthquake research to many more institutions.
Ecology is another discipline developing a blueprint for a network that will span the continent, and beyond. The National Ecological Observatory Network--NEON--is being designed to answer questions that cross space and time. It will enable ecological forecasting--helping to chart how climate change will alter forests and crops, how an infectious disease like West Nile Virus emerges and spreads, and how a foreign species, like the snakehead fish, disrupts a native ecosystem.
On a local scale, we can envision a forest site like this, embedded with sensors. Ideally, NEON will make measurements simultaneously at sites across the country, from forest ecosystems to prairie to coast, spanning biological scales from molecular to ecosystem.
Shortly we'll see a deer.
It knocks out a sensor, but the sensing and transmission network can heal itself.
I turn now from land to sea. Two centuries ago, a poet observed, "There is nothing so monotonous as the sea."3 How differently the world ocean appears now, while striking oceanographic discoveries are taking place even as scientific and environmental challenges mount.
The poet could have known nothing of the unique communities at deep-sea vents, the huge uncharted diversity of microbial life
living under the sea floor, the dynamic geophysics of plate boundaries, and the growing recognition of the ocean's role in so many global-scale processes.
To explore these grand challenges and even those that lie beyond, we'll need new research tools, such as the one portrayed in this animation.
This is the deeper-diving, human-occupied submersible that will replace the venerable Alvin. To be completed in 2008, the vessel will be able to reach more than 99% of the sea floor. It will have better visibility, sensors, and collection capabilities, among many other improvements.
Even the little known Arctic Ocean is being explored at the North Pole. This animation shows a research camp on the sea ice with an oceanographic mooring beneath. The mooring stretches more than two and a half miles down, and is anchored to the seafloor beneath the ice.
It is hung with instruments that record temperature, salinity, current speed and direction, and ice thickness and movement. This site at the top of the world is creating a benchmark to track fast-moving Arctic change.
Oceanographers are developing a broad and integrated vision for observatories, like the hypothetical one we see here off the state of Washington, that will take targeted samples and observe over space and time. For example, it will be possible to have observatories sample when triggered by actual events, such as the discharge of a hydrothermal vent. We'll be able to ask new kinds of questions, like: How much do vent emissions contribute to the earth's carbon budget?
We can begin to study biologically rich, but ephemeral, fronts in ocean waters that are so vital to open-ocean species. As NOAA oceanographer Robert Schick recently told Science, "There are a lot of well-developed ideas about spatial ecology on land. But we're just beginning to get more sophisticated about how these things work in the sea."4
Information technology is playing a key role for the oceans. A number of institutions are banding together to create a prototype grid to link oceanographers with wireless and optical networks to ocean observatories off Mexico, the United States and Canada.
Astronomical research is already moving to access on a global scale--with the National Virtual Observatory. It will offer access to all astronomy data and literature to anyone, wherever they may be. That all-important astronomical commodity, "instrument time," is no longer exclusive.
As this video shows, an astronomer will be able to integrate data from the entire observation spectrum. The Internet will become, as astronomers put it, "the world's best telescope,"5 a supercomputer at your desktop.
By enabling comparison of massive amounts of data from diverse sources--space and ground, and radio, optical, infrared, and other wavelengths--the NVO will bring the power to look at the most fundamental questions, such as the evolution of the universe.
This is a step toward what observers have termed the democratization of the conduct of science. Social scientists need to be part and parcel of studying this change.
Massive amounts of data present another challenge to astronomy--doubling in quantity every year in this exponential world.
As Alex Szalay, involved with the NVO, puts it, "The quantity of scientific data is so enormous that dealing with data is a whole new discipline in itself--that is happening in every branch of science."6
I'll turn to physics for one more example of a new, ambitious instrument that we expect to let us see a phenomenon never before detected-and that is gravitational waves. As this slide suggests, we have probed space using visible and infrared light, x-rays, and the cosmic microwave background, and each has opened up a new dimension of the universe. Gravitational waves may be the next window.
(from sound track)
"According to Einstein, every object causes a bending in the fabric of space.
"Einstein also realized that this bending of space could produce waves.
"He saw that if two objects moved around each other in orbit, they will create ripples in the curvature of space, ripples that expand throughout the universe, carrying clues about their origin."
The new Laser Interferometer Gravitational Wave Observatory (LIGO) consists of two interferometer detectors, one at Hanford, Washington and one at Livingston, Louisiana, which are separated by the length of the country. Ultimately, the LIGO detectors might reveal astrophysical phenomena not detectable any other way.
Tonight, I have surveyed some large, networked observation systems now employed to study phenomena from gravitational waves to invasive species to earthquakes. These systems have helped to draw scientists into interdisciplinary teams that span the globe.
The latest in these efforts is the Teragrid, which seeks to transcend the boundaries of place to accelerate collaborative science on complex challenges.
As NSF's chief computer scientist, Peter Freeman, puts it, "We are planning for facilities that a chemist can use this morning, a physicist can use this afternoon, and an earthquake engineer can use tonight."
The spectacular new tools of science and engineering are only part of the picture. I'd like to turn now to the final part of my talk7, with reflections on some implications of these trends for science and scientists.
