<p>The present paper examines the compatibility between scientific realism, naturalism, and empiricism, arguing that these three perspectives form an uneasy triangle: they can be coherently defended in pairs but not all three together. My primary focus is to show that radical empiricism and naturalism—whose combination constitutes eliminative naturalism—are incompatible with scientific realism. I distinguish between moderate and radical empiricism to show that while moderate empiricism avoids eliminativism, it nevertheless tends toward conventionalism, thereby blocking a robust version of scientific realism. This allows me to isolate radical empiricism for further analysis. I then argue that the view that truth plays an active role in explaining scientific progress (scientific realism) is compatible with the view that scientific theories are natural phenomena (naturalism) only if we reject radical empiricism. To support this claim, I use Boyd’s account, which combines scientific realism and eliminative naturalism, as a case study.</p>