PHILOSOPHER-SCIENTIST: Could you give me an example from human life?
TINNY: I'll give an example that will illustrate the chaining of responses, and where the stimuli perform discriminative and reinforcing functions at the same time. There would be many more responses in the chain that I could include so I'll simplify the links in the chain. Suppose a man was going for a drive in his car. First he walks up to the car, reaches out, and opens the door. He would not reach out to open the door until he was beside the car. The car door being within reach serves as a cue, a discriminative stimulus, that the response of reaching out is likely to be reinforced. The responses which open the door are rewarded, or reinforced, by the door opening. The open door serves as a cue to move into the car and sit down. The responses of moving into the car and sitting down are rewarded by being in the car, ready to drive. Sitting in the driver's seat serves as a cue for turning on the ignition. Turning on the ignition is reinforced by the car starting. The car starting serves as a cue to put the car in gear. Putting the car in gear is rewarded by the car being ready to move forward. The car being ready to move forward is a cue for pushing on the accelerator pedal. Pushing on the accelerator pedal is rewarded by the car moving forward. The car moving forward is a cue for turning the steering wheel. I won't go any further with this example, but you can see the sequential links in the chain. Each of the cues also serves as a reinforcer, and each reinforcer also serves as a cue, a discriminative stimulus.
PHILOSOPHER-SCIENTIST: I would never have thought of those various responses as occurring because they have been reinforced.
TINNY: It is undoubtedly through the conditioning process that every one of those responses came to occur. It sounds overly simplistic when put into words, but the reason people emit the responses which open doors is because behaving that way results in open doors. If door opening responses didn't open doors no one would perform those behaviours.
PHILOSOPHER-SCIENTIST: What you say is obviously true. It's surprising with such a direct relationship between that response and its resulting rewarding or punishing stimulus that the nature of the conditioning process goes largely unrealised.
TINNY: It is due to the incredible ignorance of learning principles throughout human society.
PHILOSOPHER-SCIENTIST: If such obvious conditioning relationships can exist unnoticed there would be little chance for the subtle conditioning relationships to be known.
TINNY: The average person wouldn't be aware of even one per cent of the conditioning which shapes their lives. What we have no awareness of, we have little control over. Because of our blindness to the learning process our destinies are governed mainly by unseen, or at least unrealised, influences. When our minds are opened by the knowledge of learning principles we shall determine our own destinies through the exercise of freewill. We shall guide our development through self-directed inner influence.
PHILOSOPHER-SCIENTIST: Is there much difference between discrimination training, differential reinforcement, and shaping?
TINNY: There is little difference between the three. If there is any real difference it would be in the degree of significance that the discriminative stimulus has in the conditioning process.
PHILOSOPHER-SCIENTIST: Would you give me an example of differential reinforcement that doesn't have an obvious discriminative stimulus?
TINNY: Would you prefer an example of animal learning or human learning?
PHILOSOPHER-SCIENTIST: Give me an example of differential reinforcement that occurs naturally during human development.
TINNY: A good example would be of an infant learning to pick up objects. When the infant, perhaps a baby girl, makes the first attempts in her life to pick up an object she wouldn't be very successful. She would see the object she desired to grasp, but as she reached for it her hand would wave around, almost aimlessly. All the arm movements could be considered attempts to grasp the object, but the holistic analysis has not yet learned to relate certain muscular movements to particular positions of the arm and hand. With a young infant we describe this lack of ability as being uncoordinated.
PHILOSOPHER-SCIENTIST: Doesn't physical co-ordination have something to do with how complete the development of the nervous system is?
TINNY: Many factors enter into all stages in our development. Within the limits of physical ability, conditioning determines the degree to which any characteristic is manifested. In the case of the baby in this example, she does have the physical ability to grasp and pick up an object; but, she has had virtually no training in the use of that ability. She would learn to use her natural ability through differential reinforcement. Of all the possible movements of the arm and hand only one particular set would be reinforced. The only responses being reinforced would be the ones that got the hand to the object. You could say those responses were differentially rewarded. Since no other movements of the arm and hand would get the hand to the object so it could be picked up, no other set of muscular movements would be followed by reinforcement. The sequence of movements that got the hand to the object would increase in the future, and those which did not get the hand to the object would decrease. It takes many reinforced reaching responses to become proficient at picking up objects. This process of differential reinforcement, when successful responses are rewarded and unsuccessful responses are not rewarded, is the way that co-ordination is learned.
PHILOSOPHER-SCIENTIST: What takes place in the holistic analysis during those attempts to pick up some object?
TINNY: Remember that the holistic analysis consists of perception, and comparison with previously stored information. A very young infant, no matter how well perception works, will not be able to interpret those perceptions very well and will have very little stored information to make comparisons with. For instance, the baby girl in our example may perceive some object that because of its nature, perhaps it is shiny, attracts her attention. There may be so little stored information available to the holistic analysis that it is not possible to know what the object is, or exactly how to get it. Without ever previously realising that objects can often be reached by arm and hand movements there would still have been past experiences when moving the arm around has brought the hand in contact with objects that have had some reinforcing value. This would have occurred to some degree even before birth. Holistic analysis wouldn't be certain there was a relationship between hand movements and being able to touch objects, but some hint of that relationship, even if unconscious, would exist. Holistic analysis may advise waving an arm around when the desire to pick up or touch some object occurs.
