Our aim in this paper is to propose an ontology for numbers that is compatible with an epistemology that does not invoke mysterious faculties. On the basis of my explanatory system, we find objects capable of being classified: horses, colors, etc. Once grouped, they can be reclassified in units, pairs, and so on. When they use expressions like “three horses”, in fact, I believe that what they mean is “a threesome of horses”. I call nonmathematical numbers those reclassification, and I (...) suggest that arithmetical numbers could be regarded as pictures of nonmathematical numbers, that are closely connected with empirical facts. (shrink)