In his cutting-edge treatment of paradigm shifts in science, Kuhn ends up denying the correspondence view of truth. That is, he cannot bring himself to say that a more recent scientific pardigm (say, oxygen chemistry) is more correct than the paradigm it replaced (phlogiston theory). For Kuhn, the newer paradigm is merely more practical, better at making predictions, or simply more popular, than the older one--it is not a truer understanding of the world in which we live.
Such a denial is not a necessary part of Kuhn's thesis, but an unnecessary addition. And I am not the first to notice and be troubled by this Kuhnian aspect. In fact, he apparently caught a lot of flack for this position, since scientists and philosophers alike think that correspondence to reality is a pretty important part of learning generally and of scientific advance in particular.
In response to criticism, Kuhn attempts (in the postscript to the second edition) to defend this quasi-relativism, in what I find to be a bizarre and illogical way... he draws the analogy of an evolutionary tree, this one containing not living things but scientific theories. (The idea that living things can be represented by an evolutionary tree has fallen upon hard times, even among committed naturalists, but that's beside the present point.) If we can come up with criteria that would enable us to distinguish a more recent theory from an earlier one--without appeal to correspondence to truth--
then scientific development is, like biological, a unidirectional and irreversible process.In other words, if we can accept a non-teleological view of life, than it ought to be a simple task to accept that what makes a scientific theory better than a previous one has little to do with the reality of the universe it describes.
This will, of course, be unsatisfactory to those of us who do not accept a non-teleological view of the world, who--when faced with overwhelming evidence for design in the universe in general and in living things in particular--cannot kid themeselves into thinking of design as only apparent or illusion. But many non-theists (among scientists and philosophers) are likewise unsatisfied by Kuhn's reasoning here. There just doesn't seem to be any logical link (a link apparently assumed by Kuhn, since never spelled out) between a speculative (and now rejected) biological tree and the epistemological tree he would plant in the path of a serious question, to wit, "Why shouldn't we expect a scientific theory to better approximate reality than the theory it supplants?"