Manner of speaking verbs (MoS) are said to induce strong island effects, in contrast with verbs of communication, which allow extraction. The main studies which tried to account for this distinction focused either on the existence of a manner component, of an added layer of meaning, or on that of a nominal element, corresponding to the resulting emitted noise. However, these intuitions according to which they simultaneously denote both manner and result would induce a violation of the Manner Result Complementarity (Levin & Rappaport 2011). What’s more, a closer look at the data shows that there are at least some cases where extraction out of the complement of MoS verbs is actually allowed. The goal of this paper is to put forth an analysis which first of all accounts for the ban on extraction, but also for the variable behaviour that these verbs evince. By postulating two structurally distinct subclasses of MoS verbs, I not only manage to reconcile the two intuitions present in the literature without violating the MRC, but also explain the syntactic behaviour of these verbs with respect to extraction from the post-verbal clause.