A well-fed chicken shows why science can never prove its general laws
Every morning the farmer brings food, so the chicken expects breakfast tomorrow. It is reasonable, and it may be fatally wrong: perhaps the farmer is fattening it up. That small parable captures the problem of induction, one of several basic questions about science on which philosophers still have no consensus.
Nobody has shown how piling up specific cases can prove a general statement. One response accepts that induction only makes conclusions more probable, though turning evidence into a probability raises fresh puzzles. Another says scientists reason abductively, choosing the best explanation, often guided by Occam's razor. But simplicity is slippery: most philosophers accept there is no theory-neutral measure of it. Nicholas Maxwell argues that unity, not simplicity, is what really steers theory choice.
Separating science from non-science is known as the demarcation problem, which Karl Popper called the field's central question. The logical positivists tied meaning to observation; Popper countered that a scientific claim must be falsifiable in principle. No single answer has won general acceptance, and Martin Gardner suggested recognising pseudoscience the way a US judge once described obscenity: you know it when you see it. Richard Feynman coined cargo cult science for work that looks like research but lacks rigorous honesty.
Observation is not neutral either. Before general relativity, astronomers viewing the Einstein cross would probably have counted five separate objects; with the theory, they see two, one of them imaged four times by gravity. Such observations are called theory-laden, and some argue all observation is. Thomas Kuhn's 1962 book The Structure of Scientific Revolutions challenged the picture of steady, cumulative progress, arguing that advances make sense only within a paradigm, while Paul Feyerabend denied there is any single scientific method.
Then there is the question of what science is for. Scientific realists say good theories are true or approximately true, even about unobservables like electrons. Anti-realists doubt science reaches such truths, and instrumentalists say truth is beside the point: theories should be judged on whether they predict well and enable working technology.
Source: Philosophy of science
