For several decades now a set of researches from a wide range of different sectors has been developed which goes by the name of “science of complexity” and is opposed point by point to the paradigm of classical science. It challenges the idea that world is “simple.” To the reductionist idea that each process is the sum of the actions of its components it opposes a holistic view (the whole is more than the sum of the parts). The aim of the present article is to analyze the epistemological status attributed in the science of complexity to several fundamental ideas, such as those of scientific law, objectivity, and prediction. The aim is to show that the hope of superseding reductionism by means of concepts such as that of “emergence” is fallacious and that the science of complexity proposes forms of reductionism that are even more restrictive than the classical ones, particularly when it claims to unify in a single treatment problems that vary widely in nature such as physical, biological, and social problems.