Social philosophy of language has recently focused on toxic speech and counterspeech. Roughly, toxic speech is discourse that engages (or incites people to engage) in bigoted practices; counterspeech is communication that tries to counteract the harm that other speech may bring about. In this talk, I consider a counterspeech strategy specifically tailored to address implicit toxic discourse – what Langton (2018) calls blocking. In its most paradigmatic examples, blocking amounts to spelling out and challenging the objectionable contents implicitly conveyed by a given utterance, thus preventing toxic material from entering the common ground by default. First, I’ll illustrate the cognitive and social costs and the drawbacks typically associated with blocking. Then, I’ll consider online conversations and raise the question of whether and how people engage in blocking on social media, in light of its costs, risks, and repercussions. To address this issue, I present the qualitative study I ran on Facebook and Twitter, which resulted in a taxonomy of online blocking, based on real instances from ecological contexts. My investigation shows how each kind of blocking modulates in a distinctive way the costs and drawbacks usually associated with this practice.