The course of acquisition of a line-tilt discrimination by rhesus monkeys.
Fade stimulus differences a little at a time to cut errors and build sharper visual discriminations.
01Research in Context
What this study did
Researchers taught rhesus monkeys to tell a tilted line from a straight one. They used two ways: jump straight to a big tilt, or slide in tiny tilts step by step.
Each monkey sat in a box with two buttons. Press the button under the tilted line, get a food pellet. Wrong button, no food.
What they found
Monkeys who got the gentle stair-step of tilts learned faster and made fewer mistakes. When the tilt popped in all at once, errors shot up.
The slow group ended up spotting tiny tilts the fast group never mastered.
How this fits with other research
Pritchard et al. (1987) saw the same error-cut with rats. They added guided paws and fading lights, showing the rule works across species and gear.
Kodera et al. (1976) went deeper in pigeons. They proved fading moves control in two clear hops, not one smooth slide. So probe early to catch the hop.
Madsen et al. (1968) used pigeons and parallelograms. They found bimodal generalization—birds peaked at two orientations. The monkey study shows fading can sharpen that blur.
Why it matters
Your learner’s error history sticks. Start with easy differences they can win, then inch toward the tough one. Whether you’re teaching letter reversals or coin sizes, fade the gap, don’t jump it. Track each step—control can hop, not slide.
Want CEUs on This Topic?
The ABA Clubhouse has 60+ free CEUs — live every Wednesday. Ethics, supervision & clinical topics.
Join Free →Break your next visual discrimination into five tiny steps; introduce each step only after the learner hits 90%.
02At a glance
03Original abstract
Each of four groups of monkeys were trained on a different simultaneous discrimination procedure involving a vertical line as the correct choice. Each group, after acquiring the discrimination, was tested for generalization along the dimension of line tilt. Monkeys that learned to select the vertical line when the alternative choices were distinguished from the correct choice by two aspects (brightness and absence of line) showed almost complete tilt generalization (flat gradient). Monkeys that learned to select the line when the alternatives were distinguished only by the absence of the line showed poor tilt discrimination (generalization gradient slightly peaked at vertical). Monkeys developed a good tilt discrimination when nonvertical lines were gradually introduced by progressively darkening them on the previously blank alternatives. Monkeys developed a tilt discrimination with the lowest error rate when only horizontal alternatives were gradually introduced and then pairs of alternatives progressively closer to vertical were made available.
Journal of the experimental analysis of behavior, 1967 · doi:10.1901/jeab.1967.10-17