Science Philosophy and Practice: The Scientific Method
Science Philosophy and Practice: The Scientific Method
The scientific method is the approach used by scientists in the discovery of new scientific knowledge. A simplified outline of this approach, reduced to the making of observations, the formation of hypotheses (possible cause-and-effect explanations), and the testing of hypotheses by further observations, is often taught to students as the scientific method. Although there is in fact no single scientific method, no universal way of conducting scientific research, all scientists' approaches to knowledge discovery have certain elements in common that distinguish science from other ways of knowing—the ways of knowing characteristic of religious belief, common sense, personal relationships, and so on. It is generally agreed among scientists and philosophers that scientific claims must be capable of being falsified by other scientists, must fit into some framework of explanatory ideas (a theory), and must make meaningful predictions about the observable universe. The making of observations, the formation of cause-and-effect explanations, and the comparison of these hypotheses to further observations are basic to the creation of new scientific knowledge, though they are far from the whole story and are rarely practiced in a straightforward, mechanical way.
Historical Background and Scientific Foundations
Scientific ways of thinking have developed over a period of about 2,000 years, starting from forms of commonsense problem-solving used by people in everyday life. Today, scientific knowledge is a large, ever-growing system to which individual scientists can add only by following some form of the procedure known as the scientific method.
The Greek philosophers of several centuries BC, especially Aristotle (384–322 BC), were among the first persons to think carefully about how we acquire knowledge about the natural world—scientific knowledge. Aristotle taught that science depends on two basic forms of reasoning, induction and deduction.
Induction is the inference (reasoning-out) of general principles from specific observations. For example, if we observe without exception that heavy objects fall straight downward when released, we may reason inductively that all objects have some property in common—mass—on which some force associated with the Earth acts—gravity. (Other inferences are possible from this simple observation, and the Greeks did not, in fact, reason in terms of “forces.”) Further, careful observations of exactly how quickly objects fall may allow us to induce a strict, mathematical law describing how the force of gravity accelerates bodies. This was done by European physicists in the seventeenth century at the dawn of the Scientific Revolution.
Deduction, on the other hand, is the prediction of particular events or observations from general principles or laws: It is like induction working backwards. For example, once we have proposed a law of gravitation, we can deduce from it how a space probe should behave en route from Earth to Mars. If the probe behaves as predicted, the law is confirmed, at least thus far.
Both induction and deduction are part of the scientific method. From observations, laws may be produced using inductive reasoning. From these laws, predictions may be deduced. These predictions can be tested by arranging further observations. From these further observations, adjustments to the proposed scientific laws may be made.
A version of this scientific method was described by the Arab scientist Ibn al Haytham (965–1039) in the eleventh century. The English philosopher and Franciscan monk Roger Bacon (1219–1294) proposed a version of al Haytham's method that even more closely prefigured the modern ideal: observe phenomena, propose a hypothesis to explain what is observed, make fresh observations to test the hypothesis, and publish your work so that others can check it. In the 1600s, physicists (scientists who study the fundamental laws governing all physical objects), including Isaac Newton (1643–1727), proposed further standards for scientific thought. In the nineteenth century, several philosophers refined these standards into a series of ideal steps that became known as the scientific method. In the twentieth century, the older, simplistic view of science as merely turning the crank on the scientific method was challenged by many philosophers and historians.
The classic scientific method still taught commonly to high-school and college students today is as follows:
- Observe natural phenomena.
- Propose a possible cause-and-effect explanation, a hypothesis, to account for the observations.
- Use the hypothesis to predict phenomena not already observed.
- Arrange observations of the predicted phenomena.
- If the new observations do not agree with prediction, go back to step 2 and revise your hypothesis. Repeat these steps until your hypothesis accounts for all known observations.
Many philosophers, historians, and scientists have pointed out that in real life, science does not always follow the steps of the scientific method in an orderly way. Scientists often make intuitive guesses based on very slight observations; existing theories influence which observations are made out of the infinite number of observations that could be made, and anomalous observations that seem to conflict with an otherwise well-supported theory may be ignored or put aside for a time.