As this arm waving occurs there are more perceptions and interpretations taking place. The infant can see the spatial relationships between the hand movements and the desired object. Eventually the hand touches and grasps the object. At the moment that occurs the responses which brought the hand to the object are reinforced. Those responses become more likely to occur in the future. Holistic analysis now has more information to use in advising appropriate behaviour next time the infant desires to pick up some object. This additional information contains visual cues and memory of muscular positions. Each time the infant successfully picks up some object the holistic analysis becomes more able to advise appropriate behaviour in the future. As this conditioning process takes place through differential reinforcement the infant slowly changes from an uncoordinated baby to a coordinated young child.
PHILOSOPHER-SCIENTIST: So even our most basic behaviour is developed according to the laws of learning. In that example, about learning to grasp an object, could the visual perception of the relative positions of the hand and object be considered as discriminative stimuli that provide cues to which hand movements will be rewarded by coming in contact with the desired object?
TINNY: Those visual cues about relative hand and object position do function as discriminative stimuli. There is actually no conditioning which has no discriminative stimuli in the chain of responses. This was an example where the discriminative stimuli are not so obvious as the light being switched on or off during the example of differential conditioning with a rat.
PHILOSOPHER-SCIENTIST: Is the ability to interpret the information which comes from perception also learned?
TINNY: Interpretation is a learned response. We can, throughout our lives, learn to make increasingly accurate interpretations of the things we perceive.
PHILOSOPHER-SCIENTIST: Human learning seems to always be a gradual process.
TINNY: All learning is a gradual process. Even the classical conditioning we first discussed took place gradually. Learning takes place through a series of small steps. The effect of the conditioning which takes place at each step in the learning process changes the information holistic analysis has to work with next time a similar opportunity presents itself. The more holistic analysis knows about the circumstances the better the decisions which can be made as to appropriate behaviour in those circumstances.
PHILOSOPHER-SCIENTIST: What do you mean when you call certain behaviour appropriate?
TINNY: Appropriate behaviour is that which is in the best overall interest of the animal or person, given the circumstances.
PHILOSOPHER-SCIENTIST: To act in one's best interest in any situation would be a very important survival tactic.
TINNY: It is not surprising then that holistic analysis has been so highly developed and forms such a large portion of our genetic and cultural heritage.
PHILOSOPHER-SCIENTIST: When we were discussing classical conditioning you introduced a concept called extinction. Is there a similar process in operant conditioning?
TINNY: There is a similar process in operant conditioning. It is also called extinction. Extinction could almost be called unlearning. When a response rate has been increased by reinforcement and the reinforcing stimulus is no longer presented after that response, the rate of responding will drop back to roughly the level it occurred before reinforcement. We can go back to the example of the rat which was conditioned to press a lever at a high rate by following each lever press with food pellets. A rat placed in an experimental learning chamber for the first time will press the lever a few times each hour due to natural inquisitiveness. If each lever press raises the food hopper, thereby presenting the rat with a food pellet, the rate of lever pressing will soon increase to hundreds of times per hour. That high rate of lever pressing has been conditioned by positive reinforcement. If at this point the contingencies are changed so that lever pressing no longer results in the food hopper being raised then the response of lever pressing will begin to decrease. This decrease of response rate, which occurs when a response that had previously been followed by a reinforcing stimulus is no longer followed by a reinforcing stimulus, is called the extinction process. Just as the conditioning process is gradual, so is the extinction process gradual. When the extinction process is complete in this example the rat would resume pressing the lever at a rate of several presses each hour, which was its original natural level.
PHILOSOPHER-SCIENTIST: Do all responses have some natural level of occurrence which exists prior to any conditioning?
TINNY: All behaviours except perhaps certain reflex actions have been conditioned. When I say the rat's natural rate of lever pressing is several presses each hour, I mean that is the rate which has been conditioned by a rat's normal experiences during its early development. When I say those several lever presses each hour are due to inquisitiveness I am humanising the motivation behind a general behaviour that rats have developed to aid in survival. Exploring the environment and manipulating objects found in their environment is a general tactic which has proven its usefulness in providing for the rat's basic needs. The conditioning that took place during the experiment, when lever presses were reinforced, is building upon a base which had previously undergone conditioning. During extinction the high rate of lever pressing, which was conditioned by reinforcement during the experiment, is unlearned because it is no longer reinforced. All of that gain is lost, but the original few lever presses each hour remain because the conditions which reinforced that level of responding are still in effect.
PHILOSOPHER-SCIENTIST: The effect of conditioning on behaviour can be quite complicated.
TINNY: We are still discussing only the simplest instances of conditioning. The overall interactive relationship based on the laws of learning between each human being and their environment is complex beyond present comprehension. To fully understand these interconnections would be like trying to understand the totality of interconnections of material forms in the physical universe. For all practical purposes the number of interconnections, in both cases, can be considered to be infinite.
PHILOSOPHER-SCIENTIST: Wouldn't that place a limit on the degree to which we can control our own destiny by the exercise of freewill.
TINNY: It's true that since we can't perfectly know every effect and interconnection of conditioning in our lives we don't have perfect control over our own destiny. The better we understand the laws of learning, the greater will be our awareness of the effects and interconnections of conditioning in our lives. The greater our awareness of the effects and interconnections of conditioning in our lives, the more fully we can express our freewill and control our own destinies. As perfection is approached, complete understanding of the effects and interconnections of conditioning on our lives is approached. Perfected existence has absolute and unlimited knowledge of the innumerable effects and interconnections of conditioning and is therefore able to perfectly manifest freewill
PHILOSOPHER-SCIENTIST: The envolution of material form is a shaping process isn't it?
