An analysis of concept learning: simple conceptual control and definition-based conceptual control.
Treat concept learning as a chain of tiny stimulus-response steps, not one big insight.
01Research in Context
What this study did
Lalli et al. (1995) wrote a theory paper. They asked how people learn concepts through definitions. They broke it into two kinds of control: simple and definition-based.
They said both kinds are just chains of stimulus-response links. No inner pictures needed. They gave rules for building these chains when you teach.
What they found
The paper does not give new data. It gives a teaching recipe. Show a rational set of examples and non-examples. Present the next item right after the learner's last error.
The chain grows step-by-step. Each step adds either a critical or a variable feature. The final link produces the correct concept name.
How this fits with other research
Marr (1989) wanted Newton-style math for all behavior. S et al. keep the math spirit but aim it at concept learning. Both stay at the white-board level.
Fahmie et al. (2013) showed rats shift response bias fast but slow to see the real stimulus. The rat data fit the chaining view: early links change before later ones.
Wilkie (1973) swapped contiguity for correlation in the law of effect. S et al. make a similar swap for concepts: no single "aha" moment, just correlated links stacking up.
Why it matters
Next time you run a concept program, stop asking "Does the child get it?" Ask "Which link is missing?" Pick examples that differ by one feature at a time. Place the next trial right after an error. You are building a chain, not flipping a light bulb.
Want CEUs on This Topic?
The ABA Clubhouse has 60+ free CEUs — live every Wednesday. Ethics, supervision & clinical topics.
Join Free →After an error, present the next example within three seconds and change only one relevant feature.
02At a glance
03Original abstract
Concept learning can involve either contingency shaping of stimulus-class discriminations or the application of definitions of the concepts. Experimental behavior analysts have studied contingency shaping, whereas educational psychologists have studied definitional concept training. In this paper, we analyze definition-based concept learning in terms of stimulus-response chains. Then we apply this chaining analysis to principles of instruction proposed by educational psychologists. These principles include (a) stating the definition in terms of critical and variable attributes, (b) using examples and nonexamples, (c) using a rational set of examples and nonexamples, (d) presenting coordinate concepts simultaneously, and (e) presenting the next instance based on the learner's previous error.
The Analysis of verbal behavior, 1995 · doi:10.1007/BF03392898