In the 1930s, American philosopher Karl Popper (1902–1994) proposed that the distinguishing mark of a truly scientific idea is that it is falsifiable—that is, it makes predictions that can be tested against observation (step 4 of the classic scientific method). In recent decades, other philosophers have pointed out that Popper's definition is inadequate: Simply making a falsifiable prediction is not the same thing as doing science. For example, a person might claim, on a whim, that cars run on air. This claim might be falsified by observing that cars need to be supplied with liquid fuel or electricity in order to run. However, even though it is falsifiable, the cars-run-on-air hypothesis is not scientific: It is merely a free-floating claim that does not follow from specific observations or from any coherent cause-and-effect theory. All scientific claims are falsifiable, but not all falsifiable claims are scientific.
Explanations based on magic, gods, miracles, or other supernatural causes are never scientific, because if supernatural forces do exist, they might, in principle, cause anything at all to happen. Such forces are potentially compatible with all possible observations and so their existence cannot ultimately be tested. Science studies only natural explanations because they are the only explanations that can be ruled in or out by the scientific method.
Modern Cultural Connections
Not all scientists practice the scientific method in exactly the same way. For example, scholars often distinguish between the historical sciences and experimental sciences. Historical sciences, such as astronomy, geology, evolutionary biology, and astrophysics, seek to explain chains of events that occurred in the past. Experimental sciences, such as physics, test their hypotheses in controlled settings (e.g., laboratories). Both types of science are completely scientific and operate according to the basic principles of scientific method. The difference is that historical scientists usually predict new observations of naturally occurring evidence or events, rather than arranging experiments.
IN CONTEXT: OBSERVATION VS. THEORY
In the late 1990s, some scientists observed an effect that was dubbed the Pioneer anomaly. The Pioneer 10 and Pioneer 11 space probes, launched by the United States in 1972 and 1973, respectively, to explore the outer solar system (they still sail through space) were observed to be traveling along paths that could not be exactly explained using existing gravitational laws. Scientists studied every cause for the Pioneer anomaly that they could think of—fuel leaks acting like weak rockets, resistance from dust and gas floating in space, measurement errors due to waves breaking on the shore miles from radar stations tracking the spacecraft, and many more—but could not account for the anomaly. The same effect has been observed, although with a wider margin for error, for two other probes flying through the outer solar system, Ulysses and Galileo.
Since no force could be identified that might cause the anomalous movement of these spacecraft, it appeared that these observations might contradict the prevailing theory of gravitation, general relativity. However, general relativity has been supported by many other observations, and no scientist has been willing to reject it based only on the Pioneer anomaly. This illustrates the principle that a few anomalous observations are not necessarily enough to overturn a well-supported scientific theory.
As of 2008, the Pioneer anomaly remained unexplained, while astronomers continued to make many other observations that confirmed predictions of general relativity.
Confusion sometimes arises when people assume that making observations means performing experiments—that is, manipulating objects and forces in a laboratory to produce a certain outcome. For example, people who disbelieve the biological theory of evolution (in science, a “theory” is not a guess but any well-supported explanation for a body of facts) sometimes argue that because nobody was present to observe the evolution of life, and because evolution cannot be repeated in a laboratory, scientists' claims about evolution are a matter of faith, not science.
However, this objection is based on a misunderstanding of the scientific method. Observations do not need to be based on laboratory experiments to support or contradict a scientific theory. For example, in the twentieth century, fossils gathered by paleontologists (scientists who study fossils) showed that the earliest known four-limbed land-dwelling animals appeared in the Late Devonian, about 360 million years ago. This corresponds to step 1 of the scientific method: Observation. Evolutionary theory predicts that such animals evolved from four-limbed fish living in shallow coastal waters just before the appearance of the first definitely land-dwelling animals, and that transitional (or, in-between) animals must have developed at that time. This corresponds to step 2: Create a Hypothesis. In the early 2000s, several paleontologists, reading in a geology textbook that shallow-shore rocks dating to about 375 million years ago are found on Ellesmere Island in northern Canada, reasoned that fossils of transitional animals should be found there. This corresponds to step 3: Make a Prediction Based on Hypothesis. These scientists arranged an expedition to Ellesmere Island—step 4: Arrange New Observations—and in 2004 discovered fossils of just such a transitional animal, now known as Tiktaalik. The discovery of Tiktaalik fossils was an observation that confirmed the hypothesis of evolution, as many other observations have done over the last century and a half.
