The word hypothesis once meant the plot summary of a Greek play
In ancient Greek the word literally meant placing something underneath, and theatre audiences used it for a synopsis of a drama. Today it names a proposed explanation that must risk a testable prediction. Along the way it was even used to muzzle Galileo, who was told to treat a moving Earth as merely hypothetical.
Plato's Meno shows Socrates borrowing a trick from mathematicians: reasoning forward from an assumption to see where it leads. In that older sense a hypothesis was a convenient shortcut rather than a claim about reality. Cardinal Robert Bellarmine leaned on exactly this meaning in the early 17th century, warning Galileo that the Earth's motion could be entertained only as a calculating device, never asserted as fact.
The modern scientific version is stricter. A hypothesis should rest on observation and yield a prediction others can reproduce. When independent experiments keep confirming it, it may graduate into a theory, which is why scientists bristle when the two words are swapped in everyday speech. Karl Popper argued that a proposal which cannot, even in principle, be shown false does not count as scientific at all, though other philosophers preferred verifiability or coherence as the test. Imre Lakatos and Paul Feyerabend, Popper's colleague and student, later tried to fuse the competing approaches.
Researchers choosing between rival explanations, according to Schick and Vaughn, weigh testability, parsimony in the spirit of Occam's razor, scope, fruitfulness and how well a proposal fits established knowledge. Some deliberately adopt a working hypothesis they suspect is incomplete or wrong, simply to push a stalled project in roughly the right direction. In logic the term is narrower still: in an if-then statement, the if-part is the hypothesis.
Statistics formalises the idea by pitting a null hypothesis, which says no relation exists, against an alternative. Conventional thresholds for wrongly rejecting a true null are 0.10, 0.05 and 0.01, and the level must be fixed before the data are seen; choosing it afterwards invalidates the test. Carl Hempel pictured a theory as a net floating above the plane of observation, tied down by strings of interpretation. There is even a Mount Hypothesis in Antarctica, named to honour the role these guesses play in research.
Source: Hypothesis
