"... in science there is no 'knowledge', in the sense in which Plato and Aristotle understood the word, in the sense which implies finality; in science, we never have sufficient reason for the belief that we have attained the truth. ... This view means, furthermore, that we have no proofs in science (excepting, of course, pure mathematics and logic). In the empirical sciences, which alone can furnish us with information about the world we live in, proofs do not occur, if we mean by 'proof' an argument which establishes once and for ever the truth of a theory."
Sir Karl Popper, The Problem of Induction, 1953
"If you thought that science was certain — well, that is just an error on your part."
Richard Feynman (1918-1988).
"A religious creed differs from a scientific theory in claiming to embody eternal and absolutely certain truth, whereas science is always tentative, expecting that modification in its present theories will sooner or later be found necessary, and aware that its method is one which is logically incapable of arriving at a complete and final demonstration."
Bertrand Russell, Grounds of Conflict, Religion and Science, 1953.
"It is the aim of science to establish general rules which determine the reciprocal connection of objects and events in time and space. For these rules, or laws of nature, absolutely general validity is required — not proven."
Albert Einstein, in Science, Philosophy and Religion, A Symposium, 1941.
hat is meant by scientific evidence and scientific proof? In truth, science can never establish 'truth' or 'fact' in the sense that a scientific statement can be made that is formally beyond question. All scientific statements and concepts are open to re-evaluation as new data is acquired and novel technologies emerge. Proof, then, is solely the realm of logic and mathematics (and whiskey). That said, we often hear 'proof' mentioned in a scientific context, and there is a sense in which it denotes "strongly supported by scientific means". Even though one may hear 'proof' used like this, it is a careless and inaccurate handling of the term. Consequently, except in reference to mathematics, this is the last time you will read the terms 'proof' or 'prove' in this article.
Though science formally cannot establish absolute truth, it can provide overwhelming evidence in favor of certain ideas. Usually these ideas are quite unobvious, and often they clash with common sense. Common sense tells us that the earth is flat, that the Sun truly rises and sets, that the surface of the Earth is not spinning at over 1000 miles per hour, that bowling balls fall faster than marbles, that particles don't curve around corners like waves around a floating dock, that the continents don't move, and that objects heavier-than-air can't have sustained flight unless they can flap wings. However, science has been used to demonstrate that all these common sense ideas are wrong.
The primary function of science is to demonstrate the existence of phenomena that cannot be observed directly. Science is not needed to show us things we can see with our own eyes. Direct observation is not only unnecessary in science; direct observation is in fact usually impossible for the things that really matter. In fact, the most important discoveries of science have only be inferred via indirect observation. Familiar examples of unobservable scientific discoveries are atoms, electrons, viruses, bacteria, germs, radio-waves, X-rays, ultraviolet light, energy, entropy, enthalpy, solar fusion, genes, protein enzymes, and the DNA double-helix. The round earth was not observed directly by humans until 1961, yet this counterintuitive concept had been considered a scientific fact for over 2000 years. The Copernican hypothesis that the earth orbits the sun has been acknowledged virtually ever since the time of Galileo, even though no one has ever observed the process to this day. All of these "invisible" phenomena were elucidated using the scientific method of inference. When the term "evidence" is used in this article, it is used strictly with respect to this scientific method.
What exactly is the scientific method? This is a complex and contentious question, and the field of inquiry known as the "philosophy of science" is committed to illuminating the nature of the scientific method. Probably the most influential philosopher of science of the 20th century was Sir Karl Popper. Other notables are Thomas Kuhn, Imre Lakatos, Paul Feyerabend, Paul Kitcher, A. F. Chalmers, Wesley Salmon and Bas C. van Fraassen. This is not the place to delve into an explication of the various philosophies represented by these scholars. For more information I refer you to their works and to the discussion presented by John Wilkins in his Evolution and Philosophy FAQ. Personally, I take a Bayesian view of the scientific method in principle (Jaynes 2003; Salmon 1990), and a Likelihoodist stance on evidence in practice (Burnham and Anderson 2002; Edwards 1972; Royall 1997), and these views will come through in how I present the evidence for common descent.
Now, to answer the question "What is the scientific method?" - very simply (and somewhat naively), the scientific method is a program for research which comprises four main steps. In practice these steps follow more of a logical order than a chronological one:
Because scientists are constantly making new observations and testing via those observations, the four "steps" are actually practiced concurrently. New observations, even if they were not predicted, should be explicable retrospectively by the hypothesis. New information, especially details of some process previously not understood, can impose new limits on the original hypothesis. Therefore, new information, in combination with an old hypothesis, frequently leads to novel predictions that can be tested further.
Examination of the scientific method reveals that science involves much more than naive empiricism. Research that only involves simple observation, repetition, and measurement is not sufficient to count as science. These three techniques are merely part of the process of making observations (#1 in the steps outlined above). Astrologers, wiccans, alchemists, and shamans all observe, repeat, and measure — but they do not practice science. Clearly, what distinguishes science is the way in which observations are interpreted, tested, and used.
