In this paper, an attempt is made to solve various problems posed to current theories of verisimilitude: the problem of linguistic variance; the problem of which are the best scientific methods for getting the most verisimilar theories; and the question of the ontological commitment in scientific theories. As a result of my solution to these problems, and with the help of other considerations of epistemological character, I conclude that the notion of 'Tarskian truth' is dispensable in a rational interpretation of (...) the scientific enterprise. As a logical result, however, falsificationism will be vindicated. (shrink)