In this study we examined asymmetric semantic activation patterns as people listened to conversations and narratives that promoted causal inferences. Based on the hypothesis that understanding the unique features of conversational input may benefit from or require a modified pattern of conceptual activation during conversation, we compared semantic priming in both hemispheres for inferences embedded in conversations and in narratives. Participants named inference-related target words or unrelated words presented to the left visual field-right hemisphere (lvf-RH) or to the right visual field-left hemisphere (rvf-LH) at critical coherence points that required an inference in order to correctly understand an utterance in the context of the conversation or narrative. Fifty-seven undergraduates listened to 36 conversations or narratives and were tested at 100 target inference points. During narrative comprehension, inference-related priming was reliable and equally strong in both hemispheres. In contrast, during conversation comprehension, inference-related priming was only reliable for target words presented to lvf-RH. This work demonstrates that priming for inference-related concepts can be measured with input in conversational form and suggests the language processing style of the RH is advantageous for comprehending conversation. (Contains 1 figure and 2 tables.)