Posted: Apr 10, 2011 7:19 am
by Mr.Samsa

B. F. Skinner argued that the science of psychology is the study of behaviour and not the psyche or mind, and in doing so he challenged the lay understanding of human nature by asserting that hypothetical concepts like the mind cannot be used as explanations for behaviour as they themselves are behaviours that require causal explanations. So radical behaviourism is the philosophical approach that underpins the science of behaviour (Skinner, 1974) and it is an extension of methodological behaviourism which aimed to turn psychology into a natural science.

As this metascientific position claims that behaviour is caused by its relation to the environment through a method of selection by consequences, then the question of free will and determinism is necessarily raised. That is, if a science of human nature were possible and that science were able to predict and control our choices, then how truly free can our will be? The three main variations of determinism discussed in this essay are hard determinism, soft determinism and libertarianism – with free will being compatible only with the latter two categories. The importance of examining the implications that radical behaviourism has on free will is not merely an academic exercise, rather the question of its existence significantly impacts the way we should develop and control how society is run.

Radical Behaviourism

The notion that psychology should rely on objective measures, as opposed to the subjective approach of introspection, was outlined in John B. Watson’s “Psychology as the behaviorist views it” (Watson, 1913). It was here that he extended upon Darwin’s theory of evolution and argued that since there is the continuity of species then research from humans should be able to be applied to animals and vice versa, thus creating a general science of behaviour instead of the narrower anthropocentric model of pre-Watson psychology. He hoped that by rejecting hypothetical constructs such as references to the mind or consciousness (and avoiding anthropomorphism) then he could pioneer a truly objective approach to understanding behaviour.

Mazur (2002) argues that the integral point of contention between Watson’s methodological behaviourism and Skinner’s radical behaviourism is that whilst Watson was critical of using unobservable events as psychological data, Skinner was more concerned with the inappropriate use of unobservable events in psychological theories. In other words, although Skinner accepted the methodological advances made by Watson, he did not feel it was necessary to reject unobservable (“private”) events but simply acknowledged that these events were also behaviours that needed explaining.

These intervening variables, Skinner believed, were not necessary in order to understand behaviour. For example, take the situation where a rat is deprived of water and then placed in an operant chamber where it can work for water, then we would expect to find a relationship between the level of deprivation and the rate of lever pressing. Instead of claiming that the rate of lever pressing was a result of “thirst” that was caused by the hours of water deprivation, Skinner argued that the variable “thirst” adds nothing to our ability to predict the rat’s behaviour because our rule/equation works equally well without it (with the bonus of not needing to appeal to an unobservable event). The problem with hypothetical constructs is that we can easily deceive ourselves into thinking that we have uncovered the causal relation when in fact all we have done is create a plausible description of what preceded the behaviour and halted any inquiry into what the actual cause is – in effect, we have simply devised a “just-so” story (an ad hoc fallacy).

Gilbert Ryle described the use of intervening variables as an explanation for behaviour as a category mistake (Ryle, 1984). There are numerous category mistakes that one can make, such as mistakenly assigning an instance to the wrong category (for example, naming “carrot” as an example of “fruits”) – but the one that is pertinent to radical behaviourism is the use of a category label as an example of a specific instance of that category. Continuing with the example above, this would be like saying “vegetables” or, even worse, “fruits” as an example of “fruits”. “Vegetables” is clearly a category mistake as it is not an example of fruits as well as being a category label in itself, however, “fruits” is an even greater logical error as it is not only a category label but it is the very label of the group we are attempting to name instances of.

In terms of behaviours, this would be like changing the game to naming instances of intelligent behaviour. The game may begin well, with the players suggesting things such as chess playing, mathematical ability, literacy, and other various skills, but then suppose someone suggests “intelligence” as an example of intelligent behaviour. The mistake here is one of mentalism; that is, assuming that there is some underlying quality beneath the intelligent behaviours called “intelligence”, rather than understanding that “intelligence” is a label that describes these behaviours. This means that if someone attempted to understand a person’s chess playing behaviour by claiming it is caused by their intelligence, then this would be a category mistake – it adds nothing to our understanding and so cannot be used to make further predictions, and worst of all, it is circular as the intelligent behaviour is being “explained” by naming the category it belongs to.
In essence, the behaviourist position is that the study of human behaviour should progress in the same way all other sciences have; by moving away from speculation about possible internal states to describing observable events. We now realise that the Aristotelian approach to understanding the acceleration of falling objects being due to the increasing jubilance that the object feels as it nears home is no longer a valid inference and radical behaviourism argues that we should do the same in psychology.

