Thomas Samuel Kuhn (July 18 1922 - June 17 1996) wrote extensively on the history of science, and developed several important notions in the philosophy of science.
Kuhn obtained his Ph.D in physics from Harvard University in 1949, and taught a course in the history of science at Harvard from 1948 to 1956. After leaving Harvard, Kuhn taught at the University of California, Berkeley until 1964, then at Princeton University until 1979, and then at the Massachusetts Institute of Technology (MIT) until 1991.
Kuhn is most famous for his book The Structure of Scientific Revolutions (SSR) in which he presented the idea that science does not "evolve gradually toward truth", but instead undergoes periodic revolutions which he calls paradigm shifts.
Kuhn's analysis of the history of science suggests to him that the practice of science comes in three phases. The first phase, which is undergone only once, is the "pre-scientific phase," in which there is no consensus on any theory of explanation. This phase is generally characterized by several incompatible and incomplete theories.
Eventually one of the theories becomes so generally accepted that practicioners of science begin to successfully use it in methodical ways. Other knowledge such as common terminology, common experimental methods and equipment, and to a greater or lesser degree a common interpretation of scientific phenomena all develop into a paradigm. After this ocurrs, "normal science" begins. Kuhn explains that normal science is what scientists spend most of their careers doing. It can only be performed under a specific paradigm, and its goal is to explain and expand the paradigm. Kuhn explained normal science as a process of puzzle solving. Armed with knowledge provided by a paradigm, scientists can begin to make well founded and trusted assumptions about what they are studying. This may seem to violate long held ideals about objectivity in science, however it is extremely difficult to study anything without making at least a few basic assumptions (see Naive Empiricism). The challenge of normal science is to see how well one can apply all his or her knowledge and assumptions to a certain problem.
It is important to note that there are two main advantages and disadvantages to using a paradigm to make assumptions about a particular topic. The advantage is that if all scientists are using similar assumptions, then their methods, terminology, and analyses will all be very homogenous and easily compared; it allows for greater communication and cooperation between people. However, if many scientists use similar assumptions, which are not entirely correct, it may lead scientists astray for a very long time before an anomaly occurs which brings attention to the problem. When this happens there is usually a period of disagreement between scientists, and the theory is modified in an ad hoc way to accommodate experimental evidence which might seem to contradict the original theory.
Eventually, the current explanatory theory fails to explain some phenomenon or group thereof, and someone proposes a replacement or redefinition of the theory. This is what Kuhn calls a paradigm shift, which ushers in a new period of revolutionary science. Kuhn believes that all scientific fields go through these paradigm shifts multiple times, as new theories supplant the old.
One well known Kuhnian example involves Copernicus' suggestion that the Earth revolves around sun, rather than the Ptolemaic suggestion that the Sun (and the other planets and stars) revolve around the Earth. Prior to Copernicus there was an elaborate set of epicycles (circles on top of circles) which were used to predict the movements of the "heavenly bodies." Ptolemy's original epicyclic combinations were, by the Middle Ages, becoming noticeably less adequate, and "fixes" by later astronomers were more and more elaborate. Copernicus offered a return to an alternative view (suggested by many in Antiquity) but with rather better data to support it; this new account decreased the complexity of theory necessary to account for the available observations. Of course, once Copernicus' theory was accepted by other astronomers, it ushered in a new period of "normal science." Refinements added by Kepler and Newton adhered to the new paradigm.
Other more recent examples are the acceptance of Einstein's general relativity to replace Newton's account of gravity in the 1920s and 1930s and Suess and Wegener's plate tectonics the 1960s by geologists.
According to Kuhn, the science before and after a paradigm shift are so much different that their theories are "incomparable" - the paradigm shift does not just change a single theory, it changes the way that words are defined, the way that the scientists look at their subject, and perhaps most importantly the questions that are considered valid, and the rules used to determine the truth of a particular theory.
This incommensurability applies not just before and after a paradigm shift, but between conflicting paradigms. It is simply not possible, according to Kuhn, to construct an impartial language that can be used to compare conflicting paradigms. It is not possible to perform a neutral comparison between conflicting paradigms, because the very terms used belong within the paradigm, and are therefore different in different paradigms.
This has important implications for other attempts to explain scientific progress. Kuhn (SSR, section XII) points out that the probabilistic tools used by verificationistsists are in themselves inadequate to the task of deciding between conflicting theories, since they are a component of the very paradigms they seek to compare. Similarly, observations intended to falsify a statement will be part of one or other of the paradigms they seek to compare, and so inadequate to the task. Advocates of such paradigms are in an insidious position. 'Though each may hope to convert the other to his way of seeing science and its problems, neither may hope to prove his case. The competition between paradigms is not the sort of battle that can be resolved by proof' (SSR, p. 148).
Kuhn attributes the success of science to the way in which scientists are able to work within a paradigm, removing the need to repeatedly work from first principles. For Kuhn, it is that scientists work within a particular kind of community that explains the astonishing success of science. 'The scientific community is a supremely efficient instrument for maximising the number and precision of the problem solved through paradigm change' (SSR, p. 169).
There is an old observation by Max Planck:
Kuhn is interpreted by postmodern and post-structuralist thinkers as having undermined the enterprise of science by showing that scientific knowledge is dependent on the culture of groups of scientists rather than on adherence to a specific, definable method. In this regard he is considered a precursor to the more critical thinking of Paul Feyerabend. Kuhn's work has also been interpreted as blurring the demarcation between scientific and non-scientific enterprises because it describes scientific progress without reference to an idealised scientific method that can be used to distinguish science from non-science. In the years after the publication of The Structure of Scientific Revolutions debate raged with adherents of Popper's falsificationism such as Imre Lakatos.