TINNY: That's correct, envolution takes place through shaping. From the first subatomic particles at the beginning of the physical universe, until the most highly envolved animal species, the contingencies which shape development are mainly externally determined. The human species represents a new phase in the envolution of material form. For the first time the potential exists for the contingencies which shape further development to be internally directed through self-reflective consciousness.
PHILOSOPHER-SCIENTIST: Tell me some more about the laws of learning.
TINNY: Another important concept in conditioning is called generalisation. Generalisation is related to the process of discrimination. I'll explain it by another example. I'll use the example of classical conditioning when a dog was conditioned to salivate to the sound of a bell. In that situation the bell was a conditioned stimulus which elicited salivation. For the purposes of this example we will say that each time the bell was sounded the dog produced fifty drops of saliva. The dog had been conditioned using a bell with a certain tone. Some bells have higher tones and some bells have lower tones than the one used in the conditioning procedure. If a bell with a slightly higher tone or a bell with a slightly lower tone is sounded the dog would still salivate, even though neither of those bell tones had been used in the conditioning procedure. The ability of stimuli similar, but not identical, to the stimulus used in the conditioning process to exert an influence on the response is called generalisation. The dog generalises from the stimulus it is most familiar with to other stimuli which are similar.
PHILOSOPHER-SCIENTIST: Would the similar stimuli have exactly the same effect as the stimulus originally used in the conditioning process?
TINNY: In the salivation experiment it was found that the further the bell tone varies, either higher or lower than the original bell, the less the dog would salivate. If the frequency of the tone varied, say, ten percent higher or lower, the dog might produce only forty drops of saliva. If the frequency of the tone varied by twenty per cent, the dog might only produce thirty drops of saliva. If the tone of the bell varied too far from the original to be identified as similar, no drops of saliva would be produced.
PHILOSOPHER-SCIENTIST: Is there a generalisation effect in operant conditioning also?
TINNY: There is. Generalisation in operant conditioning takes place in regards to the discriminative stimulus. The discriminative stimulus provides a cue to whether or not a response will be reinforced. In the example of discrimination training we discussed the rat was reinforced for pressing the lever when the light was on, and would not be reinforced for pressing the lever when the light was off. The light functioned as a discriminative stimulus. A tone would have served equally well as a discriminative stimulus in this learning situation. When the tone was on, pressing the lever would be rewarded; and, when the tone was off pressing the lever would not be rewarded. The rat having learned to discriminate between these two conditions would respond at perhaps five hundred lever presses an hour if the tone was sounding, and just several lever presses an hour when the tone was not sounding. If tones were sounded that were higher or lower than the tone involved in the original conditioning procedure the rat would still press the lever, but at rates less than five hundred presses an hour. The further from the original tone the new tone was, the fewer times the rat would press the lever. If the tone sounded was so different from the tone originally involved in the conditioning process that it could not be identified as similar then the rat would not be influenced by the tone at all, and would only press the lever several times each hour, just as it would do if there was no tone sounding.
PHILOSOPHER-SCIENTIST: What takes place in the holistic analysis which results in generalisation?
TINNY: Holistic analysis becomes aware of a certain stimulus which provides cues to appropriate future behaviour. After the cue has proven itself as an accurate predictor of what is to come, the holistic analysis accepts the certainty of the relationship of the cue to reinforcing or punishing events which will follow the response. When another stimulus appears, which is similar to the one that holistic analysis has become certain about, the common characteristics are recognised. Since the match in characteristics is not perfect, holistic analysis is somewhat less than certain that this similar stimulus is a useful cue to future events. Being less certain of the value of this new but similar stimulus, holistic analysis expresses those reservations by advising a reduced response. The further the new stimulus is from the original stimulus the greater the reservations expressed by the holistic analysis. If the characteristics of the new stimulus are too far removed from those of the original stimulus, the new stimulus will not be taken into consideration in the holistic analysis. In this case the new stimulus will be neutral in respect to the response and will have no influence on the response.
PHILOSOPHER-SCIENTIST: Could you give me an example of generalisation in human learning.
TINNY: Human examples are often most obvious with young children. When very young children are learning to talk they are sometimes shown different animals and told what each animal is called. For instance the child might be show a small dog and told it is called a puppy. Several times the child might point out other little dogs and say "puppy". The parents reinforce this behaviour by showing their approval. In this learning sequence a little dog is the discriminative stimulus, saying "puppy" is the response, and the parent's approval is the reinforcing stimulus. The child has been conditioned by positive reinforcement to call little dogs 'puppy'. Now the child sees another small animal, with fur, a tail, and running on four legs. Once again, the child says "puppy". This time though the small animal was actually a cat. The mistake made by the child was due to generalisation. There were enough similar characteristics between the original stimulus, the little dog, and the new stimulus, the cat, for the child to respond to the cat in a way which would have been the appropriate response upon seeing a dog. If the characteristics of the cat had not been so similar, it would be unlikely the child would have responded by saying "puppy". The child, upon seeing an elephant, is not likely to generalise and call it a puppy. Although the elephant shares some characteristics with a dog, there would not be enough to allow generalisation to take place. Young children have also been known to call the wrong adult "Mummy" or "Daddy". This happens because of generalisation also. The child, seeing enough similar characteristics, responds to an unrelated adult as it had learned to do when seeing its real parents.
PHILOSOPHER-SCIENTIST: During that explanation you said that a parent's approval was a reinforcer.
TINNY: We humans and some animals find approval, affection, friendliness, smiles, a kind word, or even attention to be desirable. They can all be positive stimuli and serve as reinforcers.
