Matters of Gravity
Gravity's Shadow: The Search for Gravitational Waves. Harry
Collins. xxiv + 870 pp. University of Chicago Press, 2004. Cloth,
$100; paper, $39.
A reader who has avidly followed the popular science of the last
decade could be forgiven for forgetting that physics is an
experimental science. Strings, evaporating black holes, higher
dimensions, supersymmetry and quantum gravity are all beautiful
ideas, but we should never forget that so far they are only ideas.
They make for great reading, but I sometimes wonder how many of the
books and papers I have devoured in my efforts to keep up with the
latest trends in particle physics will be worth reading 10 years hence.
The whole point of science—and the feature to which it owes
all its success—is that no idea is to be believed until it has
been rigorously tested by experiment. Thus it is worrying that these
days many theorists spend their whole careers on ideas that are
But if all is not well with theory, at least experiment must be
healthy. After all, experiment is a simple thing: Just as in high
school, you follow the procedure, and you either get the result
predicted by the theory you are testing or you don't. Given enough
resources, how could this approach fail? And, although you wouldn't
know it from the popular literature, hundreds of millions of dollars
a year are spent on experimental physics—much more than is
spent on theory.
But despite the several billions invested in particle accelerators
and detectors, there have been few truly major experimental
discoveries in fundamental physics in the past 20 years. The fields
that have continued to amaze are astronomy and cosmology, which are
obviously healthy. But the only major addition to our knowledge of
the elementary particles these past two decades was the discovery
that neutrinos have mass. The list of new particles or effects that
have been looked for—and so far not found—is longer: the
Higgs particle, supersymmetric particles, dark-matter particles,
proton decay, the fifth force, evidence of extra dimensions.
It is, then, very timely that Harry Collins has written a
first-class study of how contemporary experimental physics operates.
Collins is a distinguished sociologist, and in Gravity's
Shadow he demonstrates why it is important to go beyond
superficial characterizations of science to study how groups of
scientists actually work together and make decisions. Collins has
taken as his subject the search for gravitational radiation.
Gravitational waves are ripples in the geometry of spacetime,
analogous to electromagnetic waves. Just as moving charges and
magnets produce light waves, masses when they accelerate inevitably
produce waves in the gravitational field—a field that, as
Einstein discovered in working out his theory of general relativity,
is exactly the same as the geometry of space and time.
The existence of gravitational waves (unlike that of strings or
extra dimensions) is undoubted by experts. The waves are a
consequence of the basic physical principles that go into Einstein's
theory of relativity—principles that have been precisely
tested in numerous experiments. Moreover, the effects of these waves
have been seen: Two stars orbiting each other radiate gravitational
waves. As a result, they lose energy and spiral in toward each
other. For most binary star systems this decay happens very slowly,
but for neutron stars, which are extremely compact and dense, and
can orbit very close to each other at close to the speed of light,
the effect is significant. Moreover, the radio pulses astronomers
detect from some spinning neutron stars can be used to track the
decay of their orbits. A major triumph of the general theory of
relativity is that its predictions of how fast orbiting neutron
stars lose energy by radiating it away in gravitational waves have
been confirmed to an accuracy of many decimal places.
Of course it would be much better to observe the gravitational waves
directly. The challenge is that they are extremely weak. To make
waves large enough to be detectable with any instrument we can
imagine building on Earth, the most violent events in the universe
are required: supernova explosions, the formation of black holes or
the collisions of stars. Even so, the effects are tiny: The geometry
changes so little that a distance of several kilometers changes by
less than the diameter of a proton. To detect such changes requires
enormous machines that bounce light back and forth over paths many
kilometers long. Efforts to build these gravitational-wave
interferometers are currently under way in the United States, Italy,
Germany, France, the United Kingdom, Australia,
Japan and other countries, whose governments have invested so far
more than $300 million in the search to detect gravitational waves.
But despite the enormous effort and expense, no waves have been seen.
Harry Collins's book offers an opportunity to understand whether
experimental physics is really as simple as we were taught in high
school. Naively, we talk about using a particle detector to
"see" a particle, or the Wilson Microwave Anisotropy Probe
to "see" the cosmological microwave background. But these
are complex instruments, built and run by large groups of people.
What is "seeing" when the image we are finally shown on
the front page of a newspaper is the output of a computer program
that is itself the outcome of a decades-long process involving large
numbers of people and machines? This is the central question Collins
is concerned with.
