Two experiments were designed to replicate and extend previous findings on the transformation of avoidance response functions in accordance with the relational frames of Same and Opposite. Participants were first exposed to nonarbitrary and arbitrary relational training and testing. Next, during avoidance conditioning, one stimulus from the relational network signaled a simple avoidance response that cancelled a scheduled presentation of an aversive image and sound. The majority of participants who met the criteria for conditioned avoidance also demonstrated derived avoidance. Experiment 1 showed that detailed instructions were not necessary for derived transformation to occur. Experiment 2 showed that more complex patterns of transformation may emerge when another stimulus from the relational network signaled the avoidance response. Implications for understanding clinically significant avoidance behavior are discussed.