The paper analyzes four different aspects of abduction when used in science and philosophy. As a preliminary step, it presents the main kinds of abductive arguments in terms of their logical form and extralogical criteria of evaluation. Subsequently, it discusses the standard functions of abductive inference. Finally, it pays attention to four levels of difference in comparing the use of abduction in science with that in philosophy. In particular, it is argued that the difference between abduction in science and philosophy involves the distinct kinds of hypotheses, a different notion of explanation, a difference in the contexts of abductive reasoning, and different criteria for the hypothesis-selection.