How science knows: observation, models, measurement and testing: Facts, models, laws and theories
All courses / Understanding Science / Module 1: How science knows: observation, models, measurement and testing Part 3 of 6Facts, models, laws and theories
Scientists use a few words with sharper meanings than in daily speech. A fact, a model, a law and a theory each do a different job, and mixing them up causes most public arguments about science. This part sorts them out and shows how a community of researchers keeps the whole structure honest.
3.1 Facts are checked observations, and models are deliberate simplifications
A road map of Mumbai shows main roads and railway lines but not every lamp-post and not the colour of the buildings. It is wrong in a thousand ways and still extremely useful, because it keeps what matters for finding your way and drops the rest. A model in science is like this. It is a simplified stand-in for something too complicated to handle whole.
Start with facts. A scientific fact is an observation that careful, independent observers agree on, and that can be repeated: water at sea level boils at about 100 degrees Celsius. Facts are not infallible. A fact can be revised when better instruments or conditions show it was only approximately right. From the 1920s, textbooks gave as fact that human cells hold 48 chromosomes. In 1956, with better ways of spreading cells on a slide, Joe Hin Tjio and Albert Levan counted 46, and the fact was corrected.
A model goes further than a fact. It supplies a picture or a set of rules from which facts follow. Atoms drawn as little balls can explain why gases fill containers and why pressure rises with heating. A hydrograph that treats a river as a pipe with inflow and outflow can forecast a flood level. A climate model slices the atmosphere into millions of boxes and calculates how air, heat and water move between them.
Every model has a purpose and a range. The ball picture of atoms fails the moment you ask about light colours, where a quantum picture is needed. A map of Mumbai's roads cannot tell you where it floods. So the sensible questions are: what does this model leave out, and does the thing I am asking about depend on what was left out? A model should be as simple as the question allows, and no simpler.
- Hewitt, Lyons, Suchocki and Yeh, Conceptual Integrated Science, 2nd edition (Pearson, 2013) · ch. 1, 'About Science'
- Science History Institute, Joe Hin Tjio counts chromosomes · web
3.2 Laws describe patterns and theories explain why they hold
Drop a stone from a roof and it falls. Drop another and it falls the same way. The pattern repeats so reliably that we can write it as a rule. A scientific law is such a rule: a concise statement, often an equation, of a regularity found in nature. Newton's law of gravitation states how the pull between two masses depends on the masses and the distance. Boyle's law states that, for a fixed amount of gas at constant temperature, pressure times volume stays constant. A law tells you what happens.
A theory tells you why. It is a broad, well-tested explanation that ties many facts and laws into one structure and predicts new things. The kinetic theory of gases says a gas is a crowd of tiny moving molecules. From that picture, Boyle's law follows: squeeze the same number of molecules into half the volume and they hit the walls twice as often, so the pressure doubles. Evolution by natural selection explains the pattern of fossils, the similarities of related species and the spread of drug resistance in bacteria. Plate tectonics explains earthquakes, volcanoes and mountain belts.
Laws and theories are not ranks on a ladder, with theories rising to become laws. They are different kinds of statement. Laws are descriptions. Theories are explanations. Many theories never become laws because they do not reduce to one short equation, and that is no weakness.
The word 'theory' causes confusion. In daily speech it means a hunch, as in 'my theory is the train is late because of signals'. In science it means the best-tested explanations we have. When someone says evolution or relativity is 'just a theory', the scientific meaning has been swapped for the casual one. The germ theory of disease is a theory too, and nobody sensible calls it a hunch.
Theories are not frozen. Newton's gravity has worked for over three centuries, and still it was refined, not discarded, by Einstein for strong gravity and high speeds. NASA sends spacecraft to Mars using Newton's equations, with relativity as a correction when great precision is required. A good theory stays valid inside its tested range and a better one extends that range.
| Word | What it is | Example | Can it change? |
|---|---|---|---|
| Fact | A checked observation | Water boils near 100 C at sea level | Yes, if better measurement shows otherwise |
| Model | A simplified stand-in | A river as inflow minus outflow | Yes, improved or replaced |
| Law | A compact statement of a pattern | Pressure x volume constant at fixed temperature | Yes, with limits of range |
| Theory | A tested explanation | Molecules in motion explain gas behaviour | Yes, refined or extended |
- Hewitt, Lyons, Suchocki and Yeh, Conceptual Integrated Science, 2nd edition (Pearson, 2013) · ch. 1, 'About Science'
3.3 Honest science is checked by a community, and pseudoscience avoids checking
One researcher can fool herself. A hundred researchers who dislike each other's ideas are much harder to fool. Scientific knowledge is trustworthy less because individuals are honest and more because the system makes it costly to be wrong. Results are published with methods in enough detail to repeat. Other teams try to reproduce them. Journals send papers to independent experts for review. Rivals look for the weak point.
In 1989 two chemists announced 'cold fusion', nuclear fusion in a jar of heavy water at room temperature, which would have solved the energy problem. Within weeks groups around the world tried to repeat it. Most could not, and careful checks of the heat measurements found ordinary errors. The claim failed. That was the system working, even though the episode was embarrassing.
Pseudoscience looks like science but dodges this process. Telltale signs include claims that cannot be tested, excuses added each time a test fails, reliance on testimonials in place of controlled comparison, appeals to ancient wisdom or to suppressed truths, and a refusal to share data. In a double-blind test published in Nature in 1985, astrologers matched birth charts to personality profiles no better than chance. Water 'memory' claims in homeopathic dilutions collapse when you calculate that a 30C dilution, a factor of 10^60, leaves, on average, no molecule of the original substance in the bottle.
Honesty includes admitting uncertainty. Good science papers state limits. Retraction when an error is found is a feature. Fraud occurs, as in any human activity, but it tends to be caught because other teams build on the results and notice when things fail.
- Hewitt, Lyons, Suchocki and Yeh, Conceptual Integrated Science, 2nd edition (Pearson, 2013) · ch. 1, 'About Science'
- Carlson, A double-blind test of astrology, Nature 318, 419 (1985) · web
See it in the table
H₂O WaterCheck your understanding
What is the difference between a scientific law and a theory?
Laws describe what happens, often as an equation. Theories supply the explanation. One does not turn into the other, and both can be refined.A model of a river as a simple pipe predicts water levels well in normal flow but poorly during floods. The most sensible response is to
Every model has a range. The pipe leaves out overflow onto flood plains, which matters in floods, so a model with that included is needed there.Which sign most strongly suggests pseudoscience?
A claim that can always explain away failure cannot be tested, so it can never be shown wrong. The other three are habits of real science.