Determinism and Free Will

The idea that there can be a science of behavior implies that behaviour, like any scientific subject matter, is orderly, can be explained, with the right knowledge can be predicted, and with the right means can be controlled. This is determinism, the notion that behavior is determined solely by hereditary and environment. (Baum, 2005, p.12)

The basic definition of determinism is presented by Baum above – if a behaviour can be predicted to the degree that it can be controlled by altering some of the variables, then that behaviour is described as being determined. Free will, in contrast, is the position that despite genetic inheritance and learning histories, a person who behaves one way could still have chosen to behave another way. Sappington (1990) categorises the determinism/free will debate into three positions; hard determinism, soft determinism and libertarianism. Hard determinism, the idea that human behaviour is controlled entirely by external factors outside of the person, is contrasted here with soft determinism, which argues that free will is not necessarily incompatible with deterministic factors as the person will always choose what they ‘want’ even if those desires are ultimately determined by external causes themselves. The soft determinism position avoids the free will issue by redefining the concept of freedom and differs from the traditional libertarianism definition of free will – that is, humans are exempt from natural laws as the choices they make are not determined by external factors and so people are viewed as active agents in the world.

Radical behaviourism is a hard determinist philosophy as Skinner argues that human behaviour is a result of our phylogeny (evolutionary history) and ontogeny (environmental history) and it does not include a variable that can accommodate free will (Skinner, 1966). Although it may seem presumptuous for Skinner to assume that our behaviour is the result of genetic and environmental factors, it is necessary not from a philosophical standpoint but rather from a scientific one. If we can describe and predict human choices without reference to a mind or free will, then such a notion would be extraneous to our understanding and it would be more parsimonious to exclude it until such time evidence arises in favour of free will or until it becomes necessary to include factors other than the external.

The soft determinist position is characterised by Daniel Dennett (1984) who defined free will as the deliberation that occurs before a behaviour. This form of free will is compatible with determinism because the deliberation itself forms part of the causality chain that extends backward in time. This is in conflict with the libertarianism position which claims that free will is separate from any causality and it seems to be a position that is unfalsifiable as any action, whether it is predictable or not, could be argued to have been a result of free will. A more intriguing argument for free will, however, is one that arises from chaos theory which basically states that even though complex physically systems may be composed of identifiably deterministic parts, the system as a whole may be inherently unpredictable in some aspects (Duke, 1994).

This approach essentially frames the concept of free will as ‘unpredictability’, which may superficially satisfy our definition of free will but Baum (2005) suggests that this unpredictability cannot be used as evidence of free will because there are many natural systems that we cannot perfectly predict but this does not mean that the weather, for example, has free will. The logical error, pointed out by Baum, is that although free will implies some degree of unpredictability, the converse is not equally true; that is, unpredictability does not imply free will.

Why is it important to consider the concept of free will when discussing the idea of radical behaviourism in society? The first is that for a science of human behaviour to be possible we need to be able to accept (conditionally, at least) that human behaviour is predictable and to be able to reject claims that human behaviour cannot be understood because of the unpredictable nature of free will. The second consideration is that in order to construct a society that runs efficiently, we must implement laws and practices that are compatible with the causal laws of human behaviour.

Responsibility and Justice

As discussed above, the idea of free will ultimately seems to be founded in a form of ignorance – that is, ignorance of the underlying contingencies. When we learn of a politician who has accepted a bribe, we are no longer confident that his actions are a result his own free will. Equally, if we are to learn that an artist had supportive parents and committed teacher then we are less likely to wonder his creative abilities came from. Skinner (1971) argued that the mentalistic concepts of credit and blame being awarded to certain actions, like the ones presented above, prevent us from solving important societal problems.

Baum (2005) notes that the concept of responsibility generally applies to the cause of events, for example “The bad wiring was responsible for the fire” is equivalent to saying “The bad wiring caused the fire”. So if we were to suggest that a person is responsible for the fire, then we are saying that this person caused the fire. The difference between these two explanations, however, is that we can readily accept the idea that the bad wiring was the result of environmental effects (such as weathering) but when we replace the wiring with a person, the causal chain seems to end there – that is, the person caused the fire because they wanted to. There seems to be a distinction made between objects and people which becomes apparent when we consider the same situation as above, but this time with the person being threatened at gun point to start the fire. We may feel compelled to say that this person no longer has a choice but the only thing that has changed is that the contingencies behind his actions are now obvious. In the original example, we may discover that the person had a traumatic childhood or that he is a pyromaniac – as we discover more and more about he history, the less likely we are to attribute his actions to his free will and the more likely we are to recognise the environmental factors that resulted in the fire.

