constructive role in the scientific realism debate. It is the aim of the present paper to tackle this question by investigating the relationship between Carnap's approach toward the realism issue, on the one hand, and the (presumably) realist positions defended by his logical empiricist fellows Hans Reichenbach and Eino Kaila, on the other. It will be shown that Carnap agreed with Reichenbach that realism has essentially to do with language, but that he disagreed with him over the significance of probability in defending the scientific realist stance. My point will be that realism is not a "problem of language.' Furthermore, it will be argued that Carnap was correctly criticized by Reichenbach for neglecting the role of probability in science. Nevertheless, what can be learned from Carnap's approach toward the realism issue is that scientific realism cannot be defended in the way that Reichenbach himself suggested, namely by arguing inductively for the adequacy of the realistic "language form.' Rather, scientific realism can only be defended on the ground that language itself is dependent on the "structure of the world.' And this is exactly the point where, eventually, Kaila will enter the scene." /> Realism as a problem of language – from Carnap to Reichenbach and Kaila - Neuber Matthias | sdvig press