Inhibitory control and errorless discrimination learning.

The team tested Terrace’s fading procedure. This method slowly changes the wrong choice until the learner almost never picks it.

They wanted to see if the wrong picture stays neutral. In other words, does the learner still learn to avoid it?

Even with the gentle fading, the wrong picture gained “don’t pick” power. The bird still treated it as a stop signal.

Earlier work showed flat response curves and called the wrong picture neutral. The new data say those flat curves were just floor effects—very low response rates that hid the underlying inhibition.

Fantino (1968) and Barlow et al. (1973) reported near-zero errors with fading and praised the method. Aragona et al. (1975) agree errors drop, but add the warning: the wrong stimulus is no longer neutral.

Frederiksen et al. (1978) later found fewer reversal errors after errorless training, seeming to clash with the negative finding. The difference is focus—W counted errors during reversals, while J looked at hidden inhibition after learning.

Mulder et al. (2020) still call fading the go-to tactic for learners with ID/DD. The 1975 paper does not overturn the tool; it simply tells us the tool leaves a mark we must plan for.

You can keep using fading to cut mistakes, but do not assume the wrong card, photo, or word stays harmless. Check for inhibitory control in probe tests and program plenty of review trials so the “don’t pick” strength does not block later learning.