The ‘quality’ of a scientific publication is not an absolute but must be assessed in relation to a journal’s mission. It should be judged primarily by its disciplined intellectual rigor, bearing in mind what course the publication aims to take, as between the Scylla of hidebound conservatism and the Charybdis of mindless speculation. One commonly applied measure of the ‘quality’ of a periodical, perhaps most overtly in the social sciences, is its rejection rate: the higher that rate, the higher the presumed quality. In the natural sciences, a somewhat similar criterion is invoked, perhaps less overtly and certainly less quantitatively, when kudos comes for being published in journals ‘hard to get into’, such as Nature or Science or the Journal of the American Chemical Society. But consider the professed, or implicitly taken-for-granted, aim of these ‘top’ journals. Actually, there are 2 aims: that what gets published should be ground breaking; and that what gets published should not be in error. But it seems not to be commonly understood that these aims are incompatible. The first implies a willingness to be often wrong, at least to some degree, because it is always difficult to judge the validity of something that is without precedent. On the other hand, the second places high barriers in the way of anything so novel as to call into question ideas that have hitherto been widely accepted. Between these 2 incompatible aims, no journal can avoid making its own choice, at least implicitly, in which direction to lean. The judgment of a journal’s quality should then be based upon how well it performs its chosen task, not according to whether one agrees or differs with the journal’s aim of emphasizing novelty over reliability or vice versa.