As a
philosopher of
science, Imre Lakatos was less concerned with the
philosophy of
theory (like
Carl Hempel,
Pierre Duhem and others) and more of the research process within the
scientific world. His description of a good
research program consists of four basic concepts: the
hard core, the
protective belt, and the positive and negative
heuristics.
The hard core is a theory or a set of theories that has been tried and tested again and again throughout the years, so much that for all intents and purposes it as accepted as scientific truth. In the process of experimentation, this hard core should not be questioned or attempted to be disproven. An example of a hard core theory would be Newton's gravitational theory. Time and time again this theory was tested, and time and time again its predictions came true; so much so that it became a hard core: "irrefutable" by Lakatos' standards.
The process of testing, examining, and qualifying a hard core creates what Lakatos calls a protective belt around the theory. The protective belt consists of less expansive theories and assumptions that, unlike the hard core, are subject to change when a prediction fails. Under the hard core of Newtonian theory, if a planet's orbit does not behave in a pattern predicted, there is a problem within the protective belt - some assumption is wrong or there is a flaw in measurement.
Lakatos derived positive and negative heuristics to create a universal standard by which scientists should conduct research. They instructs scientists in what they should do when a prediction fails. He claims that the positive heuristic "saves the scientist from becoming confused by an ocean of anomalies." It dictates that if a prediction derived from the hard core fails, that there is some factor unaccounted for, some missing element, and that the researching scientist should propose additional factors to account for the failure in prediction. The negative heuristic simply states that scientists should leave the hard core alone. They should operate under the assumption that the hard core is true, and any discrepancy in an experiment is a result of flaws in an external theory that the hard core is not dependent on.
An example of these heuristics in action is the discovery of Neptune. Scientists of Newtonian theory noticed that there were perturbations in Uranus's orbit that their current observations could not explain. They came up with the idea of another planet, further away from the sun than Uranus, that would account for the odd orbit. After the subsequent calculations were complete, they pointed their telescopes at the area in space these calculations directed and - voila! - discovered Neptune there, exactly as predicted. Throughout this process they did not question the validity of Newtonian physics(the hard core).
Lakatos' research programs have not gone without criticism. Alan Musgrave especially, a colleague of Lakatos, refuted his points by claiming that there was no such thing as a hard core theory - everything in science is questionable and refutable. Even Lakatos' example of Newtonian theory fell with the advent of the notion of relativity. Musgrave also claimed that there were heuristics within scientific research, but they were not so clearly defined and inflexible as Lakatos claimed they should be. The common philosophy today is that semblances of a hard core theory exist, like evolutionary theory in biology and quantum theory in physics, but a fair amount of skepticism should be involved and that nothing in the scientific world be taken as pure truth.
Sources: Imre Lakatos, Falsification and the Methodology of Scientific Research Programmes, 1970.