PHILOSOPHER-SCIENTIST: Does this class of reinforcing stimuli have a name?
TINNY: They are called social reinforcers. Social reinforcers are one of the most common and effective rewards. Most human learning involves some form of social reinforcement.
PHILOSOPHER-SCIENTIST: Social reinforcement seems like it would be a very convenient way to reward desirable behaviour.
TINNY: Social reinforcement is also a very nice way to reward desirable behaviour.
PHILOSOPHER-SCIENTIST: Couldn't social reinforcement also be used to reward undesirable behaviour?
TINNY: Unfortunately that's true. Much social reinforcement is used to develop behaviours which are wrong and harmful.
PHILOSOPHER-SCIENTIST: That will change.
TINNY: It has to if the human species is to survive.
PHILOSOPHER-SCIENTIST: Let's go on with our discussion on the laws of learning. What more can you tell me about operant conditioning?
TINNY: When I began explaining about operant conditioning I said there were four main types. They were positive reinforcement, negative reinforcement, presentation punishment, and withdrawal punishment. I want to discuss one of those four types further.
PHILOSOPHER-SCIENTIST: Which one?
TINNY: Negative reinforcement. Negative reinforcement increases the rate of a response which can terminate an ongoing aversive stimulus. If a response puts an end to something that is unpleasant, then that response is more likely to occur again in the future. Getting out of an unpleasant situation is a form of escape. Sometimes negative reinforcement is called escape training. There is another way to deal with an unpleasant situation. If the unpleasant situation has already begun then escape is the appropriate behaviour; but if the unpleasant situation hasn't yet begun then the appropriate behaviour is to avoid the forthcoming unpleasant situation. There is a type of reinforcement called avoidance. It can be thought of as a special instance of negative reinforcement. In avoidance conditioning the response postpones an imminent aversive stimulus. The response keeps something unpleasant from taking place. Any response which keeps something unpleasant from happening is more likely to occur again.
PHILOSOPHER-SCIENTIST: Would you give an example of avoidance conditioning?
TINNY: Once again I'll use the example of a rat in an experimental chamber. This time the food hopper won't be involved in the conditioning process. The floor of the chamber is divided into two sections by a low wall. The floor on each side is electrified and hooked up so that only one side is switched on at a time. When the rat is on one side and the electric shock begins on that side it jumps the fence to the other side where the electricity is not on. By jumping the wall to the other side the rat is able to escape the shock. The rat learns this escape response very quickly by negative reinforcement. The electric shock is the ongoing aversive stimulus and the response of jumping the wall terminates that aversive stimulus. The electricity is periodically switched on, always on the side of the chamber the rat is in. Now something new is added to this situation. A light is placed on the wall in each compartment. The light is turned on several seconds before the shock begins. This light, since it always precedes the electric shock, provides information which can be used to predict when the shock is coming. After only a few pairings between the light and the electric shock the rat learns to avoid being shocked by jumping over the wall as soon as the light comes on and before the electricity is switched on.
PHILOSOPHER-SCIENTIST: In a similar situation I would do the same thing; but, it is unclear to me why you say this is a special instance of negative reinforcement. In negative reinforcement the response puts an end to an ongoing aversive stimulus. In this case there is no ongoing aversive stimulus so, although the response avoids the shock, it doesn't stop an ongoing shock. Wouldn't this type of conditioning be something other than negative reinforcement?
TINNY: In this example we see how various conditioning processes can be interconnected. A sharp electric shock to the feet of a rat elicits certain reflexive behaviours. The rat would jump around and squeal. Using terms from our discussion on classical conditioning, the shock would be called an unconditioned stimulus. The jumping and squealing would be called an unconditioned response. The shock naturally elicits a jumping and squealing response. The light would, before conditioning, be a neutral stimulus to the rat; but, when it is turned on a few seconds before the shock, the light become a conditioned aversive stimulus. The light, by being paired with the shock, gains an aversive quality from that association. As we look at the example now, it can be seen that when the light goes on it is treated by the rat as an ongoing aversive stimulus. The rat's response of jumping over the wall to the other compartment terminates this ongoing conditioned aversive stimulus. The rat has learned to jump over the wall to escape the aversive association with the light. In this case the rat learned that behaviour by the process of negative reinforcement. The rat is now escaping the aversive associations of the light and, therefore, avoiding completely the electric shock.
PHILOSOPHER-SCIENTIST: When we discussed extinction in classical conditioning you said that when the conditioned stimulus is presented a number of times, no longer followed by the unconditioned stimulus, the conditioned stimulus would lose its influence over the response.
TINNY: That's true.
PHILOSOPHER-SCIENTIST: Then in this example if the rat always avoided the shock by jumping over the wall each time the light went on, the light would no longer be followed by the shock. Wouldn't the light lose the aversive quality it had gained by being paired with the shock?
TINNY: You might be right; it could begin to lose its influence as an conditioned aversive stimulus. When this happens the rat would probably not jump over the wall when the light came on. It would then once again receive an electric shock and the light would almost immediately regain its conditioned aversive character.
PHILOSOPHER-SCIENTIST: Wouldn't that cycle go on forever? The light would gain aversive qualities by being paired with the electric shock, and then lose those aversive qualities when it is no longer paired with the shock?
TINNY: That cycle wouldn't continue endlessly. Eventually the rat would become so powerfully conditioned both to consider the light an aversive stimulus and to jump the wall when the light came on, that extinction would probably no longer take place. In human terms the rat would become so sure that the light would usually be followed by an electric shock that it would never again take the risk of not jumping the wall to escape the light and avoid the forthcoming shock.