Like particle accelerators and dark-matter detectors,
gravitational-wave observatories require collaborations of hundreds
of scientists, working in many different universities spread across
the globe. But even with huge investments and large teams of
scientists, the experiments are hard to do. The signals being looked
for are rare and must be distinguished from millions of spurious
events. Often you cannot tell just from the raw data whether
anything has been seen or not. Very likely the signature you are
seeking will be, just by chance, mimicked by noise in your
instrument. How do you decide when you have registered a rare signal
often enough to be sure that you are seeing something besides the
random effects of noise in your detector?
When the difficulty of the experiments is combined with the social
dynamics of large teams, things get very far from high school
science. To take just one question, raised by the historian of
science Peter Galison: How does a large collaboration decide when an
experiment is finished? Do they vote? That seems hardly good
enough—scientific truth is not a matter of the view of the
majority. But if consensus is sought, what about the few cautious
holdouts who will always insist on waiting for more data before
announcing a result? And should we worry that the group may be more
likely to embrace an analysis of the data that agrees with what the
theorists want than one that disagrees?
Indeed, what happens when a good scientist believes that he or she
has made a discovery but fails to convince colleagues? The shadow in
Gravity's Shadow is just such a story.
The pioneer of gravitational-wave science was an American
experimentalist named Joe Weber, who announced in 1969 that he had
detected gravitational waves—not with enormous machines, but
with furniture-sized devices that he and a few students had built
and operated in their lab. The problem was, when other physicists
tried to replicate Weber's results, most of them couldn't. Even more
unhappily, when they analyzed his experiments, they concluded that
his electronics were too noisy to have seen what he claimed. The
outcome, sensitively related by Collins in what is the emotional
core of the book, was tragic. Had Weber been willing to accept the
judgments of his colleagues, he would have ended his life loved and
admired as the pioneering founder of a field of science. Instead,
his insistence that he was the only one who could do the experiment
right forced him to embrace a series of more and more unlikely and
implausible hypotheses, which alienated him from those who had
followed in his path.
Collins's book is long and full of detail, but it is an entirely
rewarding read. Just like the physicists he studies, he has made the
search for gravitational radiation his life's work. For three
decades, he embedded himself in the community of gravitational-wave
researchers, interviewing all the key scientists, visiting their
labs, and attending many conferences, workshops and meetings.
Funding agencies and scientists opened their files and archives to
him. The resulting narrative is as provocative as it is convincing.
There is a lot written by philosophers and others about how science
is supposed to work. But this is the one of the very few books I've
read that tries to help the reader understand what really goes on
these days in the world of big science.
Not surprisingly, the answers to the question of how science works
are not simple. One response is that scientists apply well-tested
methods, which necessarily lead to an increase in truth; politics
and sociology play minimal roles, for the processes by which
scientists arrive at truth are foolproof enough to mitigate against
the complexities of other human endeavors. At the other end of the
spectrum is the relativist's belief that contemporary science is
nothing but academic politics—to be comprehended as a
sociological phenomenon that no more approaches truth than do the
activities of other organized communities. In the early 1990s there
was a very public argument between these two poles, which was dubbed
the Science Wars.
Collins's book is an overdue antidote to both kinds of
naiveté. He dubs himself a "methodological
relativist," but he insists that this means only that as an
observer of the processes by which scientists make decisions and
come to agreement, he cannot himself try to reach his own views as
to what is true or false in the science. The aim of his study is not
to decide whether Weber or his critics were right. Rather, it is to
understand the processes by which a community of scientists came,
uncomfortably but almost unanimously, to the conclusion that its
founding member was wrong.
The second half of Gravity's Shadow is about how the search
for gravitational waves was taken up after Weber. With the consensus
that Weber fooled himself came the realization that it would take
huge machines, costing hundreds of millions of dollars, to actually
see gravitational waves. Thus the search for these waves, which
started as small science in one professor's laboratory, ends up as a
paradigmatic example of big science. The present American detectors,
which are part of the Laser Interferometer Gravitational-Wave
Observatory (LIGO) project, have been built and are run by a process
that seems more akin, in organizational style, to a big industrial
project than to our romantic notions of what happens in an
individual scientist's laboratory.
As Collins tells the story of how science by lone seeker became
science by international committee, he provides us with the material
to think hard about the meaning and cost of the transition. And
these are questions we should be asking. No one believes that a
large advertising agency or a Hollywood film studio can easily
produce art with the lasting value that a single painter may achieve
in his or her studio, even if people of equal talent are employed.
