Transient effects of acquisition history on generalization in a matching-to-sample task.

Researchers taught 12 preschoolers to match pictures using three different methods. Some kids got step-by-step instructions. Others watched an adult do it first. The rest learned through shaping—getting praise only when they picked the right picture.

After each child learned the first task, the team flipped the rules. Now the 'correct' picture became wrong. They timed how fast each child adjusted to the new rule.

Kids taught by shaping caught on to the rule change about a large share faster. Children who got instructions took longest to drop the old rule. But after a few rounds, all groups performed the same.

The effect was small and faded quickly. Enough practice erased any early advantage.

Bruns et al. (2004) saw a similar fade-out when they tried listener-only training. Toddlers first sorted pictures after just hearing names, but the skill vanished without extra tact practice. Both studies show initial method matters less than repeated consequences.

Brown et al. (1994) went the other direction—proving pigeons can form equivalence classes too. That widens the basic process beyond children. The bird data fit here: once any learner has a solid reinforcement history, the original training path doesn't lock them in.

Paul (1983) found pigeons reacted to ratio requirements as cues, not just reward size. L et al. now show children also read subtle cue changes faster after shaping. Together they hint that contingency-shaped learning tunes attention to environmental shifts better than rule following.

If you run matching-to-sample or equivalence programs, don't worry about picking the 'perfect' first teaching style. Shaping may give a tiny head start when you later change contingencies, but plenty of reinforced practice washes the difference away. Focus on clear feedback and enough trials rather than the initial path into the task.