The defining characteristic of science is the concept of the testable hypothesis. A testable hypothesis must make predictions that can be validated by independent observers. By "testable", we mean the predictions must include examples of what is likely be observed if the hypothesis is true and of what is unlikely to be observed if the hypothesis is true. A hypothesis that can explain all possible data equally well is not testable, nor is it scientific. A good scientific hypothesis must rule out some conceivable possibilities, at least in principle. Furthermore, a scientific explanation must make risky predictions — the predictions should be necessary if the theory is correct, and few other theories should make the same necessary predictions. These scientific requirements are the essence of Popperian falsifiability and corroboration.
For instance, the solipsistic hypothesis that the entire universe is actually an elaborate figment of your imagination is not a scientific hypothesis. Solipsism makes no specific or risky predictions, it simply predicts that things will be "as they are". No possible observations could conflict with solipsism, since all observations always may be explained away as simply another detailed creation of your imagination. Many other extreme examples can be thought of, such as the hypothesis that the universe suddenly came into existence in toto five minutes ago, with even our memories of "earlier" events intact. In general, creationist and "intelligent design" conjectures fail scientifically for these same reasons. Both can easily explain all possible biological observations, and neither one makes risky, specific predictions.
In contrast, Newton's scientific theory of universal gravitation makes specific predictions about what should be observed. Newton's theory predicts that the force between two masses should be inversely proportional to the square of the distance between them (otherwise known as the "inverse square law"). In principle, we could take measurements which indicated that the force is actually inversely proportional to the cube of the distance. Such an observation would be inconsistent with the predictions of Newton's universal theory of gravitation, and thus this theory is testable. Many anti-evolutionists, such as the "scientific" creationists, are especially fond of Karl Popper and his falsifiability criterion. These cynics are well known for claiming that evolutionary theory is unscientific because it cannot be falsified. In this article, these accusations are met head on. Each of the evidences given for common descent contains a section providing examples of potential falsifications, i.e. examples of observations that would be highly unlikely if the theory is correct.
"Testability" is not an either-or concept; some hypotheses are more testable than others. Contrary to some anti-evolutionist claims, not all hypotheses are equally valid scientific "interpretations" of the evidence. Some hypotheses are more successful in terms of the scientific method. Based on the scientific method, valid and useful hypotheses explain the observed facts simply, predict many previously unobserved phenomena, and withstand many potential falsifications. From a Bayesian perspective (and according to Popper's corroboration measure), the best hypothesis available explains the most facts with the fewest assumptions, makes the most confirmed predictions, and is most open to testing.
In scientific practice, a superior and well-supported hypothesis will be regarded as a theory. A theory that has withstood the test of time and the collection of new data is about as close as we can get to a scientific fact. An example is the aforementioned notion of a heliocentric solar system. At one time it was a mere hypothesis. Although it is still formally just a well-supported theory, validated by many independent lines of evidence, it is now widely regarded as scientific "fact". Nobody has ever directly observed an electron, stellar fusion, radio-waves, entropy, or the earth circling the Sun, yet these are all scientific facts. As Stephen J. Gould has said, a scientific fact is not "absolute certainty", but simply a theory that has been "confirmed to such a degree that it would be perverse to withhold provisional consent".
"As far as the laws of mathematics refer to reality, they are not certain; and as far as they are certain, they do not refer to reality."
Albert Einstein, addressing the Prussian Acadamy of Science, Berlin , Jan 27, 1921
The validity of a hypothesis does not stand or fall based on just a few confirmations or contradictions, but on the totality of the evidence. Often, data that initially may seem to be inconsistent with a theory will in fact lead to new important predictions. The history of Newtonian physics gives a clear example. The abnormal movement of Uranus was initially considered inconsistent with Newton's new theory. However, by claiming the existence of an unseen planet, the anomaly was explained within Newton's paradigm. In general, an explanation for anomalous behavior should be considered ad hoc unless it is independently verifiable. Positing a new, unseen planet might be considered hedging if there were no independent way to detect if a new planet actually existed. Nevertheless, when technology had advanced enough to reliably test the new prediction, the unseen planet was found to be Neptune.
The lesson to be learned is that alternate explanations for "anomalies" should be treated like any other hypotheses: they should be weighed, tested, and either ruled out or confirmed. But a hypothesis should not be considered invalidated until thorough testing has produced multiple lines of positive evidence indicating that the hypothesis is truly inconsistent with the empirical data.
A crucial related point is that modern scientific theories are probabilistic. This means that all testing of scientific predictions is carried out in a statistical framework. Probability and statistics pervade modern scientific theories, including thermodynamics (statistical mechanics), geology, quantum mechanics, genetics, and medicine. While the mathematics of probability can be daunting for some, a working knowledge of statistics is absolutely essential for judging the fit between observed data and the predictions of any theory.
Lakatos, I. (1974) "Falsification and the Methodology of Scientific Research Progammes." in Criticism and the Growth of Knowledge. I. Lakatos and A. Musgrave. Eds. Cambridge; Cambridge University Press: 91-196.