In operant conditioning, punishment is any change in a human or animal's surroundings that occurs after a given behavior or response which reduces the likelihood of that behavior occurring again in the future. As with reinforcement, it is the behavior, not the animal, that is punished. Whether a change is or is not punishing is determined by its effect on the rate that the behavior occurs, not by any "hostile" or aversive features of the change. For example, a painful stimulus which would act as a punisher for most people may actually reinforce some behaviors of masochistic individuals.
There are two types of punishment in operant conditioning:
- positive punishment, punishment by application, or type I punishment, an experimenter punishes a response by presenting an aversive stimulus into the animal's surroundings (a brief electric shock, for example).
- negative punishment, punishment by removal, or type II punishment, a valued, appetitive stimulus is removed (as in the removal of a feeding dish). As with reinforcement, it is not usually necessary to speak of positive and negative in regard to punishment.
Punishment is not a mirror effect of reinforcement. In experiments with laboratory animals and studies with children, punishment decreases the likelihood of a previously reinforced response only temporarily, and it can produce other "emotional" behavior (wing-flapping in pigeons, for example) and physiological changes (increased heart rate, for example) that have no clear equivalents in reinforcement.
Punishment is considered by some behavioral psychologists to be a "primary process" – a completely independent phenomenon of learning, distinct from reinforcement. Others see it as a category of negative reinforcement, creating a situation in which any punishment-avoiding behavior (even standing still) is reinforced.
Positive punishment occurs when a response produces a stimulus and that response decreases in probability in the future in similar circumstances.
- Example: A mother yells at a child when he or she runs into the street. If the child stops running into the street, the yelling ceases. The yelling acts as positive punishment because the mother presents (adds) an unpleasant stimulus in the form of yelling.
- Example: A barefoot person walks onto a hot asphalt surface, creating pain, a positive punishment. When the person leaves the asphalt, the pain subsides. The pain acts as positive punishment because it is the addition of an unpleasant stimulus that reduces the future likelihood of the person walking barefoot on a hot surface.
Negative punishment occurs when a response produces the removal of a stimulus and that response decreases in probability in the future in similar circumstances.
- Example: A teenager comes home after curfew and the parents take away a privilege, such as cell phone usage. If the frequency of the child coming home late decreases, the privilege is gradually restored. The removal of the phone is negative punishment because the parents are taking away a pleasant stimulus (the phone) and motivating the child to return home earlier.
- Example: A child throws a temper tantrum because he wants ice cream. His mother subsequently ignores him, making it less likely the child will throw a temper tantrum in the future when he wants something. The removal of attention from his mother is a negative punishment because a pleasant stimulus (attention) is taken away.
Simply put, reinforcers serve to increase behaviors whereas punishers serve to decrease behaviors; thus, positive reinforcers are stimuli that the subject will work to attain, and negative reinforcers are stimuli that the subject will work to be rid of or to end. The table below illustrates the adding and subtracting of stimuli (pleasant or aversive) in relation to reinforcement vs. punishment.
|Rewarding (pleasant) stimulus||Aversive (unpleasant) stimulus|
|Adding/Presenting||Positive Reinforcement||Positive Punishment|
|Removing/Taking Away||Negative Punishment||Negative Reinforcement|
Aversive stimulus, punisher, and punishing stimulus are somewhat synonymous. Punishment may be used for (a) an aversive stimulus or (b) the occurrence of any punishing change or (c) the part of an experiment in which a particular response is punished. However, some things considered aversive can become reinforcing. In addition, some things that are aversive may not be punishing if accompanying changes are reinforcing. A classic example would be mis-behavior that is 'punished' by a teacher but actually increases over time due to the reinforcing effects of attention on the student.
Primary versus secondary
Pain, loud noises, foul tastes, bright lights, and exclusion are all things that would pass the "caveman test" as an aversive stimulus, and are therefore primary punishers. The sound of someone booing, the wrong-answer buzzer on a game show, and a ticket on your car windshield are all things you have learned to think about as negative, and are considered secondary punishers.
Contrary to suggestions by Skinner and others that punishment typically has weak or impermanent effects, a large body of research has shown that it can have a powerful and lasting effect in suppressing the punished behavior. Furthermore, more severe punishments are more effective, and very severe ones may even produce complete suppression. However, it may also have powerful and lasting side effects. For example, an aversive stimulus used to punish a particular behavior may also elicit a strong emotional response that may suppress unpunished behavior and become associated with situational stimuli through classical conditioning. Such side effects suggest caution and restraint in the use of punishment to modify behavior. (Further reading: Ayotte, R.; Muster, H.; Morais, F.; et al. "Positive and negative reinforcement and punishment effectiveness". Cognitive Sciences Stack Exchange. Retrieved 10 May 2017.)