PHILOSOPHER-SCIENTIST: In that example, the rat learned to consider the light an aversive stimulus by classical conditioning, and learned to jump the wall by a form of operant conditioning called negative reinforcement. The light, which in the process of classical conditioning served as a conditioned aversive stimulus, also served as a discriminative stimulus in the process of negative reinforcement.
TINNY: Exactly right.
PHILOSOPHER-SCIENTIST: Could you give me an example of avoidance learning in human behaviour?
TINNY: Avoidance conditioning is very common in human development. A familiar example involves the behaviour of people crossing roads. When someone begins crossing a road then suddenly sees a fast approaching vehicle they would quickly jump back off the roadway. This behaviour has been learned by avoidance conditioning. In this case the fast approaching vehicle is the conditioned aversive stimulus and jumping out of the way is the response which terminates that ongoing aversive stimulus. And at the same time, the fast approaching vehicle serves as a discriminative stimulus in an avoidance learning process where it is a cue that the response of jumping out of the way would be appropriate be the appropriate behaviour.
PHILOSOPHER-SCIENTIST: But jumping out of the way of a fast approaching vehicle is just normal behaviour. How can you say it is conditioned?
TINNY: We are not born knowing that getting out of the way of approaching vehicles is appropriate behaviour. That is why little children have to be taught that running onto the roadway can be dangerous.
PHILOSOPHER-SCIENTIST: But it is obvious a fast approaching vehicle would cause great harm if we were hit. How can that knowledge be considered to have been conditioned?
TINNY: It is only obvious that a fast approaching vehicle is dangerous because we have so successfully learned that truth.
PHILOSOPHER-SCIENTIST: Although little children might not have learned that it can be dangerous to run out in the road, most if they saw a vehicle rushing at them would try to get out of the way.
TINNY: Even if what you say is true, that doesn't prove that the aversive character of an onrushing vehicle has not been conditioned to be seen as dangerous.
PHILOSOPHER-SCIENTIST: If a person had never seen an onrushing vehicle before they would still jump out of the way if approached by one. How then could either the aversive quality of the approaching vehicle or the response of jumping out of the way have been conditioned? There was no past experience of either.
TINNY: I still claim the approaching vehicle gained its aversive quality by conditioning; and, that the response of jumping out of the way was learned by conditioning.
PHILOSOPHER-SCIENTIST: Would you explain how?
TINNY: We live in a world of objects, many of them moving objects. That is the nature of the physical universe. The objects of material existence relate to each other in certain ways. Human bodies during their development experience relationships of many sorts with many different objects. One of these classes of relationships is with moving objects. Since we can't experience every possible relationship with every different object we come to general conclusions about the nature of these interactions. The general conclusions we arrive at are determined by the particular experiences we undergo. We learn collisions with hard objects usually hurt more than with soft objects. We learn collisions with big objects often hurt more than with small objects. And we learn that collisions with fast moving objects often hurt more than with slow moving objects. It would be fair to say that a collision with a big, hard, fast moving object would virtually always cause more harm than a collision with a small, soft, slow moving object. Since this is true, big, hard, fast moving objects become conditioned as aversive stimuli. We do not inherit this knowledge, we learn it by various experiences during our interactions with the environment. That is the reason there would be few people in the world, regardless of whether or not they had ever previously seen a fast approaching vehicle, who would not recognise the potential danger of that oncoming vehicle and would act appropriately by jumping out of the way.
PHILOSOPHER-SCIENTIST: Would that be due to generalisation?
TINNY: It would. The onrushing vehicle would exhibit a sufficient number of similar characteristics to the varied objects in the past which had played a part in conditioning big, hard, fast moving objects as aversive stimuli for holistic analysis to be treat the fast approaching vehicle as an aversive stimulus.
PHILOSOPHER-SCIENTIST: I understand. It seems no matter how natural and obvious our behaviour patterns appear to be, that in truth, they have all been affected by various conditioning processes.
TINNY: Conditioning is a most ubiquitous process.
PHILOSOPHER-SCIENTIST: What does that mean?
TINNY: The laws of learning are always present everywhere. There is no time and no place where the laws of learning are not in effect.
PHILOSOPHER-SCIENTIST: Getting back to the rat in the example of avoidance conditioning. Wouldn't that rat be quite upset being in a situation where it is constantly in danger of receiving a painful electric shock?
TINNY: It would be a terrible experience. The rat would certainly undergo emotional stress.
PHILOSOPHER-SCIENTIST: Would suffering from emotional stress always be part of escape and avoidance conditioning procedures?
TINNY: There would always be some undesirable emotional side effects. Sometimes the emotional difficulties could go virtually unnoticed, but in other circumstances the emotional problems could be devastating.
PHILOSOPHER-SCIENTIST: Is there any name for these various expressions of emotional stress?
TINNY: There is a word which quite accurately describes that emotionally distressed state. The word is anxiety. Anxiety could be defined as emotional distress or uneasiness of mind, caused by the apprehension of danger or misfortune.
PHILOSOPHER-SCIENTIST: I suppose that would describe the experience the rat underwent.
TINNY: Well, that definition described anxiety in human terms; but although less developed, all the same mental and emotional states exist at the animal level. The higher the consciousness of the animal the closer to the human experience.
PHILOSOPHER-SCIENTIST: Do rats have very highly developed consciousness?
TINNY: Yes. Among those in the animal kingdom rats have developed a quite high level of consciousness.
