Play with me!
I recently gave a lecture on play - how to define it and how to teach it in children for whom play is not a natural occurring behaviour. Achieving a definition of play continues to challenge evolutionary scientists, ethologists and psychologists alike, and many would argue that such quest is futile. As Power (2000) notes: “Given its elusive nature, it is unlikely that researchers will ever come up with a satisfactory definition of play.” (p.391)
Is a definition needed? At the end of the day we all recognise play behaviour when we see it. When does defining play become important? In my line of work, the application of behavioural principles to evoke significant positive changes in a person’s behaviour, definitions are fundamental. Behavioural definitions provide not only a description of a phenomenon, but, in some cases, also an explanation – an analysis of controlling variables. Understanding controlling variables leads to effective environmental manipulations (i.e., teaching) so that the target behavioural repertoire can be established.
I asked my students to give me their own definition: “If you had to explain to an alien what play was, what would you say to him?” Most of the students began by talking about imitation, imagination and pretend play. I showed them videos of play in dogs, cats and monkeys to illustrate the point that play is not limited to humans. They eventually came up with something along the lines of “doing something for fun, for one’s own enjoyment, without instruction.” My job was now to guide them toward achieving a behavioural description of both form and function.
When defining play, it helps to think of it as a complex behavioural repertoire that has different topographies (e.g., solitary, imitative, imaginative, cooperative) but which share the same underlying source of reinforcement: “one’s own enjoyment” meaning that engaging in the behaviour itself is the reinforcer. One of my students pointed out that stereotypical behaviour also produces its own reinforcer and asked “how do we differentiate it?” “Play is varied” – another replied. (I smiled. Maybe they have been listening after all for the past year…) Variability is one of the distinctive features and reinforcing dimensions of play. My students are studying to be applied behaviour analysts, and they find my theoretical ramblings interesting only in as much as they are followed by a detailed explanation of procedures. The question that was on their mind was “How do you make play a reinforcer if the child doesn’t play?”
In my head I was hearing “How do I make play a reinforcer if my dog doesn’t play?”
Darwin used to love tugging until he began teething. Once he got his big boy’s teeth I assumed that his interest in tugging would resume. Unfortunately, this was not the case. I bought every possible tug toy out there, made of fabric, rope, felt, made of real and synthetic fur, crinkley squeaky, rubbery, with a long and a short bungee, with a tennis ball, with a hollow ball with a squeaky toy inside and without. I even got the local farmer to give me a cow chute that stank up the whole house. I tried to work on my technique, I animated the toy, made it move like prey, ran around the house with it for him to chase it, attached it to a flirt pole, worked on my body position, didn’t make eye-contact as to not put pressure on him. The result was always the same. He would pick it up half-heartedly and drop it immediately. I just never even got to the point that he would hold it long enough to let him win the toy, let alone the first tug.
Fetching or retrieving was a non-starter. Darwin simply did not read his breed-specific job description. He must have read a beagle job description instead, because his nose was constantly on the ground. At the park he might chase a stick I threw, only to lie down and enjoy the taste of it. Balls received the same treatment. Darwin seemed to enjoy only two kinds of play: wrestling with other dogs and having them chase him. In comparison to his play with a human, his play skills with members of his own species were well developed. With me, on the other hand, he only liked one kind of play: chase first, then anchor paws on hips and attempt to hump. That was one topography I was never going to reinforce.
Luckily, we had made some progress with food. In the concurrent schedule arrangement (play food games vs free food) he consistently chose the food games, which made me slightly more optimistic about the potential effectiveness of food to build novel responses. I must admit that teaching some basic instructions or begin platform work was very tempting, but I resisted and decided to use half the food I reserved for food games to focus on teaching how to play.
My experience of shaping animal behaviour is limited to two rats. The first time I visited Tim Hackenberg’s lab, he had kindly set aside two rats for me to practice shaping lever pressing, something many of his first-year psychology students achieve in their first session. I had a PhD in Behaviour Analysis, I taught postgraduate students advanced behavioural concepts, I have shaped pretty complex verbal behaviour in children with severe language delay. How hard would it be to shape a rat to lever press? The end of session criteria was 60 pellets or 30 minutes. Having run out of all the pellets in the first 10 minutes, the tummy of my first rat was so enormous she could hardly move, let alone press. With my second rat I made the shaping steps so big that she lost interest and fell asleep in the first 5 minutes. Not a good start to my animal shaping career.
I took it very slowly with Darwin. My ultimate goal was fetching the ball and dropping to hand. I backward chained it, shifting criterion very gradually and being prepared to go back to the previous step immediately if there was any hesitation. I set the trials at no more than 10 trials per 1-minute session, no more than twice a day. I did not write down the shaping steps. I decided to use an inductive process, letting him guide me, shifting criterion after 3 consecutive correct responses at the established step. Each next step received a bigger reinforcer, a kind of differential reinforcement by reinforcer magnitude . Returns to previous steps still got reinforced, with one lower value treat.
1) Turning to ball – mark - reinforce
2) Touching ball with nose – mark - reinforce
3) Teeth on ball – mark (my hand was below ready to catch the ball) – reinforce
4) Teeth on ball – drop to hand below – mark – reinforce
5) Teeth on ball – hand to the side – drop – mark – reinforce
6) Hand moved higher (I wanted his head up)
7) Moving one step with ball in mouth – drop to hand
8) Increasing distance very gradually
We got to this stage but I still wasn’t getting any sign of “enjoyment”. I wanted anticipation, speed, I wanted more momentum. Borrowing the term from physics, the term behavioural momentum describes the general relation between resistance to change and the rate of reinforcement in a given context. The High-Probability sequence is an antecedent strategy that relies on this process and it involves asking an individual to perform behaviours associated with a high rate of reinforcement first and then presenting the demand that has a low probability of occurrence. Essentially, the momentum built over the previous high-P responses carries over to the low-P one. This procedure is often used to address non-compliance, but it can also be used as a teaching strategy.
The High-P behaviour was chasing food and returning to me for the next bit. So I began with a few repetitions of that and then threw the ball as I would throw the food. The fact that the topography was very similar clearly helped. Over time I reduced the antecedent food chasing. Once the basic rules of the game were established, I brought in response variability by throwing the ball in different ways. I am currently still reinforcing with food on a FR1 schedule, but I expect to move it to a variable schedule soon. Over time it might become reinforcing in its own right and I might then be able to use it to shape other behaviours.
Play as a reinforcer
Having learned the basics of fetching a ball, Darwin quickly demonstrated generalization to a stick and we finally cracked tugging. Like the infant who likes to bang pots and pans rather than play with the expensive educational toys his mother dutifully bought, Darwin’s favourite tug toy is something that we have always had: a football. He began stealing my kids’ ball when they were playing. They would chase him and one day he stopped and offered it to them. When they grabbed it he pulled hard, and a tug of war game began. He now patiently waits for one of them to kick the ball so he can go and catch it and will then bring it back for a quick tug. We might just train him to be the goalkeeper.
In the end, the answer to my students was that sometimes we have to teach learners how to play before they can begin to enjoy it. Thanks to Darwin, I now know this to be as true for dog learners as human ones.