Translated from the Portuguese original. Read the original (Português)
Some reinforcers work from the start, without any learning: food, water, warmth. These are unconditioned. Others only start to work after appearing together with reinforcers that already worked — praise, the token, money, the click of the clicker. These are conditioned. And when a conditioned reinforcer is paired with many different reinforcers, it becomes generalized: it works in many more situations, because its effectiveness isn't tied to one specific deprivation — it doesn't depend on the person being hungry, thirsty, or tired at that moment. Much of what moves human behavior day to day is a conditioned reinforcer — and, because it's learned, it can be taught.
In Cooper, Heron & Heward (Applied Behavior Analysis), an unconditioned reinforcer (or primary reinforcer) is a stimulus change that strengthens behavior without needing any prior pairing with another reinforcer — it is a product of the species' history. A conditioned reinforcer (or secondary reinforcer) is a previously neutral stimulus that acquires a reinforcing function by having been paired with an unconditioned reinforcer, or with another already-established conditioned reinforcer — it is a product of each person's life history. And a generalized conditioned reinforcer is one that has been paired with many different reinforcers, so that its effectiveness no longer depends on any one specific deprivation condition at the moment.
Two paths for a stimulus to become a reinforcer
In Behavior Analysis, nothing is a reinforcer by nature or by good intention: whatever strengthens the behavior that produced it is a reinforcer — that is a matter of effect, verified in the data, as explained in positive reinforcement. The question here is different: where does that power to reinforce come from.
- Through the species' history. Food, water, oxygen, warmth, sensory stimulation: they work from the very beginning of life, with no training, because over the course of evolution organisms sensitive to these consequences survived and left descendants. This is the unconditioned reinforcer.
- Through life history. The sound of a coin, the click of a clicker, a little star sticker in a notebook: none of these had a reinforcing function before it was learned. It came to reinforce because it showed up together, many times, with something that already reinforced. This is the conditioned reinforcer.
The second path is, by far, the more important one for understanding adult human life. No one goes to work because the salary itself feeds them — money is a piece of paper or a number on a screen. It moves behavior because it is linked to everything it can buy.
The three types, side by side
Tap each tab to see what defines each type, what weakens it, and where it shows up:
Works without ever having been taught
Food, water, warmth, sensory stimulation. They did not need to be paired with anything: sensitivity to these consequences is part of the species' biological makeup. There are few of them, and they share an important practical limit — they depend heavily on the moment. Food reinforces someone who is hungry; right after lunch, the same cookie loses much of its effect. That back-and-forth is the subject of motivating operations: deprivation raises a reinforcer's value, satiation lowers it.
Learned by pairing, and sustained by it
Praise, attention, a grade on a test, the click of a clicker, the sound of a bag of pet food being opened. They gained their reinforcing function because they repeatedly showed up together with something that already reinforced — a learning process based on the association between stimuli, close to respondent conditioning. The trade-off is that this function fades if the backup disappears for good: a conditioned stimulus presented over and over with nothing behind it goes back to being just a sound, just a word.
Paired with a lot, which is why it works in more situations
Generalized reinforcers are a special case of conditioned reinforcer — not a separate third kind. Money and tokens are the textbook examples; attention, approval, and affection are the social ones. What sets them apart is not strength, but the variety of reinforcers they have been paired with. Since a token is worth a snack, a toy, and time at the park, it keeps being worth something right after the child has just had a snack — its effectiveness is not held hostage to one specific deprivation. That is exactly why the generalized reinforcer is so practical for anyone teaching: it works during more hours of the day and in more contexts.
How a new reinforcer is taught
Because the conditioned reinforcer is learned, it can also be deliberately built — and this is a common teaching procedure in early intervention. The principle is simple: present the stimulus you want to turn into a reinforcer immediately before delivering something that already reinforces, many times in a row.
- Praise and attention. The adult says "that's it!" and, right after, delivers the preferred item. With repetition, "that's it!" comes to sustain behavior on its own.
- The click of the clicker. A short, always-identical sound is paired with the reinforcer. Because the click can be delivered at the exact instant of the response, it marks what is being reinforced with a precision that delivering the item itself could not allow.
- Tokens. Each token is traded, at first, almost immediately for a backup reinforcer; later the exchange gets more spaced out, as the token gains its own value.
A point that is easy to lose: pairing is not a one-time event, it is maintenance. Conditioned reinforcers need to keep being backed up, from time to time, by the backup reinforcers. That is why a token economy whose exchange store "closes" tends to stop working: the token stops being worth anything.
Praise doesn't come with a guaranteeIt is tempting to assume that "everyone likes praise." That is not how it works: for some children — and also for some adults — praise and social attention have not yet acquired a reinforcing function, and insisting on them only produces frustration on both sides. Whether or not this is a reinforcer for that particular person is an empirical question, answered by preference assessment and by the effect observed in behavior. When praise does not work yet, it becomes a teaching target — something to be established through pairing — not a requirement.
Why the generalized reinforcer is so powerful
Skinner, in Science and Human Behavior (1953), drew attention to a small set of generalized reinforcers that sustains a huge part of human behavior: attention, approval, affection, the submission of others, and tokens — including money. What they all have in common is having been paired with a wide range of other consequences over the course of a lifetime.
From this come two practical advantages:
- They resist satiation. Since they are not tied to one specific deprivation, they stay effective throughout the day — no one "gets full" on money after lunch.
- They allow delivery to be delayed. The token can be given right now, at the exact moment of the behavior, and traded in later. This solves a real problem for anyone teaching: the reinforcer that actually matters is almost never at hand at the right time.
There is also a quieter role, and perhaps the most pervasive one: within a sequence of steps, each completed step produces a change in the environment that works at the same time as a conditioned reinforcer for the response that just occurred and as a signal for the next response. It is this dual function that holds an entire behavior chain together — brushing your teeth, getting dressed, making a snack — without needing to reinforce each step with something from outside.
Tokens in the lab, since the 1930sThe idea is not recent or improvised. In classic studies at the Yerkes Laboratories, Wolfe (1936) and Cowles (1937) showed that chimpanzees worked to obtain coin-like tokens, later exchangeable for food — and that the tokens kept up responding even before the exchange. Kelleher picked up this line of research in the 1950s, and the experimental literature accumulated since then was reviewed by Hackenberg (2009) in the Journal of the Experimental Analysis of Behavior.
Spotting it in everyday life
Tap the cards to see which category each example belongs to — and why:
"Conditioned" doesn't mean artificialCalling a reinforcer conditioned does not make it any less legitimate or less powerful than food. A hug from someone we love, a coworker's recognition, the satisfaction of seeing a job finished — much of what gives meaning to human daily life is a conditioned reinforcer. The label describes the origin of the function (learned, not inherited), not its value.
Test your understanding
Quick quiz
Three questions to lock in the concept. You'll see the explanation right away.
Further reading
- Cooper J. O., Heron T. E. & Heward W. L. Applied Behavior Analysis. 3rd ed. Pearson, 2020 (chapters on positive reinforcement and chaining).
- Skinner B. F. Science and Human Behavior. Macmillan, 1953 (generalized reinforcers: attention, approval, affection, submission, and tokens).
- Hackenberg T. D. “Token reinforcement: a review and analysis”. Journal of the Experimental Analysis of Behavior, 91(2), 2009.
- Ayllon T. & Azrin N. H. The Token Economy: A Motivational System for Therapy and Rehabilitation. Appleton-Century-Crofts, 1968.
This entry is educational and does not replace assessment, diagnosis or treatment by a qualified health professional. Every person is unique — for a specific case, seek individual guidance.