It's really a method with just one enter, situation, and only one output, motion (or actions) a. You can find neither a individual reinforcement input nor an information input with the environment. The backpropagated value (secondary reinforcement) would be the emotion toward the consequence