We challenge Folli and Harley’s (2020) account of Talmy’s (2000) typology, as it does not derive the unavailability of satellite-framed constructions in verb-framed languages, and it fails to predict the existence of weak satellite-framed languages like Latin and Slavic (Acedo-Matellán 2016). We propose an alternative approach based on Spanning (Svenonius 2016). Variation stems from the distribution of PF-interpretable features on the resultative head Res. Depending on the language, this head may or may not bear features triggering lexical access (w) and linearization (@). Conforming with Minimalist desiderata, our account captures a wider range of Talmyan patterns and produces some novel successful predictions.