Negative reinforcement vs punishment, and the word Skinner changed his mind about

Negative reinforcement and punishment have opposite effects: negative reinforcement makes a behaviour more likely by removing something unpleasant, and punishment makes a behaviour less likely. The word “negative” does not mean bad. It means something was taken away. If the confusion still feels natural, there is a reason for it: in 1938 B. F. Skinner himself used “negative reinforcement” for what is now called punishment, and changed the definition fifteen years later.

• Reinforcement always increases a behaviour; punishment always decreases it. “Positive” means a stimulus is added, “negative” means one is removed.
• That gives four combinations: positive reinforcement, negative reinforcement, positive punishment and negative punishment.
• Two questions sort any example: did the behaviour become more or less frequent afterwards, and was something added or removed?
• The categories are defined by the effect on behaviour, not by what anyone intended. A “punishment” that makes a behaviour more frequent is a reinforcer.
• Skinner’s 1938 book used “negative reinforcement” for punishing stimuli. His 1953 book reversed it, and behaviour analysts have debated since 1975 whether the positive/negative split is needed at all.

Negative reinforcement vs punishment: the short answer

Negative reinforcement is not a kind of punishment. It is a kind of reinforcement, and reinforcement, by definition, strengthens the behaviour it follows. What makes it “negative” is how it strengthens: by ending or removing something unpleasant. The open textbook Psychology 2e from OpenStax defines it as a case where “an undesirable stimulus is removed to increase a behavior”.

Punishment does the opposite job. It weakens the behaviour it follows, either by adding something unpleasant or by taking away something pleasant.

The standard example is a car’s seatbelt chime. The chime is annoying, fastening the seatbelt stops it, and so you fasten the seatbelt sooner next time. Something unpleasant was removed, and the behaviour that removed it became more frequent. That is negative reinforcement. Nobody was punished.

The four quadrants of operant conditioning

Reinforcement and punishment belong to operant conditioning, the term Skinner introduced in 1937 for behaviour shaped by its consequences. Crossing the two effects with the two operations gives four terms, and every one of them is defined by two things at once.

Read the table by row first. The row tells you whether the consequence strengthened or weakened the behaviour, which is the part that matters most. The column only tells you how it did it.

A two-question test that sorts any example

The order matters. Ask the questions in this sequence:

1. 1.Did the behaviour become more frequent or less frequent afterwards? More means reinforcement, less means punishment.
2. 2.Was a stimulus added or removed? Added means positive, removed means negative.

The first question does most of the work. Once you know a consequence is a punishment, it cannot be negative reinforcement, whatever was removed.

The last row is deliberately awkward. A fine can be described as adding a penalty or as removing money, so the second question has no clean answer. The first question still does: if speeding becomes less frequent, the fine is a punishment. That is the practical reason to ask about the behaviour first.

Is time-out negative reinforcement or punishment?

Time-out is usually classed as negative punishment: it removes access to attention, play and other things the child finds rewarding, with the aim of making the behaviour less frequent. But the definition is functional, not a matter of intent. Punishment is defined by whether the behaviour actually decreases, as the Wikipedia entry on punishment in psychology summarises. If a child is sent out of a task they wanted to escape, and the behaviour that got them sent out becomes more frequent, the time-out has worked as negative reinforcement instead. The label follows the result.

Negative reinforcement vs negative punishment

This is the pair people actually mix up, because both involve removing something. The difference is what gets removed, and in which direction the behaviour moves.

• Negative reinforcement removes something unpleasant, and the behaviour increases. The headache ends, so you reach for the painkiller again next time.
• Negative punishment removes something pleasant, and the behaviour decreases. The phone is confiscated, so the late-night scrolling stops.

A quick way to hold them apart: in negative reinforcement, the person is relieved by the removal. In negative punishment, they lose something they wanted.

The other pair Google’s autocomplete shows people confusing is negative reinforcement vs positive punishment. Both involve an unpleasant stimulus, which is where the confusion starts. In positive punishment the unpleasant thing arrives after the behaviour; in negative reinforcement the behaviour makes it go away. They point in opposite directions.

Escape and avoidance: the two forms of negative reinforcement

Negative reinforcement comes in two forms, and they are worth naming because examples often mix them.

• Escape: the behaviour ends an unpleasant stimulus that is already present. Buckling up to stop the chime is escape.
• Avoidance: the behaviour prevents or postpones an unpleasant stimulus before it starts. Buckling up before the chime ever sounds, because you know it is coming, is avoidance.

Both strengthen the behaviour, so both are negative reinforcement.

Escape is also the procedure behind one of psychology’s best-known experiments. In Seligman and Maier’s 1967 experiment on learned helplessness, dogs could end a shock by stepping over a low barrier in a shuttle box. The dogs that had earlier been given shocks they could not control largely failed to learn that escape response, even though it was available.

