Untold hours have been spent trying to sort out the differences between these ideas. Should we bother?
Ask what the differences between these concepts are and you’re likely to encounter a raft of distinctions; typically with charts and ladders of generality leading from hypotheses to theories and, ultimately, to laws. Countless students have been exposed to and forced to learn how the schemes are set up. Theories are said to be well-tested hypotheses, or maybe whole collections of linked hypotheses, and laws, well, laws are at the top of the heap, the apex of science having enormous reach, quantitative predictive power, and validity. It all seems so clear.
Yet there are many problems with the general scheme. For one thing, it is never quite explained how a hypothesis turns into a theory or law and, consequently, the boundaries are blurry, and definitions tend vary with the speaker. And there is no consistency in usage across fields, I’ll give some examples in a minute. There are branches of science that have few if any theories and no laws – neuroscience comes to mind – though no one doubts that neuroscience is a bona fide science that has discovered great quantities of reliable and useful information and wide-ranging generalizations. At the other extreme, there are sciences that spin out theories at a dizzying pace – psychology, for instance – although the permanence and indeed the veracity of psychological theories are rarely on par with those of physics or chemistry.
Some people will tell you that theories and laws are “more quantitative” than hypotheses, but the most famous theory in biology, the Theory of Evolution, which is based on concepts such as heritability, genetic variability, natural selection, etc. is not as neatly expressible in quantitative terms as is Newton’s Theory of Gravity, for example. And what do we make of the fact that Newton’s “Law of Gravity” was superceded by Einstein’s “General Theory (not Law) of Relativity?”
What about the idea that a hypothesis is a low-level explanation that somehow transmogrifies into a theory when conditions are right? Even this simple rule is not adhered to. Take geology (or “geoscience” nowadays): We have the Alvarez Hypothesis about how an asteroid slamming into the earth caused the extinction of dinosaurs and other life-forms ~66 million years ago. The Alvarez Hypothesis explains, often in quantitative detail, many important phenomena and makes far-reaching predictions, most remarkably of a crater, which was eventually found in the Yucatan peninsula, that has the right age and size to be the site of an extinction-causing asteroid impact. The Alvarez Hypothesis has been rigorously tested many times since it was proposed, without having been promoted to a theory.
But perhaps the Alvarez Hypothesis is still thought to be a tentative explanation, not yet worthy of a more exalted status? It seems that the same can’t be said about the idea that the earth’s crust consists of 12 or so rigid “plates” of solid material that drift around very slowly and create geological phenomena, such as mountain ranges and earth-quakes, when they crash into each other. This is called either the “Plate Tectonics Hypothesis” or “Plate Tectonics Theory” by different authors. Same data, same interpretations, same significance, different names.
And for anyone trying to make sense of the hypothesis-theory-law progression, it must be highly confusing to learn that the crowning achievement of modern physics – itself the “queen of the sciences” – is a complex, extraordinarily precise, quantitative structure is known as the Standard Model of Particle Physics, not the Standard Theory, or the Standard Law! The Standard Model incorporates three of the four major forces of nature, describes many subatomic particles, and has successfully predicted numerous subtle properties of subatomic particles. Does this mean that “model” now implies a large, well-worked out and self-consistent body of scientific knowledge? Not at all; in fact, “model” and “hypothesis” are used interchangeably at the simplest levels of experimental investigation in biology, neuroscience, etc., so definition-wise, we’re back to the beginning.
The reason that the Standard Model is a model and not a theory seems basically to be the same as the reason that the Alvarez Hypothesis is a hypothesis and not a theory or that Evolution is a theory and not a law: essentially it is a matter of convention, tradition, or convenience. The designations, we can infer, are primarily names that lack exact substantive, generally agreed-on definitions.
So, rather than worrying about any profound distinctions between hypotheses, theories, laws (and models) it might be more helpful to look at the properties that they have in common:
1. They are all “conjectural” which, for the moment, means that they are inventions of the human mind.
2. They make specific predictions that are empirically testable, in principle.
3. They are falsifiable – if their predictions are false, they are false – though not provable, by experiment or observation.
4. As a consequence of point 3., hypotheses, theories, and laws are all provisional; they may be replaced as further information becomes available.
“Hypothesis,” it seems to me, is the fundamental unit, the building block, of scientific thinking. It is the term that is most consistently used by all sciences; it is more basic than any theory; it carries the least baggage, is the least susceptible to multiple interpretations and, accordingly, is the most likely to communicate effectively. These advantages are relative of course; as I’ll get into elsewhere, even “hypothesis” is the subject of misinterpretation. In any case, its simplicity and clarity are why this website is devoted to the Scientific Hypothesis and not the others.