Do stimulus classes exist before they are tested?

The authors asked a simple question: do stimulus classes exist before we test them?

They wrote a conceptual paper, not an experiment.

Their goal was to stop people from treating stimulus classes like invisible boxes in the brain.

They found that classes are not things; they are patterns of behavior we produce during training and testing.

Talking about classes as if they pre-exist makes us sound like we are doing mentalism, not behavior analysis.

Vollmer et al. (1996) extends this idea. They show that when you teach a child to name pictures, the child later matches new pictures without extra training. The class is created by the naming procedure, not discovered.

Critchfield (2018) also extends the point. He tells graduate programs to teach stimulus-relation engineering so BCBAs can build emergent skills instead of drilling every single target.

Lattal (1984) is a predecessor. He urged us to explain stimulus control with plain associative learning, six years before the target paper warned us not to reify classes. The two papers book-end the same debate.

Stop saying "the child has a stimulus class." Say "the child responds by matching and naming." This keeps your language behavioral and your interventions lean. When you write protocols, describe the exact training and test steps you will use to build the performances you want.