Someone says, “That's just a theory,” meaning a loose suspicion. A scientist uses the same word for an explanation that has survived years of testing. The difference can make public debates sound as if two people disagree about evidence when they actually disagree about a definition. In science, a theory is a coherent explanation of a range of observations, supported by evidence and capable of being tested against the world.¹
Gravity, evolution by natural selection and the germ theory of disease are theories in this sense. Calling them theories does not place them halfway between “guess” and “fact”. It describes the job they do: they explain why many facts fit together. The amount and quality of support for particular theories can still differ greatly.
From question to explanation
Imagine noticing that plants near a window lean towards it. A guess might be “they like the view”. A testable hypothesis might be that directional light changes growth on opposite sides of a stem. You could rotate plants, vary the direction of light and compare their growth with controls. A broader theory would connect those results with other observations through a mechanism that works beyond one windowsill.²
A hypothesis is often a focused proposal for a test. A theory is usually a larger explanatory framework that draws together multiple lines of evidence. Neither word guarantees correctness. An elegant explanation can fail when new measurements contradict it. What makes the work scientific is its exposure to such checks.
A useful theory also reaches beyond the observations that inspired it. It may predict what happens under new conditions or tell researchers what measurement could distinguish it from a rival account. Successful prediction is powerful, especially when the result was not built into the theory after the fact. Replication by independent investigators matters too: a claim that depends on one apparatus, one sample or one researcher's judgement is less secure.³
Why theories do not graduate into laws
A familiar ladder says that a hypothesis becomes a theory, which finally becomes a law. That is misleading. A law and a theory answer different kinds of questions. A law summarises a pattern or relationship; a theory explains the processes behind it. Theories do not earn promotion to laws by accumulating enough evidence.²
Newton's law of gravitation, for example, describes how objects attract in an enormously useful mathematical form. Relativity gives a different and deeper account of gravity as spacetime geometry. The law did not vanish when the explanation changed, nor did relativity “graduate” into a law. Scientists still use Newton's equations where they are accurate enough, such as many ordinary engineering problems.
Provisional does not mean flimsy
Scientists say theories are revisable because they allow reality to overrule them. Sometimes new findings refine a theory's range: a model works at familiar speeds but needs correction near the speed of light. Sometimes an explanation is replaced. That openness is a strength, not a confession that all ideas are equally likely.¹
Evidence can also be indirect. No one needs to watch every individual past event in order to test an explanation of them. The explanation can be checked against independent traces, predictions and measurements. The important question is whether an account survives attempts to prove it wrong and whether it explains the evidence better than alternatives.³
What makes evidence strong?
A theory becomes persuasive when different methods converge. If an explanation predicts one result in a laboratory and another in a field survey, agreement across both is more valuable than repeating the same measurement with the same potential error. Scientists also examine whether the proposed mechanism is consistent with established findings in neighbouring areas. A theory that explains one anomaly by creating ten new contradictions needs more work.
Theories vary in scope. Some explain a narrow phenomenon; others connect large parts of a discipline. Broad scope is impressive only when the theory continues to make precise, risky predictions. An explanation that can be adjusted to fit any conceivable outcome is hard to test and therefore less informative.³
Consensus is earned, not voted into existence
When many specialists accept a theory, the reason should be the accumulated evidence and the failure of competing explanations to do as well. Consensus is a useful clue for non-specialists because no one can personally repeat every experiment. It is not a guarantee against future revision and is not a substitute for asking what the evidence is.
Scientific disagreement is also normal at the edges. Researchers may agree on a well-supported core while debating details, mechanisms or the limits of its application. Public arguments sometimes present any open question as if the entire field were guessing. That is like claiming we know nothing about weather because tomorrow's exact cloud pattern is uncertain.
A good theory gives scientists work to do. It suggests measurements, reveals gaps and invites challenge. If it survives, confidence grows. If it fails, the failure teaches us something about the world. That is a much more demanding role than an ordinary guess can play.¹
So when you hear “only a theory”, ask what sort of theory is meant. In ordinary conversation, it may indeed be a hunch. In science, the word usually signals a serious explanatory achievement. Its credibility comes from tests, not from the label itself.