The current article reports the first attempt to test the Implicit Relational Assessment Procedure (IRAP), as a group-based measure of natural verbal relations, using both response-latency and event-related potentials as dependent variables. On each trial of the IRAP, participants were presented with I of 2 attribute stimuli ("Pleasant" or "Unpleasant"), a positive (e.g., "Love") or negative (e.g., "Murder") target stimulus, and 2 relational terms, "Similar" and "Opposite," as response options. Participants were required to respond as quickly and accurately as possible across blocks of trials, with half of the blocks requiring responses that were deemed consistent (e.g., Pleasant-Love-Similar), and the other half inconsistent (e.g., Pleasant-Love- Opposite), with natural verbal relations. Shorter mean latencies were predicted for consistent than for inconsistent blocks. Two separate experiments supported this prediction. Event-related potentials, gathered during the second experiment, also proved to be sensitive to the IRAP, yielding more negative waveforms for inconsistent relative to consistent blocks of trials. A theoretical interpretation of the IRAP effect is offered, and important directions for future research are highlighted.