So the notions of credit and blame are essentially just different ways of talking about causes, yet with the addition of approval or disapproval attached to each, respectively. When we are caught doing something we know that we should not be doing, we try to assign blame to an environmental factor but we tend to shy away from illuminating the environmental factors that resulted in us doing something good in order to have the credit attributed to our self. Baum (2005) suggests that the reason for this disparity is not difficult to uncover – if we assign blame to the environment we can avoid punishment, and if we resist credit being assigned to the environment then we can receive reinforcement in the form of social praise or even monetary rewards. Credit and blame, and ultimately responsibility as a whole, is just a subjective way of dividing up causes of our behaviours.

This concept seems, superficially at least, to be a fairly simple one to grasp so why is it that has not been adopted into societal practices already? Why do we still base our society around the notion of free agents? Staddon (2001) suggests that the reason for this is that if the criminal behaviour of a person is perfectly predictable, then punishment for such actions seems unjust as it is not possible for a person to behave any other way than his hereditary and environment dictates. This concept, however, is not as controversial as it may seem as defence attorneys will frequently suggest ‘mitigating factors’ that resulted in the defendant committing the crime. Staddon presents such a case where two brothers, accused of murdering their father, were defended in court through the use of testimony that suggested a history of child abuse (this trial ended in a hung jury, but in the next trial they were convicted and sentenced to life imprisonment).

The obvious question that needs to be asked here, however, is what do we mean by punishment? If by punishment we mean the moral condemnation of an action, then we might be inclined conclude that it is in fact unreasonable for us to judge a criminal as his actions were determined by genetic and environmental factors. This appears to be traditional usage of the term punishment and the judicial system seems to be based on the concept of retribution. From a behavioural perspective though, this is not what we mean by punishment. Baum (2005) argues that a greater recognition of extenuating circumstances will necessarily lead to a more practical system, one that is concerned with either changing the behaviour or, failing such an option, creating measures that ensure the safety of the public by removing the incorrigible criminal from society. (This will be discussed in more detail in the design of culture section).

Values and Morals

As an extension of responsibility and justice, we have to consider what impact the determinist position of radical behaviourism has on values and morals. Although most religions will argue that our morality and values are inherent to us and come from god, but as an explanation of human behaviour it falls short as it is unfalsifiable and offers no predictive power. Lewis (1960) asserted that science can only describe how we do behave but not how we should behave. C. S. Lewis starts with the assumption that god exists and from there argues that science cannot tell us what god finds good or bad, and this much is true insofar as it is impossible for science to understand the whims and preferences of an entity for which there is no solid evidence for. Science can, however, make judegements based on what people find good or bad by assuming moral relativism. Moral relativism is the antithesis of absolute morality as it contends that values differ across people, places, times and contexts. The standard argument against relativism is that if there is no absolute standard of what is right and wrong, how are we to decide who is correct when definitions of right and wrong conflict?

Baum (2005), among others, argues that social conventions dictate who is correct. Good and bad is decided by the group and this convention is then translated to the individual’s situation. The question then becomes, how does the group decide what is to be considered good and bad, right and wrong? The answer from Skinner (1971) was, in a nutshell, things that are considered good are positive reinforcers and things that are considered bad are punishers. Money is good, ill health is bad – the associations between the behaviours/events with the verbal behaviour are made as a function of the consequences they yield. This approach is not only plausible but it can also account for what Lewis (1960) terms “The law of human nature”. His argument begins by appealing to a situation that is common to all of us; when a conflict arises we tend to hear somebody make claims such as “That isn’t fair, I was here first”, or “How would you like it if someone treated you like that?” and so on. To this he says:

Now what interests me about all these remarks is that the man who makes them is not merely saying that the other man’s behaviour does not happen to please him. He is appealing to some kind of standard of behaviour which he expects the other man to know about. And the other man very seldom replies: “To hell with your standard.” Nearly always he tries to make out that what he has been doing does not really go against the standard, or that if it does there is some special excuse... It looks, in fact very much as if both parties had in mind some kind of Law or Rule of fair play or decent behaviour or morality or whatever you like to call it, about which they really agreed. And they have. (Lewis, 1960, p. 17)

Although Lewis takes this as evidence that there is an absolute morality, Skinner argues that this is simply the product of learning – when a child is playing with another child and shares his toys, then this behaviour is likely to be followed by various forms of reinforcement such as social praise from the parents as well as possible reciprocal reinforcement in the form of a toy being shared in exchange.

Design of culture

If we were to assume that radical behaviourism is the philosophy upon which society should be formed, then what exactly does that mean? Essentially, it is an extension of the scientific movement, the science of human behaviour – the idea that systems such as governments and prisons should be based on evidential claims and should be able to readily change under the light of new findings. The justice and penal system, as mentioned earlier, should not be based on mentalistic notions of retribution and “punishment”, instead it should be based on changing problem behaviours and creating a more functional society. For example, there are a number of experimental studies looking at the effectiveness of punishment that have not been utilised in societal institutions; partially due to the impracticality of some but also due to refusal to accept that human behaviour is predictable and controllable, to some degree at least.

