The naive Baconian view of science is essentially a paradigm which holds natural law as it is discovered by scientific experiment to be the reality which it discerns. In short, it is the idea that because observational data can experimentally demonstrate the validity of a scientific theory, that theory thus reveals an objective fact, a law, about nature. It basically asserts that nature is lawful and we can discern its laws by "listening" as nature tells us its objective facts through science.
The logician W. V. O. Quine, among others, noted that the ideal experiment necessary to the Baconian view would contain two testable hypotheses, of which either was either absolutely true or could categorically be rejected as false. Although the Baconian view is correct insofar as we can assume an experiment which would wholly reject one hypothesis in favor of the other, it is not so clear that we are in fact capable of generating such an experiment.
The nature of scientific hypothesis itself is the ground of inquiry which makes it difficult for us to accept the Baconian view of ideal experiments and absolute scientific laws. The problem is known in the philosophy of science as the underdetermination of theory, a philosophical issue which asks us to examine exactly what is constituted by a scientific hypothesis and whether it expresses or can express a relation to any objective truth.
The Quine-Duhem thesis states that
"any seemingly disconfirming evidence can always be accommodated to any theory."
It is an indictment of the correctness of any hypothesis, which implies that when we form a hypothesis and seek to confirm it through experiment, there is no guarantee that we are not simply seeing a pattern in accord with what we want to see, and that if we hold such a preconceived notion of our hypothesis' correctness, the data can be correlated to confirm the theory. This is possible because our experiment will not contradict the whole system of laws involved in our hypothesis, only certain pieces of it. We might contradict a large "chunk" of theory, but some of the assumptions underlying our hypothesis will still hold, which are compatible with the observational data.
Quine's probable intention in advancing the thesis in the first place was to make a "metaphysical" point: it is human manipulation and not natural law which determines how a theory is interpreted and adjusted in light of experimental evidence. The problem posed is ultimately a call for pragmatism - that a theory should be adjusted in the most useful way, so as to allow prediction and accurate modeling of reality.
The thesis caused a great deal of emotional stirrings and controversy among the establishment, and not only by scientists and philosophers of science. Some have taken a very radical interpretation of the idea, reading it to mean that there is no ground for scientific research whatsoever and that science can tell us only what we wish to be true. Larry Laudan interpreted Quine-Duhem to mean that there is no way to determine a superior theory when competing hypotheses are consistent with experimental data. This creates an unresolvable "egalitarian" version of the thesis, where all such scientific hypotheses are equally true -- i.e. all theories are epistemologically equal.
Politically, the thesis has been taken in hand as a blunt instrument by opponents of science, as well as social constructivists, and has been used in various attacks on the credibility of science itself.
If there is no rational ground for science, there is no reason to treat it as anything other than another political process, which acts to invisibly govern the individual attitudes of scientific professionals by "irrational" and non-obvious emotional, economic, social, and even psychological factors.
sources:
Klee, Robert. "Introduction to the Philosophy of Science", Oxford University Press, 1997.