Nonhumans have not yet shown stimulus equivalence.

Hayes (1989) looked at every paper that said animals show stimulus equivalence.

The author checked each one against Sidman’s three rules: reflexivity, symmetry, and transitivity.

No study passed the test. All left out at least one rule or only showed trained links.

The review found zero clean demos of true equivalence in non-humans.

Monkeys, pigeons, and rats solved tasks, but only after direct teaching of every link.

Without untaught, derived links, the animals did not meet the definition.

O'Reilly et al. (2004) seems to fight this view. Crows showed transitive choices after size-ordered feedback. The birds picked the middle cup when they had never been taught that pair.

The fight is only skin-deep. Hayes (1989) asked for full equivalence classes. F et al. gave one derived relation tied to a visible line. That is a step, not the whole ladder.

Bromley et al. (1998) backs C’s style. Both papers blame weak methods, not the species, for past null results.

When you test for equivalence with clients, check all three parts. If a learner only passes taught trials, you have strong links, not an equivalence class. Add probes for untaught pairs before you claim the skill is yours.