Sociologist Robert Merton, who really founded the sociology of science in the 1940s, identified norms that science embodied. Sociologists today study how those norms are changing, and why--a discussion inseparable, I would offer, from the transformation in the tools we invent to do science.
Social scientists point to changes in the conduct of science driven by information technology. I have already touched upon the democratization of scientific access theoretically made possible by networks like the National Virtual Observatory or the Network for Earthquake Engineering Simulation.
Social scientists stress, however, that in changing the communication of science, electronic publishing, as opposed to
paper journals, means that it is now possible not only to publish but to "depublish"--to remove material from the Web with nary a trace. Information technology can be used to increase access to data--or to tighten access.
I've also alluded to the massive avalanche of data now pouring down on many disciplines. The capability to collect such quantities of data indeed changes the questions we can ask, making it possible to deal with higher orders of complexity. It also creates new needs for techniques like data mining and data integration.
A fascinating example of utilizing IT for "knowledge discovery" and for broader access to existing databases comes from geoscience. Given that a number of researchers believe that we could be in the throes of another mass extinction event on earth, understanding the geologic records of past events becomes more than academic.
A previous extinction event--perhaps the most extensive--occurred at the boundary of the Permian and Triassic periods some 250 million years ago, when some 90 percent of marine species and 70 percent of terrestrial vertebrate species went extinct.
Various explanations have been proposed, from meteorite impact to supernova event to volcanism. "Each model is dominated by a subset of disciplinary knowledge, and is able to provide only partial resolution for the cause of the extinction event," explains geoscientist Krishna Sinha of Virginia Tech.
Integration of existing knowledge from sub-disciplines is essential to assessing the cause of the event as well as its duration, he points out. Here, IT pulls together existing data in new ways to advance discovery and transform science.
Such studies also raise questions of how to train and reward workers in untraditional areas such as the "reuse" of data.
The growing collaborations within large interdisciplinary teams raise other issues. In a world where the reward structure is still oriented toward the individual, for example, how does one reward researchers in large teams? Sociologists and philosophers of science should be part of those teams, and could help them to articulate new ethical guidelines.
Sociologists of science also advocate creating guidelines for communication across disciplines. In interdisciplinary teams, the issue of whose expertise counts is very real--not all members may be treated equally. In addition, each discipline has its own conventions to govern authorship of articles, and for how to treat junior members of a team.
Other questions arise regarding the potential diffusion of ethical responsibilities in a large team. How do we ensure the integrity of data in a physics paper--one that describes, for example, an experiment at CERN or Fermilab--that could have four-to-five-hundred authors?
I will cite one more aspect of the changing conduct of science, and that is the increased public engagement and participation in science, fueled in no small part by information technology. The old model advocated the public understanding of science, which implies a one--way flow of knowledge to a seemingly passive public.
Now we think, instead, of engagement and exchange. Today, in Costa Rica, biodiversity surveys are being assisted by "parataxonomists" who carry hand-held computers into the field, which help them to identify species on the spot. At the Space Physics and Aeronomy Research Collaboratory website, visitors can view real-time data on the solar wind, the aurora, and the ionosphere--one of many such research sites.
Scientific engagement with the public may be the grandest challenge of all--and is sure to change the conduct of science in ways we can hardly imagine.
I'll end now by turning back to Sherlock Holmes for a bit of concluding wisdom. In a way he prefigured the new conduct of science when he said, "One's ideas must be as broad as Nature if they are to interpret Nature."
The new tools that enable us to join together in interdisciplinary teams, to tackle huge questions that span many fields of knowledge, are broadening our ideas about Nature, even as we are challenged by expanded societal expectations, and by the need to integrate social science perspectives in all of this work. I think the good detective would be amazed at the new ways we are able to see the world.
1 "New Imaging Tools Put the Art Back Into Science," by Andrew Lawler, Science, vol. 292, issue 5519, 11 May 2001; and "Origins of Aircraft-Damaging Clear-Air Turbulence during the 9 December 1992 Colorado Downslope Windstorm: Numerical Simulations and Comparison with Observations," Clark et al., Journal of the Atmospheric Sciences, vol. 57, 15 April, 2000.
Return to speech.
2 NSF social scientist Rachelle Hollander, pers. comm.
Return to speech.
3 American poet James Russell Lowell.
Return to speech.
4 "New Tools Reveal Treasures at Ocean Hot Spots," by David Malakoff, Science, vol. 304, issue 5674, 21 May 2004.
Return to speech.
5 "The Worldwide Telescope," by Alexander Szalay and Jim Gray, Science, vol. 293, issue 5537, 14 Sept. 2001.
Return to speech.
6 Quoted in IEEE Spectrum Online, "Downloading the Sky," by Jonathan C. McDowell, Aug. 4, 2004.
Return to speech.
7 This section draws upon insights from several NSF-supported social scientists: Vivian Weil, Center for the Study of Ethics in the Professions, Illinois Institute of Technology; Geoffrey Bowker, Center for Science, Technology and Society, Santa Clara University; Melissa Anderson, Department of Educational Policy and Administration, University of Minnesota; and Stephen Hilgartner, Department of Science and Technology Studies, Cornell University.
Return to speech.