Drawing on the assumptions made in Construction Grammar(s), the present proposal addresses the debate between formulating broad-scale generalizations of the type postulated by Goldberg (1995) or finer-grained analyses, heavily based on lexical-class identification, as those put forward by Boas (2010, 2011), who claims that Goldberg’s account leads to the over-generation of ungrammatical examples. The position taken here is that, although Goldberg’s theory has largely overlooked the role of verb meaning(s), generalizations in the form of constraints are still necessary to build a fully principled account of lexical-constructional fusion. Taking the family of resultative constructions as a case study, I employ the analytical tools (i.e. the apparatus of so-called internal and external constraints on constructional integration) developed by the Lexical Constructional Model in order to show that a fine-nuanced description can go hand in hand with the postulation of macro-generalizations.