Hypothesis, Theory and Law: Three Different Jobs

One word doing two very different jobs

In everyday English, a theory is a hunch. It is the thing you say when you are guessing and want to sound organised about it. It carries no requirement to have checked anything.

In science the word means close to the opposite. Germ theory, atomic theory, the theory of plate tectonics, the kinetic theory of gases, evolution by natural selection: each of these is a framework that has absorbed enormous quantities of evidence and is used to make predictions that keep coming out right. Nobody refuses antibiotics on the grounds that germ theory is only a theory, and nobody hesitates at the idea that continents move.

The confusion is not the fault of students. It is a genuine collision between two meanings of one word, and the useful move is to notice which room you are standing in.

What each of the three actually does

The clearest way to hold hypothesis, law and theory apart is not by how certain each one is, but by the question each one answers.

Read that list again and the standard misconception collapses on its own. A theory does not become a law after enough evidence. They are not two rungs of one ladder — they are answers to what happens and why it happens. Boyle's law states that pressure and volume are inversely related for a fixed mass of gas at constant temperature. Kinetic theory explains why, in terms of particles colliding with the container walls. More evidence for Boyle's law will never turn it into kinetic theory, because the two are not the same kind of statement.

• A hypothesis predicts. It is specific, testable and disposable: cress seedlings kept in the dark for five days will be taller and paler than those grown in the light.
• A law describes. It states a relationship that holds, usually as an equation, without saying why — Ohm's law, Boyle's law, Newton's law of universal gravitation.
• A theory explains. It says why the laws hold and ties together facts, laws and tested hypotheses into one account — the kinetic theory of gases, germ theory, plate tectonics.

How a hypothesis has to be written to be worth anything

Most marks lost on this topic are lost writing hypotheses, not defining them. Plants need light is not a hypothesis. It names no variable to change, no variable to measure, and no direction of effect, so no experiment could contradict it.

A usable hypothesis names three things: what you will change, what you will measure, and which way you expect the measurement to go. Increasing the light intensity from 0 to 60 W will increase the mean height of cress seedlings after five days does all three, and it can be shown to be wrong — which is its most important property.

That last point deserves stating outright, because it is counter-intuitive. A hypothesis is valuable in proportion to how easily an experiment could kill it. A prediction that fits every possible outcome has told you nothing, no matter how confidently it is written.

Why just a theory is a category error

The phrase treats theory as a rank below fact, as though a theory were an idea waiting for its promotion. The National Academy of Sciences puts it the other way round: theories are the end points of science, developed from extensive observation, experimentation and reflection, and incorporating a large body of scientific facts, laws, tested hypotheses and logical inferences. A theory is where the evidence ends up, not where it starts.

None of which means a theory is beyond revision. Newtonian gravitation was replaced as the fundamental account by general relativity, and it is still used every day for anything that is not moving very fast or sitting in a very strong field. That is how the replacement happened: a better explanation covering everything the old one covered plus what it could not. It did not happen by someone pointing out that it was only a theory.

For a student sitting a paper, the practical version is short. Do not write that a theory is a guess. Write that it is an explanation supported by a large body of evidence, and that a hypothesis is a testable prediction made before the evidence is gathered.