Approximate Truths: A New Ground for the Pillars of Scientific Realism

The role of approximate truth (AT) seems quite marginal to the scientific realism debate, especially in light of the fact that AT has often been depicted as a truth surrogate. To the contrary, my project will rethink the role of AT as essential to scientific realism, and will show that AT shouldn't be conceptualized as a truth surrogate only, but as a complex phenomenon of which a qualitative account will be provided.
First, I will critically assess the history of the concept in contemporary philosophy of science. On the one hand, there have been reflections on the concept of truth approximation emerging from the traditional debate on scientific realism; on the other, there is another stream of reflections associated with the concept of truthlikeness, often carried out through formal methods. Still, I will show that a partial reconciliation between these two debates is possible: both acknowledge that approximate truth involves false beliefs - even if this is disputable - and consider approximate truth (or truthlikeness) as pivotal to justify scientific realism.
After considering this common ground, I will analyze the dynamic relationship between truth and AT that this history displays. As a matter of fact, AT can be seen as putting under scrutiny the correspondence theory of truth and other core tenets of realism, such as the metaphysical thesis, namely the idea that 'the world has a definite and mind-independent natural-kind structure', and the semantic stance, according to which theories are 'truth-conditioned descriptions of their intended domain, both observable and unobservable'. For instance, while sympathizing with Tarski's theory of truth, Popper introduced a holistic understanding of truth in order to justify his account of verisimilitude; similarly, Boyd challenged Tarski's account of truth in order to make room for the role of approximation in science, and dwelled on what can be considered as a more radical form of approximation: indeterminacy. In light of these considerations, I will argue that there is a sense of truth approximation in which the latter does not work as a mere surrogate, inasmuch there is no final truth working as a benchmark. Therefore, drawing from Teller, I will introduce the idea that a qualitative, multidimensional concept of approximation has the potential to question the basic tenets of realism and ground a more modest position.
Subsequently, I will analyze the problem of approximate truth from a particular viewpoint, that of natural kinds - which is considered one of the most important battlegrounds for the dispute between realists and antirealists. I will consider the problems that stem from the intersection between the concepts of natural kinds and approximate truth, and in particular the issue of borderline cases. In light of the previous observations, and by drawing from other realist authors, I will try to deflate the importance of natural kinds for the realist's cause.
Finally, I will consider the relationship between approximate truth and the more recent developments of realism, which can be seen as addressing many conventionalist concerns. I will take into account what can be considered the recent "pragmatist turn" of realism, as the concept of truth has become increasingly more sensitive to intentional factors in different realist authors, thus absorbing conventionalist insights. My contention is that these conventionalist elements could be rephrased in terms of approximate truth and that these renewed forms of realism could therefore benefit from a stronger emphasis on truth approximation as presented earlier in the thesis. The hypothesis is that resorting to "Roschian concepts", "exemplars" and similar categories used in the philosophy of psychiatry, which are more sensitive to the phenomenon of truth approximation, might be useful to ground modest forms of realism.