<p>Digital philosophy of science has split into a normative, model-driven first wave and a descriptive, text-driven second wave. This paper argues for a synthesis of these traditions. We propose a framework grounded in normative naturalism for evaluating scientific norms based on their role in solving local, context-specific problems. This framework is advanced through three complementary digital methods: tracking norm effectiveness, computational modeling research systems, and analyzing scientific discourse. To demonstrate our approach, we conduct a detailed analysis of how reviewers use the term “pseudoscience” in a large corpus of open peer review reports. This case study shows how combining quantitative methods with close reading can uncover the operative, and often implicit, norms that guide scientific communities. Our approach suggests that digital philosophy of science, by systematically examining scientific communication, can open new possibilities for empirically-informed normative guidance in practice.</p>