Abstract
This article argues that Thomas Kuhn's views on the existence of the world have undergone significant change in the course of his philosophical career. In Structure, Kuhn appears to be committed to the existence of the ordinary empirical world as well as the existence of an independent metaphysical world, but realism about the empirical world is abandoned in his later writings. Whereas in Structure the only relative worlds are the scientific worlds inhabited by the practitioners of various paradigms, the later Kuhn puts the non-scientific worlds of particular groups or cultures on the same footing as the paradigm-related scientific worlds. The article shows that, on what Ian Hacking called the "new-world problem", the later Kuhn has moved to a more radical antirealist position. It is also argued that the earlier and later solutions to the "new-world problem" face insuperable difficulties, which render Kuhn's account of scientific change implausible.