We move through environments and interact with objects constantly. The scenes we interact with are thereby not separated from us; we are agents in the world we experience. In line with this notion, scenes are suggested to be best categorized by the actions they afford: their functions or affordances (Greene et al., 2016). However, scenes are not holistic entities; they can be separated into meaningful clusters (“phrases”) consisting of an anchor object around which other objects are located (e.g., soap positioned on top of the sink; Vo, 2021). Anchor objects are important for visual search, but their role in affordance understanding is unclear. In the current study, we explore the influence of action related and unrelated anchors on scene affordance understanding. For a given target action (e.g., taking a shower), we presented indoor scenes in which we either masked an action related anchor (shower), an action unrelated anchor (sink), or a random non-anchor object (mirror). Participants then performed a lexical decision task on pseudo-words vs. action-words, the latter of which could be either consistent (a bathroom and the action showering) or inconsistent (a living room and the action showering) with the presented stimulus. Reaction times were fastest when both anchors were visible and the stimulus was consistent with the word. The contrast between conditions in which action related and unrelated anchors were masked was not significant. The fact that an anchor’s absence impacts the lexical access to action words suggests that anchor objects impact scene affordance perception and understanding.