Importance of contingency and contiguity
One variable affecting punishment is contingency, which is defined as the dependency of events. A behavior may be dependent on a stimulus or dependent on a response. The purpose of punishment is to reduce a behavior, and the degree to which punishment is effective in reducing a targeted behavior is dependent on the relationship between the behavior and a punishment. For example, if a rat receives an aversive stimulus, such as a shock each time it presses a lever, then it is clear that contingency occurs between lever pressing and shock. In this case, the punisher (shock) is contingent upon the appearance of the behavior (lever pressing). Punishment is most effective when contingency is present between a behavior and a punisher. A second variable affecting punishment is contiguity, which is the closeness of events in time and/or space. Contiguity is important to reducing behavior because the longer the time interval between an unwanted behavior and a punishing effect, the less effective the punishment will be. One major problem with a time delay between a behavior and a punishment is that other behaviors may present during that time delay. The subject may then associate the punishment given with the unintended behaviors, and thus suppressing those behaviors instead of the targeted behavior. Therefore, immediate punishment is more effective in reducing a targeted behavior than a delayed punishment would be. However, there may ways to improve the effectiveness of delayed punishment, such as providing verbal explanation, reenacting the behavior, increasing punishment intensity, or other methods.
Applied behavior analysis
Punishment is sometimes used for treatment programs in applied behavior analysis in the most extreme cases, to reduce dangerous behaviors such as head banging or biting exhibited most commonly by children or people with special needs or disabilities. Punishment is considered one of the ethical challenges to autism treatment and is one of the major reasons for discussion of professionalizing behavior analysis. Professionalizing behavior analysis through licensure would create a board to ensure that consumers or families had a place to air disputes. (see Professional practice of behavior analysis)
Controversy regarding ABA persists in the autism community. A 2017 study found that 46% of people with autism spectrum undergoing ABA met the criteria for post-traumatic stress disorder (PTSD), and the risk of developing it as a result of ABA is more than 86% higher in people with autism spectrum than in non-autistic people. According to the researchers, the risk of PTSD was increased when ABA was used, regardless of the age of the patient (adults and children were examined). However, the quality of this study is disputed.
- Positive reinforcement: includes praise, superficial charm, superficial sympathy (crocodile tears), excessive apologizing, money, approval, gifts, attention, facial expressions such as a forced laugh or smile, and public recognition.
- Negative reinforcement: may involve removing one from a negative situation
- Intermittent or partial reinforcement: Partial or intermittent negative reinforcement can create an effective climate of fear and doubt. Partial or intermittent positive reinforcement can encourage the victim to persist - for example in most forms of gambling, the gambler is likely to win now and again but still lose money overall.
- Punishment: includes nagging, yelling, the silent treatment, intimidation, threats, swearing, emotional blackmail, the guilt trip, sulking, crying, and playing the victim.
- Traumatic one-trial learning: using verbal abuse, explosive anger, or other intimidating behavior to establish dominance or superiority; even one incident of such behavior can condition or train victims to avoid upsetting, confronting or contradicting the manipulator.
Traumatic bonding occurs as the result of ongoing cycles of abuse in which the intermittent reinforcement of reward and punishment creates powerful emotional bonds that are resistant to change.
- D'Amato, M. R. (1969). Melvin H. Marx (ed.). Learning Processes: Instrumental Conditioning. Toronto: The Macmillan Company.
- Solnick, J. V., Rincover, A. and Peterson, C. R. (1977), Some Determinants Of the Reinforcing and Punishing Effects of Timeout. Journal of Applied Behavior Analysis, 10: 415-424. doi:10.1901/jaba.1977.10-415
- Skinner, B. F. "Science and Human Behavior" (1953)McMIllan, New York
- Solomon, R. L. (1964). "Punishment." American Psychologist, 19(4), 239-253.
- Azrin, N. H. (1960). Effects of punishment intensity during variable-interval reinforcement. Journal of the Experimental Analysis of Behavior, 3(2), 123–142.
- Schwartz, B, Wasserman, E. A., & Robbins, S. J. "Psychology of Learning and Behavior" (5th Ed) (2002) Norton, New York
- Meindl, J. N., & Casey, L. B. (2012). Increasing the suppressive effect of delayed punishers: A review of basic and applied literature. Behavioral Interventions, 27(3), 129–150. https://doi.org/10.1002/bin.1341
- Henny Kupferstein. Evidence of increased PTSD symptoms in autism exposed to applied behavior analysis . "Advances in Autism". 4 (1), pp. 19-29, 2018.
- Justin Barrett Leaf. Evaluating Kupferstein’s claims of the relationship of behavioral intervention to PTSS for individuals with autism . "Advances in autism". 4 (3), pp. 122-129, 2018.
- Braiker, Harriet B. (2004). Who's Pulling Your Strings ? How to Break The Cycle of Manipulation. ISBN 0-07-144672-9.
- Dutton; Painter (1981). "Traumatic Bonding: The development of emotional attachments in battered women and other relationships of intermittent abuse". Victimology: An International Journal (7).
- Chrissie Sanderson. Counselling Survivors of Domestic Abuse. Jessica Kingsley Publishers; 15 June 2008. ISBN 978-1-84642-811-1. p. 84.
- Skinner, B. F. (1938) The behavior of organisms. New York: Appleton-Century-Crofts.
- Chance, Paul. (2003) Learning and Behavior. 5th edition Toronto: Thomson-Wadsworth.
- Holth, P. (2005). Two Definitions of Punishment. The Behavior Analyst Today, 6(1), 43- 55 BAO .
- Chance, Paul. (2009) "Learning and Behavior: Active Learning Edition." 6th edition Belmont, CA: Wadsorth/Cengage Learning.