We propose quantitative experimental approaches to the question of whether positive and negative reinforcement are functionally different, and discuss scientific and ethical concerns that would arise if these approaches were pursued.