Linguistic mitigation is a pragmatic phenomenon that has been profusely treated in the literature, and yet there are few studies dedicated exclusively to offering methodological criteria for its recognition and analysis (but see Albelda, 2010 and Albelda et al., 2014). Consequently, the researcher must infer the methods for its recognition through problems arising during the analysis or through the examples and reflections offered by the authors who have addressed the issue. The purpose of this article is twofold. Firstly, to present some of the main keys for the recognition of mitigation presented by the bibliography, such as the catalogues of mitigation devices, the context, the face and the illocutionary force. Secondly, this text problematizes Villalba's (2018) proposal for the identification of mitigation based on three tests: the absence test, the commutation test, and the solidarity test. In order to do that, we apply this test to a corpus of examples of mitigation extracted from research articles that study this phenomenon. After that, we reflect on the problems the tests or the examples may present to the researcher. The proposal, in conjunction with the traditional criteria, aims to offer strategies in order to facilitate the analysis of mitigation.