PHILOSOPHER-SCIENTIST: Rats are not very well liked animals.
TINNY: They have their place in the ecology of the world.
PHILOSOPHER-SCIENTIST: Is anxiety always a bad thing?
TINNY: Interestingly, like many things, anxiety can be beneficial in moderation. A little bit of anxiety stimulates particularly efficient performance. Performance with a small amount of anxiety usually shows an improvement over performance with no anxiety at all.
PHILOSOPHER-SCIENTIST: What effect does a high degree of anxiety have?
TINNY: With a large amount of anxiety, performance drops below the level of performance with no anxiety. While a small amount of anxiety seems to stimulate behaviour, a large amount of anxiety seems to inhibit behaviour. The level of anxiety can become so high that all behaviour comes to a stop.
PHILOSOPHER-SCIENTIST: That might be the origin of the saying, "frozen with fear".
TINNY: Almost surely.
PHILOSOPHER-SCIENTIST: Do you really think anxiety is a good thing if only in small doses?
TINNY: I said a small amount of anxiety can be beneficial and will stimulate behaviour. From the relative point of view, this would be considered good. From the absolute perspective anxiety is never good, since the influence of anxiety is through the imposition of will by force. Anxiety always involves the influence of some form of negative control. It would be better if the improved performance came about from the influence of positive factors.
PHILOSOPHER-SCIENTIST: What is the most harmful aspect from the use of conditioning procedures using negative influence?
TINNY: While there are many harmful effects from the use of negative means to condition behaviour, there is one which out shadows all others. Every time conditioning by negative means is successful it reinforces the response of using negative means. Each successful use of negative control increases the use of negative control in the future.
PHILOSOPHER-SCIENTIST: When you refer to negative means of conditioning, does that include both negative reinforcement and presentation punishment?
TINNY: Yes, those two forms of conditioning are harmful because of the use negative stimuli to influence behaviour. I would also class withdrawal punishment as a negative means of conditioning, although this form involves a positive stimulus.
PHILOSOPHER-SCIENTIST: Why would you include a form of conditioning, withdrawal punishment, which involves the use of a positive stimulus as a negative means of conditioning?
TINNY: Because of the effect on the emotional state of the subject being conditioned. No one is happy to have something they want taken from them.
PHILOSOPHER-SCIENTIST: If negative reinforcement and presentation punishment are harmful because they both involve the use of a negative stimulus, and withdrawal punishment is harmful because it also results in an unhappy emotional state, there is only one of the four basic forms of operant conditioning left.
TINNY: That's right, positive reinforcement is the only form of conditioning which has no inherent harmful side effects. It would be best if all conditioning took place through positive reinforcement.
PHILOSOPHER-SCIENTIST: Would it be possible for all conditioning to take place through positive reinforcement?
TINNY: Not at present; but, it would be possible for a much greater portion of all conditioning to take place through positive reinforcement.
PHILOSOPHER-SCIENTIST: You said the most harmful aspect of conditioning using negative control was that it reinforces the use of negative control. Would you give me an example of that happening?
TINNY: Sometimes parents attempt to control their child's behaviour by the use of physical punishment. If the child does something the parent thinks is wrong the parent will hit the child. The parent follows the child's wrong behaviour with a negative stimulus. This is a form of presentation punishment. The parent uses this form of negative control to reduce or eliminate some behaviour which they consider to be wrong. If hitting the child is successful in stopping that behaviour, the parent is reinforced for hitting the child. The parent's behaviour is being conditioned by negative reinforcement. The ongoing aversive stimulus is the child's wrong behaviour, and the response that terminates the ongoing aversive stimulus is hitting the child. Each time the parents succeed in stopping behaviour they consider wrong by hitting the child the probability increases that the parents will in the future hit the child to stop wrong behaviour. Successfully using physical punishment increases the use of physical punishment. In this example the parents are not the only ones being conditioned to use hitting as a means to attain desired ends. The child also learns hitting is an appropriate way to get what you want.
PHILOSOPHER-SCIENTIST: How does the child learn to use negative control by that experience?
TINNY: That involves another concept in learning called modeling that I would rather explain later.
PHILOSOPHER-SCIENTIST: Whenever you think best. Would you give me a different example of how using negative control perpetuates the use of negative control?
TINNY: In some countries the government keeps people in a deprived and oppressed condition. The government maintains this control over the people by negative means. Some of the oppressed people ask in a reasonable way to be granted their basic human rights. The government ignores this quiet and reasonable request. The response of asking quietly and reasonably is not reinforced. Therefore, asking quietly and reasonably is not likely to occur again in the future. Some of the oppressed people now take up arms and fight violently against the government. The government cannot continue to ignore the oppressed people. The violent struggle becomes widespread, becomes a revolution, and the oppressive government is overthrown. The control is now in the hands of the people. The response of violent struggle has been powerfully reinforced. Violent struggle is likely to be looked on favourably and is also likely to reoccur in the future. Revolution through violent struggle will be promoted as the best way, perhaps the only way, to attain positive social change.
PHILOSOPHER-SCIENTIST: It isn't though, is it?
TINNY: Revolution through violent means may bring about positive social change; but, this social change will contain a fatal flaw.
PHILOSOPHER-SCIENTIST: What is that flaw?
TINNY: The seeds of negative control and violent action are spread throughout the new government and among the people. This seed will grow and overwhelm whatever positive social change had been obtained. This fervour for revolution through violent action could destroy the government, destroy the people, and even destroy the planet.
