Languages differ in how events are described, but little is known about how language interacts with online event processing. To explore this question we examined placement events in Swedish. Swedish has three obligatory placement verbs, sätta, ‘set’, ställa ‘stand’, lägga, ‘lay’, and lacks a superordinate general term like English put (Gullberg & Burenhult, 2011; Viberg, 1999). Every placement event in Swedish must be labelled by one of the three verbs, whose choice depends on object properties, and the object's relationship to the ground. The current study investigates how sensitive Swedes are to the relationship between event properties and verb labels. Native speakers (N = 20, 18-35-years) watched images of a hand placing an object on a table followed by visually presented sentences that were either congruent or incongruent with the images while event-related potentials were recorded and time-locked to the placement verbs. We varied object properties such as ± base (e.g., glass/orange), spatial extension (e.g., tall/short glass), and orientation (vertical/horizontal). The three verbs were combined with each image in a cross-subject design. The results showed that, as expected, incongruent picture-verb combinations elicited an increased centro-medial N400 modulated by verb appropriateness. Congruent picture-verb combinations also elicited an N400 when objects were placed in non-canonical positions (e.g. laying a glass on its side), suggesting that native placement event processing may depend on an interaction between actual and canonical event properties and language. A follow up study presenting auditory sentences simultaneously with images will explore this hypothesis further.