The operant-respondent distinction: Future directions.
The operant-respondent split still works if you loosen the seams for autoshaping and avoidance.
01Research in Context
What this study did
Pear et al. (1984) wrote a think-piece, not an experiment.
They looked at cracks in the operant-respondent wall.
Autoshaping and avoidance data were bending the rules.
What they found
The old two-box system still sorts most behavior.
Some birds peck because food predicts light, not because food follows pecking.
The authors say: keep the boxes, but widen the lids.
How this fits with other research
Craig (2023) makes the same move for behavioral momentum.
Both papers tell us: patch the theory, don’t junk it.
Branch et al. (1980) saw the field drifting toward mental talk; J et al. answered by fixing the map, not the language.
Why it matters
When your client’s avoidance looks like both reflex and reward, you no longer have to pick sides.
Use the operant label for what you can reinforce, the respondent label for what you can’t, and write a note when the case straddles the line.
This keeps your treatment plan clean and your data honest.
Want CEUs on This Topic?
The ABA Clubhouse has 60+ free CEUs — live every Wednesday. Ethics, supervision & clinical topics.
Join Free →During prep, tag each target behavior as O, R, or O/R so your team knows which procedure to run.
02At a glance
03Original abstract
The operant-respondent distinction has provided a major organizing framework for the data generated through the experimental analysis of behavior. Problems have been encountered, however, in using it as an explanatory concept for such phenomena as avoidance and conditioned suppression. Data now exist that do not fit neatly into the framework. Moreover, the discovery of autoshaping has highlighted difficulties in isolating the two types of behavior and conditioning. Despite these problems, the operant-respondent framework remains the most successful paradigm currently available for organizing behavioral data. Research and theoretical efforts should therefore probably be directed to modifying the framework to account for disparate data.
Journal of the experimental analysis of behavior, 1984 · doi:10.1901/jeab.1984.42-453