PHILOSOPHER-SCIENTIST: From some of your examples I feel I'm not only learning about the laws of learning, but I'm also learning about social influence.
TINNY: The laws of learning are the basis of social influence. Society as a whole sets up reinforcement and punishment contingencies which control individual behaviour.
PHILOSOPHER-SCIENTIST: Would you give me an example of this form of social conditioning?
TINNY: Through much of human history living, speaking, or even thinking which varies from the social norm has not been considered acceptable. Adherence to the prevailing social norm was required of every member of the society. Strict adherence to the current religious beliefs and standards was a common area subject to social control.
PHILOSOPHER-SCIENTIST: Which of the four basic types of operant conditioning does society use to insure its members don't vary from the accepted norms?
TINNY: Positive and negative reinforcement, presentation and withdrawal punishment are all used to condition socially acceptable behaviour. I'll briefly present an example of each of those four types of conditioning as I describe how strict adherence to a certain system of religious belief could be ensured by social control. By positive reinforcement those actions which are in accord with religious standards are increased. Those whose response is to speak or act most appropriately, as determined by the religious beliefs, are given special social recognition, respect, and power in the community. By negative reinforcement those actions in accord with religious standards are also increased. Those who are seen as being in conflict with religious standards are subject to constant disapproval. This disapproval is expressed in many ways. Social disapproval can be a strong ongoing aversive stimulus. When a member of society who has been subject to disapproval begins speaking and acting in accord with religious standards the disapproval vanishes, the dissenter is rewarded by being welcomed back into the community. By presentation punishment those actions in opposition to religious standards are decreased. Those whose words and actions conflict with religious beliefs may come to verbal abuse, physical harm, or even death. And, with withdrawal punishment those actions in opposition to religious standards are also decreased. Those whose words and actions oppose religious beliefs may find their jobs, social standing, or material goods taken from them.
PHILOSOPHER-SCIENTIST: In a closed society all those conditions and sanctions must have been extremely effective in controlling the religious beliefs of the people.
TINNY: It is a process which has been successfully used for thousands of years.
PHILOSOPHER-SCIENTIST: Are the use of those controls to insure adherence to certain religious beliefs linked to any particular religion?
TINNY: Followers of all religions have used variations of those four types of conditioning to control their members. Some emphasise the use of one type of influence more than others, but all have been commonly used.
PHILOSOPHER-SCIENTIST: Do followers of religions today still attempt to control their members by those same conditioning methods?
TINNY: In some parts of the world those same methods are still very much in use. In most of the world today, though, even if religions still attempt to use those methods, the nature of modern society makes those methods much less compelling.
PHILOSOPHER-SCIENTIST: Does any other institution in society rely on those same methods to control the masses?
TINNY: Governments have relied heavily on all those same conditioning methods to maintain power and control. Sometimes it is hard to distinguish between political ideologies and religions.
PHILOSOPHER-SCIENTIST: None of those social influences seem like they provide the right reasons to accept any certain religion or support any government.
TINNY: They don't. Religions should attract followers only because they are good and true. Political ideologies should also attract supporters only because they are good and true.
PHILOSOPHER-SCIENTIST: What comes next in the logical progression of the knowledge of the laws of learning?
TINNY: Next I want to explain schedules of reinforcement.
PHILOSOPHER-SCIENTIST: What are schedules of reinforcement?
TINNY: When I described operant conditioning I said it involved a response that is followed by some stimulus which, depending on whether it is a positive or negative stimulus and whether it is presented or removed, will increase or decrease the probability of that response occurring in the future. Schedules of reinforcement explain the different possible relationships between the response and the stimulus; and, although called schedules of reinforcement, these laws pertain not only to conditioning through reinforcement but also to conditioning through punishment. The different relationships which can exist between the response and the following stimulus usually involve either the number of responses which must be emitted before being followed by the reinforcing or punishing stimulus, or the amount of time which must pass before a response can be followed by a reinforcing or punishing stimulus. In the examples of reinforcement and punishment we have already discussed, every response was followed by a reinforcing or punishing stimulus. That relationship between the response and the following stimulus is one particular schedule of reinforcement. When every response is followed by the reinforcing or punishing stimulus the relationship is called continuous reinforcement. Often however a response is not reinforced or punished every time it occurs. The response may only be reinforced occasionally. This relationship is called intermittent reinforcement. In one type of intermittent reinforcement the response is reinforced only after a certain number of responses have occurred. Reinforcement of this type is called a ratio schedule, because there is some ratio between the number of responses required and each reinforcement which is presented. In another type of intermittent reinforcement the response is reinforced only after a certain amount of time has passed before the response occurs. Reinforcement of this type is called an interval schedule, because there is some interval of time between reinforced responses.
PHILOSOPHER-SCIENTIST: So schedules of reinforcement can be divided into those which are continuous, when reinforcement follows every response, and those which are intermittent, when reinforcement only follows certain responses.
TINNY: And the two intermittent schedules of reinforcement can also be divided into two classes.
PHILOSOPHER-SCIENTIST: Those two types of intermittent reinforcement are ratio schedules, when reinforcement only comes after some number of responses, and interval schedules, when some amount of time must pass between reinforced responses.
TINNY: That's right. There is one more major division in the schedules of reinforcement. Both the ration schedules and the interval schedules can be either fixed or variable.
PHILOSOPHER-SCIENTIST: What does it mean for a schedule of reinforcement to be fixed or variable?
TINNY: In the case of ratio schedules, a fixed ratio refers to reinforcement occurring always after the same number of responses; and, variable ratio refers to reinforcement occurring after different numbers of responses, which vary around some average.
