Stimulus generalization and equivalence classes: a model for natural categories.

Fields et al. (1991) taught adults to match line lengths. First they learned A→B, then B→C. No one got money for the new mixes.

After the classes locked in, the team showed brand-new line lengths. They watched if picks stayed inside the class.

Every adult kept choosing the same class even with unseen lines. Generalization spread to look-alikes without extra teaching.

The result says equivalence classes act like natural categories. One example can stand for the whole group.

Ayres‐Pereira et al. (2018) stretched the idea to preschoolers. Kids first matched photos, then real 3-D toys. Only two of six passed the toy test, showing young learners may need more help.

Emmelkamp et al. (1986) warns that children without language may never build the classes at all. Check signs or speech before you start equivalence lessons.

Marin et al. (2024) adds a caution: lab-perfect classes can fall apart in noisy homes or classrooms. Probe in real settings before you trust the skill.

You can use this when teaching concepts like big/little or same/different. Train a few examples, then test with new items that look close. If the learner sorts them correctly, the class is solid. If not, add more exemplars or review the baseline relations. Always check generalization in the place where the skill must work.