This review was prompted by the publication of John Gunnell’s latest book, “Conventional Realism and Political Inquiry: Channeling Wittgenstein.” (Chicago, London: The University of Chicago Press, 2020). Theoretically and methodologically, his approach develops the popular current agenda of practical and linguistic turns, and also illustrates well its drawbacks. The review aims to identify the contribution of Gunnell’s work and problems in this context. We fulfilled three objectives – reviewing the book’s main ideas; its particular shortcomings; and the problems representative of the aforementioned broad topic. The first two objectives are fulfilled in parallel, followed by the third. Drawing on Ludwig Wittgenstein’s ideas, Gunnell unfolds his position of ‘conventional realism’ as applied to politics. Gunnell systematically showed that politics largely resides in human collective linguistic practices, which are plural and changeable. They cannot be reduced to any dogmatic paradigm abusing the rhetoric of metaphysics, ‘the natural’, or ‘science’ to justify a specific policy. The making of multiple practices and their change manifest humanity’s collective creative freedom; it also implies their responsibility. However, we note the vulnerability of the general idea, which overstates the unity of thinking and language. Three closely related, representative problems directly concerning politics are shown. The first and broadest one is a combination of trivialness and relativism. It is not clear enough how to compare different practices and grasp relations between them, and the risks remain of reifying the social ‘given’ or trivially noting an arbitrary ‘switch’ between ‘givens’; more attention should be paid to conceptions of rationality that are critical, and neither relativist nor dogmatic. The second problem lies in lessening the role of philosophy and political science in a rational and critical assessment of political practices. The third issue is that Gunnell himself firstly criticised the grounding of practices by scholars, but then largely made the same mistake himself by defending the normative benchmark of democracy.