||If â€˜scientific methodâ€™ is taken to mean â€˜a way of reaching the truth about natural phenomenaâ€™, there is no such thing. There are many â€˜truthsâ€™, and many â€˜methodsâ€™. Nothing can be taken for granted. Even the view, put forward by Francis Bacon in the 16th century, that logical inference based on observation is better than guesswork, does not hold true in every instance. It used to be believed that there was a fixed body of truth about the universe, and that if we used the right method and the right amounts of effort and sophistication, we might one day discover, own, all of it. This no longer seems to be the case; in fact the more scientists discover, the more evidence they find for chance, randomness and unpredictability. Statistically, the amount of such things is still minute but we have to remember two things, first that one aberrant result in a billion billion identical tests of a theory is enough to cast doubt on the theory, and second that we only know, have only probed, a tiny fraction of what is there to be investigated. Laws and paradigms (models, such as physics) are useful in helping us to make sense of natural phenomena, but their value lasts only until we start taking them, in whole or in part, as absolute truths. (Darwin\'s theory of evolution, for example, does not match the paradigm of physics. Newton\'s laws of motion failed to explain the orbit of the planet Mercury, and it was only with the development of the theory of relativity that it was seen that the problem was not that Mercury\'s orbit was an aberration, but that Newton\'s laws were wrong. The theory of relativity itself is incompatible with Quantum mechanics; eventually some discovery will be made which resolves the inconsistencies, but in the meantime the two ideas cover enough ground, and enable enough work to proceed, for us to accept their incompatibility and study it.)
In ancient intellectual investigation of the natural world, the method used was deductive reasoning (see deduction). Starting from simple premises (ones regarded as self-evident, such as that parallel lines never meet), one proceeded by logical steps, each dependent on the one before, until a conclusion was reached. The assumption was that if the initial premise was true, and the logic was properly carried out, the conclusion must also be true.
This is the method of mathematical calculation, and underlies much work in the physical sciences, where behaviour and relationships can be precisely measured and quantified. But it suffers from two major flaws. First, someone must decide on the truth of the initial premise of a line of reasoning, and if that decision is wrong, then the whole line of reasoning is invalidated. (Parallel lines, for example, may never meet in Euclid\'s two-dimensional geometry; but in the real, curved universe they do.) Second, the method tends to advance under its own intellectual impetus, the steps following the processes of logic and sidestepping the real world. This is fine so long as the real world does not throw up examples which conflict with the logic. Failure to relate the intellectual rationalization of science to practical observation, the belief that if observed phenomena conflicted with the events in a logical progression it was the phenomena which were â€˜wrongâ€™, hindered the development of science for two millennia.
Since the Renaissance, the deductive method in science has been replaced by (or gone hand in hand with) the method of induction. Here one starts from an observation (such as that the Sun has risen every day, in human experience, so far), and induces the conclusion that the Sun will rise tomorrow. From a large set of particular statements one goes to a single general statement. The chief flaw in this method is that, once again, it depends on the ability and application of the person doing the work. From a given set of observations, I may induce one general statement, you may induce quite another. The problem is that, unlike the conclusion in a deductive argument (which is, so to speak, â€˜containedâ€™ in the original premise), that in an inductive argument is not, and could in theory be anything at all. In practice, the scientist must spend time gathering evidence to support the thesis (the assumption, or theory, he or she makes based on the original observation).
In the 18th century, David Hume pointed out that the inductive method, though attractive and useful, was logically entirely invalid; in the 20th century Karl Popper said that the way to restore rigour was to seek not to confirm theories, but to refute them. This has led to general and healthy scepticism in science. Nothing is taken on trust, all experiments are repeated and repeated, and everything we â€˜knowâ€™ is taken to be, as it were, a temporary acquisition based on information at present available and a useful basis for speculation and analysis, but by no means absolute truth. This means that science is now two-headed. In speculation it is imaginative, intuitive, a matter of inspiration; in execution it is methodical, mechanical and unimaginative, a matter of perspiration. Its incarnation as a speculative enterprise which uses analytical methods as accurate as can be devised is, some argue, a departure from the real world almost as remarkable as that of the ancient Greek â€˜natural philosophersâ€™ who saw no need to test their reasoning by observation, and a very long way indeed from the simple â€˜Start with a hypothesis, do experiments to test it, and then announce a conclusionâ€™ of the average school exercise-book. KMcL
See also mathematics; philosophy of science; science; scientific laws.