A scientific theory is fundamentally an explanation of phenomena. These phenomena may be observed in the world around us, for example, the tides in the oceans or the different varieties of living things. They may be observed with the aid of devices made for the purpose, for example, observing distant galaxies with a telescope or bacteria through a microscope. The observations we want to explain may also be the result of highly contrived experiments, for example, the collisions of particles in a particle accelerator or chemical reactions in glassware.
In ordinary life, we may look for explanations of single events, such as “who left the door open?”, “how did that batch get contaminated?” or “how did that person die?” We may even talk of having a ‘theory’ about how something happened. Usually this means that we are not sure about the explanation. If we are sure, then we would usually describe the explanation as a fact.
Scientists use the word “theory” in a very different way. A scientific theory attempts to do two things as well as explain. It attempts to generalise: to cover all the occurrences of a particular phenomenon. In the theory of gravity, there are not separate explanations of each individual case of gravitational attraction, for example, falling objects on the earth’s surface. The theory describes what is common to all kinds of gravitational attraction, including objects falling on the moon, the tides, the motions of the planets and the structures of the galaxies. If you were looking at these different things without the benefit of a modern education you might not recognise that there was a single explanation linking these various phenomena. It took Newton’s genius to do that.
A good theory also unifies many different phenomena, where the connection may be very difficult to see. For example, the theory of plate tectonics became accepted during the 1960s because it explained not only old and otherwise inexplicable phenomena such as the existence of long mountain chains (like the Himalaya and Andes chains) that are mostly made up of tens of kilometres thickness of rocks formed from the erosion of older mountains, but also the new observations that were being made of long volcanic ridges and very young crustal rocks under the oceans.
Observations are tests of theories. Most times when scientists build an apparatus to do an experiment, they have a theory in mind. If the observations that are made with an apparatus are not consistent with the theory, then of course the experimenter will check the design of the apparatus first, but if nothing can be found wrong with it, then the theory is refuted, and a new theory, or an improved version of the old theory, is required.
It is important to realise that, in science, the word ‘theory’ has no implication about how well understood a theory is, about how certain it is, or how many practising scientists in that field accept it*. A new theory must obviously go through a stage where most people regard it as provisional and not properly tested. This was the case with the theory of plate tectonics in the 1950s-1960s, although the theory had been around for a few decades before that. Nowadays, every practising geologist accepts plate tectonics, although there are always details that are in dispute. For example, we know that Iceland sits on a spreading ridge in the Atlantic, but there is currently disagreement over exactly what is happening under Iceland to make it such a prominent feature.
Similarly, there are no practising biologists, in fields related to evolution, who doubt that evolution by natural selection is correct, although there are always important details to be discovered and clarified. The important thing to remember is, that although we can never be 100% certain about a theory, there are many theories that we have no scientific reasons to doubt are fundamentally true. These include the theory of gravitation, which is used to make predictions of eclipses and other movements in the solar system accurate to fractions of a second. Well-established theories also include quantum theory and the theory of evolution by natural selection. The reason why these theories are accepted is that there is no other explanation in each case that fits the full range of obsrvations so well.
*Sometimes the word ‘hypothesis’ is used to describe a new theory that has not much observational support, but there is no consistency about this usage.