In particular, we need to consider factors such as manner of introduction and immediacy of punishment implemented to assess whether they are the most effective methods we can use. Azrin and Holz (1966) suggest that if the goal is to obtain a large, permanent decrease in a problem behaviour then the punisher must be delivered immediately and at its full intensity to avoid habituation to successive mild punishers. Azrin, Holz, & Hake (1963) demonstrated that an immediate large shock of 80-volts following a pigeon’s keypeck was sufficient to result in a complete suppression of that behaviour, however, if that intensity was not used from the outset and the punishment began at lower intensities that gradually increased, then the pigeons would continue to respond even when the intensity was raised to 130 volts. This is an issue with the current system where infringements are usually punished with warnings or minor fines that increase as the number of infringements increase.

The immediacy of punishment is also a key factor in suppression behaviours. Baron, Kaufman, and Fazzini (1969) found an orderly relationship between the punishment delay and response rate when looking at rats responding on a Sidman avoidance task. The delay between a response and punishment was varied between 0 to 60 seconds and they found that the more immediate the punishment then the greater the decrease in responding. The reason for this is probably due to creating effective contiguity between the response and consequence – the less delayed the punishment, the less likely the consequence will be attributed to an intervening, but irrelevant, response. The implications this holds for things such as speeding fines should be readily apparent; the greater the delay in the act of speeding and receiving the ticket (then having to pay it) then the less effective the speeding ticket is in actually changing behaviour. Instead a much more immediate form of punishment is necessary – although possibly impractical, it may be prudent to install ticket devices in cars that will automatically print out a ticket the instant the speeding camera records an infringement.

Punishment is not the only tool of the behaviour analyst though, and reinforcement is also a powerful manipulator of behaviour. In education settings, for example, a study by Thomas, Presland, Grant and Glynn (1978) suggested that teachers tend to use very low levels of approval (approximately once every five minutes) and about three times as many disapproving statements. These findings lead us to question not only what the reinforcement contingencies are for the children in the classroom, but also what environmental factors are influencing the teachers’ behaviour. The Thomas et al. study argued that the attention given to off-task students acted both as a reinforcement for the child to continue being off-task, but also as a reinforcer for the teacher as their reprimand resulted in momentary on-task behaviour.


By understanding the deterministic nature of human behaviour and accepting that it can be predicted and controlled, we move closer toward building a more functional society – from restructuring the penal system to more effectively suppress problem behaviours, to creating a more productive learning environment that benefits both the student and the teacher. The philosophy of radical behaviourism is not necessarily about pushing operant techniques on society, but rather it is the notion that human behaviour should be understood in the same way we understand the rest of our world and through experimentation and science we can build on our knowledge to improve our way of life.


Azrin, N. H., & Holz, W. (1966). Punishment. In W. K. Honig, Operant behavior: Areas of research and application. Englewood Cliffs, NJ: Prentice-Hall.

Azrin, N. H., Holz, W., & Hake, D. F. (1963). Fixed-ratio punishment. Journal of the Experimental Analysis of Behavior , 6, 141-148.

Baron, A., Kaufman, A., & Fazzini, D. (1969). Density and delay of punishment of free-operant avoidance. Journal of the Experimental Analysis of Behavior , 12, 1029-1037.

Baum, W. M. (2005). Understanding Behaviorism: Behavior, Culture, and Evolution (2nd ed.). Oxford: Blackwell Publishing.

Duke, M. P. (1994). Chaos theory and psychology: Seven propositions. Genetic, Social, and General Psychology Monographs , 120, 267-286.

Lewis, C. S. (1960). Mere Christianity. New York: Macmillan.

Mazur, J. E. (2002). Learning and Behavior (5th ed.). New Jersey: Prentice Hall.

Ryle, G. (1984). The Concept of Mind. Chicago: University of Chicago Press.

Sappington, A. A. (1990). Recent Psychological Approaches to the Free Will Versus Determinism Issue. Psychological Bulletin , 108, 19-29.

Skinner, B. F. (1974). About Behaviorism. New York: Random House.

Skinner, B. F. (1971). Beyond Freedom and Dignity. Middlesex: Penguin Books Ltd.

Skinner, B. F. (1966, September 9). The Phylogeny and Ontogeny of Behavior. Science , pp. 1205-1213.

Staddon, J. E. (2001). The New Behaviorism: Mind, Mechanism, and Society. Philadelphia: Psychology Press.

Thomas, J., Presland, I., Grant, M., & Glynn, T. (1978). Natural rates of teacher approval and disapproval in grade 7 classrooms. Journal of Applied Behavior Analysis , 11, 91‐94.

Watson, J. B. (1913). Psychology as the behaviorist views it. Psychological Review , 20, 158-177.