Language users employ creative and innovative means to refer to novel concepts. One example is place-for-event metonymy as in “How many bands played at Woodstock?” where the place name is used to refer to an event. We capitalize on the observation that place-for-event metonymy can on the one hand result in the conventionalization of the event reading (as is the case for “Woodstock”) but on the other hand can also be relatively short-lived as a function of the socio-cultural or historical impact of the respective event (e.g., “Egypt” to refer to one of the sites of the Arab Spring). We use place-for-event metonymy as a test case to tap into discrete stages of conventionalization and compare the processing of the place and the event reading of particular expressions, with ratings of the degree of conventionalization as predictors. In an event-related potential (ERP) reading study, we observed a modulation of the Late Positivity between 500 and 750 ms post-onset by condition (event vs. place reading) and degree of conventionalization. The amplitude of the positivity was most pronounced for event readings with a low degree of conventionalization (similar to previous findings from ad-hoc metonymy). Interestingly, place readings with a high degree of (event) conventionalization also evoked a pronounced positivity. The Late Positivity is viewed to reflect processing demands during reconceptualization required for proper utterance interpretation. Overall, the data suggest that stages of meaning evolution are reflected in the underlying neurophysiological processes. Display omitted •Place-for-event metonymy comes with different degrees of conventionalization.•Late Positivity is modulated by condition and degree of conventionalization.•Non-conventional event readings of place names pattern with ad-hoc metonymy.•Place names with highly conventionalized events also show Late Positivity.•Stages of meaning evolution are reflected in neurophysiological signals.