At the end of the previous century a remarkable reformulation of epistemology emerged, i.e., naturalized epistemology. What is characteristic of this epistemology is its rejection to infallibilism and apriorism. It asserts that scientific empirical results are crucial to solve traditional inquiries about knowledge. Quinean naturalized epistemology claimed that we should abandon traditional epistemology and replace it with psychology. Another brand of naturalized epistemology is evolutionary epistemology, an approach to knowledge aiming to answer traditional epistemological questions based on the theory of evolution by natural selection. It has two different but interrelated programs, the first of them (EET). accounts for scientific theory change as resembling the mechanisms of natural selection theory. The second program (EEM) studies the development of our cognitive capacities and structures as well as their fixation in our brain along evolution. It is the extension of biological theory of evolution to cognitive activity and its apparatus like the brain and sensory and motor systems. After evolutionary epistemology, neurophilosophy, another brand of naturalized epistemology arouse, one which is continuing much of the theoretical work started by evolutionary epistemology. It tries to answer traditional epistemic enquiries analyzing the place where knowledge is produced: the brain. A particular representative of this project is P. S. Churchland ‘neurophilosophy’, a program that looks forward to reduce and eliminate traditional epistemology. Churchland is convinced that neurology is all we need to elucidate questions about knowledge. These three programs have accomplished some kind of reductionism, mainly an eliminative one, and this fact has brought them major impasses. What I will do is to review the mentioned programs, going deep into the last one. The twofold purpose is to show that there is no way back once naturalized epistemology has arrived, as long as knowledge is a subject to be disentangled interdisciplinarily. However, it is also pertinent not being captive into exaggerated versions of it, the ones that don’t take into consideration human reasoning. The tool for doing so will be the study of reductionism that the mentioned naturalized epistemologies accomplish. In the end I will hazard a theory of reductionism useful for naturalized epistemologies in general.