Why everyone mixes them up: the definition Skinner reversed

The usual explanation for the confusion is that students hear “negative” as “bad”. That is true, but it is not the whole story. For fifteen years, the technical meaning of “negative reinforcement” was punishment.

In The Behavior of Organisms (1938), Skinner divided reinforcing stimuli into two kinds, as Jack Michael documented in a 1975 paper that quotes both of Skinner’s books side by side:

The negative kind was shock after lever pressing, a stimulus whose presentation made the behaviour less frequent. Skinner described one of his own studies, the so-called bar-slap experiment, as showing “the effect of negative reinforcement upon extinction”. In today’s terms that is positive punishment.

In Science and Human Behavior (1953), he changed the definition. Reinforcers that add something, such as food or water, became positive. Reinforcers that remove something, such as a loud noise, a bright light, extreme heat or cold, or electric shock, became negative. And both, he wrote on page 73, work the same way: “In both cases the effect of reinforcement is the same — the probability of response is increased.” Punishment moved to its own category. On pages 184–185 Skinner defined it as withdrawing a positive reinforcer or presenting a negative one: “An example of the former would be taking candy from a baby; an example of the latter, spanking a baby.”

So the everyday reading, negative reinforcement equals punishment, is not an invention of confused students. It is Skinner’s 1938 usage, still circulating. Michael was blunt about what the switch cost:

He also pointed to the second trap, which is built into ordinary language. Reward and punishment sound like opposites, so, as he put it, “if reward is positive reinforcement isn’t it reasonable that punishment should be negative reinforcement?” And “reinforce” means strengthen, so “negative reinforcement” sounds like negative strengthening, which is to say weakening.

The confusion is documented well beyond one paper. Lilienfeld and colleagues included the pair in a 2017 compendium of 50 frequently confused term pairs in psychology, citing a 1988 paper by Tauber on its “ubiquitous confusion by generations of undergraduate psychology students”. It is the textbook case of the problem described in our guide to memorising psychology terminology: an everyday word, “negative”, already installed with the wrong meaning, sitting next to a near-synonym it has to be told apart from.

Should the positive/negative distinction exist at all?

Michael’s 1975 paper did not only diagnose the confusion. It proposed a remedy: call every strengthening consequence reinforcement and every weakening one punishment, and stop dividing reinforcement by whether something was presented or removed. In his words, that distinction “simply can be dropped”.

Some summaries online report the opposite, that Michael concluded the distinction is necessary. They quote a sentence from the paper in which he sets out the position he goes on to reject. His conclusion is the one above.

The argument was revisited thirty years later in The Behavior Analyst. Alan Baron and Mark Galizio asked in 2005 whether the distinction should be preserved, restating Michael’s case that the terms are confusing and ambiguous. Replies followed in 2006. Brian Iwata argued for keeping the distinction for practical reasons in applied work, while acknowledging that some changes, such as a shift in temperature or in shock intensity, can be described as either adding or removing. Philip Chase argued that the distinction is useful mainly as a teaching tool rather than as a marker of two different processes.

In practice the debate is unresolved, and introductory textbooks still teach all four terms. The ambiguous cases are a real reason experts disagree; they are not a reason to treat the distinction as wrong. For a student, the useful conclusion is narrower: the reinforcement/punishment split is the solid one, and the positive/negative split is the one to handle with care.

Extinction is a third thing

A third term often gets pulled into the same comparison. Extinction is what happens when a behaviour that used to be reinforced stops being reinforced, and becomes less frequent as a result. Nothing is added and nothing is taken away as a consequence of the behaviour; the consequence simply stops arriving.

That makes extinction neither reinforcement nor punishment. It lowers the behaviour, like punishment, but by withholding a reinforcer rather than applying a consequence. It is the answer to “negative reinforcement vs punishment vs extinction”: one strengthens, one weakens by consequence, one weakens by absence.

Where the terms come from

The idea is older than the vocabulary. Edward Thorndike’s law of effect, set out in Animal Intelligence in 1911, held that responses followed by “satisfaction” become more firmly connected with the situation, and responses followed by “discomfort” have those connections weakened. Skinner turned that into operant conditioning, moved the definitions from feelings to observable effects on behaviour, and in 1953 supplied the definitions the four modern terms rest on. The same behavioural vocabulary is the ancestry that CBT and several related therapies still carry in their names and methods.

How to keep the four terms apart

Four terms that differ by one word each are exactly the kind that fade when they are reread and stick when they are retrieved. The four quadrants are also a ready-made list: in Psychology World you can save them to a list of their own and drill them in the multiple-choice mode, which needs at least four words in a list, or in the typing mode, which will not accept an answer you merely recognise. What each study mode does is set out on the home page. Whichever tool you use, test the pair you confuse most in the same sitting, not on separate days.