Already in the "classical' Greek culture a partition of the "sciences' was recognized and established either by considering their different aim, or their different subject matter. This was the first appearance of "territories' in science which, however, did not entail a differentiation in the cognitive approach. A new model of science was introduced in the age of Renaissance with the Galilean revolution based on the proposal to delimit the inquiry to the behavior of physical bodies and, moreover, by considering only a restricted amount of their properties (the measurable magnitudes). This delimitation of "territory' was accompanied by a significant novelty in the admitted tools of inquiry, that were concentrated in the use of mathematical descriptions and the adoption of the experimental method. By means of these tools a display of mappings was developed during a couple of centuries, thanks to which the different "sub-territories' of physics were investigated by using the methods originally proposed for mechanics. This was a reductionist program whose aim was that of overcoming the partition in territories and recover the unitary totality at least of the physical world. The failure of this program at the end of the 19th century was interpreted by certain philosophers as the evidence that science is unable to offer a reliable knowledge of reality. This pessimistic view can be avoided by recognizing that science is constituted by a rich variety of "territories' (including also the broad display of the social sciences and humanities), that is, of really distinct disciplines each having its own domain of objects which is also its specific "conceptual space'. Within this conceptual space several scientific theories are proposed offering different ways of mapping the territory. Both moments rely upon the presence of operational procedures that are essential for linking the maps with the territories and justify a realistic conception of science.