Primary Source Connection
Two millennia of scientific discovery are summarized and ranked into a “best of” list in this article by New York Times reporter George Johnson. Johnson is also the author of Miss Leavitt's Stars and received a Templeton-Cambridge Journalism Fellowship in Science and Religion in 2005.
HERE THEY ARE, SCIENCE'S 10 MOST BEAUTIFUL EXPERIMENTS
Whether they are blasting apart subatomic particles in accelerators, sequencing the genome or analyzing the wobble of a distant star, the experiments that grab the world's attention often cost millions of dollars to execute and produce torrents of data to be processed over months by supercomputers. Some research groups have grown to the size of small companies.
But ultimately science comes down to the individual mind grappling with something mysterious. When Robert P. Crease, a member of the philosophy department at the State University of New York at Stony Brook and the historian at Brookhaven National Laboratory, recently asked physicists to nominate the most beautiful experiment of all time, the 10 winners were largely solo performances, involving at most a few assistants. Most of the experiments—which are listed in this month's Physics World—took place on tabletops and none required more computational power than that of a slide rule or calculator.
What they have in common is that they epitomize the elusive quality scientists call beauty. This is beauty in the classical sense: the logical simplicity of the apparatus, like the logical simplicity of the analysis, seems as inevitable and pure as the lines of a Greek monument. Confusion and ambiguity are momentarily swept aside, and something new about nature becomes clear.
The list in Physics World was ranked according to popularity, first place going to an experiment that vividly demonstrated the quantum nature of the physical world. But science is a cumulative enterprise—that is part of its beauty. Rearranged chronologically and annotated below, the winners provide a bird's-eye view of more than 2,000 years of discovery.
Eratosthenes' measurement of the Earth's circumference
At noon on the summer solstice in the Egyptian town now called Aswan, the sun hovers straight overhead: objects cast no shadow and sunlight falls directly down a deep well. When he read this fact, Eratosthenes, the librarian at Alexandria in the third century BC, realized he had the information he needed to estimate the circumference of the planet. On the same day and time, he measured shadows in Alexandria, finding that the solar rays there had a bit of a slant, deviating from the vertical by about seven degrees.
The rest was just geometry. Assuming the earth is spherical, its circumference spans 360 degrees. So if the two cities are seven degrees apart, that would constitute seven-360ths of the full circle—about one-fiftieth. Estimating from travel time that the towns were 5,000 “stadia” apart, Eratosthenes concluded that the earth must be 50 times that size—250,000 stadia in girth. Scholars differ over the length of a Greek stadium, so it is impossible to know just how accurate he was. But by some reckonings, he was off by only about 5 percent. (Ranking: 7)
Galileo's experiment on falling objects
In the late 1500's, everyone knew that heavy objects fall faster than lighter ones. After all, Aristotle had said so. That an ancient Greek scholar still held such sway was a sign of how far science had declined during the dark ages.
Galileo Galilei, who held a chair in mathematics at the University of Pisa, was impudent enough to question the common knowledge. The story has become part of the folklore of science: he is reputed to have dropped two different weights from the town's Leaning Tower showing that they landed at the same time. His challenges to Aristotle may have cost Galileo his job, but he had demonstrated the importance of taking nature, not human authority, as the final arbiter in matters of science. (Ranking: 2)
Galileo's experiments with rolling balls down inclined planes
Galileo continued to refine his ideas about objects in motion. He took a board 12 cubits long and half a cubit wide (about 20 feet by 10 inches) and cut a groove, as straight and smooth as possible, down the center. He inclined the plane and rolled brass balls down it, timing their descent with a water clock—a large vessel that emptied through a thin tube into a glass. After each run he would weigh the water that had flowed out—his measurement of elapsed time—and compare it with the distance the ball had traveled.