Why do we believe that organizing scientists into groups and teams
run by professional managers will easily result in discoveries of
the same import as used to be made by heroic pioneers such as
Galileo and Faraday?
Indeed, the story Collins tells is very much about the conflict
between science as done by individuals and science as done by modern
large-scale organizations. And unfortunately, Weber is not the only
tragic figure in the drama. Some of the key ideas and techniques
that make the present detectors possible were invented by the
Scottish physicist Ron Drever. But halfway through the development
of LIGO, he was forced out of the collaboration because he refused
to stop voicing his objections to the decisions of the project
directors, who necessarily spent more time with schedules, budgets
and reports than they could with equipment in the lab. Drever was
convinced that the right path was to build incrementally, through
the careful testing of larger and larger prototypes. Other
scientists disagreed and insisted that the need to keep the project
on time and within budget required them to freeze the design, even
if everyone involved knew it was suboptimal.
Perhaps the greatest achievement of Collins's book is that he tells
this part of the story in a way that makes the reader sympathize
with all participants. This balance is the fruit of his many years
of interaction with the scientists: He understands that the issue is
not good and bad individuals but an inevitable conflict between
different modes of organizing creative work.
One of the most fascinating parts of the story is what has happened
since the conflict was resolved. After spending $300 million, LIGO
is on schedule and is starting to report data. But the version
presently being run, the initial LIGO, is not expected to be
sensitive enough to ensure that it will detect gravitational waves.
Given presently accepted estimates of the strengths and frequencies
of events that cause the waves, there is a reasonable chance that it
will see nothing.
There is an upgrade called Advanced LIGO that could make detection
of gravitational waves more likely. It is not yet funded, and we can
only hope that the U.S. Congress will have the vision to support it.
The reader is bound to ask, however, How was a decision reached to
spend hundreds of millions of dollars of the public's money to build
an experiment when it was understood that it might well see nothing?
Part of the answer is that the scientists involved were able to
assure the National Science Foundation (NSF) that the upgrade to
Advanced LIGO, which was to be based on technologies then under
development, would be able to increase sensitivity to the point that
detection of gravitational waves would become likely. But the
decision to go ahead before these technologies were ready and before
Advanced LIGO had been funded certainly contained elements of risk
both for the scientists and for the NSF.
Collins makes it clear that it is not his job to second-guess the
science or the decisions of policymakers. We will know in a few
years whether the risk paid off. But it is Collins's job to show us
how scientists and decision makers invent the contexts, intellectual
and organizational, within which decisions about science get made,
and Gravity's Shadow does this better than any book I have
As Collins tells it, only part of the answer lies in the need to
keep to schedules and budget. At least as influential was that the
scientists leading the project feared repeating the embarrassment of
Weber's false discovery. Weber's shadow thus appears to darken the
whole enterprise. If you build too ambitiously, your experiment may
not work, or you may not understand the data well enough to know
whether you have seen anything. Better to be very careful and see
nothing, but know for certain that you see nothing, than to risk
announcing a false discovery. When one takes into account these
kinds of risks, it becomes understandable how, within the LIGO
collaboration, it came to seem more rational to first build a
machine that had a low chance of achieving success.
Of course, even if the initial LIGO finds nothing, some science will
have been done. One of the arguments made for the huge investment in
LIGO, when there was already indirect evidence that gravitational
waves exist, is that it might discover something completely
unexpected: Perhaps the dark matter produces huge bursts of
gravitational radiation, or perhaps there are unknown processes that
produce more energy in gravity waves than light. It was argued that
every time a new technology has opened an observational window,
something surprising has been seen.
In the end, this is a book that raises important questions about big
science without pretending to answer them. Is the present system
really the best way that funds could be allocated and priorities set
in science? There is certainly room to wonder whether, if it is not
okay to invest heavily in theories that make no experimental
predictions, there might also be something troubling about huge
investments in experiments that are not sufficiently sensitive to
ensure that they can discover anything. On the other hand, perhaps
the path that the LIGO scientists chose is the only way the project
could have been run. After all, if science cannot progress without
risks being taken, then big scientific collaborations must be able
to take big risks as well.
What is certain is that we cannot address these kinds of questions
without a detailed understanding of how decisions are actually made
within the large communities that make up contemporary science as
well as within the government agencies that support scientists.
Harry Collins has shown us in this book that it is possible to do
the kind of careful, sensitive scholarship that makes mature
reflection on such questions possible. This is a book that everyone
who cares about the future of science should read.