SCIENCE / MEDICINE : Since Proof Is Elusive, Science Seeks Best Odds : Research: Theories are embraced through a process of consensus, and they hold up until a better one comes along.
How do scientists know when they have collected enough evidence to feel comfortable with their conclusions?
The word proof practically never appears in scientific literature because most scientists do not think it is appropriate beyond the narrow confines of the pure world of mathematics. At best, in many areas of great public concern, scientists can only offer odds and probabilities, not proof.
“The way science works is that you have a social process,” said James Woodward, philosopher of science at Caltech. “You have people who reach at least initially different judgments about the merits of theories, they argue and produce new evidence, and finally sometimes things get sorted out and you move toward consensus.
“There’s no oracle” on the mountaintop ready to proclaim the arrival of truth, he added.
“There is a lot of judgment involved,” said Peter Galison, who teaches physics and philosophy at Stanford University.
“But judgment is not necessarily something arbitrary,” he said. “Judgment is something you learn over the years.
“You might think a scientific experiment is objective (only) if a raving lunatic could look at it and tell what the experiment is saying. But experiments aren’t like that. They don’t yell an answer to you. There’s not a meter that goes over to the red.
“Judgment has a bad name, a bad reputation. Judgment makes it sound like it’s anybody’s call. But scientific judgment is much more important than that.”
It is, however, something that the untrained mind has trouble grasping. In Galison’s field of high-energy physics, the evidence consists of graphics and data that are meaningless to anyone who has not been trained in physics. So the world at large understands the inner workings of the atom not because the evidence is clear but because physicists have declared that’s the way it is.
That blind faith in the judgment of scientists is easier to accept in an arcane field such as high-energy physics because most people accept the fact that they will probably never really understand what is going on. If they are going to understand anything, they will simply have to believe what they are told.
The problem gets much murkier, however, in fields that people think they should be able to understand and in which they have a considerable stake.
Physics, Galison insisted, is “simple” compared to multidisciplinary fields in which there are many variables and great uncertainties.
“Look at something like weather,” he said. “People don’t appreciate what a difficult job meteorologists have. It’s an immensely complicated science.
“One of the reasons is it depends on phenomena that range from very tiny areas, such as what happens at the intersection of sea air and land masses, that can impact on huge weather systems that may span several hundred miles in diameter.
“It is that vulnerability to tiny changes that can lead to huge changes, global-sized phenomena, that makes meteorology an immensely difficult problem.”
Proof in such fields “becomes much more elusive,” Galison said. “It’s still a science where prediction is enormously difficult.”
But while people believe fields such as physics are extremely complicated, requiring great machines and complex experiments, the “weather is something we experience every day, and it does not seem to be an arcane subject,” Galison said.
“Medicine is the same way. We all live in our bodies, and we have a frustration that medicine can’t do more. It can’t cure the common cold.
“There’s this funny inversion that takes place. In the very arcane subjects like particle physics, people think if scientists can’t predict something it would be understandable. Whereas, for God’s sake, if science can do anything, why can’t it cure the common cold, predict the weather and tell me what the GNP and unemployment rate will be?
“But we really ought to look at it in the opposite way. It’s much more understandable why physics comes to a precise, quantitative prediction. It’s about the collisions of protons and antiprotons, whereas understanding the enormously interconnected processes that go into the body’s response to a cold or the complexity of weather or earthquakes” is far more elusive, he said.
There are many areas today that seem very uncertain, and people are being asked to make considerable sacrifices because scientists believe the failure to act now will probably be more costly down the road. One of the best examples is global warming.
It may well be, as many scientists have noted, that the evidence that will prove global warming is as catastrophic as some have warned will come too late. By the time scientists are able to document that the world is indeed getting warmer and the seas are indeed rising because of glacial melt, global warming will have advanced to the point where little can be done to head it off.
Yet, “to wait until the point at which definitive evidence is available may well be to wait until the effects are irreversible,” said Robert McGinn, vice chairman of Stanford University’s values, technology, science and society program.
McGinn suggested that the time for action will be “when there is a significant body of opinion that this (global warming) is not a spurious event.”
But how do scientists reach that point in a field as complex as meteorology?
The same way as in any other field of science, according to Caltech’s Woodward.
In science, “you have a variety of different alternative mechanisms that could have produced the evidence,” Woodward said. “What you try to do is proceed by eliminating all of the alternatives except one.”
In addition, “you try to find evidence that is so distinctive and so unique that there is only one plausible explanation for it.
“Typically, proof in science is a matter of relative competition between two or more theories in which you are really saying Theory A does better than its competitors and therefore I’m going to accept Theory A, at least for the time being,” Woodward said.
“But then someone may come along and introduce some possibility that hadn’t even occurred to you, and you have to start all over again.”
As UCLA chemist David Eisenberg noted, “Science is a self-correcting process.”
Another approach, said Stanford’s Galison, “is to try to see if the phenomenon is stable. You vary things in certain ways to see if the result changes. The search for stability and the search for directness are two ways for getting at the reality.”
But that is becoming increasingly difficult because many of the problems confronted by science today are global in scope, and major experiments may involve hundreds of scientists scattered all over the world.
That applies to research with the giant machines used for modern science, including the supercollider that the U.S. government wants to build in Texas, as well as efforts to determine the impact of deforestation over wide ranges of the planet, acid rain, atmospheric and oceanic pollution and various threats to public health.
“The single hardest problem facing modern scientific collaborations is it is no longer possible for any single individual to be in full command of all parts of the experiment,” Galison said. “It’s simply not humanly possible for somebody to be an expert on everying. You can’t know everything. So you have to rely on a team which depends on the performance of each person and each person must know enough about what the next person is doing so that you can piece it all together.”
That means, he added, “you can have errors that come in on one part of the experiment that are not recognized by the rest of the team. Or, more subtly, you can have everybody doing their job right but not making the interconnections properly.”
Yet while such a process will not lead to “the proof of a mathematical equation,” it can still lead to “something you could bet your bottom dollar on.”
“I think there is a certain stage,” Galison said, “where scientific demonstration gets to that point where the scientific community does come to a consensus and the information is secure.”
In such matters as global warming, he added, “we will reach a certain point when failure to act will be irresponsible. We will have to foot the bill.”
Galison believes there are bound to be many issues like global warming in the years ahead that no one has even anticipated, and society will be asked to act on the basis of very inconclusive evidence. He said that is not because science has changed as much as it is because humans “have managed to get better at interfering with the environment.”
“It’s not just a matter of better data collection,” he said. “It’s a matter of what they (scientists) are collecting the data about. They weren’t testing hydrogen bombs in the 19th Century. We can do more damage. We have done more damage.”
Ultimately, decisions on which course to follow in the years ahead may well have little to do with science, even if scientists have been asked to “prove” which course would be less hazardous.
“I think the issue really becomes ultimately a philosophical one,” Stanford’s McGinn said.
And it may be that the hard decisions will simply not be made.
“I think it probable that even when the answer is known (about global warming) there won’t be adequate cooperation,” Eisenberg said.
“Even if we can determine what to do, the organizational and political difficulties may be the formidable ones. It may be that everyone would rather sink into the mire together than build a structure that would allow us to get out.”
And to that must be added the uncertainty that is the constant companion of the working scientist.
Issues such as global warming are so complex, McGinn observed, that “I would be very surprised if all the practitioners in climatology ever reach agreement.”
So the most that public policy-makers should expect from the world of science is a consensus among the experts that the evidence truly points toward a conclusion that should, for now, be accepted--until someone with sound credentials and new evidence can show that the consensus was probably wrong.