Aristotle would have predicted that the velocity of a rolling ball was constant: double its time in transit and you would double the distance it traversed. Galileo was able to show that the distance is actually proportional to the square of the time: Double it and the ball would go four times as far. The reason is that it is being constantly accelerated by gravity. (Ranking: 8)
Newton's decomposition of sunlight with a prism
Isaac Newton was born the year Galileo died. He graduated from Trinity College, Cambridge, in 1665, then holed up at home for a couple of years waiting out the plague. He had no trouble keeping himself occupied.
The common wisdom held that white light is the purest form (Aristotle again) and that colored light must therefore have been altered somehow. To test this hypothesis, Newton shined a beam of sunlight through a glass prism and showed that it decomposed into a spectrum cast on the wall. People already knew about rainbows, of course, but they were considered to be little more than pretty aberrations. Actually, Newton concluded, it was these colors—red, orange, yellow, green, blue, indigo, violet and the gradations in between—that were fundamental. What seemed simple on the surface, a beam of white light, was, if one looked deeper, beautifully complex. (Ranking: 4)
Cavendish's torsion-bar experiment
Another of Newton's contributions was his theory of gravity, which holds that the strength of attraction between two objects increases with the square of their masses and decreases with the square of the distance between them. But how strong is gravity in the first place?
In the late 1700's an English scientist, Henry Cavendish, decided to find out. He took a six-foot wooden rod and attached small metal spheres to each end, like a dumbbell, then suspended it from a wire. Two 350-pound lead spheres placed nearby exerted just enough gravitational force to tug at the smaller balls, causing the dumbbell to move and the wire to twist. By mounting finely etched pieces of ivory on the end of each arm and in the sides of the case, he could measure the subtle displacement. To guard against the influence of air currents, the apparatus (called a torsion balance) was enclosed in a room and observed with telescopes mounted on each side.
The result was a remarkably accurate estimate of a parameter called the gravitational constant, and from that Cavendish was able to calculate the density and mass of the earth. Erastothenes had measured how far around the planet was. Cavendish had weighed it: 6.0 × 1024 kilograms, or about 13 trillion trillion pounds. (Ranking: 6)
Young's light-interference experiment
Newton wasn't always right. Through various arguments, he had moved the scientific mainstream toward the conviction that light consists exclusively of particles rather than waves. In 1803, Thomas Young, an English physician and physicist, put the idea to a test. He cut a hole in a window shutter, covered it with a thick piece of paper punctured with a tiny pinhole and used a mirror to divert the thin beam that came shining through. Then he took “a slip of a card, about one-thirtieth of an inch in breadth” and held it edgewise in the path of the beam, dividing it in two. The result was a shadow of alternating light and dark bands—a phenomenon that could be explained if the two beams were interacting like waves.
Bright bands appeared where two crests overlapped, reinforcing each other; dark bands marked where a crest lined up with a trough, neutralizing each other.
The demonstration was often repeated over the years using a card with two holes to divide the beam. These so-called double-slit experiments became the standard for determining wavelike motion—a fact that was to become especially important a century later when quantum theory began. (Ranking: 5)
Last year when scientists mounted a pendulum above the South Pole and watched it swing, they were replicating a celebrated demonstration performed in Paris in 1851. Using a steel wire 220 feet long, the French scientist Jean-Bernard-Léon Foucault suspended a 62-pound iron ball from the dome of the Panthéon and set it in motion, rocking back and forth. To mark its progress he attached a stylus to the ball and placed a ring of damp sand on the floor below.
The audience watched in awe as the pendulum inexplicably appeared to rotate, leaving a slightly different trace with each swing. Actually it was the floor of the Panthéon that was slowly moving, and Foucault had shown, more convincingly than ever, that the earth revolves on its axis. At the latitude of Paris, the pendulum's path would complete a full clockwise rotation every 30 hours; on the Southern Hemisphere it would rotate counterclockwise, and on the Equator it wouldn't revolve at all. At the South Pole, as the modern-day scientists confirmed, the period of rotation is 24 hours. (Ranking: 10)
Millikan's oil-drop experiment
Since ancient times, scientists had studied electricity—an intangible essence that came from the sky as lightning or could be produced simply by running a brush through your hair. In 1897 (in an experiment that could easily have made this list) the British physicist J. J. Thomson had established that electricity consisted of negatively charged particles—electrons. It was left to the American scientist Robert Millikan, in 1909, to measure their charge.
