Abstract. After a brief presentation of what I take to be the representational démarche in science, I stress the fundamental role of true judgements in model construction. The success and correctness of a representation rests on the truth of judgements which attribute properties to real targeted entities, called "ontic judgements". I then present what van Fraassen calls "the Loss of Reality objection". After criticizing his dissolution of the objection, I offer an alternative way of answering the Loss of Reality objection by showing that the contact of our models with reality is grounded on the truth of ontic judgements. I conclude by examining the consequence of this view on the issue of scientific realism.Keywords: Representation; realism; van Fraassen; model; judgements.In Scientific Representation. Paradoxes of Perspective, Bas van Fraassen (2008) claims that scientific theories are in the first place representations and that the aim of science is to construct empirically adequate representations of phenomena. The novel terminology in which he chooses to couch the aim of science should not mislead us. Such a view essentially accords with the manner in which knowledge has been construed by modern thinkers. In the glorious 17 th century, when modern mathematical science blossomed, empiricist and rationalist philosophers alike proclaimed that to know is to represent. Galileo, Descartes and Locke, among others, had then to address the vexing philosophical problem of how our representations -which for them were ideas in our minds -can in some sense correspond to real external entities. They all bestowed the privilege of hitting on external entities to a specific class of representations, namely mathematical representations, further restricted to geometrical representations. The supposed real correlates of mathematical representations, the so-called primary qualities or properties, acquired then a particular aura, at the expenses of the visual, auditory, olfactory, tactile and gustative secondary qualities that were degraded as mere "appearances". Modern thinkers were well aware that a "representation in my mind" is in itself devoid of truth or falsity. Only claims that an external entity resembles in certain respects my representation of it run the risk of falsity; in compensation they might carry the prize of truth. Despite strong disagreements at the epistemological level, Galileo, Descartes and Locke all embraced the metaphysical thesis that the "real world" is a mechanical, geometrical world of systems, namely a world of entities made of geometrical elements structured by relations of positions and motions. If Principia 15(3): 461-474 (2011).