PHILOSOPHER-SCIENTIST: How about with interval schedules?
TINNY: In the case of interval schedules, fixed interval refers to reinforcement always following the first response after the same period of time; and, variable interval refers to reinforcement following the first response after different periods of time, which vary around some average.
PHILOSOPHER-SCIENTIST: That explanation needs to be made more clear.
TINNY: I'll give an example of each of those schedules of reinforcement.
PHILOSOPHER-SCIENTIST: I'm sure that will help me to understand schedules of reinforcement better.
TINNY: To make these examples as simple as possible I'll explain each of the different schedules of reinforcement using the example of the rat in the experimental chamber. These examples will refer to conditioning through reward rather than punishment; but, remember that the same schedules apply to conditioning through punishment. We have already discussed continuous reinforcement. In continuous reinforcement every time the rat presses the lever it receives a food pellet as a reward.
PHILOSOPHER-SCIENTIST: That seems simple enough.
TINNY: I'll explain the ratio schedules first. The ratio schedule can be either fixed or variable. In a fixed-ratio schedule there is some fixed number of responses which must occur before reinforcement is given. For example every fifth response is reinforced, but no others are. In this case when the rat is pressing the bar, the first, second, third, and fourth responses are not followed by reinforcement. When the fifth response occurs, the rat is rewarded with a food pellet. Every subsequent fifth response is also reinforced. That would be called a fixed-ratio (five) schedule since every fifth response is reinforced.
PHILOSOPHER-SCIENTIST: Does it matter how high or low the fixed number of responses is?
TINNY: That fixed number could range from one to thousands.
PHILOSOPHER-SCIENTIST: If the schedule was fixed-ratio (one), would that mean every response would be reinforced?
TINNY: That's what a fixed-ration (one) schedule means.
PHILOSOPHER-SCIENTIST: But that's the same as continuous reinforcement.
TINNY: Continuous reinforcement must then be a special case of a fixed-ratio schedule. It's usually considered a separate schedule, though, because it is not intermittent.
PHILOSOPHER-SCIENTIST: And how about the high fixed-ratio schedules. It's hard to imagine a rat pressing the lever a thousand times for one food pellet.
TINNY: It's not common, but it's possible.
PHILOSOPHER-SCIENTIST: How would you get the rat to press the lever a thousand times for just one reinforcement?
TINNY: By using the shaping procedure. First you would start with a continuous reinforcement schedule. After the rat learns to press the lever for food the number of responses needed to receive a food pellet can be slowly increased.
PHILOSOPHER-SCIENTIST: Must the schedule be increased slowly?
TINNY: The shaping process is a series of steps toward some goal. If any step is made too large the progressive series may break down. The steps must be kept small enough to make sure each step can be successfully taken. They are the links in a chain. If one of the links is broken the chain falls to pieces.
PHILOSOPHER-SCIENTIST: What determines how large each step may be?
TINNY: In this example the critical factor, which determines how large a step may be taken, involves the extinction process. The increase in the number of responses required over the previous level must not be so great as to allow extinction of the response to take place before the next reinforcement is given. After the rat had been conditioned to press the lever on a continuous reinforcement schedule you wouldn't make the next fixed-ratio schedule require one hundred responses to obtain reinforcement. A fixed-ratio (100) schedule would probably be far too large a step. The rat having learned to press the lever to obtain food during continuous reinforcement would realise that pressing the lever was no longer being followed by the reinforcing stimulus. This realisation would take place due to the ongoing holistic analysis. During continuous reinforcement each response, when followed by the presentation of a food pellet, confirms the relationship whereby lever pressing is known as a sure way to obtain food. If the schedule was suddenly changed from continuous reinforcement to a fixed-ratio (100) schedule, there would be ninety-nine responses in a row which were not followed by a reinforcing stimulus. Each of those non-reinforced responses would, from the perspective of holistic analysis, be cues that there was no longer a relationship between lever pressing and obtaining food. Holistic analysis would advise the lever pressing response be greatly reduced, as pressing at any high rate would be an inefficient use of energy.
PHILOSOPHER-SCIENTIST: But there would still be a reinforcement schedule in effect. The rat could still receive reinforcement for responding.
TINNY: The experimenter may know that, but the rat doesn't. The most logical analysis given the information the rat has to go on would be to conclude there was no reinforcement available for lever pressing.
PHILOSOPHER-SCIENTIST: It seems that any response conditioned by continuous reinforcement would be quickly extinguished.
TINNY: As we will see later, each of these schedules of reinforcement has some weakness, but they also have their strengths.
PHILOSOPHER-SCIENTIST: We can leave the strengths and weakness of the other schedules until later; but would you tell me now what the strength of continuous reinforcement is?
TINNY: Well, as you pointed out, the weakness of continuous reinforcement is that it can so easily be extinguished. The reason that continuous reinforcement can be so easily extinguished is because when a reinforcer is expected after each response, it becomes very obvious when that reinforcer does not appear. The strength of continuous reinforcement comes from the same factor as does the weakness. Continuous reinforcement's strength is that learning takes place so quickly. The reason that learning takes place so quickly during continuous reinforcement is because it is most obvious that the response brings reinforcement when every response is reinforced.
PHILOSOPHER-SCIENTIST: It sounds like the strength and weakness of continuous reinforcement is due to the great amount of information available about the relationship between the response and stimulus in the conditioning process.
TINNY: That's the case. In continuous reinforcement schedules every response provides useful information.