Using a perfume atomizer, he sprayed tiny drops of oil into a transparent chamber. At the top and bottom were metal plates hooked to a battery, making one positive and the other negative. Since each droplet picked up a slight charge of static electricity as it traveled through the air, the speed of its descent could be controlled by altering the voltage on the plates. (When this electrical force matched the force of gravity, a droplet—“like a brilliant star on a black background”—would hover in midair.) Millikan observed one drop after another, varying the voltage and noting the effect. After many repetitions he concluded that charge could only assume certain fixed values. The smallest of these portions was none other than the charge of a single electron. (Ranking: 3)
Rutherford's discovery of the nucleus
When Ernest Rutherford was experimenting with radioactivity at the University of Manchester in 1911, atoms were generally believed to consist of large mushy blobs of positive electrical charge with electrons embedded inside—the “plum pudding” model. But when he and his assistants fired tiny positively charged projectiles, called alpha particles, at a thin foil of gold, they were surprised that a tiny percentage of them came bouncing back. It was as though bullets had ricocheted off Jell-O.
Rutherford calculated that actually atoms were not so mushy after all. Most of the mass must be concentrated in a tiny core, now called the nucleus, with the electrons hovering around it. With amendments from quantum theory, this image of the atom persists today. (Ranking: 9)
Young's double-slit experiment applied to the interference of single electrons
Neither Newton nor Young was quite right about the nature of light. Though it is not simply made of particles, neither can it be described purely as a wave. In the first five years of the 20th century, Max Planck and then Albert Einstein showed, respectively, that light is emitted and absorbed in packets—called photons. But other experiments continued to verify that light is also wavelike.
It took quantum theory, developed over the next few decades, to reconcile how both ideas could be true: photons and other subatomic particles—electrons, protons, and so forth—exhibit two complementary qualities; they are, as one physicist put it, “wavicles.”
To explain the idea, to others and themselves, physicists often used a thought experiment, in which Young's double-slit demonstration is repeated with a beam of electrons instead of light. Obeying the laws of quantum mechanics, the stream of particles would split in two, and the smaller streams would interfere with each other, leaving the same kind of light- and dark-striped pattern as was cast by light. Particles would act like waves.
According to an accompanying article in Physics Today, by the magazine's editor, Peter Rodgers, it wasn't until 1961 that someone (Claus Jönsson of Tübingen) carried out the experiment in the real world.
By that time no one was really surprised by the outcome, and the report, like most, was absorbed anonymously into science. (Ranking: 1)
Correction: September 27, 2002, Friday. An article in Science Times on Tuesday about the experiments selected by physicists as the 10 most beautiful in history referred incorrectly at one point to the magazine edited by Peter Rodgers, which first printed the list. It is Physics World, not Physics Today.
Correction: October 7, 2002, Monday. An article in Science Times on Sept. 24 about physicists' selections of the 10 most beautiful experiments misstated a portion of Newton's theory of gravity, cited in a discussion of Cavendish's torsion-bar experiment. Newton held that the strength of attraction between two objects increases with the product of their masses, not with the square of their masses.
johnson, george. “here they are, science's 10 most beautiful experiments.” new york times (september 24, 2002).
Gauch, Hugh G., Jr. Scientific Method in Practice. Cambridge, UK: Cambridge University Press, 2002.
Cleland, Carol E. “Methodological and Epistemic Differences Between Historical Science and Experimental Science.” Philosophy of Science 69 (2002):474–495.
Johnson, George. “Here They Are, Science's 10 Most Beautiful Experiments.” New York Times (September 24, 2002).
Wolfs, Frank. University of Rochester (New York). “Introduction to the Scientific Method.” http://teacher.pas.rochester.edu/phy_labs/AppendixE/AppendixE.html (accessed February 6, 2008).